Société en commandite Gaz Métro
Cause tarifaire 2010, R-3690-2009
Appendix G:
Evidence of Dr. Kolbe and Dr. Booth in Previous Proceedings
Original : 2009.07.27
Gaz Métro - 7, Document 17
(356 pages)
NATIONAL ENERGY BOARD
IN THE MATTER OF the National Energy Board Act and the
Regulations made thereunder;
AND IN THE MATTER OF an Application by Trans Québec &
Maritimes Pipeline Inc. for orders pursuant to Part I and Part IV of the
National Energy Board Act.
APPENDICES TO WRITTEN REPLY EVIDENCE
OF
A. LAWRENCE KOLBE
FOR
TRANS QUÉBEC & MARITIMES PIPELINE INC.
The Brattle Group
44 Brattle Street
Cambridge, MA 02138
September 2008
2
APPENDICES TO WRITTEN REPLY EVIDENCE OF
A. LAWRENCE KOLBE
Table of Contents
GUIDE TO LOCATIONS OF POINTS ALREADY ADDRESSED IN PREVIOUS PROCEEDINGS . . . . . . 4
Appendix R-Booth1: Booth Evidence in RH-2-2004, Phase II . . . . . . . . . . . . . . . . . . . . . . . . . . 5
Appendix R-Kolbe1: Kolbe Reply Evidence in RH-2-2004, Phase II . . . . . . . . . . . . . . . . . . . 122
Appendix R-Booth2: Booth Evidence in EB-2005-0520 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 165
Appendix R-Kolbe2: Kolbe Reply Evidence in EB-2005-0520 . . . . . . . . . . . . . . . . . . . . . . . . 295
Note: Page numbers within the Booth Evidence or the Kolbe Reply Evidence within each Appendix
are those from the original documents, which correspond to the page numbers in the Guide on page
4. The above page numbers refer to cover pages in the present document that match the page
numbers within this PDF file.
3
APPENDICES TO WRITTEN REPLY EVIDENCE OF
A. LAWRENCE KOLBE
GUIDE TO LOCATIONS OF POINTS ALREADY ADDRESSED IN PREVIOUS PROCEEDINGS
TransCanada Mainline
(RH2-2004, PHASE II)
Topic of Kolbe Reply Evidence
1) ATWACC A TOOL, NOT AN
OUTCOME
Written Reply
of
Evidence of
Dr. Kolbe
Dr. Booth (Appendix D)
Union Gas
(EB-2005-0520)
TQM
Evidence of
Dr. Booth
Written
Reply of Dr.
Kolbe
Evidence of
Dr. Booth
(Appendix B)
pp. 9-10
pp. 4, 28
pp. 3-4, 64,
70
pp. 2-4
pp. ii, 79, 104
pp. 62-76, 8790
pp. 4-12
pp. 70-81, 83- pp. 11-12,
89, 96
Appendix R-A
3) INADEQUATE REVIEW OF CAPITAL
STRUCTURE LITERATURE
pp. 81-87
pp. 13-16
pp. 89-95
pp. 12-16
pp. 20-26
4) RELIANCE ON SELECTED
REGULATORY DECISIONS RATHER
THAN CAPITAL STRUCTURE
LITERATURE
pp. 3,74-76,
78-80
pp. 17-19
pp. ii, 2, 82,
88, 97, 100,
104
pp. 16-17
pp. 1,15,20,28
pp. 83-85
pp. 19-20
pp. 93-94
pp. 18-19
p. 25
6) INCORRECT CLAIM THAT MY
EQUATION FOR THE INTERACTION
BETWEEN THE COST OF EQUITY
AND CAPITAL STRUCTURE
PRODUCES THE HIGHEST POSSIBLE
CHANGES
pp. 85-86, 89
pp. 21-22
pp. 93-94
pp. 19-20
pp. 18, 24, 25,
27
7) INCORRECT IMPLICATION THAT MY
PROCEDURES IGNORE NON-TAX
COSTS TO DEBT
pp. 16-17
pp. 22-24
p. 23
pp. 21-22
p. 25
8) CHARACTERIZATION OF THE WAY
REGULATION WORKS
INCONSISTENT WITH THE EVIDENCE
p. 79
pp. 24-26
p. 87
pp. 22-24
pp. 11-13
9) FINANCIAL RISK DEPENDENT ON
MARKET VALUES, NOT BOOK
VALUES
pp. 15-17, 81
pp. 26-30
pp. 21-23,
87-89
pp. 24-30
p. 19
10) MY PROCEDURES TO CALCULATE
APPROPRIATE DEEMED EQUITY
RATIOS ARE REASONABLE
pp. 89-91
pp. 30-31
pp. 98-100
pp. 30-31
p. 28
2) FLAWED NUMERICAL EXAMPLE
5) INCORRECT CLAIM THAT MY
EVIDENCE RELIES ON THE 1977
MILLER MODEL
4
pp. 3-7, 8-14,
16, 20-23
APPENDICES TO WRITTEN REPLY EVIDENCE OF
A. LAWRENCE KOLBE
APPENDIX R-BOOTH1: BOOTH EVIDENCE IN RH-2-2004, PHASE II
(Page numbers in Booth Evidence are from the original document)
5
BUSINESS RISK AND CAPITAL STUCTURE FOR THE
TRANSCANADA MAINLINE
Evidence of
Laurence D. Booth
BEFORE THE
National Energy Board
October 19 2004
1
2
3
TABLE OF CONTENTS
4
5
6
EXECUTIVE SUMMARY .......................................................................................................................2
7
1.0
INTRODUCTION ..........................................................................................................................5
8
2.0
REGULATORY TOOLS................................................................................................................7
9
3.0
BUSINESS RISK .........................................................................................................................19
10
4.0
FINANCIAL RISK.......................................................................................................................39
11
5.0
LEVERAGE ADJUSTMENTS ....................................................................................................77
12
1
EXECUTIVE SUMMARY
2
The Canadian Association of Petroleum Producers (CAPP) has asked me to provide an
3
independent assessment of the appropriate common equity ratio for the TransCanada Mainline
4
(the Mainline), to assess its business risk, and to discuss whether “leverage adjustments” are
5
needed to the firm’s common equity ratio to ensure that the Mainline’s rates are fair and
6
reasonable. This latter issue involves a discussion of the relevance of the weighted average cost
7
of capital (WACC) and whether the appropriate common equity ratio can be considered
8
separately from the allowed ROE, which is not itself a formal a part of this hearing.
9
My overall assessment is as follows:
10
•
The short term business risk of the Mainline is very low. The Mainline continues
11
to earn its allowed ROE with the same precision as previously and there is no
12
indication that the impact of the Board’s policy of allowing deferral accounts and
13
a forward test year has exposed the Mainline’s shareholder to any increase in risk.
14
•
There has been a marginal increase in the longer term “supply” and “competitive”
15
risks since the 1994 multi-pipeline hearing as the Western Canadian Sedimentary
16
Basin (WCSB) has matured and Alliance has become operational. However, the
17
exposure of the Mainline to this maturing of the WCSB has been offset by a
18
significant increase in the allowed depreciation rate, so that the capital at risk has
19
not been affected. Overall I see no change in the business risk of the Mainline
20
since RH-4-2001.
21
•
In RH-4-2001 Dr. Berkowitz and I recommended a 30% common equity ratio, up
22
from the 28% recommended by us in RH-2-94. This increased common equity
23
ratio recommendation in part reflected the retirement of the Mainline’s preferred
24
shares and their replacement with junior subordinated debentures prior to the RH-
25
4-2001 hearing. Nothing has changed in this regard since then, when the Board
26
correctly classified the JSDs as debt. I continue to view the JSDs as debt and do
27
not think that it appropriate to impute a 30% common equity component to them.
2
1
•
I would judge the Mainline to have a good investment grade bond rating with its
2
current allowed ROE and common equity ratio. Over the last two years there has
3
been concern expressed over “increasing” credit standards motivated by the
4
introduction into Canada of S&P’s US credit standards and their experience with
5
US utilities and pipelines like Enron and Aquilla. However, there is no indication
6
that the Canadian capital markets have reflected these US concerns. Spreads on
7
utility and pipeline debt over the last five years have reflected normal cyclical
8
concerns and do not indicate that the market has re-evaluated the regulatory
9
protection accorded utility and pipeline debt.
10
•
Overall conditions in the bond market indicate that spreads are tighter now than
11
they were at the time of RH-4-2001, so that utilities can access debt markets more
12
easily. Further the dramatic increase in the income trust market over the last five
13
years has opened up another source of financing. In my judgment the Mainline
14
has just as much, if not more, financial flexibility than at the time of RH-4-2001
15
and there is no need to make capital structure changes to improve its access to
16
capital, particularly since the rate base is declining.
17
•
My own judgement is that the Mainline’s currently allowed ROE formula,
18
combined with a 33% common equity ratio, is generous. However, like the Board
19
I believe that common equity ratios should only be revisited after a fundamental
20
shift in business risk and I see no signs of such a shift in the last three years. I
21
therefore would recommend that the Board continue with its current allowed
22
common equity ratio.
23
•
In terms of the ATWACC approach advocated and used implicitly in the
24
company’s filed testimony I would point out the fundamental contradiction in its
25
use in regulatory filings in that it is the mirror image of shareholder value
26
maximisation. That is, earning more than the WACC is synonymous with the
27
creation of shareholder value, whereas the Board’s responsibility is not to create
28
or maintain shareholder value, but to ensure that rates are fair and reasonable. The
29
Alberta EUB felt it would be “derelict” in its responsibilities to recognise market
3
1
capitalisation ratios, an assessment I agree with. In my judgment setting the equity
2
ratio or ROE implicitly by awarding a regulated firm its (AT)WACC partially
3
“rubberstamps” existing market values that may in turn reflect unfair and
4
unreasonable rates. I therefore see no value to introducing ATWACC into a
5
regulatory setting, either explicitly or implicitly.
6
•
Leverage adjustments should be made when Boards set both the allowed ROE and
7
the common equity ratio, as is done for example by the BCUC. In this way the
8
Board makes sure that it does not “double count” the impact of changes in
9
business risk. However, given the Board’s practise of allowing a common ROE
10
formula and setting differential common equity ratios to adjust for differences in
11
business risk, I see no point to making further leverage adjustments.
12
•
Overall, I would recommend that the Board ignore this back door ATWACC
13
approach and simply continue with its current practise of setting common equity
14
ratios based on changes in business risk and an assessment of the financial
15
flexibility and capital market access of the regulated firm in question.
4
1
1.0
INTRODUCTION
2
Q.
PLEASE DESCRIBE YOUR QUALIFICATIONS AND EXPERIENCE.
3
A.
Laurence Booth is a professor of finance and finance area co-ordinator in the Rotman
4
School of Management at the University of Toronto, where he holds the CIT Chair in Structured
5
Finance. Professor Booth (with the late Professor M. K. Berkowitz1) has previously filed
6
testimony with the Board in the multi-pipeline hearing, RH-2-94, and the Mainline’s last full cost
7
of capital hearing, RH-4-2001. In both cases a detailed resume was filed with the Board. If
8
needed, Professor Booth’s current CV can be downloaded from his web site.2
9
Q.
PLEASE DISCUSS HOW YOUR TESTIMONY IS ORGANISED AND THE
ISSUES THAT YOU DEAL WITH.
10
11
A.
12
parts of the Mainline’s pre-filed evidence and to recommend a financial structure consistent with
13
the company’s business risk and the requirement that rates be fair and reasonable. My
14
understanding is that the allowed ROE is not formally an issue in this hearing, as this is dealt
15
with according to the Board’s ROE formula. Consequently, fair ROE testimony is not required.
16
However, in its pre-filed testimony, the Company has filed rate of return testimony through the
17
evidence of Dr. Vilbert. Dr’ Kolbe has justified this on the basis that what is important is the
18
market determined weighted average cost of capital (WACC or what has been referred to before
19
this Board as the ATWACC). Dr. Kolbe by assuming an appropriate WACC for the Mainline
20
and that this WACC is constant, and invariant to changes in the debt ratio, is then able to base his
21
common equity ratio on Dr. Vilbert’s rate of return evidence.
22
In response I have organised my testimony as follows. First, I will review my understanding of
23
Board policy and how it relates to the issues at hand. Second, I will discuss the business risk of
24
the Mainline. In doing this I will take a capital markets perspective, Dr Safire and Mr Johnson
1
2
The Canadian Association of Petroleum Producers (CAPP) has asked me to comment on
Professor Berkowitz died August 8, 2004.
http://www.rotman.utoronto.ca/~booth.
5
1
will be independently filing traditional business risk evidence and a risk positioning matrix. I
2
have consulted with them on the most important issues and my judgment has been informed by
3
these discussions as well as by conversation with members of CAPP. Third, I will discuss typical
4
regulated financial structures in Canada, credit concerns, capital market access and provide my
5
recommended common equity ratio. Fourth, I will discuss the issues raised by the company’s
6
pre-filed testimony of the importance of WACC; why it occupies such a prominent position in
7
the finance literature and how, if at all, it relates to the Board’s responsibilities for setting fair
8
and reasonable rates. Finally, I will discuss the procedures that have been suggested for going
9
from the WACC to an implied equity cost with a fixed equity ratio, or an implied equity ratio
10
from a fixed equity cost. This latter section is critical for the company’s pre-filed testimony, but
11
as I will show I do not recommend that the Board place any weight on these techniques as they
12
are unreliable and do not offer any value added over the traditional, well accepted, techniques
13
used by the Board.
6
1
2.0
2
Q
WHAT RISKS DO INVESTORS FACE?
3
A.
Investors are interested in the rate of return on the market value of their investment. This
4
value can be represented by the standard discounted cash flow model:
5
REGULATORY TOOLS
P0 =
ROE * BVPS * (1 − b)
( K − g)
(1)
6
where P0 is the stock price, ROE the return on equity, BVPS the book value per share, b the
7
retention rate (how much of the firm’s earnings are ploughed back in investment). The product of
8
the ROE, BVPS and payout rate determine the dividend per share, which is then assumed to grow
9
at the rate g, which determines the future cash flow stream. This is then discounted back at the
10
investors cost of equity, or required rate of return, K.
11
The simple discounted cash flow (DCF) model is useful for thinking of the sources of risk to the
12
investor and the tools that the Board has available to it in managing that risk. Some of these risks
13
stem from the firm's operations and financing, while others stem from the capital market's
14
perception of the firm as well as general capital market conditions. For rate of return regulated
15
utilities we add another dimension to risk, which is the impact of regulatory risk. In terms of the
16
DCF equation the actual earned return on equity (ROE) captures the business, financial and
17
regulatory risk, and together I term these income risk, whereas all the other factors are reflected
18
in investment risk, which is the way in which investors react to this income risk and other factors
19
such as the firm's growth prospects and exposure to interest rates.
20
Business risk is the risk that originates from the firm’s underlying “real” operations. These risks
21
are the typical risks stemming from uncertainty in the demand for the firm’s product resulting,
22
for example, from changes in the economy, the actions of competitors, and the possibility of
23
product obsolescence. This demand uncertainty is compounded by the method of production
24
used by the firm and the uncertainty in the firm’s cost structure, caused, for example, by
25
uncertain input costs, like those for labour or critical raw or semi-manufactured materials.
26
Business risk, to a greater or lesser degree, is borne by all the investors in the firm. In terms of
27
the firm's income statement, business risk is the risk involved in the firm's earnings before
7
1
interest and taxes (EBIT). It is the EBIT, which is available to pay the claims that arise from all
2
the invested capital of the firm, that is, the preferred and common equity, the long term debt, and
3
any short term debt, such as debt currently due, bank debt and commercial paper.
4
If the firm has no debt or preferred shares, the common stock holders “own” the EBIT, after
5
payment of corporate taxes, which is the firm’s net income. This amount divided by the funds
6
committed by the equity holders (shareholder’s equity) is defined to be the firm's return on
7
invested capital or ROI, and reflects the firm's operating performance, independent of financing
8
effects. For 100% equity financed firms, this ROI is also their return on equity (ROE), since by
9
definition the entire invested capital has been provided by the equity holders. The uncertainty
10
attached to the ROI therefore reflects all the risks prior to the effects of the firm’s financing and
11
is commonly used to measure the business risk of the firm.
12
As the firm reduces the amount of equity financing and replaces it with debt or preferred shares,
13
two effects are at work: first the earnings to the common stock holder are reduced as interest and
14
preferred dividends are deducted from EBIT and, second the reduced earnings are spread over a
15
smaller investment. The result of these two effects is called financial leverage. The basic
16
equation3 is as follows:
ROE = ROI + [ ROI − Rd (1 − T )]
17
D
S
(2)
18
where D, and S are the amounts of debt, and equity invested respectively, T is the corporate tax
19
rate and Rd is the embedded debt cost. If the firm has no debt financing (D/S =0), the return to
20
the common stockholders (ROE) is the same as the return on investment (ROI). In this case, the
21
equity holders are only exposed to business risk. As the debt equity ratio increases, the spread
22
between what the firm earns and its borrowing costs is magnified. This magnification is called
23
financial leverage and measures the financial risk of the firm.
3
Note this equation captures how the actual ROE varies with operating profits. It does not show how the
investors required rate of return varies with financial leverage. This requires a valuation model to
understand how debt tax shields for example affect value.
8
1
The common stockholders in valuing the firm are concerned about the total “income” risk they
2
have to bear, which is the variability in the ROE. This reflects both the underlying business risk
3
as well as the added financial risk. If the firm operates in a highly risky business, the normal
4
advice is to primarily finance with equity. Otherwise the imposition of fixed financial charges by
5
the firm on top of the uncertainty in the firm’s EBIT might force the firm into serious financial
6
problems. Conversely, if there is very little business risk, as is the case with most regulated
7
utilities, the firm can afford to carry large amounts of debt financing, since there is very little risk
8
to magnify in the first place.
9
In this fundamental sense business risk and financial risk work in opposite directions. Firms in
10
industries with very high business risk tend to finance primarily with equity, while firms with
11
very low business risk tend to finance with more debt. The best examples of the latter are the
12
banks and regulated utilities. Before going on it is important to note that the financial leverage
13
equation deals with the firm’s financial statements, that is, its balance sheet and income
14
statement to determine income risk. In the case of a regulated utility this is what needs the
15
approval of the Board. However, how investors react to the firm’s income risk in determining
16
market prices, that is, investment risk, is beyond the direct control of either the firm or the
17
regulator. As a result, even if the income risk is removed completely, and the firm’s net income
18
is similar to the interest income on a long Canada bond, there is still investment risk, since the
19
market price will still fluctuate due to these other risks.
20
Q.
HOW CAN A REGULATOR RESPOND TO THE RISKS FACED BY THE
REGULATED FIRM?
21
22
A.
23
jurisdiction. In BC Electric Railway Co Ltd., vs the Public Utilities Commission of BC et al
24
([1960] S.C.R. 837), the Supreme Court of Canada had to interpret the following statute:
25
Regulators respond according to the particulars of the legislation that establish their
(a)
rate:
26
27
28
The Commission shall consider all matters which it deems proper as affecting the
(b)
The Commission shall have due regard, among other things, to the protection of
the public interest from rates that are excessive as being more than a fair and
9
1
reasonable charge for services of the nature and quality furnished by the public
2
utility; and to giving to the public utility a fair and reasonable return upon the
3
appraised value of the property of the public utility used, or prudently and
4
reasonably acquired, to enable the public utility to furnish the service:
5
This statute articulated the "fair and reasonable" standard in terms of rates, and that the
6
regulatory body should consider all matters that determine whether or not the resulting charges
7
are "fair and reasonable." Most rate of return regulation is based on similar provisions, for
8
example, section 62 of the National Energy Board Act requires that “all tolls should be just and
9
reasonable.”
10
Most recently NEB Board member Quarshie4 in a presentation indicated the Board’s response to
11
the Federal Government’s “smart regulation” “enabling and protection” initiative which
12
13
14
15
“involves using the regulatory system to generate social and environmental benefits while
enhancing the conditions for a competitive and innovative economy that will attract
investment and skilled workers and sustain a high quality of life for Canadians. It is about
making regulation as effective as possible.”
16
Board member Quarshie went on to say that “enabling implies a responsibility to ensure that
17
projects in the public interest can proceed,” while protecting “has an economic component that
18
is, protecting shipper interests, while being fair to investors.”
19
These provisions and discussion of recent initiatives have one meaning to an economist: that
20
rates reflect the operation of the regulated utility at minimum long run average cost. Costs in a
21
competitive market naturally gravitate towards minimum long run average cost and by definition
22
do not include charges that are unfair and unreasonable (or unjust), while this cost ensures that
23
the regulated services are provided at a cost that promotes the overall efficiency of the economic
24
system. Further the BC. statute articulated the “prudently and reasonably acquired” test in terms
25
of rate base assets, which would also seem to be assets acquired in the public interest.
4
Presentation to PCRI 9 August 2004.
10
1
Of the costs included in the Mainline’s rates approximately 77% are fixed financial charges
2
comprising 22.8% return of capital (depreciation), 10.6% income taxes and 43.2% return to debt
3
holders and equity holders.5 Clearly “fair and reasonable” rates have to ensure that the total
4
dollars of return, ie., the $796.2mm included in the interim application for return are “fair and
5
reasonable.” This means that both the amount of various forms of capital, as well as their
6
allowed returns, are “fair and reasonable.”
7
In the BC Electric decision the Supreme Court of Canada adopted Mr. Justice Lamont's
8
definition of a fair rate of return as enunciated in the Northwestern Utilities Limited v. City of
9
Edmonton ([1929] S.C.R. 186) decision that:
"By a fair return is meant that the company will be allowed as large a return on
the capital invested in its enterprise (which will be net to the company) as it
would receive if it were investing the same amount in other securities possessing
an attractiveness stability and certainty to that of the company's enterprise."
10
11
12
13
14
Mr. Justice Lamont's definition embodies what a financial economist would call a risk-adjusted
15
rate of return or "opportunity cost."
16
The Board has accepted this requirement by linking the allowed ROE to conditions in the long
17
Canada bond market through its adjustment mechanism. In RH-2-94 the Board decided that
18
allowed ROEs should adjust by 75% of the change in the long Canada bond yield. Implicitly the
19
Board acknowledged that the risk premium demanded by investors for pipeline investments was
20
not constant and varied inversely with long Canada bond yields. The Board thus took into
21
account both conditions in the long Canada bond market and their implication for low risk,
22
interest sensitive, investments like utility shares. In RH-4-2001 the Board reaffirmed its
23
adjustment mechanism and stated (page 54 reasons for Decision)
“in the Board’s view this provides a strong confirmation that the ROEs resulting from the
RH-2-94 Formula represent a reasonable estimate of the cost of equity capital for the
Mainline.”
24
25
26
27
The allowed ROE was therefore set at 9.53% for 2002.
5
Based on interim toll application for 2004 Schedule 2.1 and revenue requirement net of transportation by others.
11
1
In RH-4-2001 my colleague, Dr. Berkowitz, and I recommended that the Board continue its
2
formula but rebase it based on a recommended ROE of 8.50%, so it was our judgment that the
3
allowed ROE formula was generous. Below is a table summarising the main differences between
4
our testimony at that time and the situation today.
(December) 2001
5
6
Long term Canada bond yield
(September)
5.31
5.10
2004
7
Consensus
5.95
5.506
8
Dr. Booth
6.00
5.50
Real Canada Yield
3.60
2.30
10
Beta estimates
0.42-0.60
0.45-0.557
11
Pipeline risk premium
250bp
250bp
12
Inflation forecast
2.0%
1.90%
9
13
14
Our RH-4-2001 testimony was prepared and filed in the wake of the 9/11 terrorist attack on New
15
York. At that time, there was a definite slowing in the US economy and in response the US
16
Federal Reserve Board dramatically dropped short term interest rates. Three years later the US
17
economy still has not fully recovered and the state of financial markets is remarkably similar.
18
Current nominal long Canada bond yields are 21 basis points lower, while the yield on the
19
inflation index bond has dropped more precipitously by 1.30% to 2.30%. However, despite this
20
widening of the gap between the real and nominal yield the short run inflation forecast is
21
essentially the same at just under 2.0% indicating more uncertainty about the long run inflation
22
forecast. Apart from the fact that equity markets are stronger and we are marginally further along
23
the business cycle, there is nothing that would indicate any significant changes in financial
24
markets that would warrant a change in Board policy.
6
48 basis point spread over the consensus ten year forecast of 4.8% for August 2004 and 5.3% for May
2005 based on May Consensus.
7
Range used in the Alberta generic hearing (September 2003)
12
1
Further confirmation for this is the recent decision of the Ontario Energy Board (RP-2002-0158,
2
paragraph 142) confirming its adjustment mechanism, which gives allowed ROEs very similar to
3
those of the Board’s formula
“Therefore, with respect to the first and primary issue of whether a new benchmark ROE
should be established for EGDI and Union, we find that the current ROE Guidelines
methodology continues to produce appropriate prospective results. We have not found
any demonstrated need to set a new benchmark ROE.”
4
5
6
7
8
9
What is of particular importance in the OEB hearing is that neither Union Gas nor Enbridge Gas
10
Distribution (Consumers) put forward any testimony indicating that their business risk had
11
increased; or that their allowed common equity ratio should be increased. There was almost a
12
complete absence of any discussion of supply coming from the WCSB or elsewhere or any other
13
fundamental characteristic that might affect an LDC operating a long way from a supply basin.
14
In fact I have testified in Union and Consumers Gas rate hearings for many years and cannot
15
remember the last time they asked for an increase in the common equity ratio.8
16
Also the recent EUB generic hearing resulted in similar conclusions. After receiving evidence
17
from six sets of expert witnesses including Drs. Kolbe and Vilbert and myself, the Board
18
concluded (Decision 2004-052, page 21)
19
20
21
“Based on the above-determined risk-free rate of 5.68%, MRP of 5.50%, beta of 0.55,
and allowance for flotation costs of 0.50%, the Board concludes that a reasonable CAPM
estimate for 2004 is 9.20%.”
22
The EUB then considered a variety of other factors. It considered that other equity risk premium
23
studies, pension fund data and other Board allowed ROEs, including the 9.56% allowed by this
24
Board, argued for higher returns, while observations from market to book ratios and income trust
25
data argued for lower returns. The Board felt that limited or no weight should be applied to
26
results from DCF tests, US data, FERC incentive ROEs, or the allowed ROEs for Alliance and
27
MN &P. The EUB seemed especially critical of alternative investment opportunities available to
28
the parent holding company, stating
8
Union’s common equity ratio was changed when it merged with Centra Gas Ontario.
13
1
2
3
“The Board concludes that there is no basis on which to place any weight, other than
already reflected in earlier tests, on other specific investment opportunities potentially
available to utility investors or on stated expectations of return from such opportunities.”
4
I agree with this judgment for theoretical, as well as practical reasons, which I develop later. In
5
conclusion the Board went on (page30)
6
7
8
9
“In consideration of the impact of the above factors, it is the judgment of the Board that it would
be appropriate to establish the 2004 ROE at a level that is 40 basis points above the Board’s
CAPM estimate. Therefore, the Board concludes the generic ROE for 2004 should be set at
9.60%.”
10
However, what is important to note is that the EUB’s estimate of the investor’s required rate of
11
return was 8.70%, their CAPM estimate before adding 0.50% for flotation costs, and before
12
taking into account other factors. This is the investors discount rate applied to valuing a typical
13
utility’s shares. Overall it is my judgment that application of the Board formula generates an
14
allowed ROE that continues to be generous and is above the cost of equity capital for Mainline
15
transmission assets.
16
Q.
WHY HAVE YOU DISCUSSED THE ROE WHEN IT IS NOT AN ISSUE IN THIS
HEARING?
17
18
A.
Dr. Vilbert on behalf of the Mainline has filed 285 pages of rate of return testimony. I
19
have the same concerns with this testimony as I had when similar evidence was filed in RH-4-
20
2001. Further this evidence is used to support the estimates of Dr. Kolbe and Vilbert as to the
21
appropriate ATWACC or WACC for the Mainline. The main new twist is that instead of
22
recommending an allowed ROE and equity ratio based on this testimony, with a fixed allowed
23
ROE the full impact is now felt on the deemed common equity ratio recommendation alone.
24
However, the testimony of Drs. Kolbe and Vilbert remains firmly rooted in ROE testimony.
25
Consequently all the estimation problems that Dr. Berkowitz and I had with this testimony in
26
RH-4-2001 continue today, particularly since some of Dr. Vilbert’s beta estimates are based on
27
estimates as of May 2000 and do not seem to have been updated since RH-4-2001 anyway.
28
Consequently, the reason that I have provided this very brief background on capital market
29
conditions and recent Board decisions is simply to point out that relatively little has changed
30
since the time of RH-4-2001. Further, were I to provide ROE estimates for the Mainline today,
14
1
my judgment would be that the cost of capital estimates used by Drs. Kolbe and Vilbert, as the
2
base for their deemed common equity ratio recommendation, continue to be significantly higher
3
than those I would arrive at.
4
Q.
IF YOU DO NOT AGREE WITH THE APPROACH OF DRS. KOLBE AND
5
VILBERT, HOW DO YOU RECOMMEND THE BOARD SET COMMON
6
EQUITY RATIOS?
7
A.
I would recommend that the Board continue to use its existing policy. In RH-2-94 the
8
Board stated (Decision page 24)
9
10
11
12
13
14
“The Board is of the view that the determination of a pipeline’s capital structure starts
with an analysis of its business risk. This approach takes root in financial theory and has
been supported by the expert witnesses in this hearing. Other factors such as financing
requirements, the pipeline’s size and its ability to access various financial markets are
also given some weight in order to portray, as accurately as possible, a complete picture
of the risks facing a pipeline ”
15
I agree 100% with this assessment, since it follows the prior discussion of the impact of financial
16
leverage. To repeat the previous financial leverage equation
17
ROE = ROI + [ ROI − Rd (1 − T )]
D
S
(2)
18
If this equation is rearranged we can express the variability of the ROE as a function of the
19
variability in the operating income or
STDEV ( ROE ) = STDEV ( ROI ) * (1 +
20
D
)
S
(3)
21
where the standard deviation of the actual ROE is that on the ROI times one plus the debt equity
22
ratio. So if the Board wants to equalise the risk to equity holders (STDEV(ROE)) investing in
23
different pipelines with different business risk ((STDEV(ROI)) in principle it can alter the
24
deemed debt equity ratio.
25
At this point it is important to point out that the above equation is based on the firm’s financial
26
statements. It is an accounting relationship that has nothing to do with how the stock market
15
1
reacts to the firm’s use of financial leverage. As far as I know no-one has ever disputed the
2
above equations, as they are simply a rearrangement of the flow of income through a firm’s
3
financial statements. That is, the ROE is not that required by investors (the cost of equity capital)
4
it is simply the actual ROE earned by the firm on the book value of its equity. Using these
5
relationships is consistent with the fact that the Board can only control these accounting values.
6
The Board can alter business risk through the use of deferral accounts and the financial risk
7
through changes in the deemed equity ratio, but it can not change stock market risk, as the
8
market, not the Board, determines market values.
9
This last point should be emphasised: the financial leverage equation is not equivalent to the
10
formulae used by Dr. Kolbe. Dr. Kolbe’s equity cost adjustment formulae are based on
11
assumptions about how the stock market values the use of financial leverage. All of the work of
12
Drs. Kolbe and Vilbert is based on market values and their equations are based on assumptions
13
about how the stock market values the effects of financial leverage. Unlike the financial leverage
14
equation which indicates how the Board can alter financial risk to offset business risk, the
15
equations used by Dr. Kolbe attempt to answer the question of how the rate of return required by
16
an investor changes as the financial leverage based on market values changes. This adjustment
17
requires a theory of how the market values financial leverage, which is not required for the
18
Board to change deemed equity ratios in response to changes in business risk.
19
To illustrate in RH-2-94 several experts submitted testimony on how the allowed ROE should
20
change as the capital structure changes along the lines of the current testimony of Drs. Kolbe and
21
Vilbert. Dr. Sherwin and Ms. McShane, who provided testimony on behalf of the companies,
22
concluded (page 24)
23
24
“The finance models, even when adapted to the real world of Canadian utility regulation,
cannot provide the basis for determining a pipeline’s optimal capital structure.”
25
More importantly Dr. Berkowitz and I used models similar to those used by Dr. Kolbe, but
26
expressed little support for them. As the Board noted in its Reasons for Decision (page 24)
27
28
29
“Dr. Booth and Berkowitz concluded that these estimates are approximately the increases
in ROE required by investors. However, they noted the estimates are subject to error
since they are based on valuation formulas, which are as yet unproven. Moreover, they
16
1
2
noted that these formulas ignored the non-tax advantages of debt financing and the
effects of financial distress.”
3
Finally, the Board also noted Dr. Waters’ testimony (a frequent witness before the Board at that
4
time) where he indicated that “To date empirical testing to more clearly describe the relationship
5
(between capital structure and the investors required return) has not been done successfully.
6
The Board’s summary from ten years ago is an accurate assessment of my views today and it is
7
my judgment that the misgivings expressed by experts ten years ago continue, since the issues
8
have still not been resolved. I would therefore recommend that the Board continue its practise of
9
making capital structure changes based on its qualitative assessment of a pipeline’s business risk.
10
Q.
IS
BUSINESS
RISK
THE
ONLY
FACTOR
IN
SETTING
CAPITAL
STRUCTURES?
11
12
A.
13
pipeline can access capital on reasonable terms. If, for example, the Board has not sufficiently
14
increased the common equity ratio in response to an increase in business risk then the stock
15
market will discount the pipeline’s stock price and make it difficult for the regulated firm to
16
access capital on reasonable terms. In Federal Power Commission et al v. Hope Natural Gas Co.
17
[320 US 591, 1944], the United States Supreme Court decided that a fair return
18
19
No. Ultimately the litmus test of whether the Board has “got it right” is whether the
"should be sufficient to assure confidence in the financial integrity of the
enterprise so as to maintain its credit and to attract capital."
20
Although the Hope “financial integrity” criteria flows from considering a fair return it applies
21
equally to the deemed common equity ratio. In my judgment an appropriate common equity ratio
22
is one which, in conjunction with the allowed return, allows a pipeline to maintain its credit and
23
attract capital.
24
The Hope criterion would therefore support the view that after examining business risk, the
25
Board consider factors such as size, financing requirements and market access, since all of these
26
are important for financial integrity. However, note that “maintaining credit” is not the same as
27
maintaining a particular credit rating. Credit standards constantly change as does the market’s
28
appetite for certain types of credits. This means that there is no need to target a particular credit
17
1
rating. What is important is that a pipeline can access the capital markets with conventional
2
financial securities when it needs to raise capital to provide service.
3
Q.
IS THERE ANY OTHER RECENT DECISION THAT SUPPORTS THE
BOARD’S VIEW?
4
5
A.
Yes, the recent Alberta Generic Hearing followed in the wake of the Board’s policy
6
expressed in RH-2-94 and established not just an adjustment formula to set the allowed ROE, but
7
also the allowed common equity ratios for eleven distinct regulated entities in a range of ROE
8
regulated businesses including pipelines (ATCO Pipe and NGTL). The EUB stated (Generic
9
Cost of Capital Decision page 35)
10
11
12
“To determine the appropriate equity ratio for each Applicant, the Board will consider the
evidence and, where applicable, the experts’ views and rationales in each of the following
topic areas:
13
1.
The business risk of each utility sector and Applicant;
14
2.
The Board’s last-approved equity ratio for each Applicant (where applicable);
15
3.
Comparable awards by regulators in other jurisdictions;
16
4.
Interest coverage ratio analysis; and
17
18
5.
Bond rating analysis.”
19
This approach of the EUB seems to be substantially the same as the traditional approach used by
20
this Board. I, therefore, first look at the business risk of the Mainline and then consider financial
21
market access and the EUB’s points 2-5.
18
1
3.0
BUSINESS RISK
2
Q.
HOW DO YOU VIEW THE BUSINESS RISK OF A PIPELINE UNDER THE
BOARD’S JURISDICTION
3
4
A.
In the discussion of risk in Section 2 I pointed out that income risk to the investor is a
5
function of business risk and financial risk. However, I then clarified that financial risk has been
6
set by this Board to modify the underlying business risk of a pipeline. In this sense financial risk
7
is a tool used by the Board. However, the Board has other tools in addition to simply setting the
8
deemed equity ratio and allowed ROE. In fact the whole regulatory process changes the risk of
9
investing in regulated industries, to the extent that when finance researchers try to model what
10
determines capital structures they frequently either exclude regulated industries completely or
11
add a “dummy” (zero vs one) variable for a regulated firm, simply because the act of regulation
12
itself explains more than can be explained by independent variables such as revenue variability.9
13
Q.
14
A.
15
year pipelines like the Mainline and TQM and the full cost of service pipelines like Foothills and
16
the TCPL BC System (the former ANG now the “BC System”). Full cost of service utilities have
17
their revenue requirement recovered from a limited number of customers and can true up actual
18
costs with revenues, so that they exactly earn their allowed ROE. This is the case for Foothills
19
and the BC System where both can exactly earn their allowed ROE, to the extent that in
20
surveillance reports to the Board they sometimes do not break out actual from allowed ROE,
21
since they are the same number! In the RH-2-94 hearing Dr. Berkowitz and I recommended a
22
25% deemed equity ratio for the full cost of service pipelines due to the absence of any income
23
risk. We made the same recommendation for TQM since its costs were recovered through the
24
Mainline. At the time TQM had a 25 % common equity ratio, ANG 28% and Foothills 30%.
WHAT OTHER TOOLS HAS THE BOARD USED?
Of the gas transmission pipelines there is the basic distinction between the forward test
9
The regulatory process determines the revenue requirement, so that variability in revenues caused by a
changing allowed ROE, for example, is not a risk factor in the way it would be for a competitive firm.
19
1
Compared to full cost of service regulation, the Mainline recovers its revenue requirement from a
2
larger number of shippers, so in practise rates are not trued up on a frequent basis. The regulatory
3
response to this has been to allow deferral accounts where any revenue deviations can be
4
captured in a temporary account with the balance allocated to future cost of service. Revenues
5
are then set based on forecast costs over the test period together with the disposition of the
6
balances in deferral accounts. In this way any unexpected revenue losses due to contract non-
7
renewals etc are recovered in a future test year and the costs are borne not by the equity holders,
8
but by future customers in the same way as for full cost of service regulation. However, since
9
deferral accounts are not allowed for all items, for example O&M costs, the shareholders are
10
liable for some forecasting risks with forward test year utilities that they are not with full cost of
11
service regulation.
12
In RH-2-94 Dr. Berkowitz and I filed Schedule 1 as Schedule 19 to our part B testimony.
13
Schedule 1 reports the actual versus allowed ROE for the major pipelines in Canada from 1989-
14
1993. Of note is that the two full cost of service pipelines (Foothills and ANG) exactly earned
15
their allowed ROE each year from 1989-1993. In contrast TQM, Westcoast (WEI) and the
16
Mainline (TCPL) over earned their allowed ROE in each year. It is difficult to see how over
17
earning the allowed ROE can be classified as “more risky,” but other forward test year pipelines
18
such as the two oil pipelines, Interprovincial (IPL now Enbridge) and Transmountain (TMP now
19
part of Terasen), did fail to earn their allowed return in some years. On this basis Dr. Berkowitz
20
and I recommended 28% common equity for the Mainline and 30% for Westcoast. The Board
21
actually set all of these pipelines on a 30% common equity basis for their mainline gas
22
transmission operations. Further the Board stated (Reasons for Decision page 25)
23
24
25
26
27
“With regard to the argument that regulation shields pipelines from risk, the Board
believes that its regulation provides pipelines with a degree of assurance of cost recovery
which is absent for non-regulated industrials. However, the Board believes that the
realities of market forces cannot be discounted when addressing pipelines’ business
risks.”
28
My interpretation of the Board’s decision is that they accept that regulation lowers pipeline risk
29
but that it may not be able to shield them completely from market forces.
20
1
Q.
WHAT HAS CHANGED SINCE RH-2-94?
2
A.
In terms of the ability of the Mainline to earn its allowed ROE, very little. In Schedule 2
3
is data on the earned vs actual ROEs for the major gas transmission pipelines since 1990. Note
4
that Foothills continues to be a full cost of service pipeline and exactly earns its allowed ROE.10
5
Similarly despite the changing supply position in the WCSB and the introduction of Alliance, the
6
TCPL Mainline continues to over earn its allowed ROE, in fact it has never earned less than its
7
allowed ROE except in 1994 when the NEB disallowed some costs related to fuel imbalances.11
8
Based on the objective data of whether risk has changed, there is no indication of any change in
9
the ability of the Mainline to earn its allowed ROE since RH-2-94.
10
In Schedule 3 is the same earned vs allowed ROE for the two premier gas local distribution
11
companies in Canada and NGTL (NOVA at the time of RH-2-94). The data for the two Ontario
12
Gas LDCs is based on weather normalised ROE’s since these utilities are not allowed deferral
13
accounts for variances due to weather. Further the NGTL data is not strictly comparable since for
14
a significant amount of time NGTL operated under incentive regulation and the early data
15
conflicts with that reported by CBRS in 1994. However, the message is very similar regulated
16
utilities on a forward test year consistently over earn their allowed ROE’s. In practical terms
17
there is very little risk involved in operating an ROE regulated utility in Canada.
18
Given the very low, if not non-existent, income risk, ROE regulated utilities in Canada have the
19
very stable ROI necessary to support large amounts of tax efficient debt financing.12
20
Traditionally I have always recommended higher common equity ratios for gas LDCs in Canada
10
My understanding is that the BC System’s deviation from allowed in 2001 was part of the
agreement on the change of control of ANG. In CAPP 31(a) no explanation was provided for the
BC System’s results in 2003.
11
See answer to CAPP 82(b).
12
Interest on debt is tax deductible at the corporate level and only taxed once, whereas income
earned by equity investors is double taxed. To compare the cost of equity and debt returns we
have to put them on the same tax basis and compare the pre-tax ROE with the corporate interest
rate. In this case a 3% spread between a 9% allowed ROE and a 6% interest cost, at a 40% tax
rate translates into a pre-tax cost of 15% for equity (9%/(1-.4) versus 6% for debt. So that the
spread on a same tax basis is 9%.
21
1
simply because they are subject to weather variation risk, which can, and sometimes does, affect
2
their coverage ratios and access to financial markets.13 For this reason I continue to recommend
3
that gas LDCs have the 35% allowed common equity ratios of the two Ontario gas LDCs. If this
4
weather risk is removed, as it has been for Terasen Gas (formerly BC Gas), then they can finance
5
with the 33% common equity that the BCUC allows Terasen. Finally Westcoast has recently
6
agreed to a 31% common equity ratio, up from the 30% awarded (mainline transmission
7
pipeline) previously. This is a small 1% increase since 1994 for a pipeline, like the Mainline, that
8
is also tied into the WCSB.
9
In my judgment the TransCanada Mainline has lower risk than the Ontario gas LDCs because the
10
LDCs have fewer deferral accounts, the variation of their earned ROEs is greater and historically
11
they have at times failed to earn their allowed ROEs. If the Terasen Gas allowed common equity
12
ratio of 33% and the Ontario Gas LDCs allowed common equity ratios of 35% are seen as upper
13
limits, it would indicate that the current allowed common equity ratio for the Mainline of 33%
14
continues to be generous. Further I also see the Mainline as lower or at least similar risk to
15
Westcoast, so that Westcoast’s 31% common equity ratio is also a valid benchmark.
16
It is also interesting to contrast this performance of regulated assets with the utility holding
17
companies (UHC) that actually face the market. For the major UHCs Schedule 4 gives their
18
earned ROEs along with those of the TCPL Mainline and Foothills. For example, what investors
19
invest in as "TransCanada" or TCPL is not the Mainline, but the combined entity including non-
20
regulated and regulated assets. This can be seen in the greater variability of its ROE. For 1993-
21
1997 TCPL consistently earned more than the Mainline, but then in 1998-2000 as TCPL
22
reorganised it earned less than the Mainline. Throughout this period the Mainline has
23
underpinned TCPL's results and been a beacon of stability. One way of assessing this greater risk
24
is simply to estimate the standard deviation in each firm's ROE. For the TCPL Mainline this was
25
1.05%, whereas for TCPL itself it was 2.47%, so the Mainline's ROE was only 43% as variable
26
as that for the whole company. However, as we have seen this variability in the Mainline's ROE
27
is not "risk," since it largely reflects the fluctuation in the Mainline's allowed ROE.
13
Most gas LDCs, unlike pipelines, have an interest coverage restriction in their bond indentures
that requires them to have a 2.0X interest coverage before they can issue debentures.
22
1
I can estimate the risk relative to the allowed ROE by taking Foothills ROE as the benchmark
2
and then look at both the average deviation of actual ROEs from this benchmark as well as the
3
variability of this deviation. This is included in the table in Schedule 5, where, by definition,
4
Foothills has a zero deviation every year. For example, I showed earlier that the Mainline has
5
over earned by about 0.24% relative to its allowed ROE, the above table shows that relative to
6
Foothills allowed ROE, it over earned by about 30 basis points and the standard deviation of this
7
deviation was 0.29%. In contrast TCPL has earned about the same 0.30% more than the
8
Foothill's allowed ROE, but its standard deviation or variability is much greater at 1.91%. This
9
fluctuation in TCPL's actual ROE reflects the greater risk stemming from its periodic forays (and
10
retreats) into non-regulated areas. The only one of these UHCs that is close to Foothills in terms
11
of risk is GMI. As a partnership it receives its 50% share of TQM and its Quebec gas distribution
12
assets seemingly without actually paying the income tax it collects, so that its ROE is
13
consistently higher than for the other UHCs. However, what is important is that the variability in
14
its ROE around the Foothills benchmark is 0.63%, only slightly higher than the TCPL Mainline.
15
Based on this variability around the Foothill’s allowed ROE, I would rank these UHCs in the
16
following order, from lowest to most risky, compared to the Foothills benchmark: GMI, Emera,
17
Fortis, TCPL, PNG & CU, Enbridge, Terasen (BC Gas Inc) and TransAlta. However, what it
18
also points out is that these UHCs are riskier than their underlying regulated assets and
19
adjustments have to be made in applying insights from the UHCs to the regulated assets.
20
Q.
IN YOUR JUDGMENT SHOULD THE BOARD BASE COMMON EQUITY
21
RATIOS SOLELY ON THE ABILITY OF A PIPELINE TO EARN ITS
22
ALLOWED ROE?
23
A
No. As the Reasons for Decision in RH-2-94 indicate the Board may not be able to shield
24
a pipeline completely from its underlying business risk. However, before examining this further
25
it should be pointed out that long term risks eventually become short term risks. Consequently
26
any long term risks must eventually be thought of as leading to a situation where a pipeline fails
27
to earn its allowed ROE. Of importance is that we have objective evidence from Schedules 1 & 2
28
that over the last 15 years NONE of the longer term risks put forward by the Mainline in various
29
hearings have actually materialised as short term risks. In RH-2-94 the company put forward a
23
1
series of longer term risks, such as competitive gas pricing, contract non-renewal, increased
2
competition from alternative pipelines and supplies, the diversification of the Ontario LDCs gas
3
supply, problems with negative salvage etc, in fact many of the same “risks” that it believes exist
4
today. However, from 1994 until 2003 none of these risks have actually materialised in the sense
5
of hurting the Mainline, since over this whole period it continued to over earn its allowed ROE.
6
It could be argued that it is simply a timing issue and that eventually some of these risks may
7
harm the Mainline. However, to harm the Mainline it has to be a situation of a “death spiral,”
8
where the tolls have to be so high that the Mainline is uneconomic and shippers shut in their gas
9
or find alternative uses for it, rather than ship through the Mainline. As long as the Mainline is
10
regulated through a forward test year with deferral accounts, the risk of contract non-renewals
11
etc are borne by the shippers through the readjusted tolls: the only risk to the Mainline is if the
12
shippers are unable or unwilling to bear that risk.
13
Further, the risk of “stranded assets” and the non- recovery of capital is simply another way of
14
viewing the death spiral. In terms of the charges to the income statement depreciation is a prior
15
charge to income, that is, before we get to the interest to bondholders or the income to the
16
shareholders the Mainline deducts depreciation as a return of capital. As a result, as long as the
17
Mainline earns its allowed ROE it also of necessity recovers its capital through depreciation.
18
Schedules 1-2 indicate that the Mainline has consistently over-earned its allowed ROE, which
19
also means that it has consistently recovered the capital invested in the Mainline. If there is no
20
risk attached to the Mainline earning its allowed ROE, then there is also no risk attached to the
21
Mainline recovering its capital. Again the objective evidence is that despite these longer term
22
risks being repeatedly raised in rate hearings, and even in some cases materialising, there is no
23
evidence of the Mainline ever failing to earn its allowed ROE. Further in my judgment the
24
impact of these longer term risks is not just small but it is also decreasing, not increasing.
25
Q.
WHY
IS
THE
IMPORTANCE
OF
THESE
LONGER
TERM
RISKS
DECREASING?
26
27
A.
Simply because the Mainline is a declining rate base pipeline. What this means is that the
28
capital invested in the Mainline is being recovered and not being reinvested. As a result, there is
24
1
less capital at risk at some arbitrary future point in time, than there was prior to RH-1-2002. As a
2
result, the Mainline is less exposed to these risks.
3
It is important to point out that risk is composed of two parts. The first part is the risk itself
4
stemming from the economy, such as the variability in exchange rates, while the second part is
5
the extent to which a company is affected or exposed to this risk. For example in the case of
6
exchange rate risk, all companies operating in Canada face the risk that the Canadian dollar may
7
increase or decrease in value. However, not all firms are equally exposed to this risk. Natural
8
resource companies pricing in US dollars are more exposed to exchange rate changes, than say a
9
consumer products firm buying and selling in Canadian dollars. The risk to the firm is then the
10
economy wide source of risk times its exposure to that risk. For the Mainline it may be that the
11
risks that have been repeatedly mentioned since at least RH-2-94 have now increased, however
12
this does not mean that the Mainline’s business risk has necessarily increased, since what is
13
unambiguous is that the Mainline’s exposure to those risks has decreased.
14
Q.
CAN YOU ILLUSTRATE WHY THE MAINLINE’S EXPOSURE HAS GONE
DOWN?
15
16
A.
The easiest way to see this is to work through a simple example. In Schedule 6 is a
17
spreadsheet of the cash flows on a very simple “pipeline.” The rate base is $100 and assumed to
18
last 25 years based on the “supply basin,” so $4 of depreciation or capital is recovered each year.
19
For simplicity the pipeline is assumed to be 100% equity financed with a 10% allowed ROE, the
20
same as the investor’s cost of equity capital or required return. Discounting the total cash flows,
21
the income return (ROE times rate base) plus the return of capital (depreciation) gives a present
22
value of $100. However, note that this $100 is comprised of three components. The first is the
23
return of capital or depreciation worth $36.30, the second the income return for years 1-10 worth
24
$52.29 and the final the value of the cash flow stream for years 11-25 worth $11.40. The market
25
value is the sum of these three components and is always split between return of capital and
26
income, for reasons I will come to later I have split the income into the value of that before and
27
after year 10. Note that this is completely standard; for example; part of the value of investing in
28
a long Canada bond is the return of capital at maturity. Only for perpetuities, where the
29
depreciation is constantly reinvested and the capital never recovered, is the value of the
25
1
investment simply the present value of the income stream. For a declining rate base pipeline like
2
the Mainline a very important part of its value is the return of capital.
3
Fair regulation requires that if investors commit $100 in capital the regulator allows them an
4
opportunity to earn a fair return. In this case, the market value of the investment will be very
5
close to $100. As I will discuss later, it is inherently unfair for the regulator to set the allowed
6
ROE above the investor’s required rate of return, so that $100 invested in the regulated assets is
7
immediately worth more than $100. That is, fair and reasonable rates imply that the regulated
8
firm’s market to book ratio should be around 1.0X14 and that in the weighted average cost of
9
capital (WACC) market values should be very close to book values. For a regulator to accept
10
market value weights significantly different from book value weights implies that regulation has
11
failed.
12
Now consider what happens if there is increased risk after year 10. This might arise because
13
although the assets are expected to last 25 years, there is increased uncertainty about supply
14
starting in year 11. As a result there is increased risk attached to the firm earning its allowed
15
return for years 11-25. As I indicated previously all the talk about stranded assets and longer
16
term risks must manifest themselves into increased uncertainty around the firm’s ability to earn
17
its allowed return. If all these “risks” have no impact on the regulated firm’s ability to earn its
18
allowed ROE, then by definition they are not risks. To capture this increased “supply” risk or
19
“competitive” risk or “operational” risk, I assume that the correct discount rate that reflects the
20
uncertainty of the regulated firm earning its allowed ROE for years 11-25 is 15%, not 10%, so I
21
am assuming a huge increase in risk. If the regulator does not recognise this higher risk for years
22
11-25, and investors believe that the firm will continue to earn 10% for the cash flows in these
23
years, then the stock price will drop $5.36 to $94.63, since the cash flows for years 11-25 are less
24
valuable now that they are more risky. In this way longer term risks can affect the current stock
25
price and the financial integrity of the regulated firm.
14
I say around since market prices will fluctuate and the firm is entitled to recover its floatation costs etc,
so a cushion above the required return to allow the stock to trade above its book value is fair and
reasonable.
26
1
There are several very important implications from this simple example. First, as always the
2
litmus test is the firm’s market to book ratio. If investors and the stock market are worried about
3
these longer term risks, then they will be impounded in the current stock price and reflected in
4
the market to book ratio. As always a market to book ratio below 1.0 indicates problems for the
5
regulated firm.15 In this case it is perfectly legitimate for the firm to point out to the regulator that
6
its financial integrity is at risk and that “regulator there is a problem.”
7
A second major implication is the magnitude of the price drop, even though the earnings are
8
more risky and warrant a huge increase in the allowed ROE to 15% from 10%, the impact on the
9
market value is only $5: a slightly more than 5% drop in share price even though the required
10
risk premium has increased dramatically. The reason for this is that the present value of the
11
return of capital is unaffected,16 as is the present value of the income for years 1-10. From
12
Schedule 6 only 11.40% of the market value of the company comes from the present value of the
13
riskier income from years 11-25. Of course the numbers are specific to this example, but the
14
point is simply that the contribution of the income coming fifteen or twenty years in the future to
15
the current market value is very small. As a result, the significance of these longer term risks is
16
much lower than more immediate risks.
17
Q
YOUR EXAMPLE ASSUMES THAT THE MARKET BELIVES THAT THERE
WILL BE NO REGULATORY RESPONSE, IS THIS REASONABLE?
18
19
A.
No. Regulation is not a static exercise it is dynamic. If problems occur, then firms bring
20
these problems to the regulator and frequently “compromises” are worked out. This is part of the
21
regulatory bargain and only regulated firms have this capability. For example if a competitive
22
firm suffers a supply shock then the stockholders are directly affected, but in contrast a regulated
23
firm can have losses put in a deferral account and allocated to future customers or apply to the
24
regulator for other means of protecting the stockholders from loss. Consequently it is
15
A recent example of this phenomenon is Pacific Northern Gas where its market to book ratio below 1.0
reflected the market’s view on PNG’s future operations and ability to earn its allowed ROE, as much, if
not more, than its current ability to earn its allowed ROE.
16
The riskyness of the return of capital is mainly determined by the depreciation rate as is discussed later.
27
1
unreasonable to expect no action on the part of the regulator to the increased risk after year 11 in
2
the above example.
3
If the market understands the regulatory bargain then it would expect the pipeline to approach
4
the regulator in year 10 for an increase in its ROE (or equity ratio) to reflect the increase in risk.
5
In this case, knowing that this will happen, the impact on the current stock price will be much
6
less. In this respect regulatory risk is simply the market believing that the regulator will not
7
respond efficiently to the situation at hand in the future. Again, there may be increased supply
8
risk or competitive risk or other risks for dates distant in the future, but if the market believes the
9
regulator will respond fairly at that time in terms of increased allowed ROE, then the impact of
10
these future events on today’s market value is much reduced and in many cases minimal.
11
A corollary to this efficient regulation is that regulated firms that operate with infrequent rate
12
hearings are clearly riskier than ones with annual reviews or those that have the ability to file
13
information to review near term events. In this way a pipeline that is known to be unable or
14
unwilling to file a rate hearing to increase the allowed ROE in year 11 will be more risky than a
15
pipeline, where a Board is more willing to closely monitor events, such as is the practise in
16
Canada. In this respect, the Board clearly stated in its RH-1-2002 Reasons for Decision (page 43)
17
18
19
20
21
22
23
24
“The Board notes the importance of performing depreciation studies on a timely basis
and of ensuring that depreciation rates reflect up-to-date information. The Board notes
TransCanada’s expert witness recommended that depreciation studies be performed every
three to five years and that TransCanada accepted this recommendation. Accordingly, the
Board would expect the filing of TransCanada’s next comprehensive depreciation study
to be within this time frame.”
I would judge these remarks to indicate that the Board is fully aware of the importance of
25
“dynamic” regulation and would find it difficult to believe that the stock market does not notice
26
this. Consequently the impact of these longer term risks on the Mainline today is muted.
27
The important point is that as long as the regulator is flexible and deals with situations as they
28
arise, and markets know this, there is less reason to allow higher allowed ROEs or thicker equity
29
cushions today. Paradoxically, the main reason for allowing a higher ROE or thicker equity
30
cushion today is if the market believes that the regulator is not up to the job. That is, the market
31
believes that future events will not or cannot be dealt with fairly at that time.
28
1
A final implication of the example is for fairness and the basic principal that tolls should reflect
2
the cost of providing service at that time. If for some reason the market believes that regulation is
3
inefficient and the market to book drops below 1.0, the regulator may increase the equity
4
thickness or the allowed ROE today on the grounds that current equity holders are not being
5
treated fairly. However, this raises two questions: the first is that current customers are paying
6
the uncertainty “cost” attached to future cash flows largely because of regulatory risk, not supply
7
or competitive risks. In this respect it is important to note that it is not current cash flows that are
8
riskier in the example, but the cash flows in years 11 and beyond. Yet increasing the current tolls
9
requires current shippers to pay the uncertainty costs attached to future deliveries. The second
10
implication is that in year 11 when the uncertainty materialises and the utility does have
11
problems earning its allowed ROE, it is important to note that the ROE has already been
12
increased to reflect this fact. It is then important that these risks not be double counted.
13
Q.
HOW DOES THE DEPRECIATION RATE AFFECT THE MAINLINE’S
EXPOSURE TO RISK?
14
15
A.
The depreciation rate changes the amount of capital and income coming from earlier
16
versus later years. As a result, whenever the depreciation rate is changed it affects the Mainline’s
17
exposure to shorter term versus longer term risks. For example, suppose a “supply” study
18
indicated that the assets in my simple example have a longer useful life of 35 instead of 25 years.
19
In this case the $100 capital cost is spread over a longer period, since the depreciation rate is
20
2.86% instead of 4.0%. That is the capital invested is recovered over a longer time period. In
21
Schedule 7 are the cash flows for this “pipeline” serving a “stronger” supply basin, where a
22
2.86% depreciation rate is deemed appropriate. Note that as long as regulation is fair and
23
efficient this is the correct assessment and the market value of the assets remains at $100.
24
However, the three components are different. First, the present value of the return of capital
25
drops from $36.3 to $27.6, in fact for an infinite supply basin the present value of the return of
26
capital would approach zero, as is the case for a perpetuity bond.17 The second implication is that
27
the present value of the first ten year’s of income goes up to $54.9 from $52.29. This is because
17
Note the assumption here is that depreciation is based on useful economic life, the example abstracts
from engineering life and reinvestment etc, since these do not affect the significance of the results.
29
1
the undepreciated asset base is greater during these early years, so the income earned on that rate
2
base also goes up. Finally with ten more years of income the share of the present value coming
3
from income beyond year 10 also increases from $11.40 to $17.5.
4
This is the most important point to note. It is a matter of simple arithmetic that as the
5
depreciation rate goes down more of the value of the asset comes from cash flows far off in the
6
future. In the original example of a 25 year useful life only $11.40 of value was exposed to the
7
probability of the firm not earning is allowed ROE beyond year 10, now with a 35 year life that
8
share increases to $17.5. Now consider again what happens if the ability of the pipeline to earn
9
its allowed ROE decreases at year 10. Assuming the same increase to a 15% required return, the
10
market value decreases by $$8.80 instead of $5.37 to $91.2. The fact that more of the income is
11
generated beyond year 10 means that the market value is more sensitive to the increased longer
12
term uncertainty. Intuitively these longer term risks are more significant for pipelines with lower
13
depreciation rates and more “at risk” at that time.
14
Conversely the Mainline has been in the opposite situation to the example. At the time of RH-2-
15
94 the Mainline’s average depreciation rate was 2.58%, but it was increased in RH-1-2002 to
16
3.42%. The economic useful life was essentially reduced from about 35 years to 25 years, and as
17
a result less of the current market value of the Mainline comes from income exposed to these
18
longer term factors that might cause the Mainline not to earn its allowed ROE. In this sense since
19
many of the risk factors raised by the Mainline today are the same factors that were raised in RH-
20
2-94 (as well as other hearings), to the extent that there is any validity to them, what is
21
undeniable is that the Mainline is less exposed to them now than before.
22
Q.
THE COMPANY PUTS A 25% WEIGHT ON SHORTER TERM AND 75% ON
LONGER TERM FACTORS, DO YOU AGREE WITH THIS WEIGHTING?
23
24
A.
No. As the previous discussion indicates as the depreciation rate increases and the
25
economic useful life of the asset shortens the proportion of the firm’s market value exposed or
26
sensitive to these longer term risks obviously declines. For example if we assume a 25 year
27
useful life, no significant reinvestment and a 10% ROE the adjacent graph shows the relative
28
importance of the future cash flows to $1 of current market value. For example the first year’s
29
cash flow constitutes 12.72% and the first two years 23.96% of every dollar of market value. The
30
1
first five years of cash flow
2
then constitute 50% of the
3
market
then
120
4
obviously, 100% is recovered
100
5
over 25 years, since after that
80
6
there is no recapture of capital
60
7
and also no income since the
40
8
rate base is fully depreciated.
20
9
One undeniable implication of
0
10
this is that shortening the
11
economic
12
increasing the depreciation rate reduces a firm’s exposure to longer term risks.
13
It is uncertain what shorter term represents, but the supply evidence put forward by the Mainline
14
indicates that production from the WCSB has plateaued at about 16 bcf a day and under its base
15
case declines to 11.6bcf a day by 2025. This forecast may be conservative, since it does not
16
include Alaskan Gas and there is significant longer term potential attached to supplies of coal
17
bed methane. However with the Mainline’s base case forecast and northern gas it is about 20
18
years until the Mainline utilisation drops to 50%. Further Alliance is fully contracted for another
19
fifteen possibly 20 years depending on renewal behaviour. I leave it to others to question these
20
long run forecasts, but I would judge “longer term” to be beyond year 10.
21
Based on these supply forecasts and the contracting on Alliance I see very little change in the
22
Mainline’s ability to earn its allowed ROE over the next ten years. This means that with a 25
23
year useful life, no reinvestment and a 10% allowed ROE about 75% of the value of a dollar
24
invested in the Mainline is sensitive to shorter term factors and 25% to longer term. Obviously
25
the breakdown can not be quantitatively assessed with the accuracy the above calculation
26
indicates, but in my judgment the longer term risk factors can not possible affect 75% of the
27
market value of the Mainline. Further putting such a high weight on these longer term factors
28
simply removes the obvious fact that in the shorter term the Mainline has very little risk, as
29
reflected in its continuing ability to earn its allowed ROE.
useful
life
25
23
21
19
17
15
13
9
11
7
5
% Value
3
and
1
value
% Value
and
31
1
Q.
DO YOU BELIEVE THAT THE DEPRECIATION RATE AFFECTS BUSINESS
RISK?
2
3
A.
Yes. It is important to recognise that depreciation rates are set based on useful economic
4
lives. If the underlying supply basin matures and there is less ability to supply a pipeline over say
5
a 35 year life then the correct response is to shorten the useful life and increase the depreciation
6
rate. In this way depreciation rates must perforce reflect supply and other risks. In RH-1-2002
7
the Board considered the correct depreciation rate for the Mainline based on supply forecasts and
8
stated (Reasons for Decision page 32)
“In the Board’s view, the Mainline’s depreciation rate should provide TransCanada with
a reasonable opportunity to recover its invested capital. Further, where feasible and
practical, the Board should attempt to ensure that companies under its jurisdiction are
treated on a similar basis.”
9
10
11
12
13
14
My interpretation of this decision is that by increasing the depreciation rate the Board recognized
15
that it has increased the Mainline’s ability to recover its capital from what would have been the
16
case using the depreciation rates at the time of RH-2-94 and RH-4-2001. In this sense the longer
17
term risks that affect capital recovery have already been, in part, dealt with by the RH-1-2002
18
decision. Further, by putting the pipelines on a “similar” basis the implication is that the capital
19
recovery risk is similar for pipelines under the Board’s jurisdiction. As such the longer term
20
capital recovery risks faced by the Mainline are similar to those of other pipelines serving the
21
WCSB.
22
The recency of the RH-1-2002 decision combined with the slow evolution of supply forecasts
23
makes it difficult to believe that longer term factors affecting the ability of the Mainline to
24
recover its capital have changed significantly since then. Even if they have, the correct response
25
is to consider a change in depreciation rates. It may be double counting to consider the impact of
26
these longer term risks in setting both the Mainline’s depreciation rate and in assessing its
27
business risk to set its deemed equity ratio.
28
Q.
CAN YOU EXPLAIN WHY IT CAN BE DOUBLE COUNTING?
29
A.
Yes. If the Board believes that supply factors have increased the risk of the non-recovery
30
of capital it should shorten the economic useful life of the asset and increase the depreciation
32
1
rate. If this is done correctly then the risk attached to the recovery of capital should be held
2
constant. This is why in Schedules 6 and 7 I discounted the depreciation cash flows (return of
3
capital) at the same 10% discount rate, despite the increased uncertainty around the pipeline’s
4
ability to earn its allowed ROE. In terms of the firm’s income statement depreciation (return of
5
capital) is a prior charge and the assumption in those schedules was that the depreciation rate
6
adjusted for any risk in the return of capital.
7
If the economic useful life is set incorrectly then this will affect business risk. For example, if the
8
Board were to set the economic useful life of the Mainline at 50 years and only allow a 2.0%
9
return of capital, then the risk of failing to earn the allowed ROE in the future and failing to
10
recover capital both increase. This is because future rates have to recover a bigger depreciation
11
component and a bigger income component (since the rate base is larger) and as a result the risk
12
of a death spiral is that much higher. Conversely, if the Board allows a shorter economic useful
13
life and increases the depreciation rate say to 5%, then there is less risk of a future death spiral
14
since there is less value at stake in the future.
15
Further as I showed earlier in Schedules 6 and 7 as the depreciation rate increases the share of
16
the market value reflecting “lower risk” principal recovery increases and the share of market
17
value exposed to business risk decreases. Consequently, and despite the maturing of the WCSB
18
and the existence of a “competitor” pipeline in Alliance, in my judgment the ability of the
19
Mainline to recover its capital is largely unchanged. Further, since the shorter term income risk
20
faced by the Mainline is also unchanged, I would judge that together the recovery of capital
21
through depreciation and the share of income from years 1-10 make up around 75% of the
22
market value of the Mainline. This leaves around 25% of the market value exposed to these
23
longer term risks beyond year 10. This would be a lower percentage than at the time of RH-2-94,
24
so the Mainline is less exposed to these longer term risks than at that time, overall I am therefore
25
skeptical that the business risk of the Mainline has increased materially either since then, or even
26
more so since RH-4-2001.
27
Q.
28
ARE THEIR OTHER PARTIES APART FROM THE STOCKHOLDERS
INTERESTED IN THE DEPRECIATION RATE?
33
1
A.
Yes. Depreciation represents a return of capital, but in the example I assumed 100%
2
equity financing for simplicity. In reality 67% of the Mainline’ capital is provided by bond
3
holders. Consequently the bondholders have a vested interested in assessing the impact of these
4
longer term risks since they are lending to the Mainline in some cases on 25 year maturities.
5
Further, as a declining rate base company, the Mainline should be structuring its borrowing
6
maturity profile such that its debts can be paid off with the depreciation cash flow as they come
7
due. Consequently if these longer term risks have indeed increased, then bond holders should
8
also have noticed that the risk of failing to get their capital back has also increased.
9
The reaction of the bond holders to increased risk is the same as that of the stockholders, they
10
sell their investments driving down their prices and driving up their yields until the new prices
11
reflect these new risks. For equity holders I have stressed repeatedly that the market to book
12
ratio signals any problems. However, the Mainline is not traded, so the signal from looking at
13
TransCanada Corporation’s market to book ratio is “murky.” In contrast, we can directly observe
14
the yield on TCPL’s debt, which reflects the risk of not earning interest and not receiving the
15
return of capital. After all if a death spiral sets in then this affects the bondholders as well as the
16
equity holders. I will discuss the empirical observations on TCPL’s yield spread later, but despite
17
this also being affected by non-regulated operations and thus also being “slightly murky,” it is
18
difficult to see any increased risk in the bond market since the spread on TCPL’s debt has
19
tightened since RH-4-2001, not widened. The bond market is signaling, possibly imperfectly,
20
that there has been no increase in business risk.
21
Q.
ARE THERE OTHER FACTORS THAT ARE RELEVANT?
22
A.
Yes. The risk to the Mainline comes from not enough gas from the WCSB, and other
23
alternative uses for natural gas, such as being transported on a competitor pipeline. At the time of
24
RH-4-2001 the position of Dr. Berkowitz and I was that Alliance was not a competitor. The
25
reason was simply that Alliance is essentially fully contracted under long term contracts, so that
26
it does not have the ability to compete for incremental supplies or take supplies away from the
27
Mainline. Further even though Alliance can expand relatively cheaply by compression, I
28
assumed that the Board would not countenance expansion to increase capacity that was not
34
1
needed. My view on this latter issue has been strengthened in part by the statements made by
2
Board member Quarshie.
3
In a presentation to the PCRI, Board member Quarshie laid out the five goals of the NEB one of
4
which is for “Canadians to derive the benefits of economic efficiency.” Further Board member
5
Quarshie stressed that “enabling implies a responsibility to ensure that projects in the public
6
interest can proceed.” I imagine that the opposite is also true: that it is not the goal of the Board
7
to create economic inefficiency or proceed with projects that are not in the public interest. In this
8
regard I would judge wasting scarce resources building pipeline facilities that are not needed as
9
promoting economic inefficiency and it is hard to see how this can be in the public interest. I am
10
therefore reinforced in my view that Alliance is not a “competitor” pipeline for the foreseeable
11
future and not a significant risk factor for the market value of the current investment in the
12
Mainline.
13
The foregoing comments are not meant to imply that the Mainline has no business risk. At the
14
time of RH-2-94 Dr. Berkowitz and I laid great stress on the fact that the Mainline was operating
15
at full capacity; that production from the WCSB was expanding and that there was shut in gas
16
due to the shortage of takeaway capacity. Consequently without even considering the impact of
17
regulation, the Mainline was very low risk. This situation has changed somewhat: the Mainline
18
tolls are now higher than they would be if it were operating at 100% capacity, while there are
19
stronger signals that the WCSB is maturing. However, a constant theme of the previous
20
discussion is the regulatory dynamic. As the WCSB has matured the Board has revisited the
21
Mainline’s depreciation rate to adjust the risk of capital (non) recovery and asked to be kept
22
updated. Similarly the Board has reviewed its ROE Formula. The fact is that investors in the
23
Mainline do not face the hypothetical risk of investing in an unregulated Mainline. Investors face
24
the risk of investing in the Mainline, given the policies of this Board. On this basis the risk is still
25
very low.
26
Q.
WHAT COMPARATORS WOULD USE FOR THE MAINLINE?
27
A.
before the Alberta EUB last year I compared the different utilities in the Alberta generic
28
hearing on the following basis:
35
1
The major short term risks caused by cost and revenue uncertainty:
2
3
4
5
•
On the cost side since regulated utilities are capital intensive most of their costs
are fixed. The major risks are in operations and maintenance expenditures.
However, over runs are usually under the control of the regulated firm and can be
time shifted between different test years.
6
•
On the revenue side the risks largely stem from rate design, critical features are:
7
8
9
10
11
12
o Who is the customer and what credit risk is involved. For example, electricity
transmission operators who recover their revenue requirement in fixed
monthly payments from the provincially appointed TA, who is responsible for
system integrity, have less exposure than the local gas and electricity
distributors who recover their revenue requirement from a more varied
customer mix involving industrial, commercial and retail customers.
13
14
15
16
o Is there a commodity charge involved? The basic distribution function is very
similar to transmission, except when the distributor buys the gas or electricity
wholesale and then also retails the commodity. The distributor is then exposed
to weather and price fluctuations depending on rate design.
17
18
19
20
o Even if there is no commodity charge, how much of the revenue is recovered
in a fixed versus a variable usage charge? Utilities that recover their revenue
in a fixed demand charge face less risk than those where the revenues have a
variable component based on usage.
21
The medium and long term risks are mainly as follows:
22
23
24
25
26
27
28
29
30
31
•
Bypass risk. The economics of regulated industries are as natural monopolists
involved in “transportation” of one kind or another. However, one utility may not
own all the transportation system so that it may be economically feasible to
bypass one part of the system. This happens for local gas distributors, when a
customer can access the main gas transmission line directly, rather than through
the LDC, or when a large customer may be able to bypass part of the transmission
system. This is often a rate design issue: a postage stamp toll clearly leads to
uneconomic tolls and potential bypass problems, whereas distance or usage
sensitive tolls will discourage it. Similarly, rolled in tolling will encourage
predatory pricing by potential regulated competitors.
32
33
34
35
36
37
•
Capital recovery risk. Since most utilities are transportation utilities, the critical
question is the underlying supply and demand of the commodity. If supply or
demand does not materialise then tolls may have to rise and the utility may not be
able to recover the cost of its capital assets. Depreciation rates are set to mitigate
this risk to ensure that the future revenues are matched with the future costs of the
system.
36
1
A common thread running through the above brief discussion of utility risks is rate design and
2
regulatory protection. There can be significant differences in underlying business risk that are
3
moderated by the regulator in response to those differences. The lowest risk utility is then one
4
with the strongest underlying fundamentals and the least need to resort to regulatory protection.
5
In contrast, another utility may have similar short term income risk, but only because of its need
6
to resort to more extensive regulatory protection, so that it faces more problematic longer term
7
risks.
8
On this basis I judged the lowest risk regulated utilities in Canada to be electricity transmission
9
assets, since these have the following characteristics:
10
*
Minimal forecasting risks attached to O&M
11
*
Revenue recovery via the TA through fixed monthly charges
12
*
Limited (non existent) by-pass problems
13
14
*
Minimal capital recovery problems, since there are many suppliers of electricity
as a basic commodity.
15
*
Deferral account for capital expenditures
16
and recommended 30% common equity ratios.
17
I then placed the gas transmission pipelines as the second lowest risk group. Here I classified
18
Foothills and the TCPL BC System (formerly ANG) as of equivalent risk to electricity
19
transmission assets with NGTL having marginally more risk than Foothills and the TCPL BC
20
System, since it is exposed to bypass and recovers its revenues through a forward test year from
21
a greater variety of shippers. However, the combination of distance sensitive tolls, the ability to
22
offer load retention service and a more rapid depreciation rate significantly reduce any increase
23
in risk NGTL may have faced since 1995. I therefore judged that on its own NGTL could
24
maintain its financial flexibility on the same 30% common equity ratio allowed mainline gas
25
transmission assets. However, because NGTL was then allowed 32% and was almost
26
“indistinguishable” from the TCPL Mainline, I recommended the same 33% common equity
27
ratio that this Board currently allows the Mainline.
37
1
I then judged the local distribution companies (LDCs), including both gas and electric as the next
2
riskiest. These companies are distinguished by their retail operations, which mean that their
3
revenues are recovered from a large number of industrial, commercial and residential consumers.
4
This exposes them to both the business cycle and weather fluctuations. This revenue recovery is
5
also a function of their rate design that may expose them to commodity charges and a fixed and
6
variable recovery charge. Within this group the conventional yardstick for LDCs is that
7
Consumers (Enbridge Gas Distribution Inc or EGDI) and Union Gas are both allowed 35%
8
common equity by the Ontario Energy Board. However, whereas the Ontario Energy Board
9
allows a purchased gas variance account (PGVA) to ensure that the full costs of gas are
10
recovered, they are still subject to volume related variances. In contrast, the BCUC allows BC
11
Gas (Teresen Gas) a more comprehensive deferral account, but limits the allowed common
12
equity ratio to 33%. With these yardsticks I recommended 35% common equity ratio for a
13
typical local distribution companies.
14
Finally, I recommended 42% as the upper end of a reasonable range for the common equity of
15
ATCO pipelines, given that the BCUC allows PNG, a smaller and much riskier pipeline, 36%
16
common equity. However, this ranking was provisional being dependent on the EUB developing
17
clear rules on intra Alberta pipeline competition and a rate design that lowers ATCO Pipeline’s
18
risk. It was, and remains, my judgement that none of the Alberta utilities were as risky as Pacific
19
Northern Gas (PNG) with a 36% common equity ratio or Gaz Metropolitain (GMI) with a 38.5%
20
common equity ratio, where I continue to regard these two as the riskiest regulated utilities in
21
Canada.
22
On this basis I continue to believe that the Mainline deserves a 30% common equity ratio based
23
on its underlying business risk. At the very least I see nothing to indicate any changes since the
24
RH-4-2001 decision that increased the allowed common equity ratio from 30% to 33%. In fact,
25
as I pointed out early the change in depreciation rates resulting from RH-1-2002 has reduced the
26
Mainline’s exposure to these longer term risks. Given that I support the Board’s view that capital
27
changes should be made only rarely in response to significant changes in business risk, I would
28
recommend that the Mainline common equity ratio remain at 33%.
38
1
4.0
FINANCIAL RISK
2
Q.
ARE CAPITAL MARKET CONDITIONS AT PRESENT MORE DIFFICULT
THAN THEY WERE AT THE TIME OF RH-4-2001?
3
4
A.
Over the past several years, the Canadian economy has experienced low and stable
5
inflation together with reasonably strong economic growth. Schedule 8 graphs the year over year
6
inflation rate along with the yields on long Canada bonds and Treasury Bills since 1990. The
7
graph shows the impact of the severe recession in Canada during the early 1990s when interest
8
rates declined from over 10% and inflation from the 5% level. Since the economy recovered in
9
the mid 1990s inflation has stayed around the mid–point of the Governor of the Bank of
10
Canada’s 1.0-3.0% range with much of the volatility coming from volatility in natural resource
11
prices.
12
We can clearly see in the graph that prior to 9/11 yields on 91 day treasury bill yields were
13
around 4.0% but trending downwards, a trend that was accelerated after the terrorist attacks.
14
After 9/11, 91 Treasury bill yields have stayed at low levels, but were recovering until the middle
15
of 2003 when they started to weaken again. Throughout 2004 yields on short term bonds have
16
been around 2.0%, similar to where they were in December 2001. However, Treasury Bill yields
17
reflect short run conditions in the economy, since by definition they reflect the opportunity cost
18
of investing over a short (91 day) horizon. Long Canada yields, in contrast, reflect the
19
opportunity cost of investing over a thirty year horizon. Consequently yields are much less
20
volatile. As Schedule 8 shows long Canada yields have been at the 5.0-6.0% level for the past six
21
years and currently at 5.15% (end of August) are marginally lower than December 2001, when
22
they were at 5.72%.
23
Interest rates, however, are only one part of the Bank of Canada’s of monetary policy; the other
24
part is the foreign exchange (FX) rate. To incorporate both the interest rate and FX rate effects
25
the Bank publishes a monetary conditions index (MCI), which is graphed in Schedule 9. Since
26
1990 the MCI has dropped precipitously as the Bank tried to dampen the impact of the early
27
1990s recession. It then held steady throughout the late 1990s, but dropped again during late
28
2001 and early 2002 at the time of the RH-4-2001 hearing. The reason for this was the easy
29
money policy adopted to offset the impact of the terrorist attacks. As well this was a period of
39
1
significant weakness in the value of the Canadian dollar. The Bank of Canada uses a trade
2
weighted exchange rate in its MCI and this bottomed out in October 2002. In Schedule 10 is the
3
more familiar graph of the Canadian dollar on a US dollar basis and in October 2002 the
4
Canadian dollar was worth only slightly over 63 cents. The pickup in the value of the Canadian
5
dollar since then has lead to the increase in the MCI observed in Schedule 9, as exporters get less
6
for their US dollar exports and thus demand is squeezed. As of August 2004 the C$ was over 76c
7
US compared to 63.3c US in December 2001.
8
Q.
HOW DOES THE GENERAL STATE OF THE CAPITAL MARKET AFFECT
COMPANIES?
9
10
A.
The recent low level of interest rates reflects the weakness in the economy over the last
11
several years, relative to the “go-go” years of the late 1990s. One way to see this is the graph in
12
Schedule 11 of capacity utilisation. This shows the extent to which factories are operating at
13
close to capacity. The graph shows the recovery from the recession of the early 1990s and the
14
dramatic increase in capacity utilisation up till the beginning of 2000. The slow-down from the
15
86% level for manufacturing in 2000 is then clearly evident, as it falls almost continuously as the
16
economy slows down through to the fourth quarter of 2000 when it bottomed out at under 80%.
17
The hiccough in the third quarter of 2003, when capacity dropped again, reflects the worsening
18
situation that caused interest rates to drop again. Since then capacity utilisation has recovered
19
and we are again approaching 85% capacity utilisation levels in manufacturing. Compared to the
20
time of RH-4-2001, instead of a worsening economic situation, the real economy is now much
21
stronger.
22
Changes in capacity utilisation affect corporate profitability since for each firm there is a break-
23
even point at which they start to make money and below which they lose money. For the
24
financial markets it is the state of corporate profits that is key since this is reflected in share
25
prices and the default riskyness of corporate debt. In Schedule 12 is a graph of after tax corporate
26
profits against GDP since 1962. Typically after tax profits are around 6% of GDP, so the drops
27
in the early 1980s and 1990s, when after tax profits dropped to lows close to 2.0% are clearly
28
evidence of the recession’s impact. In contrast the recent slow-down doesn’t merit the word
29
“recession,” since although evident in the data the impact of the slow-down on corporate profits
40
1
has been muted in Canada.18 The recent after tax profit low of 5.7% of GDP in the fourth quarter
2
of 2000 represents a minor fluctuation compared to past slowdowns and recession and corporate
3
profits have quickly recovered to the 7.50% level of a strong economy. However, again relative
4
to the time of RH-4-2001 the profit picture is one of an improving situation, rather than a
5
deteriorating one.
6
The combination of declining interest rates and booming corporate profits clearly affects the
7
stock market. From Schedule 13 the S&P/TSX Composite Index
8
As stock prices fell throughout the remainder of 2000, the S&P/TSX Composite finished the year
9
with a modest gain of 6.0%. Since the beginning of 2001, the drop in technology stocks that
10
marked the fourth quarter of 2000 continued with the tech heavy Composite continuing to fall.
11
By August 27, 2001 the S&P/TSX Composite was at 7,399, a 35% decline from the 11,389 level
12
of a year earlier. In September 2002 the index fell to the 6,000 level, indicating losses of almost
13
50% from the cyclical high and one of the worst bear markets since the 1930s.
14
Since September 2002 the TSX/S&P Composite has rebounded and most of the market
15
movements have been forward three steps followed by backwards two steps with more sustained
16
upward movement since March 2003. As of August 2004, the TSX/S&P Composite index was
17
back to 8,377, much below its peak but above where it was in December 2001 when it stood at
18
7,638. Overall the general trend in the equity market has been one of cautious optimism, rather
19
than the despondency existing at the time of RH-4-2001.
20
One indicator of the above is that spreads on corporate issuers have narrowed over the last few
21
years. In Schedule 14 is a graph of the spreads on BBB and A rated bonds minus the yield on
22
long Canadas. These spreads measure the default riskyness of corporate debt and are sensitive to
23
the business cycle. We can see, for example, that the BBB spreads widened dramatically during
24
the recession of the early 1990s and then narrowed as the economy recovered, before widening
25
again during the recent slowdown. Since December 2001 when the BBB spread was at 226 basis
19
peaked in September 2000.
18
Note that the data is from tax returns and so does not include the accounting write-offs that have
skewed the profit picture when looking at reported profits.
19
The composite is the old TSE300, data for September 2001 is not available..
41
1
points and the A spread 124 basis points, the spreads have dropped to 115 and 90 respectively.
2
Again the bond market is indicating that the economy is on a different trajectory than at the time
3
of RH-4-2001: the situation is improving not getting worse.20 Overall the capital market seems
4
more receptive to bond and equity issues now than at the time of RH-4-2001.
5
The strengthening economy also has implications for the government as it affects tax revenues
6
and counter cyclical expenditures. In Schedule 15 is a graph of government lending as a
7
percentage of GDP. In accounting jargon when the government “lends” it is running a surplus
8
and paying off prior debts, conversely government borrowing is when it is running a deficit. The
9
importance of Schedule 15 is simply that throughout the almost 25 year period from 1973 till
10
1998 governments at all levels in Canada ran persistent deficit and became the most dominant
11
actors in the capital market. This reached a ridiculous scenario when in the depths of the early
12
1990s recession governments were borrowing the equivalent of almost 10% of GDP. As these
13
deficits added to existing government debt being refunded, it is easy to see how government debt
14
“crowded out” other sectors from the capital markets. In turn this made the government’s fiscal
15
position the critical factor in the state of the capital markets.
16
However, by 1997 government lending had become genuine lending and governments were in
17
surplus for the first time in twenty-three years. In 2000 all layers of government in aggregate ran
18
a surplus of $32 billion as tax revenues soared and expenditures on welfare, unemployment, etc.,
19
declined along with the unemployment rate. This amounted to over 3.0% of GDP, the biggest
20
surplus since 1951, when governments were still actively paying down the war debt. Although
21
the weakening economy has eroded the aggregate surplus, it is remarkable that the weakening
22
economy has not presented more pressure on government finances. Currently aggregate
23
government surpluses are still about 0.8% of GDP.
24
The overall decline in government “lending” has opened up room for private sector borrowing as
25
corporations have returned to the equity and bond markets, following the strengthening of
20
Note that these are general spreads on A and BBB bonds. As I have previously testified before this
board, spreads on utility bonds are not as sensitive to the business cycle and the market recognises that a
BBB utility bond is not as risky as a BBB non-utility bond. I deal with utility spreads later.
42
1
corporate balance sheets. In Schedule 16 is a table showing the importance of various issuers in
2
the Canadian markets. Public sector issuers were the most important issuers and reached a peak
3
of 71% of all new security issues in 1994. Since then the taming of government deficits have
4
allowed private sector issuers to increase their ability to access the capital markets as the
5
government share has now dropped to under 50%. Of particular note is that with the decline in
6
interest rates the markets have been hungry for yield and income trusts have filled the void. As
7
recently as 1994 income trusts and limited partnerships represented less than 0.10% of total new
8
securities issues, but in 2003 they were over 10.0%.
9
Q.
DO YOU FEEL THAT INCOME TRUSTS ARE A GOOD SURROGATE FOR
UTILITIES?
10
11
A.
12
income gets passed through without a corporate income tax and there are less agency problems
13
attached to them. By the last remark I mean that with regular equities, there is always the
14
problem that managers think of the corporate cash as belonging to “themselves,” rather than the
15
stockholders. Consequently, there is a reinvestment risk attached to some firms, since they have
16
a history of making extremely bad corporate investments. This has happened previously for
17
several holding companies of Canadian regulated utilities, which have had poor investment
18
experience outside their regulated businesses. Income trusts and limited partnerships in part
19
shield the investor from these reinvestment risks by paying out substantially all of their free cash
20
flow and leaving it up to the investor to decide where to reinvest the cash.
21
However, apart from these organisational differences, income trusts are as varied as
22
corporations. As a result making a comparison with a regulated utility requires that particular
23
trusts be chosen. In Schedule 17 is a table of critical data for the major “comparables” in the
24
business trust segment of the Income Trust market.21 Note that there are nine different segments
25
ranging from restaurants to natural resources and the business risk characteristics of each trust
26
will flow from the underlying assets placed in it. From an investor’s point of view, apart from the
27
differing business risk, the most important features are liquidity represented by the market
21
Income Trusts are equities under a different legal format. The main difference is that the
Apart from business s trust there are real estate (REITs) and oil and gas trusts.
43
1
capitalisation and the yield. Reflecting the new issue data from Schedule 16 the market cap for
2
each set of comparables is significant with $10billion in the resource group alone.
3
The yield on the different business trusts ranges from 9.7% for restaurants and “other” to a low
4
of 7.1% for energy infrastructure. These yields will in turn reflect how much of the trust’s cash
5
flow is paid out and the underlying risk of the assets. As in all valuations, the more that is paid
6
out to investors, the less left for reinvestment and future growth. The fact that the payout ratio
7
averages 88% for these trusts, with several of them over 100%, indicates that there is limited
8
growth potential since very little is being reinvested. Further the fact that not all of the cash
9
distribution is taxable indicates that the yield normally includes some return of capital and would
10
exceed the investor’s required rate of return (cost of capital). Overall, the yield on all these
11
different businesses of 8.3% means a premium over the ten year long Canada bond of 3.6% and
12
less than that over the 30 year bond.
13
In the second part of Schedule 17 is more detail on some of the energy infrastructure business
14
trusts. This data includes Gaz Metro which indirectly includes gas distribution assets in Quebec
15
and part of TQ &M; Fort Chicago which indirectly includes 50% of Alliance plus a Chicago
16
NGL plant; and Enbridge Income Fund, which indirectly includes 50% of Alliance plus Enbridge
17
Pipelines (Saskatchewan). For these three trusts investors are requiring a cash yield of 6.5-8.1%,
18
which means that pipeline assets earning the Board’s allowed ROE can immediately be put into a
19
business trust and sell above book value.22 In the valuation summary the analyst valued Gas
20
Metro using a required yield of 6.8% and TransCanada Power’s LP using a required yield of
21
8.2%
22
Overall the expanding nature of the income trust market, and the low yields on similar assets
23
required by investors through income trusts both strengthen my previous observation that the
24
allowed ROE formula is generous. However, the important point is that this market is robust and
25
indicates that financing is readily available for low risk utility assets at currently allowed ROEs.
26
There is nothing in the behaviour of Canadian capital markets over the last few years that
22
If these assets also continue to earn an income tax component, even if it is not being paid, then the
structure is even more advantageous for investors.
44
1
indicate any access problems for low risk utility assets. Further, the growth of the income trust
2
market indicates a readily available source of financing source, which has been tapped by other
3
pipelines regulated by the Board.23
4
Q..
IF ROE AWARDS ARE GENEROUS WHY ARE SOME BOND RATING
AGENCIES EXPRESSING CONCERN?
5
6
A.
The bond rating agencies are concerned with accurately predicting the credit quality of a
7
firm’s debt. In this task they face an asymmetry; if they get it right they get little credit, but if
8
they make errors and firms default, that they previously rated investment grade, then they get
9
severely criticised. They therefore take a conservative approach and sometimes over react. In this
10
respect the most dramatic re-evaluation has occurred on the part of Standard and Poors (S&P) as
11
it has “harmonised” ratings between the US and Canada as a consequence of its takeover of the
12
Canadian Bond Rating Service (CBRS). In doing this harmonisation S&P has taken a
13
quantitative approach and seemingly simply taken standard ratios and judgments drawn from the
14
US and applied them in Canada with little qualitative adjustment for the different institutional
15
environment.
16
On March 5, 2003 S&P indicated that it was putting several Canadian utilities on credit watch
17
and was re-evaluating their ratings, primarily as it re-evaluated the nature of Canadian regulatory
18
protection. Subsequently several of them were downgraded. However, TransCanada Pipelines,
19
and Enbridge have an S&P A- and CU Inc an S&P A rating.
20
Q.
HAVE THESE DOWNGRADES AND REVIEWS HAD AN IMPACT ON
BORROWING COSTS?
21
22
A
Not as far as I can determine. Schedule 18 tracks spreads of major utility issues against
23
the equivalent maturity long Canada bond. The absolute level of the spread is not significant,
24
since the spread is affected by the bond’s maturity, what is important is the trend since the end of
25
2002 and from March 2003 around the time of the S&P announcement. The average spread is in
23
The extensive on and off again discussion of putting Manitoba Telecom into an income trust indicates
that even higher risk telecom assets can be financed through this structure.
45
1
the final row. The average spread was 91 basis points at December 1999, increased to 138 and
2
129 basis points by December 2000 and December 2001; it then increased again to 152 basis
3
points at the end of 2002. Since December 2002 it decreased throughout 2003 to finish at 95
4
basis points in December and since then it has been marginally above 90 basis points. Overall
5
utility spreads were slightly higher in December 2001 than they are today, since the trend in
6
profits and economic conditions is now generally up rather than down. As a result, the current
7
business environment results in marginally lower credit risk.
8
If we look at the individual issues, the three A rated issuer spreads at the end of December 2001
9
were 90-130 basis points, at the time of the S&P review (about March 2003) they were 80-123
10
basis points and are now 50-108 basis points. For the A- issuers the spreads were 80-150, 70-201
11
and then 49-119. For the BBB+ issuers the spreads were 80-191; 105-205 and 60-135, while for
12
the BBB issuers they were 88-235; 150-250, and 70-141. Overall it is clear that spreads have
13
tightened since the announcement of the S&P review and they are now tighter than either at that
14
time or the time of RH-4-2001. Overall, it seems that the S&P announcement and review (and
15
downgrades) has had a negligible impact on these issuers.
16
From this spread data there is no indication that S&P’s decision to impose harsher credit
17
standards on Canadian utility holding companies has had any impact on either their borrowing
18
costs or presumably the marketability of future debt issues. Spreads for issues rated A- and
19
below, the ones that presumably would have been affected the most by S&P’s announcement,
20
have almost all declined since the end of 2002 and the time of S&P’s announcement.24
21
The fact that the market has shrugged off S&P’s new harsher credit standards should not be
22
surprising given that they received a negative reaction when they were announced. In S&P’s
23
May 9, 2003 teleconference on TCPL when Ms. Dathorne of S&P referred to her observation
24
Note that since July 2001 S&P has downgraded CU Inc, Newfoundland Power, Westcoast, Fortis BC
Gas (Teresen), Epcor Utilities Enbridge Consumers Gas, and Union Gas.
46
1
that “based on this review that the business risk associated with TransCanada is increasing,”
2
Maureen Howe the equity analyst at RBC-DS, Canadas’ largest investment bank,25 said
3
4
5
6
7
8
9
10
11
12
13
“Well Michelle, with all due respect, if I can just move on, I guess, I don’t know what
business risk you can possibly be thinking of, but if you think the business risk of
TransCanada today – and I’d like to know where you think the business risk of
TransCanada, today, is higher than it was four years ago -- you know, I have to say that
you have a very different understanding of the company, I would say than I think, the
general investment community; because certainly, the company today, is a much more
focused business; you know, much higher proportion of its revenue is either locked in
under contract, if we’re talking about the power side or under regulated sector, if we’re
talking about the transmission side; and you know, has divested of – as I said, and let me
repeat myself – fractionation operations, in key operations, operations that made no
money.
14
So can you explain to me where this higher business s risk is coming from?”
15
In answer Ms. Dathorne talked about a “small” movement in terms of business risk but
16
emphasised that TCPL’s financial profile is significantly weaker than the average for the rating
17
classification of its North American, European and Australian peers. In other words, S&P did not
18
emphasise any increase in business risk, but simply stated that TCPL’s ratios are lower than its
19
peers: a quantitative assessment without any deep qualitative analysis. The trend in spreads since
20
the S&P report indicate that the Canadian capital markets have not bought into S&P’s simple
21
statistical comparison with US companies.
22
S&P is known to be a harsher judge of credit quality than either DBRS or Moodys. For example,
23
in a February 1998 report, Credit Ratings in Canada, DBRS indicated that its ratings for the same
24
companies were 0.23 categories, that is, a “high” or “low” qualifier, higher than Moodys and
25
0.55 higher than S&P. This would indicate that approximately in four ratings three would be the
26
same between Moodys and DBRS with one Moodys a category lower, whereas for S&P two
27
would be the same and two a category lower. However, in comparing S&P and DBRS ratings in
28
the companies in Schedule 18 in the prior table the S&P ratings are a total of nineteen “notches”
29
below those of DBRS, so that S&P rates these utilities on average at least a notch below DBRS,
30
where a notch is a modifier such as a high or low or plus or minus.
25
Maureen Howe was rated Canada’s top utilities analyst in Brendan Woods International’s annual
survey, Globe and Mail, August 20, 2003.
47
1
For two companies Union Gas and Terasan Gas Utility (BC Gas) the differences are marked and
2
S&P rates them three notches lower than DBRS. In both cases DBRS rates them as A, whereas
3
S&P rates them BBB. In March 2004 Terasen announced the discontinuance of its engagement
4
with S&P, stating that there is no benefit to an S&P rating, since it does not agree with S&P’s
5
credit assessment and does not plan to issue US dollar debt. Since currently both these
6
companies’ long term debt issues are selling at similar spreads to S&P A- rated credits the
7
market seems to agree with the companies rather than S&P. However, the fact that S&P could
8
come up with ratings so profoundly different from DBRS indicates that something has
9
“spooked” S&P, that has not “spooked” the Canadian rating company or the Canadian market.
10
Q.
WHY DO YOU THINK S&P HAS TAKEN SUCH A HARD LINE WITH
CANADIAN UTILITY HOLDING COMPANIES?
11
12
A.
One can only speculate, after all when Maureen Howe challenged S&P on their
13
perception of TCPL’s increased business risk she did not get a clear answer. However, DBRS
14
provided six reasons for their higher average ratings than the US agencies:
15
*
less penalty for size
16
*
better knowledge of Canadian companies
17
*
treatment of sovereign rating principle
18
*
technical factors in the treatment of holding companies
19
*
specific industry biases
20
*
unsolicited ratings not generally done by DBRS
21
Only some of these six factors are relevant for utilities. For example, it is known that the US
22
agencies now have a policy of rating all public market debt and on average assign marginally
23
lower ratings for unsolicited ratings, compared to when one that is solicited and paid for.
24
Similarly, the industry bias has traditionally focused on oil and gas companies not utilities, while
25
the sovereign treatment refers to rating government debt. What is left is: size, Canadian
26
knowledge and the holding company problems.
48
1
S&P has specific rules on how to rate the parent of a holding company that issues debt versus a
2
subsidiary and to illustrate I will refer to some examples from S&P’s Corporate Ratings Criteria
3
handbook with respect to telecom companies that have motivated this policy.
4
5
6
•
Frontier Telephone was rated AA- and was purchased by Global Crossing and the
rating subsequently lowered to BB+. The New York Public Service Commission
did not prevent the acquisition.
7
8
9
10
•
Cincinatti Bell was rated AA- when its parent acquired IXC Communications,
which had a B- rating, and subsequently Cincinatti Bell’s rating was dropped to
BBB-. The Public Utilties Commission of Ohio did not create any roadblocks or
impose any penalties on Cincinatti Bell.
11
12
13
•
Qwest acquired US West Communications, which was rated A+ and S&P warned
its rating would be cut to BBB- but regulatory concern was on service quality not
protection of bond holders.
14
Each of the three prior examples were taken from the “go go” era of the late 1990s when telecom
15
was in a state of flux and the internet revolution was going to transform telecom. Hence, these
16
were deals involving local telecom providers with stellar bond ratings and riskier internet and
17
broadband providers. In each case, the local public utilities commission evaluated the acquisition
18
and reflected on service questions but did not step in to protect the bond holders from significant
19
credit downgrades. The downgrades were the direct result of a change in the riskyness of the
20
holding companies which “infected” the subsidiary, since the parent had control of the sub and
21
could over-leverage it or dividend cash out to support parent company debt. In response, S&P
22
now has a policy that the credit rating of a regulated telecom can not be higher than the credit
23
rating of its parent.
24
For non-telecom utilities S&P
25
26
27
28
“rarely view(s) the default risk of an unregulated subsidiary as being substantially
different from the credit quality of the consolidated entity. Regulated subsidiaries can be
treated as exceptions to this rule – if the specific regulators involved are expected to
create barriers that insulate a subsidiary from its parent.”
29
In other words there is a cross subsidy from the regulated to the unregulated entity unless the
30
regulated entity is “ring fenced” so that any problems on the non-regulated side do not impact the
31
regulated side. S&P refers to this as “structural insulation techniques” which may involve:
49
1
•
separate incorporation of the sub
2
•
independent directors
3
•
minority ownership stakes
4
•
regulatory oversight to insulate the subsidiary
5
•
Restrictions on holding company cash management programs
6
S&P is very forthright in that the onus lies on the regulators. It states
7
8
9
10
“the bar has been raised with respect to factoring in expectations that regulators would
interfere with transactions that would impair credit quality. To achieve a rating
differential for the subsidiary requires a higher standard of evidence that such
intervention would be forthcoming.”
11
My reading of these remarks is that having been “burned” with these US telecoms and the lack
12
of reaction from US public service commission S&P is now taking a tougher line on all utilities;
13
an approach that they are bringing into Canada. The fact that this reaction was before the Enron
14
scandal and the problems with other “pipeline” utilities in the US will simply have reinforced
15
this attitude. The observation that S&P could not refer to any specific incidents of increasing risk
16
for TransCanada, when challenged by Maureen Howe, seems to confirm that S&P is simply
17
reacting to a situation in the US and that what is important to them are average ratios, that is,
18
conforming to ratios appropriate to the US, not the Canadian, experience.
19
Q.
IS THERE OTHER EVIDENCE OF THE IMPORTANCE OF “PARENT
COMPANY RISK” TO BOND RATINGS?
20
21
A.
Yes, if we return to the Terasen Gas and Westcoast ratings we have the following more
22
detailed information:
23
Rating
S&P
DBRS
24
Terasen Inc
BBB Stable
A(low) Stable
25
Terasen Gas
BBB Stable
A Stable
26
Terasen Pipelines
BBB Stable
A(low) Stable
27
Duke Capital
n/a
BBB
50
1
Westcoast Energy
BBB Stable
A(low)
2
Union Gas
BBB Stable
A Stable
3
Terasen Inc is the holding company for both Terasen Gas and Terasen Pipelines
4
(TransMountain), and note that S&P ranks them all as BBB Stable consistent with the earlier
5
remarks that without specific-ring fencing the parent’s rating gets assigned to the operating
6
subsidiaries. On the other hand DBRS rates the holding company more poorly than Terasen Gas.
7
Similarly Duke Capital owns Union Gas and other subsidiaries and gets a DBRS BBB stable,
8
while Union Gas gets an A stable. Again S&P rates both Union Gas and Westcoast as BBB.
9
It is clear from the above that as a general rule S&P does rate all subsidiaries the same as that of
10
their parent, consistent with the view that the parent can always “raid” the sub and thus threaten
11
the credit of the subsidiary’s bondholders. In contrast, DBRS views the credits as having more
12
independence, so that the bond ratings of Terasen Gas and Union Gas reflect their individual
13
risks, rather than that of their parent. In essence the S&P approach is closest to that of rating the
14
bonds on the basis of their lowest common denominator. This means that there is little that the
15
regulator for Union Gas or Terasen Gas can do to get the S&P rating up to investment grade
16
unless they substantially intervene in parent-subsidiary relationships and ring-fence the
17
subsidiary.
18
Q.
IS THE US EXPERIENCE RELEVANT FOR CANADA?
19
A.
No. It is extremely difficult to look at US energy infrastructure companies and draw risk
20
conclusions for their Canadian equivalents. On October 3, 2001 Maureen Howe of RBC-
21
Dominion looked at a sample of US and Canadian energy infrastructure companies. The
22
Canadian sample included ATCO Ltd, BC Gas, Canadian Utilities, Emera, Fortis, TransAlta,
23
Enbridge, TCPL and Westcoast. The US companies were El Paso Energy, Kinder Morgan and
24
Williams as pipelines, Enron and Dynergy as Energy marketing and trading and AES, Calpine,
25
NRG and Mirant as US generators.
26
Maureen Howe provided four reasons as to why the market was valuing the Canadian and US
27
utility holding companies differently:
28
Χ
The US energy companies have greater commodity price exposure.
51
1
Χ
The US companies have a greater share of their earnings from energy marketing
and trading.
2
3
Χ
The US companies lack dividend yield support.
4
Χ
The regulated nature of the Canadian energy infrastructure operations.
5
6
The latter two factors are particularly important. She thinks that the Canadian companies can be
7
thought of as “like convertible bonds. When interest rates are low, as they currently are, the
8
companies trade on their bond value and are supported by tax-efficient dividend yields. When
9
the 10-year GOC yield rises above 6%-6.5%, the Canadian companies trade on the basis of their
10
underlying earnings and P/E.” She goes on to say that “Although the regulated nature of the
11
Canadian energy infrastructure operations do not imply the same “sizzle” as the dynamic
12
businesses of their US cousins, the earnings that emanate from the regulated businesses tend to
13
be very low risk and largely immune to fluctuations caused by economic cycles.”
14
Ms. Howe mentions the lower risk of Canadian companies as a result of more protective
15
regulation. DBRS also notes that compared with US pipelines Canadian ones regulated by the
16
NEB have the benefit of:
17
•
Cost of service rather than market pricing
18
•
NEB policies promote stability rather than short term contracting
19
20
•
NEB rolled in tolling rather than marginal cost tolling promotes rate base
expansion.
21
•
FERC’s policies promote pipeline competition and discounting
22
In each case, Canadian pipelines are of lower risk than those in the US before we consider the
23
fact that most US pipelines are part of more diversified utility holding company. However,
24
despite this, US companies have been aided by a benign regulatory environment. The major US
25
federal regulator, the Federal Energy Regulatory Commission (FERC), has established very
26
generous ROEs by a formula approach. The FERC’s formula approach is based on a DCF
27
dividend yield plus growth formula where the growth is estimated from analyst’s short and
28
medium term forecasts, which are know to be biased high. However, an additional problem with
52
1
the FERC analysis is that until recently its sample of six comparables included (with their
2
estimated fair ROEs circa 2000):
FERC ROE Sample
3
4
•
Coastal Corporation
10.74%
5
•
El Paso Energy
12.10%
6
•
Enron Corporation
13.97%
7
•
PanEnergy Corporation
12.79%
8
•
Sonat Inc
12.32%
9
•
Williams Companies Inc
13.93%
10
Apart from the fact that some of these companies have been acquired, the mention of Enron and
11
Williams should raise huge red flags, as only a small part of their profits prior to bankruptcy or
12
financial difficulty came from pipeline operations. Further in 2000 the FERC moved to weight
13
the short term forecast at twice the long term forecast and then with its sample of electric
14
companies it adjusted the weights further to target a particular ROE. However, the bankruptcy
15
and/or acquisition of several of the companies in the FERC sample makes this approach
16
infeasible and some analysts expect allowed ROEs for regulated firms in the US to drop
17
dramatically.26
18
Finally FERC has been less forthcoming than some expected in reining in the utilities. After
19
Enron siphoned off $1.5billion from its two natural gas pipelines, the FERC instituted a review
20
of inter-affiliate transfers. Many expected FERC to impose minimum equity ratios of 30% and
21
requirements such as maintaining an investment grade bond rating before the parent could
22
manage the subsidiary’s cash. However, when the FERC announcement was made in November
23
2003 it fell far short of S&P’s expectations. As S&P noted
26
In the July 2003 issue of the Public Utilities Fortnightly, the editor Richard Stavros (Falling Down the
Earnings Curve) refers to a report by Lehman Brothers, Daniel Ford A Blast from the Past, which notes
that “Historically allowed returns or return on equity (ROE) have been 393 basis points above the 10 –
year Treasury yield (+/- 153 basis points)26 which implies decisions in the 9 per cent range could lie
ahead.”
53
1
2
3
4
5
6
“the degree of oversight by the FERC has traditionally been less than sufficient to justify
insulation. That the FERC took almost two years to respond to the Enron pipeline
situation indicates that timely intervention that would protect bondholder interests is not
likely when a regulated utility’s parent is experiencing financial problems. It seems clear
to Standard and Poors that the new rule falls far short of providing the requisite insulation
to justify any ratings separation for utilities regulated primarily by FERC”
7
It is clear from this comment from S&P that it is their disenchantment with events in the US that
8
has triggered their review of regulatory protection in Canada. Further they are not the only ones.
9
In a recent article in Public Utilities Fortnightly (August 2004) two members of the New Jersey
10
Board of Public utilities state
11
12
13
14
15
16
17
“ring fencing holds out the prospect for insulating regulated utilities from the traditional
failed diversification investments of the parent holding company….. Successful ring
fencing is even more critical considering that state regulators are facing challenges
created by failures of corporate governance, accounting scandals, and in some cases
alleged criminal conduct in energy markets. Ring fencing may be the only regulatory
device capable of levelling the playing field and forcing the holding companies to absorb
the consequences of failed non-utility investments.”
18
With FERC failing to implement ring fencing and these types of concerns being raised in the US
19
it is hardly surprising that S&P has adopted a negative tone towards US utilities, but that does
20
not mean that this evaluation is important in the Canadian market.
21
Q.
HOW DOES THIS AFFECT THE MAINLINE’S BOND RATING?
22
A.
The Mainline is part of TransCanada Pipelines, which with NGTL are the two main
23
subsidiaries of TransCanada Corporation. Currently DBRS rates both TCPL and NGTL as A
24
stable, while S&P rates them both as A- negative. However, the above discussion indicates that
25
the biggest risk to the Mainline’s S&P bond rating is probably what TransCanada Corporation
26
does, not necessarily what TCPL does. For example if the acquisition of GTN for US$1.7b ends
27
up being financed imprudently then this will affect the Mainline’s S&P bond rating. As Merrill
28
Lynch noted (26 July 2004)
29
30
31
“We view M&A as one of the larger risks to credit quality. Canadian pipelines find
themselves in a position of strength relative to US peers and may be tempted to pursue
higher risk/return opportunities in the US.”
54
1
Further the Mainline is clearly affected by the financing of other ventures within TCPL. For
2
example, the purchase of a 31% interest in Bruce Power A under the TCPL umbrella in February
3
2003 for $451mm clearly has the potential to affect the Mainline’s bond rating.
4
The important point is that with the S&P rating philosophy the Mainline is affected by other
5
corporate activities within TCPL, as well as within TransCanada Corporation. Currently, this
6
would indicate that these other activities are more likely to have a negative impact on the
7
Mainline’s “independent” bond rating in the same way that Terasen and Union Gas do not have
8
better S&P ratings than their parents. However, if TransCanada believes that the Mainline is
9
subsidising other corporate activities either within TCPL or TransCanada Corporation, then there
10
is a simple answer: either the non-regulated power and other assets can be moved out of TCPL or
11
the Mainline can be ring-fenced. By ring fencing the Mainline and having it issue debt under its
12
own credit, the market pricing of that debt will reveal immediately whether or not there is any
13
cross subsidisation.
14
Q.
DO REGULATED FIRMS HAVE FINANCIAL FLEXIBILITY WITH YOUR
RECOMMENDATIONS?
15
16
A.
Yes. Currently spreads on A- rated long term debt (Enbridge Gas Distribution
17
(Consumers) and TCPL) are 100-114 basis points over the equivalent long Canada yield. A
18
typical regulated firm will then have a mix of shorter and medium term debt, so that the average
19
debt cost would be perhaps 80 basis points over the long Canada forecast. Using the Board’s
20
forecast for 2004 of a 5.68% long Canada yield and 9.56% allowed ROE along with the current
21
allowed common equity ratio this means an overall cost of capital (ATWACC) of 5.760% with a
22
40% corporate tax rate. This is estimated as follows:
WACC and Interest Coverage
23
Pre-Tax
24
After-Tax
Weighting
25
Senior debt
6.48%
3.88%
67%
26
Common equity
15.93%
9.56%
33%
27
The pre-tax cost of capital is the EBIT cost of 9.60% (5.76/(1-Tax rate of 40%)), so the interest
28
coverage ratio is simply this EBIT divided by the interest cost or 2.11X.
55
1
An interest coverage ratio of 2.11X is reasonable considering the historic levels for Canadian
2
pipeline companies as indicated below:27
Interest Coverages
3
4
2001
2000
1999
1998
1997
1996
1995
1994
5
Foothills
2.16
2.16
2.06
1.88
2.27
2.34
2.17
2.10
6
TQM
2.15
1.99
1.69
1.57
1.90
1.92
1.88
1.64
7
NGTL
2.31
2.33
2.24
2.08
2.12
1.82
N/A
N/A
8
Westcoast
1.74
1.47
1.29
1.33
1.61
1.65
N/A
N/A
9
These companies are predominantly pure utilities with Foothills and TQM primarily reflecting
10
the impact of the NEB formula approach, NGTL the impact of negotiated settlements that have
11
allowed it to earn more than the NEB regulated pipelines and Westcoast’s more diversified
12
operations.
13
Q.
WHY HAVE YOU USED A CURRENT DEBT COST AND A 40% TAX RATE?
14
A.
These are reference numbers, since tax rates change. Further the current cost of debt
15
produces an interest coverage ratio consistent with the current Board allowed ROE and common
16
equity ratio. It makes no sense to target a particular interest coverage ratio and allow a higher
17
ROE simply because a pipeline has a high embedded cost of debt. For example, suppose a
18
pipeline has a current embedded cost of debt of 10.0% as the debt was raised some time ago
19
when interest rates were higher. In this case it would obviously have a lower interest coverage
20
ratio. If this lower coverage ratio were then used to justify a higher ROE or deemed common
21
equity ratio, then it means that rate payers are paying for the higher debt cost twice. They pay
22
once through the higher embedded debt cost and then again in a higher allowed ROE or deemed
23
equity ratio due to the lower coverage ratio.
24
I have consistently argued that if the Board believes that the allowed ROE and deemed common
25
equity ratios are fair, but the resulting coverage ratio is too low and there are access problems,
27
It is also consistent with historic coverages for the Mainline as indicated in NEB 1.35(a)
56
1
then the solution is to allow a pipeline to use some preferred shares. In the Summer 2004 issue of
2
their Preferred Share Quarterly BMO-Nesbitt Burns provided the following yields:
3
Preferred Share vs Canada Yield Comparisons
June 2004
4
5
Retractable Preferreds (%)
6
Dividend yield
4.01
7
Mid Canada yield
4.09
8
After tax spread (corp)
1.77
9
After tax spread (indiv)
0.63
10
11
Straight Preferreds (%)
12
Dividend yield
5.48
13
Long Canada yield
5.34
14
After tax spread (corp)
2.54
15
After tax spread (indiv)
0.98
16
17
Floating Rate Preferreds (%)
18
Dividend yield
3.42
19
BA (3 month)
2.12
20
After-tax spread (corp)
2.25
21
After-tax spread (indiv)
1.22
22
The retractable preferreds are compared to mid Canada bonds since the retraction feature
23
shortens their maturity compared to a long bond. The traditional straight preferreds are compared
24
to long Canada bonds, while the floating rate preferreds are compared to 91-day Bankers
25
acceptances (BAs), since their dividends are usually reset quarterly.
26
The important point about the comparison is that what we observe in the capital market is a
27
yield. This is determined by both risk and taxes. Take the straight preferreds, for example, in
28
June 2004 the long Canada bond had a yield of 5.34%, while straight preferreds had a yield of
29
5.48%. Clearly the preferreds would be regarded as riskier than the long Canada bond, since the
30
corporate issuer can default. However, the yield on the preferred shares was only 0.14% higher,
57
1
much less than the spread on equivalent corporate bonds. The reason for this tiny spread (which
2
is usually negative) is that dividend income gets more favourable tax treatment than interest
3
income. As a result, the correct comparison is the after tax yield difference, which Nesbitt-Burns
4
gives as 2.54% in favour of the preferred shares. This is the correct result: that on an after tax
5
basis the riskier preferred shares have a higher yield.
6
Currently the Nesbitt Burns data shows that retractable preferred shares are yielding 4.01%, so
7
for a company in a 40% tax bracket their pre tax cost, equivalent to bond interest, is 6.68% only
8
marginally higher than the 6.48% that I used in the interest coverage example. If the company
9
were to use floating rate preferred with a current cost of 3.42%, obviously the pre-tax cost would
10
be even lower. Consequently, if a pipeline has market access problems due to high embedded
11
interest costs, then it can use retractable preferreds to boost its interest coverage ratio until the
12
high embedded cost debt can be refinanced. For example, if the pipeline has a 10% embedded
13
debt cost and the Board, for some reason, targets a 2.11X interest coverage ratio, it can be
14
obtained as follows:
WACC with Preferred Shares
15
16
Pre-Tax
After-Tax
Weighting
17
Senior debt
10.00%
6.00%
55%
18
Preferred shares
6.68%
4.01%
12%
19
Common equity
15.93%
9.56%
33%
20
21
The overall before tax cost of capital increases from 9.60% to 11.6% but the interest coverage
22
ratio is 2.11X. The rate payer is still paying the high embedded cost of debt, since this cannot be
23
avoided. However, the allowed ROE would have to be increased to 13.5%, or almost 4.0% above
24
the fair ROE, to get the same coverage ratio with an ROE adjustment. Further in this case the
25
pre-tax WACC increases 14.13%.
26
The basic problem is that a high embedded interest cost lowers any interest coverage ratio; if this
27
causes access problems then the answer lies in using preferred shares, rather than giving the
28
equity holder a bonus to the fair ROE or equity ratio. Problems in the bond market can be met
58
1
more efficiently though preferred shares than common equity adjustments.
2
Q.
HOW DO YOU TREAT PREFERREDS IN A REGULATORY CONTEXT?
3
A.
In RH-2-94 Dr. Berkowitz and I recommended that traditional preferred shares be
4
replaced with a mixture of 75% debt and 25% common equity.28 The problem is that as the
5
Nesbitt-Burns data shows different types of preferred shares have different debt and equity
6
characteristics. For example if a pipeline issues five year retractable preferreds shares the
7
question is what happens after five years when the preferreds are retracted or the investor asked
8
to be paid off? If the preferreds stipulate that the payoff is in common shares then this is regarded
9
as a soft retraction and as a result the retractables have more equity like characteristics. If on the
10
other hand, the payoff is in cash, then they are more debt-like. This example illustrates how
11
companies traditionally structured preferreds to mimic debt like characteristics and yet claim
12
them as equity for reporting purposes. It is this flexibility that accountants have now clamped
13
down on.
14
However, despite their presentation for reporting purposes, the key legal element is that the
15
dividends on preferreds still have to be declared, while interest on debt is a legal commitment.
16
This means that a tranche of preferreds supports higher interest coverage ratios and may increase
17
a regulated firm’s market access and ability to issue senior debt. Consequently, they are an added
18
tool in a regulated firm’s financial arsenal.
19
Q.
THE MAINLINE IS ASKING FOR AN INCREASE IN ITS COMMON EQUITY
20
RATIO BASED ON THE RETIREMENT OF ITS TRADTIONAL PREFFERED
21
SHARE COMPONENT, DO YOU AGREE WITH THIS?
22
A.
No. Despite the above remarks the critical question is what constitutes a fair package of
23
capital structures and allowed return. The discussion of Dr. Berkowitz and I in RH-2-94 was in
24
the context of making adjustments to overall capital structures and comparisons across
25
companies. The Mainline had retired the preferred shares by 1999 and replaced them with junior
28
In its ATCO Gas GRA decision (2003-072), the Alberta EUB noted that ATCO has replaced its
preferred shares in the ratio debt 80% and common equity 20%
59
1
subordinated debentures (JSD), which it refers to as preferred securities. However, the JSDs are
2
simply debt issues that rank behind senior debt issues both in terms of interest and any winding
3
up of claims in a bankruptcy proceeding.
4
It is important to recognise that the contractual features of the JSDs identifies them as debt and
5
not preferred shares. First, the interest on the JSDs, like that on any debt instrument is tax
6
deductible, whereas the dividends on preferred shares have to be declared by the Board of
7
Directors and is not a legal liability until they are. Consequently, preferred shares are treated as
8
equity. For example, in CAPP 34(a) and (b) the Mainline correctly compared the after tax cost of
9
the JSDs of 4.64% to the yield on the preferred shares that were being replaced of 7.08%.
10
Second, TCPL’s senior debt contains cross default clauses (CAPP 34(d), the normal meaning is
11
12
“cross default provisions exist such that the moment any series of debt goes into default,
all series are so captured.”
13
As a result, default on the JSDs would trigger the cross default clause and immediately impact
14
the senior debt. In contrast non-payment of a dividend on a preferred share does not trigger the
15
cross default clause. Consequently the JSDs do not have the same cushion effect supporting the
16
senior debt as do preferred shares; their only advantage is in the ranking of claims in the unlikely
17
event of a liquidation of the Mainline.
18
In RH-4-2001 Dr. Berkowitz and I recognised the retirement of the Mainline’s preferred shares
19
and specifically stated in our recommendations (page iii of our executive summary)
20
21
22
23
24
25
“We put the common equity ratio of low risk transmission assets in the range 2530% and would recommend that TransCanada be awarded the upper end of this
range. This 2.0% increase in our recommended common equity component,
compared to 1994, compensates for the removal of the preferred share component
and other changes since 1994. “
26
Hence, Dr. Berkowitz and I felt at the time that we had dealt with the removal of the preferred
27
shares in RH-4-2001. It is my judgment that JSDs are debt and do not carry the same
28
characteristics as the Mainline’s recommended replacement package of 30% common equity and
29
70% debt. Further to allow this financing package now simply double counts what has already
30
been adjusted for in RH-4-2001.
60
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
Further in the Reasons for Decision (RH-4-2001, page 59) the Board stated
34
preferred shares but did not regard them as common equity and specifically set the Mainline’s
35
capital structure at 33% common equity and 67% debt, where the JSDs were included in the debt
36
component with no equity characteristics. Further the Board recognized that within the debt class
37
there were a variety of different types of debt, such as first mortgage bonds, MTNs, senior
38
debentures and JSDs, but that they were all debt.
39
Finally, I would like to simply mention that in the previous discussion of spreads and market access, the
40
yields on the TCPL bonds since 1998 would include the effect of the retirement of preferred shares and
41
their replacement with the JSDs.
“At the end of 2001, TransCanada’s consolidated common equity ratio stood at 33%
compared with 31% in 2000. While TransCanada indicated that its consolidated equity
ratio at 31 December 2001 was 35%, this calculation included 2% capital arising from
perpetual preferred shares. The Board does not consider that perpetual preferred shares
have liability and reward attributes which are comparable to common equity. The Board
has therefore excluded perpetual preferred shares from its definition of the deemed
common equity ratio.
In light of the above, the Board is of the view that it would be appropriate to increase the
Mainline’s deemed common equity ratio from 30% to 33%.
Decision
The Board approves an increase in the Mainline’s deemed common equity ratio
from 30% to 33%.”
Further in on considering the Mainline’s debt costs the Board went on to state
“TransCanada indicated that for 2001, the Mainline’s funded debt would amount to
$6,302,367,000, or 68.38% of the Mainline’s capitalization and that the average cost of
this debt would be 8.97%. This debt is made up of First Mortgage Pipe Line Bonds,
Debentures, Medium Term Notes, Junior Subordinated Debentures.
Decision
The Board approves a percentage of debt in the Mainline’s capital structure of
67%. The Board also approves a cost of funded debt and pre-funded debt of 8.97%
for 2001.”
My interpretation of the above is that the Board was well aware of the Mainline’s previous use of
61
1
5.0
WACC
2
Q.
WHAT IS THE BASIS FOR WACC?
3
A.
Before considering a firm with debt, first consider the simplest problem in finance that of
4
a 100% equity financed firm. I will use K as the investors required return, so that if the firm’s
5
earnings are a perpetuity its value, V, is its forecast perpetual earnings of X divided by K or
V=
6
X
K
(4)
7
If the earnings are expected to be $1mm and the typical investor requires a 10% return then the
8
firm’s value is $10mm. Note that I said the investor “requires,” since this is the meaning of a fair
9
return: it is the return required by the investor relative to other investment opportunities.
10
Only the investors knows what they require, but if we take the above equation we can reverse it
11
and infer what the investor requires from the current market value and the firm’s expected
12
earnings as
13
K=
X
V
(5)
14
With the example numbers, $1mm in earnings valued at $10mm implies that the investor
15
requires a 10% return. Since the firm has to meet this requirement we also refer to this 10% as
16
the firm’s cost of equity capital: what the firm has to pay to raise equity. The interpretation of
17
this is straight forward, if the stock market requires a $1mm in earnings to support the $10mm in
18
market value, the market value will change if the earnings expectation changes.
19
The above perpetuity equation is used to value conventional preferred shares, but it illustrates the
20
standard result in finance that market values and required returns or “costs of” vary inversely.
21
For example, we value bonds using the same formula adjusted for the fact that the fixed
22
payments do not go on in perpetuity, but instead are truncated at the bond’s maturity date.
23
However, we still get the same result: when market interest rates fall, and with them investor
24
required rates of return, then bond market values go up and vice versa.
62
1
2
Further the expected earnings of the firm can be broken out into the rate of return, r, on the book
3
value of the firm’s assets A or
V=
4
rA
K
(6)
5
if the firm’s $1mm in earnings are the result of earning a 10% return on a $10mm book value
6
then we get the additional insight that the firm’s market to book ratio is simply
V
r
=
A K
7
(7)
8
In this case, since the firm is expected to earn 10% and the investor requires a rate of return of
9
10%, then the market value is equal to what the investor has contributed, that is, the firm’s
10
market to book ratio is equal to 1.0.
11
The above results are perfectly general and indicate that there are two basic ways in which a firm
12
can increase its market value: the first is by increasing its earnings through a higher rate of return
13
r, the second is through lowering its cost of capital, K. In both cases, the market value goes up.
14
For example, if the firm’s expected rate of return increases from 10% to 11% its earnings
15
increase to $1.1mm and, all else constant, its market value increases to $11mm. Alternatively, if
16
it lowers its cost of capital to 9%, then its market value increases to $11.11mm. These two ways
17
of increasing value are the heart of corporate finance and reflect the value of investment
18
decisions and financing decisions respectively. Note also that given the higher stock market
19
value we can see that the market to book ratio now exceeds 1.0, indicating that the investor has
20
bid up the value of the shares since they are getting more than they originally anticipated when
21
the firm raised and invested the $10mm in the firm’s assets.29
29
The example assumes a 100% ROE regulated firm, as it involves more non-regulated assets it is more
difficult to look at the market to book ratio as a signal. It is then in the interests of the regulated firm to
make observing this market to book ratio as difficult as possible. I refer to this as “looking through a dirty
window,” where there is no incentive for the utility to clean the window. It may not be an accident that
there are so few pure regulated utilities left.
63
1
The above implications for the market to book ratio are important for understanding how capital
2
markets work. It is a central result in finance that when investors receive more than they require,
3
market values increase, and with them the market to book ratio. Conversely, when they receive
4
less than what they require; the market to book ratio drops below 1.0. This is observed every day,
5
for example, in the pricing of debt securities, where government bonds with higher coupons
6
(interest rates) than current market rates require sell at a premium to their par value and those
7
with smaller coupons sell at a discount. In the debt markets these bonds are referred to as
8
premium or discount bonds, but we could just as easily refer to them as bonds with market to
9
book ratios above 1.0 and below 1.0. The information and meaning is exactly the same.
10
I developed the previous example to emphasise the importance and generality of the market to
11
book ratio and the important relationship between what the firm is earning, r, and what the stock
12
market requires, K. In the example, the firm is earning 10% and the investor requires 10%, so the
13
stock market is telling the firm “only invest in new projects that earn at least a 10% rate of
14
return, otherwise you are wasting our money.” If, for example, the firm raised $10mm from
15
investors and invested in a new $10mm project earning 8% in perpetuity, its new stock market
16
value would be $18mm
17
$18mm =
$1mm + $0.8mm
0.1
18
Its value has gone up by $8mm, but only at the cost of investing an additional $10mm, so the
19
firm has wasted $2mm of stockholder’s money. This change in value is called the net present
20
value (NPV) of the investment. Corporate finance is focussed on the firm making decisions that
21
increase shareholder value and not waste it. The key message is that the firm should only invest
22
in projects earning at least the firm’s cost of capital, which in the example is 10%.
23
Now suppose in the example market interest rates drop and with them the investor’ required rate
24
of return. This has been one of the central features of capital markets over the last twenty years.
25
Long-term interest rates have been on a long-term downward trend since peaking at over 18.0%
26
in September 1981. Suppose in our example the firm’s cost of capital drops from 10% to 7%. In
27
this case, the firm’s market value would increase to $14.3mm and the market to book ratio
28
increases to 1.43X. Again the market to book ratio indicates that the firm is now earning 10%,
64
1
when investors only require 7%. Notice, however, that the original investors have earned a 43%
2
capital gain that they did not anticipate, since they expected to earn their 10% rate of return
3
through the $1mm in earnings on their $10mm investment. If this were a utility the stockholders
4
would have earned an excess return and we can note this by looking at the market to book ratio:
5
any market to book ratio significantly above 1.0 indicates that investors have earned a return
6
that is above a fair and reasonable rate of return.30
7
Further, the 8% investment that previously destroyed $2mm in value now increases it, since 8%
8
exceeds the new 7% cost of capital. The new value for the firm with the investment is
$25.7mm =
9
$1mm + $0.8mm
0.07
10
This $25.7mm is the $14.3mm market value of the original investment, plus the $10mm cost of
11
the new investment plus an additional $1.43mm NPV from the new investment.
12
The example indicates that the firm has to constantly monitor its cost of capital: a project that it
13
would not accept when its cost of capital was 10%, it will accept when it drops to 7%. Moreover,
14
this investment yardstick is NOT a rate of return earned by other firms, that is, other firms “r,”
15
but the market cost of capital, K. The rate of return, r, earned by other firms is irrelevant in
16
corporate finance, since the firm has to satisfy its stockholders, not other firms.31 Moreover,
17
non-regulated firms will have a variety of projects earning more than the cost of capital; some
18
will earn say 12%, and some 11% all the way down to the last project that is accepted with a 7%
19
return. Consequently, the expected average return of a competitive firm should always be higher
20
than its cost of capital, reflecting the fact that it will have some projects earning above average
21
rates of return.32
30
If the firm is sold for $14.3 mm, new investors only earn their 7% required, since their return is based
on the $14.3mm investment cost, that is., $4.3mm of “goodwill” plus the $10mm original cost.
31
This is the core reason why comparable earnings estimates, which simply try to estimate r, are of no
value in public utility regulation.
32
This further shows that if corporate rates of return (comparable earnings) are to be relevant it is the
marginal rate of return that matters not the average rate that is estimated from financial statement data.
65
1
Q.
CAN YOU FURTHER EXPLAIN WHY THE FIRM’S OTHER INVESTMENT
OPPORTUNITIES IS NOT RELEVANT TO THE FAIR RETURN?
2
Yes. Suppose a firm has four projects A, B, C and D open to them, each represents a
3
A.
4
perpetuity, has the same risk and the firm correctly estimates its cost of capital at 7%.
5
Cost
R
Value
NPV
M/B
6
Project A
$10mm
20%
28.6mm
16.6mm
2.86
7
Project B
15
15
32.14
17.14
2.14
8
Project C
12
10
17.14
5.14
1.43
9
Project D
10
9
12.85
2.85
1.28
10
The first project A costs $10mm and is expected to earn 20% so it will add $2mm a year, which
11
at 7% is worth $28.6mm, so the net present value is $16.6mm. What this means is that by
12
undertaking this investment at a cost of $10mm the firm’s market value is expected to increase
13
by a further $16.6mm over and above this $10mm. We can view this project’s market to book
14
ratio as 2.86X. Projects B, C and D then have NPVs of $17.14, $5.14mm and $2.85mm and
15
market to book ratios of 2.14X, 1.43X and 1.28X respectively.
16
The important point to note is that all four of these projects should be accepted by the firm, since
17
they all have positive NPVs. Even project D with the lowest expected return of 9.0% should be
18
accepted since it has a positive NPV of $2.85mm and makes the stockholders better off. We call
19
these project rates of return marginal rates of return on investment and the basic result in
20
corporate finance is that any firm interested in maximising shareholder welfare accepts projects
21
that have positive NPVs or alternatively projects where the marginal rate of return exceeds the
22
cost of capital.
23
Further it is completely irrelevant that the firm has project A where the marginal rate of return of
24
20% vastly exceeds the 7% cost of capital: it undertakes project D simply because its marginal
25
rate of return exceeds the 7% cost of capital. If the firm didn’t have project A it would still
26
undertake project D, or conversely if there were another project E with a 50% rate of return it
66
1
would still undertake project D. The objective criterion is simply that project D on its own
2
increases the market value of the firm, regardless of what other investment opportunities the firm
3
has. What matters is the investor’s opportunity cost of 7.0%, not corporate opportunities.
4
The only qualification to this is if for some reason the firm is constrained in its access to capital
5
and can not finance all the available projects. In this case, the firm’s available funds may only
6
allow it to finance projects earning as low as 10%. In this case it might turn down the 9% project
7
D, since its available funds should then be spent on higher marginal rate of return projects.
8
However, for regulated utilities in general and the Mainline in particular this is not a factor since
9
they are not normally constrained for capital, in fact the Mainline as a decreasing rate base
10
company has “surplus” funds and can not possibly be constrained.
11
What this means is that TransCanada may have much better investment opportunities than the
12
Mainline elsewhere, that is, lots of projects like A, B and C. However, in financial terms this has
13
nothing to do with the fair rate of return for the Mainline or whether or not TransCanada will
14
reinvest in it. As long as the Mainline offers a rate of return above the investor’s required rate of
15
return, then it is in TransCanada’s best interests to invest, regardless of its investment
16
opportunities elsewhere. Further there has been no evidence presented that TransCanada is
17
capital constrained, which given the cash generating status of the Mainline, might be viewed as
18
violating the “standalone” principle even if it were.33
19
Q.
HOW DOES ADDING DEBT CHANGE THIS?
20
A.
The basic principles are exactly the same. Suppose the firm continues to have $10mm in
21
assets earning 10%, but there is now $5mm in debt at a cost of 5% and $5mm in equity at a cost
22
of 15%.34 The firm’s weighted average cost of capital (WACC) is then 10% and the firm’s equity
23
value is determined by the net income to the common shareholders discounted at the equity cost.
24
As discussed earlier, the rate of return on the firm’s assets is called its return on investment
33
Note that TransCanada indicated in CAPP 8(a) that it has always been able to raise funds on reasonable
terms.
34
15% is used simply for ease of calculation.
67
1
(ROI), which is the return prior to meeting interest payments. In the example, the ROI is 10% so
2
the equity market value is determined as the earnings before interest of $1mm (ROI times the
3
$10mm in assets) minus the interest of $0.25mm (5% times $5mm) or $0.75mm. If this
4
$0.75mm in net income is discounted at 15%, then we get the equity market value of $5mm.
5
That is,
6
7
$1mm − $0.25mm
0.15
$5mm =
ROI * A − K b * B
Ke
or algebraically
8
9
$5mm =
where A is the total book value of assets, B is the amount of debt financing ($5mm) and I have
10
subscripted the two costs, b for debt and e for equity.
11
In this example the market value of the equity is $5mm so that the total enterprise value (V), or
12
overall market value of the firm (debt plus equity), is the $5mm equity value plus the $5mm of
13
debt or $10mm. This calculation is an example of the flows to equity (FTE) method of valuation,
14
where the flows to the equity holder are discounted at the cost of equity capital to directly
15
determine the value of the equity. However, to do this calculation we need the value of the debt
16
financing and most corporate investment decisions are separated from these financing decisions.
17
Consequently, it is conventional to rearrange this equation to get the WACC. First multiply
18
through by the cost of equity,
E * K e = ROI * A − K b * B
19
20
where I have substituted E for the equity market value. Second, group the equity and debt costs
21
and factor for the overall market value to get,
V (K e
22
23
E
B
+ K b ) = ROI * A
V
V
dividing through we get
68
V =
1
ROI * A
E
B
Ke + Kb
V
V
(8)
2
where the market value is equal to the pre-interest earnings discounted at the weighted average
3
cost of capital (WACC). In our example $1mm discounted at the WACC of 10% gives the total
4
market value of $10mm. The equity value is then this $10mm value minus the $5mm in debt or
5
the same $5mm as calculated using the flows to equity approach.
6
The WACC simply recognises the different sources of finance available to the firm and averages
7
them to get an overall cost of capital. In this sense the cost of capital is a blended cost of
8
financing to the firm, but once this is done all the previous results hold just as before. For
9
example, if the ROI exceeds the WACC then the market value increases and the market to book
10
ratio is again above 1.0. For example if the ROI increases to 11% and the WACC stays at 10%
11
then the total market value increases to $11mmm. The equity value is then $11mm minus the
12
$5mm in debt, so the equity value increases to $6mm and the equity market to book ratio
13
increases to 1.2X. Similarly, if the cost of equity declines to 13% from 15%, then all else
14
constant, the WACC drops to 9% and the market value increases to $11.11mm. Again with
15
$5mm in debt the equity value increases to $6.11mm for a market to book of 1.22X. Finally, if
16
the WACC drops to 9%, this becomes the hurdle rate for new investments; that is, the ROI on
17
new investments has to be greater than 9%.
18
Adding corporate income taxes does not materially change any of the basic results. All that
19
happens is that the after interest earnings become taxable and it is this after tax net income that is
20
discounted by the equity cost, that is,
Equity =
21
( ROI * A − K b * B)(1 − T )
Ke
22
where T is the corporate tax rate. Rearranging, as before, means that the after tax ROI has to beat
23
the after tax WACC or
69
V =
1
ROI (1 − T ) * A
E
B
K e + K b (1 − T )
V
V
(9)
2
The only difference is that since interest is tax deductible, whereas equity costs are not, the after
3
tax ROI has to exceed an after tax WACC. This is normally referred to as the WACC since most
4
firms pay tax, but Drs. Kolbe and Vilbert refer to it as the ATWACC to differentiate it from the
5
normal utility WACC where the return is included in the revenue requirement.
6
It is fundamental to corporate finance that the WACC uses market values. This means for
7
example, that the debt and equity ratios use market or target values for debt and equity divided
8
by the total enterprise value. The WACC then gives the current yardstick that the firm has to beat
9
in order to create shareholder value, that is, to increase the firm’s market value. Only by
10
calculating the WACC in this way can the firm be sure that it is accepting projects that enhance
11
shareholder value, that is, have positive NPVs, rather than destroying it.
12
Q
SINCE YOU FIRMLY BELIEVE IN WACC (OR ATWACC), WHY DON’T YOU
USE IT IN YOUR TESTIMONY?
13
The basic difference is that regulators are not concerned with maximising or enhancing
14
A.
15
shareholder value; their mandate is to set “fair and reasonable” rates and frequently this sets
16
them at variance with maximising shareholder value, since regulation should never be designed
17
to rubber stamp market values. This means that the regulator sets rates and through them the
18
firm’s ROI, whereas for non-regulated firms the ROI is determined in the marketplace.
19
Consequently, the ROI is changed for the regulated firm to make sure that the return to the
20
stockholders (ROE) is fair.
21
To continue with the previous (no tax purely for simplicity) example, where the WACC is 10%
22
and the equity cost 15%, suppose the regulator institutes some risk reduction techniques such as
23
the use of a forward, instead of an historic test year,35 or the use of deferral accounts. As a result,
35
Forward test years remove any inflationary bias involved in the use of an historic test year adjusted for
specific identifiable changes. With the decline in inflation most of the need for forward test years is
removed.
70
1
the equity cost drops to 11%. This is the critical example the Board needs to be aware of, since
2
as I indicated earlier, we have been in a period of long run declining interest rates since 1981 and
3
the problem that Boards have been largely dealing with is declining equity costs.
4
In the example, everything else is held constant, so the debt is still $5mm and the rate base (total
5
assets) $10mm; the only thing that has changed is the equity cost. The equity value can be
6
determined from the flows to equity formula as
$6.818mm =
7
$1mm − $0.25mm
0.11
8
In this case the equity holders recognise the reduction in risk and bid up the stock market value
9
from $5mm to $6.818mm for an extra capital gain of 36.4% over and above their fair return.36 If
10
the firm is now in a rate hearing to adjust its ROE, a tip off to the regulator is that the market to
11
book ratio has gone from 1.0 to 1.364X (6.818/5), so intuitively by lowering the firm’s risk and
12
seeing the market value increase the regulator knows that the allowed ROE has to be cut. The
13
obvious thing for the regulator to do is simply get expert opinion estimating the equity cost, and
14
if this is unbiased, notice and cut the allowed ROE to 11%. The equity value will then return to
15
$5mm and the stockholders will continue to earn a fair return on their $5mm investment. The
16
question is then what does estimating the WACC add to this process?
17
Assuming there is no bias to estimating the equity cost at 11.0% the new WACC is
WACC = 0.11 *
18
6.818
5
+ .05 *
11.818
11.818
19
or 8.46%.37 The most important thing to note is that the WACC uses market value weights and
20
since the equity market value has gone up to $6.818mm, the WACC uses an equity ratio of
21
57.7% and a debt ratio of 42.3%, rather than the assumed regulated weights of 50% for both.
36
As perpetuities they get their fair return as the earnings are paid out as a dividend.
37
Note that discounting the $1mm in pre interest earnings by 8.46% gives the total enterprise value of
$11.818mm. This is the new cut off rate for evaluating the firm’s investments.
71
1
The reason for the use of market value weights is that the WACC is the minimum rate of return
2
the firm has to earn to maintain its market value, which has increased from $10mm to
3
$11.818mm. Theoretically, it makes no difference whether this $11.818mm is the result of
4
actually raising $11.818mm, or whether it’s the current market value of the original $10mm
5
investment as it is in this case. The point is simply that using WACC as a cut off rate is simply
6
that it reflects what the firm has to earn to sustain current market values. In particular, the new
7
WACC of 8.46% is applied to the market value of $11.818mm. In contrast, the regulator should
8
not be interested in sustaining current market values, since in the example it is clear that the
9
allowed ROE has to be cut and implicitly that market value has to fall. Moreover, the regulator
10
has to determine a fair return on the book value of the investment, that is, the rate base, not the
11
firm’s market value. In this sense, there is a fundamental contradiction to applying the
12
conventional WACC to regulated firms.
13
However, suppose the Board tries to apply WACC. First, note that this exercise is much more
14
prone to error, and as a result subjective, than just estimating the fair return directly. This is
15
because as well as estimating the equity cost correctly, you have to estimate the market cost of
16
debt, not the embedded cost, the financing weights and the appropriate tax rate. All of these
17
components are subject to error, since many issues of debt are not traded and as a result it is
18
difficult to estimate either their cost or their market value. However, assuming all these
19
additional estimation problems away, suppose the correct 8.46% WACC is estimated and
20
awarded the regulated firm, what does this do?
21
If the regulator accepts this WACC, the equity value is $5.42mm or
V =
22
.0846 * $10 − .25
.11
23
Although the ROI is reduced from 10% to 8.46%, it is not reduced to the correct ROI of 8.0%,38
24
so the equity market value is still $0.42mm higher than it needs to be. The reason for this is that
25
using market value weights in the WACC puts greater emphasis on the higher equity cost than the
38
The correct regulated WACC is the average of the debt and equity costs using regulated book value
weights, in this case 50%.
72
1
debt cost. For non-regulated firms this is correct since the objective is to maintain these market
2
values and create wealth. However, it is totally incorrect for a regulator who is tasked with
3
awarding fair allowed returns and implicitly causing market values to drop when allowed ROEs
4
are too high. By estimating and applying a market based WACC the effect of the higher allowed
5
ROE is perpetuated by its impact on the higher equity market value.
6
Over time, if nothing else changes, the excess value will be removed. For example, with the new
7
market value of $5.42mm the new WACC is
WACC = 0.11 *
8
5.42
5
+ .05 *
10.42
10.42
9
or 8.12%. Again if everything remains constant, in the next rate hearing the regulator would cut
10
the allowed ROI to this level and the equity market value would fall again until after successive
11
rounds it ends up at the $5mm fair value for the equity. Note that the regulated firm, as well as
12
others with a vested interest in the firm as an investment, may complain about the regulator
13
being tough by repeatedly cutting the allowed ROE, but it is not being tough at all. The ROE is
14
still above the fair ROE. However, by using market value weights in the WACC and by shifting
15
the focus from the ROE to the WACC, this adjustment process is drawn out to the stockholders’
16
benefit. Further it gives the regulated firm an opportunity to bring up other arguments that may
17
delay even this adjustment. Consequently the adoption of WACC based regulation delays the
18
adjustment process to the stockholders’ benefit.
19
The basic insight from this discussion is that by using market values in WACC, the resulting cost
20
of capital is higher than a fair return, since the higher equity cost is given a greater weight.
21
Further if the firm is a pure ROE regulated utility it tends to “rubberstamp” the use of market
22
values and is contrary to “fair and reasonable” regulation.
23
Q.
CAN YOU ELABORATE ON THE LAST COMMENT?
24
A.
As previously explained, the market to book ratio is the basic signal as to whether or not
25
investors are being fairly treated. When we add in flotation costs and a desire to allow the
26
regulated firm to access markets at all time, most would believe that the market to book ratio
27
should be marginally above 1.0, say 1.10. In this way, after incurring flotation costs, the firm
73
1
should be able to net out the equity book value per share and avoid any dilution of the existing
2
shareholder’s interests. However, apart from this minor deviation from book values, the essential
3
point is that the correct financing weights for a regulated firm should be the regulated capital
4
structure weights, not the market value weights. To incorporate into the regulatory process a
5
regulated firm’s market value is to rubberstamp investor expectations, however unrealistic, and
6
delay the adjustment to a fair and reasonable value for the allowed ROE.
7
The Alberta EUB has directly addressed this question on a number of occasions. For example, in
8
connection with comparable earnings testimony the EUB stated (Generic Cost of Capital
9
Decision U-200452, page 24)
10
11
12
“The Board considers that the application of a market required return (i.e. required
earnings on market value) to a book value rate base is appropriate in the context of
regulated utilities.”
13
That is, you estimate a market opportunity cost, such as that from the CAPM, and apply it to
14
book values, not market values as is the assumption in WACC.
15
In explicitly considering the usefulness of ATWACC the EUB stated (Decision U-99099, page
16
300)
17
18
19
20
21
22
“The Board observes that the intrinsic long-run value of a pure play regulated entity is
best represented by book value. In other words, the present worth of future regulated
earnings, discounted at the allowed return, is by definition equal to book value assuming
achieved regulated earnings on average equal allowed regulated earnings. Accordingly,
the Board considers that book capitalization represents the best indicator of the long-run
market capitalization for a pure play regulated firm.”
23
It is difficult to see how a regulator could say anything other than what the EUB stated above,
24
since to accept a market to book much above 1.0 is in effect to rubberstamp unrealistic investor
25
expectations or to admit that allowed ROEs are too high. The EUB further recognised this when
26
it went on to say (U99099, page 303)
27
28
29
“The Board would be derelict in its statutory responsibilities to recognize market
capitalization ratios that are derived from a market value capitalization that deviates from
the intrinsic long-run value of the regulated firm.”
30
This is the clearest possible statement by a regulator of the fundamental grounds for rejecting
31
ATWACC and its emphasis on market values.
74
1
2
3
4
5
6
7
8
9
10
Further the EUB went on to say
“In essence, a regulated company’s earnings are driven by the portion of the original cost
rate base deemed to be financed by common equity. This fact results in a fundamental
disconnect to the theory that market capitalization ratios, which have deviated
significantly from book capitalization ratios, reflect the appropriate financial risk
necessary to determine a fair composite return to be applied to the original cost rate base
of a pure play regulated utility. This is because the earnings of a pure play regulated
utility are governed by and driven by the regulated return allowed on book equity. In
other words, it is the book equity that reflects the appropriate financial risk necessary to
determine a fair composite return for a pure play regulated utility.”
11
This means that the correct financial risk measure is that which I discussed earlier under
12
financial leverage. It is also the approach pioneered by this Board, where financial risk
13
adjustments using the deemed common equity ratio are made for differences in business risk.
14
The EUB went on to calculate an ATWACC using regulated book value capital structure weights
15
and the embedded debt costs. In this case (Decision U-99099, page 303)
16
17
18
19
An ATWACCBV would be suitable for a regulated utility whose profit, by legislation, is
limited to a fair return on the book value (i.e. original cost) of its assets. The Board notes
that an ATWACCBV is consistent with the logic of the traditional method of determining
fair return.
20
In our example, the ATWACCBV is the 5% debt and 11% equity cost weighted with the 50%
21
regulated capital structure weights. In this case the ATWACCBV is 8.0% and awarding this 8%
22
cost of capital means that the value of the equity is
V =
23
.08 * $10 − .25
.11
24
or $5mm. This is the exact same result that would arise if the firm were simply given its 11%
25
lower ROE.
26
The EUB ATWACCBV correctly recognised that apart from any estimation error attached to the
27
equity cost, the WACC approach is inconsistent with allowing a fair return to a regulated firm.
28
The only approach consistent with allowing “fair and reasonable” rates is to estimate the
29
regulated, “comparable,” sample of firm’s ATWACC using book value weights and embedded
30
debt costs. In this case the exercise comes down to the normal problem of whether or not the
75
1
estimated equity cost is accurate or not. I will not enter testimony on this question, since it is not
2
part of the hearing, but suffice it to say that I judge the estimates of Dr. Vilbert as being high.
3
The final step is to adjust for differences in the financial leverage between the calculated WACC
4
estimates and the firm in question. That is, given the use of market value weights in the
5
calculated WACC, which in my example were 57.7% common equity and 42.3% debt, are
6
leverage adjustments need to apply the estimates to the regulated book equity, which in our
7
example was 50%?
76
1
5.0
LEVERAGE ADJUSTMENTS
2
Q.
WHAT IS THE BASIS FOR MAKING A LEVERAGE ADJUSTMENT?
3
A.
One basic proposition in finance is that investors don’t like risk, and as I showed in
4
Section 2.0, increasing the amount of debt financing magnifies risk. Logically, therefore as firms
5
finance with more debt they magnify business risk and investors respond to this increased risk by
6
requiring a higher rate of return. This logic is unassailable and not in dispute. Therefore, if an
7
expert estimated an equity cost from a sample of regulated firms and applied that estimate to a
8
firm with the same business risk, but much higher financial risk as represented by the regulated
9
or book debt equity ratio, then the estimated return would be below a fair return. For example,
10
estimating a fair return from a sample of normal utilities that happened to have no debt at all and
11
applying that to the Mainline with 67% debt would be patently unfair.
12
However, note two things. First of all leverage adjustments have nothing to do with ATWACC.
13
Leverage adjustments are theoretically necessary in the normal estimation approach where the
14
equity cost is estimated from a sample of firms and then a recommendation made for the firm in
15
question. Drs. Kolbe and Vilbert could, for example, take their equity costs from their sample
16
and make leverage adjustments without going through the trouble of estimating their ATWACC.
17
Second, in the above example there is obviously a problem, since it is extremely difficult to find
18
a sample of regulated firms with substantially different debt ratios in the first place. In reality
19
most regulated firms in Canada have similar debt ratios, or their debt ratios have been
20
specifically set, as is the practise of this Board, to offset differences in business risk.
21
To elaborate on this last point, as late as August 1997 in testimony before the CRTC Dr.
22
Berkowitz and I were using two samples of regulated firms for estimating the fair return. The
23
first sample consisted of six energy distribution companies and the second six telcos. However,
24
the CRTC had increased the business risk of the telcos by opening their long distance markets to
25
competition and there was no doubt that their business risk was higher than that of the energy
26
distribution companies. Normally this would have caused comparability problems. However, the
27
CRTC also allowed their common equity ratios to increase to 55% to compensate for the
28
increased business risk. Consequently at that time we judged the overall risk of the six telco
29
sample to be useful in comparisons with energy distribution companies. Explicitly it was our
77
1
judgment that no leverage adjustments were needed going from a telco sample with 55%
2
common equity, that is, 45% debt to an energy distribution sample with much greater financial
3
leverage.
4
By and large this continues to be my judgment: that the actions of regulators, like this Board, to
5
equalise risk obviates the need for leverage adjustments. In fact, in the recent Alberta generic
6
hearing the EUB specifically followed the lead of this Board and set common equity ratios for a
7
large sample of ROE regulated companies such that they could all earn the same formula
8
allowed ROE.
9
Q.
WHAT IF THERE ARE SIGNIFICANT LEVERAGE DIFFERENCES?
10
A.
The first question the Board has to ask is: are these leverage differences real, that is, were
11
they set to equalize overall risk or not? The second question the Board has to ask is: are the
12
leverage differences based on market or regulated book weights? If the answer to the first
13
question is that the leverage differences just offset business risk differences, then no action is
14
needed. If the answer to the second is the differences are only due to “temporary” market value
15
differences then they should also be ignored.
16
To continue with the previous example where the equity cost dropped from 15% to 11% and as a
17
result the equity market value increased and the equity ratio at market values increased to 57.7%
18
from the regulated 50%. Suppose this were the sample average from say twenty companies and
19
the results had to be applied to a non-traded regulated firm with 50% common equity and the
20
same business risk as the sample. Dr. Kolbe would seem to argue that the sample average has
21
less financial risk and that to apply the estimated equity cost, assuming it is accurate, to the
22
regulated firm in question underestimates its fair ROE, since it has a 50% debt not 42.3%. As a
23
result, he would increase the recommended ROE from the 11% estimated from the sample or
24
conversely recommend a higher common equity ratio. I will show later that their leverage
25
adjustment gives the highest plausible leverage adjustment. However, the approach itself is
26
wrong for two reasons.
27
First, if the regulated firm is earning approximately the same allowed ROE and has the same
28
capital structure, there is no reason to believe that its implicit market valued equity ratio is any
78
1
different from that of the sample. That is, the non-traded regulated firm in all likelihood has an
2
implicit market valued debt ratio the same as those of the sample firms, so there is no leverage
3
difference that needs to be adjusted for. Further allowing a higher ROE just increases the market
4
value of the equity, causing the market valued debt ratio to drop further, creating an even bigger
5
internal contradiction.
6
Second it is important to remember that in the example the market valued debt ratio fell not
7
because the firm substituted equity for debt and reduced the amount of fixed interest payments
8
and financial risk, but because the equity cost fell. This is the reality of the declining interest rate
9
scenario we have lived with since 1981 and the effect of a persistent decline in equity costs
10
coupled with regulatory lag. However, there is a big difference between the impact of
11
substituting equity for debt in a regulated capital structure and an increase in equity value as a
12
cause of the fall in the market valued debt ratio.
13
In the example, the equity is obviously riskier at $6.818mm and a market to book of 1.36X than
14
it is at $5mm and a market to book of 1,0X. This is because it is highly unlikely that the regulator
15
will cut the allowed ROE when the stock is trading at book value. In contrast, the higher the
16
market to book ratio the more likely the regulator will cut the allowed ROE and thus the riskier
17
the stock. In contrast, Drs Kobe and Vilbert would have us believe that the equity in a regulated
18
firm is less risky when it is trading at a market to book well above 1.0, since the debt ratio is
19
lower. I don’t accept this.
20
Second and more significant, the financial leverage risk premium stems from the imposition of
21
fixed interest charges. That is, prior to receiving their equity return the firm has to pay these
22
interest charges. This risk does not change as the market value of the firm changes; it only
23
changes as the book debt equity ratio changes. For example, if the Mainline moved to a 60-40
24
debt equity ratio in terms of book amounts, then there would be less interest expense.
25
Consequently, the financial risk, both to the bond-holder and the stock-holder, would decline and
26
with them the both the debt and equity costs. In this case, a leverage adjustment would indicate a
27
lower equity cost, since the financial risk has declined. As Standard and Poors have stated,
28
29
“Similarly ratios using market value of a company’s equity in calculations of leverage are
given limited weight as analytical tools. The stock market emphasises growth prospects
79
1
2
3
and has a short time horizon; it is influenced by changes in alternative investment
opportunities and can be very volatile. A company’s ability to service its debt is not
affected directly by such factors (italics added).”
4
That is, S&P is basically saying book value leverage is important, when it is assessing the default
5
or credit risk in debt, whereas market values don’t count, or at least don’t count as much. If it is
6
book values and interest payments that affect credit risk and the cost of debt then this is the risk
7
that also affects utility equity investors.
8
Following on from the Alberta EUB’s decision that to accept market value weights would be a
9
“dereliction” of duty, the obvious implication is that the weights in the sample WACC should
10
also be book value weights. In my example this means that the regulated book value of 50%,
11
rather than the market value debt ratio of 42.3% is what matters. Hence in comparing this 50%
12
regulated debt ratio with the firm in hand that also has 50% debt ratio means that no adjustment
13
is necessary. Making an adjustment based on market values is then inappropriate for a regulated
14
firm. As the EUB again noted (Decision U99099, page 301)
15
16
17
18
19
“the Board considers that beta and the cost of equity do not change to the extent
necessary for an ATWACC, determined from market capitalization weights, to remain
constant when applied to the book capitalization for a pure play regulated utility. The
increase required to the cost of equity to achieve a constant ATWACC would be
excessive and violate the fair return standard.”39
20
In my judgment that the only time that a leverage adjustment is needed for a regulated firm is
21
either, when its overall risk differs from that of a sample of regulated firms from which an ROE
22
estimate is derived, or when its business risk changes and the Board wants offset this change so it
23
can continue to award a formula allowed ROE. In these cases, as I indicated earlier, in my
24
judgment the literature has not reached “closure” on how to make leverage adjustments and the
25
Board is best advised to base its decision on the business risk of the pipeline and its access to
26
financial markets. These are the factors discussed in Section 2.0, which the Board has used in the
27
past.
39
Note that in CAPP 32(c) TCPL indicated that its requested fair return is equivalent to 13.3% on the
same 30% common equity allowed Foothills. It is difficult to see how the Mainline deserves almost 4.0%
more as a risk premium than another pipeline accessing the WCSB.
80
1
Q.
SUPPOSE THE BOARD FEELS THAT A LEVERAGE ADJUSTMENT BASED
2
ON MARKET VALUES IS NECESSARY ANYWAY, DO YOU AGREE WITH
3
DR.KOLBE’S ADJUSTMENT METHOD?
No. Again, it is well accepted that financial risk magnifies business risk. The basic
4
A.
5
financial leverage equation indicates that the accounting return to the stockholder is determined
6
as follows
ROE = ROI + ( ROI − Rd ) D
7
S
(2)
8
where these are all book values, that is, D and S are the book values of debt and equity and Rd is
9
the embedded cost of debt. The equation simply comes from manipulating the firm’s financial
10
statements. It means, for example, that with a fixed stock of assets, as revenues and the ROI
11
changes, then the greater the amount of debt the greater is the variation in the accounting return
12
to the stock holders. However, the above equation says absolutely nothing about how the stock
13
market reacts to this financial risk, that is, how market values change, or how the cost of equity
14
changes as the firm uses debt.
15
To understand how the investor’s required rate of return or equity cost varies with the use of debt
16
we need a valuation model. The first valuation attempt was by Franco Modigliani and Merton
17
Miller (M&M) who in 1958 developed an arbitrage model to show that the total enterprise value
18
was independent of the use of debt. This was their famous “no magic in debt argument.” If
19
individuals can borrow on the same terms as the firm, then investors will not pay a premium for
20
firms that use debt, since the firm is not adding value. Consequently, they derived the following
21
formula
Ke = K0 + (K0 − Kb ) B
22
E
(10)
23
where the K’s indicate the cost of equity and debt, not accounting returns, and B and E represent
24
the market values of debt and equity respectively. The subscript 0 then indicates what the equity
25
cost would be if the firm had no debt outstanding, which is often referred to as the unlevered
26
equity cost.
81
1
Note two things about this equation. First, apart from redefining returns and debt ratios, in form
2
it is the same as the leverage equation I used earlier. This is because in the accounting model
3
total assets are fixed, whereas in this valuation model M&M “proved” that the value of the firm
4
was fixed. As a result, changes in the book and market debt ratios have the same impact. Second
5
M&M “proved” that as the market value was constant the weighted average cost of capital was
6
also constant, which in this case means that it is equal to the unlevered equity cost. However,
7
note that I italicised “proved,” since this was a mathematical proof that followed from their
8
assumptions, not a description of reality.
9
In the M&M equation changes in the market valued debt equity ratio (B/E) are multiplied by the
10
spread between the WACC and the cost of debt. It is this coefficient that determines how much
11
changes in the debt equity ratio affect the equity cost since it is this coefficient that determines
12
the risk. This is the important point: people who believe that changes in the debt equity ratio
13
have a big impact on the equity cost believe that the coefficient on the market valued debt equity
14
ratio is high and vice versa.
15
However, the overall market value in the M&M model is only fixed by their assumptions. To
16
emphasise, remember that from equation (9)
17
V =
After − tax operating income
WACC
18
The total firm value is after tax operating income divided by the after tax WACC. Given that
19
M&M were discussing capital structure not operating changes, the after tax operating income,
20
the numerator above, is by definition constant. What M&M “proved” was that with firm value
21
constant the WACC must also be constant. In this case, given that the WACC is a weighted
22
average of the debt and equity costs, the equity cost has to increase with more debt to offset the
23
impact of more “cheaper” debt. This is what equation (10) indicates.
24
However, if the market value increases with more debt then from equation (9) the cost of capital
25
will decrease and vice versa. In this case, the equity cost may then increase or possibly even
26
decrease with the use of debt, it all depends on the valuation model and the assumptions that are
27
made. The critical question is how the use of debt affects the overall firm value; the impact on
82
1
the WACC and the equity cost then follow directly.
2
M&M’s “no magic in debt” result was controversial in 1958 and remains so today. This is
3
because of the assumptions required to “prove” their result. The most important are that:
4
5
6
7
8
9
10
11
12
13
•
•
•
•
•
•
•
there are no taxes of any kind;
there are no transactions costs;
there are no information asymmetries between borrowers and lenders;
everyone can borrow on the same terms, that is., if the company can issue 25 year
bonds or access the swap market, then so too can other individuals;
all firms are perpetuities that pay out 100% dividends;
there are no bankruptcy or financial distress costs;
there are two or more identical firms with different levels of debt that can be
arbitraged.
14
All of these assumptions have been disputed to a greater or lesser extent and if any of them are
15
incorrect then the total value of the firm is affected by the use of debt. Hence, so too is the cost
16
of capital.
17
M&M’s result is a classic in corporate finance and they won the Noble prize in economics for it.
18
However, its great strength lies not in its result, which few accepted then or now, but the fact
19
they focused corporate finance on the implications of their assumptions. For example, in 1963
20
they recognised that they made a mistake in their treatment of corporate income taxes and
21
corrected their original paper. They then showed that, all else constant, the value of the firm
22
increases due to the tax shield generated by the tax deductibility of interest payments. The reason
23
is simply that what we term value is the private value and by reducing corporate income taxes
24
the private value of the firm increases at the expense of the government. Hence from equation
25
(9), if the private market value increases the WACC of necessity must decline.
26
In fact in the M&M (1963) model the WACC declines continuously since the corporation can
27
issue risk free debt and the average and marginal tax advantage to debt are the same. In this case,
28
the equity cost changes in the following way with the use of debt,
29
K e = K 0 + (1 − T )( K 0 − K b ) B
83
E
(11)
1
There is still a financial leverage risk premium but it is now smaller, since the use of debt also
2
generates a valuable tax shield. Note that in M&M (1963) changes in the market valued debt
3
equity ratio are now multiplied by (1-T), so are smaller than in M&M (1958). Thus assuming a
4
40% corporate tax rate, people who believe in M&M (1963) would estimate a leverage impact
5
only 60% the size of those who believe in M&M (1958).
6
Since 1963 all the other assumptions of M&M have been relaxed and every time an assumption
7
has been relaxed there is another leverage equation similar to equations (10) and (11) and
8
another estimate of the leverage effect. However, two main theories of capital structure have
9
emerged: the static trade off (STO) model and the pecking order hypothesis (POH). The STO is a
10
static model that assumes that firms trade off the tax advantages of using debt against the loss of
11
financial flexibility that arises due to excessive leverage. It is this model that develops the
12
familiar “U” shaped WACC function below as the firm increases its debt ratio.
13
The U shaped WACC
14
WACC
15
16
17
18
19
Significant tax
advantages
Loss of financial
flexibility severe
risk of distress
20
21
Initially the WACC declines due to the tax advantages of debt. In the M&M (1963) model, for
22
example each dollar of debt increases the firm’s market value by the value of the corporate tax
84
1
rate,40 the WACC then starts to increase as the loss of financial flexibility sets in. Obviously
2
there has to be some offset to the tax deductibility of interest, otherwise all firms would try to
3
finance with 100% debt. The offset comes as the debt becomes riskier and has to be sold on
4
higher and higher yields and the firm loses its financial flexibility.
5
In contrast, the POH, developed in 1963 by Gordon Donaldson at Harvard, is a dynamic model
6
of financing based on the fact that firms are controlled by managers. In this case, firms raise
7
capital by issuing securities that impose the least restrictions on management. Consequently,
8
firms primarily rely on internal funds and only after these are exhausted do they go outside for
9
capital, where then they initially rely on bank debt and bonds, rather than new equity.
10
I have reviewed these basic ideas on capital structure since the flat ATWACC approach of Drs,
11
Kolbe and Vilbert is essentially the 1958 M&M model as extended to include corporate and
12
personal taxes by Miller (1977). This is a very important model and for the last 26 years I have
13
taught corporate financing to second year MBAs with the first five weeks devoted almost
14
exclusively to these ideas,41 as well as to the implication that if this model holds there is no value
15
to the activities of investment bankers and they should all study marketing! I then spend the
16
balance of my course explaining how companies add value by adopting different financing
17
decisions. The fact is that financial theory has come a long way since 1958 and is now better
18
harmonised with practise: no one believes the flat WACC model fits reality; it is simply a good
19
starting point to discuss how investment bankers, like Mr Lackenbauer, can create value for
20
firms.42
21
However, a flat ATWACC does have the advantage that it gives the largest possible leverage
40
This simple model has been dubbed adjusted present value (APV) by Professor Myers. In Principles of
Corporate Finance (2nd Canadian edition, 1991 pages 490-493 they work an example and the base case
NPV of $170,000 is then increased by $592,000 by the tax advantages to debt. In this case, Professor
Myers, who Dr. Kolbe references throughout his testimony, clearly believes in the tax advantages of debt.
41
This is MGT2300. A course outline can be downloaded from my web page at
http://www.rotman.utoronto.ca/~booth
42
It would be interesting to ask why investment bankers are so well paid if corporate financing decisions
as represented by a flat ATWACC have no value and firms can do whatever they want.
85
1
effect, that is, the coefficient on the market valued debt equity ratio in the equity cost equation is
2
as large as possible. I showed earlier that the M&M 1958 flat WACC model gives a bigger equity
3
cost adjustment (equation (10)) than if the WACC declines with leverage in the conventional
4
way (equation 11). However, Dr. Kolbe goes further by assuming a flat ATWACC in the presence
5
of corporate taxes, which gives an even bigger coefficient on the market valued debt equity ratio.
6
To illustrate Drs. Kolbe gets his leverage adjustment by assuming a flat, that is, constant
7
WACC.43 Dr. Vilbert first calculates the WACC using market value weights from his sample:44
Ke
8
E
B
+ K b (1 − T ) = WACC = K A
V
V
9
Dr. Kolbe then assumes that the WACC (KA) is constant and then either alters the equity ratio to
10
get a new equity cost or alters the equity cost to get a new equity ratio, holding everything else
11
constant. In terms of the equity cost, implicitly Dr. Kolbe is rearranging this WACC equation to
12
solve for the equity cost (Ke) at any leverage ratio,
K e = K A + ( K A − (1 − T ) K b ) B
13
E
(12)
14
Since the WACC is assumed constant, it has the same no leverage equity cost (K0) as before, the
15
only difference is that it is this cost minus the after tax cost of debt that determines the leverage
16
coefficient. With a constant WACC this coefficient is larger than either the M&M (1958) no tax
17
case or the M&M (1963) tax case as a simple comparison with equations (10) and (11) indicates.
18
In fact, as far as I am aware it is the largest coefficient possible, since I have not seen an equity
19
cost equation with a larger coefficient.
20
The reason for the very large leverage adjustment in equation (12) is that the model is internally
21
inconsistent. Equation (12) and the flat WACC assumes the tax deductibility of interest which
43
Note that a flat ATWACC requires in part that personal taxation offsets the corporate tax shield, yet in
CAPP 85(b) TCPL indicated that it has never commissioned any study of its marginal investor’s tax rate
and only considers corporate taxes in its capital budgeting procedures.
44
Note that as explained earlier the use of market values is not appropriate for regulated firms, either
directly or indirectly through WACC estimates from samples of regulated firms.
86
1
causes the WACC to fall, but there is no explicit account of the offsetting costs that negate this to
2
keep the WACC constant. For example, if the WACC is constant it could be that as the market
3
valued debt equity ratio increases the debt cost also increases due to the higher risk of insolvency
4
and the costs of financial distress and bankruptcy. This would be particularly true as the firm
5
goes to very high debt equity ratios. In this case, what is keeping the WACC constant is an
6
increasing Kb as creditors protect themselves from the insolvency risk attached to highly debt
7
financed firms. From the spread date in Schedule 16, we know this happens. Moreover, it is
8
obvious from equation (12) that if the debt cost, Kb, increases with the debt equity ratio then the
9
equity cost does not increase so fast, which is what Solomon showed in the Journal of Finance in
10
1963.45 The intuition is simply that “debt” in highly debt financed firms has some of the same
11
characteristics as equity. 46
12
To show these principles backtrack to the previous example, where the equity cost was assumed
13
to decrease from 15% to 11% due to a reduction in risk and consequently the equity market value
14
increases from $5mm to $6.818mm. As a result, the market valued debt ratio decreases from
15
50% to 42.3%, simply because the equity value has increased due to regulatory lag. Suppose that
16
the equity cost is then accurately estimated at 11.0%, but that someone believes that a leverage
17
adjustment is needed to apply this to a firm with 50% debt; how could this be done?
18
One way is to estimate an unlevered equity cost from equation (10) by inserting the debt cost of
19
5% the debt equity ratio of .423/.577 and the equity cost of 11%. In this case, the unlevered
20
equity cost is 8.46% and the use of debt financing has increased the equity cost from the debt
21
free 8.46% to the observed 11.0%, so 2.54% is the financial leverage risk premium. The
22
coefficient on the market valued debt equity ratio in this example is 3.46% (8.46-5.0). The
23
relevered equity cost at the 50:50 debt equity ratio would then be 11.92%. So someone believing
45
The only reason for the cost of debt to increase is the risk of financial distress or bankruptcy, which
M&M ignored in their 1958 paper. Therefore, Solomon’s result is inconsistent with the M&M
assumptions. However, it is consistent with a model of bankruptcy and financial distress.
46
Note that the Mainline is claiming that its JSDs are really 30% equity and 70% debt.
87
1
in M&M (1958) would use a coefficient on the debt equity ratio of 3.46%. Further if they
2
believed that the equity cost estimated from a sample of firms with lower market valued debt
3
ratios underestimated the financial risk at the regulated firm’s debt ratio, they would increase the
4
11.0% by 92 basis points.
5
If instead the M&M (1963) with taxes equation (11) is used with a 50% tax rate, the unlevered
6
equity cost is higher at 9.39% and the financial leverage risk premium is only 1.61%. Since the
7
risk impact of financial leverage is offset in part by the tax advantages attached to debt, the
8
financial leverage risk premium is only half what it is with the flat WACC M&M 1958 model. In
9
this case the coefficient on the market valued debt equity ratio is 1.7% ((8.46-5.0)*.5).
10
Relevering to the 50% debt ratio increases the equity cost to 11.56% or 36 basis points less than
11
by using the flat WACC M&M 1958 model. Believing in M&M (1963) gives a smaller bump to
12
the ROE estimate.
13
Believing in a flat WACC gives a WACC and unlevered equity cost of a constant 7.4%.47 Hence
14
the market valued debt equity ratio is multiplied by (7.4-2.5) or approximately 5.0%. This is
15
higher than either M&M (1958) no tax or M&M (1963) with tax and givers the highest possible
16
leverage adjustment. This is because the debt cost is after tax and there are no explicit offsetting
17
costs in the model, yet the WACC is somehow held constant. Using this model the leverage
18
adjustment would not be 36 or 92 basis points but 131 basis points to move the equity cost at the
19
regulated debt ratio to 12.31%.
20
Let me make the importance of this example clear. The chain of events is that the risk of the
21
utility has declined causing its equity cost to drop from 15% to 11%. The obvious thing that the
22
regulator should do is cut the allowed ROE from 15% to 11%. This is also what would happen if
23
the regulator used the EUB’s ATWACCBV approach and recognised that it would be “derelict” in
24
using market values to rubber stamp this increase in market value. However, using the
25
“(AT)WACC approach” avoids this drop in the ROE in two ways. The first is to go directly to
26
the WACC with market values, which seals in the higher equity ratio and delays the drop in the
27
allowed ROE. However, if this fails, as it has before this Board, as well as the EUB, the second
47
7.4%= 11%*0.577 + 5%(1-.5)*0.423
88
1
step is to argue for a leverage adjustment. Then the assumption of a flat ATWACC generates the
2
biggest coefficient on the debt equity ratio and the largest financial leverage risk premium. This
3
in turn provides the biggest “bump” when a sample estimate is applied to the regulated common
4
equity ratio. In my example it would give an equity cost of 12.31%, 131 basis points higher than
5
the true equity cost and higher than using any other equity cost model that I am aware of. As the
6
example shows these assumed leverage adjustments can be very large and they are totally
7
unnecessary.
8
Q.
DRS. KOLBE AND VILBERT ARE NOT MAKING ROE ADJUSTMENTS BUT
EQUITY RATIO ADJUSTMENTS, DO THESE CONCLUSIONS STILL APPLY?
9
Yes. Both the equity cost and equity ratio results both flow from a rearrangement of the
10
A.
11
same equations; it is simply easier to “see” them using the equity cost approach since that is what
12
most people are interested in. Note that Dr. Kolbe’s critical equation is B-4 in his appendix B-19,
13
which is the equation he recommends in the main body of his evidence. This equation is derived
14
from his equations B-3a,b,c on page B-10, as he points out in his footnote 66. It is useful to go
15
through how this is derived.
16
First equation B-3a indicates that the WACC, which he denotes as r*3 is equal to the all equity
17
cost of capital for the firm, which he denotes ra3 and is more commonly referred to as the
18
unlevered equity cost (my K0) minus the tax advantages to debt, which he sets equal to zero. So
19
in equation B-3a, Dr. Kolbe is specifically assuming that the WACC is constant, regardless of
20
how much debt the firm uses. Note that this assumption is not that the WACC is constant around
21
a small optimal range, but that it is constant throughout the whole range of debt ratios.48
22
Equation B-3b is then the normal WACC which Dr. Vilbert estimates.
23
Second Dr. Kolbe assumes that this WACC is equal to the WACC for the regulated firm using
24
both book value weights and the allowed ROE on book equity. He thus gets
48
Dr. Kolbe states that the flat ATWACC doesn’t hold at extremes but mathematically his equation B-3a
states the opposite.
89
1
Ke
E
B
S
D
+ (1 − T ) K d = ROE + K d (1 − T )
V
V
A
A
(13)
2
where I have been consistent with my own notation and use E and B for the market values of
3
equity and debt and S and D for their book values; V is total firm value and B total assets. I then
4
have the equity cost as Ke and the allowed ROE. Dr. Kolbe then assumes that the after tax cost of
5
debt is the same, ie., the embedded equals the market cost. Noting that the debt ratio is just 1
6
minus the equity ratio (13) can then be rearranged as
7
( K e − (1 − T ) K d )
E
S
+ K d (1 − T ) = ( ROE − K d (1 − T )) + K d (1 − T )
V
A
8
The assumption that embedded and market after tax debt costs are the same means that this cost
9
on both sides can be cancelled and the equation rearranged as
10
S E ( K e − K d (1 − T ))
= *
A V ( ROE − K d (1 − T ))
(14)
11
This is Dr. Kolbe’s equation B-4 where the book equity ratio (S/A) is estimated from the sample
12
market valued equity ratio (E/V) and the estimated debt and equity costs and regulated ROE.
13
Note that (14) is derived from (13) and the same assumption of a flat ATWACC that I used to
14
derive equation (12). That is, it is the same assumption of a constant WACC that Dr. Kolbe uses
15
to make his equity ratio recommendation that produces the very high equity cost leverage
16
adjustment in equation (12). Note that in the example it produced an extra 131 basis points above
17
the assumed fair 11.0% ROE. Further it is the same assumption that the EUB criticized in
18
Decision U99099 (page 301)
19
20
21
22
23
“the Board considers that beta and the cost of equity do not change to the extent
necessary for an ATWACC, determined from market capitalization weights, to remain
constant when applied to the book capitalization for a pure play regulated utility. The
increase required to the cost of equity to achieve a constant ATWACC would be
excessive and violate the fair return standard.”
24
Consequently since Dr. Kolbe’s common equity ratio recommendation is derived from the same
25
assumption and equations as his equity cost recommendations before the EUB and this Board,
90
1
the comments of the EUB also apply to the way he derived his recommended common equity
2
ratio in this hearing. To accept his approach would in the words of the EUB be a “dereliction” of
3
duty on the part of the regulator.
4
Q.
YOU HAVE DISCUSSED THE IMPORTANCE OF THE MARKET VALUE
5
ASSUMPTIONS OF THE WACC AND THEIR IMPLICATIONS. HAVE YOU
6
ANY OTHER COMMENTS ON THE USE OF MARKET VALUES?
As I have stressed in the financial flexibility discussion, in the final analysis “fair” is
7
A.
8
determined in the stock market by the reaction of investors. If Board policies were not “fair” we
9
would see several things. First of all I would expect to see holding companies “ring fencing”
10
their regulated operations and selling parts to the stock market. For example, if TransCanada
11
genuinely believed that the Mainline’s allowed ROE and common equity ratio produced results
12
that were not fair, I would recommend that it sell say 20% to the stock market and establish a
13
public float. If the stock market agreed with TransCanada we would then see a market to book
14
ratio below 1.0 and that would provide powerful ammunition to support a higher equity ratio or
15
allowed ROE. However, I have seen very few ROE regulated utilities establishing public floats
16
in fact the reverse has happened as Consumers Gas, Island Tel, Maritime Electric etc have all
17
ceased to exist as public entities. The elimination of publicly traded pure regulated utilities is a
18
telling sign that the allowed financial parameters are too generous since under stand alone
19
regulation there should be few synergies with other corporate entities.
20
Second I would expect to see the market to book ratios of regulated utilities selling below 1.0. In
21
Schedule 19 is a table of earned ROEs, preferred stock yields and market to book ratios for a
22
sample of ROE regulated Telcos up until 1996.49 This sort of data was previously included by
23
Professor Berkowitz and I in risk premium estimates over preferred stock yields filed before this
24
Board in RH-2-94 as Appendix E. These risk premium estimates were then consistent with the
25
previously noted tax impact on dividend yields in Canada, which produces a tax preference for
26
low risk high yield utility shares.
49
Source data is from my paper, The Importance of Market to Book Ratios in Regulation, NRRI
Quarterly Bulletin, Winter 1997.
91
1
Note that for the period 1970-1983 market to book ratios for this sample of Telcos hovered
2
around 1.0 and at times were significantly below 1.0, as the combination of high inflation,
3
historic test years and regulatory lag exposed them to significant risk. As interest rates fell from
4
the early 1980s highs, the market to book ratios increased significantly, as allowed ROEs were
5
not cut sufficiently to reflect these market changes as well as the reduction in risk generated by
6
more regulatory protection.
7
The major point is that these Telcos were at that time relatively undiversified ROE regulated
8
utilities and observing the market to book ratio provided a valid way of assessing how investors
9
reacted to allowed ROEs. Dr. Kolbe makes essentially the same point,50 when he stated (page 27)
“
11
12
13
14
15
16
Equation (2.3) of course reflects strong simplifying assumptions. But the qualitative
conclusions we draw from it hold in most cases. If regulators allow the firm to expect to
earn its cost of capital, market value will equal book value (ROR=r implies ROR/r –
MV/BV =1, so that MV = BV). Conversely, if we observe MV=BV, we conclude that
investors expect regulators to allow the firm to earn its cost of capital, at least on average.
(MV=BV implies ROR/r =1, so that ROR=r).”
17
At the time that Dr. Kolbe’s book was published in 1984 the market to book ratios of regulated
18
firms in the US (his graph on page 32 of the text) were below 1.0. As a result, this observation
19
worked to increase allowed ROEs, as did my observation from the data in Schedule 17 that Telco
20
market to books in Canada were below 1.0 on many occasions in the period 1970-1982. Since
21
1983 market to books for regulated firms have generally been significantly above 1.0 as the data
22
in Schedule 19 indicates. By the mid 1990s the sample average market to book ratio for these
23
Telcos was well above any cushion allowed for flotation costs, so that the reverse held true:
24
allowed were too high.
25
In Schedule 20 is a graph of the market to book ratios for a sample of Canadian utility holding
26
companies (UHCs). The key implication is that, except for PNG, the market to book ratios are all
27
well above 1.0. For PNG it is clear that despite the efforts of the BCUC to reduce PNG's risk, the
50
(
MV
ROR
)=(
) . (2.3)
BV
r
10
L. Kolbe, J. Read and G. Hall, The Cost of Capital, Charles River Associates, MIT Press, 1984.
92
1
market is still sceptical of the company's long run prospects. These market to book ratios include
2
to a differing degree the impact of non-regulated operations, but there is a clear indication that
3
none of these companies have suffered any loss of financial flexibility as regulation has evolved
4
over the last ten years, since there is no obvious trend in the market to book ratios.
5
Further we have direct evidence of the value of regulated assets from sales of these assets
6
between firms. Unlike the UHC data in Schedule 20 these observations are not contaminated by
7
non-regulated assets. For example,
8
9
10
11
12
•
TCPL purchased the 50% of Foothills that it did not own at a market to book of
1.6 based on the common equity. Foothills, of course, accesses the WCSB, earns
the NEB formula and has 30% common equity. Moreover since TCPL already
owned 50% of Foothills the number of potential buyers was limited, which
reduced the price.
13
14
•
Aquila purchased TransAlta’s distribution and retail business at a market to book
of 1.5 based on a total rate base of $472mm (premium of $238mm);
15
16
•
Fortis purchased Aquila’s Alberta interests for a premium of $215mm over a rate
base of $601mm.
17
18
•
AltaLink purchased TransAlta’s transmission business for a $200mm premium
over a rate base of $644mm.
19
20
•
Hydro Quebec recently announced a $266mm gain on the sale of its interest in the
gas distribution assets in the province.
21
Note that in all these cases, the market to book ratio based on the equity is much greater than that
22
based on the total rate business, since the debt is normally assumed and is valued at close to its
23
book value. For example in Fortis’ purchases from Aquila it paid $1.3 billion for total rate base
24
assets of $943mm (in Alberta and elsewhere) for an overall premium of $357mm over rate base
25
and an overall market to book of 1.38X. However, it assumed the existing debt which was 60%
26
of rate base so effectively Fortis assumed about $565.8mm in debt and paid $734.2mm for the
27
book equity, so the market to book based on equity was about 1.96X. The final transaction value
28
depends on closing transactions but the point is that the market to book based on the common
29
equity was well above the indicated values based on total rate base.
30
Overall these observations on market to books are a significant indication that regulated assets in
31
Canada are worth well above their regulated book values. This observation means that $1 in
93
1
equity reinvested in rate base is immediately worth much more almost $2 as in the case of
2
Fortis’s purchase of Aquila’s assets. Further as Dr. Kolbe pointed out, and consistent with basic
3
financial theory, this means that utility allowed ROEs and common equity ratios in Canada are
4
excessive and rates include equity charges that are more than fair and reasonable. I therefore, see
5
no reason to add leverage or equity ratio adjustments that further compound these already
6
generous (excessive) charges.
7
Q.
CAN YOU QUICKLY SUMMARISE YOUR TESTIMONY?
8
A.
Yes. The Board should reject outright Drs. Kolbe and Vilbert’s implicit ATWACC
9
approach to setting the allowed common equity ratio. This is simply a back door way of
10
introducing an approach that the Alberta EUB has rejected as representing “dereliction” of duty
11
on their part if accepted. Further the approach of Dr. Kolbe and Vilbert is based on assumptions
12
that are incompatible with the way that markets work with the result that they come up with
13
“leverage“ adjustments that apart from being unnecessary are also extreme as the EUB also
14
noted.
15
Instead I would recommend that the Board continue with its existing policy of setting common
16
equity ratios based on business risk and financial market access. Here the Board should be aware
17
that by changing the Mainline’s depreciation rate it has reduced the Mainline’s exposure to
18
longer term risks. This effect is simply a reflection of the fact that less of the Mainline’s market
19
value is exposed to these longer term risks than at the time of RH-4-2001. In my judgment the
20
Board’s clear willingness to revisit issues such as depreciation in the future, means that many of
21
these longer term risks will not materialise anyway, or by the time that they materialize, will
22
have a negligible impact on the Mainline. Consequently, it is my judgment that the Mainline will
23
continue to earn its allowed ROE in the future in the way that it has undeniably done in the past
24
(plus a persistent over–earn).
25
Given that the Mainline refunded its preferred shares in 1998 and that this was recognized at the
26
time of RH-4-2001, and that it has subsequently been financed with 67% debt, including JSDs,
27
and 33% equity with no noticeable impact on its spread over equivalent term long Canadas, I can
28
see no reason to adjust its common equity ratio for any possible removal of the JSDs. The fact
94
1
that market to books for regulated assets are so high confirms the previous conclusions,
2
indicating that the Mainline allowed ROE and common equity ratio are generous.
3
Q.
DOES THIS CONCLUDE YOUR TESTIMONY?
4
A.
Yes.
5
95
1
Schedule 1
2
Pipeline Income Variability (RH-2-94)
3
(Schedule 19 in the Part B testimony of Drs. Booth and Berkowitz)
P IP ELIN E 'S ALL OWED VE RSU S ACTU AL ROE
1993
1992
1991
1990
1989
AN G
allowed
actual
deviation
12.00
12.00
0.0 0
12.00
12.00
0 .00
13.25
13.25
0 .00
13.25
13.25
0 .0 0
13.25
13.25
0.0 0
F OOTH ILLS
allowed
actual
deviation
11.73
11.73
0.0 0
13.83
13.83
0 .00
14.25
14.25
0 .00
14.25
14.25
0 .0 0
14.25
14.25
0.0 0
T QM
allowed
actual
deviation
1993
12.25
12.66
0.4 1
1992
13.75
14.62
0 .87
1991
13.75
14.44
0 .69
1990
13.75
14.87
1 .2 2
1989
13.75
14.79
1.2 4
IP L
allowed
actual
deviation
12.50
12.30
-0.20
12.50
14.40
1 .90
13.25
14.10
0 .85
13.25
14.40
1 .1 5
13.25
11.20
-2.0 5
T CP L
allowed
actual
deviation
12.25
12.42
0.1 7
13.25
13.43
0 .18
13.50
13.65
0 .15
13.25
13.34
0 .0 9
12.75
14.59
1.8 4
T MP
allowed
actual
deviation
11.50
11.32
-0.18
12.50
15.63
3 .13
14.00
16.65
2 .65
14.00
12.45
-1.55
14.00
17.64
3.6 4
WEI
allowed
actual
deviation
12.25
12.30
-0 .05
12.50
12.50
0 .00
13.75
13.60
-0 .15
13.25
13.10
-0.15
13.75
13.90
0.1 5
N OVA
allowed
actual
deviation
11.75
11.68
-0.07
12.50
12.72
0 .22
13.75
13.80
0 .05
13.65
13.66
0 .0 1
13.44
13.39
-0.0 5
SOURCE:
4
CBRS METH ODOLOGY OF RATIN G D E BT S ECU RITIES O F REGU LATED
U T ILITIE S, CBRS, May 1994, except for TQM, Surveillance Reports.
96
Schedule 2
NEB Pipelines Actual vs Allowed ROEs
(Confirmed by TCPL CAPP 31(a)
EARNED ROE vs ALLOWED
1990
1991
1992
1993
1994
1995
1996
1997
1998
1999
2000
2001
2002
2003
ovrearn
Allowed
13.25
13.5
13.25
12.25
11.25
12.25
11.25
10.67
10.21
9.58
9.9
9.61
9.53
9.79
11.16
TCPL
Actual
13.34
13.65
13.43
12.31
11.16
12.56
11.83
11.15
10.63
9.64
9.99
10.01
9.95
10.18
11.42
0.25
Allowed
14.25
14.25
13.83
11.73
11.5
12.25
11.25
10.67
10.21
9.58
9.9
9.61
9.53
9.79
11.31
Foothills
Actual
14.25
14.25
13.83
11.73
11.5
12.25
11.25
10.67
10.21
9.58
9.9
9.61
9.53
9.79
11.31
0.00
1
TCPL BC (ANG)
Allowed
Actual
Allowed
13.25
13.25
13.75
13.38
13.38
13.75
13.43
13.43
13.75
12.08
12.08
12.25
12
12
12.25
12.25
12.25
12.25
11.25
11.25
11.25
10.67
10.67
10.67
10.21
10.21
10.21
9.58
9.58
9.58
9.9
9.9
9.9
9.61
6.86
9.61
9.53
9.53
9.53
9.79
8.21
9.79
11.21
10.90
11.32
-0.31
TQM
Actual
14.87
13.94
13.97
12.5
12.55
12.65
11.83
10.94
10.32
9.94
9.96
10.21
9.8
10.21
11.69
0.37
Schedule 3
Earned vs Allowed ROEs
1990
1991
1992
1993
1994
1995
1996
1997
1998
1999
2000
2001
2002
ovrearn
Allowed
13.25
13.13
13.13
12.30
11.60
11.65
11.88
11.50
10.30
9.51
9.73
9.54
9.66
11.32
EGDI
Actual
13.60
13.29
13.40
14.43
12.49
12.66
13.14
13.00
11.97
10.77
10.83
10.03
11.81
12.42
1.10
Allowed
13.50
13.50
13.00
12.50
11.75
11.75
11.75
11.00
10.44
9.61
9.95
9.95
9.95
11.56
UNION
Actual
13.80
13.40
12.50
13.70
14.30
12.14
12.12
12.52
12.26
11.14
10.10
10.11
11.45
12.34
0.78
Allowed
13.25
13.50
12.50
11.75
11.75
11.50
11.25
10.67
10.21
9.58
9.90
9.61
9.53
11.29
NGTL
Actual
15.54
15.08
12.80
11.86
11.31
11.10
12.10
12.90
12.60
13.20
12.40
12.80
12.90
12.81
1.52
NGTL data is from CAPP-NGTL-17 where the allowed ROE is as given or the TCPL allowed ROE for the period when NGTL was not subject to an allowed ROE. The data
for EGDI and Union is taken from Appendix B Schedule 10 of the pre-filed testimony of Dr. William Cannon in RP-2002-0158, the Ontario Energy Board’s review of its ROE
guidelines.
2
Schedule 4
Earned Utility Holding Company (UHC) ROEs
CU Ltd
1993
1994
1995
1996
1997
1998
1999
2000
2001
2002
STDEV
13.37
13.71
14.12
14.86
14.87
14.75
14.54
15.44
14.96
17.56
1.15
Emera
12.02
11.90
11.55
10.59
10.56
9.47
10.83
10.88
10.58
6.65
1.55
Enbridge Fortis
17.53
9.59
16.91
14.47
14.04
13.25
13.35
15.65
14.90
10.11
2.59
11.84
10.71
10.74
9.61
9.43
7.16
8.56
9.71
12.25
11.87
1.61
PNG
GMI
19.29
19.73
19.50
19.91
18.91
19.11
17.66
17.93
17.45
18.91
0.87
3
12.92
13.44
11.77
13.32
13.32
10.14
10.79
9.75
7.50
5.94
2.61
Terasen
10.82
7.24
8.51
17.59
8.34
12.09
13.35
15.16
10.26
9.59
3.28
TransAlta TCPL
16.00
15.10
14.00
13.24
12.84
16.41
4.88
8.14
7.23
2.31
4.99
14.01
12.86
13.20
12.33
11.25
7.04
7.42
8.44
10.89
11.93
2.47
Mainline Foothills
12.31
11.16
12.56
11.83
11.15
10.63
9.64
9.99
10.01
9.95
1.05
11.73
11.50
12.25
11.25
10.67
10.21
9.58
9.90
9.61
9.53
1.00
Schedule 5
Variability in Earned UHC ROEs
CULtd
1993
1.64
1994
2.21
1995
1.87
1996
3.61
1997
4.20
1998
4.54
1999
4.96
2000
5.54
2001
5.35
2002
8.03
DEV
4.20
STDEV 1.97
Emera Enbridge
0.29
5.80
0.40
-1.91
-0.70
4.66
-0.66
3.22
-0.11
3.37
-0.74
3.04
1.25
3.77
0.98
5.75
0.97
5.29
-2.88
0.58
-0.12
3.36
1.22
2.42
Fortis
0.11
-0.79
-1.51
-1.64
-1.24
-3.05
-1.02
-0.19
2.64
2.34
-0.44
1.77
PNG
1.19
1.94
-0.48
2.07
2.65
-0.07
1.21
-0.15
-2.11
-3.59
0.27
1.97
GMI
7.56
8.23
7.25
8.66
8.24
8.90
8.08
8.03
7.84
9.38
8.22
0.63
4
Terasen TransAlta TCPL Mainline Foothills
-0.91
4.27
2.28
0.58
0.00
-4.26
3.60
1.36
-0.34
0.00
-3.74
1.75
0.95
0.31
0.00
6.34
1.99
1.08
0.58
0.00
-2.33
2.17
0.58
0.48
0.00
1.88
6.20
-3.17
0.42
0.00
3.77
-4.70
-2.16
0.06
0.00
5.26
-1.76
-1.46
0.09
0.00
0.65
-2.38
1.28
0.40
0.00
0.06
-7.22
2.40
0.42
0.00
0.67
0.39
0.31
0.30
0.00
3.65
4.25
1.91
0.29
0.00
Schedule 6
Increased Future Uncertainty in Earning Allowed ROE
Stable
Rate base
Return
Depreciation
Total
PV factor at 10%
PV Income years 1-10
PV Income years 11-25
PV return of capital
Market Value
1
100
10
4
14
0.91
52.29
11.40
36.3
100.0
2
96
9.6
4
13.6
0.83
3
92
9.2
4
13.2
0.75
4
88
8.8
4
12.8
0.68
5
84
8.4
4
12.4
0.62
6
80
8
4
12
0.56
7
76
7.6
4
11.6
0.51
8
72
7.2
4
11.2
0.47
9
68
6.8
4
10.8
0.42
10
64
6.4
4
10.4
0.39
11
60
6
4
10
0.35
12
56
5.6
4
9.6
0.32
13
52
5.2
4
9.2
0.29
14
48
4.8
4
8.8
0.26
15
44
4.4
4
8.4
0.24
16
40
4
4
8
0.22
17
36
3.6
4
7.6
0.20
18
32
3.2
4
7.2
0.18
19
28
2.8
4
6.8
0.16
20
24
2.4
4
6.4
0.15
21
20
2
4
6
0.14
22
16
1.6
4
5.6
0.12
23
12
1.2
4
5.2
0.11
24
8
0.8
4
4.8
0.10
25
4
0.4
4
4.4
0.09
0.42
0.39
0.21
0.19
0.16
0.14
0.12
0.11
0.09
0.08
0.07
0.06
0.05
0.05
0.04
0.03
0.03
Increased risk for years 6-10 correct discount rate for these years is 15% and market believes Board will not respond
PV Factor
PV Income years 1-10
PV Income years 11-25
PV return of capital
Market Value
0.91
52.29
6.03
36.3
94.63
0.83
0.75
0.68
0.62
0.56
0.51
0.47
5
Schedule 7
Increased Future Uncertainty in Earning Allowed ROE (Stronger basin)
Stable: 35 year life
1
2
3
4
5
6
7
8
rate base
100.00 97.14 94.29 91.43 88.57 85.71 82.86 80.00
Return
10.00
9.71
9.43
9.14
8.86
8.57
8.29
8.00
Depreciation
2.86
2.86
2.86
2.86
2.86
2.86
2.86
2.86
Total
12.86 12.57 12.29 12.00 11.71 11.43 11.14 10.86
PV Factor
0.91
0.83
0.75
0.68
0.62
0.56
0.51
0.47
PV return of capital
27.6
PV Income years 1-10
54.9
PV Income years 11-35
17.5
Market Value
100.0
Increased risk for years 11-35 correct discount rate for these years is 15% and market believes Board will not respond
PV Factor
0.91
0.83
0.75
0.68
0.62
0.56
0.51
0.47
PV return of capital
27.6
PV Income years 1-10
54.9
PV Income years 11-35
8.7
Market Value
91.2
9
77.14
7.71
2.86
10.57
0.42
10
74.29
7.43
2.86
10.29
0.39
11
71.43
7.14
2.86
10.00
0.35
12
68.57
6.86
2.86
9.71
0.32
13
65.71
6.57
2.86
9.43
0.29
14
62.86
6.29
2.86
9.14
0.26
15
60.00
6.00
2.86
8.86
0.24
16
57.14
5.71
2.86
8.57
0.22
17
54.29
5.43
2.86
8.29
0.20
18
51.43
5.14
2.86
8.00
0.18
19
48.57
4.86
2.86
7.71
0.16
20
45.71
4.57
2.86
7.43
0.15
21
42.86
4.29
2.86
7.14
0.14
22
40.00
4.00
2.86
6.86
0.12
23
37.14
3.71
2.86
6.57
0.11
24
34.29
3.43
2.86
6.29
0.10
25
31.43
3.14
2.86
6.00
0.09
26
28.57
2.86
2.86
5.71
0.08
27
25.71
2.57
2.86
5.43
0.08
28
22.86
2.29
2.86
5.14
0.07
29
20.00
2.00
2.86
4.86
0.06
30
17.14
1.71
2.86
4.57
0.06
31
14.29
1.43
2.86
4.29
0.05
32
11.43
1.14
2.86
4.00
0.05
33
8.57
0.86
2.86
3.71
0.04
34
5.71
0.57
2.86
3.43
0.04
35
2.86
0.29
2.86
3.14
0.04
0.42
0.39
0.21
0.19
0.16
0.14
0.12
0.11
0.09
0.08
0.07
0.06
0.05
0.05
0.04
0.03
0.03
0.03
0.02
0.02
0.02
0.02
0.01
0.01
0.01
0.01
0.01
6
19
90
M
19 01
91
M
19 01
92
M
19 01
93
M
19 01
94
M
19 01
95
M
19 01
96
M
19 01
97
M
19 01
98
M
19 01
99
M
20 01
00
M
20 01
01
M
20 01
02
M
20 01
03
M
20 01
04
M
01
Schedule 8
Interest Rates and Inflation
16
14
12
10
8
6
4
2
0
-2
T Bills
7
LTC
CPI
Schedule 9
Canadi an M CI
15
10
5
0
1990 J
1993 J
1996 J
-5
-10
-15
8
1999 J
2002 J
9
Jan-04
Jan-03
Jan-02
Jan-01
Jan-00
Jan-99
Jan-98
Jan-97
Jan-96
Jan-95
Jan-94
Schedule 10
US $ Value of FX Rate
0.8
0.75
0.7
0.65
0.6
0.55
0.5
19
94
-Q
3
19
95
-Q
2
19
96
-Q
1
19
96
-Q
4
19
97
-Q
3
19
98
-Q
2
19
99
-Q
1
19
99
-Q
4
20
00
-Q
3
20
01
-Q
2
20
02
-Q
1
20
02
-Q
4
20
03
-Q
3
Schedule 11
Capacity Utilisation
88
86
84
82
80
78
76
Manufacture
10
Non-farm
19
61
Q
19 1
64
Q
19 1
67
Q
19 1
70
Q
19 1
73
Q
19 1
76
Q
19 1
79
Q
19 1
82
Q
19 1
85
Q
19 1
88
Q
19 1
91
Q
19 1
94
Q
19 1
97
Q
20 1
00
Q
20 1
03
Q
1
Schedule 12
After tax profits (% GDP)
10
9
8
7
6
5
4
3
2
1
0
11
12
2004M08
2004M03
2003M10
2003M05
2002M12
2002M07
2002M02
2001M09
2001M04
2000M11
2000M06
2000M01
1999M08
1999M03
1998M10
1998M05
1997M12
Schedule 13
TSX Composite
12000
11000
10000
9000
8000
7000
6000
5000
13
A
BBB
10
2003
4
7
10
2000
4
7
10
1997
4
7
10
1994
4
7
10
1991
4
7
10
1988
Schedule 14
Spreads since 1988
6
5
4
3
2
1
0
14
2003Q3
2001Q1
1998Q3
1996Q1
1993Q3
1991Q1
1988Q3
1986Q1
1983Q3
1981Q1
1978Q3
1976Q1
1973Q3
1971Q1
1968Q3
1966Q1
1963Q3
1961Q1
Schedule 15
Government Lending (% GDP)
6
4
2
0
-2
-4
-6
-8
-10
-12
Schedule 16
Security Issues
Federal Crown
Federal
Provincial
Provincial Crown
Municipal
Total Public
1993
1994
1255.7
2533
43927
57522
14435
9030.85
3064.4
1088.1
693.03
1911.93
63375.13 72085.88
1995
1996
1997
1998
1999
2000
1550
207
1765
3661.25 5310.45
2917.5
48940
56535
43900
39000
42500
43000
8133
8936.4 10443.83 17193.99 17374.63 20993.25
3079.5
2361.04
2029
3533.02
3222.1 5091.47
2481.97
2479.1
2548.54
2972.97 2382.45 2497.38
64184.47 70518.54 60686.37 66361.23 70789.63 74499.6
Equities
29778.67
19081.2
18123.48 24368.73 28943.08 21928.64 21201.86 23200.01 16707.22 17971.48 22887.33
Private Debt
13419.72
9309.51
10438.31
114.69
520.72
0
0
118.14
411.03
0
Limited partnerships
Trusts
Other
Total
Public share
106688.2 100997.3
0.59
0.71
14654.6
2001
2002
2003
8513.75 20249.26 18946.14
39150
41150 41820.12
15803.1
14571 20859.96
3514.7
6184 5001.99
2819.55 3229.79 3854.96
69801.1 85384.05 90483.17
19457.4 26617.04 34701.19 39223.04 39822.19
407.7
1172.07
4264.3 10306.57
0
0
93275.43 114213.9 120565.5
0.69
0.62
0.50
Source Investment Dealers Association
15
690.33
1822.81
0
376.57
1498.08
0
211.63
2805.98
3140
32373.5 54141.67
516.93
636.32 2011.91
6996.72 11096.76 17286.93
1750 2100.01 1650.01
117420 128567.3 143080.3 135594.2 149562.1
0.57
0.55
0.52
0.51
0.57
188461
0.48
Schedule 17
Business Trust Information
Market CAP
Entertainment
Energy Infrastructure
Restaurants
Resources
Marketing
Oil and Gas Services
Power
Printing
Other
Business Trust Average
Yield
Payout
805
4070
785
10120
3605
1275
2225
4785
2040
9.5
7.1
9.7
7.4
7.2
7.4
8.1
7.6
9.7
7.3
1000
410
1060
2410
1415
660
1190
145
9
7.3
8.1
6.5
9.4
9.4
8.9
8.4
94
68
97
81
106
73
92
88
96
88
Energy Infrastructure
AltaGas Income Fund
Enbridge Income Fund
Fort Chicago
Gas Metro
Inter Pipeline
KeySpan Facilities
Pembina Income Fund
Taylor NGL
Source: RBC Capital Markets Business Trust Weekly September 3, 2004
Schedule 18
Yield Spreads on Utility/Pipeline Debt
(Source RBC Capital markets: Spread history of Canadian, Corporate and Government Issuers May 31, 2004)
1999
CU Inc
Maritime and NE Pipe
GMI
Newfoundland Power
Enbridge Pipe
TCPL
Enbridge Gas
Alliance
TQ&M
Epcor Utilities
Fortis
Nova Scotia Power
Westcoast
Union Gas
Emera
Terasen Gas
Average
S&P
A
A
A
AAAABBB+
BBB+
BBB+
BBB
BBB+
BBB
BBB
BBB
BBB
Date
2019
2019
2009
2022
2009
2026
2027
2015
2009
2029
2010
2009
2010
2025
2006
2029
2000
2001
DEC
64
DEC
110
DEC
130
70
115
63
130
93
133
90
87
100
140
90
169
130
159
120
169
225
108
107
149
90
91
150
138
90
150
80
131
136
141
105
191
235
80
88
126
100
155
129
70
2002
DEC
104
146
80
185
65
187
120
157
105
234
275
145
130
162
165
175
152
2003
MAR
120
123
80
170
70
201
135
161
105
205
250
135
150
176
165
182
152
JUN
115
138
75
160
70
157
125
108
105
182
180
125
115
127
130
169
130
2004
SEPT
100
138
70
160
70
141
100
87
85
150
180
95
80
126
115
145
115
DEC
95
119
50
130
47
100
83
95
65
140
115
72
70
95
115
129
95
MARCH
104
85
50
130
47
110
86
97
65
129
105
72
80
102
65
128
91
JUNE
107
90
50
110
49
118
103
79
49
135
105
60
92
120
70
129
92
SEPT
108
85
50
110
49
119
99
94
49
135
105
60
78
122
70
141
92
Sschedule 19
RETURN ON EQUITY AND MARKET TO BOOK RATIO
TELCO ROE
TELCO M/B*
PREF YIELD
SPREAD
1970
9.63
0.97
7.42
2.21
1971
11.00
1.07
6.98
4.02
1972
11.83
1.12
7.00
4.83
1973
11.46
1.01
7.26
4.20
1974
9.94
0.86
8.90
1.04
1975
11.80
0.84
9.48
2.32
1976
12.84
0.93
9.28
3.56
1977
13.37
1.06
8.39
4.98
1978
13.43
1.17
8.34
5.09
1979
14.09
1.19
8.64
5.45
1980
13.68
1.05
9.89
3.79
1981
14.06
0.92
12.02
2.04
1982
15.08
0.91
13.78
1.30
1983
15.58
1.16
10.16
5.42
1984
14.82
1.24
9.89
4.93
1985
14.11
1.39
9.26
4.85
1986
13.16
1.41
8.92
4.24
1987
13.03
1.31
8.51
4.52
1988
12.90
1.27
8.37
4.60
1989
12.79
1.32
8.46
4.33
1990
12.68
1.26
9.20
3.48
1991
12.72
1.34
8.54
4.18
1992
12.41
1.35
8.20
4.21
1993
11.98
1.41
7.73
4.25
1994
11.49
1.50
7.96
3.53
1995
10.25
1.33
7.76
2.49
1996
11.22
1.47
7.51
3.71
*
Average high low price divided by average book value per share.
Schedule 20
M ar k et t o Book R at i os of U H Cs
3
CU L
2.5
Em er a
Enbr i dge
2
F or t i s
GM I
1.5
PN G
1
T er asen
T AU
0.5
Aver age
0
1993
1994
1995
1996
1997
1998
1999
2000
2001
2002
APPENDICES TO WRITTEN REPLY EVIDENCE OF
A. LAWRENCE KOLBE
APPENDIX R-KOLBE1: KOLBE REPLY EVIDENCE IN RH-2-2004, PHASE II
(Page numbers in Kolbe Reply Evidence are from the original document)
122
NATIONAL ENERGY BOARD
IN THE MATTER OF the National Energy Board Act, R.S.C. 1985,
c. N-7, as amended, (Act) and the Regulations made under it; and
IN THE MATTER OF an Application by TransCanada PipeLines
Limited (TransCanada) pursuant for orders pursuant to Part IV of the
National Energy Board Act for approval of tolls for 2004.
WRITTEN REPLY EVIDENCE
OF
A. LAWRENCE KOLBE
FOR
TRANSCANADA PIPELINES LIMITED
The Brattle Group
44 Brattle Street
Cambridge, Massachusetts 02138
617.864.7900
November 2004
WRITTEN REPLY EVIDENCE OF
A. LAWRENCE KOLBE
Table of Contents
I.
INTRODUCTION AND SUMMARY . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
II.
UNFOUNDED RELIANCE ON MARKET-TO-BOOK RATIO . . . . . . . . . . . . . . . . . . . 6
A.
MUTUALLY CONTRADICTORY STATEMENTS ON STOCK MARKET
VALUES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
B.
INVALIDITY OF MARKET-TO-BOOK RATIO AS A TEST OF FAIR
RETURN . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8
1.
Implied Cost of Equity Values Far Too Low, Often Less Than the
Board’s Benchmark Interest Rate . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9
2.
Market-to-Book Ratio Test Inconsistent with the Way the Market
Behaves . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15
III.
FLAWED COMMENTS ON THE MAINLINE’S RISK . . . . . . . . . . . . . . . . . . . . . . . . 19
A.
INCORRECT RETURN ADEQUACY STANDARD . . . . . . . . . . . . . . . . . . . . 20
B.
WRONGLY FOCUSED NUMERICAL EXAMPLES THAT ACTUALLY
SUPPORT A MARKED INCREASE IN DEEMED EQUITY RATIO . . . . . . . 23
C.
OTHER PROBLEMS IN RISK-RETURN STATEMENTS . . . . . . . . . . . . . . . . 39
1.
Depreciation Adjustments No Substitute for Equity Return
Adjustments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39
2.
Declining Rate Base Not the Same as Declining Risk . . . . . . . . . . . . . . 41
3.
No Double-Counting if Recognize Impact of Future Risk Increases
Both Now and When They Occur . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 42
4.
Past Not Necessarily Prologue . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43
5.
“Soft” Capital Constraints Relevant in Practice . . . . . . . . . . . . . . . . . . . 45
IV.
FLAWED STATEMENTS ON CAPITAL STRUCTURE PRINCIPLES . . . . . . . . . . . 47
V.
CONCLUSION . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51
Appendix D: DETAILED RESPONSE ON CAPITAL STRUCTURE PRINCIPLES . . . . . D-1
A.
THE USE OF ATWACC FOR ONE PURPOSE NOT INCONSISTENT
WITH ITS USE FOR OTHER PURPOSES . . . . . . . . . . . . . . . . . . . . . . . . . . D-2
B.
BOOTH EVIDENCE NUMERICAL EXAMPLE INTERNALLY
INCONSISTENT, INCONSISTENT WITH THE WAY CAPITAL
MARKETS WORK, AND MISINTERPRETS MY
RECOMMENDATIONS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . D-4
C.
BOOTH EVIDENCE’S INADEQUATE AND INACCURATE REVIEW OF
THE CAPITAL STRUCTURE LITERATURE . . . . . . . . . . . . . . . . . . . . . . D-13
D.
BOOTH EVIDENCE RELIANCE ON SELECTED REGULATORY
DECISIONS RATHER THAN THE CAPITAL STRUCTURE
LITERATURE . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . D-17
E.
INCORRECT CLAIM THAT MY EVIDENCE RELIES ON THE 1977
MILLER MODEL . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . D-19
WRITTEN REPLY EVIDENCE OF
A. LAWRENCE KOLBE
F.
G.
H.
I.
J.
INCORRECT CLAIM THAT MY EQUATION FOR THE INTERACTION
BETWEEN THE COST OF EQUITY AND CAPITAL STRUCTURE
PRODUCES THE HIGHEST POSSIBLE CHANGES . . . . . . . . . . . . . . . . .
INCORRECT IMPLICATION THAT MY PROCEDURES IGNORE NONTAX COSTS TO DEBT . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
BOOTH EVIDENCE CHARACTERIZATION OF THE WAY
REGULATION WORKS INCONSISTENT WITH THE EVIDENCE . . . .
FINANCIAL RISK DEPENDENT ON MARKET VALUES, NOT BOOK
VALUES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
MY PROCEDURES TO CALCULATE APPROPRIATE DEEMED
EQUITY RATIOS REASONABLE . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
ii
D-21
D-22
D-24
D-26
D-30
WRITTEN REPLY EVIDENCE OF
A. LAWRENCE KOLBE
1
Appendix D: DETAILED RESPONSE ON CAPITAL STRUCTURE PRINCIPLES
2
Q1.
What is the purpose of this appendix?
3
A1.
This appendix provides the details of my response to the Booth Evidence’s comments on
4
capital structure principles.
5
Q2.
What specific issues does the appendix address?
6
A2.
It covers the ten points summarized in Section IV of my reply evidence in order.
7
Q3.
Do you wish to make any prefatory comments before you start?
8
A3.
Yes. I would note that the amended evidence filed by Dr. Vilbert and myself in this
9
proceeding focuses on the deemed equity ratio that corresponds to the Board’s current
10
9.56 percent formula rate of return on equity. To do so, we take advantage of decades of
11
financial research to address the issue quantitatively, not merely qualitatively. That
12
necessarily involves an analysis of what the research says about how the return on equity
13
and capital structure interact.
14
The Booth Evidence responds to this analysis with comments of its own, many
15
of them involving the overall weighted-average cost of capital. Here, to keep the
16
discussion from becoming unnecessarily complex, I respond to those comments as stated,
17
without constantly trying to translate both the Booth Evidence’s comments and my reply
18
into deemed equity ratio terms. That should not be misread as implying that I am
19
anything other than totally committed to addressing the issue the Board has set forth in
20
this proceeding, i.e., to finding the appropriate deemed equity ratio given the current
21
value of the Board’s formula rate of return on equity.
WRITTEN REPLY EVIDENCE OF
A. LAWRENCE KOLBE
1
2
A.
THE USE OF ATWACC FOR ONE PURPOSE NOT INCONSISTENT
WITH ITS USE FOR OTHER PURPOSES
3
Q4.
What is the first issue you discuss here?
4
A4.
The Booth Evidence at p. 64 notes that firms use the after-tax weighted-average cost of
5
capital (“ATWACC”)35 as a discount rate to try to maximize the value of the firm. At p.
6
3-4 and 70, it suggests this somehow invalidates the use of ATWACC (and implicitly the
7
principles associated with it) by regulators, because regulators’ goal should not be to
8
maximize the value of the firm. For example, at p. 3 it states:
9
10
11
12
13
14
15
In terms of the ATWACC approach advocated and used implicitly in the
company’s filed testimony I would point out the fundamental
contradiction in its use in regulatory filings in that it is the mirror image
of shareholder value maximisation. That is, earning more than the WACC
is synonymous with the creation of shareholder value, whereas the
Board’s responsibility is not to create or maintain shareholder value, but
to ensure that rates are fair and reasonable.
16
Q5.
What is your reaction to this view?
17
A5.
I have two reactions.
18
First, I would note again that while I do advocate that the Board approach the task
19
of setting a deemed equity ratio by taking full advantage of the last 50 years of financial
35
A word about terminology: the ATWACC is the market-value weighted average of the cost of equity and
the after-tax cost of debt. The Booth Evidence often uses the uses the conventional corporate finance
term “WACC” to refer to the overall cost of capital calculated with market-value weights. Depending
on whether its discussion at that time includes or excludes taxes, it may or may not correspond to the
ATWACC as I use the term. However, I have not found a place in the part of the evidence that discusses
capital structure principles in which it corresponds to the conventional regulatory weighted-average cost
of capital. In this appendix, for convenience I sometimes use “WACC” in the sense of the Booth
Evidence, and I never use it in the sense of the traditional regulatory weighted-average cost of capital.
D-2
WRITTEN REPLY EVIDENCE OF
A. LAWRENCE KOLBE
1
research, I do not advocate the use of ATWACC as a rate of return standard in this
2
proceeding.
3
Second, with apologies to Prof. Booth, the quoted statement simply makes no
4
sense at all. The ATWACC is a tool, not a result. It is as though someone said,
5
“unregulated firms use hammers to try to build houses. Regulators want to build garages,
6
so they shouldn’t make use of hammers, because if they use hammers, they will
7
automatically build houses.” There is no contradiction at all in unregulated firms using
8
ATWACC to try to maximize the value of the firm and regulators using the financial
9
research that underlies the ATWACC, among other things, to try to find the just-fair rate
10
of return. The output of a use of the ATWACC tool, like any other tool, depends on what
11
it is used to do, not what the tool itself is.36
12
Q6.
Can you put this point directly in regulatory terms?
13
A6.
Yes. To a firm making unregulated investments, projects must at least have a present
14
value of future cash flows equal to the project’s cost when discounted at the ATWACC.
15
Projects that generate present values higher than their costs are preferred. Regulation,
16
however, does not take the cash flows as determined by the market conditions facing the
17
investment, it determines the cash flows itself, by deciding on the revenue requirement.
18
In regulation, the ATWACC can be used to decide on the set of cash flows that provide
19
a present value just equal to the amount the regulated firm invests. The fact that
36
For the convenience of the reader, I note also that Appendix B of my direct evidence at p. B-32 disposes
of the incorrect but possibly related view that the use of market-value weights to calculate the ATWACC
somehow automatically translates book-value regulation into market-value regulation.
D-3
WRITTEN REPLY EVIDENCE OF
A. LAWRENCE KOLBE
1
unregulated firms use the ATWACC differently in no way precludes regulators from
2
using it for their purposes, too.37
3
4
5
B.
BOOTH EVIDENCE NUMERICAL EXAMPLE INTERNALLY
INCONSISTENT, INCONSISTENT WITH THE WAY CAPITAL
MARKETS WORK, AND MISINTERPRETS MY RECOMMENDATIONS
6
Q7.
What numerical example do you discuss in this section?
7
A7.
The example begins on p. 62 and serves general expository purposes. The part I need to
8
address begins on p. 67, where debt is added. It ignores taxes for simplicity and assumes
9
investors price a regulated firm’s equity as a perpetuity, that is, as the present value of a
10
constant annual sum of money that investors believe will come in at the same level
11
forever. It sets the amount of money that equity gets as the difference between the overall
12
return on the assets and the amount that gets paid out for interest expense. This leads to
13
the formula,
Value of
Equity
14
15
16
17
=
(Return On × (Asset !
(Cost of × (Debt
Investment) Value)
Debt)
Value)
Cost of Equity
or in the Booth Evidence’s notation (p. 68),
18
19
E
=
ROI × A ! Kb x B
Ke
20
where asset value, A, equals the book value of utility assets, which initially equals the
21
sum of the market values of equity, E, plus debt, B. It initially postulates that the return
37
It could hardly be otherwise, given that regulators in Australia, New Zealand and the U.K., which have
come late to the model of private ownership with public oversight have, to my knowledge, universally
adopted ATWACC as the appropriate rate of return standard.
D-4
WRITTEN REPLY EVIDENCE OF
A. LAWRENCE KOLBE
1
on investment, ROI, is 10 percent, the value of the assets, A, is $10 million, the cost of
2
debt, Kb, is 5 percent, the amount of debt, B, is $5 million, and the cost of equity, Ke, is
3
15 percent. This gives a value of equity equal to,38
=
ROI × A ! Kb x B
Ke
6
7
=
0.10 × $10 million ! 0.05 × $5 million
0.15
8
9
=
$1 million ! $ 0.25 million
0.15
10
=
$5 million
E
4
5
=
$0.75 million
0.15
11
If I adopt the convention the Booth Evidence uses of “mm” for million, the overall
12
weighted-average cost of capital (“WACC”) is the weighted average of the costs of equity
13
and debt, or
14
WACC
=
Ke × [E/(E+B)]
+
Kb × [B/(E+B)]
15
=
0.15 × ($5mm/$10mm) + 0.05 × ($5mm/$10mm)
16
=
0.15 × 0.5 + 0.05 × 0.5 = 0.10 = 10 percent.
17
Starting on p. 70, the Booth Evidence postulates that a change in regulatory rules
18
drops the cost of equity from 15 percent to 11 percent. It does not, however, explain
19
directly at what capital structure this 11 percent is supposed to obtain, and alternative
38
See p. 68 of the Booth Evidence, which reaches this result but which need not spell out the steps in this
level of detail because some of them were explained earlier.
D-5
WRITTEN REPLY EVIDENCE OF
A. LAWRENCE KOLBE
1
interpretations of that key fact would produce different criticisms of its example. In what
2
follows, I take my lead from its statement at p. 71,
3
4
5
6
7
The obvious thing for the regulator to do is simply get expert opinion
estimating the equity cost, and if this is unbiased, notice and cut the
allowed ROE to 11%. The equity value will then return to $5mm and the
stockholders will continue to earn their fair return on their $5mm
investment.
8
I therefore interpret the 11 percent to be the cost of equity that obtains at a market-value
9
capital structure of 50-50 equity-debt. In that case, the true new WACC in its example
10
is
11
True WACC =
Ke × [E/(E+B)]
+
Kb × [B/(E+B)]
12
=
0.11 × ($5mm/$10mm) + 0.05 × ($5mm/$10mm)
13
=
0.11 × 0.5 + 0.05 × 0.5 = 0.08 = 8 percent.
14
From this point through p. 73, its example postulates a series of partial
15
adjustments by regulators that are internally inconsistent with the underlying assumptions
16
of the model. It also mischaracterizes the way capital markets work and the results that
17
would follow if regulators used the WACC to set the overall returns.
18
Q8.
Please explain what you mean in the previous paragraph.
19
A8.
There are two main problems with its example. The first, which is the less important
20
problem, is that it assumes investors price the stock as though the current cash flow would
21
be unchanged forever, yet its example imagines a series of annual changes by regulators
22
that gradually converge on a new equilibrium WACC equal to 8 percent. It is hard to
D-6
WRITTEN REPLY EVIDENCE OF
A. LAWRENCE KOLBE
1
imagine that investors would not catch on that their valuation model was woefully
2
inadequate at this point, since each year of the adjustment they would bear a major capital
3
loss, cumulatively undoing the major gain they received when the cost of equity initially
4
fell from 15 to 11 percent.
5
More fundamental is its failure to recognize that if its imagined gradual
6
adjustment process nonetheless took place as postulated, the actual cost of equity would
7
be below 11 percent until equilibrium was restored. Its example’s assumption that the
8
cost of equity would remain at 11 percent regardless of the market-value capital structure
9
of the company in questions is false.
10
Q9.
Please show how this affects its steps.
11
A9.
First (on p. 71) it calculates the new value of equity on the assumption that equity
12
investors expect the same $0.75mm forever, but now discount it at 11 percent:
=
ROI × A ! Kb × B
Ke
15
16
=
0.10 × $10mm ! 0.05 × $5mm
0.11
17
18
=
$1mm ! $ 0.25mm
0.11
19
=
$6.818mm
13
14
20
E
=
$0.75mm
0.11
It calculates the new WACC using this value of equity, which it finds to be 8.46 percent:
D-7
WRITTEN REPLY EVIDENCE OF
A. LAWRENCE KOLBE
1
Booth WACC =
Ke × [E/(E+B)]
+
Kb × [B/(E+B)]
2
=
0.11 × ($6.8mm/$11.8mm) + 0.05 × ($5mm/$11.8mm)
3
=
0.11 × 0.577 + 0.05 × 0.423 = 0.0846 = 8.46 percent.
4
It imagines regulators setting the ROI equal to 8.46 percent for the next year, which drops
5
the “perpetual” equity cash flow to (ROI x A - Kb x B) = (0.0846 x $10mm - 0.05 x
6
$5mm) = $0.596mm. That produces a new, lower value of equity ($5.42mm, on p. 73),
7
a new, lower calculated WACC (8.12%, on p. 73), and so on, until eventually the
8
postulated true WACC of 8 percent is reached.
9
The key mistake in logic in this example takes place at p. 71, where its evidence
10
states that the first estimation of the new WACC is done “[a]ssuming there is no bias to
11
estimating the equity cost at 11%.” However, if there were no bias in estimating the
12
equity cost, the cost of equity actually estimated would be consistent with the true market
13
WACC, not some mistaken version. The 11 percent cost of equity is correct in the Booth
14
Evidence’s example at a 50-50 market-value capital structure. It would be different at a
15
different market-value capital structure. In particular, if for whatever reason the market
16
value of equity were initially the $6.818mm its example supposes, the true cost of equity
17
would be:
D-8
WRITTEN REPLY EVIDENCE OF
A. LAWRENCE KOLBE
=
True WACC ! Kb × [B/(E+B)]
[E/(E+B)]
3
4
=
0.08 ! 0.05 × [$5mm/($6.8mm+$5mm)]
[$6.8mm/($6.8mm+$5mm)]
5
6
=
0.08 ! 0.05 × [0.423]
[0.577]
7
8
=
0.08 ! 0.0211
0.0588
9
=
0.102 = 10.2 percent
True Ke
1
2
10
In that case, in the world of the Booth Evidence’s example, the estimated WACC would
11
equal 8 percent the first time out,39 and regulators would converge on the right answer
12
after one false start, based on the market’s initial misestimation of the new value of equity
13
at $6.818 million.40
14
Based on this flawed example, the Booth Evidence mischaracterizes the situation
15
in several ways.
16
Q10. In what ways does the Booth Evidence mischaracterize the situation?
17
A10.
First, I reiterate that despite the focus on ATWACC regulation in the Booth Evidence,
18
neither I nor TransCanada is recommending regulation based on ATWACC in this
19
proceeding. In some ways, the Booth Evidence’s whole example is addressed to a
20
recommendation that is irrelevant to this proceeding.
39
That is, if the cost of equity is estimated without bias, Estimated WACC = True Ke × [E/(E+B)] + Kb
× [B/(E+B)] = 0.102 × 0.577 + 0.05 × 0.423 = 0.08 = 8 percent, the true WACC.
40
If the market immediately recognized that regulators were about to change the allowed rate of return to
the new true WACC, 8 percent, the market value of equity would fall to $5 million and the cost of equity
if estimated without bias would be the postulated 11 percent immediately, under this valuation model.
D-9
WRITTEN REPLY EVIDENCE OF
A. LAWRENCE KOLBE
1
More fundamental is the claim at p. 71 that its example would produce an
2
estimated ATWACC of 8.46 percent the first year. That statement wrongly assumes the
3
market would set the cost of equity at 11 percent despite the fact that the company in
4
question has much less financial risk at a 57.7-42.3 equity-debt ratio than it does at a 50-
5
50 equity-debt ratio. The market will recognize the actual level of financial risk,
6
however, whether the Booth Evidence’s example does or not. In the world of its
7
example, the true cost of equity is 10.2 percent at the postulated 57.7-42.3 equity-debt
8
ratio. If the true cost of equity were estimated without bias, the estimate of the WACC
9
would be 8 percent the first time, not 8.46 percent.
10
Next, the Booth Evidence states at pp. 72-73,
11
12
13
14
15
16
17
18
19
20
21
22
23
24
Although the ROI is reduced from 10% to 8.46%, it is not reduced to the
correct ROI of 8.0%,38 so the equity market value is still $0.42mm higher
than it needs to be. The reason for this is that using market value weights
in the WACC puts greater emphasis on the higher equity cost than the
debt cost. For non-regulated firms this is correct since the objective is to
maintain these market values and create wealth. However, it is totally
incorrect for a regulator who is tasked with awarding fair allowed returns
and implicitly causing market values to drop when allowed ROEs are too
high. By estimating and applying a market based WACC the effect of the
higher allowed ROE is perpetuated by its impact on the higher equity
market value.
25
(Italicized emphasis in the original.)
38
The correct regulated WACC is the average of the debt and equity costs using
regulated book value weights, in this case 50%.
26
This statement contains five errors:
27
28
•
First, the estimated WACC will be 8 percent, not 8.46 percent, because the market
will recognize the effect of financial risk on the cost of equity;
D-10
WRITTEN REPLY EVIDENCE OF
A. LAWRENCE KOLBE
1
2
3
•
Second, that means the greater equity weight in the Booth Evidence’s example
(57.7 versus 50 percent) is exactly offset by a lower market cost of equity (10.2
versus 11 percent), so there is no net “greater emphasis on the higher equity cost.”
4
5
6
7
8
9
•
Third, the use of market-value weights in the WACC has nothing to do with
trying to “maintain these market values and create wealth,” whether for
unregulated or regulated firms.41 The use of market-value weights is required
because the cost of equity is also based on the level of financial risk implied by
market values. Use of any other weights simply mis-estimates the overall risk and
cost of capital of the firm.
10
11
12
•
Fourth, since the true market-based WACC is the same 8 percent that the Booth
Evidence’s example postulates as correct, a regulatory body that did use this
method of setting overall returns would not allow an excessive ROE.
13
14
15
16
17
18
19
•
Regarding the footnote, the correct cost of equity at the regulated 50-50 bookvalue weights is 11 percent, which does reach the correct value of 8 percent for
the WACC, because at those weights the cost of equity is 11 percent. It is higher
because of the greater financial risk at that capital structure. But that is not the
only way to calculate the correct overall return. The same WACC results from
use of the 10.2 percent cost of equity at the market-value equity-debt ratio of
57.7-42.3.
20
Finally, the Booth Evidence concludes at p. 73,
21
22
23
24
25
The basic insight from this discussion is that by using market values in
WACC, the resulting cost of capital is higher than a fair return, since the
higher equity cost is given a greater weight. Further if the firm is a pure
ROE regulated utility it tends to “rubberstamp” the use of market values
and is contrary to “fair and reasonable” regulation.
26
Of course, this conclusion does not follow because an accurate estimate of the new cost
27
of equity would be 10.2 percent, not the postulated 11 percent value that would be
41
This conclusion may be related to the Booth Evidence’s erroneous statements that use of the ATWACC
tool is per se an attempt to maximize a company’s market values. The previous part of this appendix
explains the fallacy in that view. There is nothing inconsistent about an unregulated company’s using
the ATWACC to find the projects that offer the most valuable net cash flows in its market, while at the
same time regulators use the ATWACC to limit the cash flows that a regulated firm can earn. The
ATWACC is merely a tool, and the output of using the tool depends on the task for which it is employed.
D-11
WRITTEN REPLY EVIDENCE OF
A. LAWRENCE KOLBE
1
appropriate with a 50-50 equity-debt ratio. Additionally, if regulation used a 50-50
2
equity-debt ratio, the cost of equity that goes with the correctly estimated 8 percent
3
WACC would be the 11 percent the Booth Evidence’s example originally postulated.
4
There would be no inconsistency and no excess return on equity.
5
Q11. Please sum up.
6
A11.
The Booth Evidence’s numerical example postulates regulation according to a
7
recommendation I am not making in this case (i.e., regulation based directly on the
8
overall cost of capital, rather than accepting the Board’s formula rate of return on equity
9
and finding the appropriate deemed equity ratio). It assumes the market will not
10
recognize the true level of financial risk the company bears. In reality, it is the Booth
11
Evidence’s example that does not recognize the actual level of financial risk. Were I
12
making the WACC-based regulation recommendation in its example, and were it
13
implemented based on a cost of equity estimated without bias, that recommendation
14
would work without providing any excess compensation to investors in the world of the
15
Booth Evidence’s example, even if equityholders initially overestimated the market value
16
of equity. That is, the Booth Evidence complains that the WACC approach under the
17
conditions of its example will not get to the right answer in one step, while its
18
recommendation would. But if the WACC approach were implemented without bias, it,
19
too, would get to the right answer in one step. Finally, the use of market-value weights
20
to calculate the WACC is equally appropriate for regulators and unregulated firms, and
21
it leads to biases in neither application.
D-12
WRITTEN REPLY EVIDENCE OF
A. LAWRENCE KOLBE
1
2
C.
3
BOOTH EVIDENCE’S INADEQUATE AND INACCURATE REVIEW OF
THE CAPITAL STRUCTURE LITERATURE
Q12. What support for its views does the Booth Evidence draw on from the financial
4
literature?
5
A12.
Very little. The discussion covers parts of pp. 81-87. It starts with an apparent claim that
6
book rather than market values cover financial risk, which echos statements made early
7
in the Booth Evidence. This claim is not sourced to the economic literature, despite an
8
explicit invitation for CAPP to do so.42 I address this part of the discussion in Section I,
9
below.
10
Q13. Please summarize the Booth Evidence’s review of the capital structure literature.
11
A13.
The financial literature the Booth Evidence cites is sparse. It mentions the 1958 and 1963
12
papers by Modigliani and Miller, the 1977 paper by Miller, a 1963 article by Ezra
13
Solomon, and a 1963 publication Gordon Donaldson.43 It identifies no study of capital
14
structure that has been published in the last quarter-century, despite the fact that this has
15
been and remains a very active field of research.44
42
See the Booth Response to TCPL Information Request No. 9.
43
Actually, the passage does not include formal citations for any of the publications. The first three papers
are among many that are cited in my own Appendix B. I am unaware of a relevant 1963 citation to
Donaldson, although perhaps it is a version of Gordon Donaldson, Corporate Debt Capacity, Boston:
Division of Research, Graduate School of Business Administration, Harvard University (1961). The
Solomon paper is Ezra Solomon, “Leverage and the Cost of Capital,” The Journal of Finance 18:273-79
(May 1963).
44
An apparent but not real exception: it cites the 1991 Canadian edition of Brealey and Myers regarding
the Adjusted Present Value (“APV”) approach to calculation of the value of the tax advantage to debt.
However, Prof. Myers developed the APV approach in 1974: S. C. Myers, “Interactions of Corporate
Financing and Investment Decisions,” The Journal of Finance 29:1-25 (March 1974).
D-13
WRITTEN REPLY EVIDENCE OF
A. LAWRENCE KOLBE
1
The literature summary itself consists of a brief overview of the tax-based theories
2
of Modigliani and Miller and Miller, a mention of the static tradeoff model (referring to
3
calculations using APV), a mention of the pecking order hypothesis (attributing it to Prof.
4
Donaldson), and a mention that the Solomon paper showed that the cost of equity does
5
not increase as fast if the cost of debt increases when firms add debt.
6
Q14. Why do you label this review “inadequate”?
7
A14.
The last paper it mentions was published twenty-six years ago. It reviews none of the
8
more recent literature explicitly, despite the fact that Prof. Booth himself has published
9
a relevant paper that I cite in my own Appendix B. It does not address any of the
10
empirical studies I cite as support for my conclusions on the actual effects of capital
11
structure. Rather than directly addressing the findings of the research I report, it states
12
Prof. Booth’s personal views and appeals to regulatory decisions rather than the economic
13
literature.
14
Q15. Why do you label the review of the financial literature in the Booth Evidence
15
16
“inaccurate”?
A15.
At p. 87 it cites the Solomon paper as establishing that “if the debt cost, Kb, increases with
17
the debt equity ratio then the equity cost does not increase so fast...”. In fact, the
18
Solomon paper purported to show that the Modigliani-Miller theory is wrong because the
19
cost of equity could actually fall as the debt ratio rose (because the cost of debt was going
D-14
WRITTEN REPLY EVIDENCE OF
A. LAWRENCE KOLBE
1
up so fast) in a no-tax, pure Modigliani-Miller world. However, the paper was shown to
2
be in error in this regard two years after it was published.45
3
Even though the paper has long been known to be erroneous, the Booth Evidence
4
not only cites it, it appears to continue to rely on it at pp. 86-87:
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
The reason for the very large leverage adjustment in equation (12) is that
the model is internally inconsistent. [The equation that shows how the cost
of equity changes with the debt ratio in the Miller (1977) model] and the
flat WACC assumes the tax deductibility of interest which causes the
WACC to fall, but there is no explicit account of the offsetting costs that
negate this to keep the WACC constant. For example, if the WACC is
constant it could be that as the market valued debt equity ratio increases
the debt cost also increases due to the higher risk of insolvency and the
costs of financial distress and bankruptcy. This would be particularly true
as the firm goes to very high debt equity ratios. In this case, what is
keeping the WACC constant is an increasing Kb as creditors protect
themselves from the insolvency risk attached to highly debt financed
firms. From the spread date in Schedule 16, we know this happens.
Moreover, it is obvious from equation (12) that if the debt cost, Kb,
increases with the debt equity ratio then the equity cost does not increase
so fast, which is what Solomon showed in the Journal of Finance in
1963.45 The intuition is simply that “debt” in highly debt financed firms
has some of the same characteristics as equity.
28
This passage errs in several ways. First, it relies on an incorrect paper. Second,
29
it appears to copy that paper’s error in attributing maintenance of a flat ATWACC to an
30
increase in the cost of debt rather than to an increase in both the cost of debt and the cost
45
The only reason for the cost of debt to increase is the risk of financial distress
or bankruptcy, which M&M ignored in their 1958 paper. Therefore, Solomon’s
result is inconsistent with the M&M assumptions. However, it is consistent with
a model of bankruptcy and financial distress.
45
See Alexander A. Robichek and Stewart C. Myers, Optimal Financing Decisions, Englewood Cliffs, NJ:
Prentice Hall, Inc. (1965), pp. 34-36 and 48-49. The Solomon paper erred in assuming that the cost of
debt could ever exceed the overall cost of capital.
D-15
WRITTEN REPLY EVIDENCE OF
A. LAWRENCE KOLBE
1
of equity.46 The cost of debt may well increase as the market-value debt ratio increases,
2
but the cost of equity increases as well, even without taxes and with a flat ATWACC over
3
the whole range of capital structures.47 Third, the Solomon result does not address a
4
model of overall financial distress (which would increase the overall cost of capital), it
5
focuses on whether the overall cost of capital will truly remain flat in a no-tax
6
Modigliani-Miller world if the cost of debt increases as the debt ratio increases.48
7
Finally, its conclusion that the Miller model is “internally inconsistent” is not
8
adequately explained. If it means the Miller model would not hold with risky debt at high
9
debt ratios, that is true but in no way contradicts my evidence. (See Sections E and F,
10
below.) If it means that Prof. Miller somehow made a major mistake in one of the papers
11
that won him a Nobel Prize, the Booth Evidence needs to provide more detail to support
12
such a claim than it has in this short passage.49
46
The Booth Evidence’s statement that “the equity cost does not increase so fast” is somewhat ambiguous
and might mean that the cost of equity increases faster than the cost of debt, but not as fast as it would
if the cost of debt were constant. Yet if the Booth Evidence did not mean the statement in the sense of
the Solomon paper, in which WACC stays flat as the cost of debt increases because the cost of equity
eventually decreases, why juxtapose the comment with the Solomon article?
47
This is what Robichek and Myers, op. cit., showed.
48
For completeness, I would note that the facts that the risk of financial distress looms and the cost of debt
increases as the debt ratio increases are in no way inconsistent with my evidence. My Appendix B goes
well beyond the early models addressed in the Booth Evidence, to include explicit consideration of the
costs of excessive debt. Additionally, Dr. Vilbert and I explicitly considered the effect of changes in the
cost of debt as capital structure changes and concluded that it would be a mistake to reflect such changes
in this context. The current cost of debt is the correct input to our calculations, not that which would exist
under different conditions.
49
In addition to its reliance on the disproved Solomon paper and its unexplained criticism of the Miller
(1977) paper, the Booth Evidence cites an illustration of the APV method in the above-cited Brealey and
Myers textbook as evidence that “In this case, Professor Myers, who Dr. Kolbe references throughout
his testimony, clearly believes in the tax advantages of debt.” One can only assume that Prof. Booth’s
tongue is firmly in his cheek in this quotation, since textbooks routinely use simplified numerical
examples to teach a technique without holding that the results of the example are perfectly general. For
(continued...)
D-16
WRITTEN REPLY EVIDENCE OF
A. LAWRENCE KOLBE
1
2
3
D.
BOOTH EVIDENCE RELIANCE ON SELECTED REGULATORY
DECISIONS RATHER THAN THE CAPITAL STRUCTURE
LITERATURE
4
Q16. On what regulatory decisions does the Booth Evidence rely?
5
A16.
It has numerous citations in this area to the Alberta Energy and Utilities Board
6
(“AEUB”), as well as some to this Board. It appears to be particularly fond of a citation
7
from AEUB decision U99099, which it says on p. 74 states that the AEUB concluded it
8
would be “derelict in its statutory responsibilities to recognize market capitalization ratios
9
that are derived from a market value capitalization that deviates from the intrinsic long-
10
run value of the regulated firm.”
11
Q17. What is your reaction to these citations?
12
A17.
I have four comments. First, I would submit that if the Booth Evidence could refute my
13
conclusions by citing scholarly research performed by financial economists, it should
14
have done so. My evidence includes an extensive discussion of the financial literature,
15
which I believe solidly supports my procedures. The Booth Evidence does not challenge
16
that evidence on its own terms.
17
Second, Section IV of my Appendix B, which runs from p. B-24 to B-32,
18
explicitly discusses the issues that arose in the U99099 decision, which appears to have
19
been based in part on what amounted to expert evidence introduced for the first time in
49
(...continued)
the record, I explicitly asked Prof. Myers about the above quotation from the Booth Evidence, and he
indicated that it was not his view that debt conveyed significant tax advantages and he did not see how
such a claim [if made seriously] could be squared with his publications.
D-17
WRITTEN REPLY EVIDENCE OF
A. LAWRENCE KOLBE
1
post-hearing argument. The Booth Evidence does not mention that discussion at all. I
2
believe that that discussion addresses the principal concerns raised by the AEUB.
3
Third, I would note that the “intrinsic long-run value” cited in the above quotation
4
is book value, as indicated by another quotation to the U99099 decision on p. 74 of the
5
Booth Evidence:
6
7
8
9
10
11
12
“The Board observes that the intrinsic long-run value of a pure play
regulated entity is best represented by book value. In other words, the
present worth of future regulated earnings, discounted at the allowed
return, is by definition equal to book value assuming achieved regulated
earnings on average equal allowed regulated earnings. Accordingly, the
Board considers that book capitalization represents the best indicator of
the long-run market capitalization for a pure play regulated firm.”
13
Section II of this reply evidence shows that the market-to-book ratio test of the adequacy
14
of a utility’s returns is inconsistent with any reasonable value for its cost of equity and
15
has been disproved by actual market behavior.
16
Finally, I would note that if regulatory decisions by other bodies rather than the
17
findings of the economic literature are to be the arbiter of this Board’s decisions about
18
capital structure principles, there are plenty of examples that contradict those the Booth
19
Evidence cites.50 Other countries learned late that private ownership with public
50
This issue arose in the Alberta EUB’s generic cost of capital proceeding, and TCPL’s response to CAPP’s
Information Request 108 in that proceeding cites a number of such decisions. They include the following
decisions of the Australian Competition and Consumer Commission (“ACCC”), the primary federal rate
regulatory body. Please note that the ACCC uses procedures consistent with the capital structure
principles I discuss, including the use of market-value weights to assess financial risk, in its: Draft
Statement of Principles for the Regulation of Transmission Revenue (1999 and 2004), Final Decision on
Central West Pipeline (June 2000), Draft Decision on the Moomba to Adelaide Pipeline (August 2000),
Draft Decision on the Moomba to Sydney Pipeline (December 2000), and Draft Decision on the NT Gas
Pipeline (May 2001). Other examples of which I am aware include the Electricity Distribution Price
Determination 2001-05, Volume I, Statement of Purpose and Reasons, by the Office of the
Regulator-General (now the Essential Services Commission), Victoria (September 2000) and the Final
(continued...)
D-18
WRITTEN REPLY EVIDENCE OF
A. LAWRENCE KOLBE
1
oversight is more efficient than public ownership. As a result, they were able to design
2
regulatory procedures from scratch, with the full benefit of access to the modern financial
3
research and the experience of other countries in mind.51
4
5
E.
6
INCORRECT CLAIM THAT MY EVIDENCE RELIES ON THE 1977
MILLER MODEL
Q18. What does the Booth Evidence claim to be the basis of your capital structure
7
procedures?
8
A18.
9
At p. 85, the Booth Evidence claims that “the flat ATWACC approach of Drs. Kolbe and
Vilbert is essentially the 1958 M&M model as extended to include corporate and personal
10
taxes by Miller (1977).”
11
Q19. Is this claim correct?
12
A19.
13
No, and it is hard to see how my evidence could be so badly misread. The 1977 Miller
model is a purely tax-based model in which the personal tax disadvantage of debt fully
50
(...continued)
Decision on the Proposed Access Arrangement for the Dampier to Bunbury Natural Gas Pipeline,
Independent Gas Pipelines Access Regulator Western Australia, 23 May 2003. In New Zealand, the
Commerce Commission has recognized the necessity to use market value weights quite explicitly for
many years. For example, the Treasury’s handbook, “Estimating the Cost of Capital for Crown Entities
and State-Owned Enterprises” (October 1997) recognizes the need for market weights. Recent regulatory
reports confirm that the Commerce Commission continues to use and advocate the use of market weights
for determining the WACC. See, for example, Final Report Part IV Inquiry into Airfield Activities at
Auckland, Wellington and Christchurch International Airports, 6 August 2002, Gas Control Inquiry,
Draft Framework Paper, July 16, 2003; and Regulation of Electricity Lines Businesses, Targeted Control
Regime, Draft Assessment and Inquiry Guidelines (Process and Analytical Framework), 7 August 2003.
51
Please note that I am not recommending that the Board adopt the regulatory frameworks these other
boards use, merely that it take notice of the fact that there are numerous regulatory decisions that reach
different conclusions regarding capital structure principles from those in the decisions that the Booth
Evidence cites.
D-19
WRITTEN REPLY EVIDENCE OF
A. LAWRENCE KOLBE
1
offsets the corporate tax advantage of debt, which is an important insight. However, my
2
discussion of the capital structure literature (unlike the Booth Evidence’s -- see the
3
section before last) goes well beyond that insight.52 The Booth Evidence never addresses
4
my actual conclusions on how capital structure affects firm value. Those who wish to
5
know my actual views may safely disregard the Booth Evidence’s discussion entirely, and
6
instead refer to my own direct evidence, particularly Section III.B and pp. B-11 to B-18
7
of Appendix B.
8
The passage in the evidence itself includes Figures 10 and 11, which should make
9
clear that I do not subscribe to the model implied by the 1977 Miller paper. That model
10
would plot straight lines for all four industries for both the values of the firm in Figure
11
10 and the associated values for the ATWACC in Figure 11. I would not say, for
12
example, that if a firm in Industry 4 used no debt at all, its ATWACC would be equal to
13
its value where the cost of equity is actually estimated. However, within a normal range
14
of capital structures for an industry, the financial research consistently shows that debt
15
does not affect value in any detectable way. That finding is consistent with the use of a
16
flat ATWACC within the normal range of industry capital structures.
52
Among other things, I calculate the net tax corporate over personal tax advantage to debt in Canada,
which is not consistent with that implied by the Miller (1977) paper. See pp. B-21 to B-24 of my
Appendix B.
D-20
WRITTEN REPLY EVIDENCE OF
A. LAWRENCE KOLBE
1
2
3
4
F.
Q20. How does the Booth Evidence characterize the equation you use to quantify the way
5
6
the cost of equity changes with debt?
A20.
7
8
9
It claims at pp. 85-86 that:
However, a flat ATWACC does have the advantage that it gives the
largest possible leverage effect, that is, the coefficient on the market
valued debt equity ratio in the equity cost equation is as large as possible.
10
11
INCORRECT CLAIM THAT MY EQUATION FOR THE INTERACTION
BETWEEN THE COST OF EQUITY AND CAPITAL STRUCTURE
PRODUCES THE HIGHEST POSSIBLE CHANGES
A similar claim is made at p. 89.
Q21. Why does the Booth Evidence characterize the equation as having the “advantage”
12
of the largest possible leverage effect?
13
A21.
14
Q22. In any case, is the statement that the flat ATWACC equation provides the maximum
15
16
It doesn’t say.
possible leverage effect correct?
A22.
No, unless one considers only the papers mentioned in the Booth Evidence’s inadequate
17
review of the capital structure literature (see Section C above). It is clearly not true given
18
the overall findings of the literature.
19
Q23. Please explain.
D-21
WRITTEN REPLY EVIDENCE OF
A. LAWRENCE KOLBE
1
A23.
The last paper the Booth Evidence mentions is Prof. Miller’s 1977 Presidential Address
2
to the American Finance Association. Of the purely tax-based theories of capital
3
structure, the model implied by that paper does have the fastest changes in the cost of
4
equity as capital structure changes.
5
However, Figure 9 on p. 45 of my direct evidence and the associated discussion
6
on pp. 43-44 make clear that in reality, the cost of equity will sometimes increase faster
7
than the flat ATWACC model implies. This will happen whenever additional debt
8
decreases the value of the firm, i.e., anywhere to the right of the absolute minimum value
9
for the ATWACC curve. The Booth Evidence’s assertion to the contrary is simply
10
incorrect.
11
12
G.
13
INCORRECT IMPLICATION THAT MY PROCEDURES IGNORE NONTAX COSTS TO DEBT
Q24. What does the Booth Evidence say regarding non-tax costs of debt and the flat
14
ATWACC equation?
15
A24.
At pp. 16-17, it notes that the Board in RH-2-94 discussed the formal tax-based models
16
of capital structure and held that they ignored the non-tax costs of debt. It also states that
17
the cost of equity equation derived from the flat ATWACC equation gives the highest
18
possible leverage adjustment (a claim rebutted in the previous section) “because the debt
19
cost is after tax and there are no explicit offsetting costs in the model, yet the WACC is
20
somehow held constant.”53
53
Booth Evidence, p. 88.
D-22
WRITTEN REPLY EVIDENCE OF
A. LAWRENCE KOLBE
1
Q25. What is your reaction to these statements?
2
A25.
First, as noted above in Section E, my procedures do not rely on any of the purely tax-
3
based models. Detailed discussions of the non-tax costs of debt in my direct evidence and
4
Appendix B underlie the findings on which I rely. Therefore, the Board’s quoted
5
comment from RH-2-94 does not apply to my evidence.
6
The second statement is puzzling. If it refers to the 1977 Miller paper, Prof.
7
Miller’s insight was that if personal and corporate tax rates had the right values, the
8
personal tax disadvantage to debt could fully offset its corporate tax advantage. The
9
Miller (1977) version of the flat ATWACC equation is intrinsically a model of no net tax
10
advantage. The ATWACC is held constant precisely because debt has no net tax
11
advantage. There is no need for risky debt to achieve this outcome, only for the tax rate
12
conditions described at pp. B-7 to B-11 of my Appendix B (right before my discussion
13
of the non-tax effects of debt). The Booth Evidence’s statement seems to say that the
14
Miller model could not achieve this result without the assumption that debt is risky. If
15
so, that plainly is wrong.
16
Q26. Do your procedures imply that you assume no “offsetting costs” to the tax advantage
17
18
to debt?
A26.
No, not at all. As discussed above, much of my Appendix B directly addresses the fact
19
that the costs of excessive debt eventually offset the tax advantage to debt. As also noted,
20
part of that appendix calculates the maximum tax advantage debt could have in Canada
21
given actual corporate and personal tax rates. That net advantage remains positive,
22
although it is quite modest. If I were ignore the offsets to the tax advantage to debt, I
D-23
WRITTEN REPLY EVIDENCE OF
A. LAWRENCE KOLBE
1
would use a different equation to relate the cost of equity and capital structure, not the
2
one about which this passage complains.
3
Q27. Are you aware of any possible explanation for the Booth Evidence’s statements in
4
5
this passage?
A27.
It may have a procedure in mind that is different from the one we actually use. The very
6
next part of this discussion, in the same paragraph, returns to its mistaken numerical
7
example, which, as shown above, misinterprets both our procedure and the way capital
8
markets reflect financial risk in the cost of equity. It again (wrongly) suggests that the
9
procedures I recommend would produce a higher overall return than hypothesized as
10
correct in its example. But regardless of the reason, this passage remains puzzling. Read
11
literally as a statement about the Miller (1977) model, it is plainly incorrect.
12
13
H.
14
Q28. What does the Booth Evidence say about regulatory risk in the context of capital
15
16
17
18
19
20
21
22
23
24
BOOTH EVIDENCE CHARACTERIZATION OF THE WAY
REGULATION WORKS INCONSISTENT WITH THE EVIDENCE
structure principles?
A28.
At p. 79, it asserts,
In the example, the equity is obviously riskier at ... a market to book of
1.36X than it is at ... a market to book of 1.0X. This is because it is highly
unlikely that the regulator will cut the allowed ROE when the stock is
trading at book value. In contrast, the higher the market to book ratio the
more likely the regulator will cut the allowed ROE and thus the riskier the
stock. In contrast, Drs Ko[l]be and Vilbert would have us believe that the
equity in a regulated firm is less risky when it is trading at a market to
book well above 1.0, since the debt ratio is lower. I don’t accept this.
D-24
WRITTEN REPLY EVIDENCE OF
A. LAWRENCE KOLBE
1
Q29. Do you agree?
2
A29.
No, for several reasons. First, for this passage to be true, either investors would have to
3
be extremely short-sighted, or regulators would have to be about to behave in a way that
4
investors do not foresee. If investors predicted that regulators were about to start
5
managing the ratesetting process to focus on the market-to-book ratio, and particularly
6
on a market-to-book ratio of one, the market-to-book ratio would be close to one
7
already.54
8
Of course, this also means that the market-to-book ratios that the Booth Evidence
9
cites themselves imply that regulators have not been basing their rate of return decisions
10
on this fact up to now. Otherwise, investors would have predicted the coming action and
11
bid the market-to-book ratios down in advance.55 Thus, there is no evidence that
12
regulators do behave in the way this passage asserts they do, and as Section II of my reply
13
evidence shows, there is no reason that they should.
14
Q30. What about the part of the quotation that says you are incorrect to view the firm as
15
more risky at lower market-to-book ratios?
54
The inherent circularity of such a process would make regulation based on the market-to-book ratio
infeasible, even if we were still able to believe the market-to-book ratio test worked. See Section II of
the body of my reply evidence.
55
Also, as noted in Section II of my reply evidence, the market-to-book ratio levels the Booth Evidence
cites imply unreasonably low values for the cost of equity even if investors never expect regulators to
behave in the way that this Booth Evidence passage suggests investors should fear.
D-25
WRITTEN REPLY EVIDENCE OF
A. LAWRENCE KOLBE
1
A30.
First, that is not what I say. I (and a great deal of financial research) say that for the same
2
business risk, the firm will be more risky at a lower market-value equity ratio. It is the
3
Booth Evidence that links this statement to the market-to-book ratio, not me. Second,
4
even if the Booth Evidence’s characterization of the way rate regulation works were right,
5
and even if investors were too blind to see the cut coming, the reason the utility’s risk
6
would be higher in such circumstances would have nothing to do with financial leverage,
7
it would be based on regulatory risk. That says nothing about the impact of market-value
8
capital structure on the level of financial risk. The Booth Evidence’s conclusion is a
9
complete non sequitur.
10
11
12
I.
Q31. Apart from the market-to-book ratio passage, on what does the Booth Evidence’s
13
14
FINANCIAL RISK DEPENDENT ON MARKET VALUES, NOT BOOK
VALUES
discussion say financial risk depends?
A31.
The Booth Evidence addresses this issue in several places. Specifically, the Booth
15
Evidence at pp. 15-17 argues that the accounting relationships between book capital
16
structure and either book net income or the variability of book net income provide all the
17
information the Board needs to set capital structure. It makes a related statement on p.
18
81, where it says that the accounting relationship among the return on equity, the cost of
19
debt and capital structure says “nothing about how the stock market reacts to financial
20
risk, that is, how market values change, or how the cost of equity changes as the firm uses
21
debt.” (Italicized, boldface emphasis in the original.)
D-26
WRITTEN REPLY EVIDENCE OF
A. LAWRENCE KOLBE
1
Q32. What is your reaction to these statements?
2
A32.
It is hard to see how the accounting relationship among net income and interest expense
3
can have no impact at all on the stock market, since future equity cash flows will be
4
influenced by that interaction in future years. For example, is the Booth Evidence
5
claiming that the market will take no notice if a company has so much debt that it will
6
regularly have trouble making its interest payments out of operating income (i.e., will
7
have regularly have negative net income)? The lack of a one-to-one correspondence
8
between accounting values and market values does not mean quantities measured by
9
accounting values have no relationship at all with quantities measured by market values.
10
Nor is a view that financial risk is solely a function of book income statements
11
and balance sheets consistent with either the financial literature or everyday experience.56
12
The cost of equity depends on the risks equityholders bear. Financial risk is the extra risk
13
that equityholders bear when firms issue debt. The cost of equity is measured in the
14
market, not on the books. Therefore, a proper analysis of the impact of financial risk on
15
the returns equityholders require must measure that risk in the market, using market
16
values. The contrary view, that financial risk is a book phenomenon, is plainly contrary
17
to standard financial theory and practice.57
18
Q33. What other comments does the Booth Evidence make on this topic?
19
A33.
It says immediately after the quotation on risk and the market-to-book ratio, still on p. 79,
56
Recall the example in Section III of my amended direct evidence of how the variability of the market
value of the equity in a condo varied with the size of the mortgage.
57
Among other problems, book relationships cannot contain enough information properly to assess required
returns in the market, since the market looks ahead, not merely at current and past book relationships.
D-27
WRITTEN REPLY EVIDENCE OF
A. LAWRENCE KOLBE
1
2
3
4
5
Second and more significant, the financial leverage risk premium stems
from the imposition of fixed interest charges. That is, prior to receiving
their equity return the firm has to pay these interest charges. This risk
does not change as the market value of the firm changes; it only changes
as the book debt equity ratio changes.
6
Q34. Is this correct?
7
A34.
8
No, it is flatly wrong. The passage appears to forget that equityholders’ return comes in
the form of capital gains as well as current dividends.
9
Changes in market values are a major source of the realized risks equityholders
10
face. If a company has a market value of $10 million in equity and $10 million in debt,
11
or $20 million total, and if the value of the enterprise falls 10 percent to $18 million,
12
equity absorbs the (vast majority of)58 the $2 million loss. The rate of return on equity
13
is -20 percent (-$2 million/$10 million). However, suppose another, otherwise identical
14
company has a market value of equity of $15 million and of debt of $5 million, for the
15
same total enterprise value of $20 million. Suppose the same forces make the value of
16
that enterprise decline by 10 percent to $18 million, also. In this case, equity “only” must
17
absorb a 13.3 percent loss (-$2 million/$15 million), not 20 percent. Note also that the
18
impact on the equityholders of the two firms is unaffected by the firms’ book capital
19
structures, which we could assume to be identical without changing the example at all.
58
As the market-value equity ratio declines, decreases in the value of the enterprise begin to be absorbed
in part by debt, since there is less equity left to provide a cushion. However, the majority of the risk falls
on equity, and adding an adjustment for the impact on debt would complicate the example without
changing the conclusion.
D-28
WRITTEN REPLY EVIDENCE OF
A. LAWRENCE KOLBE
1
In short, changes in the market-value capital structure directly affect the
2
sensitivity of the equity return to changes in the value of the enterprise. The level of the
3
financial risk equityholders bear therefore depends directly on market values.59
4
Q35. Why didn’t you point this out in your direct evidence, so the Booth Evidence could
5
respond directly to it?
6
A35.
I did discuss this effect in my direct evidence, at pp. 30-33, particularly as illustrated in
7
Figure 6 on p. 32. The Booth Evidence does not address this discussion at all, it merely
8
asserts the contrary.
9
Q36. Does the Booth Evidence make an offer of proof for its contrary assertion?
10
A36.
11
12
13
14
15
16
It offers a quotation from Standard and Poor’s at p. 79-80:
Similarly ratios using market value of a company's equity in calculations
of leverage are given limited weight as analytical tools. The stock market
emphasises growth prospects and has a short time horizon; it is influenced
by changes in alternative investment opportunities and can be very
volatile. A company's ability to service its debt is not affected directly by
such factors [Emphasis added by the Booth Evidence.]
17
The Booth Evidence interprets this passage to imply,
18
19
20
That is, S&P is basically saying book value leverage is important, when
it is assessing the default or credit risk in debt, whereas market values
don't count, or at least don't count as much. If it is book values and interest
59
Note that this conclusion holds true regardless of the “true” model or models of stock prices. Whatever
the reason the market value of a firm’s assets might decline, the impact of that decline will fall (primarily)
on equity, not debt. The greater the proportion of debt in the market value of the assets, the bigger the
proportionate fall in the market value of the equity.
D-29
WRITTEN REPLY EVIDENCE OF
A. LAWRENCE KOLBE
1
2
3
payments that affect credit risk and the cost of debt then this is the risk
that also affects utility equity investors.
Q37. Is this a reasonable interpretation in the context of whether market values affect the
4
5
amount of financial risk equityholders bear?
A37.
No, the quotation could hardly be less relevant. First, the quoted passage does not say
6
anything at all about book-value leverage. That is the Booth Evidence’s addition (or at
7
the very least, it is based on material it does not quote with the rest of the passage).
8
Second, and more fundamentally, S&P in the part of this passage that the Booth
9
Evidence emphasizes is focusing on whether bondholders can count on the company to
10
service debt, not on the financial risk equityholders bear. The market-value volatility of
11
which the passage speaks definitely will affect equityholders, whether or not the company
12
can service its debt. The passage expresses no opinion at all about whether the amount
13
that equity would be affected would be different if the market debt-equity ratio were
14
different. That topic never comes up.
15
Therefore, the passage has no meaning at all for the issue of whether market value
16
capital structures affect the degree of financial risk equityholders bear.
17
18
J.
MY PROCEDURES TO CALCULATE APPROPRIATE DEEMED
EQUITY RATIOS REASONABLE
19
Q38. What does the Booth Evidence say about the fact that your procedures focus on the
20
deemed equity ratio, not the ATWACC and cost of equity it spends most of its time
21
discussing?
D-30
WRITTEN REPLY EVIDENCE OF
A. LAWRENCE KOLBE
1
A38.
At pp. 89-91, it shows that the deemed equity ratio equation on which my
2
recommendations risk is consistent with the finding that debt does not have a material
3
effect on the value of the firm within the normal range of capital structures for an
4
industry. This point is not in dispute, nor, given the views stated in my evidence, should
5
it be surprising to anyone.
6
It then essentially says that since it has found all the problems just discussed with
7
the ATWACC-cost of equity formulation of these principles, the Board should disregard
8
the deemed equity ratio formulation, too.
9
10
Q39. What is your reaction?
A39.
I would turn the point around. None of the criticisms in the Booth Evidence of my
11
procedures have any economic merit, nor are they supported by relevant citations to the
12
economic literature. Some of them flatly misstate my evidence. Given this, I would
13
respectfully suggest that the Board should disregard the Booth Evidence’s comments on
14
capital structure principles entirely.
15
Q40. Does this complete your appendix?
16
A40.
Yes, it does.
D-31
Table R-1
[a]
[b]
[c]
[d]
[e]
[f]
[g]
[h]
[i]
Assumptions
Allowed ROE, Stable
Allowed ROE, Risky
Discount Rate, Income, Stable
Discount Rate, Income, Risky
Discount Rate, Depreciation, Stable
Discount Rate, Depreciation, Risky
Remaining Life
Initial Rate Base
Year After Which Risk Shifts
[j] Years
10%
10%
10%
15%
10%
10%
25
100
10
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
[k] Rate Base
[l] Return Stable Years
[m] Return Risky Years
[n] Depreciation Stable Years
[o] Depreciation Risky Years
[p] Total
100.0
10.0
0.0
4.0
0.0
14.0
96.0
9.6
0.0
4.0
0.0
13.6
92.0
9.2
0.0
4.0
0.0
13.2
88.0
8.8
0.0
4.0
0.0
12.8
84.0
8.4
0.0
4.0
0.0
12.4
80.0
8.0
0.0
4.0
0.0
12.0
76.0
7.6
0.0
4.0
0.0
11.6
72.0
7.2
0.0
4.0
0.0
11.2
68.0
6.8
0.0
4.0
0.0
10.8
64.0
6.4
0.0
4.0
0.0
10.4
60.0
0.0
6.0
0.0
4.0
10.0
56.0
0.0
5.6
0.0
4.0
9.6
52.0
0.0
5.2
0.0
4.0
9.2
48.0
0.0
4.8
0.0
4.0
8.8
44.0
0.0
4.4
0.0
4.0
8.4
40.0
0.0
4.0
0.0
4.0
8.0
36.0
0.0
3.6
0.0
4.0
7.6
32.0
0.0
3.2
0.0
4.0
7.2
28.0
0.0
2.8
0.0
4.0
6.8
24.0
0.0
2.4
0.0
4.0
6.4
20.0
0.0
2.0
0.0
4.0
6.0
16.0
0.0
1.6
0.0
4.0
5.6
12.0
0.0
1.2
0.0
4.0
5.2
8.0
0.0
0.8
0.0
4.0
4.8
4.0
0.0
0.4
0.0
4.0
4.4
[q]
[r]
[s]
[t]
52.3
6.0
24.6
11.7
47.5
6.9
23.0
12.9
42.7
8.0
21.3
14.2
37.7
9.2
19.5
15.6
32.7
10.6
17.4
17.2
27.6
12.1
15.2
18.9
22.3
14.0
12.7
20.8
17.0
16.0
9.9
22.9
11.5
18.5
6.9
25.1
5.8
21.2
3.6
27.7
0.0
24.4
0.0
30.4
0.0
22.1
0.0
29.5
0.0
19.8
0.0
28.4
0.0
17.5
0.0
27.3
0.0
15.4
0.0
26.0
0.0
13.3
0.0
24.6
0.0
11.3
0.0
23.0
0.0
9.4
0.0
21.3
0.0
7.6
0.0
19.5
0.0
5.9
0.0
17.4
0.0
4.4
0.0
15.2
0.0
3.1
0.0
12.7
0.0
1.9
0.0
9.9
0.0
1.0
0.0
6.9
0.0
0.3
0.0
3.6
[u] Total Present Value
94.6
90.4
86.2
82.0
77.9
73.8
69.8
65.8
62.0
58.3
54.8
51.5
48.2
44.8
41.4
37.9
34.3
30.7
27.0
23.3
19.6
15.7
11.9
7.9
4.0
[v] PV(Income Stable) / Total Value
[w] PV(Income Risky) / Total Value
[x] PV(Depreciation Stable) / Total Value
[y] PV(Depreciation Risky) / Total Value
0.55
0.06
0.26
0.12
0.53
0.08
0.25
0.14
0.50
0.09
0.25
0.16
0.46
0.11
0.24
0.19
0.42
0.14
0.22
0.22
0.37
0.16
0.21
0.26
0.32
0.20
0.18
0.30
0.26
0.24
0.15
0.35
0.18
0.30
0.11
0.41
0.10
0.36
0.06
0.47
0.00
0.45
0.00
0.55
0.00
0.43
0.00
0.57
0.00
0.41
0.00
0.59
0.00
0.39
0.00
0.61
0.00
0.37
0.00
0.63
0.00
0.35
0.00
0.65
0.00
0.33
0.00
0.67
0.00
0.31
0.00
0.69
0.00
0.28
0.00
0.72
0.00
0.25
0.00
0.75
0.00
0.22
0.00
0.78
0.00
0.19
0.00
0.81
0.00
0.16
0.00
0.84
0.00
0.13
0.00
0.87
0.00
0.09
0.00
0.91
10.3%
10.4%
10.5%
10.6%
10.7%
10.8%
11.0%
11.2%
11.5%
11.8%
12.2%
12.1%
12.1%
12.0%
11.9%
11.8%
11.6%
11.5%
11.4%
11.3%
11.1%
11.0%
10.8%
10.6%
10.4%
PV Income Stable Years
PV Income Risky Years
PV Stable Depreciation Years
PV Risky Depreciation Years
[z] Overall Cost of Equity:
Sources and Notes:
[a] - [j]: Assumed.
[k], year 1 = initial rate base; afterwards [k] = last year's [i] - last years ([n]+[o]).
[l]: [a] x [k], stable years only.
[m]: [b] x [k], risky years only.
[n]: [h] / [g], stable years only.
[o]: [h] / [g], risky years only.
[p] = [l] + [m] + [n] + [o].
[q] = present value of [l] from current year through end of life at a discount rate of [c].
[r] = present value of [m] from current year through end of life at a discount rate of [d].
[s] = present value of [n] from current year through end of life at a discount rate of [e].
[t] = present value of [o] from current year through end of life at a discount rate of [f].
[u] = [q] + [r] + [s] + [t].
[v] = [q] / [u].
[w] = [r] / [u].
[x] = [s] / [u].
[y] = [t] / [u].
[z] = ([c]×[v]) + ([d]×[w]) + ([e]×[x]) + ([f]×[y]).
Table R-2
[a]
[b]
[c]
[d]
[e]
[f]
[g]
[h]
[i]
Assumptions
Allowed ROE, Stable
Allowed ROE, Risky
Discount Rate, Income, Stable
Discount Rate, Income, Risky
Discount Rate, Depreciation, Stable
Discount Rate, Depreciation, Risky
Remaining Life
Initial Rate Base
Year After Which Risk Shifts
[j] Years
10%
15%
10%
15%
10%
10%
25
100
10
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
[k] Rate Base
[l] Return Stable Years
[m] Return Risky Years
[n] Depreciation Stable Years
[o] Depreciation Risky Years
[p] Total
100.0
10.0
0.0
4.0
0.0
14.0
96.0
9.6
0.0
4.0
0.0
13.6
92.0
9.2
0.0
4.0
0.0
13.2
88.0
8.8
0.0
4.0
0.0
12.8
84.0
8.4
0.0
4.0
0.0
12.4
80.0
8.0
0.0
4.0
0.0
12.0
76.0
7.6
0.0
4.0
0.0
11.6
72.0
7.2
0.0
4.0
0.0
11.2
68.0
6.8
0.0
4.0
0.0
10.8
64.0
6.4
0.0
4.0
0.0
10.4
60.0
0.0
9.0
0.0
4.0
13.0
56.0
0.0
8.4
0.0
4.0
12.4
52.0
0.0
7.8
0.0
4.0
11.8
48.0
0.0
7.2
0.0
4.0
11.2
44.0
0.0
6.6
0.0
4.0
10.6
40.0
0.0
6.0
0.0
4.0
10.0
36.0
0.0
5.4
0.0
4.0
9.4
32.0
0.0
4.8
0.0
4.0
8.8
28.0
0.0
4.2
0.0
4.0
8.2
24.0
0.0
3.6
0.0
4.0
7.6
20.0
0.0
3.0
0.0
4.0
7.0
16.0
0.0
2.4
0.0
4.0
6.4
12.0
0.0
1.8
0.0
4.0
5.8
8.0
0.0
1.2
0.0
4.0
5.2
4.0
0.0
0.6
0.0
4.0
4.6
[q]
[r]
[s]
[t]
52.3
9.0
24.6
11.7
47.5
10.4
23.0
12.9
42.7
12.0
21.3
14.2
37.7
13.8
19.5
15.6
32.7
15.8
17.4
17.2
27.6
18.2
15.2
18.9
22.3
20.9
12.7
20.8
17.0
24.1
9.9
22.9
11.5
27.7
6.9
25.1
5.8
31.8
3.6
27.7
0.0
36.6
0.0
30.4
0.0
33.1
0.0
29.5
0.0
29.7
0.0
28.4
0.0
26.3
0.0
27.3
0.0
23.1
0.0
26.0
0.0
19.9
0.0
24.6
0.0
16.9
0.0
23.0
0.0
14.1
0.0
21.3
0.0
11.4
0.0
19.5
0.0
8.9
0.0
17.4
0.0
6.6
0.0
15.2
0.0
4.6
0.0
12.7
0.0
2.9
0.0
9.9
0.0
1.5
0.0
6.9
0.0
0.5
0.0
3.6
[u] Total Present Value
97.6
93.9
90.2
86.6
83.1
79.8
76.7
73.9
71.2
68.9
67.0
62.6
58.1
53.6
49.0
44.5
39.9
35.4
30.8
26.3
21.8
17.3
12.8
8.4
4.2
[v] PV(Income Stable) / Total Value
[w] PV(Income Risky) / Total Value
[x] PV(Depreciation Stable) / Total Value
[y] PV(Depreciation Risky) / Total Value
0.54
0.09
0.25
0.12
0.51
0.11
0.25
0.14
0.47
0.13
0.24
0.16
0.44
0.16
0.22
0.18
0.39
0.19
0.21
0.21
0.35
0.23
0.19
0.24
0.29
0.27
0.17
0.27
0.23
0.33
0.13
0.31
0.16
0.39
0.10
0.35
0.08
0.46
0.05
0.40
0.00
0.55
0.00
0.45
0.00
0.53
0.00
0.47
0.00
0.51
0.00
0.49
0.00
0.49
0.00
0.51
0.00
0.47
0.00
0.53
0.00
0.45
0.00
0.55
0.00
0.42
0.00
0.58
0.00
0.40
0.00
0.60
0.00
0.37
0.00
0.63
0.00
0.34
0.00
0.66
0.00
0.30
0.00
0.70
0.00
0.27
0.00
0.73
0.00
0.22
0.00
0.78
0.00
0.18
0.00
0.82
0.00
0.13
0.00
0.87
10.5%
10.6%
10.7%
10.8%
11.0%
11.1%
11.4%
11.6%
11.9%
12.3%
12.7%
12.6%
12.6%
12.5%
12.4%
12.2%
12.1%
12.0%
11.8%
11.7%
11.5%
11.3%
11.1%
10.9%
10.6%
PV Income Stable Years
PV Income Risky Years
PV Stable Depreciation Years
PV Risky Depreciation Years
[z] Overall Cost of Equity:
Sources and Notes:
[a] - [j]: Assumed.
[k], year 1 = initial rate base; afterwards [k] = last year's [i] - last years ([n]+[o]).
[l]: [a] x [k], stable years only.
[m]: [b] x [k], risky years only.
[n]: [h] / [g], stable years only.
[o]: [h] / [g], risky years only.
[p] = [l] + [m] + [n] + [o].
[q] = present value of [l] from current year through end of life at a discount rate of [c].
[r] = present value of [m] from current year through end of life at a discount rate of [d].
[s] = present value of [n] from current year through end of life at a discount rate of [e].
[t] = present value of [o] from current year through end of life at a discount rate of [f].
[u] = [q] + [r] + [s] + [t].
[v] = [q] / [u].
[w] = [r] / [u].
[x] = [s] / [u].
[y] = [t] / [u].
[z] = ([c]×[v]) + ([d]×[w]) + ([e]×[x]) + ([f]×[y]).
Table R-3
[a]
[b]
[c]
[d]
[e]
[f]
[g]
[h]
[i]
Assumptions
Allowed ROE, Stable
Allowed ROE, Risky
Discount Rate, Income, Stable
Discount Rate, Income, Risky
Discount Rate, Depreciation, Stable
Discount Rate, Depreciation, Risky
Remaining Life
Initial Rate Base
Year After Which Risk Shifts
[j] Years
10%
10%
10%
15%
10%
15%
25
100
10
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
[k] Rate Base
[l] Return Stable Years
[m] Return Risky Years
[n] Depreciation Stable Years
[o] Depreciation Risky Years
[p] Total
100.0
10.0
0.0
4.0
0.0
14.0
96.0
9.6
0.0
4.0
0.0
13.6
92.0
9.2
0.0
4.0
0.0
13.2
88.0
8.8
0.0
4.0
0.0
12.8
84.0
8.4
0.0
4.0
0.0
12.4
80.0
8.0
0.0
4.0
0.0
12.0
76.0
7.6
0.0
4.0
0.0
11.6
72.0
7.2
0.0
4.0
0.0
11.2
68.0
6.8
0.0
4.0
0.0
10.8
64.0
6.4
0.0
4.0
0.0
10.4
60.0
0.0
6.0
0.0
4.0
10.0
56.0
0.0
5.6
0.0
4.0
9.6
52.0
0.0
5.2
0.0
4.0
9.2
48.0
0.0
4.8
0.0
4.0
8.8
44.0
0.0
4.4
0.0
4.0
8.4
40.0
0.0
4.0
0.0
4.0
8.0
36.0
0.0
3.6
0.0
4.0
7.6
32.0
0.0
3.2
0.0
4.0
7.2
28.0
0.0
2.8
0.0
4.0
6.8
24.0
0.0
2.4
0.0
4.0
6.4
20.0
0.0
2.0
0.0
4.0
6.0
16.0
0.0
1.6
0.0
4.0
5.6
12.0
0.0
1.2
0.0
4.0
5.2
8.0
0.0
0.8
0.0
4.0
4.8
4.0
0.0
0.4
0.0
4.0
4.4
[q]
[r]
[s]
[t]
52.3
6.0
24.6
5.8
47.5
6.9
23.0
6.6
42.7
8.0
21.3
7.6
37.7
9.2
19.5
8.8
32.7
10.6
17.4
10.1
27.6
12.1
15.2
11.6
22.3
14.0
12.7
13.4
17.0
16.0
9.9
15.4
11.5
18.5
6.9
17.7
5.8
21.2
3.6
20.3
0.0
24.4
0.0
23.4
0.0
22.1
0.0
22.9
0.0
19.8
0.0
22.3
0.0
17.5
0.0
21.7
0.0
15.4
0.0
20.9
0.0
13.3
0.0
20.1
0.0
11.3
0.0
19.1
0.0
9.4
0.0
17.9
0.0
7.6
0.0
16.6
0.0
5.9
0.0
15.1
0.0
4.4
0.0
13.4
0.0
3.1
0.0
11.4
0.0
1.9
0.0
9.1
0.0
1.0
0.0
6.5
0.0
0.3
0.0
3.5
[u] Total Present Value
88.7
84.1
79.6
75.2
70.8
66.5
62.3
58.3
54.6
51.0
47.8
45.0
42.1
39.2
36.3
33.4
30.4
27.3
24.2
21.0
17.8
14.5
11.0
7.5
3.8
[v] PV(Income Stable) / Total Value
[w] PV(Income Risky) / Total Value
[x] PV(Depreciation Stable) / Total Value
[y] PV(Depreciation Risky) / Total Value
0.59
0.07
0.28
0.07
0.56
0.08
0.27
0.08
0.54
0.10
0.27
0.10
0.50
0.12
0.26
0.12
0.46
0.15
0.25
0.14
0.41
0.18
0.23
0.17
0.36
0.22
0.20
0.21
0.29
0.28
0.17
0.26
0.21
0.34
0.13
0.32
0.11
0.42
0.07
0.40
0.00
0.51
0.00
0.49
0.00
0.49
0.00
0.51
0.00
0.47
0.00
0.53
0.00
0.45
0.00
0.55
0.00
0.42
0.00
0.58
0.00
0.40
0.00
0.60
0.00
0.37
0.00
0.63
0.00
0.34
0.00
0.66
0.00
0.31
0.00
0.69
0.00
0.28
0.00
0.72
0.00
0.25
0.00
0.75
0.00
0.21
0.00
0.79
0.00
0.17
0.00
0.83
0.00
0.13
0.00
0.87
0.00
0.09
0.00
0.91
10.7%
10.8%
11.0%
11.2%
11.5%
11.8%
12.2%
12.7%
13.3%
14.1%
15.0%
15.0%
15.0%
15.0%
15.0%
15.0%
15.0%
15.0%
15.0%
15.0%
15.0%
15.0%
15.0%
15.0%
15.0%
PV Income Stable Years
PV Income Risky Years
PV Stable Depreciation Years
PV Risky Depreciation Years
[z] Overall Cost of Equity:
Sources and Notes:
[a] - [j]: Assumed.
[k], year 1 = initial rate base; afterwards [k] = last year's [i] - last years ([n]+[o]).
[l]: [a] x [k], stable years only.
[m]: [b] x [k], risky years only.
[n]: [h] / [g], stable years only.
[o]: [h] / [g], risky years only.
[p] = [l] + [m] + [n] + [o].
[q] = present value of [l] from current year through end of life at a discount rate of [c].
[r] = present value of [m] from current year through end of life at a discount rate of [d].
[s] = present value of [n] from current year through end of life at a discount rate of [e].
[t] = present value of [o] from current year through end of life at a discount rate of [f].
[u] = [q] + [r] + [s] + [t].
[v] = [q] / [u].
[w] = [r] / [u].
[x] = [s] / [u].
[y] = [t] / [u].
[z] = ([c]×[v]) + ([d]×[w]) + ([e]×[x]) + ([f]×[y]).
Table R-4
[a]
[b]
[c]
[d]
[e]
[f]
[g]
[h]
[i]
Assumptions
Allowed ROE, Stable
Allowed ROE, Risky
Discount Rate, Income, Stable
Discount Rate, Income, Risky
Discount Rate, Depreciation, Stable
Discount Rate, Depreciation, Risky
Remaining Life
Initial Rate Base
Year After Which Risk Shifts
[j] Years
10%
10%
10%
12%
10%
12%
25
100
10
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
[k] Rate Base
[l] Return Stable Years
[m] Return Risky Years
[n] Depreciation Stable Years
[o] Depreciation Risky Years
[p] Total
100.0
10.0
0.0
4.0
0.0
14.0
96.0
9.6
0.0
4.0
0.0
13.6
92.0
9.2
0.0
4.0
0.0
13.2
88.0
8.8
0.0
4.0
0.0
12.8
84.0
8.4
0.0
4.0
0.0
12.4
80.0
8.0
0.0
4.0
0.0
12.0
76.0
7.6
0.0
4.0
0.0
11.6
72.0
7.2
0.0
4.0
0.0
11.2
68.0
6.8
0.0
4.0
0.0
10.8
64.0
6.4
0.0
4.0
0.0
10.4
60.0
0.0
6.0
0.0
4.0
10.0
56.0
0.0
5.6
0.0
4.0
9.6
52.0
0.0
5.2
0.0
4.0
9.2
48.0
0.0
4.8
0.0
4.0
8.8
44.0
0.0
4.4
0.0
4.0
8.4
40.0
0.0
4.0
0.0
4.0
8.0
36.0
0.0
3.6
0.0
4.0
7.6
32.0
0.0
3.2
0.0
4.0
7.2
28.0
0.0
2.8
0.0
4.0
6.8
24.0
0.0
2.4
0.0
4.0
6.4
20.0
0.0
2.0
0.0
4.0
6.0
16.0
0.0
1.6
0.0
4.0
5.6
12.0
0.0
1.2
0.0
4.0
5.2
8.0
0.0
0.8
0.0
4.0
4.8
4.0
0.0
0.4
0.0
4.0
4.4
[q]
[r]
[s]
[t]
52.3
8.8
24.6
8.8
47.5
9.8
23.0
9.8
42.7
11.0
21.3
11.0
37.7
12.3
19.5
12.3
32.7
13.8
17.4
13.8
27.6
15.5
15.2
15.5
22.3
17.3
12.7
17.3
17.0
19.4
9.9
19.4
11.5
21.8
6.9
21.7
5.8
24.4
3.6
24.3
0.0
27.3
0.0
27.2
0.0
24.6
0.0
26.5
0.0
21.9
0.0
25.7
0.0
19.4
0.0
24.8
0.0
16.9
0.0
23.8
0.0
14.5
0.0
22.6
0.0
12.2
0.0
21.3
0.0
10.1
0.0
19.9
0.0
8.1
0.0
18.3
0.0
6.3
0.0
16.4
0.0
4.7
0.0
14.4
0.0
3.2
0.0
12.1
0.0
2.0
0.0
9.6
0.0
1.0
0.0
6.8
0.0
0.4
0.0
3.6
[u] Total Present Value
94.4
90.2
86.0
81.9
77.8
73.7
69.7
65.7
61.9
58.2
54.5
51.1
47.6
44.1
40.6
37.1
33.6
30.0
26.4
22.7
19.1
15.4
11.6
7.8
3.9
[v] PV(Income Stable) / Total Value
[w] PV(Income Risky) / Total Value
[x] PV(Depreciation Stable) / Total Value
[y] PV(Depreciation Risky) / Total Value
0.55
0.09
0.26
0.09
0.53
0.11
0.26
0.11
0.50
0.13
0.25
0.13
0.46
0.15
0.24
0.15
0.42
0.18
0.22
0.18
0.37
0.21
0.21
0.21
0.32
0.25
0.18
0.25
0.26
0.30
0.15
0.29
0.19
0.35
0.11
0.35
0.10
0.42
0.06
0.42
0.00
0.50
0.00
0.50
0.00
0.48
0.00
0.52
0.00
0.46
0.00
0.54
0.00
0.44
0.00
0.56
0.00
0.42
0.00
0.58
0.00
0.39
0.00
0.61
0.00
0.36
0.00
0.64
0.00
0.34
0.00
0.66
0.00
0.31
0.00
0.69
0.00
0.28
0.00
0.72
0.00
0.24
0.00
0.76
0.00
0.21
0.00
0.79
0.00
0.17
0.00
0.83
0.00
0.13
0.00
0.87
0.00
0.09
0.00
0.91
10.4%
10.4%
10.5%
10.6%
10.7%
10.8%
11.0%
11.2%
11.4%
11.7%
12.0%
12.0%
12.0%
12.0%
12.0%
12.0%
12.0%
12.0%
12.0%
12.0%
12.0%
12.0%
12.0%
12.0%
12.0%
PV Income Stable Years
PV Income Risky Years
PV Stable Depreciation Years
PV Risky Depreciation Years
[z] Overall Cost of Equity:
Sources and Notes:
[a] - [j]: Assumed.
[k], year 1 = initial rate base; afterwards [k] = last year's [i] - last years ([n]+[o]).
[l]: [a] x [k], stable years only.
[m]: [b] x [k], risky years only.
[n]: [h] / [g], stable years only.
[o]: [h] / [g], risky years only.
[p] = [l] + [m] + [n] + [o].
[q] = present value of [l] from current year through end of life at a discount rate of [c].
[r] = present value of [m] from current year through end of life at a discount rate of [d].
[s] = present value of [n] from current year through end of life at a discount rate of [e].
[t] = present value of [o] from current year through end of life at a discount rate of [f].
[u] = [q] + [r] + [s] + [t].
[v] = [q] / [u].
[w] = [r] / [u].
[x] = [s] / [u].
[y] = [t] / [u].
[z] = ([c]×[v]) + ([d]×[w]) + ([e]×[x]) + ([f]×[y]).
Table R-5
[a]
[b]
[c]
[d]
[e]
[f]
[g]
[h]
[i]
Assumptions
Allowed ROE, Stable
Allowed ROE, Risky
Discount Rate, Income, Stable
Discount Rate, Income, Risky
Discount Rate, Depreciation, Stable
Discount Rate, Depreciation, Risky
Remaining Life
Initial Rate Base
Year After Which Risk Shifts
[j] Years
10%
12%
10%
12%
10%
12%
25
100
10
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
[k] Rate Base
[l] Return Stable Years
[m] Return Risky Years
[n] Depreciation Stable Years
[o] Depreciation Risky Years
[p] Total
100.0
10.0
0.0
4.0
0.0
14.0
96.0
9.6
0.0
4.0
0.0
13.6
92.0
9.2
0.0
4.0
0.0
13.2
88.0
8.8
0.0
4.0
0.0
12.8
84.0
8.4
0.0
4.0
0.0
12.4
80.0
8.0
0.0
4.0
0.0
12.0
76.0
7.6
0.0
4.0
0.0
11.6
72.0
7.2
0.0
4.0
0.0
11.2
68.0
6.8
0.0
4.0
0.0
10.8
64.0
6.4
0.0
4.0
0.0
10.4
60.0
0.0
7.2
0.0
4.0
11.2
56.0
0.0
6.7
0.0
4.0
10.7
52.0
0.0
6.2
0.0
4.0
10.2
48.0
0.0
5.8
0.0
4.0
9.8
44.0
0.0
5.3
0.0
4.0
9.3
40.0
0.0
4.8
0.0
4.0
8.8
36.0
0.0
4.3
0.0
4.0
8.3
32.0
0.0
3.8
0.0
4.0
7.8
28.0
0.0
3.4
0.0
4.0
7.4
24.0
0.0
2.9
0.0
4.0
6.9
20.0
0.0
2.4
0.0
4.0
6.4
16.0
0.0
1.9
0.0
4.0
5.9
12.0
0.0
1.4
0.0
4.0
5.4
8.0
0.0
1.0
0.0
4.0
5.0
4.0
0.0
0.5
0.0
4.0
4.5
[q]
[r]
[s]
[t]
52.3
10.5
24.6
8.8
47.5
11.8
23.0
9.8
42.7
13.2
21.3
11.0
37.7
14.8
19.5
12.3
32.7
16.6
17.4
13.8
27.6
18.6
15.2
15.5
22.3
20.8
12.7
17.3
17.0
23.3
9.9
19.4
11.5
26.1
6.9
21.7
5.8
29.2
3.6
24.3
0.0
32.8
0.0
27.2
0.0
29.5
0.0
26.5
0.0
26.3
0.0
25.7
0.0
23.2
0.0
24.8
0.0
20.2
0.0
23.8
0.0
17.4
0.0
22.6
0.0
14.7
0.0
21.3
0.0
12.1
0.0
19.9
0.0
9.7
0.0
18.3
0.0
7.6
0.0
16.4
0.0
5.6
0.0
14.4
0.0
3.9
0.0
12.1
0.0
2.4
0.0
9.6
0.0
1.2
0.0
6.8
0.0
0.4
0.0
3.6
[u] Total Present Value
96.2
92.2
88.2
84.4
80.5
76.8
73.2
69.6
66.2
63.0
60.0
56.0
52.0
48.0
44.0
40.0
36.0
32.0
28.0
24.0
20.0
16.0
12.0
8.0
4.0
[v] PV(Income Stable) / Total Value
[w] PV(Income Risky) / Total Value
[x] PV(Depreciation Stable) / Total Value
[y] PV(Depreciation Risky) / Total Value
0.54
0.11
0.26
0.09
0.52
0.13
0.25
0.11
0.48
0.15
0.24
0.12
0.45
0.18
0.23
0.15
0.41
0.21
0.22
0.17
0.36
0.24
0.20
0.20
0.31
0.28
0.17
0.24
0.24
0.33
0.14
0.28
0.17
0.39
0.10
0.33
0.09
0.46
0.06
0.39
0.00
0.55
0.00
0.45
0.00
0.53
0.00
0.47
0.00
0.51
0.00
0.49
0.00
0.48
0.00
0.52
0.00
0.46
0.00
0.54
0.00
0.43
0.00
0.57
0.00
0.41
0.00
0.59
0.00
0.38
0.00
0.62
0.00
0.35
0.00
0.65
0.00
0.31
0.00
0.69
0.00
0.28
0.00
0.72
0.00
0.24
0.00
0.76
0.00
0.20
0.00
0.80
0.00
0.15
0.00
0.85
0.00
0.11
0.00
0.89
10.4%
10.5%
10.5%
10.6%
10.8%
10.9%
11.0%
11.2%
11.4%
11.7%
12.0%
12.0%
12.0%
12.0%
12.0%
12.0%
12.0%
12.0%
12.0%
12.0%
12.0%
12.0%
12.0%
12.0%
12.0%
PV Income Stable Years
PV Income Risky Years
PV Stable Depreciation Years
PV Risky Depreciation Years
[z] Overall Cost of Equity:
Sources and Notes:
[a] - [j]: Assumed.
[k], year 1 = initial rate base; afterwards [k] = last year's [i] - last years ([n]+[o]).
[l]: [a] x [k], stable years only.
[m]: [b] x [k], risky years only.
[n]: [h] / [g], stable years only.
[o]: [h] / [g], risky years only.
[p] = [l] + [m] + [n] + [o].
[q] = present value of [l] from current year through end of life at a discount rate of [c].
[r] = present value of [m] from current year through end of life at a discount rate of [d].
[s] = present value of [n] from current year through end of life at a discount rate of [e].
[t] = present value of [o] from current year through end of life at a discount rate of [f].
[u] = [q] + [r] + [s] + [t].
[v] = [q] / [u].
[w] = [r] / [u].
[x] = [s] / [u].
[y] = [t] / [u].
[z] = ([c]×[v]) + ([d]×[w]) + ([e]×[x]) + ([f]×[y]).
Table R-6
[a]
[b]
[c]
[d]
[e]
[f]
[g]
[h]
[i]
Assumptions
Allowed ROE, Stable
Allowed ROE, Risky
Discount Rate, Income, Stable
Discount Rate, Income, Risky
Discount Rate, Depreciation, Stable
Discount Rate, Depreciation, Risky
Remaining Life
Initial Rate Base
Year After Which Risk Shifts
[j] Years
10.9%
12%
10%
12%
10%
12%
25
100
10
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
[k] Rate Base
[l] Return Stable Years
[m] Return Risky Years
[n] Depreciation Stable Years
[o] Depreciation Risky Years
[p] Total
100.0
10.9
0.0
4.0
0.0
14.9
96.0
10.5
0.0
4.0
0.0
14.5
92.0
10.0
0.0
4.0
0.0
14.0
88.0
9.6
0.0
4.0
0.0
13.6
84.0
9.2
0.0
4.0
0.0
13.2
80.0
8.7
0.0
4.0
0.0
12.7
76.0
8.3
0.0
4.0
0.0
12.3
72.0
7.8
0.0
4.0
0.0
11.8
68.0
7.4
0.0
4.0
0.0
11.4
64.0
7.0
0.0
4.0
0.0
11.0
60.0
0.0
7.2
0.0
4.0
11.2
56.0
0.0
6.7
0.0
4.0
10.7
52.0
0.0
6.2
0.0
4.0
10.2
48.0
0.0
5.8
0.0
4.0
9.8
44.0
0.0
5.3
0.0
4.0
9.3
40.0
0.0
4.8
0.0
4.0
8.8
36.0
0.0
4.3
0.0
4.0
8.3
32.0
0.0
3.8
0.0
4.0
7.8
28.0
0.0
3.4
0.0
4.0
7.4
24.0
0.0
2.9
0.0
4.0
6.9
20.0
0.0
2.4
0.0
4.0
6.4
16.0
0.0
1.9
0.0
4.0
5.9
12.0
0.0
1.4
0.0
4.0
5.4
8.0
0.0
1.0
0.0
4.0
5.0
4.0
0.0
0.5
0.0
4.0
4.5
[q]
[r]
[s]
[t]
57.0
10.5
23.7
8.8
51.8
11.8
22.2
9.8
46.5
13.2
20.7
11.0
41.1
14.8
18.9
12.3
35.7
16.6
17.0
13.8
30.1
18.6
14.8
15.5
24.4
20.8
12.4
17.3
18.5
23.3
9.8
19.4
12.5
26.1
6.9
21.7
6.3
29.2
3.6
24.3
0.0
32.8
0.0
27.2
0.0
29.5
0.0
26.5
0.0
26.3
0.0
25.7
0.0
23.2
0.0
24.8
0.0
20.2
0.0
23.8
0.0
17.4
0.0
22.6
0.0
14.7
0.0
21.3
0.0
12.1
0.0
19.9
0.0
9.7
0.0
18.3
0.0
7.6
0.0
16.4
0.0
5.6
0.0
14.4
0.0
3.9
0.0
12.1
0.0
2.4
0.0
9.6
0.0
1.2
0.0
6.8
0.0
0.4
0.0
3.6
100.0
95.7
91.4
87.2
83.0
78.9
74.9
71.0
67.2
63.5
60.0
56.0
52.0
48.0
44.0
40.0
36.0
32.0
28.0
24.0
20.0
16.0
12.0
8.0
4.0
0.57
0.11
0.24
0.09
0.54
0.12
0.23
0.10
0.51
0.14
0.23
0.12
0.47
0.17
0.22
0.14
0.43
0.20
0.20
0.17
0.38
0.24
0.19
0.20
0.33
0.28
0.17
0.23
0.26
0.33
0.14
0.27
0.19
0.39
0.10
0.32
0.10
0.46
0.06
0.38
0.00
0.55
0.00
0.45
0.00
0.53
0.00
0.47
0.00
0.51
0.00
0.49
0.00
0.48
0.00
0.52
0.00
0.46
0.00
0.54
0.00
0.43
0.00
0.57
0.00
0.41
0.00
0.59
0.00
0.38
0.00
0.62
0.00
0.35
0.00
0.65
0.00
0.31
0.00
0.69
0.00
0.28
0.00
0.72
0.00
0.24
0.00
0.76
0.00
0.20
0.00
0.80
0.00
0.15
0.00
0.85
0.00
0.11
0.00
0.89
10.4%
10.5%
10.5%
10.6%
10.7%
10.9%
11.0%
11.2%
11.4%
11.7%
12.0%
12.0%
12.0%
12.0%
12.0%
12.0%
12.0%
12.0%
12.0%
12.0%
12.0%
12.0%
12.0%
12.0%
12.0%
PV Income Stable Years
PV Income Risky Years
PV Stable Depreciation Years
PV Risky Depreciation Years
[u] Total Present Value
[v] PV(Income Stable) / Total Value
[w] PV(Income Risky) / Total Value
[x] PV(Depreciation Stable) / Total Value
[y] PV(Depreciation Risky) / Total Value
[z] Overall Cost of Equity:
Sources and Notes:
[a] - [j]: Assumed.
[k], year 1 = initial rate base; afterwards [k] = last year's [i] - last years ([n]+[o]).
[l]: [a] x [k], stable years only.
[m]: [b] x [k], risky years only.
[n]: [h] / [g], stable years only.
[o]: [h] / [g], risky years only.
[p] = [l] + [m] + [n] + [o].
[q] = present value of [l] from current year through end of life at a discount rate of [c].
[r] = present value of [m] from current year through end of life at a discount rate of [d].
[s] = present value of [n] from current year through end of life at a discount rate of [e].
[t] = present value of [o] from current year through end of life at a discount rate of [f].
[u] = [q] + [r] + [s] + [t].
[v] = [q] / [u].
[w] = [r] / [u].
[x] = [s] / [u].
[y] = [t] / [u].
[z] = ([c]×[v]) + ([d]×[w]) + ([e]×[x]) + ([f]×[y]).
Table R-7
[a]
[b]
[c]
[d]
[e]
[f]
[g]
[h]
[i]
Assumptions
Allowed ROE, Stable
Allowed ROE, Risky
Discount Rate, Income, Stable
Discount Rate, Income, Risky
Discount Rate, Depreciation, Stable
Discount Rate, Depreciation, Risky
Remaining Life
Initial Rate Base
Year After Which Risk Shifts
[j] Years
9.56%
12%
9.56%
12%
9.56%
12%
25
100
10
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
[k] Rate Base
[l] Return Stable Years
[m] Return Risky Years
[n] Depreciation Stable Years
[o] Depreciation Risky Years
[p] Total
100.0
9.6
0.0
4.0
0.0
13.6
96.0
9.2
0.0
4.0
0.0
13.2
92.0
8.8
0.0
4.0
0.0
12.8
88.0
8.4
0.0
4.0
0.0
12.4
84.0
8.0
0.0
4.0
0.0
12.0
80.0
7.6
0.0
4.0
0.0
11.6
76.0
7.3
0.0
4.0
0.0
11.3
72.0
6.9
0.0
4.0
0.0
10.9
68.0
6.5
0.0
4.0
0.0
10.5
64.0
6.1
0.0
4.0
0.0
10.1
60.0
0.0
7.2
0.0
4.0
11.2
56.0
0.0
6.7
0.0
4.0
10.7
52.0
0.0
6.2
0.0
4.0
10.2
48.0
0.0
5.8
0.0
4.0
9.8
44.0
0.0
5.3
0.0
4.0
9.3
40.0
0.0
4.8
0.0
4.0
8.8
36.0
0.0
4.3
0.0
4.0
8.3
32.0
0.0
3.8
0.0
4.0
7.8
28.0
0.0
3.4
0.0
4.0
7.4
24.0
0.0
2.9
0.0
4.0
6.9
20.0
0.0
2.4
0.0
4.0
6.4
16.0
0.0
1.9
0.0
4.0
5.9
12.0
0.0
1.4
0.0
4.0
5.4
8.0
0.0
1.0
0.0
4.0
5.0
4.0
0.0
0.5
0.0
4.0
4.5
[q]
[r]
[s]
[t]
50.9
10.5
25.0
8.8
46.2
11.8
23.4
9.8
41.4
13.2
21.7
11.0
36.6
14.8
19.8
12.3
31.7
16.6
17.6
13.8
26.7
18.6
15.3
15.5
21.6
20.8
12.8
17.3
16.4
23.3
10.0
19.4
11.0
26.1
7.0
21.7
5.6
29.2
3.7
24.3
0.0
32.8
0.0
27.2
0.0
29.5
0.0
26.5
0.0
26.3
0.0
25.7
0.0
23.2
0.0
24.8
0.0
20.2
0.0
23.8
0.0
17.4
0.0
22.6
0.0
14.7
0.0
21.3
0.0
12.1
0.0
19.9
0.0
9.7
0.0
18.3
0.0
7.6
0.0
16.4
0.0
5.6
0.0
14.4
0.0
3.9
0.0
12.1
0.0
2.4
0.0
9.6
0.0
1.2
0.0
6.8
0.0
0.4
0.0
3.6
[u] Total Present Value
95.2
91.3
87.3
83.5
79.7
76.0
72.5
69.1
65.8
62.8
60.0
56.0
52.0
48.0
44.0
40.0
36.0
32.0
28.0
24.0
20.0
16.0
12.0
8.0
4.0
[v] PV(Income Stable) / Total Value
[w] PV(Income Risky) / Total Value
[x] PV(Depreciation Stable) / Total Value
[y] PV(Depreciation Risky) / Total Value
0.53
0.11
0.26
0.09
0.51
0.13
0.26
0.11
0.47
0.15
0.25
0.13
0.44
0.18
0.24
0.15
0.40
0.21
0.22
0.17
0.35
0.24
0.20
0.20
0.30
0.29
0.18
0.24
0.24
0.34
0.15
0.28
0.17
0.40
0.11
0.33
0.09
0.47
0.06
0.39
0.00
0.55
0.00
0.45
0.00
0.53
0.00
0.47
0.00
0.51
0.00
0.49
0.00
0.48
0.00
0.52
0.00
0.46
0.00
0.54
0.00
0.43
0.00
0.57
0.00
0.41
0.00
0.59
0.00
0.38
0.00
0.62
0.00
0.35
0.00
0.65
0.00
0.31
0.00
0.69
0.00
0.28
0.00
0.72
0.00
0.24
0.00
0.76
0.00
0.20
0.00
0.80
0.00
0.15
0.00
0.85
0.00
0.11
0.00
0.89
10.1%
10.1%
10.2%
10.4%
10.5%
10.7%
10.8%
11.1%
11.3%
11.6%
12.0%
12.0%
12.0%
12.0%
12.0%
12.0%
12.0%
12.0%
12.0%
12.0%
12.0%
12.0%
12.0%
12.0%
12.0%
PV Income Stable Years
PV Income Risky Years
PV Stable Depreciation Years
PV Risky Depreciation Years
[z] Overall Cost of Equity:
Sources and Notes:
[a] - [j]: Assumed.
[k], year 1 = initial rate base; afterwards [k] = last year's [i] - last years ([n]+[o]).
[l]: [a] x [k], stable years only.
[m]: [b] x [k], risky years only.
[n]: [h] / [g], stable years only.
[o]: [h] / [g], risky years only.
[p] = [l] + [m] + [n] + [o].
[q] = present value of [l] from current year through end of life at a discount rate of [c].
[r] = present value of [m] from current year through end of life at a discount rate of [d].
[s] = present value of [n] from current year through end of life at a discount rate of [e].
[t] = present value of [o] from current year through end of life at a discount rate of [f].
[u] = [q] + [r] + [s] + [t].
[v] = [q] / [u].
[w] = [r] / [u].
[x] = [s] / [u].
[y] = [t] / [u].
[z] = ([c]×[v]) + ([d]×[w]) + ([e]×[x]) + ([f]×[y]).
Table R-8
[a]
[b]
[c]
[d]
[e]
[f]
[g]
[h]
[i]
Assumptions
Allowed ROE, Stable
Allowed ROE, Risky
Discount Rate, Income, Stable
Discount Rate, Income, Risky
Discount Rate, Depreciation, Stable
Discount Rate, Depreciation, Risky
Remaining Life
Initial Rate Base
Year After Which Risk Shifts
[j] Years
10.67%
12%
9.56%
12%
9.56%
12%
25
100
10
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
[k] Rate Base
[l] Return Stable Years
[m] Return Risky Years
[n] Depreciation Stable Years
[o] Depreciation Risky Years
[p] Total
100.0
10.7
0.0
4.0
0.0
14.7
96.0
10.2
0.0
4.0
0.0
14.2
92.0
9.8
0.0
4.0
0.0
13.8
88.0
9.4
0.0
4.0
0.0
13.4
84.0
9.0
0.0
4.0
0.0
13.0
80.0
8.5
0.0
4.0
0.0
12.5
76.0
8.1
0.0
4.0
0.0
12.1
72.0
7.7
0.0
4.0
0.0
11.7
68.0
7.3
0.0
4.0
0.0
11.3
64.0
6.8
0.0
4.0
0.0
10.8
60.0
0.0
7.2
0.0
4.0
11.2
56.0
0.0
6.7
0.0
4.0
10.7
52.0
0.0
6.2
0.0
4.0
10.2
48.0
0.0
5.8
0.0
4.0
9.8
44.0
0.0
5.3
0.0
4.0
9.3
40.0
0.0
4.8
0.0
4.0
8.8
36.0
0.0
4.3
0.0
4.0
8.3
32.0
0.0
3.8
0.0
4.0
7.8
28.0
0.0
3.4
0.0
4.0
7.4
24.0
0.0
2.9
0.0
4.0
6.9
20.0
0.0
2.4
0.0
4.0
6.4
16.0
0.0
1.9
0.0
4.0
5.9
12.0
0.0
1.4
0.0
4.0
5.4
8.0
0.0
1.0
0.0
4.0
5.0
4.0
0.0
0.5
0.0
4.0
4.5
[q]
[r]
[s]
[t]
56.8
10.5
23.9
8.8
51.5
11.8
22.4
9.8
46.2
13.2
20.8
11.0
40.8
14.8
19.1
12.3
35.3
16.6
17.1
13.8
29.8
18.6
14.9
15.5
24.1
20.8
12.5
17.3
18.2
23.3
9.8
19.4
12.3
26.1
6.9
21.7
6.2
29.2
3.6
24.3
0.0
32.8
0.0
27.2
0.0
29.5
0.0
26.5
0.0
26.3
0.0
25.7
0.0
23.2
0.0
24.8
0.0
20.2
0.0
23.8
0.0
17.4
0.0
22.6
0.0
14.7
0.0
21.3
0.0
12.1
0.0
19.9
0.0
9.7
0.0
18.3
0.0
7.6
0.0
16.4
0.0
5.6
0.0
14.4
0.0
3.9
0.0
12.1
0.0
2.4
0.0
9.6
0.0
1.2
0.0
6.8
0.0
0.4
0.0
3.6
100.0
95.6
91.3
87.0
82.8
78.7
74.7
70.8
67.0
63.4
60.0
56.0
52.0
48.0
44.0
40.0
36.0
32.0
28.0
24.0
20.0
16.0
12.0
8.0
4.0
0.57
0.11
0.24
0.09
0.54
0.12
0.23
0.10
0.51
0.14
0.23
0.12
0.47
0.17
0.22
0.14
0.43
0.20
0.21
0.17
0.38
0.24
0.19
0.20
0.32
0.28
0.17
0.23
0.26
0.33
0.14
0.27
0.18
0.39
0.10
0.32
0.10
0.46
0.06
0.38
0.00
0.55
0.00
0.45
0.00
0.53
0.00
0.47
0.00
0.51
0.00
0.49
0.00
0.48
0.00
0.52
0.00
0.46
0.00
0.54
0.00
0.43
0.00
0.57
0.00
0.41
0.00
0.59
0.00
0.38
0.00
0.62
0.00
0.35
0.00
0.65
0.00
0.31
0.00
0.69
0.00
0.28
0.00
0.72
0.00
0.24
0.00
0.76
0.00
0.20
0.00
0.80
0.00
0.15
0.00
0.85
0.00
0.11
0.00
0.89
10.0%
10.1%
10.2%
10.3%
10.5%
10.6%
10.8%
11.0%
11.3%
11.6%
12.0%
12.0%
12.0%
12.0%
12.0%
12.0%
12.0%
12.0%
12.0%
12.0%
12.0%
12.0%
12.0%
12.0%
12.0%
PV Income Stable Years
PV Income Risky Years
PV Stable Depreciation Years
PV Risky Depreciation Years
[u] Total Present Value
[v] PV(Income Stable) / Total Value
[w] PV(Income Risky) / Total Value
[x] PV(Depreciation Stable) / Total Value
[y] PV(Depreciation Risky) / Total Value
[z] Overall Cost of Equity:
Sources and Notes:
[a] - [j]: Assumed.
[k], year 1 = initial rate base; afterwards [k] = last year's [i] - last years ([n]+[o]).
[l]: [a] x [k], stable years only.
[m]: [b] x [k], risky years only.
[n]: [h] / [g], stable years only.
[o]: [h] / [g], risky years only.
[p] = [l] + [m] + [n] + [o].
[q] = present value of [l] from current year through end of life at a discount rate of [c].
[r] = present value of [m] from current year through end of life at a discount rate of [d].
[s] = present value of [n] from current year through end of life at a discount rate of [e].
[t] = present value of [o] from current year through end of life at a discount rate of [f].
[u] = [q] + [r] + [s] + [t].
[v] = [q] / [u].
[w] = [r] / [u].
[x] = [s] / [u].
[y] = [t] / [u].
[z] = ([c]×[v]) + ([d]×[w]) + ([e]×[x]) + ([f]×[y]).
APPENDICES TO WRITTEN REPLY EVIDENCE OF
A. LAWRENCE KOLBE
APPENDIX R-BOOTH2: BOOTH EVIDENCE IN EB-2005-0520
(Page numbers in Booth Evidence are from the original document)
165
EB-2005-0520
EXHIBIT K2
BUSINESS RISK AND CAPITAL STRUCTURE
FOR UNION GAS LIMITED
Evidence of Laurence D. Booth
on behalf of
the Consumers Council of Canada, the Industrial Gas Users’ Association and the
Vulnerable Energy Consumers Coalition
Before the
Ontario Energy Board
April 2006
TABLE OF CONTENTS
Page
Executive Summary
i
1.0
Introduction
1
2.0
Regulatory Tools
5
3.0
Business Risk
26
4.0
Financial Risk
45
5.0
Weighted Average Cost of Capital
70
6.0
Financial Leverage Adjustments
85
1
EXECUTIVE SUMMARY
2
The Consumers Council of Canada, Industrial Gas Users’ Association, and the Vulnerable
3
Energy Consumers Coalition (the Intervenors) have asked me to provide expert advice on
4
Union's proposal to change its capital structure. To do this I have provided an independent
5
assessment of Union Gas’ business risk and financial flexibility and considered the company’s
6
testimony, particularly whether “leverage adjustments” are needed to determine the firm’s
7
common equity ratio as suggested by Drs Kolbe and Vilbert. This latter issue involves a
8
discussion of the relevance of the weighted average cost of capital (“WACC”), referred to by
9
the company witnesses as the after-tax weighted-average cost of capital (“ATWACC”), and
10
whether the appropriate common equity ratio can be considered separately from the allowed
11
ROE, which is not part of this hearing since the Board has recently confirmed the
12
appropriateness of its adjustment mechanism.
13
My overall assessment is as follows:
14
15
16
17
•
The short term business risk of Union is very low as it continues to earn its
allowed ROE. There is no indication that the impact of the Board’s policy of
testing performance based regulation has exposed Union’s shareholder to any
increase in risk.
18
19
20
21
22
23
•
In my judgment there has not been a significant change in Union’s business risk
since RP-2003-0063/87/97 when Union requested and was granted a 35%
common equity ratio in the Board’s decision dated March 18, 2004.1 In
particular the BC Utilities Commission has just set Terasen Gas’s common
equity ratio at 35% and I see no significant differences in business risk between
these two companies.
24
25
26
27
28
29
•
Overall I would recommend that Union continue to be allowed a 35% common
equity ratio. I would judge Union in isolation to have a good investment grade
bond rating with its current allowed ROE and common equity ratio. Its DBRS
rating of A has been stable for many years and the cut in its S&P rating to BBB
is due to its ownership structure and reflects Duke Energy’s rating not that of
Union. Spreads on Union’s publicly traded debt indicate that it trades as a better
1
Union Gas was a given a little bump in EBRO 499 when its common equity ratio was increased to
35% from 34% when it was consolidated with Centra Gas Ontario, which had a 36% common equity
ratio. A straight blended rate would have been 34.5%.
-i-
than BBB credit. Spreads on Canadian utility and pipeline debt over the last
several years reflect normal cyclical concerns and do not indicate that the
market has re-evaluated the regulatory protection accorded utility and pipeline
debt in Canada.
1
2
3
4
5
6
7
8
9
10
11
12
•
Notwithstanding the above, there is no guarantee that in the future Union’s debt
costs may not reflect its ownership structure and S&P BBB bond rating. In my
judgement the BCUC has taken the right steps in ensuring that Terasen Gas (BC
Gas Utility) be ring fenced on its indirect acquisition by Kinder Morgan Inc
(KMI). Ring fencing, or structural insulation as S&P refers to it, allows an
operating subsidiary to have a bond rating that reflects its risk rather than that of
its parent. If this is not done there is always the possibility that the company’s
cost of debt includes an “unfair and unreasonable” charge due to its risky parent.
13
14
15
16
17
18
•
In my judgement the most significant change in Union’s risk since 2004 has
occurred due to its ownership rather than its business risk. When the Board
agreed to Union’s requested 35% common equity ratio in its 2004 decision
Union had an A- S&P bond rating, now it is BBB. It is unfair and unreasonable
to ratepayers that Union’s common equity ratio be increased because of its
ownership structure.
19
20
21
22
23
24
25
26
27
28
29
•
In terms of the ATWACC approach used by the company’s witnesses I would
point out the fundamental contradiction in its use in regulatory filings in that it is
the mirror image of shareholder value maximisation. That is, earning more than
the WACC is synonymous with the creation of shareholder value, whereas the
Board’s responsibility is not to create or maintain shareholder value, but to
ensure that rates are fair and reasonable. The Alberta EUB felt it would be
“derelict” in its responsibilities to recognise market capitalisation ratios, an
assessment I agree with. In my judgment setting the equity ratio or ROE
implicitly by using the (AT)WACC approach can “rubberstamp” existing
market values that may in turn reflect unfair and unreasonable rates. I therefore
see absolutely no value to its introduction into a regulatory setting.
30
31
32
33
34
35
36
•
Leverage adjustments should be made when a Board sets both the allowed ROE
and the common equity ratio. In this way the Board makes sure that it does not
“double count” the impact of changes in business risk. For example, Union has
traditionally been regarded as riskier than Enbridge Gas Distribution Inc
(Consumers) with a premium over EGDI’s ROE of 15-25 basis points. Allowing
Union an additional 5% on its common equity ratio effectively increases this
premium ROE over EGDI.
37
38
39
40
•
My recommendation is to ignore the ATWACC approach entirely and continue
with best regulatory practise in Canada and set Union’s common equity ratio
based on its business risk and an assessment of its financial flexibility and
capital market access.
- ii -
1
1.0
INTRODUCTION
2
Q.
PLEASE DESCRIBE YOUR QUALIFICATIONS AND EXPERIENCE.
3
A.
Laurence Booth is a professor of finance and finance area co-ordinator in the Rotman
4
School of Management at the University of Toronto, where he holds the CIT Chair in
5
Structured Finance. Professor Booth, either alone or with the late Professor M. K. Berkowitz,2
6
has previously filed testimony with this Board in rate hearings involving Union Gas, Centra
7
Gas Ontario and EGDI, as well as in the generic hearing in 2003 to review the Board’s ROE
8
adjustment mechanism. A detailed resume has been filed previously with the Board. If needed,
9
Professor Booth’s current CV can be downloaded from his web site.3
Q.
PLEASE DISCUSS HOW YOUR TESTIMONY IS ORGANISED AND THE
ISSUES THAT YOU DEAL WITH.
13
A.
The Intervenors have asked me to provide expert advice on Union's proposal to change
14
its capital structure. To do this I have provided an independent assessment of Union Gas’
15
business risk and financial flexibility and considered the company’s testimony, particularly
16
whether “leverage adjustments” are needed to determine the firm’s common equity ratio as
17
suggested by Drs Kolbe and Vilbert. This latter issue involves a discussion of the relevance of
18
the weighted average cost of capital (WACC), referred to by the company witnesses as the
19
ATWACC, and whether the appropriate common equity ratio can be considered separately
20
from the allowed ROE.
21
The allowed ROE is not an issue in this hearing as it is determined by the Board’s ROE
22
formula. Consequently, fair ROE testimony is not required. However, in pre-filed testimony,
23
the Company has filed rate of return testimony through the evidence of Dr. Vilbert. This has
24
then been justified by Dr Kolbe on the basis that what is important is the market determined
25
weighted average cost of capital or what they refer to as the “ATWACC” for after tax WACC.
10
11
12
2
3
Professor Berkowitz died August 8, 2004.
http://www.rotman.utoronto.ca/~booth.
-1-
1
The procedure is essentially that Dr. Vilbert estimates this ATWACC using market value
2
weights and costs for the sources of financing and Dr. Kolbe then implies a deemed common
3
equity ratio, which when combined with the allowed ROE gives the same ATWACC as that
4
estimated by Dr. Vilbert.
5
In my judgment this approach has no merit for determining either the allowed ROE or common
6
equity ratio for a regulated utility. Further this was also the judgment of the Alberta Energy and
7
Utilities Board (U99099, Vol 1, page 303) when it commented on similar ATWACC testimony
8
presented by Drs. Kolbe and Vilbert on behalf of TransAlta Utilities. The Alberta EUB stated
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
“The Board would be derelict in its statutory responsibilities to recognize market
capitalization ratios that are derived from a market value capitalization that deviates
from the intrinsic long-run value of the regulated firm. For example, if the Board has
traditionally set an allowed equity return based on book equity and this has resulted in a
equity market capitalization which is considerably above a ratio of one, an ATWACC
based on market capitalization ratios would call for a higher composite return.
Circularity could then develop in the process. To the extent that the higher return was
granted, the equity market capitalization would in all probability rise even further
which, if it went outside the narrow middle range, would again call for a higher
ATWACC. The Board must ensure that this potential circularity is avoided.
For all of the above reasons, the Board rejects the concept that an ATWACC
determined using market capitalization ratios of sample regulated firms is directly
transferable to the book assets of a regulated firm when the common equity market to
book ratio significantly exceeds the debt market to book ratio. In these situations, the
direct transfer of an ATWACC weighted with market capitalization results in a higher
return than a return determined using book capitalization weightings. The danger is that
the higher ATWACC return does not meet the regulatory and legislative test of fairness.
Accordingly, the Board finds it necessary to reject TransAlta’s version of the
ATWACC model which proposes the use of market capitalization ratios. Accordingly,
the Board considers that it should place primary weight on the traditional method in the
development of a fair return for these proceedings.”
31
In the Alberta EUB generic cost of capital hearing four years later, NGTL sponsored similar
32
ATWACC testimony by Drs Kolbe and Vilbert, but this time NGTL backed off from the
33
ATWACC approach. In its decision (U2004052, page 16) the Alberta EUB noted the following
34
exchange with Mr. Brett (of NGTL)
-2-
1
2
3
4
5
6
7
8
9
Q So, again, just to be clear, you’re not asking the Board to consider ATWACC in
terms of how it would set a fair return; moreover, it is being suggested by the
company that it is one of the tools it uses as, perhaps, a check in terms of what a
fair return would be; would that be a fair statement?
A. MR. BRETT: …..I think what I said, and what I intended to say, is we have not
asked the Board to use a return on capital or ATWACC for setting a revenue
requirement. We have applied for the traditional ROE on equity thickness.
10
The Alberta EUB’s decision (U99099) contained a particularly deep and thoughtful analysis of
11
the ATWACC approach and its relevance for a regulated utility. It is not surprising in the
12
circumstances that NGTL backed off from the ATWACC approach in the EUB generic
13
hearing. As I will show the Alberta EUB conclusion that “The danger is that the higher
14
ATWACC return does not meet the regulatory and legislative test of fairness,” is absolutely
15
correct.
16
Given its rejection by the Alberta EUB and the apparent withdrawal of support by NGTL I am
17
surprised that Union Gas would sponsor ATWACC based evidence before this board. I will
18
explain in detail the underlying logic and reasons why ATWACC based testimony is
19
inappropriate for a regulated utility, but the critical result is that setting the equity ratio or ROE
20
implicitly by using the (AT)WACC approach can “rubberstamp” existing market values that
21
may in turn reflect unfair and unreasonable rates. In many ways it is the complete opposite of
22
the regulatory requirement to determine fair and reasonable or just returns.
23
In response to Union Gas’ pre-filed testimony I have organised my testimony as follows. First,
24
I will review my understanding of regulatory policy and how it relates to the issues at hand.
25
Second, I will discuss the business risk of Union Gas from a capital markets perspective since
26
this is what is needed for determining a common equity ratio. Third, I will discuss typical
27
regulated financial structures in Canada, credit concerns, capital market access. Fourth, I will
28
discuss the issues raised by the company’s pre-filed testimony of the importance of WACC;
29
why it occupies such a prominent position in the finance literature. Finally, I will discuss the
30
procedures that have been suggested for going from the WACC to an implied equity cost with a
31
fixed equity ratio, or an implied equity ratio from a fixed equity cost. This latter section is
-3-
1
critical for the company’s pre-filed testimony, but as I will show I agree with the Alberta EUB
2
decision and do not recommend that this Board place any weight on these techniques as they
3
are unreliable and do not offer any value added over the normal, well accepted, principles
4
accepted by regulatory boards in Canada.
-4-
1
2.0
2
Q
WHAT RISKS DO INVESTORS FACE?
3
A.
Investors are interested in the rate of return on the market value of their investment. This
4
value can be represented by the standard discounted cash flow model:
5
REGULATORY TOOLS
P0 =
ROE * BVPS * (1 − b)
( K − g)
(1)
6
where P0 is the stock price, ROE the return on equity, BVPS the book value per share, b the
7
retention rate (how much of the firm’s earnings are ploughed back in investment). The product
8
of the ROE, BVPS and payout rate determine the dividend per share, which is then assumed to
9
grow at the rate g, which determines the future cash flow stream. This is then discounted back
10
at the investor’s cost of equity, or required rate of return, K.
11
The simple discounted cash flow (DCF) model is useful for thinking of the sources of risk to
12
the investor and the tools that the Board has available to it in managing that risk. Some of these
13
risks stem from the firm’s operations and financing, while others stem from the capital
14
market’s perception of the firm as well as general capital market conditions. For rate of return
15
regulated utilities we add another dimension to risk, which is the impact of regulatory risk. In
16
terms of the DCF equation the actual earned return on equity (ROE) captures the business,
17
financial and regulatory risk, and together I term these income risk, whereas all the other
18
factors are reflected in investment risk, which is the way in which investors react to this
19
income risk and other factors such as the firm’s growth prospects and exposure to interest rates.
20
It is important to realise that the Board can directly control income risk by its policies towards
21
the regulated firm. However, investment risk is beyond its direct control, even though the
22
Board can influence it, it can not control it. Think for example, about a Government of Canada
23
long term bond denominated in Canadian dollars. Such bonds are referred to as being default
24
free, since the government has complete control over the currency. As a result Government of
25
Canada bonds have no income risk. However, they do have investment risk. For example,
26
interest rates may increase causing the market value of the bonds to fall, or the rate of inflation
-5-
1
may be greater than expected so that the purchasing power of the bonds falls short of
2
expectations. In both cases investors lose either in nominal or real terms. Regulatory boards
3
have the same impact on the firms they regulate, like the Government of Canada they can take
4
measures to minimise if not remove income risk but they can not remove investment risk4
5
Q
6
A.
7
These risks are the typical risks stemming from uncertainty in the demand for the firm’s
8
product resulting, for example, from changes in the economy, the actions of competitors, and
9
the possibility of product obsolescence. This demand uncertainty is compounded by the method
10
of production used by the firm and the uncertainty in the firm’s cost structure, caused, for
11
example, by uncertain input costs, like those for labour or critical raw or semi-manufactured
12
materials. Business risk, to a greater or lesser degree, is borne by all the investors in the firm. In
13
terms of the firm’s income statement, business risk is the risk involved in the firm’s earnings
14
before interest and taxes (EBIT). It is the EBIT, which is available to pay the claims that arise
15
from all the invested capital of the firm, that is, the preferred and common equity, the long term
16
debt, and any short term debt, such as debt currently due, bank debt and commercial paper.
17
If the firm has no debt or preferred shares, the common stock holders “own” the EBIT, after
18
payment of corporate taxes, which is the firm’s net income. This amount divided by the funds
19
committed by the equity holders (shareholder’s equity) is defined to be the firm’s return on
20
invested capital or ROI, and reflects the firm’s operating performance, independent of
21
financing effects. For 100% equity financed firms, this ROI is also their return on equity
22
(ROE), since by definition the entire invested capital has been provided by the equity holders.
23
The uncertainty attached to the ROI therefore reflects all the risks prior to the effects of the
24
firm’s financing and is commonly used to measure the business risk of the firm.
WHAT ARE THESE INCOME RISKS?
Business risk is the risk that originates from the firm’s underlying “real” operations.
4
Sometimes provincial bonds have poorer bond ratings and sell on higher credit spreads than regulated
utilities in the same province.
-6-
1
As the firm reduces the amount of equity financing and replaces it with debt or preferred
2
shares, two effects are at work: first the earnings to the common stock holder are reduced as
3
interest and preferred dividends are deducted from EBIT and, second the reduced earnings are
4
spread over a smaller investment. The result of these two effects is called financial leverage.
5
The basic equation is as follows:
6
ROE = ROI + [ ROI − Rd (1 − T )]
D
S
(2)
7
where D, and S are the book values of debt and equity respectively, T is the corporate tax rate
8
and Rd is the embedded debt cost. If the firm has no debt financing (D/S =0), the return to the
9
common stockholders (ROE) is the same as the return on investment (ROI). In this case, the
10
equity holders are only exposed to business risk. As the debt equity ratio increases, the spread
11
between what the firm earns and its borrowing costs is magnified. This magnification is called
12
financial leverage and measures the financial risk of the firm.
13
The common stockholders in valuing the firm are concerned about the total “income” risk they
14
have to bear, which is the variability in the ROE. This reflects both the underlying business risk
15
as well as the added financial risk. If the firm operates in a highly risky business, the normal
16
advice is to primarily finance with equity. Otherwise the imposition of fixed financial charges
17
by the firm on top of the uncertainty in the firm’s EBIT might force the firm into serious
18
financial problems. Conversely, if there is very little business risk the firm can afford to carry
19
large amounts of debt financing, since there is very little risk to magnify in the first place.
20
In this fundamental sense business risk and financial risk work in opposite directions. Firms in
21
industries with very high business risk tend to finance primarily with equity, while firms with
22
very low business risk tend to finance with more debt. The best examples of the latter are the
23
banks and regulated utilities. Before going on it is important to note that the financial leverage
24
equation deals with the firm’s financial statements. Explicit in the derivation of the financial
25
leverage equation is that the firm’s total assets are constant all that is happening is a change in
26
the relative mix of debt and equity. This equation does not show how market values change or
-7-
1
how the investor’s required rate of return or cost of equity capital changes with the firm’s use
2
of debt. To determine this we need a valuation theory to show how the firm’s market value
3
changes with the debt equity mix. Unlike total assets in the financial leverage equation there is
4
no reason for the market value of the firm to remain constant as the firm changes its financing
5
mix.5
6
Q.
HOW CAN A REGULATOR RESPOND TO THESE RISKS?
7
A.
Regulators respond according to the legislation that establishes their jurisdiction. In BC
8
Electric Railway Co Ltd., vs the Public Utilities Commission of BC et al ([1960] S.C.R. 837),
9
the Supreme Court of Canada had to interpret the following statute:
10
11
(a)
The Commission shall consider all matters which it deems proper as affecting
the rate:
12
13
14
15
16
17
(b)
The Commission shall have due regard, among other things, to the protection of
the public interest from rates that are excessive as being more than a fair and
reasonable charge for services of the nature and quality furnished by the public
utility; and to giving to the public utility a fair and reasonable return upon the
appraised value of the property of the public utility used, or prudently and
reasonably acquired, to enable the public utility to furnish the service:
18
This statute articulated the “fair and reasonable” standard in terms of rates, and that the
19
regulatory body should consider all matters that determine whether or not the resulting charges
20
are “fair and reasonable.” Most rate base, rate of return, regulated companies in Canada are
21
subject to similar provisions. This means that a firm’s capital structure, its debt equity choice,
22
should be considered by a commission or board since it directly affects the rates charged by the
23
public utility every bit as much as the allowed rate of return.
24
The above provisions have one meaning to an economist: that rates reflect the operation of the
25
regulated utility at minimum long run average cost. Costs in a competitive market naturally
26
gravitate towards minimum long run average cost and by definition do not include charges that
5
Note that the financial leverage equation is not equivalent to the equity cost equations used by Dr.
Kolbe in his testimony. In the development above using book values the total assets are fixed by
definition, but in Dr. Kolbe’s development the market values are fixed by assumption.
-8-
1
are unfair and unreasonable (or unjust), while this cost ensures that the regulated services are
2
provided at a cost that promotes the overall efficiency of the economic system. Hence, what is
3
critical for a regulated utility is its cost structure, since this defines its business risk.
4
In its 2004 annual report (March 2005, available on http://www.sedar.com) Union Gas reported
5
total assets of $3,961 million, of which $3,059 consisted of property plant and equipment. The
6
balance was mainly balancing gas for direct purchase customers, inventories (mainly gas in
7
storage for customers) and trade receivables. What is important is that property plant and
8
equipment, hard tangible assets, comprise the vast bulk of Union’s total assets and largely
9
determine its cost structure. For 2004 Union Gas’ income statement was as follows:
10
11
12
13
14
15
16
17
18
19
20
21
22
23
$ million
Gas Distribution Margin
Storage and transportation revenue
Other revenue
Revenues net of gas costs
Operating and maintenance expense
Depreciation and amortisation
Property and capital taxes
Interest expense
Gain on sale of assets
Taxable income
Income taxes
Net Income
649
171
35
855
300
152
59
164
12
192
40
152
24
With total revenues net of gas costs of $855 million Union’s capital costs related to
25
depreciation, capital and property taxes, and interest were $375 million or 44%. If we add in
26
the net income since it is largely determined by the OEB adjustment formula these fixed capital
27
costs increase to 66.3%. Further the operations and maintenance costs are largely period costs
28
unrelated to Union’s gas throughout which means that Union’s marginal costs are very low.
29
Large immobile fixed costs and negligible marginal costs are the hallmarks of a regulated
30
industry. What this means is that an incumbent can lower prices to deter any possible new
31
entrant and that average costs are declining with output. This is why these types of companies
32
are regulated to mimic the pricing behaviour of competitive firms, rather than left to act like
-9-
1
monopolists. Further due to the capital intensive nature of their operations large amounts of
2
equity capital are contributed early and “locked in” creating a subsequent requirement to ensure
3
that the shareholders are treated fairly.
4
In the BC Electric decision the Supreme Court of Canada adopted Mr. Justice Lamont’s
5
definition of a fair rate of return as enunciated in the Northwestern Utilities Limited v. City of
6
Edmonton ([1929] S.C.R. 186) decision that:
7
8
9
10
“By a fair return is meant that the company will be allowed as large a return on
the capital invested in its enterprise (which will be net to the company) as it
would receive if it were investing the same amount in other securities possessing
an attractiveness stability and certainty to that of the company’s enterprise.”
11
Mr. Justice Lamont’s definition embodies what a financial economist would call a risk-adjusted
12
rate of return or “opportunity cost.”
13
The Board has accepted this requirement to allow an opportunity cost by linking the allowed
14
ROE to conditions in the long Canada bond market through its adjustment mechanism. The
15
Board imposed its Draft Guidelines on a Formula-Based Return on Common Equity and first
16
applied them in EBRO 495 for Consumers Gas (EGDI). These guidelines were subsequently
17
confirmed in RP2002-0158 (Decision January 2004, section 142) with the decision that
18
19
20
21
22
23
24
“Therefore, with respect to the first and primary issue of whether a new benchmark
ROE should be established for EGDI and Union, we find that the current ROE
Guidelines methodology continues to produce appropriate prospective results. We have
not found any demonstrated need to set a new benchmark ROE.
25
26
Q.
WHY HAVE YOU DISCUSSED THE ROE WHEN IT IS NOT AN ISSUE IN
THIS HEARING?
27
A.
Dr. Vilbert on behalf of Union Gas has filed voluminous rate of return testimony. His
28
main testimony consists of 77 pages of ROE testimony followed by an equity risk premium
29
appendix of 27 pages, a discounted cash flow (DCF) appendix of 18 pages and about 80 pages
30
of supporting exhibits. This ROE testimony is used to support the estimates of Dr. Kolbe and
- 10 -
1
Vilbert as to the appropriate ATWACC for Union Gas. The “twist” is that instead of
2
recommending an allowed ROE based on this testimony, with a fixed allowed ROE determined
3
by the Board’s formula, the estimated ATWACC is now used to recommend a common equity
4
ratio.
5
Although my estimates of an appropriate ROE would differ from Dr. Vilbert’s it is important to
6
note that his estimates are significantly lower than those normally presented by witnesses
7
testifying on behalf of the regulated company. For example looking at the Canadian utility
8
holding companies (UHCs) in his sample his equity cost estimates are as follows:
9
Canadian Utilities
Emera
Enbridge
TransCanada
Fortis
GMI
Average
DCF MJV-4
5.7
9.6
10.0
8.5
7.4
DCF Multi MJV-6
6.8
9.1
7.8
8.0
6.9
8.24
7.72
CAPM MJV-11
8.2
5.1
9.0
7.8
6.0
6.8
7.15
10
11
Where MJV-4 uses the standard DCF model based on perpetual growth, MJV-6 uses a multi-
12
stage growth model where growth is assumed to revert to long run GDP growth and MJV-11 is
13
Dr. Vilbert’s standard CAPM estimate. A simple average of these estimates of 7.70% is very
14
similar to the 7.75% I recommended for Terasen Gas in October 2005 in the BCUC’s review of
15
its adjustment mechanism.
16
I have testified before many regulatory boards in Canada and as far as I am aware Dr. Vilbert is
17
the only “company” witness who has presented ROE estimates this close to my own. Notably
18
Dr. Vilbert’s estimate of the equity cost of a Canadian UHC is similar to the sort of estimates
19
used by security analysts but is far below current allowed returns, which are about 8.80-9.0%
20
for 2006. If Dr. Vilbert, myself and security analysts are correct and allowed ROEs exceed the
- 11 -
1
investors’ fair rate of return or cost of equity capital, then we would expect that they would bid
2
up the market price and we would observe market to book ratios above 1.0.
3
The graph in Schedule 1 contains the market to book ratios for seven UHCs and Pacific
4
Northern Gas (PNG). What is clear is that the average market to book ratio is well above 1.0
5
and has recently been slightly below 2.0. This means that a dollar of shareholder’s equity is
6
being valued in the capital market at almost twice its book or contributed value. If these were
7
all normal pure regulated utilities then a market to book ratio of almost 2.0 would clearly
8
indicate that allowed ROEs were excessive. The appropriate regulatory response to Dr.
9
Vilbert’s ROE estimates and the observed market to book ratios would then be to lower the
10
allowed ROE.
11
A reduction in the allowed ROE is clearly not in the shareholder’s interest. Instead the thrust of
12
the testimony of Drs. Kolbe and Vilbert is to “cement” these high allowed ROEs and market to
13
book ratios. This is one of the by-products of estimating a market valued ATWACC that
14
explicitly weights the equity cost more highly by using market values. Before other boards they
15
have then taken the deemed common equity ratio as given and made a “leverage adjustment” to
16
recommend a higher ROE. The combination of the higher ROE and lower deemed book equity
17
ratio then supports their ATWACC estimate which has a lower equity cost, but a much higher
18
common equity ratio. However, before this Board they take the opposite approach by taking the
19
Board’s allowed ROE as fixed to derive a common equity ratio recommendation. However, in
20
both cases their recommendations are keyed off the high market to book ratio.
21
The result is that rather than reducing the allowed ROE because of the high market to book
22
ratio, the ATWACC approach produces either a higher ROE or a higher deemed equity ratio
23
despite what at first appears to be a reasonable equity cost estimate. If a regulator were to adopt
24
these recommendations then the result would be to reward the shareholders even more and
25
generate an even higher market to book ratio. This is why the Alberta EUB decision stated
26
27
“To the extent that the higher return was granted, the equity market capitalization would
in all probability rise even further which, if it went outside the narrow middle range,
- 12 -
would again call for a higher ATWACC. The Board must ensure that this potential
circularity is avoided.”
1
2
3
4
What is clear is that this Board should not use the ATWACC approach to indirectly set Union’s
5
common equity ratio.
6
Q.
WHY IS THE COMMON EQUITY RATIO IMPORTANT?
7
A.
The firm’s capital structure has a direct impact on the overall cost of capital. This topic
8
has been the subject of enormous academic inquiry over the last forty years and has generated
9
two Nobel Prize winners in Professors Franco Modigliani and Merton Miller. However, for all
10
the sophistication of the academic models, the most important issue is that certain types of
11
financial instruments have a tax-preferred status. In Canada this status is accorded debt
12
instruments, since interest payments are tax deductible, whereas equity dividends are not. As a
13
result, there is a built-in tax advantage to any corporation using debt financing. This tax
14
advantage goes to the shareholders of unregulated firms and to the customers of regulated
15
firms, since the use of debt reduces the firm’s revenue requirement. As will be discussed later,
16
this asymmetry in benefits for the regulated firm is a motivating factor behind regulated
17
companies continually striving to increase their equity ratios.
18
The primary fact to remember is that equity costs are paid out of after-tax income, whereas
19
debt costs are tax deductible. Hence, for example, if debt costs are 7.0% and equity costs are
20
9.0%, then at a 50% tax rate (for simplicity), the pre-tax costs are actually 18.0% for the
21
equity (.09/(1-.50)) compared to 7.0% for the debt. Conversely the after tax costs are 3.5% and
22
9.0%; either way the costs of debt versus equity have to be compared on the same tax basis. It
23
is these “same tax” cost comparisons, whether before or after tax, that competitive firms make
24
in deciding their financing. This implies that there is an incentive for competitive firms to
25
finance with debt: as they replace expensive equity with “cheap” debt, their cost of capital goes
26
down. Hence, for the same fixed amount of operating income, it is the stockholders who
27
benefit from the tax advantage of debt financing for competitive firms.
- 13 -
1
That publicly traded competitive firms do not finance with extremely high debt levels is due to
2
the increased costs of financial distress that are associated with higher fixed financial charges.
3
In extreme cases, the higher fixed financial charges can force a firm to be reorganised, or taken
4
over, when it could probably have otherwise survived had it been financed with less debt. As a
5
result, it is a basic rule of corporate finance that the financial risk is layered on top of business
6
risk: firms with high business risk are advised not to issue too much debt, otherwise their
7
solvency could be jeopardised in the event of adverse market developments. This is simply the
8
financial leverage principle discussed earlier.
9
This basic discussion is relevant since publicly traded firms are constantly re-assessing their
10
capital structures (“improving their balance sheets”) in light of changing market conditions and
11
the changing risk of financial distress. It also explains why capital structures differ from one
12
firm to another, since both the nature of their assets and expected cash flows are different. One
13
firm with mainly hard tangible assets will use large amounts of debt, since these types of assets
14
are easy to borrow against. Another firm that spends significant amounts on advertising will
15
have relatively little debt, since it is harder to borrow against brand names and “goodwill.” Yet
16
another firm will use very little debt, since it is not in a tax paying position and cannot use the
17
tax shields from debt financing. In each case, the firm will solve its own capital structure
18
problem based on its own unique factors.
19
This discussion puts the utility capital structure in perspective, since utilities have the lowest
20
business risk of just about any sector in the Canadian economy. Consequently, they should
21
have the highest debt ratios. There are several reasons for this:
22
23
24
25
First, the costs and revenues from gas distribution services are very stable so
that the underlying uncertainty in operating income is very low. As such
financial leverage is as I will show essentially magnifying almost non-existent
business risk, and zero times anything is still zero!
26
27
28
29
30
Second, in the event of unanticipated risks, regulated utilities are the only group
that can go back to their regulator and ask for “after the fact” rate relief. As
effective monopolies their rates can be increased in the event of financial
problems, while demand is typically insensitive to these rate increases. In
contrast, if unregulated corporations face serious financial problems they usually
- 14 -
1
2
3
compound one another. This is because unregulated firms encounter difficulties
raising capital and frequently suppliers and customers switch to alternates in the
face of this uncertainty creating severe financial distress.
4
5
6
7
8
9
10
Third, the major offset to the tax advantages of debt is the risk of bankruptcy.
In liquidation there are significant external costs that go to neither the equity nor
the debt holders. These costs include “knock down” asset sales, the loss of tax
loss carry forwards, and the reorganisation costs paid to bankruptcy trustees,
lawyers etc. This causes non-regulated firms to be wary of taking on too much
debt, since value seeps out of the firm as a whole. In contrast, it is highly
unlikely that Union’s distribution pipes would be ripped up and sold for scrap.
11
12
13
14
15
16
17
18
19
20
Finally, most private companies have an asset base that consists largely of
intangible assets. For example, the major value of Nortel was its growth
opportunities; of Coca Cola its brand name; of Merck its R&D team. It is
extremely difficult for non-regulated firms to borrow against these assets.
Growth opportunities have a habit of being competed away; brand names can
waste away, while R&D teams have a habit of moving to a competitor.
Regulated utilities in contrast largely produce un-branded services and derive
most of their value from tangible assets. Unlike intangible assets, tangible assets
are useful for collateral, for example in first mortgage bonds, and are easy to
borrow against.
21
Consequently, utilities have very low business risk; have reserve borrowing power by being
22
able to return to the regulator, minuscule bankruptcy/distress costs and hard tangible assets that
23
are easy to borrow against. In fact, in many ways, utilities are unique in terms of their financing
24
possibilities,6 and are prime candidates for using large amounts of debt to utilise their very
25
significant tax advantages.
26
Q
27
A.
28
MIT and in fact on page 1 references a Brattle company text co-authored by Dr. Myers. Dr.
29
Myers is also a co-author of a popular finance textbook, Fundamentals of Corporate Finance,
ARE THE ABOVE IDEAS STANDARD IN FINANCE?
Yes. Dr. Kolbe provides several references in his testimony to Dr. Stewart Myers of
6
When we analyse corporate financial decisions we normally include a number of explanatory
variables and then add a “dummy” variable for whether or not the industry is regulated, since the mere
fact of regulation is frequently the most significant feature of a firm’s operations.
- 15 -
1
McGraw Hill Irwin (3rd edition with R. A Brealey and A. J. Marcus). In chapter 15 the text
2
discusses capital structure and notes the following:
•
3
4
5
6
7
8
9
•
(Page 434) “Debt financing has one important advantage. The interest that the
company pays is a tax deductible expense, but equity income is subject to
corporate tax.”
(page 434 and 435) The interest tax shield is a valuable asset. Let’s see how
much it could be worth…………………….If the tax shield is perpetual, we use
the perpetuity formula to calculate its present value:
annualtaxsheild
= Tc D
rdebt
(page 435, 436) How interest tax shields contribute to the value of stockholder’s
equity….
PV tax shields =
10
•
11
12
13
14
15
16
17
18
•
Value of levered firm = value of all-equity firm + TCD
(Page 444) For example, high-tech growth companies, whose assets are risky
and mainly intangible, normally use relatively little debt. Utilities or retailers
can and do borrow heavily because their assets are tangible and relatively safe.
19
These four particular comments are taken from the discussion of what is commonly referred to
20
as the static trade-off model, where the tax advantages of debt financing are traded off against
21
the costs of financial distress. They are referenced simply because there is little disagreement
22
within the profession, even amongst Brattle authors, that debt is valuable to the firm and why it
23
is valuable are the tax shields it generates. As the second point indicates if debt is rolled over,
24
so that the interest and tax shields are expected to continue indefinitely, then the value of the
25
tax shield is the amount of debt times the corporate income tax rate. The third quote indicates
26
that the value of the firm is increased by the present value of these tax shields. In fact the
27
equation referenced there is part of an approach pioneered by Dr. Myers called the adjusted
28
present value approach (APV). The final quotation even has Dr. Myers specifically mentioning
29
utilities as companies that should borrow heavily.7
7
The text does note an “odd fact” that profitable companies like Merck could borrow more since Merck
has the highest possible credit rating and pays income tax. However, the text fails to note Merck’s
- 16 -
1
It is obvious that academics associated with the Brattle group when writing textbooks for the
2
academic audience have no problem noting the tax advantages of debt and expounding on
3
standard ideas in finance. Further on page 437 the Myers et al text also has the standard graph
4
showing that the weighted average cost of capital (ATWACC in this hearing) declines as the
5
firm uses more debt, that is, the ATWACC is not constant as assumed by Dr. Kolbe.
6
7
8
Q.
ARE THERE SPECIAL PROBLEMS WHEN UTILITIES ARE PART OF
HOLDING COMPANIES?
9
A.
Yes and these are relevant since Union is owned indirectly by Duke Energy
10
Corporation (“Duke”) a major US energy company. As indicated above, and in the quotes from
11
Dr. Myers’ textbook, if there is a tax advantage to using debt, then competitive firms will use
12
debt. However, for ROE regulated utilities this tax advantage flows through to the consumer as
13
a lower tax charge in the revenue requirement. As a result there is no advantage to the utility
14
using debt. However, for utilities owned by UHCs the situation is worse, since the parent has
15
an incentive to finance the utility with as much equity as possible, so that the tax advantages to
16
debt are shifted to the parent. In this way it is the UHCs shareholders that get the tax
17
advantages instead of the utility ratepayers. This is often called the “double leverage” problem,
18
where the utility assets support debt at both the utility level and then again at the parent level.
19
This situation has recently become worse as some rating agencies, such as Standard and Poors,
20
rates debt based on the credit rating of the parent. The principle here is that if the parent gets
21
into trouble it will raid the subsidiary unless it is “ring fenced” or insulated from the parent in
22
some way. Without this ring fencing the subsidiary is as risky as the parent regardless of its
23
debt ratio. Consequently, double leverage can not just transfer the tax advantages to the
24
parent’s shareholders, but it also may result in lower bond ratings and a higher debt cost for the
25
utility. As a result, utility rate-payers lose part of the debt tax shield and to add insult to injury
26
may also pay for a higher cost of debt, thus getting hit twice.
potential off balance sheet liabilities. I would imagine that in the next edition potential lawsuits related
to dangerous drugs like Vioxx will be mentioned as a reason for Merck’s financing policies.
- 17 -
1
Q.
HOW DO THESE COMMENTS APPLY TO UNION GAS?
2
A.
In EBRO 493/4 testimony Dr. Berkowitz and I made the following comments (Page
3
13):
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
“Westcoast Energy (WEI), the parent of both Centra and Union for example, has partly
financed the acquisition of its holdings through borrowing against its investment in its regulated
subsidiaries. In a series of remarks in its credit reports DBRS has pointed out the “double
leveraging” by WEI of its regulated assets. In its August 1, 1995 credit report DBRS stated
“The approximately 25% common equity component of consolidated capitalization (of WEI) is
about 8 percentage points lower than the average approved common equity components of rate
base of the WEI regulated utilities.” In September 1997 DBRS (page 3) again went on to state
“WEI’s non-consolidated capital structure includes 33% debt. This is projected to rise to 39%
by the end of 1997. Given that the remaining assets are comprised of investments in
subsidiaries, this represents double leveraging at the holding company level.” Finally, in its
May 1998 report (page 3) DBRS states “Consolidated debt to capital of about 68% reflects
double -leveraging at the holding company level. While coverage ratios are adequate on a
consolidated basis interest coverage remains weak on a non-consolidated basis.”
WEI’s actual consolidated equity ratios for the last three years have been:
Ratio
Common equity
Preferred equity
Per cent
1997
24.7
6.6
1996
25.8
5.8
1994
22.3
3.3
where the common equity ratio includes minority interest as common equity. WEI is a
competitive firm freely choosing to finance its operations with 25% common equity while
maintaining an A(low) investment grade bond rating. Moreover, it has the flexibility to lower its
common equity ratio from approximately 33% down to 25% by borrowing against the regulated
assets of its subsidiaries. However, these assets have already been used to support debt at the
subsidiary level, which gives rise to to what DBRS calls “double leverage”, or what we have
referred to as the “debt capacity transfer” problem.
33
There is no question that Union Gas’ parent, at that time, Westcoast Energy (WEI), borrowed
34
against its Union Gas assets while still maintaining an investment grade bond rating. Moreover
35
this was well recognised by DBRS. Dr. Berkowitz and I went on to point out that there was a
36
$433 million shortfall in equity at the holding company level, that is, since WEI’s assets were
37
90% regulated its consolidated common equity ratio should have been approximately the 33%
38
average of its operating subsidiaries, but it wasn’t. Further we stated that:
- 18 -
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
“The second reason could be that the regulated assets are properly leveraged and WEI’s
additional leverage causes the debt to be issued at non-investment grade bond ratings. This
would imply that none of the regulated utilities’ credit is being left on the table. However,
WEI’s credit rating over this period was upgraded by CBRS from B++(high) to A(Low) in 1992
and has been A(low) with DBRS throughout the period. Evidently, throughout this period WEI
has maintained an “A” bond rating, significantly above the lowest investment grade rating in
Canada of B++(low). Interestingly, its interest coverage ratios (based on net interest)
throughout this period have been as follows:
18
I have referenced the above remarks made by Dr. Berkowitz and I almost ten years ago since
19
Union’s common equity ratio has increased from the 29% of that period to first 34% and then
20
35% when consolidated with Centra Gas Ontario, until now, with a new parent with a worse
21
credit rating than WEI to a requested 40% common equity. If for simplicity we use $4 billion
22
as Union’s rate base then this means that common equity will have increased from $1.16 billion
23
to $1.36 billion to $1.4 billion to $1.6 billion, that is, at a 40% common equity ratio Union
24
would have $440 million more in equity than it would have had with its historic 29% common
25
equity ratio or $200 million more than with its current 35% allowed common equity ratio.
26
If this increased common equity ratio simply replaced what would have been debt, then the
27
value of the tax shields transferred from Union to its parent using Dr. Myers’ tax shield
28
formula and Union’s statutory tax rate for 2004 of 36.1% would have been ($440 million times
29
36.1%) $159 million. Even taking into account that some of the common equity replaced
30
preference shares, the current requested increase in common equity of the order of $200 million
31
represents a transfer in tax shield value of about $72.2 million. This is a direct transfer in value
32
from rate-payers to Union’s shareholders.
33
It also has to be remembered that Union Gas’ new parent Duke is not as financially strong as
34
WEI. DBRS rated WEI at A-, but S&P rates Duke BBB. Duke had disastrous losses of almost
1997
DBRS
1.73
1996
1.81
1995 1994
1.81
1.62
1993
1992
1.65
1.61
These coverage ratios are significantly below the “benchmarks” that are usually stated to be
necessary for an “A” rating, yet WEI’s ratings are described as stable and have recently been
confirmed. Moreover, DBRS specifically refers to these consolidated interest coverage ratios as
“adequate.”
- 19 -
1
US$3 billion in 2003 and its stock price dropped precipitously from over $45 down to barely
2
$10 as the following chart from E-trade indicates. Since then Duke has focussed on its core
3
business and is currently proceeding with a merger with Cinergy that seems consistent with a
4
traditional utility focus. However, it remains a low rated US energy company that needs to
5
shore up its finances. In my judgment Union Gas is in more danger from Duke than it was from
6
the double leverage used by WEI. I will discuss bond ratings and ring fencing later, but it may
7
not be an accident that Union is asking for an increase in its common equity ratio so shortly
8
after being taken over by a risky, low rated, diversified, US energy company.
9
10
11
12
13
14
Q.
HOW DO YOU RECOMMEND THE BOARD SET UNION’S COMMON
EQUITY RATIO?
15
A.
Several boards have recently looked at setting common equity ratios. In RH-2-94 the
16
National Energy Board stated (Decision page 24)
- 20 -
1
2
3
4
5
6
“The Board is of the view that the determination of a pipeline’s capital structure starts
with an analysis of its business risk. This approach takes root in financial theory and has
been supported by the expert witnesses in this hearing. Other factors such as financing
requirements, the pipeline’s size and its ability to access various financial markets are
also given some weight in order to portray, as accurately as possible, a complete picture
of the risks facing a pipeline ”
7
I agree with this assessment, since it follows the prior discussion of the impact of financial
8
leverage. To repeat the previous financial leverage equation
9
ROE = ROI + [ ROI − Rd (1 − T )]
D
S
(2)
10
If this equation is rearranged we can express the variability of the ROE as a function of the
11
variability in the operating income or
STDEV ( ROE ) = STDEV ( ROI ) * (1 +
12
D
)
S
(3)
13
where the standard deviation of the realised ROE is that of the realised ROI times one plus the
14
debt equity ratio. Consistent with the previous discussion, if business risk is the variation in the
15
ROI, then financial leverage as measured by the debt-equity ratio magnifies this risk.
16
The National Energy Board then took the next step of equalising the risk to the equity holders
17
(STDEV(ROE)) by offsetting business risk differences ((STDEV(ROI)) with different capital
18
structures (debt equity ratios). After doing this they then allowed every pipeline the same
19
formula allowed ROE. Further, this is also why they recently increased the TransCanada
20
Mainline’s deemed common equity ratio from 30% to successively 33% and 36%. In both
21
cases it was in response to a perception of increased business risk due to supply concerns in the
22
Western Canadian Sedimentary basin (WCSB) and the entry into service of the Alliance
23
pipeline, both of which have caused the load on the Mainline to drop below 100%. The
24
important point is that once this perception of increased risk had been adjusted for by reducing
25
the Mainline’s financial risk, the Mainline could continue to be allowed the NEB formula ROE.
- 21 -
1
It is important to note that the above equation is based on the firm’s financial statements. It is
2
an accounting relationship that has nothing to do with how the stock market reacts to the firm’s
3
use of financial leverage. As far as I know no-one has ever disputed the above equations, as
4
they are simply a rearrangement of the flow of income through a firm’s financial statements.
5
That is, the ROE is not the investors cost of equity capital or required rate of return. It is simply
6
the ROE earned on the book value of the shareholder’s equity. Using these relationships is
7
consistent with the fact that the Board can only control these accounting values. The Board can
8
alter business risk through the use of deferral accounts and the financial risk through changes in
9
the deemed equity ratio, but it can not change stock market risk, as the market, not the Board,
10
determines market values.
11
This last point should be emphasised: the financial leverage equation is not equivalent to the
12
formulae used by Dr. Kolbe. Dr. Kolbe’s equity cost adjustment formulae are based on
13
assumptions about how the stock market values the use of financial leverage. Unlike the
14
financial leverage equation, the equations used by Dr. Kolbe attempt to answer the question
15
how does the investor’s required rate of return or cost of equity capital change with a change in
16
the use of debt? It is not necessary for the Board to answer this question or even consider it
17
when changing the deemed common equity ratio in response to a change in business risk.
18
To illustrate when the NEB set up its regulatory system in RH-2-94 several experts submitted
19
testimony on how the allowed ROE should change as the capital structure changes along the
20
lines of the current testimony of Drs. Kolbe and Vilbert. Dr. Sherwin and Ms. McShane, who
21
provided testimony on behalf of the companies, and who have testified before this Board many
22
times concluded (page 24)
23
24
25
“The finance models, even when adapted to the real world of Canadian utility
regulation, cannot provide the basis for determining a pipeline’s optimal capital
structure.”
26
Dr. Berkowitz and I also used models similar to those used by Dr. Kolbe, but expressed little
27
support for them. As the NEB noted in its Reasons for Decision (page 24):
- 22 -
1
2
3
4
5
“Dr. Booth and Berkowitz concluded that these estimates are approximately the
increases in ROE required by investors. However, they noted the estimates are subject
to error since they are based on valuation formulas, which are as yet unproven.
Moreover, they noted that these formulas ignored the non-tax advantages of debt
financing and the effects of financial distress.”
6
Finally, the NEB also noted Dr. Waters’ testimony (a frequent witness before the NEB at that
7
time) where he indicated that “To date empirical testing to more clearly describe the
8
relationship (between capital structure and the investors required return) has not been done
9
successfully.”
10
The NEB’s summary from over ten years ago is an accurate assessment of my views today and
11
it is still my judgment that the misgivings expressed by expert witnesses over ten years ago
12
continue. Despite the seeming precision of the estimates provided by Drs. Kolbe and Vilbert no
13
other expert witnesses have based their testimony on this approach, since the views of most
14
experts is that it can not be done accurately. In its Decision on EBRO 493/4 (page 198) the
15
Board stated that
16
17
18
19
20
21
22
“The Board finds Union’s capital structure, which recognises changes in preference
share capital, tax accounting, and includes a 34% common equity component as
recommended by the ADR settlement agreement to be appropriate for the 1997 test
year. Should the LGIC approve the companies’ merger application, the Board expects
Union and Centra to fully justify from first principles, in the 1998 rates case, the
proposed capital structures of the amalgamated companies.”
23
Professor Berkowitz and I provided testimony in the subsequent case, along with Dr. Cannon
24
and Ms. McShane. This was done largely on the basis of business risk assessments and I would
25
recommend that the Board continue to make its capital structure decision changes based on
26
changes in business risk.
27
28
29
- 23 -
1
2
3
4
Q.
IS BUSINESS RISK THE ONLY FACTOR IN SETTING CAPITAL
STRUCTURES?
A.
No. Ultimately the litmus test of whether a board has “got it right” is whether the
5
regulated company can access capital on reasonable terms. If, for example, a common equity
6
ratio is inadequate then the stock market will take note of the increased financial risk and make
7
it difficult for the regulated firm to access capital on reasonable terms. In Federal Power
8
Commission et al v. Hope Natural Gas Co. [320 US 591, 1944], the United States Supreme
9
Court decided that a fair return
“should be sufficient to assure confidence in the financial integrity of the
enterprise so as to maintain its credit and to attract capital.”
10
11
12
Although the Hope “financial integrity” criteria flows from considering a fair return it applies
13
equally to the deemed common equity ratio. In my judgment an appropriate common equity
14
ratio is one which, in conjunction with the allowed return, allows a regulated company to
15
maintain its credit and attract capital.
16
The Hope criterion would therefore support the view that after examining business risk, the
17
Board consider factors such as size, financing requirements and market access, since all of
18
these are important for financial integrity. However, note that “maintaining credit” is not the
19
same as maintaining a particular credit rating. Credit standards constantly change as does the
20
market’s appetite for certain types of credits. This means that there is no need to target a
21
particular credit rating. What is important is that a utility can access the capital markets on
22
reasonable terms to raise capital and provide service.
23
Q.
IS THERE ANY OTHER RECENT DECISION THAT SUPPORTS THIS?
24
A.
Yes, the recent Alberta Generic Hearing established not just an adjustment formula to
25
set the allowed ROE, but also the allowed common equity ratios for eleven distinct regulated
26
entities in a range of ROE regulated businesses including pipelines (ATCO Pipe and NGTL).
27
The EUB stated (Generic Cost of Capital Decision page 35)
- 24 -
1
2
3
“To determine the appropriate equity ratio for each Applicant, the Board will consider the
evidence and, where applicable, the experts’ views and rationales in each of the following
topic areas:
4
1.
The business risk of each utility sector and Applicant;
5
2.
The Board’s last-approved equity ratio for each Applicant (where applicable);
6
3.
Comparable awards by regulators in other jurisdictions;
7
4.
Interest coverage ratio analysis; and
8
9
5.
Bond rating analysis.”
10
This approach of the Alberta EUB seems to be substantially the same as the traditional
11
approach used by this Board and the NEB. I therefore first look at the business risk of Union
12
Gas and then consider financial market access and the EUB points 2-5.
13
- 25 -
1
3.0
BUSINESS RISK
2
Q.
HOW DO YOU VIEW THE BUSINESS RISK OF UNION GAS?
3
A.
In the discussion of risk in Section 2 I pointed out that income risk to the investor is a
4
function of business risk and financial risk. However, I then clarified that financial risk is often
5
set to modify underlying business risk, so in this sense capital structure is a regulatory tool used
6
by the Board. However, the Board has other tools in addition to simply setting the deemed
7
equity ratio and allowed ROE. In fact the whole regulatory process changes the risk of
8
investing in regulated industries, to the extent that when finance researchers try to model what
9
determines capital structures they frequently either exclude regulated industries completely or
10
add a “dummy variable” (a zero vs a one) for a regulated firm, simply because the act of
11
regulation itself explains more than can be explained by independent variables such as revenue
12
variability.8
13
For the above reasons we can not look simply at variability in returns or earnings as a complete
14
measure of risk. For regulated industries we look at the ability of the firm to earn its allowed
15
ROE. If a regulated firm consistently earns its allowed ROE this is a strong indication that the
16
utility has very little risk. On the other hand a utility that has difficulty earning its allowed ROE
17
is not generating returns to the shareholders consistent with what they could earn elsewhere. As
18
such the share price can be expected to drop and the firm may have difficulty attracting capital
19
on reasonable terms.
20
In interrogatory response J2-31 Union was asked for data on its weather normalised and actual
21
ROE back to 1980. In answer they provided data for 2003-2005 and a response to a similar IR
22
(VECC 46) from RP-2002-0158, but stated that data back to 1980 is not available. However,
23
Dr. Canon provided data back to 1985 in the same hearing, which should have been available
8
The regulatory process determines the revenue requirement, so that variability in revenues caused by
changing the allowed ROE or common equity ratio, for example, is not a risk factor in the way it would
be for a competitive firm.
- 26 -
1
to the company. The following graphs the allowed versus weather normalised actual ROE. The
2
most important insight from the graph is that Union not only earns its allowed ROE but
3
generally exceeds it. Over the entire period Union Gas earned 0.96% more than allowed based
4
on weather normalised returns. Prior to 1997 Union was only allowed a 29% common equity
5
ratio and for the period 1985-1996 Union over-earned by 0.79%, whereas since then it has
6
over-earned by 1.18%. In contrast the variability of this over-earning has decreased from
7
1.19% to 0.74%. The movement to a higher common equity ratio has coincided with both
8
persistently higher and more stable over-earning.
Union's Allowed vs Actual ROE
(Weather normalised)
Allowed
20
05
20
03
20
01
19
99
19
97
19
95
19
93
19
91
19
89
19
87
19
85
20
18
16
14
12
10
8
6
Actual
9
10
11
Over the whole period there is little evidence of risk attached to Union Gas. At its most, basic
12
risk is the probability of incurring a loss and yet on only three occasions has Union suffered a
13
loss and in two of those years it was minimal, only in 1991 when Union earned 1.0% less than
14
its allowed ROE does the loss appear significant. In contrast, in the 18 years when it over-
15
earned, Union earned over 2.0% more on four occasions and 1.0% more on 9 occasion or
16
almost half the time. Most competitive firms would die to have a worst case scenario as one
17
where they earned 1.0% less than their “target” ROE.
- 27 -
1
Q
IS THIS PERFORMANCE NORMAL FOR REGULATED FIRMS?
2
A.
Yes. ROE regulated firms have minimal risk in Canada due to the very high degree of
3
regulatory protection. In Schedule 2 is a table of earned vs allowed ROEs for the pipelines that
4
are part of TransCanada Corporation. There is a distinction between full cost of service
5
pipelines regulated by the National Energy Board and those regulated on a forward test year
6
basis similar to Union Gas. Foothills, for example, bills its shippers for its full costs and exactly
7
earns its allowed ROE, to the extent that it only reports one number in its surveillance reports
8
to the NEB. The TransCanada BC system (formerly ANG) is regulated on a similar basis to
9
Foothills and the only difference is that on its full acquisition by TransCanada there were some
10
reorganisation costs it absorbed so in 2003 it “voluntarily” under-earned its allowed ROE. I
11
have always regarded Foothills and ANG as the lowest risk regulated entities in Canada, since
12
there is NO income risk at all. They consistently earn exactly their allowed ROE so after the
13
fact there has been no business risk attached to these two pipelines. Without any business risk,
14
both these pipelines can finance with large amounts of debt, in fact prior to RH-2-94 they
15
financed with 25-28% common equity with the balance conventional debt.
16
Unlike Foothills and ANG the TransCanada Mainline and TQ&M are regulated on a forward
17
test year basis similar to Union. This leaves the companies exposed to forecasting risk where
18
the actual revenues and expenses may deviate from those expected and included in the revenue
19
requirement. However, the use of deferral accounts and long term contracting with shippers
20
that pay fixed demand charges, regardless of whether or not they ship, significantly reduces this
21
forecasting risk. The result is that both the Mainline and TQ&M consistently over-earn their
22
allowed ROEs. Over this whole period the Mainline only failed to earn its allowed ROE once
23
and on average over-earned by 0.25%, whereas TQ&M over-earned by 0.37% and never failed
24
to earn its allowed return.
25
In Schedule 3 is similar data for Union Gas, Enbridge Gas Distribution and Terasen Gas. The
26
data for the Union and EGDI is based on weather normalised ROE’s, since these utilities are
27
not allowed deferral accounts for variances due to weather. In contrast, Terasen is allowed a
28
comprehensive RSAM that takes into account not just the cost of purchased natural gas but also
- 28 -
1
volume variances due to weather. Of note is that Terasen’s over-earning is similar to that of the
2
TransCanada Mainline.9 In contrast Union and EGDI do not have as many deferral accounts
3
and over-earned to a much higher degree than either the TransCanada mainline or Terasen, let
4
alone the full cost of service pipelines.
5
If risk is the possibility of incurring harm or a loss the insight from the data in Schedules 2 & 3
6
is that regulated utilities in Canada have very little risk. It is also interesting that the degree of
7
over earning decreases with the use of deferral accounts. The full cost of service pipes can be
8
regarded as having 100% protection, since they neither over or under-earn. The Mainline and
9
TQ&M have limited room to improve their earnings, since so many of their revenues and
10
expenses are fixed. Similarly Terasen Gas with comprehensive deferral accounts looks a lot
11
like the NEB forward test year pipes in having little room to over-earn. In contrast, the two
12
Ontario LDCs with fewer deferral accounts have over-earned the most. It is hardly surprising
13
given this performance that Union does not feel that its exposure to large industrial customers
14
warrants a deferral account (interrogatory response J2-22) and that it is requesting removal of
15
some existing deferral accounts. In my judgement fewer deferral accounts in practise means
16
greater over-earning, which is why I advocate their use in areas that are largely beyond
17
management’s control such as the weather.
18
It is also interesting to contrast this performance of regulated assets with the utility holding
19
companies (UHC) that actually face the market. For the major UHCs Schedule 4 gives their
20
earned ROEs along with those of the TCPL Mainline and Foothills. For example, what
21
investors invest in as “TransCanada” or TCPL is not the Mainline, but the combined entity
22
including non-regulated and regulated assets. This can be seen in the greater variability of its
23
ROE. For 1993-1997 TCPL consistently earned more than the Mainline, but then in 1998-2000
24
as TCPL reorganised it earned less than the Mainline. Throughout this period the Mainline has
25
underpinned TCPL’s results and been a beacon of stability. One way of assessing this greater
26
risk is simply to estimate the standard deviation in each firm’s ROE. For the TCPL Mainline
9
Since 1998 Terasen’s actual ROE is prior to earning sharing.
- 29 -
1
this was 1.05%, whereas for TCPL itself it was 2.63%, so the Mainline’s ROE was only 40%
2
as variable as that for the whole company. However, as we have seen this variability in the
3
Mainline’s ROE is not “risk,” since it largely reflects the fluctuation in the Mainline’s allowed
4
ROE.
5
I can estimate the risk relative to the allowed ROE by taking Foothills ROE as the benchmark
6
and then look at both the average deviation of actual ROEs from this benchmark as well as the
7
variability of this deviation. This is included in the table in Schedule 5, where, by definition,
8
Foothills has a zero deviation every year. For example, I showed earlier that the Mainline has
9
over earned by about 0.25% relative to its allowed ROE, the above table shows that relative to
10
Foothills allowed ROE, it over earned by about 30 basis points and the standard deviation of
11
this deviation was 0.27%. In contrast TransCanada Corporation earned slightly more than the
12
Mainline but the variation in the deviation from Foothill’s allowed ROE is much greater at
13
1.98%. This fluctuation in TransCanada’s (TCPL) actual ROE reflects the greater risk
14
stemming from its periodic forays (and retreats) into non-regulated areas. The only one of these
15
UHCs that is close to Foothills in terms of risk is GMI. As a partnership it receives its 50%
16
share of TQM and its Quebec gas distribution assets seemingly without actually paying the
17
income tax it collects, so its ROE is consistently higher than for the other UHCs. However,
18
what is important is that the variability in its ROE around the Foothills benchmark is 0.60%,
19
only slightly higher than the TCPL Mainline.
20
Based on this variability around the Foothill’s allowed ROE, I would rank these UHCs in the
21
following order, from lowest to most risky, compared to the Foothills benchmark: GMI,
22
Emera, and then Fortis, TCPL, PNG & CU as a group followed by Enbridge, Terasen Gas Inc
23
and TransAlta. However, what it points out is that these UHCs are riskier than their underlying
24
regulated assets and adjustments have to be made in applying insights from the UHCs to the
25
regulated assets. In particular I have no doubt that all of these UHCs are riskier and thus have a
26
higher WACC than the underlying regulated assets.
27
28
29
Q.
WHAT DOES IT MEAN FOR THE EVIDENCE OF THE COMPANY IF UHCS
ARE RISKIER THAN THE UNDERLYING REGULATED ASSETS?
- 30 -
1
A.
The core of the testimony of Dr. Vilbert is to estimate the WACC from a sample of
2
UHCs and use them as a proxy for Union Gas. As I have demonstrated above there is little
3
doubt that the Canadian UHCs are riskier than their underlying regulated assets due to their
4
periodic misadventures in non-regulated areas. This UHC risk will be reflected in their higher
5
WACC. In turn using the methodology of Dr. Kolbe this must result in a higher deemed
6
common equity ratio. Further in interrogatory response J2-10 Dr. Vilbert was asked what
7
adjustments he made for the higher risk of the UHCs, and the answer was none. Further Dr.
8
Vilbert admitted to doing no tests to see whether his sample of US UHCs were comparable to
9
Canadian UHCs, let alone Canadian regulated assets.
10
Q.
WHAT COMPARATORS WOULD USE FOR UNION GAS?
11
A.
Before the Alberta EUB in 2003 I compared the different utilities in the Alberta generic
12
hearing on the following basis:
13
I:
The major short term risks caused by cost and revenue uncertainty:
14
15
16
17
•
On the cost side since regulated utilities are capital intensive most of their costs
are fixed. The major risks are in operations and maintenance expenditures.
However, over runs are usually under the control of the regulated firm and can
be time shifted between different test years.
18
•
On the revenue side the risks largely stem from rate design, critical features are:
19
20
21
22
23
24
o Who is the customer and what credit risk is involved. For example,
electricity transmission operators who recover their revenue requirement in
fixed monthly payments from the provincially appointed TA, who is
responsible for system integrity, have less exposure than the local gas and
electricity distributors who recover their revenue requirement from a more
varied customer mix involving industrial, commercial and retail customers.
25
26
27
28
o Is there a commodity charge involved? The basic distribution function is
very similar to transmission, except when the distributor buys the gas or
electricity wholesale and then also retails the commodity. The distributor is
then exposed to weather and price fluctuations depending on rate design.
29
30
o Even if there is no commodity charge, how much of the revenue is recovered
in a fixed versus a variable usage charge? Utilities that recover their revenue
- 31 -
in a fixed demand charge face less risk than those where the revenues have a
variable component based on usage.
1
2
3
II:
The medium and long term risks are mainly as follows:
4
5
6
7
8
9
10
11
12
13
•
Bypass risk. The economics of regulated industries are as natural monopolists
involved in “transportation” of one kind or another. However, one utility may
not own all the transportation system so that it may be economically feasible to
bypass one part of the system. This happens for local gas distributors, when a
customer can access the main gas transmission line directly, rather than through
the LDC, or when a large customer may be able to bypass part of the
transmission system. This is often a rate design issue: a postage stamp toll
clearly leads to uneconomic tolls and potential bypass problems, whereas
distance or usage sensitive tolls will discourage it. Similarly, rolled in tolling
will encourage predatory pricing by potential regulated competitors.
14
15
16
17
18
19
•
Capital recovery risk. Since most utilities are transportation utilities, the critical
question is the underlying supply and demand of the commodity. If supply or
demand does not materialise then tolls may have to rise and the utility may not
be able to recover the cost of its capital assets. Depreciation rates are set to
mitigate this risk to ensure that the future revenues are matched with the future
costs of the system.
20
A common thread running through the above brief discussion is rate design and regulatory
21
protection. There can be significant differences in underlying business risk that are moderated
22
by the regulator in response to those differences. The lowest risk utility is then one with the
23
strongest underlying fundamentals and the least need to resort to regulatory protection. In
24
contrast, another utility may have similar short term income risk, but only because of its need
25
to resort to more extensive regulatory protection, so that it faces more problematic longer term
26
risks.
27
On this basis I judged the lowest risk regulated utilities in Canada to be electricity transmission
28
assets, since these have the following characteristics:
29
30
31
32
33
34
•
•
•
•
•
Minimal forecasting risks attached to O&M
Revenue recovery via the TA through fixed monthly charges
Limited (non existent) by-pass problems
Minimal capital recovery problems, since there are many suppliers of electricity
as a basic commodity.
Deferral account for capital expenditures
- 32 -
1
and recommended 30% common equity ratios.
2
I then placed the gas transmission pipelines as the second lowest risk group. Here I classified
3
Foothills and the TCPL BC System (formerly ANG) as of equivalent risk to electricity
4
transmission assets with NGTL having marginally more risk than Foothills and the TCPL BC
5
System, since it is exposed to bypass and recovers its revenues through a forward test year
6
from a greater variety of shippers. However, the combination of distance sensitive tolls, the
7
ability to offer load retention service and a more rapid depreciation rate significantly reduce
8
any increase in risk NGTL may have faced since 1995. I therefore judged that on its own
9
NGTL could maintain its financial flexibility on the same 30% common equity ratio allowed
10
mainline gas transmission assets. However, because NGTL was then allowed 32% and was
11
almost “indistinguishable” from the TCPL Mainline, I recommended the same 33% common
12
equity ratio then allowed the Mainline.
13
I then judged the local distribution companies (LDCs), including both gas and electric as the
14
next riskiest. These companies are distinguished by their retail operations, which mean that
15
their revenues are recovered from a large number of industrial, commercial and residential
16
consumers. This exposes them to both the business cycle and weather fluctuations. This
17
revenue recovery is also a function of their rate design that may expose them to commodity
18
charges and a fixed and variable recovery charge. Within this group the conventional yardstick
19
for LDCs is that Consumers (Enbridge Gas Distribution Inc or EGDI) and Union Gas are both
20
allowed 35% common equity by the Ontario Energy Board. However, whereas the Ontario
21
Energy Board allows a purchased gas variance account (PGVA) to ensure that the full costs of
22
gas are recovered, they are still subject to volume related variances. In contrast, the BCUC
23
allows BC Gas (Terasen Gas) a more comprehensive deferral account, but limits the allowed
24
common equity ratio to 33%. With these yardsticks I recommended 35% common equity ratio
25
for a typical local distribution companies.
26
Finally, I recommended 42% as the upper end of a reasonable range for the common equity of
27
ATCO pipelines, given that the BCUC allows PNG, a smaller and much riskier pipeline, 36%
28
common equity. However, this ranking was provisional being dependent on the EUB
- 33 -
1
developing clear rules on intra Alberta pipeline competition and a rate design that lowers
2
ATCO Pipeline’s risk. It was, and remains, my judgement that none of the Alberta utilities
3
were as risky as Pacific Northern Gas (PNG) with a 36% common equity ratio or Gaz
4
Metropolitain (GMI) with a 38.5% common equity ratio, where I continue to regard these two
5
as the riskiest regulated utilities in Canada.
6
In the two years since the Alberta generic hearing I have testified in business risk hearings for
7
the TransCanada Mainline, FortisBC and Terasen Gas and have not changed the above
8
judgment. Given the very low, if not non-existent, income risk, ROE regulated utilities in
9
Canada continue to have the very stable ROI necessary to support large amounts of tax
10
efficient debt financing. The only changes since then have been that the NEB has increased the
11
Mainline’s common equity ratio to 36%. There seems to be two reasons for this first the
12
Mainline refinanced its 10% preferred share component and replaced them with junior
13
subordinated debentures and second the entry of Alliance as a “competitor” has taken load
14
from the Mainline, such that it is running at significantly less than capacity with the fear that
15
the WCSB will not generate the new supplies to allow it to run full again.10 Neither of these
16
factors are relevant for Union Gas. The only other significant change is that the BCUC has
17
recently increased the allowed common equity ratio of Terasen Gas from 33% to 35% to bring
18
it in line with Union and EGDI. Notably Westcoast Transmission (Duke Energy Transmission)
19
has negotiated a 31% common equity ratio up from the 30% allowed by the NEB under RH-2-
20
94. Overall there is nothing in recent allowed common equity ratios that cause me to change
21
my judgment concerning the appropriateness of Union’s common equity ratio.
22
Q. WHY HAVE YOU NOT DISCUSSED UNION’S INCREASED RISK FACTORS?
23
A.
24
witnesses discuss “increases” in risk faced by various regulated utilities since I first testified in
25
1985. However, the ability of regulated utilities to earn their allowed ROE has not been
I don’t think that they are material. I have heard Dr. Sherwin and other company
10
The NEB has also increased the Mainline’s depreciation rate to compensate for supply problems from
the WCSB.
- 34 -
1
significantly impaired and I have yet to see any of these risks materialise to significantly harm
2
a Canadian utility. In fact the opposite has occurred the use of forward test years, fuel pass
3
throughs, the removal of the merchant function and increasing focus on the core monopoly
4
service have all served to reduce the risk of regulated utilities in Canada. The fact is that
5
regulation is a flexible process that moderates or shares these risks even if they do materialise
6
to the extent that the regulated utility is rarely hurt. A case in point is Pacific Northern Gas
7
(PNG), which I regard as the riskiest regulated utility in Canada by far.
8
There is no doubt that PNG is extremely risky. It operates a tiny 600 kilometre pipeline from
9
the Westcoast (Duke Energy Gas Transmission) system through to Western British Columbia
10
where the economy is heavily dependent on forest products and a few cyclical industries. Until
11
November 2005 almost 70% of PNG’s throughput came from a few industrial customers with
12
one, Methanex, overwhelmingly important. Unfortunately Methanex closed its doors in
13
November 2005 and PNG lost the load. Such a loss of load dwarfs anything that could
14
conceivably affect Union. In comparison the possibility of industrial by-pass exposes Union to
15
the potential loss (interrogatory response J2-24) of $29 million in revenues, which is not
16
material compared to Union’s $2,084 million in 2005 operating revenues.
17
How has the BCUC responded to PNG’s serious problems? In the first place the BCUC has
18
allowed PNG a 0.65% premium to the ROE as well as 3% more common equity than that
19
allowed its low risk benchmark (Terasen Gas). These more favourable financial parameters
20
have been allowed on an ex ante base to reflect PNG’s potential problems, since the risks
21
attached to PNG’s dependence on a limited number of industrial customers has been known for
22
a long time. That is, PNG’s shareholders were rewarded for its greater risk ex ante. However,
23
as the risk increased the BCUC then allowed PNG a series of deferral accounts. First a
24
comprehensive RSAM to remove weather induced variability in PNG’s earnings. Second an
25
industrial customer deliveries deferral account (ICDDA) to recover any deviations of actual
26
deliveries from those forecast for PNG’s large industrial customers. PNG has also taken $5.05
27
million of Methanex related assets out of its rate base and put these into a special deferral
- 35 -
1
account to be recovered from other customers over a ten year period. Finally the BCUC has
2
approved in principle the conversion of PNG into an income trust to help reduce costs.
3
I will discuss the future of PNG shortly, but at this point the important fact to note is the active
4
participation of the regulator, the BCUC, in helping PNG cope with a huge company
5
threatening event. For example, although Methanex accounted for 62% of PNG’s throughput
6
the BCUC allowed PNG to offer a special discount rate for Methanex and rebalance its rates.
7
As a result, before it closed Methanex only accounted for 7.6% of PNG’s operating revenues,
8
even though it was 62% of PNG’s throughput. As the Methanex related assets are recovered
9
from other customers it emphasises the fact that a regulated utility only faces two basic risks:
10
short run forecasting risk and the possibility of a “death spiral.”
11
Forecasting risks can be removed by deferral accounts if the regulator sees fit as the BCUC and
12
the NEB have. If a company is not allowed deferral accounts then it can manage these risks by
13
deferring expenditures to consistently come in under forecast and over-earn. This seems to be
14
the historic record in Canada, where over-earning seems to be positively correlated with the
15
absence of deferral accounts.11 The BCUC can and has used this regulatory protection for PNG,
16
but it can not prevent a death spiral. This occurs when customers leave the system and the
17
reallocated costs can not be recovered from the remaining customers, otherwise they too would
18
leave the system or the costs would be regarded as unfair and unreasonable. For PNG this death
19
spiral remains a possibility, where for the last five years PNG’s actual and allowed ROE have
20
been as follows:
21
22
23
24
Allowed
Actual
2005 2004 2003 2002 2001
9.68 9.80 10.17 9.88 10.00
8.20 6.90
7.50 5.60
7.40
25
The PNG ROE data indicates a persistent problem with earning its allowed ROE despite the
26
high amount of regulatory protection afforded it by the BCUC. The underlying reason for this
11
Performance based regulation can then put in sharing mechanisms to allocate any over-earning
between the utility shareholders and ratepayers.
- 36 -
1
is simply that PNG is a very small utility. For 2005 PNG had property plant and equipment of
2
$171.35 million and 39,295 customers. Union in contrast had property plant and equipment of
3
$3,133 million and 1,249,000 customers. Unlike PNG if Union lost a customer of the size of
4
Methanex the associated margin and cost of stranded assets could be recovered from other
5
customers and barely be noticed. The size and diversity of Union’s operating area dramatically
6
reduces its risk.
7
However, despite the most severe problems faced by any Canadian regulated utility how have
8
PNG’s shareholders fared? First, even after a “worst case” scenario arising at year end PNG’s
9
book value was about $21.50 and its stock price $20. So an equity investor in PNG would have
10
invested approximately $21.50 in PNG’s assets, earned a premium ROE and dividend
11
throughout the period of PNG operations, and still only seen the value of this investment drop
12
by $1.50 despite the loss of 62% of PNG’s throughput threatening the very survival of the
13
company. The point of this is simply that despite huge operational problems that dwarf
14
anything faced by Union Gas, the harm suffered by PNG’s shareholders has been minimal.
15
The example of PNG illustrates the basic proposition that regulation shields the utility from
16
many of the problems it ostensibly faces. The reason is that should these risks arise the utility
17
invariably goes to the regulator and gets the costs allocated to ratepayers. PNG, for example,
18
anticipates the costs of stranded Methanex related rate base assets being recovered from other
19
ratepayers not the shareholders. As the actual versus allowed ROE data indicates none of the
20
risks advanced in regulatory hearings involving those utilities have actually harmed their
21
shareholders. In contrast, they have harmed ratepayers.
22
The above is simply to put into perspective the risks raised by Union Gas. As Union indicated
23
in answer to interrogatory response J2-24 no assets or cost of service have become stranded
24
since 1990. Further I have no doubt that if Union became subject to any significant risk it
25
would approach the Board and seek to have the associated costs allocated to ratepayers.
26
- 37 -
1
Q.
2
UNION’S RISKS?
3
A.
4
discussion on “risk factors” since these must be disclosed in filings with the OSC. In its 2002
5
Annual Report Union grouped its risks into the following where I have added a brief
6
description of each:
7
8
9
10
11
12
13
14
EVEN IF YOU DO NOT REGARD THEM AS MATERIAL CAN YOU DISCUSS
I have looked at Union Gas’s financial reports for the last several years to read the
•
•
•
•
•
•
•
Market risk: decreased average usage of natural gas
Commodity price risk: fluctuation in natural gas prices
Credit risk: gas loans to counterparties
Weather risk: weather impact on consumption
Regulatory risk: impact of PBR and inflation floating price caps
Human Resource risks: impact of strikes and attracting good employees
Other : mainly franchise renewal risks
15
Most of these risk factors are “boiler plate” in that they change little from year to year. How
16
significant they are can be judged from the fact that for the subsequent three years, Union over-
17
earned by an average of 1.38%, which significantly exceeds its long run average over-earning.
18
In my judgment none of the aforementioned risks can be regarded as risks in the economic
19
sense of causing harm to Union’s shareholders.
20
In Union’s 2005 Annual Report the same risk factors as above are repeated and some
21
additional ones discussed such as interest rate and facility risk. However, the most significant
22
changes relate to more emphasis on market risk and the impact of high natural gas prices, the
23
possibility of by-pass due to the Board’s decision on Greenfield Energy Centre’s pipeline
24
application and a discussion of political risk attached to electricity generation in the province.
25
These risk factors mirror the main points of the testimony of Dr. Carpenter and the company
26
and I will discuss all three.
27
First in terms of regulatory risk Union claims that the Board’s January 2006 decision to allow a
28
bypass of Union’s system to a 1005 MW gas fired generating plant increases regulatory risk. In
29
particular Union noted in its 2005 Annual Report that the “The Decision (OEB’s GEC
- 38 -
1
Decision) increases the risk profile faced by natural gas distribution companies in Ontario.” To
2
judge the significance of this decision I looked at Enbridge Gas Distribution Inc’s management
3
discussion and analysis (MD&A) to its 2005 Annual Report of February 2, 2006. EGDI has the
4
following much fuller discussion of the impact of the Board decision (page 9):
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
It is clear from the above that the implications of the GEC decision may not be wide ranging.
20
Also as discussed earlier, the amount of Union revenues at risk to by-pass are modest compared
21
to the sort of risks faced by genuinely risky gas LDCs like PNG. Further what has to be
22
emphasized is that Union has simply lost the ability to generate some extra revenues to reduce
23
its overall rates, nothing has happened to significantly increase its rates and make it less
24
competitive or reduce its ability to rebalance them. I therefore judge it premature to consider
25
the GEC decision as reflecting an increase in Union’s risk.
“To date, the Company has operated with the understanding that it will be the only provider of
distribution service to all natural gas end users within its franchise area. Peer companies such
as Union Gas Limited (UGL) have operated with the same understanding. On January 6, 2006,
the OEB granted Greenfield Energy Corporation, a potential power-plant customer of UGL, the
right to physically bypass UGL’s distribution network within UGL’s franchise area, in order to
serve its own power-plant. The OEB's decision to allow a party other than the local distribution
utility to self serve is unprecedented. However, the OEB characterized this decision as
transitional and specific to the particular circumstances of this case. The OEB indicated that the
Natural Gas Electricity Interface Review (NGEIR) in 2006 will address utility offerings that could
be more robust against bypass. NGEIR is a rates proceeding that will assess the service
requirements of gas fired power generation in the province of Ontario and review natural gas
utility rate and service offerings for gas fired power generators. Until the completion of the
NGEIR proceeding, any possible future financial implications cannot be predicted.”
26
27
The remaining two risk factors are the impact of high natural gas prices and the fact that 80%
28
of Ontario’s generating capacity potentially needs replacing over the next twenty years. The
29
following graph indicates the US Henry Hub prices for natural gas. There is no question that
30
natural gas prices have spiked over the last five years and peaked at $13.42 in October 2005,
31
which would be close in time to the preparation of Union’s testimony. However, the spike in
32
natural gas prices was associated with hurricane damage in the US and the temporary
33
suspension of supply. Since October the price has fallen month by month to $7.54 at the start of
- 39 -
1
February. At this level the price of natural gas is not too dissimilar to the winter of 2000/2001
2
when natural gas prices peaked at $8.95.
Henry Hub Natural Gas Prices
16.00
14.00
12.00
10.00
8.00
6.00
4.00
2.00
Nov-05
Nov-04
Nov-03
Nov-02
Nov-01
Nov-00
Nov-99
Nov-98
Nov-97
Nov-96
Nov-95
Nov-94
Nov-93
0.00
US$ Million BTU
3
4
What is important is whether these high natural gas prices impact demand, in this respect two
5
things have to be remembered: first demand is affected by relative energy prices and second
6
these prices are affected by the long run supply response. Further these are generic effects that
7
presumably also affect EGDI.
8
In terms of competing energy supplies I could not find any mention of this in EGDI’s 2005
9
MD&A to the company’s 2005 annual report. This is of itself very revealing. However, in
10
11
12
13
14
15
16
17
18
19
20
21
22
EGDI’s annual information form for 2005, EGDI states (page 9):
Price Advantage of Natural Gas
Natural gas is the predominant fuel of choice in the residential heating market
throughout the Company's franchise area. The primary competition for natural gas
remains domestic fuel oil and electricity. Natural gas has continued to provide both
environmental and price advantages, and this is expected to continue. During 2005,
natural gas in the residential market experienced, on average, a price advantage on an
equivalent annual volume basis of 40% (2004 – 38%) against electricity and 32% (2004
– 23%) against domestic fuel oil. Although natural gas commodity prices remained
historically high over the year, the concurrent run-up in oil prices and high electricity
prices have kept natural gas prices competitive with alternative energy sources. Natural
Gas prices have risen mainly due to steady increases in demand and tight short-term
- 40 -
1
2
3
4
supply. Over the longer term, supply is expected to expand with a greater North
American Liquefied Natural Gas infrastructure and Arctic gas contributions, which
bodes well for future price competitiveness.
5
This continuing competitiveness of natural gas against competing fuels is mirrored in Union
6
Gas interrogatory response J2-21, which compared natural gas against major competing fuels.
7
DBRS also noted (June 22, 2005) that
8
9
10
11
12
13
“Natural gas is the most economical fuel source for home heating and is more
environmentally friendly than oil. In addition, the new regulated price regime in Ontario
that came into effect in April 1, 2005 raising electricity prices in the province for
residential, low volume and designated consumers, ensures that natural gas remains a
very competitive option.”
14
What has to be remembered is that natural gas prices are high mainly due to increased demand
15
and not reduced supply (apart from temporary factors). In its February 2005 Natural Gas
16
Outlook to 2020, the American Gas Foundation forecast:
17
18
19
20
21
22
23
•
•
•
Natural gas prices to remain in the $5-6 per MMBTU range for most of the
period of the study,
Supply to become much more diverse including major contributions from
Alaska and LNG,
“in spite of higher prices” demand to increase with 2/3rds coming from
electricity generation and only 1% for residential.
24
What is clear is that high natural gas prices are primarily being driven by increased demand,
25
particularly from electricity generation. This is not a sign of increased risk for Union, but
26
exactly the opposite: Union is distributing an increasingly valuable commodity.
27
What is also important in this regard is that Union operates a major transmission and storage
28
system in a rapidly developing hub. 86% of Union’s transmission costs are recovered through
29
fixed demand charges similar to other major gas transmission pipelines. However, unlike the
30
TransCanada Mainline, which is suffering from lower loads as the WCSB has failed to match
31
the more optimistic assumptions needed to fill both the Mainline and Alliance, Union’s system
- 41 -
1
is close to the consuming areas and needed to balance seasonal demand with steady shipments
2
from the WCSB and elsewhere. These services will become more valuable in the future as
3
more electricity is generated from natural gas. As DBRS notes “The company will benefit from
4
any additional gas sales to gas fired generation facilities that locate in its service area as part of
5
the provinces’ RFP.”
6
Both EGDI and the AGF forecast also point out the need for LNG imports. This is also
7
reflected in the future of PNG, where PNG is currently considering a five fold expansion of its
8
pipeline system through looping. Why a risky little pipe that has lost 62% of its load would
9
consider a fivefold expansion may seem strange. However, there are concurrent plans to build
10
an LNG plant in Kitimat BC in which case the flow on the PNG pipe would be reversed so that
11
PNG would inject natural gas from the Kitimat LNG plant into the Duke system rather than
12
take it. The important point is that at current natural gas prices there are large numbers of LNG
13
proposals that will generate extra supply and moderate price pressures to maintain natural gas’s
14
competitiveness as a fuel source.
15
In my judgment natural gas will continue to be a competitive fuel source and the only risk that
16
Union faces is the standard forecasting risk: that demand will not meet company expectations.
17
In this respect in answer to interrogatory response J2-21 Union was asked to detail what
18
measures the company has taken to track customers leaving its system due to high energy
19
prices. One would expect that Union would introduce some special measures such as customer
20
surveys etc. However, in answer to interrogatory response J2-21 there was no indication of any
21
special measures undertaken by the company. I would therefore judge that high energy prices
22
will not cause any fundamental problems to Union and simply manifest itself as a forecasting
23
problem. If this assessment is wrong and Union starts to face significant losses in industrial
24
load then either Union can ask for a deferral account similar to that used by PNG or simply go
25
back to the Board for after the fact rate relief. As EGDI noted in its 2005 MD&A page 9
26
27
28
29
30
“Rate relief may be sought for significant amounts that are not forecasted, allowing the
Company to recover the costs of providing and maintaining the quality of its service, while
achieving the allowed rate of return on rate base. To the extent the OEB denies recovery of any
such costs, the Company is at risk.”
- 42 -
1
Q.
WHAT ABOUT THE IMPACT OF POORER ECONOMIC CONDITIONS?
2
A.
I will discuss economic conditions later, but the fact is that we have a very strong
3
Canadian economy at the moment. It is so strong that the Bank of Canada has been pushing up
4
short term interest rates over the last six months to try and slow it down. In part the strength of
5
the economy is due to high natural resource prices that have pushed up the value of the
6
Canadian dollar to levels not reached since the early 1990s, when it reached 90 cents US. This
7
will undoubtedly squeeze some firms in Union’s franchise areas. However, the fact is this is
8
simply a normal cyclical phenomenon that Union has faced before, as is the problem of full
9
employment and hiring difficulties. Notably the earned vs actual ROE data that goes back to
10
1985 covers the last two major booms of the late 1980s and late 1990s when broad economic
11
conditions were similar. What is of particular importance is that Union managed the impact of
12
the boom of the late 1980s and bust of the early 1990s with a 29% common equity ratio, the
13
current strong economy hardly justifies a 40% common equity ratio.
14
15
16
Q
DOES THE MOVE TOWARDS PERFORMANCE BASED REGULATION
INCREASE UNION’S RISK?
17
A.
No.
18
with rebasing through periodic rate reviews, but in my judgment this has no impact on Union’s
19
risk. Union operated under PBR from 2001-2003 and in these years it over-earned by 1.50%,
20
2.41% and 2.13%, which is above Union’s long run tendency to over-earn. Further Terasen Gas
21
has been on a PBR mechanism for sometime and continues to earn in excess of its allowed
22
ROE.
I am aware of the Board’s intention to move to multi year incentive regulation
23
24
Q.
IS THERE ANY OTHER DATA TO SUPPORT THIS ASSESSMENT?
25
A.
26
included in Union’s financial statements along with Union’s capital expenditures (Capex).
Yes. The following data indicates the amount of depreciation and amortisation (D&A)
- 43 -
1
2
3
4
D&A
Capex
2005
158
231
2004
155
147
2003 2002 2001
156 153 150
135
193
218
5
If a company’s Capex is approximately equal to its D&A then it is a stable company and
6
simply replacing assets as they wear out. However, the fact that Union’s Capex has generally
7
exceeded its D&A indicates system expansion. This is confirmed by the $250 million planned
8
expansion of the Dawn Trafalgar transmission system.
9
In my judgement the observed expansion of Union’s system, combined with the growth in its
10
customer base and generally strong economic conditions all indicate that there has been no
11
increase in the company’s business risk. Union remains one of the two premier gas LDCs in
12
Canada with a very strong franchise and unique ownership of Canada’s major storage system.
13
If anything Union is less risky today than it was prior to its merger with Centra Gas Ontario
14
and the strengthening Canadian economy. The fact that it operated at that time with 29%
15
common equity is significant given that it is now requesting 40%. I therefore recommend that
16
Union continue with a 35% deemed common equity ratio.
- 44 -
1
4.0
FINANCIAL RISK
2
3
4
Q.
ARE CAPITAL MARKET CONDITIONS AT PRESENT MORE DIFFICULT
THAN THEY WERE AT THE TIME OF EBRO 499 AND RP 2003 0063/87/97?
5
A.
Basic macroeconomic data for the last twenty plus years is provided as background in
6
Schedule 6. Economic conditions can sometimes change quite rapidly as the impact of
7
hurricanes and oil price shocks are unpredictable. However, there is a rhythm to the economy,
8
which reflects the momentum as shocks gradually work through the system; this is what is
9
generally referred to as the business cycle. The basic economic variable here is the rate of
10
economic growth. The trend line for economic growth is around 3.0% so that periods with
11
growth significantly below that are periods of contraction or recession, whereas periods of
12
growth significantly above that are expansionary periods.
13
Note that in 1983 Canada was pulling out of a slowdown and then again from 1989 until 1993
14
Canada was mired in a deep recession in response to a normal cyclical slowdown as well as
15
restructuring that accompanied the passage of the Free Trade Agreement (FTA). We can also
16
see the strong economy of the mid 1980s and again the mid to late 1990s, when real economic
17
growth was over 4.0%. Most recently, we can see the mild slowdown of the early 2000’s as
18
recession in the United States and the effects of the stock market crash in Canada weakened the
19
economy. 2001 was then a period of moderate growth, caused in part by the startling events of
20
September 11th. The Canadian economy then picked up steam in 2002 and outperformed other
21
major industrialised economies until in 2003 it was hit by a perfect storm of a strengthening
22
exchange rate, slowing growth in the United States, severe acute respiratory syndrome (SARS)
23
and a single incident of BSE or mad cow disease. These effects were largely temporary as the
24
Bank of Canada lowered interest rates in July 2003 so that some of the steam went out of the
25
strengthening Canadian dollar and economic growth picked up in 2004.
26
For 2005 we again had good economic growth as strong growth soaked up the remaining
27
available labour and the unemployment rate dropped below 7.0%. Consumer spending was
28
strong as low interest rates supported the purchase of consumer durables, as well as record
- 45 -
1
residential housing sales. Further Business investment remained strong with additional
2
rebuilding of inventory. Even the effects of the oil price increases were largely muted by
3
external interest in Canada’s oil sands and the perception that Canada has positive exposure to
4
oil and gas prices. This perception, allied to the continuing strength of the current account
5
surplus, which has been running at over 1.0% of GDP, lead to a strengthening Canadian dollar
6
which has crept above 86 cents US. The overall strength of the Canadian economy caused the
7
Bank of Canada to reverse its stimulus policy and start increasing short term interest rates as
8
the overnight rate has been increased in stages from 2.50% to the current target range of 3.75%.
9
Q.
WHAT IS YOUR OUTLOOK FOR INFLATION?
10
A.
Over the past several years, the Canadian economy has experienced low and stable
11
inflation together with reasonably strong economic growth. The graph in Schedule 7 shows the
12
average CPI inflation rate since 1951. What is clear from this graph is the enormous run up in
13
inflation from the early 1950's through to its peak in the early 1980s. Since then it dropped to
14
plateau at the 4.0% level through the 1980s before the effects of the major slow down in the
15
early 1990s caused it to drop to its cyclical low in 1994/5, where it almost touched price
16
stability. Since that time the consumer price index has remained broadly in the middle of the
17
Governor of the Bank of Canada’s 1-3% range.
18
Schedule 8 graphs the average annual inflation rate along with the average yield on long
19
Canada bonds and Treasury Bills since 1961. The graph shows that prior to 1981, inflation was
20
increasing steadily, until the Bank of Canada engineered a recession in 1982-3 to bring
21
inflation under control. Similarly, in the late 1980's there was a gradual increase in inflation
22
and wage settlements that peaked about 1991, as again, the Bank of Canada engineered a
23
recession to bring down the rate of inflation. Although the absolute rate of inflation has been
24
brought down considerably from these earlier periods, the same pattern of increasing inflation
25
from 1994-2001 is evident as in the earlier periods of 1986-1990 and 1976-1982. In each case,
26
interest rate increases slowed down the economy and with it the rate of inflation.
- 46 -
1
Schedule 10 shows that the long Canada real bond yielded 1.60% on April 4, 2006 or 2.72
2
basis points below the equivalent nominal bond yield of 4.32%. The real bond guarantees the
3
investor protection from inflation, whereas the nominal bond has built into the yield
4
compensation for both the expected rate of inflation and a real yield. As a result, the spread
5
between the nominal and real rate can be taken as one estimate of the market’s forecast of the
6
long run inflation rate. Other measures of inflation come in slightly lower as the GDP deflator
7
has been running at 2.0%, similar to the current CPI inflation rate, while “core” inflation,
8
which nets out the more volatile items, is just below 2.0%.
9
The graph in Schedule 9 shows the aggregate net lending of governments in Canada, where a
10
negative number indicates government borrowing or a fiscal deficit. What is clear from
11
Schedule 9 is the dramatic improvement in the fiscal position of all layers of government since
12
the early 1990s and their return to balanced budgets. This in turn has reduced the supply of
13
government bonds and the need for the Bank of Canada to follow accommodative monetary
14
policy which in turn has supported the drop in inflation. Consequently it seems clear that the
15
core rate of inflation will continue within the Governor of the Bank of Canada’s 1-3.0%
16
operating range.
17
Q.
WHAT ARE YOUR INTEREST RATE FORECASTS?
18
A.
Schedule 10 provides data on the full range of interest rates across the broad maturity
19
spectrum as of April 5, 2006. What is evident is that interest rates for long maturity instruments
20
are slightly higher than at the short end of the maturity spectrum; this is referred to as a
21
‘normal’ yield curve. Schedule 8 charts the history of short and long term interest rates together
22
with inflation since 1961. It is clear that short term Treasury bill yields have continued their
23
long decline from their peaks in 1981 as inflation has receded. This long run decline has been
24
punctuated by periods when Treasury bill yields have increased to support the dollar (1996) or
25
fight a too vigorous economy (late 1980’s and 1990’s). In contrast, long term rates have
26
continued their gradual year over year decline without these peaks. This is because long term
27
bond investors look not just at the next 91 days, but far off into the future. As such, long-term
- 47 -
1
bond yields reflect the long term future of the Canadian economy, while T-Bill yields reflect
2
short term expectations.
3
Another way of looking at the impact of the Bank of Canada’s monetary policy is to recognise
4
that monetary policy works through both interest rates and the exchange rate: higher interest
5
rates and a stronger dollar together slow down the economy by impacting interest sensitive and
6
export industries. To examine both of these effects, the Bank of Canada maintains a “monetary
7
conditions index” or MCI, which is reproduced in the graph in Schedule 11. Again, the
8
dramatic changes of 1981-82 and 1988-1990 are evident, as the MCI increased dramatically.
9
We can also see the long run monetary loosening ending in 1998 with the levelling off of the
10
MCI as the Bank of Canada started to worry about a too strong economy. This stance was
11
reversed by the end of 2001 as the stock market crash exposed the economy to another shock,
12
with further loosening helped by a weak dollar. It has been the subsequent strength in the value
13
of the Canadian dollar that has largely produced the upturn in the MCI. The recent upturn in
14
short term interest rates has served to further tighten recent monetary policy.
15
What is evident from the increase in short term interest rates over the last year is that the capital
16
market believes in the integrity of the Bank of Canada. There are no longer any fears that the
17
Bank will allow inflation to increase significantly. This is evident in the fact that long term
18
rates have not increased commensurate with short term rates. I therefore expect that the Bank
19
will continue to tighten the system and increase its target overnight rate, but long rates are
20
unlikely to move from around 4.50% in the immediate future.
21
Q.
WHAT HAS BEEN THE RECENT STATE OF THE CAPITAL MARKETS?
22
A.
Since the onset of the last major recession in the early 1990s, capital markets have been
23
dominated by federal and provincial government financing. Their importance, however, has
24
been receding. Overall government “lending,” representing the aggregate of all levels of
25
government, was running at the rate of over minus $60 billion during 1992 and 1993 or at its
26
peak over 9.0% of GDP. Government net lending subsequently declined almost year by year as
- 48 -
1
the economy recovered and governments finally got their spending under control. Schedule 9
2
graphs the government's net lending as a percentage of GDP.
3
The disastrous consequences of government fiscal policy starting in the early 1970s is obvious
4
in Schedule 9, as governments started to run persistent deficits (net lending was negative
5
indicating net borrowing). However, it is equally clear that since 1992 all layers of government
6
have made serious efforts to restore some sanity to their finances. By 1997 lending had become
7
genuine lending and governments in aggregate were in surplus for the first time in twenty-three
8
years. In 2000 all layers of government in aggregate ran a surplus of $32 billion as tax revenues
9
soared and expenditures on welfare, unemployment, etc., declined along with the
10
unemployment rate. This amounted to over 3.0% of GDP, the biggest surplus since 1951, when
11
governments were still actively paying down the war debt. Although the weakening economy
12
has eroded the aggregate surplus since then, it is remarkable that the weakening economy of the
13
early 2000’s did not impose more pressure on government finances.
14
The overall decline in government “lending” has opened up room for private sector borrowing
15
as corporations have returned to the equity and bond markets, following the strengthening of
16
their balance sheets. Fuelled by healthy consumer spending, corporate profits have rebounded
17
from the extreme cyclical lows of 1992-1994. Schedule 12 graphs the level of pre-tax profits to
18
GDP. In 2000 pre-tax corporate profits reached 12.0% of GDP as the economy peaked. This
19
level is higher than the last cyclical highs of 1988-1989 and only slightly below the resource
20
boom fuelled highs of the 1970s. Although pre-tax profits dropped off to 11.0% of GDP for
21
2001 and 2002 as the economy weakened, they have subsequently spurted forward again on
22
high resource prices and reached an all time high of 14% of GDP in the second quarter of 2005.
23
This profit data is mirrored in the capacity utilisation data in Schedule 13, where we can see the
24
drop in utilisation in 2001 through the middle of 2004 and the strong rebound since then with
25
utilisation rates again at close to all time highs.
26
The profit and capacity utilisation data provide the same signals as the inflation and interest
27
rate data, that the peak of the cycle was in 2000 with a minor slowdown in 2001-2003. We are
28
now in a strengthening phase of the business cycle as the economy is strong. This combination
- 49 -
1
of declining interest rates and booming corporate profits has lead to stronger equity prices and
2
a strengthening value of the Canadian dollar. Schedule 14 graphs the C$ in terms of its US
3
dollar value, where we can see clearly that its long run secular decline, when it was heading for
4
a sub 60 cent US level, was reversed in the Fall of 2002 after which is has gone from strength
5
to strength and has recently been over 86 cents US.
6
This strength has been mirrored in the performance of the TSX/S&P Composite which has
7
rebounded from its lows in 2002 with each year since showing strong equity market
8
performance. Recently the TSX Composite has been over 12,000 indicating much more
9
confidence in the stock market and the Canadian economy.
10
11
Q.
HOW DOES THE STATE OF THE ECONOMY AFFECT PROFITS?
12
A.
Schedule 12 graphs the level of pre-tax corporate profits as a percentage of GDP. These
13
profits are taken directly from corporate tax returns and so avoid all the one time only
14
accounting losses that rocked Nortel, JDS Uniphase and others. Consequently, they are a more
15
accurate measure of corporate operating profits. The graph shows that profits are currently
16
running at all time highs of over 14% of GDP.
17
Another way of assessing corporate profitability is to look at the aggregate data maintained by
18
Statistics Canada (Quarterly Financial Statistics for Enterprises). Statistics Canada started
19
reporting quarterly return on equity data in 1980 based on Standard Industrial Classifications
20
(SIC) and then moved to North American Industrial Classifications (NAICs) in 1999. Schedule
21
15 graphs this average annual ROE against the spread between the yield on BBB debt and long
22
Canada bonds from Scotia Capital's Handbook of Canadian Debt market Indices. It shows that
23
as of 1980 the average ROE was 15.05% and the yield spread that rewards investors for
24
holding BBB rated debt instead of default free Canada bonds was very low at just over 50 basis
25
points. “Corporate Canada’s ROE” then declined during the 1982 recession as the yield spread
26
widened. The ROE then hovered around the 10% level during the growth oriented 1980's with
27
a stable yield spread. As ROEs fell from 1989 onwards, investors grew concerned about credit
28
risk and the yield spread increased dramatically to almost 350 basis points in 1993. The profit
- 50 -
1
recovery during the mid 1990s then caused the yield spread to contract only to widen in the
2
early 2000s as ROEs weakened.
3
The graph indicates the way in which the business cycle affects firms. During expansions,
4
profitability increases and credit risk is lessened, causing investors to buy corporate bonds on
5
narrower spreads over similar Canada bonds. During recessions the reverse happens: as
6
profitability is reduced credit risk tends to increase causing spreads to widen. Profitability in
7
this sense affects the market access of cyclical firms.
8
Schedule 16 shows recent information on corporate spreads using the A and BBB spread data
9
from the Scotia Capital long bond indexes. The cyclical behaviour of spreads is again clearly
10
visible. The BBB and to a lesser extent A spread over equivalent Canada bonds again clearly
11
widened during the recession/slowdowns in both the early 1990s and early 2000s. However
12
both spreads have tightened over the last few years reflecting the stronger economy and lower
13
credit concerns.
14
The combination of booming corporate profits and lower credit spreads has lead to strong
15
financing activity. In Schedule 17 is the aggregate level of financing in Canada over the last
16
twelve years from data provided by the Investment Dealer’s Association (IDA). This data
17
reflects all the factors that I have discussed so far. Government borrowing was routinely 60-
18
70% of total financing as government debt crowded out private financing. However, over the
19
last several years there has been significant refinancing of existing, as well as new corporate
20
debt issues, as companies have taken advantage of lower interest rates. Corporate debt issues
21
have increased from barely 25% of the level of government debt to 50% and the process private
22
financing activity has increased from 4-5% of GDP up to the current level of 8%.
23
Schedule 18 graphs the extent of total and private sector financing as a percentage of GDP to
24
indicate how receptive the capital markets are. This data confirms the stock market, profit and
25
spread data that capital markets are currently very receptive to new financing and a prior there
26
is no indication of any financial access problems. In fact, currently Canadian capital markets
27
are very receptive to new financing activity.
- 51 -
1
Q.
WHERE ARE WE IN THE BUSINESS CYCLE?
2
A.
Up to the middle of 2000, the U.S. was deep into an extended boom and showing
3
distinct signs of an overheating economy, whereas the Canadian economy was just getting its
4
“second wind.” The Governor of the Federal Reserve then started to slow down the U.S.
5
economy to avoid incipient inflation and the Governor of the Bank of Canada followed suit,
6
although more slowly, so that monetary policy started to head off a recession. Unfortunately
7
the bursting of the tech bubble severely destroyed investor confidence as it revealed both the
8
extent of corruption at the highest level of some US corporations and the contempt with which
9
some first line US investment banks held their retail and institutional customers. The effect of
10
this loss in investor confidence has now receded as both the US and Canadian economies are
11
showing good economic growth. For 2006 both economies are expected to be strong, inflation
12
to be contained to the 2.0% middle of the Governor of Canada’s band, despite strong energy
13
prices, and the capital markets to reflect this. Barring the impact of some extreme terrorist
14
action, it is an optimistic medium term economic and financial outlook reflecting continued
15
strong economic growth and performance towards the top of the economic cycle.
16
Q.
OF EBRO 499?
17
18
19
20
21
22
23
24
25
26
27
HOW DO ECONOMIC CONDITONS COMPARE TO THOSE AT THE TIME
A.
The major capital market indicators in 1998 versus now are as follows:
GDP Growth
CPI Inflation
T Bill Yield
Long Canadas
Corporate Profits (pretax % GDP)
TSX Composite
A Spreads (December)
1998
4.10%
1.00%
4.74%
5.45%
9.41%
6485
0.92%
2006
3.0%
2.2%
3.90%
4.32%
14.0%
12,000
1.05%
28
1998 was a period of a strengthening economy very similar to now as we were approaching the
29
peak of the economic cycle. Economic growth was higher than now and inflation marginally
- 52 -
1
lower as energy prices were depressed and profits were approaching their cyclical high.
2
Monetary authorities were beginning to worry about the effects of a bull market and the TSX
3
Composite ended the year at 6485. Short-term Treasury Bill yields averaged 4.74% and the
4
yield curve was normal with long Canada yields at 5.45%. Corporate A spreads were just under
5
1.0% over similar long Canada maturities.
6
In contrast we are currently a bit later in the business cycle. Economic growth is lower at about
7
3.0% and although inflation and corporate profits are higher they largely reflect the dramatic
8
increase in oil and gas prices, rather than the stage in the business cycle. Noticeably long
9
Canada bond yields are 1.13% lower than in 1998 and shorter term yields 0.84% as the Bank of
10
Canada has increased rates to approach a flat yield curve to slow down the economy. Overall I
11
would judge us to be perhaps a year ahead of 1998 in terms of the business cycle. Apart from
12
the impact of high oil and gas prices on the foreign exchange rate, inflation rate and corporate
13
profits I would judge the economy to be very favourable compared to 1998.
14
I do not recommend that the Board change common equity ratios in response to normal cyclical
15
macro-economic conditions. Moreover, it is my judgment that there are no macro-economic
16
factors that indicate any increase in risk facing Union Gas today as compared to the time of
17
EBRO499. In many ways we are in a very stable strong economy, which itself may be
18
disturbing.
19
20
21
Q.
22
A. The bond rating agencies are concerned with accurately predicting the credit quality of a
23
firm’s debt. In this task they face an asymmetry; if they get it right they get little credit, but if
24
they make errors and firms default, that they previously rated investment grade, then they get
25
severely criticised. They therefore take a conservative approach and sometimes over react. In
26
this respect the most dramatic re-evaluation has occurred on the part of Standard and Poors
27
(S&P) as it has “harmonised” ratings between the US and Canada as a consequence of its
28
takeover of the Canadian Bond Rating Service (CBRS). In doing this harmonisation S&P has
IF CAPITAL MARKET CONDITIONS ARE GOOD WHY ARE SOME BOND
RATING AGENCIES EXPRESSING CONCERN?
- 53 -
1
taken a quantitative approach and seemingly simply taken standard ratios from the US and
2
applied them in Canada with little qualitative adjustment for the different institutional
3
environment.
4
On March 5, 2003 S&P indicated that it was putting several Canadian utilities on credit watch
5
and was re-evaluating their ratings, primarily as it re-evaluated the nature of Canadian
6
regulatory protection. Subsequently several of them were downgraded.
7
8
Q.
HAVE THESE DOWNGRADES AND REVIEWS HAD AN IMPACT ON
BORROWING COSTS?
10
A
Not as far as I can determine. Schedule 18 tracks spreads of major utility issues from
11
data collected by RBC-DS against the equivalent maturity long Canada bond.12 The absolute
12
level of the spread is not significant, since the spread is affected by the bond’s maturity, what is
13
important is the trend since the end of 2002 and from March 2003 around the time of the S&P
14
announcement. The average spread is in the final row. The average spread was 91 basis points
15
at December 1999, increased to 138 and 129 basis points by December 2000 and December
16
2001; it then increased again to 152 basis points at the end of 2002. Since December 2002 it
17
decreased throughout 2003 to finish at 95 basis points in December and since then it has been
18
marginally above 90 basis points. Overall utility spreads were slightly higher in December
19
2001 than they are today, since the trend in profits and economic conditions is now generally
20
up rather than down. As a result, the current business environment results in marginally lower
21
credit risk.
22
If we look at the individual issues, the three A rated issuer spreads at the end of December
23
2001 were 90-130 basis points, at the time of the S&P review (about March 2003) they were
24
80-123 basis points whereas at the end of this period they were 50-108 basis points. For the A-
25
issuers the spreads were 80-150, 70-201 and then 49-119. For the BBB+ issuers the spreads
9
12
Unfortunately RBC-DS no longer publishes this data. The company was asked for spread data in J242 but none was provided.
- 54 -
1
were 80-191; 105-205 and 60-135, while for the BBB issuers they were 88-235; 150-250, and
2
70-141. Overall it is clear that spreads have tightened since the announcement of the S&P
3
review. Overall, it seems that the S&P announcement and review (and downgrades) has had a
4
negligible impact on these issuers.
5
From this spread data there is no indication that S&P’s decision to impose harsher credit
6
standards on Canadian utility holding companies has had a significant impact on either their
7
borrowing costs or presumably the marketability of future debt issues. Spreads for issues rated
8
A- and below, the ones that presumably would have been affected the most by S&P’s
9
announcement, have almost all declined since the end of 2002 and the time of S&P’s
10
announcement.13 The fact that the market has shrugged off S&P’s new harsher credit standards
11
should not be surprising given that they received a negative reaction when they were
12
announced and have lead Terasen Gas Inc to cancel its engagement with S&P.
13
This data from RBC-DS is consistent with more recent data provided by BMO-Nesbitt Burns in
14
a research report (August 3, 2005) on the Kinder Morgan proposed takeover of Terasen Inc.
15
That report produced the graph reproduced in Schedule 19. The graph indicates that Union
16
Gas’s ten year debt trades on similar spreads to that of Enbridge Gas Distribution (EGDI) and
17
Terasen
18
widened two years ago, but there is no indication that Union can not access financial markets.
19
The fact is that S&P is known to be a harsher judge of credit quality than either DBRS or
20
Moodys. For example, in a February 1998 report, Credit Ratings in Canada, DBRS indicated
21
that its ratings for the same companies were 0.23 categories, that is, a “high” or “low” qualifier,
22
higher than Moodys and 0.55 higher than S&P. This would indicate that approximately in four
23
ratings three would be the same between Moodys and DBRS with one Moodys a category
24
lower, whereas for S&P two would be the same and two a category lower. However, in
25
comparing S&P and DBRS ratings in the companies in Schedule 31 the S&P ratings are a total
Gas. The spread on EGDI is generally slightly lower particularly when spreads
13
Note that since July 2001 S&P has downgraded CU Inc, Newfoundland Power, Westcoast, Fortis BC
Gas (Teresen), Epcor Utilities Enbridge Consumers Gas, and Union Gas.
- 55 -
1
of nineteen “notches” below those of DBRS, so that S&P rates these utilities on average at least
2
a notch below DBRS, where a notch is a modifier such as a high or low or plus or minus.
3
For two companies Union Gas and Terasan Gas Utility (BC Gas) the differences are marked
4
and S&P rates them three notches lower than DBRS. In both cases DBRS rates them as A,
5
whereas S&P rates them BBB. In both cases the lower rating is due to the lower rating of the
6
utility holding company that owns the utility. In March 2004 Terasen discontinued its
7
engagement with S&P, stating that there is no benefit to an S&P rating, since it does not agree
8
with S&P’s credit assessment and does not plan to issue US dollar debt. The fact that S&P
9
could come up with ratings so profoundly different from DBRS indicates that something has
10
“spooked” S&P, that has not “spooked” the Canadian rating company or the Canadian market.
11
12
Q.
WHY DO YOU THINK S&P HAS TAKEN SUCH A HARD LINE WITH
CANADIAN UTILITY HOLDING COMPANIES?
A.
DBRS provided six reasons for their higher average ratings than the US agencies:
13
14
15
16
17
18
19
20
21
•
•
•
•
•
•
less penalty for size
better knowledge of Canadian companies
treatment of sovereign rating principle
technical factors in the treatment of holding companies
specific industry biases
unsolicited ratings not generally done by DBRS
22
Only some of these six factors are relevant for utilities. For example, it is known that the US
23
agencies now have a policy of rating all public market debt and on average assign marginally
24
lower ratings for unsolicited ratings, compared to when one that is solicited and paid for.
25
Similarly, the industry bias has traditionally focused on oil and gas companies not utilities,
26
while the sovereign treatment refers to rating government debt. What is left is: size, Canadian
27
knowledge and the holding company problems.
- 56 -
1
S&P has specific rules on how to rate the parent of a holding company that issues debt versus a
2
subsidiary and to illustrate I will refer to some examples from S&P’s Corporate Ratings
3
Criteria handbook with respect to telecom companies that have motivated this policy.
4
5
6
•
Frontier Telephone was rated AA- and was purchased by Global Crossing and
the rating subsequently lowered to BB+. The New York Public Service
Commission did not prevent the acquisition.
7
8
9
10
•
Cincinnati Bell was rated AA- when its parent acquired IXC Communications,
which had a B- rating, and subsequently Cincinnati Bell’s rating was dropped to
BBB-. The Public Utilities Commission of Ohio did not create any roadblocks
or impose any penalties on Cincinnati Bell.
11
12
13
•
Qwest acquired US West Communications, which was rated A+ and S&P
warned its rating would be cut to BBB- but regulatory concern was on service
quality not protection of bond holders.
14
Each of the three prior examples were taken from the “go go” era of the late 1990s when
15
telecom was in a state of flux and the internet revolution was going to transform telecom.
16
Hence, these were deals involving local telecom providers with stellar bond ratings and riskier
17
internet and broadband providers. In each case, the local public utilities commission evaluated
18
the acquisition and reflected on service questions but did not step in to protect the bond holders
19
from significant credit downgrades. The downgrades were the direct result of a change in the
20
riskyness of the holding companies which “infected” the subsidiary, since the parent had
21
control of the sub and could over-leverage it or dividend cash out to support parent company
22
debt. In response, S&P now has a policy that the credit rating of a regulated telecom can not be
23
higher than the credit rating of its parent.
24
For non-telecom utilities S&P
25
26
27
28
“rarely view(s) the default risk of an unregulated subsidiary as being substantially
different from the credit quality of the consolidated entity. Regulated subsidiaries can
be treated as exceptions to this rule – if the specific regulators involved are expected to
create barriers that insulate a subsidiary from its parent.”
- 57 -
1
In other words there is a cross subsidy from the regulated to the unregulated entity unless the
2
regulated entity is “ring fenced” so that any problems on the non-regulated side do not impact
3
the regulated side. S&P refers to this as “structural insulation techniques” which may involve:
4
5
6
7
8
9
•
•
•
•
separate incorporation of the sub
independent directors
minority ownership stakes
regulatory oversight to insulate the subsidiary
S&P is very forthright in that the onus lies on the regulators. It states
10
11
12
13
“the bar has been raised with respect to factoring in expectations that regulators would
interfere with transactions that would impair credit quality. To achieve a rating
differential for the subsidiary requires a higher standard of evidence that such
intervention would be forthcoming.”
14
My reading of these remarks is that having been “burned” with these US telecoms and the lack
15
of reaction from US public service commission S&P is now taking a tougher line on all
16
utilities; an approach that they are bringing into Canada.
17
Finally FERC has been less forthcoming than some expected in reining in the utilities. After
18
Enron siphoned off $1.5billion from its two natural gas pipelines, the FERC instituted a review
19
of inter-affiliate transfers. Many expected FERC to impose minimum equity ratios of 30% and
20
requirements such as maintaining an investment grade bond rating before the parent could
21
manage the subsidiary’s cash. However, when the FERC announcement was made in
22
November 2003 it fell far short of S&P’s expectations. As S&P noted
23
24
25
26
27
28
29
“the degree of oversight by the FERC has traditionally been less than sufficient to
justify insulation. That the FERC took almost two years to respond to the Enron
pipeline situation indicates that timely intervention that would protect bondholder
interests is not likely when a regulated utility’s parent is experiencing financial
problems. It seems clear to Standard and Poors that the new rule falls far short of
providing the requisite insulation to justify any ratings separation for utilities regulated
primarily by FERC”
- 58 -
1
It is clear from this comment from S&P that it is their disenchantment with events in the US
2
that has triggered their review of regulatory protection in Canada. Further they are not the only
3
ones.
4
In a recent article in Public Utilities Fortnightly (August 2004) two members of the New Jersey
5
Board of Public utilities state
6
7
8
9
10
11
12
“ring fencing holds out the prospect for insulating regulated utilities from the traditional
failed diversification investments of the parent holding company….. Successful ring
fencing is even more critical considering that state regulators are facing challenges
created by failures of corporate governance, accounting scandals, and in some cases
alleged criminal conduct in energy markets. Ring fencing may be the only regulatory
device capable of levelling the playing field and forcing the holding companies to
absorb the consequences of failed non-utility investments.”
13
With FERC failing to implement ring fencing and these types of concerns being raised in the
14
US it is hardly surprising that S&P has adopted a negative tone towards both US and Canadian
15
utilities.
16
Q.
IS THE US EXPERIENCE RELEVANT FOR CANADA?
17
A.
To some extent yes. Although we have not had the problems that they have had in the
18
US that does not mean that we can’t have them. Further, with Duke Energy’s acquisition of
19
Union Gas there could always be the sort of problems that have bedevilled Enron and other US
20
UHCs, where when the parent ran into problems they looked to the regulated subsidiary to strip
21
it of cash. As of the current point of time Union’s bonds seem to trade on their DBRS rather
22
than the S&P rating. However, during 2005 Union obtained loans from and made loans to its
23
immediate parent Westcoast indicating that Union does not truly manage its own cash flow.
24
This lack of structural insulation makes it impossible for Union Gas to have an S&P bond
25
rating that reflects its own risk. At some point in the future this may cause its borrowing cost to
26
reflect Duke Energy’s BBB bond rating, rather than its own credit. I would recommend that the
27
Board take measures to structurally insulate both Union and EGDI from its parents to ensure
28
that ratepayers only pay the legitimate borrowing cost attached to the regulated activities.
29
Otherwise there may be a long run risk stemming from Union’s ownership by a risky US
- 59 -
1
energy company as well as potential short term arguments as to whether Union’s BBB rated
2
debt costs should be passed on to Union’s ratepayers.
3
4
5
Q.
DOES UNION GAS HAVE FINANCIAL FLEXIBILITY WITH YOUR
RECOMMENDATION?
6
A.
Yes. Union filed a statement with the OSC as to its interest coverage ratio on September
7
20, 2005 which stated:
8
9
10
11
12
13
14
15
16
17
18
19
20
EARNINGS COVERAGE RATIO
Earnings Coverage Ratio
After giving effect to all issues and retirements of long-term debt since December 31, 2004, the
annual interest requirements on the consolidated long-term debt of the Company for the twelve
months ended September 30, 2005 were $155 million and for the twelve months ended
December 31, 2004 were $155 million. Consolidated net income of the Company for the twelve
months ended September 30, 2005, calculated before interest on consolidated debt and income
taxes, amounted to $325 million, which is 2.10 times the Company’s annual interest
requirements on consolidated long-term debt for that period. Consolidated net income of the
Company for the twelve months ended December 31, 2004, calculated before interest on
consolidated debt and income taxes, amounted to $349 million, which is 2.25 times the
Company’s annual interest requirements on consolidated long-term debt for that period.
21
So with its current allowed ROE, embedded interest cost and 35% common equity ratio Union
22
had an ICR of 2.25 for 2004 and 2.10 for 2005 for its September year ends. These both exceed
23
the target of 2.0 in the trust indenture for issuing unsecured debt. Further in the EBRO499
24
Decision the Board accepted that Union would have the following ICRs at a 9.64% ROE
25
26
ICR
1999
200
2001
2002
2003
2.08
2.02
2.09
2.15
2.16
27
28
So the Board has accepted in the past that ICRs marginally above 2.0 and less than Union’s
29
ICRs in 2004 and 2005 were acceptable. Further Union’s marginal ICR is significantly higher
30
than these levels.
31
- 60 -
1
Q
WHAT IS UNION’S MARGINAL ICR?
2
A.
The marginal ICR is the interest coverage ratio using a consistent set of financial
3
parameters. It reflects the impact of a marginal increase in the rate base. For example, currently
4
spreads on A rated long term debt are 105 basis points over the equivalent long Canada yield.
5
This long Canada yield then implies an allowed ROE using the Board’s adjustment formula
6
and Union’s current long term borrowing cost. Using these parameters and the statutory tax
7
rate we can estimate what the ICR is for an expansion of the rate base since the funds used to
8
finance this expansion will have these parameters. The advantage of the marginal ICR is that it
9
uses a consistent set of financial parameters and thus shows the steady state ICR that is
10
consistent with the Board allowed financial parameters.
11
To be consistent with the company’s evidence I estimate the marginal ICR using the rates in
12
effect at the time that the company prepared its evidence. These are the rates used by Ms.
13
Elliott (E1, Tab 1).With long Canada yields of approximately 4.35%, Union’s long term debt
14
would cost about 5.40%. However, this assumes that all Union’s debt is financed using 30 year
15
debt which is impractical since Union like most companies staggers its debt issues to reduce
16
refinancing risk. Ms. Elliot assumes a debt cost of 5.0% for 2007 which reflects this staggered
17
bond financing. For example in September 2005 Union raised $200 million with a ten year
18
bond issue at a cost of 4.64% and plans to do the same with an estimated 4.70% cost in 2006.
19
The equivalent cost of preferred share financing was estimated by Ms. Elliott at 4.91%. Using
20
the October consensus forecast Ms. Elliot then uses an allowed ROE of 8.89%, this is what I
21
would regard as consistent with the current long Canada rates since these are used for the
22
current debt costs as well as the allowed ROE through the adjustment formula.
23
24
25
26
27
28
29
30
Union’s after tax weighted average cost of capital using the current Board approved financing
ratios is then:
WACC and Interest Coverage
Pre-Tax
After-Tax
Weighting
Debt
5.0 %
3.2%
61.5%
Preferred equity
4.91
4.91
3.5
Common equity
8.89
8.89%
35
- 61 -
1
So the after tax WACC using Board approved parameters is 5.25%. Union’s pre-tax weighted
2
average cost of capital is the then the earnings before interest and tax (EBIT) cost of 8.21%
3
(5.25/(1-Tax rate of 36%)), so the interest coverage ratio is simply this EBIT divided by the
4
interest cost or 2.67X.
5
What the marginal ICR indicates is that Union’s ICR is very favourable with current board
6
approved parameters and the company should have no problems accessing financial markets.
7
An ICR of 2.67X is well above the 2.0 level contained in Union’s trust indenture so the
8
company should be able to issue medium term notes (MTN’s). An ICR of 2.67 is significantly
9
above Union’s historic ICR which has hovered above 2.0 and is well above the ICR of its long
10
time parent Westcoast Energy Inc, which consistently maintained an A- DBRS bond rating
11
even though it rarely if ever had an interest coverage ratio above 2.0.14
12
The fact that Union’s ICR has increased from historic levels is entirely consistent with the fact
13
that its allowed ROE varies by 75% of the change in the long Canada bond yield whereas its
14
current borrowing cost varies by 100%, that is, Union’s borrowing costs over equivalent long
15
Canada bond yields are relatively stable. As a result, Union’s ICR should increase as long
16
Canada bond yields decline thereby increasing its financial flexibility and access to capital
17
marekts. I am therefore at a complete loss in understanding how the current low levels of long
18
Canada bond yields can be regarded as risk factor: they most definitely are not.
19
To illustrate how Union’s ICR varies with the long Canada bond yield assume that Union has
20
no preferred shares, its debt costs an average 65 basis points over the long Canada bond yield
21
and its allowed ROE varies by 75% of the long Canada bond yield through the ROE adjustment
22
mechanism. The ICR that results from the formula can then be graphed as a function of the
23
long Canada bond yield. With the current 8.89% forecast allowed ROE the ICR with these
24
assumptions is 2.50X, this is less than I calculated above since the 3.5% preferred share
14
Duke Energy’s ICR is not relevant since it is considerably more risky than Union and has an S&P
bond rating of BBB. Further US companies are not normally restricted by ICRs in their trust
agreements.
- 62 -
1
component increases the ICR since preferred share dividends are not funded debt and have to
2
be declared quarterly.
3
What the graph indicates is how quickly the ICR increases with a decline in long Canada bond
4
yields. For example with a 3.0% long Canada bond yield Union’s current borrowing cost is
5
assumed to be 3.65% and the allowed ROE 7.88%. In the face of a 1.35% drop in long Canada
6
bond yields the allowed ROE is assumed to have only dropped by 75% of this or 1.01% so the
7
ICR increases to 2.82X.
The ICR and Long Canada Yields
4.6
4.4
Yield
4.2
4
3.8
3.6
3.4
3.2
3
2.45
2.5
2.55
2.6
2.65
2.7
2.75
2.8
2.85
ICR
8
9
10
The graph indicates how favourable the Board’s policies are towards Union’s financing
11
flexibility in a declining interest rate environment. Ms. Elliott was asked to reproduce the
12
above results in interrogatory response J2.25 but the answer is difficult to understand and
13
clearly incompatible with what was asked in the question.
14
Q.
WHY HAVE YOU USED A CURRENT DEBT COST IN THE ICR ESTIMATE?
15
A.
These are reference numbers as the current cost of debt produces an interest coverage
16
ratio consistent with the current Board allowed ROE and common equity ratio. On the other
- 63 -
1
hand, it makes no sense to target a particular interest coverage ratio and allow a higher ROE
2
simply because one utility has a higher embedded cost of debt. This would produce the bizarre
3
result that between two otherwise identical utilities, the one with the higher embedded cost of
4
debt would get the higher ROE. Obviously such a higher ROE has nothing to do with the legal
5
requirement to offer a fair return based on Mr. Justice Lamont’s definition. However, more
6
important it implies that ratepayers that are already saddled with a higher embedded cost of
7
debt would be hit a second time with a higher ROE or more common equity. Finally low
8
coverage ratios caused by high embedded interest costs are essentially a temporary
9
phenomenon that will gradually correct itself. In these circumstances it makes no sense to
10
change the common equity ratio, which tends to be viewed as “permanent” in response to an
11
essentially temporary phenomenon.
12
For example, Union’s embedded debt cost is 7.46% which is a decline from the 8.45%
13
approved by the Board in 2004. The decline of almost 1.0% reflects the refinancing of long
14
term debt issued at historically higher interest rates with current debt costing an average of
15
5.0%. For example Union has retired 9.75% series II debentures in December 2004 and 8.85%
16
debentures in September 2005. Both were refinanced at lower rates causing the embedded
17
interest cost to fall. Similarly Union is forecasting the refinancing of 7.80% debentures in
18
December 2006 and some 5.19% MTNs in December 2007, both of which should cause the
19
embedded interest cost to decline further. According to Union Gas’ 2005 annual report in
20
footnote 10 they indicate that debentures with a 13.50% coupon will be retired in 2008, 10.75%
21
in 2009, 11.55% and MTNs of 7.20% in 2010, 10.625% and MTNs of 6.65% in 2011. So over
22
the next five years several high coupon issues will be refinanced causing the embedded interest
23
cost to fall.
24
It is important in this respect to understand that in the calculation of the ICR for issuing further
25
debentures and MTNs the interest on the retired debt is subtracted and then the interest on the
26
new debt added. As a result all else constant in a low interest rate environment Union’s access
27
to debt markets is improved as the embedded interest rate falls and the ICR increases. Ms.
- 64 -
1
Elliott was asked to confirm this in interrogatory response J2-29c, which she did with the
2
qualification that the ICR still has to be above 2.0.
3
4
Q.
WHY HAVE YOU USED A STATUTORY TAX RATE?
5
A.
Because this is Union’s steady-state tax rate. Differences between actual and statutory
6
tax rates are mainly caused by the use of tax depreciation at higher rates than the straight line
7
depreciation used for revenue requirement purposes. The use of flow through accounting then
8
means that a rapidly growing utility will generally have a lower tax rate than a declining or
9
stable rate base utility. For example in 2005 Union’s effective tax rate was 29.4% rather than
10
the statutory 36.1%. The fact that Union’s tax rate is lower than the statutory rate means that its
11
interest coverage ratio will be lower than the steady state. However, it makes no sense to award
12
a higher ROE or common equity ratio simply because a utility is expanding to serve new
13
markets, since expansion is a sign of lower not higher risk
14
Q.
WHAT IS UNION’S ICR WITH ACTUAL RATES?
15
A.
Using the actual tax rate of 29.4% and the embedded debt cost of 7.46% produces the
16
following WACC
17
18
19
20
21
22
Debt
Preferred equity
Common equity
WACC and Interest Coverage
Pre-Tax
After-Tax
7.74 %
5.46%
4.91
4.91
8.89
8.89%
Weighting
61.5%
3.5
35
23
So the after tax WACC is 6.52% reflecting the higher after tax interest cost and the pre tax
24
WACC 9.24%. This produces an ICR of 2.02X. This is consistent with basic arithmetic that
25
with a higher embedded debt cost and lower tax rate the interest coverage ratio is bound to be
26
lower. This ICR will increase as the embedded interest cost falls and Union’s tax rate rises as
27
its growth slows down. However, there is something distinctly peculiar about rewarding the
28
shareholders with a higher ROE or more common equity simply due to the fact that the
- 65 -
1
company is experiencing more rapid growth and faces a lower corporate tax rate. Normally a
2
lower tax rate means a lower revenue requirement and lower rates, which makes the company
3
even less risky as it is more “competitive,” not less.
4
5
6
7
Q.
BUT THE COMPANY SAYS IT MAY NOT BE ABLE TO ISSUE LONG TERM
DEBT ISN’T THIS A PROBLEM?
8
A.
No. In answer to interrogatory response J2.28 Ms. Elliott stated
9
“Union has not previously had a specific case where the company has not been able to
issue debt to finance capital investment. There have been situations when the company
was limited by the interest coverage test to the timing and the amount of the debt issue.”
10
11
12
13
14
In other words, Union has experienced previous periods of rapid growth and low tax rates or
15
poor ROEs due to weather fluctuation that have limited its access to the debenture market
16
without any obvious negative consequences. There is nothing new about the current situation in
17
fact it is distinctly more favourable as the low interest rate environment will improve Union’s
18
ICR over time.
19
20
It also has to be remembered that the MTN or debenture market is only one form of capital.
21
The ICR test simply restricts Union’s access to this form of capital until it can satisfy the ICR.
22
In this respect it simply adds up all the interest on funded debt, that is, debt with a maturity of
23
longer than eighteen months, and restricts the company from issuing more of this debt unless
24
the ICR is greater than 2.0 with the new debt interest included. It does not restrict Union’s
25
access to other forms of capital or its ability to raise capital to provide service. The ICR does
26
not for example apply to commercial paper, bank debt, mortgage bonds, junior subordinated
27
debt or preferred shares.
28
29
What is highly unusual is that Union’s actual financing is as aggressive as possible with respect
30
to the MTN market and the ICR restriction. Ms. Elliott in Table 2 (E1, Tab 1, page 3) indicated
31
that Union’s financing is 65.43% long term debt subject to the ICR with a surplus of 4.0%
32
carried in marketable securities. This financing strategy has existed in 2004 , 2005 and 2006
- 66 -
1
and is forecast for 2007; it is highly unusual for two reasons. First throughout this period the
2
cost of short term debt has been below that on long term debt since the yield curve has been
3
upward sloping or normal. As a result, it would have made more sense to have been heavily
4
weighted towards short term debt and had a policy of say 51.5% long term debt and 10% short
5
term debt, rather than the 65.5% long term and -4.0% short term strategy actually followed.
6
Secondly it makes absolutely no sense to consistently borrow and lend simultaneously, since
7
you incur the spread between the borrowing and lending rates. Union’s actual financing
8
strategy violates one of the most basic propositions in finance which is to avoid transactions
9
costs. I doubt very much that an independent stand alone Union would have followed such a
10
strange financing policy.
11
12
As far as the ICR is concerned Union has more funded interest than it actually needed, so that
13
its ICR appears worse than it need be. In fact Union seems to have stacked the deck against
14
itself and followed a financing strategy that cuts itself off from the debenture market when it
15
doesn’t need to be.15 In particular if Union followed a 51.5% long and 10% short term
16
financing policy its ICR would be much higher as the interest cost on the short term debt is not
17
included in the ICR calculation. Further if Union wanted to hedge its exposure to short term
18
interest rates it could simply hedge this exposure using interest rate caps or swaps into longer
19
term debt to create synthetic MTNs. In interrogatory response J2.29 Ms. Elliott stated that this
20
interest would be included in the ICR calculation. However, this conflicts with the answer
21
given in EBRO499 where in OCAP interrogatory response J22.E1.35 Union stated that
22
“interest on commercial paper swapped into fixed rate through an interest rate swap would not
23
be included in the calculation of consolidated interest requirements.” This 1998 answer is
24
consistent with other interpretations of the ICR and the fact that it only applies to interest from
25
debt with a maturity exceeding eighteen months. Again Union seems to have chosen a position
26
that puts it in the worst possible light.
15
In J2.30 Ms. Elliott was asked to break out the different factors driving Union’s ICR. She refused
saying the declining interest rate environment was the most important. It would be interesting to know
how much is due to Union’s strange financing policy.
- 67 -
1
It also has to be remembered that Union like many companies has switched to unsecured MTN
2
and debenture financing and away from mortgage bond financing. The difference is significant.
3
Consider an individual borrowing to buy a house. You can either register the bank’s claim on
4
the house through a mortgage or sign an unsecured note. If anyone has tried the latter they will
5
discover the obvious fact that the bank will lend more at a lower rate with a mortgage than
6
through an unsecured loan. The same applies to Union, although to a lesser degree. If Union
7
finds itself shut out of the unsecured bond market it is like any of us finding that the bank will
8
not finance our house purchase with an unsecured signature loan. One solution is for Union to
9
issue first mortgage bonds. Ms. Elliott confirmed (interrogatory response J2.29) that Union can
10
still issue first mortgage bonds and while it is true that the interest is included in the ICR
11
calculation for issuing MTNs, the ICR does not restrict Union from issuing mortgage bonds.
12
13
I do not advocate Union issuing first mortgage bonds, since the more obvious strategy is simply
14
to reverse its “over-funding” strategy and return to a more sensible strategy of long term debt
15
mixed with some short term debt. In this respect, if needed Union can take out short term bank
16
loans matched to maturing high coupon bonds, so that once the bonds are redeemed Union can
17
pay off the bank debt and high coupon debt with the MTNs it can then issue when it meets the
18
ICR test. Alternatively instead of bank debt Union can issue junior subordinated debt similarly
19
matched to the maturity dates of the high coupon debt. As a final resort Union can issue some
20
medium tem preferred shares, since as Ms. Elliott confirmed in interrogatory response J2.29
21
preferred share dividends are not included in the calculation of interest in the ICR. Instead they
22
tend to increase the ICR as documented earlier with Union’s marginal ICR estimate.
23
24
The basic point is that I have consistently argued that low coverage or ICR tests should not be
25
used to justify either a higher ROE or deemed common equity ratio. This is akin to trying to
26
solve debt market problems in the equity market, which doesn’t make any sense. It is also
27
fundamentally unfair and saddles ratepayers with unfair and unjust costs, since it means that
28
ratepayers are paying for the high embedded debt costs twice. They pay once and directly
29
through the higher embedded debt cost and then again indirectly in a higher allowed ROE or
30
deemed equity ratio due to the lower coverage ratio. If the Board believes that the allowed
- 68 -
1
ROE and deemed common equity ratios are fair, but the resulting coverage ratio is too low to
2
use MTNs, then the utility has to be more flexible in its financing choices. Nowhere is it
3
mandated that “fair and reasonable” regulation requires that the utility be able to finance long
4
term assets with unsecured signature loans. In this respect, PNG can not issue MTNs and relies
5
on medium term loans from RoyNat that it pays off in monthly payments like a mortgage. A
6
similar situation exists for most small utilities that do not have the size to be able to access the
7
MTN market.
- 69 -
1
5.0
WEIGHTED AVERAGE COST OF CAPITAL (“WACC”)
2
Q.
WHAT IS THE BASIS FOR WACC?
3
A.
Before considering a firm with debt, first consider the simplest problem in finance that
4
of a 100% equity financed firm. I will use K as the investors required return, so that if the
5
firm’s earnings are a perpetuity its value, V, is its forecast perpetual earnings of X divided by K
6
or
V=
7
X
K
(4)
8
If the earnings are expected to be $1mm and the typical investor requires a 10% return then the
9
firm’s value is $10mm. Note that I said the investor “requires,” since this is the meaning of a
10
fair return: it is the return required by the investor relative to other investment opportunities.
11
Only the investors knows what they require, but if we take the above equation we can reverse it
12
and infer what the investor requires from the current market value and the firm’s expected
13
earnings as
14
K=
X
V
(5)
15
With the example numbers, $1mm in earnings valued at $10mm implies that the investor
16
requires a 10% return. Since the firm has to meet this requirement we also refer to this 10% as
17
the firm’s cost of equity capital: what the firm has to pay to raise equity. The interpretation of
18
this is straight forward, if the stock market requires a $1mm in earnings to support the $10mm
19
in market value, the market value will change if the earnings expectation changes.
20
The above perpetuity equation is used to value conventional preferred shares, but it illustrates
21
the standard result in finance that market values and required returns or “costs of” vary
22
inversely. For example, we value bonds using the same formula adjusted for the fact that the
23
fixed payments do not go on in perpetuity, but instead are truncated at the bond’s maturity date.
- 70 -
1
However, we still get the same result: when market interest rates fall, and with them investor
2
required rates of return, then bond market values go up and vice versa.
3
4
Further the expected earnings of the firm can be broken out into the rate of return, r, on the
5
book value of the firm’s assets A or
V=
6
rA
K
(6)
7
if the firm’s $1mm in earnings are the result of earning a 10% return on a $10mm book value
8
then we get the additional insight that the firm’s market to book ratio is simply
V
r
=
A K
9
(7)
10
In this case, since the firm is expected to earn 10% and the investor requires a rate of return of
11
10%, then the market value is equal to what the investor has contributed, that is, the firm’s
12
market to book ratio is equal to 1.0.
13
The above results are perfectly general and indicate that there are two basic ways in which a
14
firm can increase its market value: the first is by increasing its earnings through a higher rate of
15
return r, the second is through lowering its cost of capital, K. In both cases, the market value
16
goes up. For example, if the firm’s expected rate of return increases from 10% to 11% its
17
earnings increase to $1.1mm and, all else constant, its market value increases to $11mm.
18
Alternatively, if it lowers its cost of capital to 9%, then its market value increases to
19
$11.11mm. These two ways of increasing value are the heart of corporate finance and reflect
20
the value of investment decisions and financing decisions respectively. Note also that given the
21
higher stock market value we can see that the market to book ratio now exceeds 1.0, indicating
22
that the investor has bid up the value of the shares since they are getting more than they
- 71 -
1
originally anticipated when the firm raised and invested the $10mm in the firm’s assets.16
2
3
The above implications for the market to book ratio are important for understanding how
4
capital markets work. It is a central result in finance that when investors receive more than they
5
require, market values increase, and with them the market to book ratio. Conversely, when they
6
receive less than what they require; the market to book ratio drops below 1.0. This is observed
7
every day, for example, in the pricing of debt securities, where government bonds with higher
8
coupons (interest rates) than current market rates require sell at a premium to their par value
9
and those with smaller coupons sell at a discount. In the debt markets these bonds are referred
10
to as premium or discount bonds, but we could just as easily refer to them as bonds with market
11
to book ratios above 1.0 and below 1.0. The information and meaning is exactly the same.
12
I developed the previous example to emphasise the importance and generality of the market to
13
book ratio and the important relationship between what the firm is earning, r, and what the
14
stock market requires, K. In the example, the firm is earning 10% and the investor requires
15
10%, so the stock market is telling the firm “only invest in new projects that earn at least a 10%
16
rate of return, otherwise you are wasting our money.” If, for example, the firm raised $10mm
17
from investors and invested in a new $10mm project earning 8% in perpetuity, its new stock
18
market value would be $18mm
$18mm =
19
$1mm + $0.8mm
0.1
20
Its value has gone up by $8mm, but only at the cost of investing an additional $10mm, so the
21
firm has wasted $2mm of stockholder’s money. This change in value is called the net present
16
The example assumes a 100% ROE regulated firm, as it involves more non-regulated assets it is more
difficult to look at the market to book ratio as a signal. It is then in the interests of the regulated firm to
make observing this market to book ratio as difficult as possible. I refer to this as “looking through a
dirty window,” where there is no incentive for the utility to clean the window. It may not be an accident
that there are so few pure regulated utilities left.
- 72 -
1
value (NPV) of the investment. Corporate finance is focussed on the firm making decisions that
2
increase shareholder value and not waste it. The key message is that the firm should only invest
3
in projects earning at least the firm’s cost of capital, which in the example is 10%.
4
Now suppose in the example market interest rates drop and with them the investor’ required
5
rate of return. This has been one of the central features of capital markets over the last twenty
6
years. Long-term interest rates have been on a long-term downward trend since peaking at over
7
18.0% in September 1981. Suppose in our example the firm’s cost of capital drops from 10% to
8
7%. In this case, the firm’s market value would increase to $14.3mm and the market to book
9
ratio increases to 1.43X. Again the market to book ratio indicates that the firm is now earning
10
10%, when investors only require 7%. Notice, however, that the original investors have earned
11
a 43% capital gain that they did not anticipate, since they expected to earn their 10% rate of
12
return through the $1mm in earnings on their $10mm investment. If this were a utility the
13
stockholders would have earned an excess return and we can note this by looking at the market
14
to book ratio: any market to book ratio significantly above 1.0 indicates that investors have
15
earned a return that is above a fair and reasonable rate of return.17
16
Further, the 8% investment that previously destroyed $2mm in value now increases it, since 8%
17
exceeds the new 7% cost of capital. The new value for the firm with the investment is
$25.7mm =
18
$1mm + $0.8mm
0.07
19
This $25.7mm is the $14.3mm market value of the original investment, plus the $10mm cost of
20
the new investment plus an additional $1.43mm NPV from the new investment.
21
The example indicates that the firm has to constantly monitor its cost of capital: a project that it
22
would not accept when its cost of capital was 10%, it will accept when it drops to 7%.
23
Moreover, this investment yardstick is NOT a rate of return earned by other firms, that is, other
17
If the firm is sold for $14.3 mm, new investors only earn their 7% required, since their return is based
on the $14.3mm investment cost, that is., $4.3mm of “goodwill” plus the $10mm original cost.
- 73 -
1
firms “r,” but the market cost of capital, K. The rate of return, r, earned by other firms is
2
irrelevant in corporate finance, since the firm has to satisfy its stockholders, not other firms.18
3
Moreover, non-regulated firms will have a variety of projects earning more than the cost of
4
capital; some will earn say 12%, and some 11% all the way down to the last project that is
5
accepted with a 7% return. Consequently, the expected average return of a competitive firm
6
should always be higher than its cost of capital, reflecting the fact that it will have some
7
projects earning above average rates of return.19
8
Q.
HOW DOES ADDING DEBT CHANGE THIS?
9
A.
The basic principles are exactly the same. Suppose the firm continues to have $10mm
10
in assets earning 10%, but there is now $5mm in debt at a cost of 5% and $5mm in equity at a
11
cost of 15%.20 The firm’s weighted average cost of capital (WACC) is then 10% and the firm’s
12
equity value is determined by the net income to the common shareholders discounted at the
13
equity cost. As discussed earlier, the rate of return on the firm’s assets is called its return on
14
investment (ROI), which is the return prior to meeting interest payments. In the example, the
15
ROI is 10% so the equity market value is determined as the earnings before interest of $1mm
16
(ROI times the $10mm in assets) minus the interest of $0.25mm (5% times $5mm) or
17
$0.75mm. If this $0.75mm in net income is discounted at 15%, then we get the equity market
18
value of $5mm. That is,
$5mm =
19
20
$1mm − $0.25mm
0.15
or algebraically
18
This is the core reason why comparable earnings estimates, which simply try to estimate r, are of no
value in public utility regulation.
19
This further shows that if corporate rates of return (comparable earnings) are to be relevant it is the
marginal rate of return that matters not the average rate that is estimated from financial statement data.
20
15% is used simply for ease of calculation.
- 74 -
$5mm =
1
ROI * A − K b * B
Ke
2
where A is the total book value of assets, B is the amount of debt financing ($5mm) and I have
3
subscripted the two costs, b for debt and e for equity.
4
In this example the market value of the equity is $5mm so that the total enterprise value (V), or
5
overall market value of the firm (debt plus equity), is the $5mm equity value plus the $5mm of
6
debt or $10mm. This calculation is an example of the flows to equity (FTE) method of
7
valuation, where the flows to the equity holder are discounted at the cost of equity capital to
8
directly determine the value of the equity. However, to do this calculation we need the value of
9
the debt financing and most corporate investment decisions are separated from these financing
10
decisions. Consequently, it is conventional to rearrange this equation to get the WACC. First
11
multiply through by the cost of equity,
E * K e = ROI * A − K b * B
12
13
where I have substituted E for the equity market value. Second, group the equity and debt costs
14
and factor for the overall market value to get,
V (K e
15
16
E
B
+ K b ) = ROI * A
V
V
dividing through we get
- 75 -
V =
1
ROI * A
E
B
Ke + Kb
V
V
(8)
2
where the market value is equal to the pre-interest earnings discounted at the weighted average
3
cost of capital (WACC). In our example $1mm discounted at the WACC of 10% gives the total
4
market value of $10mm. The equity value is then this $10mm value minus the $5mm in debt or
5
the same $5mm as calculated using the flows to equity approach.
6
The WACC simply recognises the different sources of finance available to the firm and
7
averages them to get an overall cost of capital. In this sense the cost of capital is a blended cost
8
of financing to the firm, but once this is done all the previous results hold just as before. For
9
example, if the ROI exceeds the WACC then the market value increases and the market to book
10
ratio is again above 1.0. For example if the ROI increases to 11% and the WACC stays at 10%
11
then the total market value increases to $11mmm. The equity value is then $11mm minus the
12
$5mm in debt, so the equity value increases to $6mm and the equity market to book ratio
13
increases to 1.2X. Similarly, if the cost of equity declines to 13% from 15%, then all else
14
constant, the WACC drops to 9% and the market value increases to $11.11mm. Again with
15
$5mm in debt the equity value increases to $6.11mm for a market to book of 1.22X. Finally, if
16
the WACC drops to 9%, this becomes the hurdle rate for new investments; that is, the ROI on
17
new investments has to be greater than 9%.
18
Adding corporate income taxes does not materially change any of the basic results. All that
19
happens is that the after interest earnings become taxable and it is this after tax net income that
20
is discounted by the equity cost, that is,
21
Equity =
( ROI * A − K b * B)(1 − T )
Ke
22
where T is the corporate tax rate. Rearranging, as before, means that the after tax ROI has to
23
beat the after tax WACC or
- 76 -
V =
1
ROI (1 − T ) * A
E
B
K e + K b (1 − T )
V
V
(9)
2
The only difference is that since interest is tax deductible, whereas equity costs are not, the
3
after tax ROI has to exceed an after tax WACC. This is normally referred to as the WACC since
4
most firms pay tax, but Drs. Kolbe and Vilbert refer to it as the ATWACC to differentiate it
5
from the normal utility WACC where the return is included in the revenue requirement.
6
It is fundamental to corporate finance that the WACC uses market values. This means for
7
example, that the debt and equity ratios use market or target values for debt and equity divided
8
by the total enterprise value. The WACC then gives the current yardstick that the firm has to
9
beat in order to create shareholder value, that is, to increase the firm’s market value. Only by
10
calculating the WACC in this way can the firm be sure that it is accepting projects that enhance
11
shareholder value, that is, have positive NPVs, rather than destroying it.
12
13
Q
SINCE YOU FIRMLY BELIEVE IN WACC (OR ATWACC), WHY DON’T
YOU USE IT IN YOUR TESTIMONY?
14
A.
The basic difference is that regulators are not concerned with maximising or enhancing
15
shareholder value; their mandate is to set “fair and reasonable” rates and frequently this sets
16
them at variance with maximising shareholder value, since regulation should never be designed
17
to rubber stamp market values. This means that the regulator sets rates and through them the
18
firm’s ROI, whereas for non-regulated firms the ROI is determined in the marketplace.
19
Consequently, the ROI is changed for the regulated firm to make sure that the return to the
20
stockholders (ROE) is fair.
21
To continue with the previous (no tax purely for simplicity) example, where the WACC is 10%
22
and the equity cost 15%, suppose the regulator institutes some risk reduction techniques such
- 77 -
1
as the use of a forward, instead of an historic test year,21 or the use of deferral accounts. As a
2
result, the equity cost drops to 11%. This is the critical example the Board needs to be aware
3
of, since as I indicated earlier, we have been in a period of long run declining interest rates
4
since 1981 and the problem that Boards have been largely dealing with is declining equity
5
costs.
6
In the example, everything else is held constant, so the debt is still $5mm and the rate base
7
(total assets) $10mm; the only thing that has changed is the equity cost. The equity value can
8
be determined from the flows to equity formula as
$6.818mm =
9
$1mm − $0.25mm
0.11
10
In this case the equity holders recognise the reduction in risk and bid up the stock market value
11
from $5mm to $6.818mm for an extra capital gain of 36.4% over and above their fair return.22
12
If the firm is now in a rate hearing to adjust its ROE, a tip off to the regulator is that the market
13
to book ratio has gone from 1.0 to 1.364X (6.818/5), so intuitively by lowering the firm’s risk
14
and seeing the market value increase the regulator knows that the allowed ROE has to be cut.
15
The obvious thing for the regulator to do is simply get expert opinion estimating the equity
16
cost, and if this is unbiased, notice and cut the allowed ROE to 11%. The equity value will then
17
return to $5mm and the stockholders will continue to earn a fair return on their $5mm
18
investment. The question is then what does estimating the WACC add to this process?
19
Assuming there is no bias to estimating the equity cost at 11.0% the new WACC is
WACC = 0.11 *
20
6.818
5
+ .05 *
11.818
11.818
21
Forward test years remove any inflationary bias involved in the use of an historic test year adjusted
for specific identifiable changes. With the decline in inflation most of the need for forward test years is
removed.
22
As perpetuities they get their fair return as the earnings are paid out as a dividend.
- 78 -
1
or 8.46%.23 The most important thing to note is that the WACC uses market value weights and
2
since the equity market value has gone up to $6.818mm, the WACC uses an equity ratio of
3
57.7% and a debt ratio of 42.3%, rather than the assumed regulated weights of 50% for both.
4
The reason for the use of market value weights is that the WACC is the minimum rate of return
5
the firm has to earn to maintain its market value, which has increased from $10mm to
6
$11.818mm. Theoretically, it makes no difference whether this $11.818mm is the result of
7
actually raising $11.818mm, or whether it’s the current market value of the original $10mm
8
investment as it is in this case. The point is simply that using WACC as a cut off rate is simply
9
that it reflects what the firm has to earn to sustain current market values. In particular, the new
10
WACC of 8.46% is applied to the market value of $11.818mm. In contrast, the regulator
11
should not be interested in sustaining current market values, since in the example it is clear that
12
the allowed ROE has to be cut and implicitly that market value has to fall. Moreover, the
13
regulator has to determine a fair return on the book value of the investment, that is, the rate
14
base, not the firm’s market value. In this sense, there is a fundamental contradiction to applying
15
the conventional WACC to regulated firms.
16
However, suppose the Board tries to apply WACC. First, note that this exercise is much more
17
prone to error, and as a result subjective, than just estimating the fair return directly. This is
18
because as well as estimating the equity cost correctly, you have to estimate the market cost of
19
debt, not the embedded cost, the financing weights and the appropriate tax rate. All of these
20
components are subject to error, since many issues of debt are not traded and as a result it is
21
difficult to estimate either their cost or their market value. However, assuming all these
22
additional estimation problems away, suppose the correct 8.46% WACC is estimated and
23
awarded the regulated firm, what does this do?
24
If the regulator accepts this WACC, the equity value is $5.42mm or
23
Note that discounting the $1mm in pre interest earnings by 8.46% gives the total enterprise value of
$11.818mm. This is the new cut off rate for evaluating the firm’s investments.
- 79 -
V =
1
.0846 * $10 − .25
.11
2
Although the ROI is reduced from 10% to 8.46%, it is not reduced to the correct ROI of 8.0%,24
3
so the equity market value is still $0.42mm higher than it needs to be. The reason for this is that
4
using market value weights in the WACC puts greater emphasis on the higher equity cost than
5
the debt cost. For non-regulated firms this is correct since the objective is to maintain these
6
market values and create wealth. However, it is totally incorrect for a regulator who is tasked
7
with awarding fair allowed returns and implicitly causing market values to drop when allowed
8
ROEs are too high. By estimating and applying a market based WACC the effect of the higher
9
allowed ROE is perpetuated by its impact on the higher equity market value.
10
Over time, if nothing else changes, the excess value will be removed. For example, with the
11
new market value of $5.42mm the new WACC is
WACC = 0.11 *
12
5.42
5
+ .05 *
10.42
10.42
13
or 8.12%. Again if everything remains constant, in the next rate hearing the regulator would cut
14
the allowed ROI to this level and the equity market value would fall again until after successive
15
rounds it ends up at the $5mm fair value for the equity. Note that the regulated firm, as well as
16
others with a vested interest in the firm as an investment, may complain about the regulator
17
being tough by repeatedly cutting the allowed ROE, but it is not being tough at all. The ROE is
18
still above the fair ROE. However, by using market value weights in the WACC and by shifting
19
the focus from the ROE to the WACC, this adjustment process is drawn out to the stockholders’
20
benefit. Further it gives the regulated firm an opportunity to bring up other arguments that may
21
delay even this adjustment. Consequently the adoption of WACC based regulation delays the
22
adjustment process to the stockholders’ benefit.
24
The correct regulated WACC is the average of the debt and equity costs using regulated book value
weights, in this case 50%.
- 80 -
1
The basic insight from this discussion is that by using market values in WACC, the resulting
2
cost of capital is higher than a fair return, since the higher equity cost is given a greater weight.
3
Further if the firm is a pure ROE regulated utility it tends to “rubberstamp” the use of market
4
values and is contrary to “fair and reasonable” regulation.
5
Q.
CAN YOU ELABORATE ON THE LAST COMMENT?
6
A.
As previously explained, the market to book ratio is the basic signal as to whether or
7
not investors are being fairly treated. When we add in flotation costs and a desire to allow the
8
regulated firm to access markets at all time, most would believe that the market to book ratio
9
should be marginally above 1.0, say 1.10. In this way, after incurring flotation costs, the firm
10
should be able to net out the equity book value per share and avoid any dilution of the existing
11
shareholder’s interests. However, apart from this minor deviation from book values, the
12
essential point is that the correct financing weights for a regulated firm should be the regulated
13
capital structure weights, not the market value weights. To incorporate into the regulatory
14
process a regulated firm’s market value is to rubberstamp investor expectations, however
15
unrealistic, and delay the adjustment to a fair and reasonable value for the allowed ROE.
16
The Alberta EUB has directly addressed this question on a number of occasions. For example,
17
in connection with comparable earnings testimony the EUB stated (Generic Cost of Capital
18
Decision U-200452, page 24)
19
20
21
“The Board considers that the application of a market required return (i.e. required
earnings on market value) to a book value rate base is appropriate in the context of
regulated utilities.”
22
That is, you estimate a market opportunity cost, such as that from the CAPM, and apply it to
23
book values, not market values as is the assumption in WACC.
24
In explicitly considering the usefulness of ATWACC the EUB stated (Decision U-99099, page
25
300)
26
27
“The Board observes that the intrinsic long-run value of a pure play regulated entity is
best represented by book value. In other words, the present worth of future regulated
- 81 -
1
2
3
4
earnings, discounted at the allowed return, is by definition equal to book value
assuming achieved regulated earnings on average equal allowed regulated earnings.
Accordingly, the Board considers that book capitalization represents the best indicator
of the long-run market capitalization for a pure play regulated firm.”
5
It is difficult to see how a regulator could say anything other than what the EUB stated above,
6
since to accept a market to book much above 1.0 is in effect to rubberstamp unrealistic investor
7
expectations or to admit that allowed ROEs are too high. The EUB further recognised this
8
when it went on to say (U99099, page 303)
9
10
11
12
“The Board would be derelict in its statutory responsibilities to recognize market
capitalization ratios that are derived from a market value capitalization that deviates
from the intrinsic long-run value of the regulated firm.”
13
This is the clearest possible statement by a regulator of the fundamental grounds for rejecting
14
ATWACC and its emphasis on market values.
15
Further the EUB went on to say
“In essence, a regulated company’s earnings are driven by the portion of the original
cost rate base deemed to be financed by common equity. This fact results in a
fundamental disconnect to the theory that market capitalization ratios, which have
deviated significantly from book capitalization ratios, reflect the appropriate financial
risk necessary to determine a fair composite return to be applied to the original cost rate
base of a pure play regulated utility. This is because the earnings of a pure play
regulated utility are governed by and driven by the regulated return allowed on book
equity. In other words, it is the book equity that reflects the appropriate financial risk
necessary to determine a fair composite return for a pure play regulated utility.”
16
17
18
19
20
21
22
23
24
25
26
This means that the correct financial risk measure is that which I discussed earlier under
27
financial leverage. It is also the approach pioneered by the National Energy Board, where
28
financial risk adjustments using the deemed common equity ratio are made for differences in
29
business risk.
30
The EUB went on to calculate an ATWACC using regulated book value capital structure
31
weights and the embedded debt costs. In this case (Decision U-99099, page 303)
- 82 -
1
2
3
4
5
An ATWACCBV would be suitable for a regulated utility whose profit, by legislation, is
limited to a fair return on the book value (i.e. original cost) of its assets. The Board
notes that an ATWACCBV is consistent with the logic of the traditional method of
determining fair return.
6
In our example, the ATWACCBV is the 5% debt and 11% equity cost weighted with the 50%
7
regulated capital structure weights. In this case the ATWACCBV is 8.0% and awarding this 8%
8
cost of capital means that the value of the equity is
V =
9
.08 * $10 − .25
.11
10
or $5mm. This is the exact same result that would arise if the firm were simply given its 11%
11
lower ROE.
12
The EUB ATWACCBV correctly recognised that apart from any estimation error attached to the
13
equity cost, the WACC approach is inconsistent with allowing a fair return to a regulated firm.
14
The only approach consistent with allowing “fair and reasonable” rates is to estimate the
15
regulated, “comparable,” sample of firm’s ATWACC using book value weights and embedded
16
debt costs. Needless to say in Dr. Vilbert’s estimate of WACC he uses market value weights
17
and implicitly derives higher estimates than are consistent with efficient regulation. If Dr.
18
Vilbert had correctly used book value weights the exercise comes down to the normal problem
19
of whether or not the estimated equity cost is accurate or not. In this case his voluminous
20
testimony would collapse to significantly fewer pages focussing on standard DCF and CAPM
21
estimates. I will not enter testimony on this question, since it is not part of the hearing, but
22
suffice it to say that I judge the estimates of Dr. Vilbert as being high. Although in all fairness
23
to Dr. Vilbert his underlying estimates, as I showed earlier, are not as high as commonly
24
presented by witnesses on behalf of the company.
25
The final step in the estimation process of Dr. Kolbe is to adjust for differences in the financial
26
leverage between the calculated WACC estimates and the firm in question. That is, given the
27
use of market value weights in the calculated WACC, which in my example were 57.7%
- 83 -
1
common equity and 42.3% debt, are leverage adjustments needed to apply the estimates to the
2
regulated book equity, which in my example was 50%? As the EUB noted a Board should not
3
take the market value weights into account and should ignore the WACC, but what of the
4
equity cost estimates themselves?
- 84 -
1
6.0
FINANCIAL LEVERAGE ADJUSTMENTS
2
Q.
WHAT IS THE BASIS FOR MAKING A LEVERAGE ADJUSTMENT?
3
A.
One basic proposition in finance is that investors don’t like risk, and as I showed in
4
Section 2.0, increasing the amount of debt financing magnifies risk. Logically, therefore as
5
firms finance with more debt they magnify business risk and investors respond to this increased
6
risk by requiring a higher rate of return. This logic is unassailable and not in dispute. Therefore,
7
if an expert estimated an equity cost from a sample of regulated firms and applied that estimate
8
to a firm with the same business risk, but much higher financial risk as represented by the
9
regulated or book debt equity ratio, then the estimated return would be below a fair return. For
10
example, estimating a fair return from a sample of normal utilities that happened to have no
11
debt at all and applying that to Union Gas with 61.5% debt would be patently unfair.
12
However, note two things. First of all leverage adjustments have nothing to do with ATWACC.
13
Leverage adjustments are theoretically necessary in the normal estimation approach where the
14
equity cost is estimated from a sample of firms and then a recommendation made for the firm
15
in question. Drs. Kolbe and Vilbert could, for example, take their equity costs from their
16
sample and make leverage adjustments without going through the trouble of estimating their
17
ATWACC. Second, in the above example there is obviously a problem, since it is extremely
18
difficult to find a sample of regulated firms with substantially different debt ratios in the first
19
place. In reality most regulated firms in Canada have similar debt ratios, or their debt ratios
20
have been specifically set to offset differences in business risk.
21
To elaborate on this last point, as late as August 1997 in testimony before the CRTC Dr.
22
Berkowitz and I were using two samples of regulated firms for estimating the fair return. The
23
first sample consisted of six energy distribution companies and the second six telcos. However,
24
the CRTC had increased the business risk of the telcos by opening their long distance markets
25
to competition and there was no doubt that their business risk was higher than that of the
26
energy distribution companies. Normally this would have caused comparability problems.
27
However, the CRTC also allowed their common equity ratios to increase to 55% to compensate
- 85 -
1
for the increased business risk. Consequently at that time we judged the overall risk of the six
2
telco sample to be useful in comparisons with energy distribution companies. Explicitly it was
3
our judgment that no leverage adjustments were needed going from a Telco sample with 55%
4
common equity, that is, 45% debt to an energy distribution sample with much greater financial
5
leverage.
6
By and large this continues to be my judgment: that the actions of regulators, like this Board, to
7
equalise risk obviates the need for leverage adjustments. In fact, in the recent Alberta generic
8
hearing the EUB specifically followed the lead of the National Energy Board and set common
9
equity ratios for a large sample of ROE regulated companies such that they could all earn the
10
same formula allowed ROE.
11
Q.
WHAT IF THERE ARE SIGNIFICANT LEVERAGE DIFFERENCES?
12
A.
The first question the Board has to ask is: are these leverage differences real, that is,
13
were they set to equalize overall risk or not? The second question the Board has to ask is: are
14
the leverage differences based on market or regulated book weights? If the answer to the first
15
question is that the leverage differences just offset business risk differences, then no action is
16
needed. If the answer to the second is the differences are only due to “temporary” market value
17
differences then they should also be ignored.
18
To continue with the previous example where the equity cost dropped from 15% to 11% and as
19
a result the equity market value increased and the equity ratio at market values increased to
20
57.7% from the regulated 50%. Suppose this were the sample average from say twenty
21
companies and the results had to be applied to a non-traded regulated firm with 50% common
22
equity and the same business risk as the sample. Dr. Kolbe would seem to argue that the
23
sample average has less financial risk and that to apply the estimated equity cost, assuming it is
24
accurate, to the regulated firm in question underestimates its fair ROE, since it has a 50% debt
25
not 42.3%. As a result, he would increase the recommended ROE from the 11% estimated from
26
the sample or conversely recommend a higher common equity ratio. I will show later that their
27
leverage adjustment gives the highest plausible leverage adjustment. However, the approach
- 86 -
1
itself is wrong for two reasons.
2
First, if the regulated firm is earning approximately the same allowed ROE and has the same
3
capital structure, there is no reason to believe that its implicit market valued equity ratio is any
4
different from that of the sample. That is, the non-traded regulated firm in all likelihood has an
5
implicit market valued debt ratio the same as those of the sample firms, so there is no leverage
6
difference that needs to be adjusted for. Further allowing a higher ROE just increases the
7
market value of the equity, causing the market valued debt ratio to drop further, creating an
8
even bigger internal contradiction.
9
Second it is important to remember that in the example the market valued debt ratio fell not
10
because the firm substituted equity for debt and reduced the amount of fixed interest payments
11
and financial risk, but because the equity cost fell. This is the reality of the declining interest
12
rate scenario we have lived with since 1981 and the effect of a persistent decline in equity costs
13
coupled with regulatory lag. However, there is a big difference between the impact of
14
substituting equity for debt in a regulated capital structure and an increase in equity value as a
15
cause of the fall in the market valued debt ratio.
16
In the example, the equity is obviously riskier at $6.818mm and a market to book of 1.36X
17
than it is at $5mm and a market to book of 1.0X. This is because it is highly unlikely that the
18
regulator will cut the allowed ROE when the stock is trading at book value. In contrast, the
19
higher the market to book ratio the more likely the regulator will cut the allowed ROE and thus
20
the riskier the stock. As a result when equity costs are estimated from a sample of firms with
21
higher market to book ratios the estimates should be reduced when applied to the regulated
22
firm to remove this “capital gains” risk premium. However, in stark contrast Drs Kobe and
23
Vilbert would have us believe that the equity in a regulated firm is less risky when it is trading
24
at a market to book well above 1.0, since the debt ratio is lower. As a result they would
25
increase the equity cost when applied to the regulated firm. I don’t accept this since it defies
26
common sense.
27
Second and more significant, the financial leverage risk premium stems from the imposition of
- 87 -
1
fixed interest charges. That is, prior to receiving their equity return the firm has to pay these
2
interest charges. This risk does not change as the market value of the firm changes; it only
3
changes as the book debt equity ratio changes. For example, if Union Gas moved to a 60-40
4
debt equity ratio in terms of book amounts, then there would be less interest expense.
5
Consequently, the financial risk, both to the bond-holder and the stock-holder, would decline
6
and with them both the debt and equity costs. In this case, a leverage adjustment would indicate
7
a lower equity cost, since the financial risk has declined. As Standard and Poors have stated,
8
9
10
11
12
“Similarly ratios using market value of a company’s equity in calculations of leverage
are given limited weight as analytical tools. The stock market emphasises growth
prospects and has a short time horizon; it is influenced by changes in alternative
investment opportunities and can be very volatile. A company’s ability to service its
debt is not affected directly by such factors (italics added).”
13
That is, S&P is basically saying book value leverage is important, when it is assessing the
14
default or credit risk in debt, whereas market values don’t count, or at least don’t count as
15
much. If it is book values and interest payments that affect credit risk and the cost of debt then
16
this is the risk that also affects utility equity investors.
17
Following on from the Alberta EUB’s decision that to accept market value weights would be a
18
“dereliction” of duty, the obvious implication is that the weights in the sample WACC should
19
also be book value weights. In my example this means that the regulated book value of 50%,
20
rather than the market value debt ratio of 42.3% is what matters. Hence in comparing this 50%
21
regulated debt ratio with the firm in hand that also has 50% debt ratio means that no adjustment
22
is necessary. Making an adjustment based on market values is then inappropriate for a
23
regulated firm. As the EUB again noted (Decision U99099, page 301)
24
25
26
27
28
“the Board considers that beta and the cost of equity do not change to the extent
necessary for an ATWACC, determined from market capitalization weights, to remain
constant when applied to the book capitalization for a pure play regulated utility. The
increase required to the cost of equity to achieve a constant ATWACC would be
excessive and violate the fair return standard.”
29
It is my judgment that the only time a leverage adjustment is needed is either when a firm’s risk
30
differs from that of a sample of regulated firms from which an ROE estimate is derived, or
- 88 -
1
when its business risk has changed and the Board wants offset this change so it can continue to
2
award a formula allowed ROE. In these cases, as I indicated earlier, in my judgment the
3
literature has not reached “closure” on how to make leverage adjustments and the Board is best
4
advised to base its decision on the business risk of the firm and its access to financial markets.
5
These are the factors discussed in Sections 2.0, 3.0 and 4.0..
6
7
8
Q.
SUPPOSE THE BOARD FEELS THAT A LEVERAGE ADJUSTMENT BASED
ON MARKET VALUES IS NECESSARY DO YOU AGREE WITH
DR.KOLBE’S ADJUSTMENT METHOD?
9
A.
No. Again, it is well accepted that financial risk magnifies business risk. The basic
10
financial leverage equation indicates that the accounting return to the stockholder is determined
11
as follows
ROE = ROI + ( ROI − Rd ) D
12
S
(2)
13
where these are all book values, that is, D and S are the book values of debt and equity and Rd
14
is the embedded cost of debt. The equation simply comes from manipulating the firm’s
15
financial statements. It means, for example, that with a fixed stock of assets, as revenues and
16
the ROI changes, then the greater the amount of debt the greater is the variation in the
17
accounting return to the stock holders. However, the above equation says absolutely nothing
18
about how the stock market reacts to this financial risk, that is, how market values change, or
19
how the cost of equity changes as the firm uses debt.
20
To understand how the investor’s required rate of return or equity cost varies with the use of
21
debt we need a valuation model. The first valuation attempt was by Franco Modigliani and
22
Merton Miller (M&M) who in 1958 developed an arbitrage model to show that the total
23
enterprise value was independent of the use of debt. This was their famous “no magic in debt
24
argument.” If individuals can borrow on the same terms as the firm, then investors will not pay
25
a premium for firms that use debt, since the firm is not adding value. Consequently, they
26
derived the following formula
- 89 -
Ke = K0 + (K0 − Kb ) B
1
E
(10)
2
where the K’s indicate the cost of equity and debt, not accounting returns, and B and E
3
represent the market values of debt and equity respectively. The subscript 0 then indicates what
4
the equity cost would be if the firm had no debt outstanding, which is often referred to as the
5
unlevered equity cost.
6
Note two things about this equation. First, apart from redefining returns and debt ratios, in form
7
it is the same as the leverage equation I used earlier. This is because in the accounting model
8
total assets are fixed, whereas in this valuation model M&M “proved” that the value of the firm
9
was fixed. As a result, changes in the book and market debt ratios have the same impact.
10
Second M&M “proved” that as the market value was constant the weighted average cost of
11
capital was also constant, which in this case means that it is equal to the unlevered equity cost.
12
However, note that I italicised “proved,” since this was a mathematical proof that followed
13
from their assumptions, not a description of reality.
14
In the M&M equation changes in the market valued debt equity ratio (B/E) are multiplied by
15
the spread between the WACC and the cost of debt. It is this coefficient that determines how
16
much changes in the debt equity ratio affect the equity cost since it is this coefficient that
17
determines the risk. This is the important point: people who believe that changes in the debt
18
equity ratio have a big impact on the equity cost believe that the coefficient on the market
19
valued debt equity ratio is high and vice versa.
20
However, the overall market value in the M&M model is only fixed by their assumptions. To
21
emphasise, remember that from equation (9)
22
V =
After − tax operating income
WACC
23
The total firm value is after tax operating income divided by the after tax WACC. Given that
24
M&M were discussing capital structure not operating changes, the after tax operating income,
25
the numerator above, is by definition constant. What M&M “proved” was that with firm value
- 90 -
1
constant the WACC must also be constant. In this case, given that the WACC is a weighted
2
average of the debt and equity costs, the equity cost has to increase with more debt to offset the
3
impact of more “cheaper” debt. This is what equation (10) indicates.
4
However, if the market value increases with more debt then from equation (9) the cost of
5
capital will decrease and vice versa. In this case, the equity cost may then increase or possibly
6
even decrease with the use of debt, it all depends on the valuation model and the assumptions
7
that are made. The critical question is how the use of debt affects the overall firm value; the
8
impact on the WACC and the equity cost then follow directly.
9
M&M’s “no magic in debt” result was controversial in 1958 and remains so today. This is
10
11
12
13
14
15
16
17
18
19
20
because of the assumptions required to “prove” their result. The most important are that:
•
•
•
•
•
•
•
there are no taxes of any kind;
there are no transactions costs;
there are no information asymmetries between borrowers and lenders;
everyone can borrow on the same terms, that is, if the company can issue 25
year bonds or access the swap market, then so too can other individuals;
all firms are perpetuities that pay out 100% dividends;
there are no bankruptcy or financial distress costs;
there are two or more identical firms with different levels of debt that can be
arbitraged.
21
All of these assumptions have been disputed to a greater or lesser extent and if any of them are
22
incorrect then the total value of the firm is affected by the use of debt. Hence, so too is the cost
23
of capital.
24
M&M’s result is a classic in corporate finance and they won the Noble prize in economics for
25
it. However, its great strength lies not in its result, which few accepted then or now, but the fact
26
they focused corporate finance on the implications of their assumptions. For example, in 1963
27
they recognised that they made a mistake in their treatment of corporate income taxes and
28
corrected their original paper. They then showed that, all else constant, the value of the firm
29
increases due to the tax shield generated by the tax deductibility of interest payments. The
- 91 -
1
reason is simply that what we term value is the private value and by reducing corporate income
2
taxes the private value of the firm increases at the expense of the government. Hence from
3
equation (9), if the private market value increases the WACC of necessity must decline.
4
In fact in the M&M (1963) model the WACC declines continuously since the corporation can
5
issue risk free debt and the average and marginal tax advantage to debt are the same. In this
6
case, the equity cost changes in the following way with the use of debt,
7
K e = K 0 + (1 − T )( K 0 − K b ) B
E
(11)
8
There is still a financial leverage risk premium but it is now smaller, since the use of debt also
9
generates a valuable tax shield. Note that in M&M (1963) changes in the market valued debt
10
equity ratio are now multiplied by (1-T), so are smaller than in M&M (1958). Thus assuming a
11
40% corporate tax rate, people who believe in M&M (1963) would estimate a leverage impact
12
only 60% the size of those who believe in M&M (1958).
13
Since 1963 all the other assumptions of M&M have been relaxed and every time an assumption
14
has been relaxed there is another leverage equation similar to equations (10) and (11) and
15
another estimate of the leverage effect. However, two main theories of capital structure have
16
emerged: the static trade off (STO) model and the pecking order hypothesis (POH). The STO is
17
a static model that assumes that firms trade off the tax advantages of using debt against the loss
18
of financial flexibility that arises due to excessive leverage. It is this model that develops the
19
familiar “U” shaped WACC function below as the firm increases its debt ratio.
- 92 -
1
The U shaped WACC
WACC
2
3
4
5
Significant tax
advantages
6
Loss of financial
flexibility severe
risk of distress
7
8
Initially the WACC declines due to the tax advantages of debt. In the M&M (1963) model, for
9
example each dollar of debt increases the firm’s market value by the value of the corporate tax
10
rate,25 the WACC then starts to increase as the loss of financial flexibility sets in. Obviously
11
there has to be some offset to the tax deductibility of interest, otherwise all firms would try to
12
finance with 100% debt. The offset comes as the debt becomes riskier and has to be sold on
13
higher and higher yields and the firm loses its financial flexibility.
14
In contrast, the POH, developed in 1963 by Gordon Donaldson at Harvard, is a dynamic model
15
of financing based on the fact that firms are controlled by managers. In this case, firms raise
16
capital by issuing securities that impose the least restrictions on management. Consequently,
17
firms primarily rely on internal funds and only after these are exhausted do they go outside for
18
capital, where then they initially rely on bank debt and bonds, rather than new equity.
19
I have reviewed these basic ideas on capital structure since the flat ATWACC approach of Drs,
20
Kolbe and Vilbert is essentially the 1958 M&M model as extended to include corporate and
25
This simple model has been dubbed adjusted present value (APV) by Professor Myers. In Principles
of Corporate Finance (2nd Canadian edition, 1991 pages 490-493 they work an example and the base
case NPV of $170,000 is then increased by $592,000 by the tax advantages to debt. In this case,
Professor Myers, who Dr. Kolbe references throughout his testimony, clearly believes in the tax
advantages of debt.
- 93 -
1
personal taxes by Miller (1977). This is a very important model and for the last 26 years I have
2
taught corporate financing to second year MBAs with the first five weeks devoted almost
3
exclusively to these ideas,26 as well as to the implication that if this model holds there is no
4
value to the activities of investment bankers and they should all study marketing! I then spend
5
the balance of my course explaining how companies add value by adopting different financing
6
decisions. The fact is that financial theory has come a long way since 1958 and is now better
7
harmonised with practise: no one believes the flat WACC model fits reality; it is simply a good
8
starting point to discuss how investment bankers can create value for firms.27
9
However, a flat ATWACC does have the advantage that it gives the largest possible leverage
10
effect, that is, the coefficient on the market valued debt equity ratio in the equity cost equation
11
is as large as possible. I showed earlier that the M&M 1958 flat WACC model gives a bigger
12
equity cost adjustment (equation (10)) than if the WACC declines with leverage in the
13
conventional way (equation 11). However, assuming a flat ATWACC in the presence of
14
corporate taxes gives an even bigger coefficient on the market valued debt equity ratio.
15
To illustrate assume a flat, that is, constant ATWACC.28 Dr. Vilbert first calculates the WACC
16
using market value weights from his sample:29
Ke
17
E
B
+ K b (1 − T ) = WACC = K A
V
V
18
Dr. Kolbe then assumes that the WACC (KA) is constant and then either alters the equity ratio to
19
get a new equity cost or alters the equity cost to get a new equity ratio, holding everything else
26
This is MGT2300. A course outline can be downloaded from my web page at
http://www.rotman.utoronto.ca/~booth
27
It would be interesting to ask why investment bankers are so well paid if corporate financing
decisions as represented by a flat ATWACC have no value and firms can do whatever they want.
28
Note that a flat ATWACC requires in part that personal taxation offsets the corporate tax shield, yet
in J2-07 Union Gas indicated that it has never commissioned a study of its marginal investor’s tax rate
and uses the conventional after tax WACC in capital budgeting. It also believes its marginal investor is
Canadian even though its parent is American?
29
Note that as explained earlier the use of market values is not appropriate for regulated firms, either
directly or indirectly through WACC estimates from samples of regulated firms.
- 94 -
1
constant. In terms of the equity cost, implicitly Dr. Kolbe is rearranging this WACC equation to
2
solve for the equity cost (Ke) at any leverage ratio,
K e = K A + ( K A − (1 − T ) K b ) B
3
E
(12)
4
If the WACC is assumed constant it has the same no leverage equity cost (K0) as before, the
5
only difference is that it is this cost minus the after tax cost of debt that determines the leverage
6
coefficient. With a constant WACC this coefficient is larger than either the M&M (1958) no
7
tax case or the M&M (1963) tax case as a simple comparison with equations (10) and (11)
8
indicates.
9
The reason for the very large leverage adjustment in equation (12) is that the model is
10
internally inconsistent. Equation (12) and the flat WACC assumes the tax deductibility of
11
interest which causes the WACC to fall, but there is no explicit account of the offsetting costs
12
that negate this to keep the WACC constant. For example, if the WACC is constant it could be
13
that as the market valued debt equity ratio increases the debt cost also increases due to the
14
higher risk of insolvency and the costs of financial distress and bankruptcy. This would be
15
particularly true as the firm goes to very high debt equity ratios. In this case, what is keeping
16
the WACC constant is an increasing Kb as creditors protect themselves from the insolvency risk
17
attached to highly debt financed firms. From the spread date in Schedule 19, we know this
18
happens. Moreover, it is obvious from equation (12) that if the debt cost, Kb, increases with the
19
debt equity ratio then the equity cost does not increase so fast. In this case even with a flat
20
ATWACC the equity cost increase with leverage is lower than by assuming a constant debt
21
cost. Solomon showed this in the Journal of Finance in 196330 and it is also graphed on page
22
433 of Dr. Myers textbook. The intuition is simply that “debt” in highly debt financed firms has
23
some of the same characteristics as equity.
30
The only reason for the cost of debt to increase is the risk of financial distress or bankruptcy, which
M&M ignored in their 1958 paper. Therefore, Solomon’s result is inconsistent with the M&M
assumptions. However, it is consistent with a model of bankruptcy and financial distress. Note the
- 95 -
1
To show these principles backtrack to the previous example, where the equity cost was
2
assumed to decrease from 15% to 11% due to a reduction in risk and consequently the equity
3
market value increases from $5mm to $6.818mm. As a result, the market valued debt ratio
4
decreases from 50% to 42.3%, simply because the equity value has increased due to regulatory
5
lag. Suppose that the equity cost is then accurately estimated at 11.0%, but that someone
6
believes that a leverage adjustment is needed to apply this to a firm with 50% debt; how could
7
this be done?
8
One way is to estimate an unlevered equity cost from equation (10) by inserting the debt cost of
9
5% the debt equity ratio of .423/.577 and the equity cost of 11%. In this case, the unlevered
10
equity cost is 8.46% and the use of debt financing has increased the equity cost from the debt
11
free 8.46% to the observed 11.0%, so 2.54% is the financial leverage risk premium. The
12
coefficient on the market valued debt equity ratio in this example is 3.46% (8.46-5.0). The
13
relevered equity cost at the 50:50 debt equity ratio would then be 11.92%. So someone
14
believing in M&M (1958) would use a coefficient on the debt equity ratio of 3.46%. Further if
15
they believed that the equity cost estimated from a sample of firms with lower market valued
16
debt ratios underestimated the financial risk at the regulated firm’s debt ratio, they would
17
increase the 11.0% by 92 basis points.
18
If instead the M&M (1963) with taxes equation (11) is used with a 50% tax rate, the unlevered
19
equity cost is higher at 9.39% and the financial leverage risk premium is only 1.61%. Since the
20
risk impact of financial leverage is offset in part by the tax advantages attached to debt, the
21
financial leverage risk premium is only half what it is with the flat WACC M&M 1958 model.
22
In this case the coefficient on the market valued debt equity ratio is 1.7% ((8.46-5.0)*.5).
23
Relevering to the 50% debt ratio increases the equity cost to 11.56% or 36 basis points less
24
than by using the flat WACC M&M 1958 model. Believing in M&M (1963) gives a smaller
25
bump to the ROE estimate.
- 96 -
1
Believing in a flat WACC gives a WACC and unlevered equity cost of a constant 7.4%.31 Hence
2
the market valued debt equity ratio is multiplied by (7.4-2.5) or approximately 5.0%. This is
3
higher than either M&M (1958) no tax or M&M (1963) with tax and gives the highest leverage
4
adjustment. This is because the debt cost is after tax and there are no explicit offsetting costs in
5
the model, yet the WACC is somehow held constant. Using this model the leverage adjustment
6
would not be 36 or 92 basis points but 131 basis points to move the equity cost at the regulated
7
debt ratio to 12.31%. If the debt cost has also increased due to the higher financial risk
8
consistent with a constant ATWACC then this 1.31% over-estimate of the equity cost is also
9
too high even under the constant ATWACC assumptions.
10
Let me make the importance of this example clear. The chain of events is that the risk of the
11
utility has declined causing its equity cost to drop from 15% to 11%. The obvious thing that the
12
regulator should do is cut the allowed ROE from 15% to 11%. This is also what would happen
13
if the regulator used the EUB’s ATWACCBV approach and recognised that it would be
14
“derelict” in using market values to rubber stamp this increase in market value. However, using
15
the “(AT)WACC approach” avoids this drop in the ROE in two ways. The first is to go directly
16
to the WACC with market values, which seals in the higher equity ratio and delays the drop in
17
the allowed ROE. However, if this fails, as it has before the EUB, the second step is to argue
18
for a leverage adjustment. Then the assumption of a flat ATWACC generates the biggest
19
coefficient on the debt equity ratio and the largest financial leverage risk premium. This in turn
20
provides the biggest “bump” when a sample estimate is applied to the regulated common equity
21
ratio. In my example it would give an equity cost of 12.31%, 131 basis points higher than the
22
true equity cost and higher than the other equity cost models as well as an internally consistent
23
flat ATWACC model. As the example shows these assumed leverage adjustments can be very
24
large and they are totally unnecessary.
25
26
31
7.4%= 11%*0.577 + 5%(1-.5)*0.423
- 97 -
1
2
Q.
DRS. KOLBE AND VILBERT ARE NOT MAKING ROE ADJUSTMENTS DO
THESE CONCLUSIONS STILL APPLY?
3
A.
Yes. Both the equity cost and deemed common equity ratio results flow from a
4
rearrangement of the same equations; it is simply easier to “see” them using the equity cost
5
approach since that is what most people are interested in. Note that Dr.Vilbert’s equation (1)
6
(E2, Tab 3, page 13) explicitly sets the sample ATWACC as constant and solves for the
7
deemed equity ratio that gives the same ATWACC with the market costs of equity and debt for
8
the regulated firm. This process implicitly assumes a constant ATWACC, otherwise the
9
ATWACC would have to vary with the change in the deemed common equity ratio and that the
10
sample ATWACC is the same as that for the regulated firm.
11
As a result the critical equation is
12
WACC = ROE
S
D
P
+ K d (1 − T ) + K p
A
A
A
(13)
13
where I have been consistent with my own notation and use S, D and P for the book values of
14
equity, debt and preferred shares respectively and A is total assets. The left hand side is the
15
sample average ATWACC estimated using market values and the right hand side is the same
16
for the regulated firm except ROE is the fixed allowed ROE leaving only the deemed equity
17
ratio to vary to set the right hand side equal to the left hand side.
18
If we solve for the deemed equity ratio noting that the debt ratio is one minus the equity ratio,
19
we get
20
P
P
S WACC − K p A − (1 − A) K d (1 − T )
=
A
ROE − K d (1 − T )
21
Dr Vilbert agreed (interrogatory response J2.17) that this was the equation he uses to estimate
22
the deemed common equity ratio for Union..
- 98 -
1
To illustrate how this equation is used let’s take Dr. Vilbert’s CAPM estimated WACC of
2
5.70%, an allowed ROE of 8.89% and the cost rates used by Ms. Elliott, that is, 5.0% for the
3
debt and 4.9% for preferred shares with a preferred share financing ratio of 3.5%. If we plug in
4
these values with the 36.1% marginal tax rate used by Dr. Vilbert we get
5
S 5.7 − 0.1715 − 3.08
= 43%
=
A
8.89 − 3.08
6
or a deemed common equity ratio of 43.0%. Dr. Vilbert confirmed these types of calculations
7
in interrogatory response J2.17 with the proviso that he assumed that preferred share dividends
8
were tax deductible and corrected his testimony. However, before using this as a basis for a
9
deemed common equity ratio consider how sensitive it is to various parameters.
10
First consider the estimate of the WACC. As I indicated before, Dr. Vilbert’s testimony is
11
essentially rate of return testimony since estimates of the equity cost are needed to derive the
12
WACC for his sample. Suffice it to say that I disagree with many of his procedures and just
13
suppose that the true sample WACC using more reasonable values is 5.2%, how does this
14
affect the deemed equity ratio? Plugging 5.2% into the above gives a deemed equity ratio of
15
34%, which is less than Union’s currently allowed common equity ratio. Small differences in
16
the WACC estimate give significantly different estimates of the deemed common equity ratio.
17
Suppose instead that Union decides to finance its rate base with longer term debt and instead of
18
a 0.65% premium over the long Canada rate it costs 1.05%. Plugging 5.4% into the equation
19
with the 5.7% WACC gives a deemed common equity ratio of 40%. If Union finances with
20
more longer term debt at a higher cost this means that less common equity is needed to reach
21
the 5.7% WACC. Strange as it seems there is an incentive for Union to lower its debt costs
22
with the WACC approach, since it then has a higher common equity component. For example,
23
if it finances as I have suggested with more short term debt and lowers its debt cost to say 4.5%
24
then its deemed common equity ratio increases to 45.0%. If Union uses more longer term debt
25
at a cost of 5.40% and the true WACC is 5.2% then Union’s deemed common equity ratio falls
26
to 30.57%. The fact is that the WACC approach is very sensitive to the chosen parameters.
- 99 -
1
Of even more importance is that the deemed common equity ratio is derived from the same
2
assumption of a flat ATWACC that I used to derive the equity cost equation with the highest
3
leverage effect. That is, it is the same assumption of a constant WACC that Dr. Kolbe uses to
4
make his equity ratio recommendation that also produces the very high equity cost leverage
5
adjustment. Note that in the example it produced an extra 131 basis points above the assumed
6
fair 11.0% ROE. Further it is the same assumption that the EUB criticized in Decision U99099
7
(page 301)
“the Board considers that beta and the cost of equity do not change to the extent
necessary for an ATWACC, determined from market capitalization weights, to remain
constant when applied to the book capitalization for a pure play regulated utility. The
increase required to the cost of equity to achieve a constant ATWACC would be
excessive and violate the fair return standard.”
8
9
10
11
12
13
Consequently since Dr. Kolbe’s common equity ratio recommendation is derived from the
14
same assumption and equations as his equity cost recommendations the comments of the
15
Alberta EUB also apply here. To accept this approach would in the words of the Alberta EUB
16
be a “dereliction” of duty on the part of the regulator.
17
18
Q.
HAVE YOU ANY OTHER COMMENTS ON THE USE OF MARKET
VALUES?
19
A.
As I have stressed in the financial flexibility discussion, in the final analysis “fair” is
20
determined in the stock market by the reaction of investors. If Board policies were not “fair”
21
we would see several things. First of all I would expect to see holding companies “ring
22
fencing” their regulated operations and selling parts to the stock market. For example, if Duke
23
Energy believes that the Union Gas common equity ratio produced results that were not fair, I
24
would recommend that it sell say 20% to the stock market to establish a public float. If the
25
stock market agreed with this unfair assessment we would then see a market to book ratio
26
below 1.0 and that would provide powerful ammunition to support a higher financial
27
parameters. However, I have seen very few ROE regulated utilities establishing public floats in
28
fact the reverse has happened as Consumers Gas, Island Tel, Maritime Electric etc have all
29
ceased to exist as public entities. The elimination of publicly traded pure regulated utilities is a
- 100 -
1
telling sign that the allowed financial parameters are too generous, since under stand alone
2
regulation there should be few synergies with other corporate entities.
3
Of importance is that regulators can not see the market to book ratio and observe how the
4
market values the financial parameters awarded regulated utilities. Observing the market to
5
book ratio provides a valid way of assessing how investors reacted to allowed ROEs. Dr. Kolbe
6
makes essentially the same point,32 when he stated (page 27)
“
7
(
MV
ROR
)=(
) . (2.3)
BV
r
8
9
10
11
12
13
Equation (2.3) of course reflects strong simplifying assumptions. But the qualitative
conclusions we draw from it hold in most cases. If regulators allow the firm to expect to
earn its cost of capital, market value will equal book value (ROR=r implies ROR/r –
MV/BV =1, so that MV = BV). Conversely, if we observe MV=BV, we conclude that
investors expect regulators to allow the firm to earn its cost of capital, at least on
average. (MV=BV implies ROR/r =1, so that ROR=r).”
14
At the time that Dr. Kolbe’s book was published in 1984 the market to book ratios of regulated
15
firms in the US (his graph on page 32 of the text) were below 1.0. As a result, this observation
16
worked to increase allowed ROEs. Since 1983 market to books for regulated firms have
17
generally been significantly above 1.0 as the graph in Schedule 1 indicates. By the mid 1990s
18
the sample average market to book ratio was well above any cushion allowed for flotation
19
costs, so Dr. Kolbe’s logic holds in reverse that allowed were too high.
20
The market to book ratios in Schedule 1 include to a differing degree the impact of non-
21
regulated operations, but there is a clear indication that none of these companies have suffered
22
any loss of financial flexibility as regulation has evolved over the last ten years, since there is
23
no obvious trend in the market to book ratios except that regulation in Canada has been
24
extremely favourable towards regulated utilities.
32
L. Kolbe, J. Read and G. Hall, The Cost of Capital, Charles River Associates, MIT Press, 1984.
- 101 -
1
As a final comment we have direct evidence of the value of regulated assets from sales of these
2
assets between firms. Unlike the UHC data in Schedule 1 these observations are not
3
contaminated by non-regulated assets. For example,
4
5
6
7
8
•
TCPL purchased the 50% of Foothills that it did not own at a market to book of
1.6 based on the common equity. Foothills, of course, accesses the WCSB, earns
the NEB formula and has 30% common equity. Moreover since TCPL already
owned 50% of Foothills the number of potential buyers was limited, which
reduced the price.
9
10
•
Aquila purchased TransAlta’s distribution and retail business at a market to
book of 1.5 based on a total rate base of $472mm (premium of $238mm);
11
12
•
Fortis purchased Aquila’s Alberta interests for a premium of $215mm over a
rate base of $601mm.
13
14
•
AltaLink purchased TransAlta’s transmission business for a $200mm premium
over a rate base of $644mm.
15
16
•
Hydro Quebec recently announced a $266mm gain on the sale of its interest in
the gas distribution assets in the province.
17
Note that in all these cases, the market to book ratio based on the equity is much greater than
18
that based on the total rate business, since the debt is normally assumed and is valued at close
19
to its book value. For example in Fortis’ purchases from Aquila it paid $1.3 billion for total rate
20
base assets of $943mm (in Alberta and elsewhere) for an overall premium of $357mm over rate
21
base and an overall market to book of 1.38X. However, it assumed the existing debt which was
22
60% of rate base so effectively Fortis assumed about $565.8mm in debt and paid $734.2mm for
23
the book equity, so the market to book based on equity was about 1.96X. The final transaction
24
value depends on closing transactions but the point is that the market to book based on the
25
common equity was well above the indicated values based on total rate base. The most
26
egregious case of excessive valuation has recently occurred with Terasen Inc which was
27
purchased by Kinder Morgan Inc at a market to book ratio of 2.7 despite the fact that the vast
28
bulk of Terasen Inc’s earnings come from ROE regulated businesses.
29
Overall these observations on market to books are a significant indication that regulated assets
30
in Canada are worth well above their regulated book values. This observation means that $1 in
- 102 -
1
equity reinvested in rate base is immediately worth much more almost $2 as in the case of
2
Fortis’s purchase of Aquila’s assets and probably more for KMI’s purchase of Terasen. Further
3
as Dr. Kolbe points out, and consistent with basic financial theory, this means that utility
4
allowed ROEs and common equity ratios in Canada are excessive and rates include equity
5
charges that are more than fair and reasonable. I therefore, see no reason to add leverage or
6
equity ratio adjustments that further compound these already generous (excessive) charges.
7
Q.
CAN YOU SUMMARISE YOUR TESTIMONY?
8
A.
Yes.:
•
9
The short term business risk of Union is very low as it continues to earn its
10
allowed ROE. There is no indication that the impact of the Board’s policy of
11
testing performance based regulation has exposed Union’s shareholder to any
12
increase in risk.
•
13
In my judgment there has not been a significant change in Union’s business risk
14
since RP-2003-0063/87/97 when Union requested and was granted a 35%
15
common equity ratio in the Board’s decision dated March 18, 2004.33 In
16
particular the BC Utilities Commission has just set Terasen Gas’s common
17
equity ratio at 35% and I see no significant differences in business risk between
18
these two companies.
•
19
Overall I would recommend that Union continue to be allowed a 35% common
20
equity ratio. I would judge Union in isolation to have a good investment grade
21
bond rating with its current allowed ROE and common equity ratio. Its DBRS
22
rating of A has been stable for many years and the cut in its S&P rating to BBB
23
is due to its ownership structure and reflects Duke Energy’s rating not that of
33
Union Gas was a given a little bump in EBRO499 when it’s common equity ratio was increased to
35% from 34% when it was consolidated with Centra Gas Ontario, which had a 36% common equity
ratio. A straight blended rate would have been 34.5%.
- 103 -
1
Union. Spreads on Union’s publicly traded debt indicate that it trades as a better
2
than BBB credit. Spreads on Canadian utility and pipeline debt over the last
3
several years reflect normal cyclical concerns and do not indicate that the
4
market has re-evaluated the regulatory protection accorded utility and pipeline
5
debt in Canada.
6
•
Notwithstanding the above, there is no guarantee that in the future Union’s debt
7
costs may not reflect its ownership structure and S&P BBB bond rating. In my
8
judgement the BCUC has taken the right steps in ensuring that Terasen Gas (BC
9
Gas Utility) be ring fenced on its indirect acquisition by Kinder Morgan Inc
10
(KMI). Ring fencing, or structural insulation as S&P refers to it, allows an
11
operating subsidiary to have a bond rating that reflects its risk rather than that of
12
its parent. If this is not done there is always the possibility that the company’s
13
cost of debt includes an “unfair and unreasonable” charge due to its risky parent.
14
•
In my judgement the most significant change in Union’s risk since 2004 has
15
occurred due to its ownership rather than its business risk. When the Board
16
agreed to Union’s requested 35% common equity ratio in its 2004 decision
17
Union had an A- S&P bond rating, now it is BBB. It is unfair and unreasonable
18
to ratepayers that Union’s common equity ratio be increased because of its
19
ownership structure.
20
•
In terms of the ATWACC approach used by the company’s witnesses I would
21
point out the fundamental contradiction in its use in regulatory filings in that it is
22
the mirror image of shareholder value maximisation. That is, earning more than
23
the WACC is synonymous with the creation of shareholder value, whereas the
24
Board’s responsibility is not to create or maintain shareholder value, but to
25
ensure that rates are fair and reasonable. The Alberta EUB felt it would be
26
“derelict” in its responsibilities to recognise market capitalisation ratios, an
27
assessment I agree with. In my judgment setting the equity ratio or ROE
28
implicitly by using the (AT)WACC approach can “rubberstamp” existing
- 104 -
1
market values that may in turn reflect unfair and unreasonable rates. I therefore
2
see absolutely no value to its introduction into a regulatory setting.
•
3
Leverage adjustments should be made when a Board sets both the allowed ROE
4
and the common equity ratio. In this way the Board makes sure that it does not
5
“double count” the impact of changes in business risk. For example, Union has
6
traditionally been regarded as riskier than Enbridge Gas Distribution Inc
7
(Consumers) with a premium over EGDI’s ROE of 15-25 basis points. Allowing
8
Union an additional 5% on its common equity ratio effectively increases this
9
premium ROE over EGDI.
•
10
My recommendation is to ignore the ATWACC approach entirely and continue
11
with best regulatory practise in Canada and set Union’s common equity ratio
12
based on its business risk and an assessment of its financial flexibility and
13
capital market access.
14
15
Q.
DOES THIS CONCLUDE YOUR TESTIMONY?
16
A.
Yes
- 105 -
Schedule 1
Market to Book Ratios for UHCs
3
2.5
2
1.5
1
0.5
0
1995
1996
1997
1998
1999
CUL
Fortis
GMI
Enbridge
TCPL
Average
Average excludes PNG
- 106 -
2000
PNG
2001
2002
Terasen
2003
TAU
2004
Schedule 2
EARNED ROE vs ALLOWED
1990
1991
1992
1993
1994
1995
1996
1997
1998
1999
2000
2001
2002
2003
ovrearn
Allowed
13.25
13.5
13.25
12.25
11.25
12.25
11.25
10.67
10.21
9.58
9.9
9.61
9.53
9.79
11.16
TCPL
Actual
13.34
13.65
13.43
12.31
11.16
12.56
11.83
11.15
10.63
9.64
9.99
10.01
9.95
10.18
11.42
0.25
Allowed
14.25
14.25
13.83
11.73
11.5
12.25
11.25
10.67
10.21
9.58
9.9
9.61
9.53
9.79
11.31
Foothills
Actual
14.25
14.25
13.83
11.73
11.5
12.25
11.25
10.67
10.21
9.58
9.9
9.61
9.53
9.79
11.31
0.00
TCPL BC (ANG)
Allowed
Actual
Allowed
13.25
13.25
13.75
13.38
13.38
13.75
13.43
13.43
13.75
12.08
12.08
12.25
12
12
12.25
12.25
12.25
12.25
11.25
11.25
11.25
10.67
10.67
10.67
10.21
10.21
10.21
9.58
9.58
9.58
9.9
9.9
9.9
9.61
6.86
9.61
9.53
9.53
9.53
9.79
8.21
9.79
11.21
10.90
11.32
-0.31
TQM
Actual
14.87
13.94
13.97
12.5
12.55
12.65
11.83
10.94
10.32
9.94
9.96
10.21
9.8
10.21
11.69
0.37
NEB Regulated pipelines controlled by TransCanada Corporation, confirmed by TCPL in CAPP 31(a) in RH-2-2004.
- 107 -
Schedule 3
Earned vs Allowed ROEs
1990
1991
1992
1993
1994
1995
1996
1997
1998
1999
2000
2001
2002
2003
2004
Average
Overearn
Allowed
13.25
13.13
13.13
12.30
11.60
11.65
11.88
11.50
10.30
9.51
9.73
9.54
9.66
11.32
EGDI
Actual
13.60
13.29
13.40
14.43
12.49
12.66
13.14
13.00
11.97
10.77
10.83
10.03
11.81
12.42
1.10
Allowed
13.50
13.50
13.00
12.50
11.75
11.75
11.75
11.00
10.44
9.61
9.95
9.95
9.95
9.95
9.62
11.21
UNION
Actual
13.40
12.50
13.70
14.30
12.14
12.12
12.52
12.26
11.14
10.10
10.11
11.45
12.36
12.08
10.45
12.04
0.83
Allowed
12.25
n/a
10.65
12.00
11.00
10.25
10.00
9.25
9.50
9.25
9.13
9.42
9.15
10.15
Terasen
Actual
9.06
11.91
9.73
12.03
11.80
11.27
9.41
10.70
10.75
9.38
10.03
10.23
9.46
10.44
0.29
Terasen data is from the company’s response to the BCUC information request #1 in the BCUC review of its adjustment mechanism. The data for EGDI
and Union is taken from Appendix B Schedule 10 of the pre-filed testimony of Dr. William Cannon in RP-2002-0158 updated with data from
interrogatory response J2-31.
- 108 -
Schedule 4
Earned Utility Holding Company (UHC) ROEs
CU Ltd
1993
1994
1995
1996
1997
1998
1999
2000
2001
2002
2003
2004
STDEV
13.37
13.71
14.12
14.86
14.87
14.75
14.54
15.44
14.96
17.56
13.71
15.19
1.09
Emera
12.02
11.90
11.55
10.59
10.56
9.47
10.83
10.88
10.58
6.65
9.77
9.80
1.43
Enbridge Fortis
17.53
9.59
16.91
14.47
14.04
13.25
13.35
15.65
14.90
10.11
17.31
16.43
2.60
11.84
10.71
10.74
9.61
9.43
7.16
8.56
9.71
12.25
12.24
12.28
11.25
1.66
PNG
GMI
19.29
19.73
19.50
19.91
18.91
19.11
17.66
17.93
17.45
18.91
18.05
18.21
0.83
- 109 -
12.92
13.44
11.77
13.32
13.32
10.14
10.79
9.75
7.50
5.94
7.59
6.97
2.75
Terasen
10.82
7.24
8.51
17.59
8.34
12.09
13.35
15.16
10.26
9.59
9.49
3.15
TransAlta TCPL
16.00
15.10
14.00
13.24
12.84
16.41
4.88
8.14
7.23
2.31
8.67
5.97
4.99
14.01
12.86
13.20
12.33
11.25
7.04
7.42
8.44
10.89
11.93
12.80
15.49
2.63
Mainline Foothills
12.31
11.16
12.56
11.83
11.15
10.63
9.64
9.99
10.01
9.95
10.18
11.73
11.50
12.25
11.25
10.67
10.21
9.58
9.90
9.61
9.53
9.79
1.05
1.00
Schedule 5
Variability in Earned UHC ROEs
1993
1994
1995
1996
1997
1998
1999
2000
2001
2002
2003
DEV
STDEV
CU Ltd
1.64
2.21
1.87
3.61
4.20
4.54
4.96
5.54
5.35
8.03
3.92
4.17
1.87
Emera
0.29
0.40
-0.70
-0.66
-0.11
-0.74
1.25
0.98
0.97
-2.88
-0.02
-0.11
1.16
Enbridge
5.80
-1.91
4.66
3.22
3.37
3.04
3.77
5.75
5.29
0.58
7.52
3.74
2.62
Fortis
0.11
-0.79
-1.51
-1.64
-1.24
-3.05
-1.02
-0.19
2.64
2.71
2.49
-0.14
1.95
GMI
7.56
8.23
7.25
8.66
8.24
8.90
8.08
8.03
7.84
9.38
8.26
8.22
0.60
- 110 -
PNG
1.19
1.94
-0.48
2.07
2.65
-0.07
1.21
-0.15
-2.11
-3.59
-2.20
0.04
2.01
Terasen TransAlta
-0.91
4.27
-4.26
3.60
-3.74
1.75
6.34
1.99
-2.33
2.17
1.88
6.20
3.77
-4.70
5.26
-1.76
0.65
-2.38
0.06
-7.22
-0.30
-1.12
0.58
0.25
3.48
4.06
TCPL
2.28
1.36
0.95
1.08
0.58
-3.17
-2.16
-1.46
1.28
2.40
3.01
0.56
1.98
Mainline Foothills
0.58
0.00
-0.34
0.00
0.31
0.00
0.58
0.00
0.48
0.00
0.42
0.00
0.06
0.00
0.09
0.00
0.40
0.00
0.42
0.00
0.39
0.00
0.31
0.00
0.27
0.00
Schedule 6
MACROECONOMIC DATA
GDP
UNEMP T BILL
GROWTH RATE YIELD
LONG EXCHANGE PROFITS
CANADAS
RATE
%GDP
AVG
ROE
1983
2.72
11.9
9.32
11.77
.811
8.93
9.34
1984
5.81
11.3
11.10
12.75
.772
10.16
10.53
1985
4.78
10.5
9.46
11.11
.733
10.24
10.47
1986
2.42
9.6
8.99
9.54
.720
8.82
9.49
1987
4.25
8.9
8.17
9.93
.754
10.36
11.19
1988
4.97
7.8
9.42
10.23
.812
10.58
12.71
1989
2.62
7.5
12.02
9.92
.845
9.07
10.88
1990
0.19
8.1
12.81
10.85
.857
6.61
5.68
1991
-2.09
10.4
8.83
9.81
.873
4.80
2.00
1992
0.87
11.3
6.51
8.77
.828
4.66
0.18
1993
2.34
11.2
4.93
7.85
.775
5.65
3.64
1994
4.80
10.4
5.42
8.58
.732
8.49
7.20
1995
2.81
9.5
6.98
8.36
.729
9.41
8.04
1996
1.62
9.7
4.31
7.54
.733
9.60
8.09
1997
4.22
9.1
3.21
6.47
.722
9.96
9.11
1998
4.10
8.3
4.74
5.45
.674
9.41
9.30
1999
5.53
7.6
4.70
5.68
.673
11.27
10.7
2000
5.23
6.8
5.48
5.92
.673
12.63
11.7
2001
1.78
7.4
3.85
5.79
.646
11.47
8.9
2002
3.09
7.7
2.56
5.67
.637
11.77
6.8
2003
2.02
7.6
2.87
5.29
.714
12.14
11.4
2004
2.93
7.2
2.27
5.08
.768
13.58
N/A
- 111 -
-2.00
"1
95
1"
"1
95
4"
"1
95
7"
"1
96
0"
"1
96
3"
"1
96
6"
"1
96
9"
"1
97
2"
"1
97
5"
"1
97
8"
"1
98
1"
"1
98
4"
"1
98
7"
"1
99
0"
"1
99
3"
"1
99
6"
"1
99
9"
"2
00
2
SCHEDULE 7
CPI I nflat ion
14.00
12.00
10.00
8.00
6.00
4.00
2.00
0.00
CPI
- 112 -
SCHEDULE 3
T .Bill, L ong Canada Yields
20.00
18.00
16.00
14.00
10.00
8.00
6.00
4.00
2.00
T.Bills
- 113 -
Canadas
CPI
20
01
"1
99
3"
"1
99
7"
"1
98
5"
"1
98
9"
"1
97
7"
"1
98
1"
"1
96
9"
"1
97
3"
0.00
"1
96
1"
"1
96
5"
%
12.00
SCHEDULE 8
- 114 -
20
04
"1
99
8"
"1
98
6"
"1
99
2"
"1
97
4"
"1
98
0"
"1
96
2"
"1
96
8"
6.00
4.00
2.00
0.00
-2.00
-4.00
-6.00
-8.00
-10.00
"1
95
0"
"1
95
6"
% of GD P
Gover nment Net L ending
SCHEDULE 9
CANADA BOND YIELDS
Overnight money market rates
3.74
Benchmark bonds
Canada
91 day Treasury Bill yield
3.89
Canada
Six month Treasury Bills
3.99
Canada
One year Treasury Bills
4.09
Canada
Two year
4.03
Canada
Three year
4.14
Canada
Five year
4.22
Canada
Seven year
4.27
Canada
Ten year
4.32
Canada
Long term (30 year)
4.37
Canada
Real return bonds
1.60
Marketable Bond Average yields
Canada
1-3 year
4.05
Canada
3-5 year
4.18
Canada
5-10
4.29
Canada
Over tens
4.36
US
Five year Treasuries
4.79
US
Long term (30 year)
4.81
Other
Source:
Bank of Canada’s web site at http://bankofcanada.ca/en/securities.htm, for April 4, 2006 except US
rates, which are for March 29, 2006.
- 115 -
-5
-10
-15
- 116 -
Jul 2005
Jan 2004
Jul 2002
Jan 2001
Jul 1999
Jan 1998
Jul 1996
Jan 1995
Jul 1993
Jan 1992
Jul 1990
Jan 1989
Jul 1987
Jan 1986
Jul 1984
Jan 1983
Jul 1981
Jan 1980
SCHEDULE 10
MCI
25
20
15
10
5
0
SCHEDULE 11
Pr eT ax Cor por at e Pr ofit s
% of GDP
16.00
14.00
12.00
10.00
8.00
6.00
4.00
2.00
0.00
"1970"
"1975"
"1980"
"1985"
- 117 -
"1990"
"1995"
"2000
19
94
19 Q3
95
19 Q2
96
19 Q1
96
19 Q4
97
19 Q3
98
19 Q2
99
19 Q1
99
20 Q4
00
20 Q3
01
20 Q2
02
20 Q1
02
20 Q4
03
20 Q3
04
20 Q2
05
-Q
1
SCHEDULE 12
Capacity Utilisation
88
86
84
82
80
78
76
Manufacture
- 118 -
Non-farm
- 119 -
2006/02
2005/07
2004/12
2004/05
2003/10
2003/03
2002/08
2002/01
2001/06
2000/11
2000/04
1999/09
1999/02
1998/07
1997/12
1997/05
1996/10
1996/03
SCHEDULE 13
FX Rate
0.9
0.85
0.8
0.75
0.7
0.65
0.6
SCHEDULE 14
BBB Spread
ROE
- 120 -
2004
2003
2002
2001
2000
1999
1998
1997
1996
1995
1994
1993
1992
1991
1990
1989
1988
1987
1986
1985
1984
1983
1982
Percent
16.00
14.00
12.00
10.00
8.00
6.00
4.00
2.00
0.00
1981
400
350
300
250
200
150
100
50
0
1980
basis points
Corporate ROE and BBB Spread
BBB
- 121 -
A
Daily data back to December 2000, monthly data before then.
27/03/2006
27/09/2005
27/03/2005
27/09/2004
27/03/2004
27/09/2003
27/03/2003
27/09/2002
27/03/2002
27/09/2001
27/03/2001
27/09/2000
27/03/2000
27/09/1999
27/03/1999
27/09/1998
27/03/1998
27/09/1997
27/03/1997
27/09/1996
27/03/1996
`
SCHEDULE 15
SPREADS
350
300
250
200
150
100
50
0
SCHEDULE 16
Financing Activity in Canada
$Millions
1999
2000
Government
63375.13 72085.88 64184.47 70518.54 60686.37 66361.22 70789.63
74499.6
Common equities
Preferred equities
27156.37 17693.14 16748.78 21151.89 25617.56 18267.19 17737.76 20805.32 12117.85 14863.99 18765.65 25941.18
2622.3 1388.06
1374.7 3216.84 3325.52 3661.45 3464.11 2394.69 4589.37 3097.35 4106.02 3314.03
Debt
13419.72
Capital trust
Limited partnership
Trust units
1993
9309.51 10438.31
1996
14654.6
1997
1998
2001
2002
2003
2004
69801.1 85379.58 93050.56 105322.1
19457.4 26617.04 34701.19 39223.04 39822.19
32373.5 54195.07 59537.24
0
0
0
0
0
0
3140
1750
2100
1650
602.75
114.69
520.37
118.14
407.7
1172.07
690.33
376.57
211.63
516.93
636.32
1876.82
1545.54
0
0
411.03
4264.3 10306.57
1822.81
1498.08
2878.83
106688.2
GDP
Private
1995
0
Total
Financing Ratio
1994
6996.72 11023.91 17097.11 16853.15
100997 93275.43 114213.9 120565.5
117420 128567.3 143153.1 135594.2 149474.7 190741.2
213116
727184
770873
810426
836864
882733
914973
982441
1076577
1108048
1154204
1216191
1290185
14.67
13.10
11.51
13.65
13.66
12.83
13.09
13.30
12.24
12.95
15.68
16.52
5.956275 3.750434 3.589589 5.221318 6.783378 5.580364 5.881036 6.377018 5.937745 5.553184 8.032511 8.354917
Source Data: http://www.ida.ca
- 122 -
SCHEDULE 17
Financing Activity
% of GDP
18
9
16
8
14
7
12
6
10
5
8
4
6
3
4
2
2
1
0
0
1993
1994
1995
1996
1997
1998
Total
- 123 -
1999
2000
Private
2001
2002
2003
2004
Schedule 18
Yield Spreads on Utility/Pipeline Debt
Source RBC Capital markets: Spread history of Canadian, Corporate and Government Issuers (various issues)
1999
CU Inc
Maritime and NE Pipe
GMI
Newfoundland Power
Enbridge Pipe
TCPL
Enbridge Gas
Alliance
TQ&M
Epcor Utilities
Fortis
Nova Scotia Power
Westcoast
Union Gas
Emera
Terasen Gas
Average
S&P
A
A
A
AAAABBB+
BBB+
BBB+
BBB
BBB+
BBB
BBB
BBB
BBB
Date
2019
2019
2009
2022
2009
2026
2027
2015
2009
2029
2010
2009
2010
2025
2006
2029
2000
2001
DEC
64
DEC
110
DEC
130
70
115
63
130
93
133
90
87
100
140
90
169
130
159
120
169
225
108
107
149
90
91
150
138
90
150
80
131
136
141
105
191
235
80
88
126
100
155
129
70
- 124 -
2002
DEC
104
146
80
185
65
187
120
157
105
234
275
145
130
162
165
175
152
2003
MAR
120
123
80
170
70
201
135
161
105
205
250
135
150
176
165
182
152
JUN
115
138
75
160
70
157
125
108
105
182
180
125
115
127
130
169
130
2004
SEPT
100
138
70
160
70
141
100
87
85
150
180
95
80
126
115
145
115
DEC
95
119
50
130
47
100
83
95
65
140
115
72
70
95
115
129
95
MARCH
104
85
50
130
47
110
86
97
65
129
105
72
80
102
65
128
91
JUNE
107
90
50
110
49
118
103
79
49
135
105
60
92
120
70
129
92
SEPT
108
85
50
110
49
119
99
94
49
135
105
60
78
122
70
141
92
Schedule 19
- 125 -
APPENDICES TO WRITTEN REPLY EVIDENCE OF
A. LAWRENCE KOLBE
APPENDIX R-KOLBE2: KOLBE REPLY EVIDENCE IN EB-2005-0520
(Page numbers in Kolbe Reply Evidence are from the original document)
295
EB-2005-0520
Exhibit E2
Supplemental
Tab 1
WRITTEN REPLY EVIDENCE
OF
A. LAWRENCE KOLBE
FOR
UNION GAS LIMITED
The Brattle Group
44 Brattle Street
Cambridge, Massachusetts 02138
617.864.7900
May 2006
WRITTEN REPLY EVIDENCE OF
A. LAWRENCE KOLBE
Table of Contents
I.
INTRODUCTION AND SUMMARY . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
II.
CAPITAL STRUCTURE TOPICS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6
A.
NET TAX ADVANTAGES TO DEBT DO REDUCE COMPETITIVE PRICES . . . . . . . . . . 6
B.
ATWACC IS A TOOL, NOT AN OUTCOME . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9
C.
FLAWED NUMERICAL EXAMPLE . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11
D.
INADEQUATE REVIEW OF THE CAPITAL STRUCTURE LITERATURE . . . . . . . . . . . . 12
E.
RELIANCE ON SELECTED REGULATORY DECISIONS RATHER THAN THE
CAPITAL STRUCTURE LITERATURE . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16
F.
INCORRECT CLAIMS AND CHARACTERIZATIONS . . . . . . . . . . . . . . . . . . . . . . . . . 18
1.
Not Relying on the Miller 1977 Model . . . . . . . . . . . . . . . . . . . . . . . . . . 18
2.
Not the Highest Possible Level of Change . . . . . . . . . . . . . . . . . . . . . . . 19
3.
Not Ignoring Non-Tax Costs of Debt . . . . . . . . . . . . . . . . . . . . . . . . . . . 21
4.
Regulation Does Not and Should Not Work as Claimed . . . . . . . . . . . . . 22
G.
FINANCIAL RISK DEPENDS ON MARKET VALUES, NOT BOOK VALUES . . . . . . . . 24
H.
PRINCIPLES DO WORK FOR DEEMED EQUITY RATIO ANALYSIS . . . . . . . . . . . . . 30
III.
MARKET-TO-BOOK TEST . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31
IV.
OTHER TOPICS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33
A.
ECONOMICALLY INCORRECT RETURN ADEQUACY STANDARD . . . . . . . . . . . . . . 33
B.
MISTAKEN COMMENTS ON DR. VILBERT’S EVIDENCE . . . . . . . . . . . . . . . . . . . . . 35
1.
Risk Implications of Holding Companies in Sample Groups . . . . . . . . . 36
2.
Unique Theory of Financial Risk . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37
APPENDIX R-A: DETAILED DISCUSSION OF DR. BOOTH’S EXAMPLE . . . . . . R-A-1
APPENDIX R-B: ISSUES RAISED IN AEUB DECISION U99099 . . . . . . . . . . . . . . . . . R-B-1
A.
MEASURED ATWACC VS. THE DEBT RATIO . . . . . . . . . . . . . . . . . . . . R-B-1
1.
Factors that Distort the Comparison . . . . . . . . . . . . . . . . . . . . . . . . . R-B-2
2.
Factors Left Out of the Measured ATWACC . . . . . . . . . . . . . . . . . . R-B-3
B.
MARKET VS. BOOK CAPITAL STRUCTURE WEIGHTS . . . . . . . . . . . R-B-5
WRITTEN REPLY EVIDENCE OF
A. LAWRENCE KOLBE
1
I.
INTRODUCTION AND SUMMARY
2
Q1.
Please state your name and address for the record.
3
A1.
My name is A. Lawrence Kolbe. My business address is The Brattle Group, 44 Brattle
4
Street, Cambridge, Massachusetts, 02138.
5
Q2.
Did you provide written evidence earlier in this proceeding?
6
A2.
Yes.
7
Q3.
What is the purpose of your reply evidence?
8
A3.
I have been asked by Union Gas Limited (“Union,” or the “Company”) to review the
9
Evidence of Laurence D. Booth on behalf of the Consumers Council of Canada, the
10
Industrial Gas Users’ Association and the Vulnerable Energy Consumers Coalition (“Booth
11
Evidence”), and, if necessary, to comment on the parts of it that address the written
12
evidence of Dr. Michael J. Vilbert and myself.
13
Q4.
Please summarize the conclusions of this review.
14
A4.
The review leads to comments in the areas of (1) capital structure principles, (2) the utility
and meaning of the market-to-book ratio for regulated companies, and (3) other topics.1
15
1
Citations to specific parts of the Booth Evidence come in the body of the my reply and are not repeated
in this initial summary.
WRITTEN REPLY EVIDENCE OF
A. LAWRENCE KOLBE
1
Capital Structure: Discussion of capital structure principles occupies the bulk of my
2
reply, covering the following points in response to Dr. Booth’s evidence:2
3
4
5
6
7
8
9
•
Dr. Booth maintains that the corporate tax advantage of debt goes to the
shareholders of unregulated firms and to the customers of regulated firms, since the
use of debt reduces the firm’s revenue requirement. The second half of the
statement is correct, but the first half is incorrect. Cost-reducing activities that are
easily copied will lead to lower prices in a competitive market. It is hard to think
of a cost-reducing activity that is easier to copy than to finance part of the firm’s
investments with debt.
10
11
12
13
14
15
16
17
•
The after-tax weighted-average cost of capital (“ATWACC”) is just a tool. It can
be used by unregulated companies to maximize shareholder value without
invalidating its use by regulators to discern the just-fair level of return. (Were this
not true, why would regulators in other Commonwealth countries use it so
widely?3) Use of the ATWACC does not automatically lead to value maximization,
any more than the fact that some people use hammers to build houses automatically
means that anyone else who decides to use a hammer will end up building a house
whether they want to or not.
18
19
20
21
22
23
24
25
26
27
28
•
Dr. Booth’s extended numerical example does not accurately portray my evidence
or recommendations. This is immediately evident from the fact that it postulates
changes in the ATWACC in response to changing capital structure, while the entire
thrust of my review of the capital structure literature is that the ATWACC for any
industry is constant over a broad middle range of capital structures. As
demonstrated in Appendix R-A, Dr. Booth’s numerical example of capital structure
and the cost of capital is internally inconsistent and inconsistent with the way
capital markets work. If implemented without the example’s built-in errors, the
ATWACC method would work without generating any of the problems Dr. Booth
claims it would. (Again, if this were not true, why is the method used so widely by
regulators in other Commonwealth countries?)
2
The evidence filed by Dr. Vilbert and myself in this proceeding focuses on the deemed equity ratio, but
takes advantage of decades of financial research to address the issue quantitatively, not merely
qualitatively. That necessarily involves an analysis of what the research says about how the return on
equity and capital structure interact. Dr. Booth responds to this analysis with comments of his own,
many of them involving the overall weighted-average cost of capital. To keep the discussion from
becoming unnecessarily complex, here and in the material in Appendix R-A I respond to those
comments as stated, without constantly trying to translate both Dr. Booth’s statements and my reply into
deemed equity ratio terms.
3
See the answer to interrogatories in Exhibit J2.06, part (a), in this proceeding.
2
WRITTEN REPLY EVIDENCE OF
A. LAWRENCE KOLBE
1
2
3
4
•
Dr. Booth’s own review of the capital structure literature is very limited and cites
no paper written in the last quarter-century. It makes some references to more
recent textbooks, but even those references stop well short of providing a complete
summary of the relevant sections.
5
6
7
8
9
10
•
Dr. Booth ultimately relies on selected regulatory decisions rather than the
scholarly literature on capital structure, and particularly on a 1999 decision by the
Alberta Energy and Utilities Board (“AEUB”). If regulatory decisions by other
bodies rather than the findings of the economic literature are to be the arbiter of
decisions about capital structure principles, my reply identifies plenty of examples
that contradict those that Dr. Booth cites.
11
•
Dr. Booth’s evidence contains a number of incorrect claims and characterizations:
<
12
13
14
15
16
17
18
19
20
21
22
23
24
<
<
<
My evidence does not rely on the particular theoretical model (the 1977
Miller model) that Dr. Booth claims it does, as is plainly demonstrated
within my evidence.
The equation I use to adjust for changes in the deemed equity ratio does not
produce the highest possible changes in the cost of equity as the equity ratio
changes.
My evidence and procedures do not ignore the non-tax costs of debt. To
the contrary, my direct evidence devotes considerable discussion to these
costs, and they have a material impact on the procedures I adopt.
Dr. Booth claims that regulation works in a way that makes utilities riskier
when their market-to-book ratios are higher. Regulation does not work this
way, nor should it. But even if it did, that would have no impact on the
capital structure principles used in my evidence.
25
26
27
28
29
30
•
Financial risk is the risk equityholders bear when the firm uses debt. That risk
affects the cost of equity, and the cost of equity is determined in the market, not on
the company’s books. Therefore, financial risk depends on market values, not book
values. Dr. Booth purports to show the contrary with quotations from Standard &
Poor’s and from a regulatory decision, but both of these quotations actually address
other topics.
31
In short, none of Dr. Booth’s criticisms of my capital structure procedures have economic
32
merit, nor are they supported by relevant citations to the economic literature. Some of
33
them flatly misstate my evidence.
3
WRITTEN REPLY EVIDENCE OF
A. LAWRENCE KOLBE
1
Market-to-Book Ratios: Dr. Booth wrongly relies on the market-to-book ratio to assess
2
the adequacy of utility returns. At one time, this was a reasonable approach. However, a
3
combination of (1) events such as the 1987 stock market crash and the “tech bubble” of the
4
turn of the century, and (2) a growing body of financial research, have taught us that stock
5
prices are more complicated than our simple models can handle. The market-to-book test
6
always was known to have material limitations, but we now know that regulators simply
7
cannot safely rely on the market-to-book ratio test.
8
Other Issues: Dr. Booth proposes an insufficient standard for the adequacy of regulated
9
returns. Based in part on an incomplete quotation from a legal decision, he proposes that
10
the ability to attract capital be the standard. But even utilities with materially inadequate
11
rates of return can still attract capital for a while, as long as they draw on the credit
12
capacity of existing assets. Economically, the return itself needs to be adequate, which
13
requires more than the mere ability to attract capital.
14
Dr. Booth also overstates the importance of the fact that some of Dr. Vilbert’s
15
sample companies have non-regulated components. Part of his conclusions in this regard
16
are based on a review of book rates of return, but book rates of return look backwards and
17
are intrinsically less volatile than the forward-looking market. Since the cost of capital is
18
set in the market, not on the books, this is a major failing. Dr. Booth also ignores the
19
tendency of measured costs of capital for troubled investments to understate their true cost
20
of capital. Nor does he acknowledge the high quality of Dr. Vilbert’s LDC sample. Dr.
21
Vilbert and I also rely on Dr. Carpenter’s finding that Union is riskier than the LDC
22
sample, which is a particularly useful benchmark, given the LDC sample’s high quality.
4
WRITTEN REPLY EVIDENCE OF
A. LAWRENCE KOLBE
1
In short, Dr. Booth’s comments do not undercut our recommendation for Union’s deemed
2
equity ratio.
3
Finally, Dr. Booth interprets some of Dr. Vilbert’s sample cost of capital numbers
4
as implying that allowed returns are in excess of utilities’ cost of capital as measured by
5
himself and Dr. Vilbert. This discussion implies that Dr. Vilbert agrees with Dr. Booth that
6
allowed returns are too high. This plainly is not Dr. Vilbert’s view, and it is misleading to
7
suggest otherwise, for a number of reasons. The most fundamental of these reasons is that
8
Dr. Booth’s interpretation of Dr. Vilbert’s cost of equity estimates simply ignores the huge
9
difference in financial risk between the sample companies at their market-value capital
10
structures and Union at its regulatory capital structure. This comparison implies that Dr.
11
Booth believes that the cost of equity is the same even at very different capital structures,
12
which implies that equity investors demand no compensation for financial risk. No existing
13
theory of capital structure in the literature implies such a result.
14
Q5.
How is the remainder of your evidence organized?
15
A5.
Section II addresses Dr. Booth’s comments on capital structure. Section III addresses his
16
use of the market-to-book ratio. Section IV covers my comments on his proposed return
17
adequacy standard and his discussion of Dr. Vilbert’s evidence. Appendix R-A provides
18
a detailed explanation of the flaws in Dr. Booth’s extended numerical example. Appendix
19
R-B discusses and resolves issues raised by the AEUB decision that Dr. Booth’s evidence
20
cites so frequently.
5
WRITTEN REPLY EVIDENCE OF
A. LAWRENCE KOLBE
1
II.
CAPITAL STRUCTURE TOPICS
2
Q6.
How is this section organized?
3
A6.
It covers the various capital structure points summarized in the introduction, in that order.
4
A.
NET TAX ADVANTAGES TO DEBT DO REDUCE COMPETITIVE PRICES
5
Q7.
What is the first issue you discuss here?
6
A7.
At p. 13, Dr. Booth claims (emphases in the original) that
7
8
9
10
11
12
13
. . . since interest payments are tax deductible, whereas equity dividends are
not . . . there is a built-in tax advantage to any corporation using debt
financing. This tax advantage goes to the shareholders of unregulated
firms and to the customers of regulated firms, since the use of debt reduces
the firm’s revenue requirement. As will be discussed later, this asymmetry
in benefits for the regulated firm is a motivating factor behind regulated
companies continually striving to increase their equity ratios.
14
Q8.
Is he right?
15
A8.
Not entirely. He is right that rate regulation sets rates at (regulatory) cost, and that this has
16
the effect of reducing the revenue requirement by the net tax advantage of debt.4 However,
17
he is wrong that shareholders get to keep this advantage in competitive industries. Any
18
cost advantage that is easily copied by other competitors will be eliminated from
19
competitive prices.
4
By “net” tax advantage, I mean the net of the benefit of the interest tax advantage at the corporate level
over (1) the interest tax disadvantage at the personal level and (2) the non-tax costs of debt. See
Appendix B of my written evidence, especially Figures B-2 and B-3. At low debt levels in most
industries, debt conveys a net advantage initially, but too much debt conveys a net cost.
6
WRITTEN REPLY EVIDENCE OF
A. LAWRENCE KOLBE
1
The basis of Dr. Booth’s error is later on the same page, where he says (emphasis
2
in the original),
3
4
5
6
7
This implies that there is an incentive for competitive firms to finance with
debt: as they replace expensive equity with “cheap” debt, their cost of
capital goes down. Hence, for the same fixed amount of operating income,
it is the stockholders who benefit from the tax advantage of debt financing
for competitive firms.
8
Q9.
Why is this the basis of his error?
9
A9.
Operating income will not stay fixed in a competitive industry if an easily copied, cost-
10
reducing strategy exists.
11
For example, suppose debt conveyed a huge net cost advantage, such that the first
12
firm to use it became much more profitable than any other company in the industry. That
13
firm could afford to cut prices and gain market share while still earning an adequate, or
14
(temporarily) above-adequate return. Other firms in the industry would take note of the
15
advantage debt conveyed as they lost market share, and they would rapidly issue their own
16
debt so they could match the first firm’s price cuts while remaining adequately profitable
17
themselves. The end result would be a fall in prices until all firms had to issue debt and
18
get the postulated large net tax advantage just to maintain adequate profitability.5 The
19
winners would be the industry’s customers, who now would get lower prices, just as under
20
rate regulation.
5
The precise definition of “adequate profitability” for shareholders would go up due to the greater
financial risk they bear with debt, but the overall effect, debt and equity combined, would be to reduce
the competitive industry’s costs and hence its prices.
7
WRITTEN REPLY EVIDENCE OF
A. LAWRENCE KOLBE
1
Q10. Did Dr. Booth attempt to clarify his position on this issue in response to the question
2
3
in Union’s Question L2.27?
A10.
4
5
6
7
8
9
10
11
12
13
14
Apparently. His concluding statement is that
However, the flaw in the chain of logic [in the earlier parts of the question,
which underlies my above reply evidence] is that markets do not operate
according to the standard model of a stable firm operating in a frictionless
world under certainty. This firm only exists on an economist’s blackboard.
The real world exists in business schools, where firms have to deal with
uncertainty and disequilibrium markets with costs. In this world debt is not
an easily replicable strategy, since it depends on generating cash flows to
support interest payments and having lenders lend funds on reasonable
terms. In the world envisaged in [the above discussion] all firms have AAA
bond ratings and there are no investment bankers earning large salaries
living in Rosedale.
15
Q11. Do you agree?
16
A11.
I agree the real world has frictions, uncertainties, disequilibria and transaction costs. I
17
agree investment bankers can earn large salaries because of such factors, and that not all
18
firms have AAA bond ratings. However, while such factors affect the details, they do not
19
affect my basic conclusion above.
20
Dr. Booth felt it necessary in his evidence to emphasize the qualification, “for the
21
same fixed amount of operating income.” In a world in which debt conveyed a material
22
net advantage for a particular industry, operating income for firms in that industry would
23
not stay fixed at the level at which it would have been if no firm used debt, any more than
24
it would stay fixed if there were a technological breakthrough that lowered all firms’
25
production costs materially.
26
Moreover, Dr. Booth’s claim that the mere issuance of debt is not “an easily
27
replicable strategy” because the firm has to be able to service the debt afterwards only
8
WRITTEN REPLY EVIDENCE OF
A. LAWRENCE KOLBE
1
demonstrates that he has unreasonably high standards for how he defines “easily
2
replicable.” The fact that most firms in most industries in the “real world” manage to
3
obtain debt financing despite Dr. Booth’s worries is proof of that. Patent protection, in
4
contrast, can provide a cost-saving technique that is hard to replicate, which permits a
5
company in an otherwise competitive industry to earn above-normal profits for years at a
6
time. Issuance of debt does not come close to being a hard-to-replicate strategy by any
7
reasonable definition, else there would be many fewer “investment bankers earning large
8
salaries and living in Rosedale.”
9
B.
ATWACC IS A TOOL, NOT AN OUTCOME
10
Q12. What is the next issue?
11
A12.
Dr. Booth at various places6 notes that unregulated firms use the after-tax weighted-average
12
cost of capital (“ATWACC”)7 as a discount rate to try to maximize the value of the firm,
13
and suggests that suggests this somehow invalidates the use of ATWACC (and implicitly
14
the principles associated with it) by regulators, because regulators’ goal should not be to
15
maximize the value of the firm.8
6
See the middle bullet on p. ii, the first full paragraph on p. 79, and the last bullet on p. 104.
7
A word about terminology: the ATWACC is the market-value weighted average of the cost of equity
and the after-tax cost of debt. Dr. Booth’s evidence often uses the uses the conventional corporate
finance term “WACC” to refer to the overall cost of capital calculated with market-value weights.
Depending on whether the discussion at that time includes or excludes taxes, it may or may not
correspond to the ATWACC as I use the term. However, I have not found a place in the part of the
evidence that discusses capital structure principles in which “WACC” corresponds to the conventional
regulatory weighted-average cost of capital. In this reply and in Appendix R-A, for convenience I
sometimes use “WACC” in the sense of the Booth Evidence, and I never use it in the sense of the
traditional regulatory weighted-average cost of capital.
8
On p. ii, for example, he says
(continued...)
9
WRITTEN REPLY EVIDENCE OF
A. LAWRENCE KOLBE
1
Q13. What is your reaction to this view?
2
A13.
It is plainly incorrect. The ATWACC is a tool, not a result. It is as though someone said,
3
“unregulated firms use hammers to try to build houses. Regulators want to build garages,
4
so they shouldn’t make use of hammers, because if they use hammers, they will
5
automatically build houses.” There is no contradiction at all in unregulated firms using
6
ATWACC to try to maximize the value of the firm and regulators using the financial
7
research that underlies the ATWACC, among other things, to try to find the just-fair rate
8
of return.
9
The output of a use of the ATWACC tool, like any other tool, depends on what it
10
is used to do, not what the tool itself is.9 That tool manifestly can be used in rate
11
regulation, because it is used in rate regulation (see the answers to interrogatories in
12
Exhibits J2.06, part (a), and J13.32, part (c), in this proceeding). The use of ATWACC
13
creates value for unregulated firms because it is the threshold rate of return an unregulated
14
firm should accept to undertake a project. That makes it the perfect concept for regulators
15
who want to discern the rate of return necessary to be fair both to regulated investors and
16
to their customers.
8
(...continued)
. . . I would point out the fundamental contradiction in [ATWACC’s] use in regulatory filings
in that it is the mirror image of shareholder value maximisation. That is, earning more than the
WACC is synonymous with the creation of shareholder value, whereas the Board’s
responsibility is not to create or maintain shareholder value, but to ensure that rates are fair and
reasonable.
9
For the convenience of the reader, I note also that Appendix R-B to this reply evidence, discussed
further below, disposes of the incorrect but possibly related view that the use of market-value weights
to calculate the ATWACC somehow automatically translates book-value regulation into market-value
regulation.
10
WRITTEN REPLY EVIDENCE OF
A. LAWRENCE KOLBE
1
C.
FLAWED NUMERICAL EXAMPLE
2
Q14. What numerical example do you discuss in this part of your reply?
3
A14.
The example begins on p. 70 of the Booth Evidence and serves general expository
4
purposes. The part I need to address begins on p. 74 of the Booth Evidence, where debt
5
is added. The example itself runs through p. 81, and further use is made of it at pp. 83-89
6
and 96. The example is lengthy, and the explanation of its flaws involves a step-by-step
7
review. I provide that explanation in Appendix R-A for the reader who wants to see the
8
details.
9
However, its basic problem may be explained more quickly. Dr. Booth’s example
10
supposes regulation based on the WACC as a rate-of-return standard. It goes through a
11
series of flawed steps that produce higher values for the regulated company’s allowed
12
WACC, as the market-value capital structure changes, than the WACC value that is
13
consistent with the underlying sample data and assumptions. But the entire thrust of my
14
review of the financial literature is that the WACC for any industry is constant over a broad
15
middle range of capital structures (which range does vary by industry). If Dr. Booth’s
16
example produces a different outcome, as it does, the example cannot be an accurate
17
representation of either the current state of the economic literature or my evidence.
18
Appendix R-A shows that the numerical example is in fact fundamentally flawed,
19
and why. In particular, the example assumes the market will not recognize the true level
20
of financial risk the company bears, and as a result, the example assumes an excessive cost
21
of equity that leads to an excessive estimate of the WACC. In reality, it is the numerical
22
example that does not recognize the actual level of financial risk. Were regulators actually
23
to adopt a WACC-based system of regulation, the starting assumption in Dr. Booth’s
11
WRITTEN REPLY EVIDENCE OF
A. LAWRENCE KOLBE
1
flawed example, the system would work without the problems Dr. Booth claims would
2
arise. In reality, a WACC-based system does not generate excess compensation to
3
investors (which evidently is also a conclusion reached by the regulatory bodies that have
4
adopted WACC-based regulation in other Commonwealth countries).
5
In short, Dr. Booth’s example is internally inconsistent, inconsistent with the way
6
capital markets work, and misinterprets my recommendations. Appendix R-A should be
7
consulted before relying on conclusions from Dr. Booth’s example for any purpose.
8
D.
9
10
INADEQUATE REVIEW OF THE CAPITAL STRUCTURE LITERATURE
Q15. What support for his views does Dr. Booth draw from the financial literature?
A15.
Very little. The discussion covers parts of pp. 89-95. The financial literature this section
11
cites is sparse. It mentions the 1958 and 1963 papers by Modigliani and Miller, the 1977
12
paper by Miller, a 1963 article by Ezra Solomon, and a 1963 publication Gordon
13
Donaldson.10 It identifies no study of capital structure that has been published in the last
14
quarter-century, despite the fact that this has been and remains a very active field of
15
research.11
10
Actually, the passage does not include formal citations for any of the publications. The first three
papers are among many that are cited in my own Appendix B. I am unaware of a relevant 1963 citation
to Donaldson, although perhaps it is a version of Gordon Donaldson, Corporate Debt Capacity, Boston:
Division of Research, Graduate School of Business Administration, Harvard University (1961). The
Solomon paper is Ezra Solomon, “Leverage and the Cost of Capital,” The Journal of Finance 18:273-79
(May 1963).
11
An apparent but not real exception: it cites the 1991 Canadian edition of Richard A. Brealey and
Stewart C. Myers, Principles of Corporate Finance, regarding the Adjusted Present Value (“APV”)
approach to calculation of the value of the tax advantage to debt. However, Prof. Myers developed the
APV approach in 1974: S. C. Myers, “Interactions of Corporate Financing and Investment Decisions,”
The Journal of Finance 29:1-25 (March 1974). Dr. Booth also cites the third edition of the
undergraduate version of the text, Fundamentals of Corporate Finance by Brealey, Myers and Marcus,
(continued...)
12
WRITTEN REPLY EVIDENCE OF
A. LAWRENCE KOLBE
1
The literature summary itself consists of a brief overview of the tax-based theories
2
of Modigliani and Miller and Miller, a mention of the static tradeoff model (referring to
3
calculations using APV), a mention of the pecking order hypothesis (attributing it to Prof.
4
Donaldson), and a mention that the Solomon paper showed that the cost of equity does not
5
increase as fast if the cost of debt increases when firms add debt.
6
Q16. Why do you label this review “inadequate”?
7
A16.
The last paper it mentions was published nearly three decades ago. It reviews none of the
8
more recent literature explicitly, despite the fact that Prof. Booth himself has published a
9
relevant paper that I cite in my own Appendix B. It does not address any of the empirical
10
studies I cite as support for my conclusions on the actual effects of capital structure.
11
Rather than directly addressing the findings of the research I report, it states Prof. Booth’s
12
personal views and appeals to regulatory decisions rather than to the economic literature.12
11
(...continued)
published in 2001, but not as a claimed source of original research on capital structure.
12
I would also note that the Solomon paper Dr. Booth cites was shown to be erroneous shortly after it was
published. (See Alexander A. Robichek and Stewart C. Myers, Optimal Financing Decisions,
Englewood Cliffs, NJ: Prentice Hall, Inc. (1965), pp. 34-36 and 48-49; the Solomon paper wrongly
claimed the cost of equity could actually decrease as the debt ratio increased.) Dr. Booth appears to say
he is citing the paper not for its original, erroneous finding, but merely for the observation that if the
cost of debt increases with the debt ratio, the cost of equity does not increase as fast as it would if the
cost of debt stayed constant. To the best of my knowledge, this point has never been in dispute, but it
still is not the point the Solomon paper attempted to make. For completeness, I would note that the
facts that the risk of financial distress looms and the cost of debt increases as the debt ratio increases
are in no way inconsistent with my evidence. My Appendix B goes well beyond the early models
addressed in the Booth Evidence, to include explicit consideration of the costs of excessive debt.
Additionally, Dr. Vilbert and I explicitly considered the effect of changes in the cost of debt as capital
structure changes and concluded that it would be a mistake to reflect such changes in this context. The
current cost of debt is the correct input to our calculations, not that which would exist under different
conditions.
13
WRITTEN REPLY EVIDENCE OF
A. LAWRENCE KOLBE
1
One point in this discussion that merits some emphasis is Dr. Booth’s use of the
2
textbooks of which Prof. Myers is a co-author. Dr. Booth in footnote 25 on p. 93 cites an
3
illustration of the APV method in the above-cited Brealey and Myers textbook as evidence
4
that “In this case, Professor Myers, who Dr. Kolbe references throughout his testimony,
5
clearly believes in the tax advantages of debt.” Of course, textbooks routinely use
6
simplified numerical examples to teach a technique without holding that the results of the
7
example are perfectly general. The view that Prof. Myers “clearly believes in the tax
8
advantages of debt” is plainly inconsistent with Prof. Myers’s body of published research.13
9
Moreover, at pp. 16-17 Dr. Booth has lengthy quotations from and discussion of
10
the Brealey, Myers and Marcus textbook, culminating with the allegation that
11
12
13
14
15
16
17
It is obvious that academics associated with the Brattle group when writing
textbooks for the academic audience have no problem noting the tax
advantages of debt and expounding on standard ideas in finance. Further
on page 437 the Myers et al text also has the standard graph showing that
the weighted average cost of capital (ATWACC in this hearing) declines as
the firm uses more debt, that is, the ATWACC is not constant as assumed
by Dr. Kolbe.
18
However, had Dr. Booth turned the page on which “the standard graph” [Figure 15.6]
19
appears, he would have found that the very next thing the authors did, having taught the
20
various techniques that the students needed to master, was to provide context to the models
21
and an overview of the current state of knowledge. That summary, under the heading,
13
A leading example is Prof. Myers’s Presidential Address to the American Finance Association: Stewart
C. Myers, “The Capital Structure Puzzle,” The Journal of Finance, 39: 575-592 (1984). See also S. C.
Myers and N. S. Majluf, “Corporate Financing Decisions When Firms Have Information Investors Do
Not Have,” Journal of Financial Economics 13:187-222 (June 1984); Stewart C. Myers, “Still
Searching for Optimal Capital Structure,” Are the Distinctions Between Debt and Equity
Disappearing?, R.W. Kopke and E. S. Rosengren, eds., Federal Reserve Bank of Boston. (1989); and
Lakshmi Shyam-Sunder and Stewart C. Myers, “Testing static tradeoff against pecking order models
of capital structure,” Journal of Financial Economics 51:219-244 (February 1999).
14
WRITTEN REPLY EVIDENCE OF
A. LAWRENCE KOLBE
1
“The Implications of Corporate Taxes for Capital Structure,” is completely consistent with
2
the views in my written evidence at pp. 17, B-7, and B-19 to B-29:
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
If borrowing provides an interest tax shield, the implied optimal
debt policy appears to be embarrassingly extreme: all firms should borrow
to the hilt. This maximizes firm value and minimizes the weighted-average
cost of capital.
[Modigliani and Miller] were not that fanatical about it. No one
would expect the gains to apply at extreme debt ratios. For example, if a
firm borrows heavily, all its operating income may go to pay interest and
therefore there are no corporate taxes to be paid. There is no point in such
firms borrowing any more.
There may also be some tax disadvantages to borrowing, for
bondholders have to pay personal income tax on any interest they receive.
Stockholders, on the other hand, can get a tax break, because some of their
returns come as capital gains. Capital gains are not taxed until the stock is
sold and then are taxed at a lower rate. [Footnote omitted.]
All this suggests that there may come a point at which the tax
savings from debt level off and may even decline. But it doesn’t explain
why highly profitable companies with large tax bills often thrive with little
or no debt. There are clearly factors beyond tax to consider.14
21
Two pages after the page with his cited Figure 15.6, on p. 439, Dr. Booth would
22
have found Figure 15.7, which plots a graph of firm value against capital structure that has
23
the same basic shape and import as Figure B-2 on p. B-10 in Appendix B of my written
24
evidence, and which does not imply a steadily declining ATWACC as the firm adds more
25
debt (as is demonstrated in turn by my Figure B-3).
14
Richard A. Brealey, Stewart C. Myers and Alan J. Marcus, Fundamentals of Corporate Finance, 3rd
edition, McGraw-Hill/Irwin (2001), p. 438.
15
WRITTEN REPLY EVIDENCE OF
A. LAWRENCE KOLBE
1
In short, a comparison of the full discussion in the Brealey, Myers and Marcus text
2
shows that it and I agree that while the ATWACC does decline initially (in most industries)
3
as firms add debt, it levels out and eventually rises as the debt ratio increases.15
4
5
E.
RELIANCE ON SELECTED REGULATORY DECISIONS RATHER THAN THE CAPITAL
STRUCTURE LITERATURE
6
Q17. On what regulatory decisions does Dr. Booth rely?
7
A17.
His evidence has numerous citations in this area to the Alberta Energy and Utilities Board
8
(“AEUB”), among others. Dr. Booth appears to be particularly fond of a citation from
9
AEUB decision U99099, referring seven times (pp. ii, 2, 82, 88, 97, 100 and 104) to a
10
statement that the AEUB concluded it would be “derelict in its statutory responsibilities
11
to recognize market capitalization ratios that are derived from a market value capitalization
12
that deviates from the intrinsic long-run value of the regulated firm.”
13
Of course, if Dr. Booth could refute my conclusions by citing scholarly research
14
performed by financial economists, he should have done so. My evidence includes an
15
extensive discussion of the financial literature, which I believe solidly supports my
16
procedures. Dr. Booth does not challenge that evidence on its own terms.
15
Dr. Booth makes a new error in his response to Union’s Question L2.32, in which he was asked to
attach pages from the Brealey, Myers and Marcus text. He says, “Note this section includes a diagram
on page 433 [Figure 15.4] that shows how the cost of equity increases at a decreasing rate as the debt
becomes riskier, something that Drs Kolbe and Vilbert have denied in the past.” He apparently does
not notice that the graph on p. 433 has the debt-to-equity ratio on the x-axis, not the debt-to-value ratio
that appears in the graphs in my evidence. Since the debt-to-equity ratio goes up very fast at high debt
ratios (e.g., it climbs from 3.0 to 4.0 as the debt-to-value ratio goes merely from 75 percent to 80
percent), a declining rate of change in response to increases in the debt-to-equity ratio does not imply
a declining rate of change in response to increases in the debt-to-value ratio. The Brealey, Myers and
Marcus text is again completely consistent with my evidence.
16
WRITTEN REPLY EVIDENCE OF
A. LAWRENCE KOLBE
1
However, since Dr. Booth relies on this particular decision so heavily, it is probably
2
worthwhile discussing explicitly the issues that arose in the U99099 decision, which
3
appears to have been based in part on what amounted to expert evidence introduced for the
4
first time in post-hearing argument.16 I do so in Appendix R-B to this reply evidence,
5
which resolves the principal concerns raised by the AEUB.17
6
Additionally, I would note that the “intrinsic long-run value” cited in the above
7
quotation is book value, as indicated by another quotation to the U99099 decision on pp.
8
81-82 of the Booth Evidence. Exhibits J2.01, part (c), and J2.02 in this proceeding,
9
particularly the former, show that the market-to-book ratio test of the adequacy of a
10
utility’s returns is inconsistent with any reasonable value for its cost of equity and has been
11
disproved by actual market behavior. I cover this topic in additional detail in Section III
12
below.
13
Finally, I would note that if regulatory decisions by other bodies rather than the
14
findings of the economic literature are to be the arbiter of decisions about capital structure
15
principles, there are plenty of examples that contradict those that Dr. Booth cites, including
16
the NEB’s Decision RH-2-2004.18
16
In particular, Dr. Booth at p. 88 cites an AEUB statement that states as a basis of the AEUB’s decision
a view on how fast beta changes as debt is added. At least part of the evidence on which that statement
is based was an analysis supplied for the first time during post-hearing argument, without the
opportunity for expert reply.
17
That is, it explains the reasons that the type of post-hearing analysis cited in the previous footnote does
not rebut the capital structure principles on which I rely, and it explains why it is appropriate to use
market value weights to calculate the WACC even for companies regulated on a book-value rate base.
18
See the answers to interrogatories in Exhibits J2.06, part (a), and J13.32, part (c), in this proceeding.
Also, as explained in Exhibit J13.32, part (f), the NEB’s Decision RH-2-2004 itself does not adopt the
position the AEUB took in 1999, despite Dr. Booth’s recommendation that it should. Instead, the NEB
accepts the principle while expressing concern about the adequacy of the evidence in that proceeding.
(continued...)
17
WRITTEN REPLY EVIDENCE OF
A. LAWRENCE KOLBE
1
F.
INCORRECT CLAIMS AND CHARACTERIZATIONS
2
Q18. What does this part of your reply cover?
3
A18.
4
Dr. Booth makes a number of incorrect claims or characterizations concerning my
evidence. This part of my reply covers them in a group.
5
1.
Not Relying on the Miller 1977 Model
6
Q19. What does Dr. Booth claim to be the basis of your capital structure procedures?
7
A19.
At pp. 93-94, Dr. Booth claims that “the flat ATWACC approach of Drs. Kolbe and Vilbert
8
is essentially the 1958 M&M model as extended to include corporate and personal taxes
9
by Miller (1977).”
10
Q20. Is this claim correct?
18
(...continued)
Exhibit J13.32, part (f), reviews these evidentiary concerns and how they have been addressed in our
evidence in this proceeding.
Lastly, I am aware of additional regulatory decisions that take different positions from that Dr.
Booth espouses. These include the following decisions of the Australian Competition and Consumer
Commission (“ACCC”), the primary federal rate regulatory body. The ACCC uses procedures
consistent with the capital structure principles I discuss, including the use of market-value weights to
assess financial risk, in its: Final Decision on Central West Pipeline (June 2000), Draft Decision on
the Moomba to Adelaide Pipeline (August 2000), Draft Decision on the Moomba to Sydney Pipeline
(December 2000), and Draft Decision on the NT Gas Pipeline (May 2001). Other examples of which
I am aware include the Electricity Distribution Price Determination 2001-05, Volume I, Statement of
Purpose and Reasons, by the Office of the Regulator-General (now the Essential Services
Commission), Victoria (September 2000) and the Final Decision on the Proposed Access Arrangement
for the Dampier to Bunbury Natural Gas Pipeline, Independent Gas Pipelines Access Regulator
Western Australia, 23 May 2003. In New Zealand, the Commerce Commission has recognized the
necessity to use market value weights quite explicitly for many years. For example, the Treasury’s
handbook, “Estimating the Cost of Capital for Crown Entities and State-Owned Enterprises” (October
1997) recognizes the need for market weights. Recent regulatory reports confirm that the Commerce
Commission continues to use and advocate the use of market weights for determining the WACC. See,
for example, Final Report Part IV Inquiry into Airfield Activities at Auckland, Wellington and
Christchurch International Airports, 6 August 2002, Gas Control Inquiry, Draft Framework Paper,
July 16, 2003; and Regulation of Electricity Lines Businesses, Targeted Control Regime, Draft
Assessment and Inquiry Guidelines (Process and Analytical Framework), 7 August 2003.
18
WRITTEN REPLY EVIDENCE OF
A. LAWRENCE KOLBE
1
A20.
No, as is plain from my direct evidence. The 1977 Miller model is a purely tax-based
2
model in which the personal tax disadvantage of debt fully offsets the corporate tax
3
advantage of debt, which is an important insight. However, my discussion of the capital
4
structure literature (unlike that in the Booth Evidence -- see above) goes well beyond that
5
insight. For example, see Figures B-2 and B-3, pp. B-10 and B-12. The Miller1977 model
6
would plot straight lines for all four industries for both the values of the firm in Figure B-2
7
and the associated values for the ATWACC in Figure B-3. My figures do not do that. I
8
would not say, for example, that if a firm in Industry 4 used no debt at all, its ATWACC
9
would be equal to its value where the cost of equity is actually estimated. However, within
10
a normal range of capital structures for an industry, the financial research consistently
11
shows that debt does not affect value in any detectable way. That finding is consistent with
12
the use of a flat ATWACC within the normal range of industry capital structures.
13
14
2.
Q21. How does the Dr. Booth characterize the equation you use to quantify the way the cost
15
16
17
18
19
20
Not the Highest Possible Level of Change
of equity changes with debt?
A21.
He claims at pp. 94 that:
However, a flat ATWACC does have the advantage that it gives the largest
possible leverage effect, that is, the coefficient on the market valued debt
equity ratio in the equity cost equation is as large as possible.
A similar claim is made at pp. 86 and 97.
19
WRITTEN REPLY EVIDENCE OF
A. LAWRENCE KOLBE
1
Q22. Why does Dr. Booth characterize the equation as having the “advantage” of the
2
largest possible leverage effect?
3
A22.
4
Q23. In any case, is the statement that the flat ATWACC equation provides the maximum
5
6
He doesn’t say.
possible leverage effect correct?
A23.
No, unless one considers only the papers mentioned in the Booth Evidence’s inadequate
7
review of the capital structure literature (see part D of this section, above). It is clearly not
8
true given the overall findings of the literature. In particular, if the ATWACC has any sort
9
of “U” shape (e.g., Dr. Booth’s graph on p. 93, which imputes far more tax advantage to
10
debt than could ever occur in reality, or the more realistic Figure B-3 on p. B-12 in my
11
written evidence), the cost of equity will increase faster than it would if the ATWACC
12
were flat, at every capital structure to the right of the bottom of the “U.”19
13
This topic is addressed in detail in the response to interrogatories in Exhibit J2.08
in this proceeding, particularly in part (f).20
14
19
Nor is it possible that the ATWACC increases only because the cost of debt increases. This was the
mistake in the Solomon paper that Dr. Booth cites.
20
I will not repeat that discussion here, since it is already part of the record in this proceeding. But for
completeness, I will incorporate that discussion herein by reference. It, like the appendices to this reply
evidence, should be considered to be part of my reply in this proceeding.
20
WRITTEN REPLY EVIDENCE OF
A. LAWRENCE KOLBE
1
2
3.
Q24. What does the Dr. Booth say regarding non-tax costs of debt and the flat ATWACC
3
4
Not Ignoring Non-Tax Costs of Debt
equation?
A24.
At p. 23, he notes that the NEB in its RH-2-94 Decision discussed the formal tax-based
5
models of capital structure and held that they ignored the non-tax costs of debt. He also
6
states that the cost of equity equation derived from the flat ATWACC equation gives the
7
highest possible leverage adjustment (a claim rebutted in the previous part of this section)
8
“because the debt cost is after tax and there are no explicit offsetting costs in the model,
9
yet the WACC is somehow held constant.”21
10
Q25. What is your reaction to these statements?
11
A25.
The most basic response is that, as noted above in part F.1 of this section, my procedures
12
do not rely on any of the purely tax-based models. Detailed discussions of the non-tax
13
costs of debt in my direct evidence (especially in Appendix B) underlie the findings on
14
which I rely. Therefore, the Board’s quoted comment from RH-2-94 simply does not apply
15
to my evidence.
16
I would add, however, that Dr. Booth’s discussion in this section does not
17
demonstrate a clear understanding of the Miller model. Prof. Miller’s insight was that if
18
personal and corporate tax rates had the right values, the personal tax disadvantage to debt
19
could fully offset its corporate tax advantage. The Miller (1977) version of the flat
20
ATWACC equation is intrinsically a model of no net tax advantage. The ATWACC is
21
held constant precisely because debt has no net tax advantage. There is no need for risky
21
Booth Evidence, p. 97.
21
WRITTEN REPLY EVIDENCE OF
A. LAWRENCE KOLBE
1
debt to achieve this outcome, only for the tax rate conditions described in my Appendix B,
2
right before my discussion of the non-tax effects of debt. Dr. Booth’s statement seems to
3
say that the Miller model could not achieve this result without the assumption that debt is
4
risky. If so, that is plainly wrong.
5
This confusion arises again in footnote 28, p. 94, where Dr. Booth suggests that the
6
fact that a flat ATWACC requires in part a personal tax offset to debt’s corporate tax
7
advantage is somehow undercut by the fact that “Union Gas indicated that it has never
8
commissioned a study of its marginal investor’s tax rate and uses the conventional after tax
9
WACC in capital budgeting.” As my Appendix B explains at p. B-23, citing Prof.
10
Taggart’s paper, the WACC is always calculated with the corporate tax rate alone. The
11
effect of personal taxes shows up in the way the market sets the costs of equity and debt.
12
No textbook of which I’m aware teaches business students that they need to estimate the
13
marginal personal tax rate in order to calculate the WACC.22
14
15
4.
Q26. What does Dr. Booth say about regulatory risk in the context of capital structure
16
17
Regulation Does Not and Should Not Work as Claimed
principles?
A26.
18
19
At p. 87, he asserts,
In the example, the equity is obviously riskier at . . . a market to book of
1.36X than it is at . . . a market to book of 1.0X. This is because it is highly
22
Brealey, Myers and Allen, Principles of Corporate Finance, 8th edition, Irwin/McGraw-Hill (2006)
makes this point explicitly at p. 527. As to whether this undercuts a flat WACC, on the very same page
the book states, “most financial managers don’t fine-tune their companies’ debt ratios, and they don’t
rebalance financing to keep debt ratios strictly constant. In effect they assume that a plot of the WACC
for different debt ratios is ‘flat’ over a reasonable range of modest leverage.”
22
WRITTEN REPLY EVIDENCE OF
A. LAWRENCE KOLBE
1
2
3
4
5
6
7
8
9
10
unlikely that the regulator will cut the allowed ROE when the stock is
trading at book value. In contrast, the higher the market to book ratio the
more likely the regulator will cut the allowed ROE and thus the riskier the
stock. . . . Drs Kolbe and Vilbert would have us believe that the equity in
a regulated firm is less risky when it is trading at a market to book well
above 1.0, since the debt ratio is lower. As a result they would increase the
equity cost when applied to the regulated firm. I don’t accept this since it
defies common sense.
Q27. Do you agree?
A27.
No, for several reasons. First, for this passage to be true, either investors would have to
11
be extremely short-sighted, or regulators would have to be about to behave in a way that
12
investors do not foresee. If investors predicted that regulators were about to start managing
13
the ratesetting process to focus on the market-to-book ratio, and particularly on a market-
14
to-book ratio of one, the market-to-book ratio would be close to one already.23
15
Of course, this also means that the market-to-book ratios that Dr. Booth cites
16
themselves imply that regulators have not been basing their rate of return decisions on this
17
fact up to now. Otherwise, investors would have predicted the coming action and bid the
18
market-to-book ratios down in advance.24 Thus, there is no evidence that regulators do
19
behave in the way this passage asserts they do, and as the answer in Exhibit J2.01, part (c),
20
in this proceeding shows, there is no reason that they should.
23
The inherent circularity of such a process would make regulation based on the market-to-book ratio
infeasible, even if we were still able to believe the market-to-book ratio test worked. See Exhibits
J2.01, part (c), and J2.02, part (c), in this proceeding.
24
Also, as noted in Section III of my reply evidence, below, the market-to-book ratio levels that Dr. Booth
cites imply unreasonably low values for the cost of equity even if investors never expect regulators to
behave in the way that this passage suggests investors should fear.
23
WRITTEN REPLY EVIDENCE OF
A. LAWRENCE KOLBE
1
Q28. What about the part of the quotation that says you are incorrect to view the firm as
2
3
more risky at lower market-to-book ratios?
A28.
First, that is not what I say. I (and a great deal of financial research) say that for the same
4
business risk, the firm’s equity will be less risky at a higher market-value equity ratio. The
5
market-value capital structure is the cause of the risk difference, which need not correspond
6
to differences the equity market-to-book ratio.25 It is Dr. Booth who turns this into a
7
statement about the market-to-book ratio, not me.
8
characterization of the way rate regulation works were right, and even if investors were too
9
blind to see the cut coming, the reason the utility’s risk would be higher in such
10
circumstances would have nothing to do with financial leverage, it would be based on
11
regulatory risk. That says nothing about the impact of market-value capital structure on
12
the level of financial risk. Dr. Booth’s conclusion is simply a non sequitur.
13
G.
14
FINANCIAL RISK DEPENDS ON MARKET VALUES, NOT BOOK VALUES
Q29. Apart from the market-to-book ratio passage cited above, on what does Dr. Booth’s
15
16
Second, even if Dr. Booth’s
discussion say financial risk depends?
A29.
Dr. Booth addresses this issue in several places. Specifically, at pp. 21-23 he argues that
17
the accounting relationships between book capital structure and either book net income or
18
the variability of book net income provide all the information the Board needs to set capital
19
structure. He makes a related statement on p. 89, where he says that the accounting
20
relationship among the return on equity, the cost of debt and capital structure says
25
For example, a company with a market-to-book ratio of 1.0 and a 60 percent market equity-to-value
ratio will have less financial risk than a company with the same business risk, less book equity, a
market-to-book ratio of 1.5, and a 50 percent market equity-to-value ratio.
24
WRITTEN REPLY EVIDENCE OF
A. LAWRENCE KOLBE
1
“absolutely nothing about how the stock market reacts to this financial risk, that is, how
2
market values change, or how the cost of equity changes as the firm uses debt.” (Italicized,
3
boldface emphasis in the original.)
4
Q30. What is your reaction to these statements?
5
A30.
It is hard to see how the accounting relationship among net income and interest expense
6
can have no impact at all on the stock market, since future equity cash flows will be
7
influenced by that interaction in future years. For example, is Dr. Booth claiming that the
8
market will take no notice if a company has so much debt that it will regularly have trouble
9
making its interest payments out of operating income (i.e., will have regularly have
10
negative net income)? The lack of a one-to-one correspondence between accounting values
11
and market values does not mean quantities measured by accounting values have no
12
relationship at all with quantities measured by market values.
13
Nor is a view that financial risk is solely a function of book income statements and
14
balance sheets consistent with either the financial literature or everyday experience.26 The
15
cost of equity depends on the risks equityholders bear. Financial risk is the extra risk that
16
equityholders bear when firms issue debt. The cost of equity is measured in the market, not
17
on the books. Therefore, a proper analysis of the impact of financial risk on the returns
18
equityholders require must measure that risk in the market, using market values. The
26
Recall the example in Section II of my written evidence of how the variability of the market value of
the equity in a dwelling varied with the size of the mortgage.
25
WRITTEN REPLY EVIDENCE OF
A. LAWRENCE KOLBE
1
contrary view, that financial risk is a book phenomenon, is plainly contrary to standard
2
financial theory and practice.27
3
Q31. What other comments does the Dr. Booth make on this topic?
4
A31.
5
6
7
8
9
He says on pp. 87-88,
Second and more significant, the financial leverage risk premium stems
from the imposition of fixed interest charges. That is, prior to receiving
their equity return the firm has to pay these interest charges. This risk does
not change as the market value of the firm changes; it only changes as the
book debt equity ratio changes.
10
Q32. Is this correct?
11
A32.
12
No, it is wrong. The passage appears to forget that equityholders’ return comes in the form
of capital gains as well as current dividends.
13
Changes in market values are a major source of the realized risks equityholders
14
face. If a company has a market value of $10 million in equity and $10 million in debt, or
15
$20 million total, and if the value of the enterprise falls 10 percent to $18 million, equity
27
Among other problems, book relationships cannot contain enough information properly to assess
required returns in the market, since the market looks ahead, not merely at current and past book
relationships.
Brealey, Myers and Allen, op. cit., makes the same point in a particularly colourful way. At pp.
504-05, it provides market-value and book-value balance sheets for the “Sangria Corporation” and
calculates the WACC using the market-value balance sheet. Then it says,
Why did we show the book balance sheet? Only so you could draw a big X through it. Do
so now.
When estimating the weighted-average cost of capital, you are not interested in past
investments but in current values and expectations for the future. Sangria’s true debt ratio is
not 50 percent, the book ratio, but 40 percent [the market-value debt ratio], because its assets
are worth $1.250 million [which is their market, not book, value].
26
WRITTEN REPLY EVIDENCE OF
A. LAWRENCE KOLBE
1
absorbs the (vast majority of)28 the $2 million loss. The rate of return on equity is -20
2
percent (-$2 million/$10 million). However, suppose another, otherwise identical company
3
has a market value of equity of $15 million and of debt of $5 million, for the same total
4
enterprise value of $20 million. Suppose the same forces make the value of that enterprise
5
decline by 10 percent to $18 million, also. In this case, equity “only” must absorb a 13.3
6
percent loss (-$2 million/$15 million), not 20 percent. Note also that the impact on the
7
equityholders of the two firms is unaffected by the firms’ book capital structures, which
8
we could assume to be identical without changing the example at all.
9
In short, changes in the market-value capital structure directly affect the sensitivity
10
of the equity return to changes in the value of the enterprise. The level of the financial risk
11
equityholders bear therefore depends directly on market values.29
12
Q33. Does Dr. Booth claim to offer any proof for his assertion?
13
A33.
14
15
16
17
18
19
He offers a quotation from Standard & Poor’s at p. 88:
Similarly ratios using market value of a company's equity in calculations of
leverage are given limited weight as analytical tools. The stock market
emphasises growth prospects and has a short time horizon; it is influenced
by changes in alternative investment opportunities and can be very volatile.
A company's ability to service its debt is not affected directly by such
factors [emphasis added in the Booth Evidence].
28
As the market-value equity ratio declines, decreases in the value of the enterprise begin to be absorbed
in part by debt, since there is less equity left to provide a cushion. However, the majority of the risk
falls on equity, and adding an adjustment for the impact on debt would complicate the example without
changing the conclusion.
29
Note that this conclusion holds true regardless of the “true” model or models of stock prices. Whatever
the reason the market value of a firm’s assets might decline, the impact of that decline will fall
(primarily) on equity, not debt. The greater the proportion of debt in the market value of the assets, the
bigger the proportionate fall in the market value of the equity.
27
WRITTEN REPLY EVIDENCE OF
A. LAWRENCE KOLBE
1
His next paragraph interprets this passage to imply,
2
3
4
5
6
7
That is, S&P is basically saying book value leverage is important, when it
is assessing the default or credit risk in debt, whereas market values don't
count, or at least don't count as much. If it is book values and interest
payments that affect credit risk and the cost of debt then this is the risk that
also affects utility equity investors.
Q34. Is this a reasonable interpretation in the context of whether market values affect the
8
9
amount of financial risk equityholders bear?
A34.
No, the quotation could hardly be less relevant. First, the quoted passage does not say
10
anything at all about book-value leverage. That is Dr. Booth’s addition (or at the very
11
least, it is based on material he does not quote with the rest of the passage).
12
Second, and more fundamentally, S&P in the part of this passage that Dr. Booth
13
emphasizes is focusing on whether bondholders can count on the company to service debt,
14
not on the financial risk equityholders bear. The market-value volatility of which the
15
passage speaks definitely will affect equityholders, whether or not the company can service
16
its debt. The passage expresses no opinion at all about whether the amount that equity
17
would be affected would be different if the market debt-equity ratio were different. That
18
topic never comes up.
19
Therefore, the passage has no meaning at all for the issue of whether market value
20
capital structures affect the degree of financial risk equityholders bear.
21
Q35. Does Dr. Booth offer any other support for his views on the basis of financial risk?
22
A35.
He appears to try to rely again on the AEUB. In particular, at p. 81, he says
28
WRITTEN REPLY EVIDENCE OF
A. LAWRENCE KOLBE
1
2
3
However . . . the essential point is that the correct financing weights for a regulated
firm should be the regulated capital structure weights, not the market value weights.
...
4
5
6
The Alberta EUB has directly addressed this question on a number of occasions.
For example, in connection with comparable earnings testimony the EUB stated
(Generic Cost of Capital Decision U-200452, page 24)
7
8
9
“The Board considers that the application of a market required
return (i.e. required earnings on market value) to a book value rate
base is appropriate in the context of regulated utilities.”
10
11
That is, you estimate a market opportunity cost, such as that from the CAPM, and
apply it to book values, not market values as is the assumption in WACC.
12
Q36. Do you agree?
13
A36.
No, Dr. Booth here treats two quite distinct concepts as though they were the same, and
14
thereby misstates the import of the AEUB passage he quotes. One concept is how to
15
estimate the cost of capital, by use of market data (as in the Capital Asset Pricing Model,
16
the Empirical Capital Asset Pricing Model, or the Discounted Cash Flow approach) or by
17
use of book data (as in the Comparable Earnings approach). The other concept is to what
18
rate base the resulting rate of return should be applied, a market-value rate base or a book-
19
value rate base.
20
Dr. Vilbert and I endorse the application of market-value-based estimates of the
21
cost of capital to a book-value rate base. Nothing we do conflicts with this standard in any
22
way.30 However, we do state that the market evidence on the cost of equity must be
23
interpreted correctly. The level of financial risk inherent in any market-based estimate of
24
the cost of equity depends on the market-value capital structure of the sample company,
25
not the book-value capital structure of that company. That principle has been part of the
30
See also Appendix R-B to this reply evidence, part B.
29
WRITTEN REPLY EVIDENCE OF
A. LAWRENCE KOLBE
1
literature on this topic from the beginning. The entire debate over the shape of the
2
ATWACC curve is not about whether the level of financial risk in an estimate of the cost
3
of equity depends on market-value capital structure, but only about how fast the cost of
4
equity goes up as the market-value debt ratio goes down.31
5
H.
6
Q37. What does Dr. Booth say about the fact that your procedures focus on the deemed
7
8
PRINCIPLES DO WORK FOR DEEMED EQUITY RATIO ANALYSIS
equity ratio, not the ATWACC and cost of equity on which he focuses?
A37.
At pp. 98-100, he shows that the deemed equity ratio calculations on which my
9
recommendations rely are consistent with the finding that debt does not have a material
10
effect on the value of the firm within the normal range of capital structures for an industry.
11
This point is not in dispute, nor, given the views stated in my evidence, should it be
12
surprising to anyone.
13
He also demonstrates that the deemed equity ratio results change when the WACC
14
inputs change, which again should be no surprise.32 He notes that if the WACC is flat, the
15
allowed return on equity is inversely related to the market debt rate, which in part depends
16
on the maturity of the company’s debt. But this is economically appropriate. Long-term
17
bondholders shoulder the risk of future changes in interest rates, which is why long-term
31
See Appendix B to my direct evidence, including the discussion of the formal tax-based models (pp.
B-13 to B-23) and the more general discussion at pp. B-6 to B-13.
32
He does characterize a change of ½ percentage point in the overall WACC as a “small” difference (p.
99). I would just note that at a ½ percentage point change in the WACC corresponds to nearly a 1½
percentage point change in the cost of equity at a 35 percent equity ratio, and to about a 1¼ percentage
point change in the cost of equity at a 40 percent equity ratio. I would not characterize such changes
as “small” in a regulatory setting.
30
WRITTEN REPLY EVIDENCE OF
A. LAWRENCE KOLBE
1
bond rates are higher than short-term rates. If a company finances with short-term debt,
2
its equityholders shoulder that risk instead, and their cost of equity is in fact higher.
3
Dr. Booth then essentially says that since he has found all the problems discussed
4
above with the ATWACC-cost of equity formulation of the underlying principles, the
5
Board should disregard the deemed equity ratio formulation, too (again citing the AEUB
6
decision).
7
Q38. What is your reaction?
8
A38.
I would turn the point around. None of the criticisms of my procedures in Dr. Booth’s
9
evidence have economic merit, nor are they supported by relevant citations to the economic
10
literature. Some of them flatly misstate my evidence. Given this, I would submit that Dr.
11
Booth’s comments on capital structure principles should be disregarded entirely.
12
III.
MARKET-TO-BOOK TEST
13
Q39. What topics does this section cover?
14
A39.
Dr. Booth at various places (e.g., pp. 11-12, 73, 81, and 100-01) draws the conclusion that
15
if the market-to-book ratios of utilities are materially above one, their returns on equity are
16
excessive. In the process, he cites (p. 101) as supporting his view a 1984 book of mine
17
(with two co-authors).
18
Q40. Does this mean you agree with him?
31
WRITTEN REPLY EVIDENCE OF
A. LAWRENCE KOLBE
1
A40.
No. Since 1984, events in the stock market and a growing body of financial research have
2
required me to reach a different conclusion today than was warranted in 1984. As noted
3
in Section II above, the answer to interrogatories in Union’s Exhibits J2.01 and J2.02,
4
particularly part (c) of the former, show that the market-to-book ratio test of the adequacy
5
of a utility’s returns is inconsistent with any reasonable value for its cost of equity and has
6
been disproved by actual market behavior. This conclusion is consistent with a growing
7
body of economic research on stock prices, cited in the first of these responses.
8
I will not repeat that discussion here, since it is part of this record, but I will
9
incorporate that material herein by reference. It should be considered to be part of my
10
reply evidence.
11
Q41. Can you sum up the current state of knowledge?
12
A41.
Yes. The basic valuation tool in corporate finance is the present value formula. That
13
formula works well in certain applications (e.g., valuing fixed-income securities).
14
However, research now reveals that it does not correctly explain stock price levels and
15
volatility. The finance profession has yet to discover the model or models that actually
16
explain stock prices and stock price volatility. No one knows why these quantities are not
17
in accord with the predictions that would follow from our standard toolkit.
18
We do know, for reasons explained in the response in Union Exhibit J2.01, part (c),
19
that the conclusions that follow from assuming the present value formula works at today’s
20
utility market-to-book ratios are unsupportable even under conditions that otherwise are
21
completely favorable to the market-to-book test. A utility’s cost of equity would have to
32
WRITTEN REPLY EVIDENCE OF
A. LAWRENCE KOLBE
1
be very low, below the risk-free interest rate and possibly negative, for the present value
2
formula to explain these stock price levels.33
3
According, regulators cannot rely on the market-to-book test to assess whether
4
utility shareholders expect to earn more than the cost of capital.
5
IV.
OTHER TOPICS
6
Q42. What topics does this section cover?
7
A42.
There are two: a mistaken standard for the adequacy of a regulated company’s return,
8
based on a partial quotation from a legal opinion, and some mistaken statements regarding
9
Dr. Vilbert’s evidence.
10
A.
ECONOMICALLY INCORRECT RETURN ADEQUACY STANDARD
11
Q43. What return adequacy standard does Dr. Booth advance?
12
A43.
13
Dr. Booth discusses his interpretation of the relevant legal standards at pp.8-13 and 24. As
part of that discussion, Dr. Booth states on p. 24 that
14
15
16
17
18
19
20
Ultimately the litmus test of whether a board has “got it right” is whether
the regulated company can access capital on reasonable terms. If, for
example, a common equity ratio is inadequate then the stock market will
take note of the increased financial risk and make it difficult for the
regulated firm to access capital on reasonable terms. In Federal Power
Commission et al v. Hope Natural Gas Co. [320 US 591, 1944], the United
States Supreme Court decided that a fair return
33
Note that this finding refutes Dr. Booth’s response to Union’s Question L2.9, which implies that the
market-to-book test works for utilities even if the market price of growth companies cannot be trusted
(unless Dr. Booth is prepared to state that the cost of utility equity is below the risk-free rate).
33
WRITTEN REPLY EVIDENCE OF
A. LAWRENCE KOLBE
1
2
3
“should be sufficient to assure confidence in the financial
integrity of the enterprise so as to maintain its credit and to
attract capital.”
4
5
6
7
8
Although the Hope “financial integrity” criteria flows from considering a
fair return it applies equally to the deemed common equity ratio. In my
judgment an appropriate common equity ratio is one which, in conjunction
with the allowed return, allows a regulated company to maintain its credit
and attract capital.
9
That is, Dr. Booth appears to assert as an economic test that the Board can tell whether the
10
overall return is acceptable by whether the company can raise capital.
11
Q44. Do you agree?
12
A44.
No, for two reasons.
13
First, while I am not an attorney, I have studied the Hope decision from an
14
economic standpoint. Dr. Booth’s interpretation is in conflict with the plain language of
15
the decision. Specifically, recall that the oft-cited passage that Dr. Booth quotes so briefly
16
reads more completely as,34
17
18
19
20
21
22
23
24
25
26
... the investor interest has a legitimate concern with the financial integrity
of the company whose rates are being regulated. From the investor or
company point of view it is important that there be enough revenue not only
for operating expenses but also for the capital costs of the business. These
include service on the debt and dividends on the stock. [Citation omitted.]
By that standard the return to the equity owner should be commensurate
with returns on investments in other enterprises having corresponding risks.
That return, moreover, should be sufficient to assure confidence in the
financial integrity of the enterprise, so as to maintain its credit and to attract
capital.
34
Hope at 603. Underlined emphasis added.
34
WRITTEN REPLY EVIDENCE OF
A. LAWRENCE KOLBE
1
Thus, the plain language of the decision includes something more than the statement on
2
which Dr. Booth focuses: in addition to capital attraction, it calls for “commensurate”
3
returns.
4
As an economist, I would note that regardless of what the quotation may mean
5
legally, more than mere capital attraction is required economically. That is, a rate-
6
regulated company can, for a time, continue to attract new capital while being materially
7
shortchanged in its required return, if it relies on the credit capacity of pre-existing assets
8
to do so. Yet in such circumstances, the company is at the same time flunking any
9
economically reasonable definition of “commensurate” returns. Therefore, Dr. Booth is
10
wrong to suggest that the mere ability to attract new capital automatically implies that the
11
return on equity is economically fair and reasonable.
12
B.
MISTAKEN COMMENTS ON DR. VILBERT’S EVIDENCE
13
Q45. What topics do you address here?
14
A45.
There are two: an overstatement of the risk implications of the fact that some of Dr.
15
Vilbert’s sample companies have unregulated components, and misinterpretations Dr.
16
Booth makes of Dr. Vilbert’s views and analyses, which imply a theory of financial risk
17
that is completely outside the boundaries of the financial literature.
35
WRITTEN REPLY EVIDENCE OF
A. LAWRENCE KOLBE
1
2
1.
Q46. What does Dr. Booth say about the risk implications of having utility holding
3
4
Risk Implications of Holding Companies in Sample Groups
companies (“UHCs”) in Dr. Vilbert’s sample groups?
A46.
5
At p. 31, he states his view that non-regulated lines of business increase the risk of Dr.
Vilbert’s Canadian sample group of UHCs:
6
7
8
9
10
11
The core of the testimony of Dr. Vilbert is to estimate the WACC from a
sample of UHCs and use them as a proxy for Union Gas. As I have
demonstrated above there is little doubt that the Canadian UHCs are riskier
than their underlying regulated assets due to their periodic misadventures
in non-regulated areas. This UHC risk will be reflected in their higher
WACC.
12
Q47. Please comment.
13
A47.
I would first note that Dr. Booth’s “demonstration” of greater risk comes in the section
14
before this quotation and is based on book evidence, not market evidence. But the cost of
15
capital is determined in capital markets, and book evidence cannot capture market’s full
16
reaction to events. Market returns look forward, not backward, and are much more volatile
17
than book returns.
18
Second, cost of capital estimates always provide “noisy” information about the
19
“signal” of the true, underlying cost of capital. Even if it were uncontested that the
20
remaining unregulated lines of business were riskier than the regulated lines, their
21
estimated costs of capital may be lower, especially if, as Dr. Booth suggests in his
22
quotation, the investments have been subject to “misadventures.” Investments in trouble
23
often end up with cost of capital estimates with a material downward bias, since unique
24
news about the status of the investment adds so much “noise” that the “signal” is simply
25
lost.
36
WRITTEN REPLY EVIDENCE OF
A. LAWRENCE KOLBE
1
Third, Dr. Booth paints Dr. Vilbert’s samples with too broad a brush. The average
2
revenue from regulated activities in Dr. Vilbert’s Canadian sample is above 85 percent,
3
while the percentage of regulated revenues for his LDC sample is about 80 percent for the
4
full sample and above 95 percent for the subsample on which he and I place the most
5
reliance.
6
The last of these samples, particularly, is as close to the ideal of publicly traded
7
pure plays as you are likely to find in any practical application. The LDC sample not only
8
does not have material unregulated lines of business, it is also more homogenous than the
9
Canadian sample. As a result, there is less “noise” to mask the “signal” of its true cost of
10
capital.
11
Finally, Dr. Carpenter, at p. 3 of his evidence, states that he finds Union to be
12
riskier than the LDC sample, a conclusion reinforced by my own understanding the ways
13
competition can affect companies still subject to regulation. The high quality of the LDC
14
sample makes this a particularly useful benchmark.
15
For these reasons, Dr. Booth’s views on UHC risk do not undercut our conclusions
16
on the appropriate deemed equity ratio for Union.
17
2.
Unique Theory of Financial Risk
18
Q48. What aspect of Dr. Booth’s evidence do you address here?
19
A48.
On p. 11, Dr. Booth provides a table and an average of some of Dr. Vilbert’s equity cost
20
estimates for the Canadian utility sample. Dr. Booth lists the average for Dr. Vilbert’s
21
CAPM estimates as 7.15 percent and the DCF numbers as 7.74 and 8.24 percent. On pp.
22
11-12, he interprets these numbers as implying that allowed returns are in excess of
37
WRITTEN REPLY EVIDENCE OF
A. LAWRENCE KOLBE
1
utilities’ cost of capital as measured by himself and Dr. Vilbert. There are two material
2
problems with this discussion.
3
First, Dr. Booth implies that Dr. Vilbert agrees with him that allowed returns are
4
too high: (“If Dr. Vilbert, myself and security analysts are correct and allowed ROEs
5
exceed the investors’ fair rate of return or cost of equity capital, . . .”). This plainly is not
6
Dr. Vilbert’s view, and it is flatly misleading to suggest otherwise, for a number of
7
reasons.35
8
The most fundamental of these reasons is that Dr. Booth’s interpretation of Dr.
9
Vilbert’s cost of equity estimates simply ignores the huge difference in financial risk
10
between the sample companies at their market-value capital structures and Union at its
11
regulatory capital structure, a difference that both Dr. Vilbert and I are at pains to
12
recognize. Dr. Booth argues elsewhere that the ATWACC is not precisely flat, but here
13
he effectively argues that the cost of equity is precisely flat even at very different capital
14
structures, which implies that equity investors demand no compensation for financial risk.
15
No existing theory of capital structure in the literature, even the Modigliani-Miller 1963
16
model, accepts such an extreme result. Here, Dr. Booth has stepped completely outside the
17
boundaries of the financial literature.
35
For example, even ignoring the capital structure adjustments Dr. Vilbert and I make, p. 75 of Dr.
Vilbert’s evidence demonstrates that Dr. Vilbert focuses on the middle ECAPM model as providing the
best estimate of the cost of equity, not the CAPM or DCF models Dr. Booth chooses to highlight. That
model provides a value of 8.2 percent for the Canadian sample, rather than the 7.15 percent from the
CAPM cited by Dr. Booth. Additionally, the 7.15 percent figure is wrong. Dr. Vilbert mistakenly
agreed that the table presented in Exhibit J2.18 in this proceeding is a correct description of his equity
cost estimates for Canadian utility holding companies. Unfortunately, the last three rows of column
three in the table are in error. Specifically, the figures cited for TransCanada, Fortis, GMI, and
consequently the Average are in error. As can be seen from Dr. Vilbert’s Workpaper #1 to Table No.
MJV-11, the cost of equity estimates for TransCanada, Fortis, and GMI are 7.6, 7.8, and 6.0 percent for
a sample average of 7.3 percent (7.7 percent without Emera, whose estimates Dr. Vilbert does not rely
on in the risk positioning model).
38
WRITTEN REPLY EVIDENCE OF
A. LAWRENCE KOLBE
1
Q49. Does this complete your reply evidence?
2
A49.
Yes.
39
WRITTEN REPLY EVIDENCE OF
A. LAWRENCE KOLBE
1
APPENDIX R-A: DETAILED DISCUSSION OF DR. BOOTH’S EXAMPLE
2
Q50. What numerical example do you discuss in this section?
3
A50.
The example begins on p. 70 of Dr. Booth’s evidence and serves general expository
4
purposes. The part I need to address begins on p. 74, where debt is added. It ignores taxes
5
for simplicity and assumes investors price a regulated firm’s equity as a perpetuity, that is,
6
as the present value of a constant annual sum of money that investors believe will come in
7
at the same level forever. It sets the amount of money that equity gets as the difference
8
between the overall return on the assets and the amount that gets paid out for interest
9
expense. This leads to the formula,
Value of
Equity
10
11
12
13
=
(Return On × (Asset !
(Cost of × (Debt
Investment) Value)
Debt)
Value)
Cost of Equity
or in Dr. Booth’s notation (p. 75),
14
15
E
=
ROI × A ! Kb x B
Ke
16
where asset value, A, equals the book value of utility assets, which initially equals the sum
17
of the market values of equity, E, plus debt, B. It initially postulates that the return on
18
investment, ROI, is 10 percent, the value of the assets, A, is $10 million, the cost of debt,
19
Kb, is 5 percent, the amount of debt, B, is $5 million, and the cost of equity, Ke, is 15
20
percent. This gives a value of equity equal to,36
36
See pp. 74-75 of the Booth Evidence, which reaches this result but which need not spell out the steps
in this level of detail because some of them were explained earlier.
WRITTEN REPLY EVIDENCE OF
A. LAWRENCE KOLBE
=
ROI × A ! Kb x B
Ke
3
4
=
0.10 × $10 million ! 0.05 × $5 million
0.15
5
6
=
$1 million ! $ 0.25 million
0.15
7
=
$5 million
1
2
E
=
$0.75 million
0.15
8
If I adopt the convention that the Booth Evidence uses of “mm” for million, the overall
9
weighted-average cost of capital (“WACC”) is the weighted average of the costs of equity
10
11
and debt, or
WACC=
Ke × [E/(E+B)]
+
Kb × [B/(E+B)]
12
=
0.15 × ($5mm/$10mm) + 0.05 × ($5mm/$10mm)
13
=
0.15 × 0.5 + 0.05 × 0.5 = 0.10 = 10 percent.
14
Starting on pp. 77-8, Dr. Booth postulates that a change in regulatory rules drops
15
the cost of equity from 15 percent to 11 percent. It does not, however, explain directly at
16
what capital structure this 11 percent is supposed to obtain, and alternative interpretations
17
of that key fact would produce different criticisms of his example. In what follows, I take
18
my lead from the statement at p. 78,
19
20
21
22
23
The obvious thing for the regulator to do is simply get expert opinion
estimating the equity cost, and if this is unbiased, notice and cut the allowed
ROE to 11%. The equity value will then return to $5mm and the
stockholders will continue to earn their fair return on their $5mm
investment.
R-A-2
WRITTEN REPLY EVIDENCE OF
A. LAWRENCE KOLBE
1
I therefore interpret the 11 percent to be the cost of equity that obtains at a market-value
2
capital structure of 50-50 equity-debt. In that case, the true new WACC in his example is
3
True WACC =
Ke × [E/(E+B)]
+
Kb × [B/(E+B)]
4
=
0.11 × ($5mm/$10mm) + 0.05 × ($5mm/$10mm)
5
=
0.11 × 0.5 + 0.05 × 0.5 = 0.08 = 8 percent.
6
From this point through p. 71, line 4, his example postulates a series of partial
7
adjustments by regulators that are internally inconsistent with the underlying assumptions
8
of the model. It also mischaracterizes the way capital markets work and the results that
9
would follow if regulators used the WACC to set the overall returns.
10
Q51. Please explain what you mean in the previous paragraph.
11
A51.
There are two main problems with this example. The first, which is the less important
12
problem, is that it assumes investors price the stock as though the current cash flow would
13
be unchanged forever, yet the example imagines a series of annual changes by regulators
14
that gradually converge on a new equilibrium WACC equal to 8 percent. It is hard to
15
imagine that investors would not catch on that their valuation model was woefully
16
inadequate at this point, since each year of the adjustment they would bear a major capital
17
loss, cumulatively undoing the major gain they received when the cost of equity initially
18
fell from 15 to 11 percent.
19
More fundamental is the failure to recognize that if the imagined gradual
20
adjustment process nonetheless took place as postulated, the actual cost of equity would
21
be below 11 percent until equilibrium was restored. The example’s assumption that the
R-A-3
WRITTEN REPLY EVIDENCE OF
A. LAWRENCE KOLBE
1
cost of equity would remain at 11 percent regardless of the market-value capital structure
2
of the company in questions is false.
3
Q52. Please show how this affects the example’s steps.
4
A52.
5
First (on p. 78) it calculates the new value of equity on the assumption that equity investors
expect the same $0.75mm forever, but now discount it at 11 percent:
=
ROI × A ! Kb × B
Ke
8
9
=
0.10 × $10mm ! 0.05 × $5mm
0.11
10
11
=
$1mm ! $ 0.25mm
0.11
12
=
$6.818mm
6
7
13
E
=
$0.75mm
0.11
It calculates the new WACC using this value of equity, which it finds to be 8.46 percent:
14
Booth WACC =
Ke × [E/(E+B)]
+
Kb × [B/(E+B)]
15
=
0.11 × ($6.8mm/$11.8mm) + 0.05 × ($5mm/$11.8mm)
16
=
0.11 × 0.577 + 0.05 × 0.423 = 0.0846 = 8.46 percent.
17
It imagines regulators setting the ROI equal to 8.46 percent for the next year, which drops
18
the “perpetual” equity cash flow to (ROI x A - Kb x B) = (0.0846 x $10mm - 0.05 x $5mm)
19
= $0.596mm. That produces a new, lower value of equity ($5.42mm, on p. 73), a new,
20
lower calculated WACC (8.12%, on p. 80), and so on, until eventually the postulated true
21
WACC of 8 percent is reached.
R-A-4
WRITTEN REPLY EVIDENCE OF
A. LAWRENCE KOLBE
1
The key mistake in logic in this example takes place at p. 78, where Dr. Booth’s
2
example postulates that the first estimation of the new WACC is done “[a]ssuming there
3
is no bias to estimating the equity cost at 11%.” However, if there were no bias in
4
estimating the equity cost, the cost of equity actually estimated would be consistent with
5
the true market WACC, not some mistaken version. The 11 percent cost of equity is correct
6
in Dr. Booth’s example at a 50-50 market-value capital structure. It would be different at
7
a different market-value capital structure. In particular, if for whatever reason the market
8
value of equity were initially the $6.818mm the example supposes, the true cost of equity
9
would be:
=
True WACC ! Kb × [B/(E+B)]
[E/(E+B)]
12
13
=
0.08 ! 0.05 × [$5mm/($6.8mm+$5mm)]
[$6.8mm/($6.8mm+$5mm)]
14
15
=
0.08 ! 0.05 × [0.423]
[0.577]
16
17
=
0.08 ! 0.0211
0.0588
18
=
0.102 = 10.2 percent
True Ke
10
11
19
In that case, in the world of Dr. Booth’s example, the estimated WACC would equal 8
20
percent the first time out,37 and regulators would converge on the right answer after one
37
That is, if the cost of equity is estimated without bias, Estimated WACC = True Ke × [E/(E+B)] + Kb
× [B/(E+B)] = 0.102 × 0.577 + 0.05 × 0.423 = 0.08 = 8 percent, the true WACC.
R-A-5
WRITTEN REPLY EVIDENCE OF
A. LAWRENCE KOLBE
1
false start, based on the market’s initial misestimation of the new value of equity at $6.818
2
million.38
3
Based on this flawed example, Dr. Booth mischaracterizes the situation in several
4
ways.
5
Q53. In what ways does Dr. Booth mischaracterize the situation?
6
A53.
A fundamental problem is the claim at pp. 78-9 that the example would produce an
7
estimated ATWACC of 8.46 percent the first year. That statement wrongly assumes the
8
market would set the cost of equity at 11 percent despite the fact that the company in
9
question has much less financial risk at a 57.7-42.3 equity-debt ratio than it does at a 50-50
10
equity-debt ratio. The market will recognize the actual level of financial risk, however,
11
whether Dr. Booth’s example does or not. In the world of Dr. Booth’s example, the true
12
cost of equity is 10.2 percent at the postulated 57.7-42.3 equity-debt ratio. If the true cost
13
of equity were estimated without bias, the estimate of the WACC would be 8 percent the
14
first time, not 8.46 percent.
15
Next, Dr. Booth states at p. 80,
16
17
18
19
20
21
22
Although the ROI is reduced from 10% to 8.46%, it is not reduced to the
correct ROI of 8.0%,24 so the equity market value is still $0.42mm higher
than it needs to be. The reason for this is that using market value weights
in the WACC puts greater emphasis on the higher equity cost than the debt
cost. For non-regulated firms this is correct since the objective is to
maintain these market values and create wealth. However, it is totally
incorrect for a regulator who is tasked with awarding fair allowed returns
38
If the market immediately recognized that regulators were about to change the allowed rate of return
to the new true WACC, 8 percent, the market value of equity would fall to $5 million and the cost of
equity if estimated without bias would be the postulated 11 percent immediately, under this valuation
model.
R-A-6
WRITTEN REPLY EVIDENCE OF
A. LAWRENCE KOLBE
1
2
3
4
5
6
7
and implicitly causing market values to drop when allowed ROEs are too
high. By estimating and applying a market based WACC the effect of the
higher allowed ROE is perpetuated by its impact on the higher equity
market value.
8
(Italicized emphasis in the original.)
24
The correct regulated WACC is the average of the debt and equity costs using
regulated book value weights, in this case 50%.
9
This statement contains five errors:
10
11
•
First, the estimated WACC will be 8 percent, not 8.46 percent, because the market
will recognize the effect of financial risk on the cost of equity;
12
13
14
•
Second, that means the greater equity weight in Dr. Booth’s example (57.7 versus
50 percent) is exactly offset by a lower market cost of equity (10.2 versus 11
percent), so there is no net “greater emphasis on the higher equity cost.”
15
16
17
18
19
20
•
Third, the use of market-value weights in the WACC has nothing to do with trying
to “maintain these market values and create wealth,” whether for unregulated or
regulated firms.39 The use of market-value weights is required because the cost of
equity is also based on the level of financial risk implied by market values. Use of
any other weights simply mis-estimates the overall risk and cost of capital of the
firm.
21
22
23
•
Fourth, since the true market-based WACC is the same 8 percent that Dr. Booth’s
example postulates as correct, a regulatory body that did use this method of setting
overall returns would not allow an excessive ROE.
24
25
26
27
28
29
•
Regarding the footnote, the correct cost of equity at the regulated 50-50 book-value
weights is 11 percent, which does reach the correct value of 8 percent for the
WACC, because at those weights the cost of equity is 11 percent. It is higher
because of the greater financial risk at that capital structure. But that is not the only
way to calculate the correct overall return. The same WACC results from use of
the 10.2 percent cost of equity at the market-value equity-debt ratio of 57.7-42.3.
39
This conclusion may be related to Dr. Booth’s erroneous statements that use of the ATWACC tool is
per se an attempt to maximize a company’s market values. The previous part of this appendix explains
the fallacy in that view. There is nothing inconsistent about an unregulated company’s using the
ATWACC to find the projects that offer the most valuable net cash flows in its market, while at the
same time regulators use the ATWACC to limit the cash flows that a regulated firm can earn. The
ATWACC is merely a tool, and the output of using the tool depends on the task for which it is
employed.
R-A-7
WRITTEN REPLY EVIDENCE OF
A. LAWRENCE KOLBE
1
Finally, Dr. Booth concludes at p. 81,
2
3
4
5
6
The basic insight from this discussion is that by using market values in
WACC, the resulting cost of capital is higher than a fair return, since the
higher equity cost is given a greater weight. Further if the firm is a pure
ROE regulated utility it tends to “rubberstamp” the use of market values
and is contrary to “fair and reasonable” regulation.
7
Of course, this conclusion does not follow because an accurate estimate of the new cost of
8
equity would be 10.2 percent, not the postulated 11 percent value that would be appropriate
9
with a 50-50 equity-debt ratio. Additionally, if regulation used a 50-50 equity-debt ratio,
10
the cost of equity that goes with the correctly estimated 8 percent WACC would be the 11
11
percent Dr. Booth’s example originally postulated. There would be no inconsistency and
12
no excess return on equity.
13
Q54. Please sum up.
14
A54.
Dr. Booth’s numerical example postulates regulation according to a recommendation I am
15
not making in this case (i.e., regulation based directly on the overall cost of capital, rather
16
than accepting the Board’s formula rate of return on equity and finding the appropriate
17
deemed equity ratio). It assumes the market will not recognize the true level of financial
18
risk the company bears. In reality, it is Dr. Booth’s example that does not recognize the
19
actual level of financial risk. Were I making the WACC-based regulation recommendation
20
in this example, and were it implemented based on a cost of equity estimated without bias,
21
that recommendation would work without providing any excess compensation to investors
22
in the world of Dr. Booth’s example, even if equityholders initially overestimated the
23
market value of equity. That is, Dr. Booth complains that the WACC approach under the
R-A-8
WRITTEN REPLY EVIDENCE OF
A. LAWRENCE KOLBE
1
conditions of his example will not get to the right answer in one step, while his
2
recommendation would. But if the WACC approach were implemented without bias, it,
3
too, would get to the right answer in one step. Finally, the use of market-value weights to
4
calculate the WACC is equally appropriate for regulators and unregulated firms, and it
5
leads to biases in neither application.
6
Q55. Does this complete Appendix R-A?
7
A55.
Yes, it does.
R-A-9
WRITTEN REPLY EVIDENCE OF
A. LAWRENCE KOLBE
1
APPENDIX R-B: ISSUES RAISED IN AEUB DECISION U99099
2
Q56. What is the purpose of this appendix?
3
A56.
Dr. Booth refers frequently to the AEUB’s Decision U99099 in his attempt to rebut the
4
principles set out in my direct evidence. While I would submit that disagreements on
5
economic principles should be resolved by reference to the economic literature, and while
6
numerous other regulatory decisions reach different conclusions than the AEUB did based
7
on the evidence before it at that time,40 it may nonetheless be helpful to address the
8
AEUB’s concerns directly. This appendix does so.
9
10
Q57. What issues do you wish to discuss?
A57.
There are two such issues: (1) if the ATWACC is flat in the middle range, as I say, why
11
might graphs of ATWACC against the debt ratio show a downward slope, and (2) should
12
book- or market-value capital structure weights be used to calculate the ATWACC for
13
companies regulated on a book-value rate base?
14
A.
MEASURED ATWACC VS. THE DEBT RATIO
15
Q58. What is the answer to the first of these questions?
16
A58.
There are a number of forces that may be responsible for a downward slope of measured
17
ATWACC against the debt ratio. They may be broadly grouped into two categories:
18
factors that distort the comparison, and factors that are left out of the measured ATWACC.
40
See Section II.E in the body of my reply.
WRITTEN REPLY EVIDENCE OF
A. LAWRENCE KOLBE
1
2
1.
Q59. Please describe the first type of force you say is responsible for the downward slope,
3
4
Factors that Distort the Comparison
“factors that distort the comparison.”
A59.
Estimation of the cost of capital is an inherently imprecise exercise. Part of this
5
imprecision is statistical, which may give rise to anomalous comparisons in any particular
6
case, and part is due to the inevitable shortfall of a real sample from the ideal sample of
7
“pure plays” identical to the company in question. This is particularly true of Dr. Vilbert’s
8
Canadian sample, which perforce includes companies in different lines of business entirely,
9
but it is a general feature of cost of capital estimation. Therefore the sample companies in
10
reality will have somewhat different ATWACCs not because of capital structure, but
11
because of differences in business risk.
12
All else equal, less business risk means the broad middle range of capital structures
13
over which the ATWACC is constant will contain more debt on average, as illustrated in
14
Figure B-3 in the Appendix B to my direct evidence. This in turn will result in a negative
15
correlation between measured ATWACC and the debt ratio, not because more debt lowers
16
the ATWACC, but because a lower ATWACC tends to lead to more use of debt. That is,
17
the negative correlation may be real, but the cause differences in business risk rather than
18
a material tax advantage to debt.
R-B-2
WRITTEN REPLY EVIDENCE OF
A. LAWRENCE KOLBE
1
2
2.
Q60. Are such distortions alone enough to explain a negative correlation between measured
3
4
Factors Left Out of the Measured ATWACC
ATWACC and the debt ratio?
A60.
No, in my view they are not. Instead, the ATWACC measured at higher debt ratios
5
understates the ATWACC that would be ideal to use in capital budgeting and in rate
6
regulation. The reason is that some of the non-tax effects of excessive debt may be hard
7
to detect and not show up in cost of capital measurement. Others may be purely cash flow
8
effects, with no impact on the cost of capital strictly defined, but with a definite impact on
9
the value of the firm. This problem is handled in capital budgeting by strict prohibitions
10
against artificially inflating the debt ratio when evaluating a project. For example, Brealey,
11
Myers and Allen, the leading graduate textbook on the subject, cautions against such
12
adjustments under the subtitle, “Mistakes People Make in Using the Weighted-Average
13
Formula.”41 This implies that the non-tax costs of excessive debt are valued by not
14
reducing the ATWACC for tax effects beyond those embodied in the ATWACC value
15
estimated from the market.42 Rate regulation using ATWACC needs to adopt similar
16
standards.
17
Q61. Why do you say the ATWACC at high debt ratios understates the ideal number for
18
use in rate regulation?
41
Brealey, Myers and Allen, op. cit, at pp. 515-16.
42
This also is consistent with the quotation from the same text cited in the body of my reply in footnote
22.
R-B-3
WRITTEN REPLY EVIDENCE OF
A. LAWRENCE KOLBE
1
A61.
The same logic used in capital budgeting also applies to rate regulation. For regulatory
2
purposes, the non-tax costs of excessive debt would wrongly be ignored if regulators who
3
rely on ATWACC were to assume the ATWACC would continue to go down as debt was
4
added. Those costs, discussed above, consist of such factors as reduced financial flexibility
5
and a higher risk the firm may have to bear the costs of financial distress. Such factors may
6
not show up when the cost of capital is estimated, but they do not show up as line items in
7
a regulated company’s revenue requirement, either. There is no place a regulatory board
8
can point to and say, “well, we’re adding to the debt ratio without holding the ATWACC
9
constant, but that’s okay because we’ve added X dollars for the costs of excessive debt to
10
the revenue requirement.” If anything, this factor implies that the true ATWACC for
11
project valuation or regulatory purposes is somewhat higher than the simple average of the
12
industry sample ATWACCs, but this refinement cannot be made with available estimation
13
techniques.
14
Firms consistently behave as if such non-tax costs matter more than the net tax
15
advantage of debt. If anything, the logic of such behaviour is stronger in Canada than in
16
countries with classical tax systems, the subject of much of the research, since equity is at
17
a bigger a corporate tax disadvantage in those countries than in Canada. Within the
18
Board’s formula-cost-of-equity, deemed-equity-ratio approach, the way to recognize these
19
principles is to treat the ATWACC of the industry as flat across a broad middle range of
20
capital structures in assessing market evidence on the appropriate deemed equity ratio.
R-B-4
WRITTEN REPLY EVIDENCE OF
A. LAWRENCE KOLBE
1
2
B.
MARKET VS. BOOK CAPITAL STRUCTURE WEIGHTS
Q62. Should book value weights be used in the formula to calculate the deemed equity ratio
3
for firms regulated on a book-value rate base?
4
A62.
No, that would be economically incorrect.
5
Q63. Why?
6
A63.
The cost of capital is determined in the market for regulated and unregulated firms alike.
7
Regulated shareholders will be unhappy if the market value of their shares falls, even if the
8
book value is constant. They will be indifferent to a fall in book value as long as the
9
market value is unaffected. In this they are no different from any other group of
10
11
shareholders.
Q64. Would use of market-value weights to calculate the deemed equity ratio for rate-
12
13
regulated companies be circular or lock in an excessive return?
A64.
No. The true beta depends on the market value of the firm’s leverage, again for regulated
14
firms just as much as for unregulated firms.43 That means the measured beta of a regulated
15
company sample will be lower when its market-value capital structure is higher, all else
16
equal. (Of course, in practice all else may not be equal.) The result is that the ATWACC
17
using market-value weights is the best estimate of the true ATWACC of the regulated
18
company, regardless of whether the regulated company’s market-to-book ratio is above or
43
Of course, the measured beta may be different in any particular calculation.
R-B-5
WRITTEN REPLY EVIDENCE OF
A. LAWRENCE KOLBE
1
below one.44 Market-value weights must be used to calculate the implications of the
2
sample equity ratios, as well, as is done in Dr. Vilbert’s and my procedures. With a
3
market-value capital structure that differs from the book-value capital structure, use of
4
book-value weights can lead to a potentially serious misestimate of the company’s true
5
required return on equity.
6
Brealey, Myers and Allen, op. cit., makes the same point. For example, at p. 516
7
the authors caution, “You cannot increase the debt ratio without creating financial risk for
8
stockholders and thereby increasing rE, the expected rate of return they demand from the
9
firm’s common stock.” Which debt ratio do they mean? Recall the quotation from
10
footnote 27 in the body of my reply, about drawing “a big X” through the book-value
11
balance sheet.
12
Professors Brealey and Myers are very familiar with the institutions of
13
ratemaking.45 If an exception were needed to a point they make this dramatically, to say
14
that book values should be used instead of market values to calculate ATWACC for
15
companies regulated on book value, it would have been discovered and included by the
16
eighth edition of the textbook. No such exception is included because none is warranted.
17
ATWACC and cost of equity adjustments for all companies should be calculated with
18
market-value weights.
44
As just discussed, the fact that the ATWACC of any given sample may display a downward slope
against the debt ratio does not change this point.
45
Some of my earliest consulting work led to papers written with Professor Myers on the implications of
the differences between capital charges under book-value regulation and those under competition. I
subsequently worked on related issues with both Professors Brealey and Myers in the context of
Heathrow Airport landing charges.
R-B-6
WRITTEN REPLY EVIDENCE OF
A. LAWRENCE KOLBE
1
The reason that market-value weights must be used to measure equity risk can be
2
illustrated with an extension to the dwelling example.
3
Q65. Please do so.
4
A65.
All right. Recall that the purchase price of the dwelling in the example was $100,000.
5
Suppose you paid that 10 years ago, and since then you’ve been renting it out. Suppose
6
depreciation has reduced the original book value from $100,000 to $75,000. Suppose also
7
that you’ve paid off about 20 percent of the original mortgage, leaving 80 percent still
8
owed. Suppose as well that your original mortgage was for 80 percent of the purchase
9
price, or $80,000. That means your mortgage balance is now ($80,000 x 0.80) = $64,000.
10
On a book value basis, you have $75,000 - $64,000 = $11,000 in equity.
11
What happens now if housing prices increase or decrease 10 percent? You cannot
12
even start to answer this question unless I tell you how housing prices have changed over
13
the last ten years. If I tell you that the market value of the dwelling is now $200,000, you
14
can calculate a 10 percent change as $20,000. A 10 percent decrease in housing prices is
15
therefore almost twice your book equity of $11,000.
16
Does that mean a 10 percent decrease will wipe you out? Of course not. Your real
17
equity is the market value equity in your dwelling. Suppose interest rates are unchanged,
18
so the market value of the mortgage equals its remaining unpaid balance. The relevant
19
measure of equity for risk-reward calculations is
20
21
22
True Equity
in Dwelling
=
Market Value of Dwelling - Market Value of Mortgage
=
$200,000 - $64,000
R-B-7
=
$136,000
WRITTEN REPLY EVIDENCE OF
A. LAWRENCE KOLBE
1
Therefore, the percentage rate of return on equity due to a 10 percent change in dwelling
2
values is
3
4
Rate of Return =
on Equity
Change in Dwelling Value
Starting Equity Value
5
6
=
+/- $20,000
$136,000
7
=
+/- 15%
8
Figure R-B-1 depicts the actual risk-return tradeoff after 10 years. A 10 percent
9
decline in dwelling values would be painful, but it wouldn’t come close to wiping you out,
10
no matter what the books say. Nor would it even show up on the books, despite its still
11
material impact on your actual investment.
Your Dwelling is Now Worth $200,000 with a $64,000 Mortgage Left;
If Dwelling Prices Rise or Fall by 10%, You Gain or Lose 15%
250,000
10% Gain in Asset Value,
15% Gain In Equity Value
225,000
$220,000
$200,000
200,000
175,000
150,000
Equity
125,000
10% Loss in
Asset Value,
15% Loss in
Equity Value
$180,000
Equity
If the Dwelling Price Rises by 10%:
$220,000 - $64,000 = $156,000
$156,000/$136,000=115%
100,000
75,000
If the Dwelling Price Falls by 10%:
$180,000 - $64,000 = $116,000
$116,000/$136,000=85%
$64,000
50,000
Mortgage
Mortgage
Condo Value after 10 Years
Change in 10-Year Value
25,000
0
Figure R-B-1
R-B-8
Your Equity Changes by +/-15%
WRITTEN REPLY EVIDENCE OF
A. LAWRENCE KOLBE
1
No landlord would assess his or her risk due to a mortgage by comparing
2
fluctuating property values to the remaining book value of the property. The risk that debt
3
imposes on the cost of equity is a function of relative market values, not relative book
4
values.
5
Q66. Is use of market values to calculate the impact of capital structure on the risk of
6
7
equity incompatible with use of a book-value rate base for a regulated company?
A66.
No, no more than it is incompatible to use market-based cost of equity estimation methods
8
(such as the Discounted Cash Flow method or the Capital Asset Pricing Model) with a
9
book value rate base. That is, the cost of capital is the fair rate of return on regulatory
10
assets for investors and customers alike. Most regulatory jurisdictions in North America
11
measure the rate base using the net book value of assets. But the jurisdictions still apply
12
market-derived measures of the cost of equity to that net book value rate base.
13
The issue here is, what level of risk is reflected in that cost of equity estimate? That
14
risk level depends on the sample company’s market-value capital structure, not its book-
15
value capital structure. That risk level would be different if the sample company’s market-
16
value capital structure exactly equaled its book-value capital structure, so the estimated
17
cost of equity would be different, too.
18
Q67. So why doesn’t this effect always show up when someone plots beta against the debt
19
20
21
ratio?
A67.
The forces outlined above and in Dr. Vilbert’s evidence are at work. Part of the problem
is in measurement. Part of the problem is the “decoupling” of beta from the market that
R-B-9
WRITTEN REPLY EVIDENCE OF
A. LAWRENCE KOLBE
1
accompanies industry transitions, financial distress and mergers. And part of the problem
2
is that some of the costs of excessive leverage don’t show up in measured beta, leading to
3
an underestimate of the appropriate ATWACC for capital budgeting and regulation when
4
sample companies with relatively high debt ratios are used. That is, the ATWACC you
5
estimate at high debt ratios may not be the one that gives the correct value for the firm or
6
investment. The problem is that the estimated ATWACC is too low to calculate the actual
7
value of the company’s operating cash flows.
8
But none of these forces imply it would be circular to use market-value weights to
9
calculate the deemed equity ratio at the formula rate of return on equity that is implied by
10
sample risk-return evidence for a company regulated on a book-value rate base. The fair
11
return on equity is just as much a market-driven parameter for regulated companies as it
12
is for unregulated firms. Use of book-value weights to calculate a regulated company’s
13
deemed equity ratio based on sample evidence when the sample’s average market-value
14
equity ratio exceeds its book-value equity ratio definitely underestimates the deemed equity
15
ratio necessary to give the regulated company a fair return on equity.
16
Q68. Would use of market-value weights to calculate an deemed equity ratio consistent
17
18
with sample evidence imply an abandonment of regulation based on book value?
A68.
Absolutely not. The deemed equity ratio is used to calculate a rate of return on the overall
19
rate base. It is absolutely standard in rate regulation, even in North America, to apply a
20
market-derived rate of return to a book-value rate base. The issue that drives the choice
21
of deemed equity ratio is how to understand what the market is telling us about the rate of
22
return investors require. The risk of shares, as with the equity in a home, depends on
R-B-10
WRITTEN REPLY EVIDENCE OF
A. LAWRENCE KOLBE
1
market values, not book values. Therefore, market values must be used to interpret the
2
sample company risk-return evidence. (If this were not true, book value rather than market
3
value would be the appropriate denominator for the dividend yield in the DCF model!)
4
North American rate regulation routinely looks to market values for every other part of the
5
rate of return calculation, and it should look to market values for the weights to use to
6
calculate the deemed equity ratio consistent with the sample evidence as well. Then, with
7
the fair overall rate of return correctly calculated based on market evidence, it can be
8
applied to the book-value rate base in the traditional way. It would be inconsistent with
9
standard regulatory practice in North America to say that a market-based rate of return
10
cannot be applied to a book-value rate base without abandoning book-value regulation.46
11
Q69. Does this complete Appendix R-B?
12
A69.
46
Yes, it does.
That, I would submit, is the real import of the AEUB’s U200452 quotation that Dr. Booth cites at p.
81 of his evidence, which I discuss at page 29 in Section II.G of the body of my reply.
R-B-11