FAIR RETURN FOR
GAZ METRO
EVIDENCE of
Laurence D. Booth
BEFORE THE
Regie de L’Energie du Quebec
July 2007
TABLE OF CONTENTS
EXECUTIVE SUMMARY .......................................................................................................................2
1.0
INTRODUCTION.........................................................................................................................4
2.0
FINANCIAL AND ECONOMIC OUTLOOK...........................................................................8
3.0
THE REGULATORY FRAMEWORK AND GAZ METRO’S RISK .................................. 18
4.0
FAIR ROE ESTIMATES...........................................................................................................44
5.0
REASONABLENESS OF THE ESTIMATES......................................................................... 59
APPENDIX A:
Professor Booth’s Curriculum Vitae
APPENDIX B:
Fair Rate of Return Standard and Comparable Earnings
APPENDIX C:
Discounted Cash Flow Estimates
APPENDIX D:
Internationalisation and the Market Risk Premium
APPENDIX E:
Canadian Market Risk Premium estimates
APPENDIX F:
US Market Risk Premium Estimates
EXECUTIVE SUMMARY
1
2
3
The Industrial Gas Users Association (IGUA) has asked me to provide an independent
4
assessment of the appropriate ROE for Gaz Metro and to assess its business risk. My overall
5
assessment is as follows:
6
7

Canada is late in the business cycle, but showing unexpected growth. However, it
8
is expected to slow soon to its trend line growth rate. Long Canada yields have
9
been stable until the last two months at just over 4.0% but have recently increased
10
by 50 basis points and expectations are now for yields to further increase over the
11
near term. Inflation is expected to remain in the middle of the 1.0-3.0% operating
12
range of the Bank of Canada.
13

Financial markets are flush with liquidity and even low rated non-investment
14
grade issuers have few access problems. Spreads for A rated issuers are at
15
reasonable levels and the TSX has reached all time highs. There are no financial
16
access problems for regulated utilities at present. Overall in my judgment there
17
has been no deterioration in the state of the financial markets since the last time
18
the Regie reviewed the fair ROE for Gaz Metro.
19

I continue to recommend that boards should adjust for changes in business risk
20
wherever possible through the use of deferral accounts and common equity ratio
21
adjustments, rather than through changes in the ROE. This allows the use of ROE
22
adjustment formulas in a mechanical way to avoid ROE hearings.
23

I would place the business risk of Gaz Metro as higher than that of either Union or
24
Enbridge Gas Distribution Inc (EGDI), which have less common equity than Gaz
25
Metro. Further, Gaz Metro benefits from relatively generous performance based
26
regulation (PBR) so that together I judge them to have equalized the overall risk
27
to shareholders as compared to the Ontario LDCs implying a similar fair ROE.
2
1

Using the traditional risk premium tests I would judge that utilities have a relative
2
risk rating of 45-55% of the overall market. Given that I make risk adjustments
3
through the common equity ratio I would use a value of 0.50 for Gaz Metro.
4

I estimate the current market risk premium consistent with long Canada bond
5
yields at the 4.5% level, to be 5.0%. Market risk premium studies support this
6
5.0% estimate. Including estimates from a multi-factor risk premium model gives
7
an average fair return of approximately 7.50%, adding in a 50 basis point
8
"cushion" gives a fair ROE of 8.00%.
9

In my judgement current formula allowed ROEs are excessive across Canada and
10
have failed to recognize that the use of an adjustment mechanism has lowered the
11
investment risk attached to Canadian utilities and converted their equity into a
12
form of floating rate preferred share, where observed yields are significantly
13
lower than current allowed ROEs.
14

It is the generosity of current allowed ROEs that has caused utility assets to be
15
valued so significantly above their book values. The evidence from takeovers of
16
Canadian utilities indicates that they are very attractive investments since the
17
takeovers are uniformly at significant premiums to book value. Once it is
18
recognised that the takeover premium is a non-earning asset it, is obvious that
19
investors are willing to “eat through” this non-earning asset simply to get the
20
return from the book assets in the rate base. In turn this implies that these assets
21
are being allowed a too-generous rate of return.
22

The generosity of the current allowed ROE to Gaz Metro is also indicated by the
23
fact that on October 10, 2006 GMi sold 2,9133,753 units of Gaz Metro for $17.16
24
a unit and recognized a gain of $16.70 million on the $50 million in proceeds.
25
Again, this sale was at a significant premium to book value indicating that the
26
allowed ROE is above the investor’s required rate of return or fair ROE.
27
3
1
1.0
INTRODUCTION
2
Q.
PLEASE DESCRIBE YOUR NAME, 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. A detailed resume is filed as Appendix A to this testimony. Further information and
6
copies of working papers by Dr. Booth can be can be downloaded from his web site at the
7
University of Toronto at http://www.rotman.utoronto.ca/~booth.
8
Dr. Booth has previously filed testimony before the Regie providing expert option on the
9
financial parameters of Gaz Metro and Hydro Quebec. He has also appeared before most of the
10
major utility regulatory boards in Canada including the National Energy Board and the CRTC.
11
Q.
PLEASE DISCUSS HOW YOUR TESTIMONY IS ORGANISED
12
A.
What is important in determining the fair ROE is that conceptually it is the investor’s
13
required rate of return, adjusted for issue costs. For debt instruments this required rate of return is
14
simply the current yield available in the capital market, for example, the cost of the Government
15
of Canada raising capital is the yield on Canada debt. These costs can then be estimated very
16
accurately, since apart from the issue costs they are reported in the financial press every day. The
17
fair ROE is then conceptually exactly the same, except that it has to be estimated since it is not
18
100% quoted in the press. This is because part of the investors’ required rate of return comes in
19
the form of a dividend yield, which is quoted, and part as a capital gain, which is not. As the
20
capital gain component becomes more important the overall problems in estimating the
21
investors’ required rate of return on equity, or fair ROE, increase.
22
However, before discussing estimation issues in detail it is important to note that Gaz Metro is a
23
stable gas distribution company with relatively little growth, so that estimation problems
24
involving Gaz Metro are intrinsically smaller than for regular corporations. In Gas Metro’s 2006
25
annual report, page 26 (AR2006, P26), Gaz Metro indicates that its policy is to “distribute
26
virtually all its income.” As a result it’s dividend yield is a much more accurate estimate of the
27
investor’s required rate of return than would be the case for a growth company. On June 19,
4
1
2007, units of Gaz Metro limited partnership, which owns 29% of Gaz Metro with the remainder
2
owned by Gaz Metro Inc, were selling for $16.84. These units are quite actively traded and
3
reflect all the normal concerns of investors with interest rates, the business cycle etc. The
4
following graph tracks the price of GMLP.
5
6
Of note is the relative stability in GMLP’s market price. Over the past year the price has
7
fluctuated in a relatively narrow range around $16-18 and apart from the price drop after the
8
Minister of Finance imposed a distribution tax on income trusts on October 31, 2006, that also
9
affected limited partnerships like GMLP, it has been very stable.
10
The stability of GMLP’s unit price and the absence of significant growth options implies that
11
Gaz Metro would be a dividend or income stock, even if it were not organized as a limited
12
partnership. Gaz Metro’s partial organization as an LP, where the bulk of the income is passed
13
through to the unit holders implies that its dividend yield is a relatively good indicator of the
14
investor’s required rate of return. On June 19, 2007 these limited partnership units were selling
15
for $16.84 for a 7.3% dividend yield. This 7.3% dividend yield may be “contaminated” by
16
GMLP’s other businesses, but most of them are also regulated. According to AR2006, P57 the
17
gas distribution assets were 77% of GMLP’s total assets. And on P8 of the AR2006 it is stated
18
that the gas distribution assets are the core business. I would therefore expect that this 7.3%
19
dividend yield would be very close to the investor’s required rate of return.
5
1
This conclusion is reinforced by comments made by Gaz Metro in its 2006 Annual Information
2
Form, page 24 where the company states
3
4
5
Given the relative stability of Gas Metros’ unit price this yield would be very close to the
6
investor’s actual return. The yield spread of 2-3% would indicate the relative risk of Gaz Metro
7
units as compared to the mid term Canada bond yield.
8
Further as of September 2006 the book value of GMLP was $7.87 so at $16.84 GMLP was
9
selling for over 2X book value. I will discuss the importance of market to book ratios as signals
10
to regulators of the fair ROE later, but the fact that GMLP is selling at such a premium to book
6
1
value indicates that the return earned by GMLP is in excess of what investors require. From this
2
market data on GMLP it is clear that the fair ROE is significantly less than the 10.19% estimate
3
of Dr. Chretien on behalf of Gaz Metro.
4
This discussion of GMLP’s unit price and its dividend yield indicate that financial issues such as
5
the fair ROE revolve around the capital market and the economy. This has been explicitly
6
recognized by the Regie in setting Gaz Metro’s fair ROE through a formula adjustment model
7
tied to the forecast long Canada bond yield. In Section 2 I therefore discuss where we are in the
8
economy in terms of basic indicators such as the GDP growth rate, inflation and interest rates.
9
Although I consider the Regie’s adjustment mechanism to be generous, I can see no economic
10
and/or financial market changes over the last few years to justify changing it. Currently corporate
11
profitability is at record highs, but in Appendix B I discuss in detail the fair return standard and
12
why “comparable earnings” testimony should be ignored.
13
In Section 3 I discuss the process of regulation and the business risk attached to Canadian
14
utilities. In section 4 I consider the fair return and draw on Appendix C where I attempt a DCF
15
analysis based on the S&P US utility data. I look at US evidence since other witnesses often
16
make the claim that Canada now has to “compete” for capital in an international market. This is
17
less true for utilities than any other sector of the capital market and further in Appendix D I show
18
that internationalisation of capital markets reduces the market risk premium. It takes
19
pathological assumptions for the Canadian market risk premium to increase as it becomes more
20
integrated into a world market. I therefore make the conservative assumption of largely ignoring
21
evidence from other capital markets; otherwise I would be forced to reduce my market risk
22
premium estimates. Section 5 comments on the reasonableness of the recommendations in light
23
of the effect of a formula ROE on the required rate of return and the observed valuation of utility
24
assets in Canada.
7
1
2.0
FINANCIAL AND ECONOMIC OUTLOOK
2
Q.
WHAT ARE CAPITAL MARKET CONDITIONS AT PRESENT?
3
A.
Basic macroeconomic data for the last twenty plus years is provided as background in
4
Schedule 1. Economic conditions can sometimes change quite rapidly as the impact of hurricanes
5
and oil price shocks are unpredictable. However, there is a rhythm to the economy, which
6
reflects the momentum as shocks gradually work through the system; this is what is generally
7
referred to as the business cycle. The basic economic variable here is the rate of economic
8
growth. The trend line for economic growth is around 3.0%, while the Conference Board of
9
Canada has recently estimated that potential GDP can grow at 3.2% due to increases in total
10
factor productivity, largely resulting from the application of information technology. So that
11
periods with growth significantly below that level are periods of contraction or recession,
12
whereas periods of growth significantly above that are expansionary periods.
13
Looking back over the last twenty years indicates that from 1989 until 1993 Canada was mired in
14
a deep recession in response to a normal cyclical slowdown as well as restructuring that
15
accompanied the passage of the Free Trade Agreement (FTA). We can also see the strong
16
economy of the mid 1980s and again the mid to late 1990s, when real economic growth was over
17
4.0%. Most recently, we can see the mild slowdown of the early 2000’s as recession in the
18
United States and the effects of the stock market crash in Canada weakened the economy. The
19
recovery was then slowed in 2003 as Canada was hit by a “perfect storm” of a strengthening
20
exchange rate, slowing growth in the United States, severe acute respiratory syndrome (SARS)
21
and a single incident of BSE or mad cow disease. These effects were largely temporary as the
22
Bank of Canada lowered interest rates in July 2003 and economic growth picked up and has
23
remained largely on track since.
24
Most recently we have again had good economic growth as strong growth soaked up the
25
remaining available labour and the unemployment rate dropped to 6.6%, marginally above the
26
natural or non-inflation increasing rate of 6.0%.1 Consumer spending has been strong as low
1
As estimated by the Conference Board of Canada.
8
1
interest rates supported the purchase of consumer durables, as well as record residential housing
2
sales. Further Business investment remained strong with additional rebuilding of inventory. Even
3
the effects of the oil price increases have been largely muted by external interest in Canada’s oil
4
sands and the perception that Canada has positive exposure to oil and gas prices. This perception,
5
allied to the continuing strength of the current account surplus, which has been running at 1.0%
6
of GDP, lead to a strengthening Canadian dollar, which recently has been above 95 cents US.
7
The overall strength of the Canadian economy caused the Bank of Canada to reverse its stimulus
8
policy at the start of 2006 and start increasing short-term interest rates as the overnight rate
9
increased in stages from 2.50% to the current 4.25%. This tightening monetary policy, coupled
10
with what seemed to be a slowdown in the US, partly caused by housing market concerns, led to
11
a minor slowdown in economic growth through 2006. Until recently it seemed that the Canadian
12
economy was operating at trend line growth and most of the inflationary pressures were in check.
13
The expectations were therefore for a loosening of monetary policy and a decline in interest
14
rates. However, over the last two months this policy stance has changed quite dramatically as
15
both the economies of Canada and the US have remained strong. As Business Week (June 25,
16
2007) recently mentioned “Look at May sales: despite the bite from $3 a gallon gas, they blew
17
away expectations rising 1.4% from April, the biggest gain since January 2006.” As a result of
18
this continuing strength in the economy expectations for interest rate cuts have reversed quite
19
sharply.
20
Q.
WHAT IS YOUR OUTLOOK FOR INFLATION?
21
A.
Over the past several years, the Canadian economy has experienced low and stable
22
inflation together with reasonably strong economic growth. The graph in Schedule 2 shows the
23
average CPI inflation rate since 1951. What is clear from this graph is the enormous run up in
24
inflation from the early 1950's through to its peak in the early 1980s. Since then it dropped to
25
plateau at the 4.0% level through the 1980s before the effects of the major slow down in the
26
early 1990s caused it to drop to its cyclical low in 1994/5, where it almost touched price stability.
27
Since that time changes in the consumer price index have remained broadly in the middle of the
28
Governor of the Bank of Canada’s 1-3% range.
9
1
Schedule 3 graphs the average annual inflation rate along with the average yield on long Canada
2
bonds and Treasury Bills since 1961. The graph shows that prior to 1981, inflation was
3
increasing steadily, until the Bank of Canada engineered a recession in 1982-3 to bring inflation
4
under control. Similarly, in the late 1980's there was a gradual increase in inflation and wage
5
settlements that peaked about 1991, as again, the Bank of Canada engineered a recession to bring
6
down the rate of inflation. Although the absolute rate of inflation has been brought down
7
considerably from these earlier periods, the same pattern of increasing inflation from 1994-2001
8
is evident as in the earlier periods of 1986-1990 and 1976-1982. In each case, interest rate
9
increases slowed down the economy and with it the rate of inflation. We can also see the effects
10
of the Bank of Canada’s tightening during 2006 as the 91 day Treasury Bill yield increased so
11
that by the end of the year it was almost at the same level as the long Canada bond yield, so that
12
we had a flat yield curve indicating a slowing economy.
13
Schedule 5 shows that the long Canada real bond yielded 2.15% on June 20, 2007, or 2.40%
14
below the equivalent nominal bond yield of 4.55%. The real bond guarantees the investor
15
protection from inflation, whereas the nominal bond has built into the yield compensation for
16
both the expected rate of inflation and a real yield. As a result, the spread between the nominal
17
and real rate marginally overstate’s the market’s inflationary expectations. Other measures of
18
inflation come in slightly lower as the GDP deflator has been running at under 2.0% as Canada’s
19
terms of trade have been impacted by exchange rate changes. Recently the Conference Board of
20
Canada estimated the long run inflation rate at 2.0% right in the middle of the Bank of Canada’s
21
operating band of 1.0-3.0%. What this means of course is simply that financial markets accept
22
the determination of the Bank of Canada not to let inflation get out of hand again.
23
The graph in Schedule 4 shows the aggregate net lending of governments in Canada, where a
24
negative number indicates government borrowing or a fiscal deficit. What is clear from Schedule
25
4 is the dramatic improvement in the fiscal position of all layers of government since the early
26
1990s and their return to balanced budgets. This in turn has reduced the supply of government
27
bonds and the need for the Bank of Canada to follow accommodative monetary policy, which in
28
turn has supported the drop in inflation. The recent Monetary Policy Update by the Bank of
29
Canada (January 2007) indicates confidence that core inflation will remain at the 2.0% level
30
through 2007.
10
1
Q.
WHAT IS YOUR INTEREST RATE FORECAST?
2
A.
Schedule 5 provides data on the full range of interest rates across the broad maturity
3
spectrum as of June 12-20, 2007. What is evident is that interest rates for long maturity
4
instruments are essentially the same as they are at the short end of the maturity spectrum; this is
5
referred to as a ‘flat’ yield curve. Schedule 3 charts the history of short and long term interest
6
rates together with inflation since 1961. It is clear that short term Treasury bill yields have
7
continued their long decline from their peaks in 1981 as inflation has receded. This long run
8
decline has been punctuated by periods when Treasury bill yields have increased to support the
9
dollar (1996) or fight a too vigorous economy (late 1980’s and 1990’s). In contrast, long-term
10
rates have continued their gradual year over year decline without these peaks. This is because
11
long-term bond investors look not just at the next 91 days, but far off into the future. As such,
12
long-term bond yields reflect the long-term future of the Canadian economy, while T-Bill yields
13
reflect short-term expectations.
14
Another way of looking at the impact of the Bank of Canada’s monetary policy is to recognise
15
that monetary policy works through both interest rates and the exchange rate: higher interest
16
rates and a stronger dollar together slow down the economy by impacting interest sensitive and
17
export industries. To examine both of these effects, the Bank of Canada maintains a “monetary
18
conditions index” or MCI, which is reproduced in the graph in Schedule 6.2 Again, the dramatic
19
changes since the early 1990’s are evident, as the MCI increased dramatically. We can also see
20
the long run monetary loosening ending around 1998 with the levelling off of the MCI as the
21
Bank of Canada started to worry about a too strong economy. This policy stance was reversed by
22
the end of 2001 as the stock market crashed, and the effects of 9/11 exposed the economy to
23
another shock, with further loosening helped by a weak dollar. It has been the subsequent
24
strength in the value of the Canadian dollar that has largely produced the upturn in the MCI
25
along with the recent upturn in short term interest rates.
26
What is evident from the increase in short-term interest rates over the last year or so is that the
27
capital market believes in the integrity of the Bank of Canada. There are no longer any fears that
2
The bank has recently downplayed the importance of the MCI and no longer relies on it.
11
1
the Bank will allow inflation to increase significantly. However, capital market conditions have
2
changed over the last two months. As recently as the end of March long Canada bond yields
3
were still at the 4.2% level of the last year or and capital markets were predicting interest rate
4
cuts. Since then the whole yield curve has moved up by about 50 basis points as Schedule 5
5
indicates. The fact that long Canada bond yields have increased in line with short-term rates
6
indicates that the market expects further tightening by the Bank of Canada.
7
Earlier this year in line with the predictions of the yield curve, I was predicting that short-term
8
rates would fall by the middle of the year and long Canada bond yields would not exceed 4.50%.
9
I now believe that this is optimistic and that the overnight rate will increase by 50 basis points to
10
4.75% with a significant risk that it will go higher. In this case long-term rates will marginally
11
increase from current levels to around 5.0%.
12
Q.
WHAT HAS BEEN THE RECENT STATE OF THE CAPITAL MARKETS?
13
A.
Since the onset of the last major recession in the early 1990s, capital markets have been
14
dominated by federal and provincial government financing. Their importance, however, has been
15
receding. Overall government “lending,” representing the aggregate of all levels of government,
16
was running at the rate of over minus $60 billion during 1992 and 1993 or at its peak over 9.0%
17
of GDP. Government net lending subsequently declined almost year by year as the economy
18
recovered and governments finally got their spending under control. Schedule 4 graphs the
19
government's net lending as a percentage of GDP.
20
The disastrous consequences of government fiscal policy starting in the early 1970s is obvious in
21
Schedule 4, as governments started to run persistent deficits (net lending was negative indicating
22
net borrowing). By the early 1990s interest payments were eating up over 30% of federal
23
government revenues and government spending at over 50% of GDP was unsustainable. Since
24
then it is clear that all layers of government have made serious efforts to restore some sanity to
25
their finances. By 1997 lending had become genuine lending and governments in aggregate were
26
in surplus for the first time in twenty-three years. In 2000 all layers of government in aggregate
27
ran a surplus of $32 billion as tax revenues soared and expenditures on welfare, unemployment,
28
etc., declined along with the unemployment rate. This amounted to over 3.0% of GDP, the
29
biggest surplus since 1951, when governments were still actively paying down the war debt.
12
1
Although the fluctuations in the economy have eroded the aggregate surplus since then, it is
2
remarkable that the weakening economy of the early 2000’s did not impose more pressure on
3
government finances.
4
The overall decline in government “lending” has opened up room for private sector borrowing as
5
corporations have returned to the equity and bond markets, following the strengthening of their
6
balance sheets. Fuelled by healthy consumer spending, corporate profits have rebounded from
7
the extreme cyclical lows of 1992-1994. Schedule 7 graphs the level of pre-tax profits to GDP. In
8
2000 pre-tax corporate profits reached 12.0% of GDP as the economy peaked. This level was
9
higher than the last cyclical highs of 1988-1989 and only slightly below the resource boom
10
fuelled highs of the 1970s. Although pre-tax profits dropped off to 11.0% of GDP for 2001 and
11
2002 as the economy weakened, they have subsequently spurted forward again on high resource
12
prices and reached a high of 14% of GDP in 2006. This profit data is mirrored in the capacity
13
utilisation data in Schedule 8, where we can see the drop in utilisation in 2001 through the
14
middle of 2004 as the economy slowed and the strong rebound since then with utilisation rates at
15
all time highs until the recent levelling off in response interest rate increases and the strong value
16
of the Canadian dollar.
17
The profit and capacity utilisation data provide the same signals as the inflation and interest rate
18
data: the last peak in the business cycle was 2000 with a minor slowdown in 2001-2003. Since
19
then we have been in the strengthening phase of the business cycle as the economy has been
20
strong. This combination of relatively low interest rates and booming corporate profits has lead
21
to stronger equity prices and a strengthening value of the Canadian dollar. Schedule 9 graphs the
22
C$ in terms of its US dollar value, where we can see clearly that its long run secular decline,
23
when it was heading for a sub 60 cent US level, was reversed in the Fall of 2002 after which is
24
has gone from strength to strength and has recently been over 95 cents US.
25
This strength has been mirrored in the performance of the TSX/S&P Composite, which has
26
rebounded from its lows in 2002 with each year since showing strong equity market
27
performance. Recently the TSX Composite has hit all time highs of well over 14,000, despite the
28
collapse of the income trust sector, indicating much more confidence in the stock market and the
29
Canadian economy. However, similar to the pattern of interest rate changes, the TSX has shown
13
1
considerable volatility since the end of March with days of plus or minus 100 points becoming
2
increasingly common. This may very well indicate the top of the market as investors have no
3
clear idea where the market is going.
4
5
Q.
HOW DOES THE STATE OF THE ECONOMY AFFECT PROFITS?
6
A.
Schedule 7 graphs the level of pre-tax corporate profits as a percentage of GDP. These
7
profits are taken directly from corporate tax returns and so avoid all the one time only accounting
8
losses that rocked Nortel, JDS Uniphase and others. Consequently, they are a more accurate
9
measure of corporate operating profits. The graph shows that profits are currently running at all
10
time highs at about 14% of GDP.
11
Another way of assessing corporate profitability is to look at the aggregate data maintained by
12
Statistics Canada (Quarterly Financial Statistics for Enterprises). Statistics Canada started
13
reporting quarterly return on equity data in 1980 based on Standard Industrial Classifications
14
(SIC) and then moved to North American Industrial Classifications (NAICs) in 1999. Schedule
15
15 graphs this average annual ROE against the spread between the yield on BBB debt and long
16
Canada bonds from Scotia Capital's Handbook of Canadian Debt market Indices. It shows that as
17
of 1980 the average ROE was 15.05% and the yield spread that rewards investors for holding
18
BBB rated debt instead of default free Canada bonds was very low at just over 50 basis points.
19
“Corporate Canada’s ROE” then declined during the 1982 recession as the yield spread widened.
20
The ROE then hovered around the 10% level during the growth oriented 1980's with a stable
21
yield spread. As ROEs fell from 1989 onwards, investors grew concerned about credit risk and
22
the yield spread increased dramatically to almost 350 basis points in 1993. The profit recovery
23
during the mid 1990s then caused the yield spread to contract only to widen in the early 2000s as
24
ROEs weakened.
25
The graph indicates the way in which the business cycle affects firms. During expansions,
26
profitability increases and credit risk is lessened, causing investors to buy corporate bonds on
27
narrower spreads over similar Canada bonds. During recessions the reverse happens: as
28
profitability is reduced credit risk tends to increase causing spreads to widen. Profitability in this
29
sense affects the market access of cyclical firms.
14
1
Schedule 12 shows recent information on corporate spreads using the AA, A and BBB spread
2
data from the Scotia Capital long bond indexes. The cyclical behaviour of spreads is again
3
clearly visible. The BBB and to a lesser extent A and AA spreads over equivalent Canada bonds
4
again clearly widened during the recession/slowdowns in the early 2000s. However both spreads
5
have tightened over the last few years reflecting the stronger economy and lower credit concerns.
6
The combination of booming corporate profits and lower credit spreads has lead to strong
7
financing activity. In Schedule 13 is the aggregate level of financing in Canada for the period
8
1993-2005 from data provided by the Investment Dealer’s Association (IDA). This data reflects
9
all the factors that I have discussed so far. Government borrowing was routinely 60-70% of total
10
financing as government debt crowded out private financing. However, over the last several
11
years there has been significant refinancing of existing, as well as new corporate debt issues, as
12
companies have taken advantage of lower interest rates. Corporate debt issues have increased
13
from barely 25% of the level of government debt to almost 70% and in the process private
14
financing activity has increased from 4-5% of GDP up to the current level of 8%.
15
Schedule 14 graphs the extent of total and private sector financing as a percentage of GDP to
16
indicate how receptive the capital markets are. This data confirms the stock market, profit and
17
spread data that capital markets are currently very receptive to new financing and a priori there is
18
no indication of any financial access problems. In fact, currently Canadian capital markets are
19
very receptive to new financing activity. As the Bank of Canada reported (January 2007)
20
“Against this backdrop, financing conditions in Canada remain favourable. Financial institutions
21
and markets continue to be willing to lend, since business profitability and overall financial
22
health remain strong.”
23
In the US the situation in financial markets is even more favourable. Business Week recently
24
reported (January 29, 2007) that new issues of non-investment grade debt have been running at
25
$127 billion a year: twice the level of 2002. Business Week attributes this to the “enormous
26
amount of money sloshing around and the changing structure of the debt market. Foreign
27
investors are shipping gobs of cash into the US.” As a result Business Week concluded
28
“Together these factors have combined to create unheard of pools of liquidity. Not only has that
15
1
helped keep the lid on interest rates – holding debt payments down – it has also made funding
2
readily available even for struggling companies.”
3
Q.
WHERE ARE WE IN THE BUSINESS CYCLE?
4
A.
Up to the middle of 2000, the U.S. was deep into an extended boom and showing distinct
5
signs of an overheating economy, whereas the Canadian economy was just getting its “second
6
wind” after spending so many years adjusting to the FTA and government over spending. The
7
Governor of the Federal Reserve then started to slow down the U.S. economy to avoid incipient
8
inflation and the Governor of the Bank of Canada followed suit, although more slowly, so that
9
monetary policy started to head off a recession. Unfortunately the bursting of the tech bubble
10
severely destroyed investor confidence as it revealed both the extent of corruption at the highest
11
level of some US corporations and the contempt with which some first line US investment banks
12
held their retail and institutional customers. The effect of this loss in investor confidence lasted
13
into 2003, but has now receded as both the US and Canadian economies have shown good
14
economic growth. The strength of commodity prices stimulated the Canadian economy through
15
2006 even as the US economy weakened.
16
For 2007 both economies are expected to be strong at trend line economic growth rates, inflation
17
to be contained to the 2.0% middle of the Governor of Canada’s band, despite strong energy
18
prices, and the capital markets to reflect this. Barring the impact of some extreme terrorist action,
19
it is an optimistic medium term economic and financial outlook reflecting continued strong
20
economic growth and performance at trend line performance. The only serious potential problem
21
is the fall out from the sub prime mortgage market in the US. Currently Bear Sterns is attempting
22
a multi-billion dollar bailout of its hedge funds that are heavily exposed to sub prime mortgages
23
in the US. The capital markets have brushed off these concerns so far, but should the bail out run
24
into problems and the problem itself be more extensive than at first anticipated then this could
25
yet be the defining moment of this business cycle top.
16
1
Q.
DOES YOUR PROFITABILITY DATA HAVE ANY IMPLICATIONS FOR THE
FAIR ROE?
2
3
A.
Yes. The stage in the business cycle affects the level of corporate profits as Schedule 7
4
clearly indicated. However, expressing profits as percentage of GDP isn’t useful for indicating
5
what firms typically earn as ROEs. In Appendix B I provide data on the ten year average ROE
6
for all the firms with full coverage provided by the Financial Post and the firms included in the
7
TSX60 sub-index. This appendix also includes a full discussion of the fair return standard and
8
how these ROEs relate to he market opportunity cost or fair return.
9
For the 675 firms in the Financial Post data base for which they provide coverage, the average
10
and median ROEs were as follows:
11
average
median
12
Clearly the typical (median) firm only earns about 6-7% ROE much less than that requested by
13
Gaz Metro. For the TSX60 firms the data is
14
average
Median
15
and reflects the higher level of profitability you would expect from the largest most powerful
16
firms in Canada.
17
Overall the FP data reinforces the aggregate profitability data that we are at the top of the
18
business cycle and profits are peaking.
2006
-0.88
6.91
2006
20.39
22.62
2005
-2.73
7.06
2005
16.33
14.94
2004
-0.38
6.37
2004
13.20
14.72
2003
-0.79
5.14
2003
11.87
11.29
17
2002
-5.49
2.64
2002
6.80
9.63
2001
-7.31
2.62
2001
6.78
11.30
2000
14.11
6.12
2000
7.97
12.74
1999
-4.46
5.26
1999
9.69
9.66
1998
-4.78
3.07
1998
6.30
8.70
1997
-0.35
5.40
1997
8.13
13.16
1
3.0
THE REGULATORY FRAMEWORK AND GAZ METRO’S RISK
2
Q.
WHY IS GAZ METRO REGULATED?
3
A.
Gaz Metro is a natural monopoly, which provides gas distribution services in Quebec and
4
delivers about 97% of the natural gas delivered in the province. The operation of a distribution
5
system is a geographic monopoly, since it makes little sense having two parallel distribution
6
lines. As a result, while there may be competition in the product (natural gas) delivered through
7
the distribution system, there is no competition for the distribution services. As a result there is a
8
residual power to abuse a dominant market position, which means the system has to be regulated.
9
The reason for the natural monopoly feature is the high amount of fixed costs, that is, the system
10
11
12
13
14
15
16
17
is capital intensive. In 2006 Gaz Metro’s gross margin (AR2006, P57) was allocated as follows:
$ Million
485
168
117
73
127
Gross Margin (distribution)
Operation and maintenance
Depreciation and amortisation
Interest
Income and taxes
18
Of significance is that the financing costs (equity and debt) are largely fixed costs set and
19
approved by the Regie, while depreciation and amortisations are not only a fixed cost but also a
20
non-cash charge, largely reflecting prior investment. These costs are all known in advance and
21
are independent of the demand for Gaz Metro’s distribution services. Moreover, operating and
22
maintenance expenses are also period costs and again largely independent of operating demand.
23
These costs largely increase due to annual wage increases approved by the Regie. Without
24
getting into a detailed analysis of Gaz Metro’s cost structure, it is clear that most of its costs are
25
“period” or fixed costs invariant to demand with very little variable or marginal costs.
26
In a competitive market it is difficult for a new entrant to enter a market where costs are largely
27
fixed, since the incumbent can lower prices to deter entry and drive a new entrant out of the
28
market. This is because with fixed cost production long run average costs are constantly
29
declining up to the capacity limit. As well as predatory pricing, the incumbent can cross
18
1
subsidise losses in one area with profits in another to enhance its position in the market. As
2
competition is reduced the incumbent can then charge higher prices and earn monopoly rents.
3
The relatively high fixed costs and low marginal costs can create risk for a company if the
4
product is a tangible good that faces direct competition with other high fixed cost producers. As
5
a result manufacturers with high fixed costs, like cars, airplanes and pulp and paper companies
6
are inherently risky. However, when the commodity is an intangible service that cannot be resold
7
or arbitraged away, like a distribution company, there is no feasible competition. Consequently
8
they have market power.
9
For these reasons the economics of a fixed cost “service” industry is such that a single firm
10
usually survives in the market with the potential for abuse of its dominant position.
11
Consequently, industries like natural gas transmission and distribution, electricity generation,
12
transmission and distribution, telecommunications, railways etc have traditionally been subject to
13
government regulation either through direct ownership or direct regulation. The presumption is
14
that without such regulation, the activities and prices of the dominant firm would be
15
unreasonable. In this respect it is important to note that it is regulation that follows the
16
underlying economics, not vice versa. Changing the regulation does not, in and of itself, change
17
the underlying economics or the dangers for the abuse of a dominant position as recent
18
experience with "deregulation" indicates. This economic imperative is reflected in the statutes
19
under which regulated companies operate, where firms are regulated to mimic the actions of a
20
competitive firm and yet reap the scale economies of the natural monopolist.
21
The litmus test for the competitive firm is the absence of monopoly profits. Conversely, the
22
regulated firm only earns normal profits and the equity holders earn a fair return on their
23
investment. Although legal statutes differ marginally from one jurisdiction to another, they are
24
similar to the regulations by which the Supreme Court of Canada came to determine a fair rate of
25
return. In BC Electric Railway Co Ltd., vs the Public Utilities Commission of BC et al ([1960]
26
S.C.R. 837), the Supreme Court of Canada had to interpret the following statute:
27
28
(a)
The Commission shall consider all matters which it deems proper as affecting the
rate:
19
(b)
1
The Commission shall have due regard, among other things, to the protection of
2
the public interest from rates that are excessive as being more than a fair and
3
reasonable charge for services of the nature and quality furnished by the public
4
utility; and to giving to the public utility a fair and reasonable return upon the
5
appraised value of the property of the public utility used, or prudently and
6
reasonably acquired, to enable the public utility to furnish the service:
7
This statute articulated the "fair and reasonable" standard in terms of rates, and that the
8
regulatory body should consider all matters that determine whether or not the resulting charges
9
are "fair and reasonable." To an economist, "fair and reasonable" means minimum long run
10
average cost, since these are the only costs which satisfy the economic imperative for regulation
11
and by definition do not include unreasonable and unfair cost allocations. The statute also
12
articulated the “prudently and reasonably acquired” test in terms of the assets included in the rate
13
base.
14
Q.
WHAT RISKS DO INVESTORS FACE IN INVESTING IN UTILITIES?
15
A.
Investors are interested in the rate of return on the market value of their investment. This
16
investment can be represented by the standard discounted cash flow model,
PO 
ROE * BVPS *(1 b)
K g
17
18
where P0 is the stock price, ROE the return on book equity, BVPS the book value per share, b the
19
retention rate (how much of the firm’s earnings are ploughed back in investment) and K and g
20
are the investor’s required rate of return and growth expectation respectively.3
21
Of the different sources of risk, we normally focus on the firm’s business risk, its financial risk,
22
and its investment risk. For regulated utilities we also add a fourth dimension, namely its
23
regulatory risk. In terms of the above equation the firm's accounting return on equity (ROE)
3
This equation is in every introductory finance textbook as d/(K-g) where d is the dividend or
ROE*BVPS*(1-b).
20
1
captures the business, financial and regulatory risk, which together we term income risk, whereas
2
all the other factors are reflected in investment risk, which is the way in which investors react to
3
the income risk and other macroeconomic variables. The regulator primarily affects income risk,
4
whereas investment risk is determined in the capital market and reflects, for example the impact
5
of changing interest rates.
6
Business risk is the risk that originates from the firm’s underlying “real” operations. These risks
7
are the typical risks stemming from uncertainty in the demand for the firm’s product resulting,
8
for example, from changes in the economy, the actions of competitors, and the possibility of
9
product obsolescence. This demand uncertainty is compounded by the method of production
10
used by the firm and the uncertainty in the firm’s cost structure, caused, for example, by
11
uncertain input costs, like those for labour or critical raw or semi-manufactured materials.
12
Business risk, to a greater or lesser degree, is borne by all the investors in the firm. In terms of
13
the firm's income statement, business risk is the risk involved in the firm's earnings before
14
interest and taxes (EBIT). It is the EBIT, which is available to pay the claims that arise from all
15
the invested capital of the firm, that is, the preferred and common equity, the long-term debt, and
16
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 financing
21
effects. For 100% equity financed firms, this ROI is also their return on equity (ROE), since by
22
definition the entire capital investment has been provided by the equity holders. The uncertainty
23
attached to the ROI therefore reflects all the risks prior to the effects of the firm’s financing and
24
is commonly used to measure the business risk of the firm.
25
As the firm reduces the amount of equity financing and replaces it with debt or preferred shares,
26
two effects are at work: first the earnings to the common stock holder are reduced as interest and
27
preferred dividends are deducted from EBIT and, second the reduced earnings are spread over a
28
smaller investment. The result of these two effects is called financial leverage. The basic
29
equation is:
21
1
Erreur ! Des objets ne peuvent pas être
créés à partir des codes de champs de
2
3
where D, and S are the amounts of debt, and equity respectively in terms of book values. If the
4
firm has no debt financing (D/S =0), the accounting return to the common stockholders (ROE) is
5
the same as the return on investment (ROI). In this case the equity holders are only exposed to
6
business risk. As the debt equity ratio increases, the spread between what the firm earns and its
7
borrowing costs is magnified. This magnification is called financial leverage and measures the
8
financial risk of the firm. The simplest way to measure this financial risk is through the debt
9
equity ratio.
10
The common stockholders in valuing the firm are concerned about the total “income” risk they
11
have to bear, which is the variability in the accounting ROE. This reflects both the underlying
12
business risk as well as the added financial risk. If the firm operates in a highly risky business,
13
the normal advice is to primarily finance with equity, otherwise the resulting increase in financial
14
risk might force the firm into serious financial problems. Conversely, if there is very little
15
business risk, as is the case with regulated utilities, the firm can afford to carry large amounts of
16
debt financing, since there is very little risk to magnify in the first place.
17
Business risk is then equivalent to variability in EBIT or the ROI, both of which reflect the
18
variability in the firm’s operating costs and revenues. To analyse this we normally look at how
19
easy it is to forecast operating costs and how stable revenues are.
20
These comments mean that a regulatory board has a variety of tools to manage the regulated
21
firm’s income risk. The first is that the Board can manage the different components of business
22
risk. The basic way that a board can do this is by establishing deferral accounts. The essence of
23
deferral accounts is simply to capture major forecasting errors. Instead of having the utility’s
24
stockholders “eat” any cost over runs in terms of a lower earned rate of return, the regulator can
25
simply pass the extra costs to a balance sheet deferral account. The value of the deferral account
26
is then charged to the ratepayers over some future time period. In this way “ratepayers” always
27
pay the full cost of service and stockholder risk is lowered.
22
1
A second tool is for the regulator to alter the amount of debt financing. If the regulator feels that
2
the firm’s business risk has increased (decreased) it can reduce (increase) the amount of debt
3
financing so that the total risk to the common stockholder is the same. Both of Canada’s national
4
regulators, the National Energy Board and the CRTC, have recognized this. When the CRTC
5
opened up Canada’s telecommunications market to long distance competition it specifically
6
increased the allowed common equity component of the Telcos to 55% to offset their increased
7
business risk. Similarly, when the National Energy Board decided to go to a formula based
8
approach for the return on equity in 1994 it reviewed all the capital structure ratios for the major
9
oil and gas pipelines and set the oil pipelines at 45% common equity, Westcoast at 35%, and the
10
remaining mainline gas transmission companies at 30%. In each case the different equity ratio
11
adjusted for differences in perceived business risks.4 Most recently the Alberta EUB has also
12
established different common equity ratios for a variety of different regulated utilities that
13
include local gas distribution companies, pipelines, electricity Discos and electricity transmission
14
companies.
15
The third tool available for the regulator is to directly alter the allowed rate of return, so that the
16
stockholder only earns a rate of return commensurate with the risks undertaken. The CRTC, for
17
example, has historically allowed Northwestel 0.75% more than the other Telcos primarily due
18
to the “ruggedness” of its operating region. The BC Utilities Commission has allowed Pacific
19
Northern Gas a 0.65% premium over its low risk utility (Terasen Gas) and the Ontario Energy
20
Board has allowed Union Gas a small premium over Enbridge Gas Distribution Inc.
21
Q.
WHICH TOOLS DO YOU ADVOCATE USING?
22
A.
It makes sense that any significant forecasting risks that are largely beyond the control of
23
the firm should be managed though the use of deferral accounts. The reason for this is simply
24
that they do not affect the efficiency of the utility and there are diversification gains by spreading
25
the variability over a large number of customers. As a result, deferral accounts are a “win-win”
26
solution as they reduce the operating risk faced by the company, thereby allowing a higher debt
4
Westcoast was allowed a higher common equity ratio because of the greater share of non-mainline
assets in its rate base. The mainline tolls were based on a 30% deemed common equity.
23
1
ratio and they lower overall cost of capital thereby benefiting customers. For this reason I have
2
long argued that companies should have deferral accounts for the cost of short term debt, for
3
example, since no-one can predict short term interest rates and otherwise there may be a
4
tendency to over estimate them.
5
With a choice between capital structure versus ROE adjustments; my preference is to adjust for
6
business risk in the capital structure for two main reasons. First, the market seems to consider
7
any changes in the allowed capital structure to be a more permanent change, while it expects the
8
ROE to change with capital market conditions. Since business risk is the primary determinant of
9
capital structure, it is to be expected that a board will change an allowed capital structure
10
relatively infrequently in response to significant changes in business risk. Second, allowing firms
11
to chose their capital structure and then adjusting the ROE to a fair return runs the risk that
12
although the equity holders are getting a fair rate of return the overall utility income and thus
13
rates are too high and unfair. An extreme example here would be a firm that “chooses” 100%
14
equity financing. The Board might then give a fair return, but rates are still unfair and
15
unreasonable, since the company is forgoing the advantages of using debt financing.
16
One corollary to the decision of many boards such as the National Energy Board and the Alberta
17
EUB to adjust capital structures in response to business risk differences is that the risk faced by
18
shareholders in utilities is very similar. To a great extent regulators have reduced differences in
19
business risk by allowing the use of deferral accounts and altering equity ratios.
20
Q.
WHY IS THE COMMON EQUITY RATIO IMPORTANT?
21
A.
The firm’s capital structure has a direct impact on the overall cost of capital as
22
conventionally defined in finance as the weighted average of the after tax sources of funds to the
23
firm. Note that this is not the same thing as the utility weighted average cost of capital that does
24
not consider these tax effects. In the following discussion wherever I use the phrase cost of
25
capital I am referring to the conventional, that is, non-utility definition
26
This topic has been the subject of enormous academic inquiry over the last forty years and has
27
generated two Nobel Prize winners in Professors Franco Modigliani and Merton Miller.
28
However, for all the sophistication of the academic models, the most important issue is that
24
1
certain types of financial instruments have a tax-preferred status. In Canada this status is
2
accorded debt instruments, since interest payments are tax deductible to the firm, whereas equity
3
dividends are not. As a result, there is a built-in tax advantage to any corporation using debt
4
financing. This tax advantage goes to the shareholders of unregulated firms and to the
5
customers of regulated firms, since the use of debt reduces the firm’s revenue requirement. As
6
will be discussed later, this asymmetry in benefits for the regulated firm is a motivating factor
7
behind regulated companies continually striving to increase their equity ratios.
8
The primary fact to remember is that equity costs are paid out of after-tax income, whereas debt
9
costs are tax deductible. Hence, for example, if debt costs are 7.0% and equity costs are 9.0%,
10
then at a 50% tax rate (for simplicity), the pre-tax costs are actually 18.0% for the equity
11
(.09/(1-.50)) compared to 7.0% for the debt. Conversely the after tax costs are 3.5% and 9.0%;
12
either way the costs of debt versus equity have to be compared on the same tax basis. It is these
13
“same tax” cost comparisons, whether before or after tax, that competitive firms make in
14
deciding their financing. This implies that there is an incentive for competitive firms to finance
15
with debt: as they replace expensive equity with “cheap” debt, their cost of capital goes down.
16
Hence, for the same fixed amount of operating income, the stockholders benefit from the tax
17
advantage of debt financing for competitive firms.
18
Q.
HOW DO WE KNOW THERE IS A TAX ADVANTAGE TO THE USE OF
DEBT?
19
20
A.
Apart from the fact that a huge amount of corporate financing revolves around tax
21
motivated transactions the recent announcement by the Government of Canada changing the tax
22
status of income trusts is a vivid reminder of their importance.
23
Income trusts invest in both the debt and equity of an operating company, where the debt is
24
structured to remove the income tax liability of the operating company. The trust is then non-
25
taxable, since it is legally the same as a mutual fund, and flows the interest on the debt, the
26
dividends on the equity, plus other non-cash charges like depreciation, through to the trust unit
27
holders. The income trust structure, therefore effectively removes the corporate income tax.
25
1
Income trusts have been incredibly popular in Canada, since the absence of the corporate income
2
tax allows more income to flow through to investors. However, government has lost increasing
3
amounts of corporate income tax. Even though the conservative government in Ottawa
4
campaigned on ‘no changes to the tax treatment of income trusts,’ their hand was forced by the
5
announcement of Bell Canada that it was following the lead of Telus and converting to an
6
income trust. There were also rumours that Encana and Suncor were planning $40 billion in
7
income trust conversions of their oil and gas assets. The result was that on October 31, 2006 after
8
the markets closed the Federal Minister of Finance, Mr. Jim Flaherty, announced that all new
9
trusts would be subject to a 31.5% distribution tax to put them on the same tax status as
10
corporations and that existing trusts would pay this tax in five year’s time.
11
The importance of the income tax changes can be understood from the following graph that
12
tracks the price of the exchange traded income trust fund, XTR.
13
14
Before the Minister of Finance’s decision the income trust ETF was at $15 and the day after it
15
had dropped to $13.25 and then on November 2 even further to $12.75 before rebounding
26
1
slightly. Most analysts predicted that the tax changes would cause income trusts to drop in value
2
by 20-25%, but the effect varies across different trusts depending on the proportion of Canadian
3
to foreign income and the type of income, that is, how much is return of capital and how much
4
newly taxable income. Plus the existing trusts would only be taxed after a four year grace period,
5
that is, in five year’s time.
6
Regardless the carnage on Bay Street caused by the changing tax rules vividly demonstrates that
7
the corporate income tax has a huge impact on the valuation of shares. Another way of saying
8
this is that removing the corporate income tax by financing with debt adds of the order of 15-
9
20% to the market value of the firm. We can see this from the fact that the exchange traded fund
10
would sell for $15 without the corporate tax and about $13 with the tax levied in five year’s time.
11
The impact of the time until the tax is levied means that the true value of removing the corporate
12
income tax is much greater than these price changes indicates.
13
Q
IF DEBT IS SO MUCH CHEAPER THAN EQUITY WHY DON’T FIRMS USE
MORE DEBT?
14
15
A.
They try to use as much debt as they can, but unlike income trusts the debt is held by
16
third parties. The beauty of the income trust structure is that the debt and equity is held by the
17
same part (the trust) so if a firm has trouble making an interest payment it negotiates with the
18
same party that owns the equity. However, for regular corporations the debt is owned by banks
19
and public institutions, like pension funs etc., that are not identical to its shareholders. As a
20
result, there are limits to the amount that firms can borrow due to the increased costs of financial
21
distress that are associated with higher fixed financial charges. In extreme cases, the higher fixed
22
financial charges can force a firm to be reorganised, or taken over, when it could probably have
23
otherwise survived had it been financed with less debt. As a result, it is a basic rule of corporate
24
finance that the financial risk is layered on top of business risk: firms with high business risk are
25
advised not to issue too much debt, otherwise their solvency could be jeopardised in the event of
26
adverse market developments.
27
This basic discussion is relevant since publicly traded firms are constantly re-assessing their
28
capital structures (“improving their balance sheets”) in light of changing market conditions and
29
the changing risk of financial distress. It also explains why capital structures differ from one firm
27
1
to another, since both the nature of their assets and expected cash flows are different. One firm
2
with mainly hard tangible assets will use large amounts of debt, since these types of assets are
3
easy to borrow against. Another firm that spends significant amounts on advertising will have
4
relatively little debt, since it is harder to borrow against brand names and “goodwill.” Yet
5
another firm will use very little debt, since it is not in a tax paying position and cannot use the
6
tax shields from debt financing. In each case, the firm will solve its own capital structure
7
problem based on its own unique factors.
8
This discussion puts the utility capital structure in perspective, since utilities have the lowest
9
business risk of just about any sector in the Canadian economy. Consequently, they should have
10
the highest debt ratios. There are several reasons for this:
11
12
13
14
First, the costs and revenues from distributing natural gas 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!
15
16
17
18
19
20
21
22
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
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.
23
24
25
26
27
28
29
30
31
32
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 Gaz
Metro’s distribution pipes would ever be ripped up and sold for scrap. In fact
Standard and Poors (Rating Nov 27, 2006) specifically states that it “believes that
if GMi were to default it would continue to operate as part of a reorganized entity
because of the essential service nature of its business.”
33
34
35
36
37
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,
28
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.
1
2
3
4
5
Consequently, utilities have very low business risk; have reserve borrowing power by being able
6
to return to the regulator, minuscule bankruptcy/distress costs and hard tangible assets that are
7
easy to borrow against. In fact, in many ways, utilities are unique in terms of their financing
8
possibilities,5 and are prime candidates for using large amounts of debt to utilise their very
9
significant tax advantages.
10
Q
ARE THE ABOVE IDEAS STANDARD IN FINANCE?
11
A.
12
Irwin (3rd edition) by Brealey, Myers and Marcus). In chapter 15 the text discusses capital
13
structure and notes the following:
Yes. A popular finance textbook is Fundamentals of Corporate Finance, McGraw Hill

14
15
16
17
18
19
20

(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 =
21

22
23
24
25
26
27
28
29
30
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.
5
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.
29
1
These four particular comments are taken from the discussion of what is commonly referred to as
2
the static trade-off model, where the tax advantages of debt financing are traded off against the
3
costs of financial distress and loss of financial flexibility. They are referenced simply because
4
there is little disagreement amongst academics that debt is valuable to the firm due to the tax
5
shields it generates. This consensus has then been amply verified by the stock market’s reaction
6
to the changing status of income trusts. As the second point indicates if debt is rolled over, so
7
that the interest and tax shields are expected to continue indefinitely, then the value of the tax
8
shield is the amount of debt times the corporate income tax rate. At a 36.12% tax rate this means
9
that every dollar of debt adds 36.12 cents in value to the common shareholders. The third quote
10
indicates that the value of the firm is increased by the present value of these tax shields. In fact
11
the equation referenced there is part of an approach called adjusted present value approach
12
(APV), which focuses heavily on the tax advantages to debt and which has been widely used to
13
value financial engineering strategies involving leveraged buyouts etc that remove the corporate
14
income tax. The final quotation specifically mentions utilities as companies that should borrow
15
heavily.6
16
Q.
PARTNERSHIP?
17
18
WHY HAVE YOU DISCUSSED TAXES WHEN GAZ METRO LP IS A LIMITED
A.
In Gaz Metro’s AR2006, P38, it states
“For regulatory purposes, cost of service includes deemed income taxes, large
corporations tax, and capital tax. These deemed income and other taxes are computed as
though Gaz Metro was a taxable Canadian corporation, notwithstanding the tax status and
the tax rate of the partners.”
19
20
21
22
6
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
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.
30
1
This means that Gaz Metro Inc pays taxes on its 71% share of Gaz Metro, whereas ordinary
2
investors who own the residual 29% currently do not.7 As a result there is still a tax advantage to
3
debt financing, as the more debt that is used the less income taxes that are imputed in rates.
4
Q.
IF UTILITIES ARE FINANCED WITH A LARGE AMOUNT OF DEBT
DOESN’T THIS MAKE THEIR EQUITY RISKIER?
5
6
A.
Not in practice. While financial leverage (the use of debt) magnifies the business risk to
7
the common shareholder, there has to be business risk to magnify in the first place. In practice
8
the monopoly position of most public utilities and the effect of protective regulation in Canada
9
has not allowed utilities to be put at risk so that high amounts of debt have not magnified the risk
10
to the shareholder in any material way.
11
In Schedule 15 is a table of earned vs allowed ROEs for the pipelines that are part of
12
TransCanada Corporation. There is a distinction between full cost of service pipelines regulated
13
by the National Energy Board and those regulated on a forward test year basis similar to Hydro
14
One. Foothills, for example, bills its shippers for its full costs and exactly earns its allowed ROE,
15
to the extent that it only reports one number in its surveillance reports to the NEB. The
16
TransCanada BC system (formerly ANG) is regulated on a similar basis to Foothills and the only
17
difference is that on its full acquisition by TransCanada there were some reorganisation costs it
18
absorbed so in 2003 it “voluntarily” under-earned its allowed ROE. I have always regarded
19
Foothills and ANG as the lowest risk regulated entities in Canada, since there is NO income risk
20
from their regulated operations at all. They consistently earn exactly their allowed ROE so after
21
the fact there has been no business risk attached to these two pipelines. Without any business
22
risk, both these pipelines can finance with large amounts of debt, in fact prior to RH-2-94 they
23
were financed with 25-28% common equity with the balance conventional debt.
24
Unlike Foothills and ANG the TransCanada Mainline and TQ&M are regulated on a forward test
25
year basis similar to Gaz Metro. This leaves the companies exposed to forecasting risk where the
7
In similar situations the 29% owned by a non-taxable entity would not have income taxes imputed and
recovered in rates. For example the Alberta EUB denied Altalink a tax component on the share owned by
Ontario Teachers Pension Fund, since it is non-taxable.
31
1
actual revenues and expenses may deviate from those expected and included in the revenue
2
requirement. However, the use of deferral accounts and long term contracting with shippers that
3
pay fixed demand charges, regardless of whether or not they ship, significantly reduces this
4
forecasting risk. The result is that both the Mainline and TQ&M consistently over-earn their
5
allowed ROEs. Over this whole period the Mainline only failed to earn its allowed ROE once
6
and on average over-earned by 0.27%, whereas TQ&M over-earned by 0.36% and never failed
7
to earn its allowed return.
8
In Schedule 16 is similar data for Union Gas, EGDI and Terasen Gas. This data is more difficult
9
to get since it does not appear to be publicly available the way that surveillance reports on the
10
NEB pipelines usually are. The data for Union and EGDI is based on weather normalised ROE’s,
11
since these utilities are not allowed deferral accounts for variances due to weather. In contrast,
12
Terasen Gas is allowed a comprehensive RSAM, which is a complete weather normalization
13
account, which takes into account not just the cost of purchased natural gas but also volume
14
variances due to weather. Of note is that Terasen’s “over-earning” is similar to that of the
15
TransCanada Mainline.8 In contrast Union and EGDI do not have as many deferral accounts and
16
over-earned to a much higher degree than the TransCanada Mainline or Terasen, let alone the
17
full cost of service pipelines.
18
If risk is the possibility of incurring harm or a loss the insight from the data in Schedules 15 & 16
19
is that regulated utilities in Canada have very little risk. It is also interesting that the degree of
20
over earning decreases with the use of deferral accounts. The full cost of service pipes can be
21
regarded as having 100% protection, since they neither over nor under-earn. The Mainline and
22
TQ&M have limited room to improve their earnings, since so many of their revenues and
23
expenses are fixed. Similarly Terasen Gas with comprehensive deferral accounts looks a lot like
24
the NEB forward test year pipes in having little room to over-earn. In contrast, the two Ontario
25
LDCs with fewer deferral accounts have over-earned the most.
26
It is also interesting to contrast this performance of regulated assets with the utility holding
27
companies (UHC) that actually face the market. For the major UHCs Schedule 17 gives their
8
Since 1998 Terasen’s actual ROE is prior to earning sharing.
32
1
earned ROEs along with those for Foothills. For example, what investors invest in as
2
“TransCanada” or TCPL is not the Mainline, but the combined entity including non-regulated as
3
well as regulated assets. This can be seen in the greater variability of its ROE. For 1993-1997
4
TCPL consistently earned more than the Mainline, but then in 1998-2000 as TCPL reorganised it
5
earned less. Throughout this period the Mainline underpinned TCPL’s results and was a beacon
6
of stability. One way of assessing this greater risk is simply to estimate the standard deviation in
7
each firm’s ROE. For Foothills as a full cost of service pipeline this was 0.98%, whereas for
8
TCPL it was 3.03%
9
Q.
WHAT COMPARATORS WOULD YOU USE FOR GAZ METRO?
10
A.
Before the Alberta EUB in 2003 I compared the different utilities in the Alberta generic
11
hearing on the following basis:
12
I:
The major short term risks caused by cost and revenue uncertainty:
13
14
15
16

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.
17

On the revenue side the risks largely stem from rate design, critical features are:
18
19
20
21
22
23
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.
24
25
26
27
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.
28
29
30
31
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.
32
II:
The medium and long term risks are mainly as follows:
33
1
2
3
4
5
6
7
8
9
10

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.
11
12
13
14
15
16

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.
17
A common thread running through the above brief discussion is rate design and regulatory
18
protection. There can be significant differences in underlying business risk that are moderated by
19
the regulator in response to those differences. The lowest risk utility is then one with the
20
strongest underlying fundamentals and the least need to resort to regulatory protection. In
21
contrast, another utility may have similar short-term income risk, but only because of its need to
22
resort to more extensive regulatory protection, so that it faces more problematic longer term
23
risks.
24
On that basis and at that time I judged the lowest risk regulated utilities in Canada to be
25
electricity transmission assets, since they had the following characteristics:
26
27
28
29
30
31
32





Minimal forecasting risks attached to O&M
Revenue recovery via the Transmission Administrator as a fixed monthly charge
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
33
and recommended 30% common equity ratios.
34
I then placed the gas transmission pipelines as the second lowest risk group. Here I classified
35
Foothills and the TCPL BC System (formerly ANG) as of equivalent risk to electricity
34
1
transmission assets with NGTL having marginally more risk than Foothills and the TCPL BC
2
System, since it was exposed to bypass risk and recovered its revenues through a forward test
3
year from a greater variety of shippers. I therefore judged that on its own NGTL could maintain
4
its financial flexibility on the same 30% common equity ratio allowed mainline gas transmission
5
assets. However, because NGTL was then allowed 32% and was almost “indistinguishable” from
6
the TCPL Mainline, I recommended the same 33% common equity ratio then allowed the
7
Mainline.
8
I then judged the local distribution companies (LDCs), including both gas and electric as the next
9
riskiest. These companies were distinguished by their retail operations, which mean that their
10
revenues are recovered from a large number of industrial, commercial and residential consumers.
11
This exposes them to both the business cycle and weather fluctuations. This revenue recovery is
12
largely a function of their rate design that may expose them to commodity charges and a fixed
13
and variable recovery charge. Within this group the conventional yardstick for LDCs was that
14
Enbridge Gas Distribution Inc and Union Gas were both allowed 35% common equity by the
15
Ontario Energy Board. In contrast, whereas the Ontario Energy Board allowed a purchased gas
16
variance account (PGVA) to ensure that the full costs of gas were recovered, both were still
17
subject to volume variances due to weather. In contrast, the BCUC through its RSAM removed
18
this risk from BC Gas (Terasen Gas), but only allowed it a 33% common equity ratio. With these
19
yardsticks I recommended a 35% common equity ratio for a typical local distribution companies.
20
Finally, I recommended 42% as the upper end of a reasonable range for the common equity of
21
ATCO pipelines, given that the BCUC allowed PNG, a smaller and much riskier pipeline, 36%
22
common equity. However, this ranking was provisional being dependent on the EUB developing
23
clear rules on intra Alberta pipeline competition and a rate design that lowered ATCO Pipeline’s
24
risk. Further it was my judgement that none of the Alberta utilities were as risky as Pacific
25
Northern Gas (PNG) with a 36% common equity ratio or Gaz Metropolitain (GMI) with a 38.5%
26
common equity ratio, where I regarded those two as the riskiest regulated utilities in Canada.
27
Q
WHAT DID THE EUB ALLOW?
28
A.
The Board decision can be summarised in the following table:
35
Table 13 Board Approved Equity Ratios
ATCO TFO
AltaLink
EPCOR TFO
NGTL
ATCO Electric DISCO
FortisAlberta (Aquila)
ATCO Gas
ENMAX DISCO
EPCOR DISCO
AltaGas
ATCO Pipelines
Last BoardApproved
Common
Equity Ratios
(%)
32.0
34.0
35.0
32.0
35.0
N/A
37.0
N/A
N/A
41.0
43.5
2004 Board
Approved
Common
Equity Ratios
(%)
33.0
35.0
35.0
35.0
37.0
37.0
38.0
39.0
39.0
41.0
43.0
Change in Approved
Common Equity Ratio
(%)
1.0
1.0
0.0
3.0
2.0
N/A
1.0
N/A
N/A
0.0
(0.5)
1
2
The Board’s risk ranking was essentially the same as mine although they allowed higher
3
common equity ratios than I recommended. Electricity transmission facilities operators (TFO)
4
were allowed 33% common equity, NGT was next with 35%, then electric distributors with 37%,
5
gas distribution 38% and finally ATCO pipelines was allowed the highest common equity ratio
6
at 43%. In each case non-taxable utilities were allowed more common equity due to the absence
7
of the dampening effect of corporate income taxes. AltaGas is a very small rural utility and was
8
allowed 41% common equity due this small size.
9
With risk adjusted through the common equity ratio the Alberta EUB then allowed all the
10
utilities the same ROE determined through an annual adjustment mechanism similar to that used
11
by this board.
12
Q.
WHAT HAS CHANGED SINCE 2003?
13
A.
In the four years since the Alberta generic hearing I have testified in business risk hearings
14
for the TransCanada Mainline, FortisBC, Terasen Gas, Union Gas, EGDI and Hydro One
15
Transmission and have not changed this basic ranking. The main change since then has been the
16
increased supply risk out of the Western Canadian Sedimentary Basin (WCSB) as its maturity
17
has progressed largely on track. This combined with the introduction into service of Alliance has
18
meant that the extra Alberta pipeline capacity exceeds current demand. For this reason the NEB
19
has successively allowed the TransCanada Mainline to increase its common equity ratio from
36
1
30% to 36% and increased its allowed depreciation rate to keep the stranded asset risk constant.
2
This common equity ratio has now been accepted as part of settlement agreements with
3
Westcoast (Duke) Transmission, Foothills and the TransCanada BC System. Although this
4
represents four different “companies” it is the same factor that has lead the NEB to allow higher
5
common equity ratios.
6
In the only other Board decision that I was involved with, the BCUC increased the allowed
7
common equity ratio of Terasen Gas from 33% to 35% to bring it into line with Union and
8
EGDI. In a 2006 settlement Union Gas negotiated an increase in its common equity ratio from
9
35% to 36%, while EGDI has recently requested an increase to 38% and is awaiting a board
10
decision.
11
With these common equity ratio changes, boards across Canada have reaffirmed the validity of
12
their adjustment formula. The EUB decision was announced in 2004, which was also when the
13
OEB announced its decision to continue its ROE formula after a generic hearing in 2003. The
14
NEB also reaffirmed its adjustment formula while increasing the TransCanada Mainline’s
15
common equity ratio. The BCUC marginally changed its ROE formula when it increased
16
Terasen Gas’s common equity ratio to 35%. The substantive change was to change the ROE with
17
75% of the long Canada yield to bring it into line with other boards, rather than the 100%
18
adjustment it used previously. Finally the OEB reaffirmed its formula as recently as December
19
2006 with a decision involving local electric distribution companies.
20
The important point is that almost all the boards across Canada that have looked at their ROE
21
adjustment formula have reaffirmed the fact that they are fair and reasonable. I therefore find it
22
difficult to imagine that the Regie’s adjustment formula does not also continue to give fair and
23
reasonable estimates of the ROE.
24
Q.
25
A.
26
“increases” in risk faced by various regulated utilities since I first testified in 1985. However, the
27
ability of Canadian regulated utilities to earn their allowed ROE has not been significantly
28
impaired and I have yet to see any of these risks materialise to significantly harm a Canadian
29
utility. In fact the opposite has occurred: the use of forward test years, fuel pass throughs, the
WHY HAVE YOU NOT DISCUSSED GAZ METRO’S INCREASED RISK?
I don’t think that they are material. I have heard many company witnesses discuss
37
1
removal of the merchant function and increasing focus on the core monopoly service have all
2
served to reduce the risk of regulated utilities in Canada. The fact is that regulation is a flexible
3
process that moderates or shares these risks even if they do materialise to the extent that the
4
regulated utility is rarely hurt. A case in point is Pacific Northern Gas (PNG), which I regard as
5
the riskiest regulated utility in Canada.
6
There is no doubt that PNG is extremely risky. It operates a tiny 600 kilometre pipeline from the
7
Westcoast (Duke Energy Gas Transmission) system through to Western British Columbia where
8
the economy is heavily dependent on forest products and a few cyclical industries. Until
9
November 2005 almost 70% of PNG’s throughput came from a few industrial customers with
10
one, Methanex, overwhelmingly important. Unfortunately, Methanex closed its doors in
11
November 2005 and PNG lost the load. Such a loss of load dwarfs anything that could
12
conceivably affect Gaz Metro.
13
How has the BCUC responded to PNG’s serious problems? In the first place the BCUC has
14
allowed PNG a 0.65% premium to the ROE as well as 3% more common equity than that
15
allowed its low risk benchmark (Terasen Gas). These more favourable financial parameters have
16
been allowed on an ex ante base to reflect PNG’s potential problems, since the risks attached to
17
PNG’s dependence on a limited number of industrial customers have been known for a long
18
time. That is, PNG’s shareholders were rewarded for its greater risk ex ante. However, as the risk
19
increased the BCUC then allowed PNG a series of deferral accounts. First a comprehensive
20
RSAM to remove weather induced variability in PNG’s earnings. Second an industrial customer
21
deliveries deferral account (ICDDA) to recover any deviations of actual deliveries from those
22
forecast for PNG’s large industrial customers. PNG has also taken $5.05 million of Methanex
23
related assets out of its rate base and put these into a special deferral account to be recovered
24
from other customers over a ten year period. Finally the BCUC has approved in principle the
25
conversion of PNG into an income trust to help reduce costs.9
26
I will discuss the future of PNG shortly, but at this point the important fact to note is the active
27
participation of the regulator, the BCUC, in helping PNG cope with a huge company threatening
9
This is now obviously dead. In fact PNG had dropped the idea before Mr. Flaherty’s Halloween
announcement.
38
1
event. For example, although Methanex accounted for 62% of PNG’s throughput the BCUC
2
allowed PNG to offer a special discount rate for Methanex and rebalance its rates. As a result,
3
before it closed Methanex only accounted for 7.6% of PNG’s operating revenues, even though it
4
was 62% of PNG’s throughput. As the Methanex related assets are recovered from other
5
customers it emphasises the fact that a regulated utility only faces two basic risks: short run
6
forecasting risk and the possibility of a “death spiral.”
7
Forecasting risks can be removed by deferral accounts if the regulator sees fit as the BCUC and
8
the NEB have. If a company is not allowed deferral accounts then it can manage these risks by
9
deferring expenditures to consistently come in under forecast and over-earn. This seems to be the
10
historic record in Canada, where over-earning seems to be positively correlated with the absence
11
of deferral accounts.10 The BCUC can and has used this regulatory protection for PNG, but it
12
cannot prevent a death spiral. This occurs when customers leave the system and the reallocated
13
costs cannot be recovered from the remaining customers, otherwise they too would leave the
14
system or the costs would be regarded as unfair and unreasonable. For PNG this death spiral
15
remains a possibility, where PNG’s actual and allowed ROE have recently been as follows:
16
17
18
19
Allowed
Actual
2005
9.68
8.20
2004
9.80
6.90
2003 2002
10.17 9.88
7.50 5.60
2001
10.00
7.40
20
The PNG ROE data indicates a persistent problem with earning its allowed ROE despite the high
21
amount of regulatory protection afforded it by the BCUC. The underlying reason for this is
22
simply that PNG is a very small utility. For 2005 PNG had property plant and equipment of
23
$171.35 million and 39,295 customers. Gaz Metro has a relatively high industrial load but if it
24
lost a customer the size of Methanex the associated margin and cost of stranded assets could
25
quite easily be recovered from other customers and barely noticed. The size and diversity of the
26
Quebec economy and the fact that Gaz Metro serves 97% of the province dramatically reduces
27
its risk.
10
Performance based regulation can then put in sharing mechanisms to allocate any over-earning between
the utility shareholders and ratepayers.
39
1
However, despite the most severe problems faced by any Canadian regulated utility how have
2
PNG’s shareholders fared? First, even after a “worst case” scenario arising, at year-end 2006
3
PNG’s book value was about $23.00 and its stock price $18. So an equity investor in PNG would
4
have invested approximately $23.0 in PNG’s assets, earned a reasonable ROE and yet still only
5
seen the value of this investment drop by $5.00 despite the loss of 62% of PNG’s throughput
6
threatening the very survival of the company.
7
The example of PNG illustrates the basic proposition that regulation shields the utility from
8
many of the problems it ostensibly faces. The reason is that should these risks arise the utility
9
invariably goes to the regulator and gets the costs allocated to ratepayers. PNG, for example,
10
anticipates the costs of stranded Methanex related rate base assets being recovered from other
11
ratepayers, not the shareholders. Another more recent example is the potential liability to EGDI
12
caused by the Supreme Court of Canada with respect to late payment penalties and the July 20,
13
2006 settlement. On page 3 of the October 31, 2006 MD&A EGDI simply states
“The company intends to apply to the OEB for recovery of the proposed payments
resulting from the settlement of this action.”
14
15
16
Again the major inference is that this is a “risk” not born by the company, but by the ratepayers.
17
As the actual versus allowed ROE data for the major utilities indicates none of the risks
18
advanced in regulatory hearings involving those utilities have actually harmed their shareholders.
19
In practice even when these risks have materialized they have harmed ratepayers. I would agree
20
with S&P that given the importance of Gaz Metro’s distribution system it is unlikely to ever go
21
bankrupt and would simply be reorganized. However a long time before this happened I would
22
expect the Regie to step to ensure that shareholders are not harmed.
23
Q
CAN YOU DISCUSS GAZ METRO’S RISKS EVEN IF THEY ARE NOT
MATERIAL?
24
25
A.
Yes. I will discuss the different factors but what is important is the sum of these factors
26
not qualitative weights placed on each one. Here the most important factor is Gaz Metro’s
27
answer to information request doc08 of the Regie, where it provided the allowed and incentive
40
1
ROE and the actual ROE. This data is graphed below where the sum of the allowed and incentive
2
return is labelled “target” ROE.
3
4
Allowed vs Actual ROE
15
14
5
6
13
12
11
7
10
9
8
8
1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006
9
Allowed
Target
Actual
10
11
Before 1996 the allowed equals the actual ROE after that except for 1995 Gaz Metro has always
12
exceeded its allowed ROE. Over the whole time period the over-earning averages 0.46% slightly
13
more than Terasen Gas and the Mainline, but less than EGDI and Union. “Over-earning” is a bit
14
of a misnomer, since some of this comes from the effect of incentive regulation, where gains are
15
shared between shareholders and ratepayers. If the incentive return and the allowed ROE are
16
taken to be the target then the over earning is reduced to 0.16%
17
On the basis of the historic ability of Gaz Metro to earn its allowed ROE there is no question that
18
it is no riskier than any other regulated gas distribution utility in Canada. Further there is no
19
indication that incentive regulation has hurt the company in any way. In this sense PBR has had
20
the same effect on Gaz Metro as it has had on Fortis BC and Terasen Gas: it has allowed them to
21
over-earn the fair ROE. It would be nice to think that a regulated utility would operate efficiently
22
without PBR, but the evidence is to the contrary. It is for this reason that the Ontario Energy
23
Board is considering the introduction of PBR for both Union Gas and EGDI.
24
One problem with PBR is that once the easy gains have been made a utility finds it increasingly
25
difficult to generate incentive returns. In this case there is a tendency to ask for some rebasing or
41
1
one time only improvements. In its management discussion and analysis of its recent results
2
(page 3) Gaz Metro states
3
4
5
6
7
8
“Gaz Metros efforts to have its performance mechanism revised were rewarded on April
27, 2007 when the Regie de l’energie (the Regie) approved the changes proposed earlier
in April by Gaz Metro in collaboration with interested parties. This is excellent (bold
italics added) news for Gaz Metro which expects to be in a stronger position to benefit
from the new performance incentive mechanism considering the evolution of market
conditions.”
9
This gleeful announcement to investors hardly indicates that PBR is a risk factor for Gaz Metro.
10
In fact it sounds like Gaz Metro negotiated a good deal.
11
Q.
IS GAZ METRO’S LOW RISK RESULT A SURPRISE?
12
A.
No. As I discussed earlier Canadian regulators tend to protect “their” utilities from risk so
13
that they can carry more debt. This is a substantive difference from US utilities where it seems
14
that rate hearings are less frequent, there is less use of deferral accounts and interventions are
15
less common.
16
In the case of Gaz Metro its main risk, that of weather fluctuations on demand, has been
17
minimised by a comprehensive deferral account. This account is more similar to that used by
18
Terasen Gas than either Union Gas or EGDI. In its AR2006 P 14 Gaz Metro notes
19
20
21
22
23
“The partnership benefits from a revenue normalization mechanism that is a function of
normal temperatures for the distribution of natural gas in Quebec. Gaz Metro normalies
deliveries (for temperatures) and reflects the adjustment in its revenues through rate
stabilization accounts…. The regulatory mechanism provides that the partnership will
recover these amounts from customers over five years.”
24
Gaz Metro is still subject to some weather risk, since heating use varies with prevailing winds as
25
well as temperatures, but the evidence in terms of earned ROEs is that this “wind risk” appears to
26
be small.
27
The other major risk mitigation tool used by the Regie is Gaz Metro’s allowed common equity
28
ratio. As indicated earlier Terasen Gas with a similar weather deferral account now has 35%
29
common equity versus the 38.5% allowed Gaz Metro. In its March 2, 2006 decision on Terasen
30
Gas (TGI), the BCUC stated
42
1
2
In this case the BCUC felt that Terasen’s RSAM deferral account was worth 0-3% on its
3
common equity ratio. However, Gaz Metro is also deemed a 7.5% preferred equity ratio for a
4
total equity ratio of 46% one of the largest of any Canadian utility since almost all Canadian
5
utilities have been retiring their preferred shares. This is because of recent accounting changes
6
that cause debt like preferred shares with hard retractions to be treated like debt for reporting
7
purposes regardless f heir legal characteristics. With a 46% equity ratio Gaz Metro has a very
8
large offset to any remaining risk differences between it and the two large Ontario Gas LDCs.
9
Q.
WHY THEN DO YOU THINK THAT GAZ METRO IS RELATIVELY RISKY?
10
A.
In my judgment Gaz Metro’s basic business risk is higher than either Union Gas or
11
EGDI, but this is before the application of regulatory protection. In assessing the evidence before
12
it in its generic hearing, the Alberta EUB compared the gas and electricity function on four
13
criteria as follows:
14
15
16

Credit risk: The Board believes that AltaLink (electric transmission) faces lower
credit risk compared to gas pipelines since its sole customer is a provincial
authority;
17
18
19

Supply risk: Gas pipelines have greater supply risk due to depletion of gas basins.
By contrast electricity generation is not a primary industry such as gas extraction
and therefore more stable in output.
20
21
22

Competition risk: Pipe on pipe competition is a reality for many gas pipelines,
whereas for electricity transmission assets, such risks are non-existent under the
current and foreseeable regulatory environment in Alberta.
23
24
25
26
27

Deferral accounts: The typical gas pipeline company has both capital and
operational deferral accounts that shield it from forecasting and unanticipated
errors. By contrast, AltaLink has only capital deferral accounts, and therefore
faces somewhat higher capital expenditure forecasting risk for a portion of its
capital projects.
43
1
These four categories are also useful for comparing Gaz Metro with EGDI and Union. One
2
critical difference here is that Gaz Metro has greater exposure to industrial customers. Overall I
3
would judge this to marginally increase its credit exposure and its competition risk. In its
4
AR2006, page 25 Gaz Metro notes
5
6
7
8
9
10
11
12
13
“Gaz Metro’s ability to achieve sound results is dependent on the competitiveness of
natural gas in relation to other energy sources, such as fuel oil and electricity. In Quebec
electricity has the largest share of the residential market. In the commercial sector natural
gas is generally competitive. The large industrial interruptible service customers, which
can generally use other energy sources during interruptions, currently prefer fuel oil.
However, large industrial firm service customers continue to prefer natural gas because
the potential savings are generally insufficient to justify the expenditures required to
adopt substitute energy. They also prefer natural gas to heavy fuel oil for its
environmental impacts.”
14
This points to greater competition risk facing Gaz Metro since there are more alternative
15
competitive fuels due to lower cost electricity in the residential market and fuel oil in the
16
industrial market. This has required subsidies in certain areas to attract and retain customers.
17
This competition risk may be decreasing as Gaz Metro also notes that the Quebec government
18
has questioned the use of electricity for space and water heating. Gaz Metro points as support for
19
this the Regie’s approval of a relatively large increase in electricity rates in 2006. On supply risk
20
Gaz Metro would be marginally riskier than Union and EGDI, since it is further along the
21
TransCanada Mainline system. However regulatory protection and the way that the TQ&M tolls
22
are blended into the TransCanada Mainline’s tolls mitigates this. Similarly its higher business
23
risk is mitigated by the more extensive weather deferral account.
24
Overall I would judge that the greater regulatory protection to have equalized the risk of Gaz
25
Metro with other gas distribution utilities, so it can be allowed the same ROE.
26
27
4.0
FAIR ROE ESTIMATES
44
1
Q.
HOW DO YOU ASSESS THE RISK OF A REGULATED UTILITY RELATIVE
TO THE MARKET AS A WHOLE?
2
3
A.
4
for the major utility holding companies and pure play utilities in Canada. Of note is that although
5
we use variability as a measure of risk, for utilities it is not a measure of business risk. This is
6
because as we have seen for companies like Foothills that exactly earn the allowed ROE, all the
7
variability comes from the variability in the forecast long Canada yield and the application of the
8
NEB formula to determine the allowed ROE. However several points are important: first for
9
TransCanada (TCPL) the holding company has more variability than the regulated Mainline;
10
second in comparing these variability measures with those in Appendix B it is clear than even
11
these UHCs are very low risk compared to Corporate Canada as a whole.
12
Note that the average standard deviation of the annual ROE for the TSX60 firms in Appendix B
13
is 12.87%, but this is pulled up by the short history for Fording Canadian Coal Trust. The median
14
is 9.20% so 9-10% seems reasonable for the typical standard deviation of the ROE for a large
15
TSX60 company. With this base of reference Foothills would have relative risk of 10% of a
16
typical TSX60 firm and TransAlta 46%, with most of the UHCs at around 20%. This is
17
supported by the observation that the only firms with more than a few years data with similar
18
standard deviations to the UHCs in Schedule 17 are the Chartered banks, Loblaws, Thomson and
19
Canadian Tire. This relative risk assessment of about 20% based on the standard deviation of the
20
ROE has been stable over time.
21
The weakness of this risk assessment is that it is based on the variability of the firm’s accounting
22
earnings, or total income risk. What investors are interested in is the risk of the securities they
23
hold, which includes investment risk independent of the income risk. Moreover, since investors
24
rarely hold single investments, they are interested in how the risk of their overall portfolio
25
changes as a result of holding a particular security. This measure of risk is called the security's
26
beta coefficient. The most common risk premium model is the capital asset pricing model
27
(CAPM), which says,
28
Schedule 17 shows the estimates of the variability of the ROE over the period 1993-2005
K RF MRP * 
45
1
that the investors required return (K) is equal to the risk free rate (RF) plus a risk premium, which
2
is the market risk premium (MRP) times the security’s beta coefficient (β).
3
Why the CAPM is so widely used is because it is intuitively correct. It captures two of the major
4
“laws’ of finance: the time value of money and the risk value of money. I will discuss the third
5
law of finance the tax value of money later, but the time value of money is captured in the long
6
Canada yield as the risk free rate. The risk value of money is captured in the market risk
7
premium, which anchors an individual firm’s risk. As long as the market risk premium is
8
approximately correct the estimate will be in the right “ball park.” Where the CAPM gets
9
controversial is in the beta coefficient since risk is constantly changing so too are beta
10
coefficients, which makes testing the model difficult. However, it measures the right thing:
11
which is how much does a security add to the risk of a diversified portfolio, which is the central
12
idea of modern portfolio theory.
13
The CAPM is the premier model for estimating required or fair rates of return. However, when it
14
was originally tested early results showed that it tended to over estimate returns for high-risk
15
(β>1) stocks and under-estimate for low risk (β<1) stocks. However these tests suffered two
16
major problems, which have never been overcome. First they used the Treasury bill yield as the
17
risk free rate, which is only appropriate for very short horizon (91 days) investments. In
18
regulatory hearings it is customary to use the CAPM with the long Canada yield, since equities
19
have longer time horizons than even the longest maturity long Canada bond as they have no
20
maturity date. The use of the CAPM with a long Canada yield will be referred to as the “classic”
21
CAPM even though this is not the way that it is discussed in finance textbooks or tested. To the
22
extent that long Canada bonds earn a maturity premium of about 1.0% over the average Treasury
23
bill yield, this classic CAPM automatically increases the risk free rate and adjusts for the bias
24
noted in these early tests of the CAPM.
25
The second problem is that these test used actual betas and were simply mechanical: whatever
26
was the beta over the previous five year period was used in the test as a forecast beta. As we will
27
see this is not how betas have ever been used in a regulatory context, where more judgment
28
based or adjusted betas are used.
46
1
To illustrate the betas for the major Canadian utilities for each of the 5-year periods ending 1985
2
through 2005 are as follows:
TCPL
3
Terasen
CUL
TAU
EMERA
Fortis
PNG
AVG
AVG(No PNG)
31/12/1985
0.79
0.205
0.477
0.617
0.659
0.545
0.55
0.55
31/12/1986
31/12/1987
0.845
0.588
0.136
0.465
0.471
0.315
0.531
0.222
0.517
0.253
0.384
0.459
0.48
0.38
0.50
0.37
30/12/1988
29/12/1989
0.634
0.599
0.524
0.561
0.373
0.381
0.201
0.221
0.301
0.248
0.449
0.424
0.41
0.41
0.41
0.40
31/12/1990
0.588
0.558
0.392
0.273
0.213
0.469
0.42
0.40
31/12/1991
31/12/1992
0.543
0.55
0.538
0.471
0.368
0.465
0.275
0.399
0.25
0.383
0.457
0.353
0.41
0.44
0.39
0.45
31/12/1993
30/12/1994
0.445
0.575
0.47
0.597
0.511
0.593
0.469
0.557
0.537
0.367
0.446
0.555
0.449
0.47
0.54
0.45
0.55
29/12/1995
31/12/1996
0.528
0.632
0.489
0.581
0.541
0.509
0.446
0.53
0.55
31/12/1997
0.478
0.344
0.567
0.483
0.546
0.625
0.573
0.46
0.506
0.4
0.376
0.31
0.288
0.437
0.48
0.44
0.51
0.44
31/12/1998
31/12/1999
0.563
0.254
0.461
0.331
0.596
0.516
0.529
0.273
0.562
0.43
0.491
0.343
0.593
0.523
0.54
0.38
0.53
0.36
29/12/2000
0.181
0.23
0.354
0.067
0.294
0.238
0.491
0.27
0.23
31/12/2001
31/12/2002
-0.051
-0.069
0.158
0.104
0.244
0.186
0.078
0.095
0.223
0.171
0.155
0.151
0.448
0.467
0.18
0.16
0.13
0.11
31/12/2003
31/12/2004
-0.423
-0.207
0.009
-0.002
0.04
0.029
-0.064
0.138
-0.051
-0.012
-0.043
0.031
0.362
0.464
-0.02
0.06
-0.09
0.00
30/12/2005
-0.183
0.076
0.408
0.059
1.09
0.477
0.32
0.29
4
The average beta excludes Gaz Metro (GMI) and are provided both with and without PNG due to
5
that company’s particular problems. For the market as a whole the beta is 1.0, so these beta
6
estimates indicate that these utilities and utility holding companies (UHCs) are lower risk than
7
the typical stock which is what we would expect given their ability to earn their allowed ROE
8
and the associated income certainty.
9
When betas are estimated there is always measurement error, since unique events can just
10
happen to coincide with stock market movements to exaggerate or moderate the underlying risk.
11
The beta estimates for TransCanada Corporation, for example (TCPL) reflect the collapse of its
12
stock market price as it divested non-regulated assets in the last 1990s just as the stock market
13
was booming and then its recovery in 2000-2001 as the market dived. As a result from 2001-
14
2005, TransCanada’s beta was negative. What this means is that during this period adding
15
TransCanada to a diversified portfolio reduced its risk since it went up when others went down
16
and vice versa. However it would be naïve to expect TransCanada to go through a similar
17
restructuring over the next five years so in this way its historic beta is a poor indicator of its
18
future risk.
47
1
It is for this reason that betas are usually grouped into industries and examined over time. In this
2
way the random behaviour of one firm is reduced in importance. The last column in the prior
3
table gives the average for the UHCs, which can be regarded as an “industry” beta. This average
4
beta is then graphed below along with that for Gaz Metro. The data since February 1993 reflects
5
Gaz Metro LP and the data before January 1987 Gaz Metro.
6
GMI and Average Utility Betas
7
0.7
0.6
8
0.5
0.4
9
0.3
0.2
10
0.1
Utility beta
J an-05
J an-04
J an-03
J an-02
J an-01
J an-00
J an-99
J an-98
J an-97
J an-96
J an-95
J an-94
J an-93
J an-92
J an-91
J an-90
J an-89
J an-88
J an-87
-0.2
12
J an-86
-0.1
J an-85
0
11
GMI
13
14
The data shows that for the five-year period ending in 1985 the average beta was 0.5511 . The
15
average then drops through to 1992 before increasing back to 0.55 for the period 1991-1995
16
before dropping from the 0.50 level in the late 1990s to negative for 2003 before increasing back
17
to 0.29-0.32 for the most recent five year period. Over this long period the average beta for the
18
utilities has been in a range from a negative number to 0.55. The top of this risk assessment is
19
higher than that obtained by examining the variability of accounting ROEs alone, reflecting the
20
fact that some of the risk is investment risk, independent of the income risk. The bottom of the
21
range reflects some unique factors from the stock market bubble of the late 1990s. However,
22
what is obvious is that Gaz Metro despite being an income stock and a limited partnership
11
Betas are estimated over five year periods of monthly data so the 1985 estimate covers the period 19801985.
48
1
behaves almost identically to the average for the other utilities. This is because they too are
2
income stocks.
3
Another way of looking at the data is to look at the betas of the relevant TSX/S&P Composite
4
sub-indexes. These are graphed in Schedule 18. The great advantage of the sub-index betas is
5
that they include more companies than the individual estimates and the data is more readily
6
available.12 This is particularly important due to the fact that a large number of regulated firms,
7
like Consumers Gas, Maritime Electric, Island Tel etc., have disappeared through corporate
8
reorganisation. Although, this means that their individual company betas have also disappeared,
9
it does not mean that their economic impact has disappeared. Consumers Gas now shows up as
10
part of Enbridge, Island Tel as Aliant and BCE etc., so their economic impact continues to show
11
up in the sub index betas. However, there are two disadvantages: the first is the impact of BCE's
12
non regulated operations on the sub index betas; the second is that the sub indexes are weighted
13
according to the TSE weights for each company. Consequently, these are not simple averages but
14
market value weighted averages, so that big companies like BCE have a disproportionate weight.
15
The Telco, Gas and Electric, Pipeline and utility sub-index betas up to the end of 2002 when the
16
TSE sub indexes were changed are as follows:
17
G as/E l ect r i c
D E C /96
0.52
D E C /97
0.47
D E C /98
0.53
D E C /99
0.37
D E C /00
0.21
D E C /01
0.17
D E C /02
0.14
T el co
0.60
0.61
0.80
0.96
0.82
0.87
0.85
Pi p es
0.54
0.44
0.42
0.18
0.06
-0.14
-0.18
U t ility
0.60
0.59
0.83
0.96
0.80
0.83
0.80
18
The sub-index betas largely tell the same story: Telco risk has undoubtedly increased as
19
competition has been introduced, particularly long distance, and consequently they have been
20
removed from ROE regulation. This has caused the betas for both the Telcos and the Utility sub-
12 Index data is available at the end of the month, whereas company data is only available in May-June
of the following year. The TSX sub index data ends in May 2002. The Telcos were removed from the
utility sub index as part of this reorganisation.
49
1
index to increase, since BCE has been such a large part of the Utility index. This has been
2
exaggerated by the fact that the sub indexes are based on market value weights so that BCE has a
3
huge influence on both the Telco and the Utility sub-indexes. However, the recent behaviour of
4
the Gas and Electric and Pipeline sub-indices require explanation.
5
It is important to remember that betas are simply a statistical estimate of the extent to which a
6
stock moves with the general market over a particular period of time. By convention, betas are
7
estimated over a five-year period. This means that if a critical event happens during the
8
estimation period, then the beta estimate will pick it up. However, once the event “passes out” of
9
the five-year estimation window, the impact of the event will disappear from the beta estimate.
10
For example, the graph in Schedule 18 shows that beta estimates were trending to a common
11
average until 1987, after which the pipeline beta increased and the others decreased. This lasted
12
for five years until they again came together.
13
If I had estimated betas during the period ending say in 1990, I would have estimated that gas
14
and electric betas had dropped and pipeline betas increased. However, is it reasonable to say that
15
gas and electric risk dropped during this period? The answer is no. What happened was that there
16
was a large stock market crash in October 1987 (-22.0%) and this was such a significant factor
17
that whatever happened in that one month affected all the beta estimates for the next five years
18
until October 1992, when the October 1987 results were no longer in the sample period.
19
Professional judgement would indicate that it is unreasonable to just use the statistical estimate
20
without recognising the underlying events that caused it, and then to make appropriate
21
adjustments. It is my judgement that betas tend to revert to their long run average levels: for the
22
market as a whole this is 1.0, but for regulated firms from Schedule 18, this is about 0.5-0.6.13
23
There is no indication from Schedule 18 that the non-Telco betas are reverting to 1.0.14
24
Consequently it is illogical to weight them with 1.0, since there is no expectation that their risk is
25
increasing to that of an average firm. So what explains the current betas?
13
This is also the accepted in the literature. Gombola and Kahl, “Time series properties of utility Betas,”
Financial Management, 1990, come to the same conclusion.
14 The Telcos have been reclassified out of utilities since they are no longer ROE regulated.
50
1
The answer is Nortel and the Internet bubble. During the late 1990s, the technology and internet
2
boom were driving North American markets. Nortel was controlled by BCE, so that BCE's stock
3
price was being driven by Nortel and the internet boom. In fact, this was driving the entire
4
Canadian stock market as Nortel and JDS Uniphase became an increasing part of the market and
5
at one point made up almost 35% of the value of the TSE300. As the prices of Nortel and JDS
6
Uniphase stock increased, so did the Telco and Utility indices and the TSE300. When this boom
7
turned into a crash and Nortel declined from $124 to under $1, Nortel took the Canadian market
8
and the Telco and Utility indices down with it. This is what caused the high beta estimates for
9
the Telco and Utility indexes in both 2000 and 2001.
10
In contrast, the gas and electric and pipeline betas declined. The reason for this was that as the
11
market went on a technology driven boom and bust, these stocks were largely ignored. In the
12
case of the Pipeline sub index, the collapsing share price of TransCanada Pipelines during 1999
13
and its recovery during 2000, was against a strong equity market in 1999 and a weak one in
14
2000. This movement of TransCanada’s share price against the general market movement
15
induced a negative correlation and the low beta estimate for the pipeline sub index.15
16
For the last several years the story in the Canadian equity markets has been recovery from the
17
“bubble” in Nortel’s stock price. Unless a similar bubble is expected in the next few years, taking
18
the recent beta estimates at face value makes little sense. It is my professional judgement that
19
after examining the behaviour of the betas we will not have another Internet bubble in the stock
20
market over the next few years. Further, the betas of gas and electric companies will revert to
21
their recent range of around 0.50 once the data from this anomalous period has passed out of the
22
estimation window as they now seem top be doing.
23
Q.
HAVE THESE INDEX BETA ESTIMATES CONTINUED TO BE AT LOW
LEVELS?
24
25
A.
Yes. The tables of individual beta estimates go to the end of 2005 and show that betas are
26
still at relatively low levels as the period 2001-2005 still includes the effect of the stock market
15 This stock market reaction was due to the poor performance of TransCanada’s non-regulated
operations in 1999 and the programme of retrenching and selling them off in 2000.
51
1
crash. In addition although the TSE discontinued the most useful sub indexes in 2002, the new
2
S&P/TSX indexes do have a utility index. There are problems in the coverage of the new
3
S&P/TSX sub indexes since they reflect S&P’s world wide view of what constitutes a sub index
4
with little concern for regional differences as a result there are many anomalies as S&P tries to
5
squeeze a square peg into a round hole. However, Schedule 19 shows that the betas of the new
6
utility subindex continued to decline through 2003 before trending upwards to a zero beta by the
7
end of 2005.
8
For further information on the effect of the stock market bubble on betas I have graphed the
9
betas of all the major TSX sub indexes from 1992 until the end of 2005 in Schedule 20. The most
10
important thing to note is simply that the average beta for the market as a whole is 1.0. However
11
we can see the dramatic impact of the information and technology (think Nortel and JDS
12
Uniphase) sub index beta, which increased dramatically from about 1.5 to over 3. As this beta
13
increased, by construction other betas had to decrease, and we can see that the effects were felt
14
by almost all the other sub index betas. Consequently the Nortel effect is not just a utility
15
phenomena.
16
Q.
WHAT ADDITIONAL EVIDENCE HAVE YOU LOOKED AT?
17
A.
One of the most important investment characteristics of utilities is their high dividend
18
payouts. As Gaz Metro indicates it is an income stock. This is why they appeal to Canadian
19
investors who can use the dividend tax credit and why their shares are generally held by
20
Canadian and not foreign investors. This means that utility share prices are driven by interest
21
rates as well as common market factors and suggests a two-factor risk premium model, where
22
there are two risk premiums: the market risk premium and a term spread risk premium that
23
reflects exposure to interest rate risk. Interest rate risk is the risk of investing in long Canada
24
bonds, instead of treasury bills. As interest rates increase returns from long Canada bonds go
25
down and vice versa. This exposure to interest rate risk also characterises utility stocks since
26
there dividend rich returns makes them “interest sensitive.”
27
I therefore estimated a two factor model for utilities where their returns were driven by the
28
common market factor, the TSX Composite return, as well as the return on the long Canada
29
bond. The beta from this two-factor model along with the conventional beta estimate is graphed
52
1
in Schedule 21. As can be seen the one and two factor beta estimates for the gas and electric and
2
pipeline subindexes show essentially the same behaviour over time. Given the measurement
3
error involved in any statistical estimation and the sensitivity of the estimates to economic
4
conditions, I discount the current estimates and judge a reasonable range for normal market
5
conditions going forward to be 0.45-0.55.
6
Q.
WHAT IS YOUR RISK PREMIUM OVER BONDS ESTIMATE?
7
A.
From Appendix E the Canadian market risk premium of equities over long-term bonds
8
since 1956 has been in a range 1.86-3.06% based on annual holding periods. If I extend the data
9
back to 1924 the range increases to 4.87-5.36%. However, conditions in the bond market prior to
10
1956 were substantially different from what they have been since and most of the decline in the
11
market risk premium has been caused not by a decline in equity returns but an increase in bond
12
market returns, commensurate with their increased risk. My Appendix F shows that similar
13
changes have occurred in the US, where the US market risk premium since 1956 has similarly
14
been in a range 3.63-4.61%, which is a substantial drop from the estimates from 1926.
15
My assessment is that much of the drop in the market risk premium has been caused by an
16
increase in the risk of investing in long government bonds. The twin problems of government
17
deficits and inflation drove up market yields in the 1970s and 1980s and caused the risk of
18
investing in government bonds to approach that of investing in equities. One way of looking at
19
this is to chart the yield on the real return bonds, which is in Schedule 22. Of note is that from
20
1991 through the end of 1996 the yield on the real return bond was around the 4.50% level. This
21
is the period when the government deficit and borrowing was approaching 10% of GDP. This
22
crowding out in the bond market created a significant risk that the government would inflate
23
itself out of its deficit problems causing bond investors to demand higher yields to protect
24
themselves. Significantly, as the government deficit began to fall so too did the yield on the real
25
return bond. Notably since government moved into surplus the yield on the real return bond has
26
been in free fall and has recently been well under 2.0%.
27
The impact of government financing problems has primarily been in the government bond
28
market where this inflation risk has been most obvious. In Appendix F Schedule 5, I graph
29
government bond betas from 1926-35 until the end of 2004. From this data it is clear that bond
53
1
betas increased dramatically until the mid 1990s when they peaked at over 0.50. Since deficits
2
have been tamed (at least in Canada) government bond betas have decreased accordingly and this
3
reduction in risk has lead to commensurate declines in real and nominal government bond yields.
4
At a bond beta of 0.50, at their peak, government bonds had at least a 200 basis point risk
5
premium embedded in them, a level similar to that of low risk utilities. This is why at that time I
6
was recommending very low risk premiums. This risk premium has now largely been removed
7
from government bond yields, as the yield on real return bonds has declined by a similar amount.
8
I currently estimate the market risk premium at 5.0%. This is significantly higher than the
9
experienced market risk premium earned in Canada over the last 48 years, but takes into account
10
the influence of the earlier data, the recent unexpected performance of the bond market, due to
11
declining long Canada bond yields, and the reduction in risk in the bond market compared to a
12
few years ago. From the previous discussion of the risk of a typical regulated utility, I would
13
place a reasonable beta estimate at 0.50. This would imply a risk premium of 2.5%. Adding this
14
risk premium to the long Canada yield forecast of 5.0% produces an estimate of the required rate
15
of return for investing in a typical utility stock at approximately 7.5%.
16
Q.
HAVE YOU ESTIMATED ANOTHER RISK PREMIUM MODEL?
17
A.
Yes. The CAPM is a single factor model, where all that matters is the risk of holding
18
securities in a diversified portfolio. However, the two-factor model indicates that the CAPM
19
does not capture all of the risks that affect securities. It has been known for some time that the
20
CAPM, when used with Treasury Bill yields as the risk-free rate, tends to give low estimates for
21
certain types of securities, which is partly why for regulatory reasons it is normally used with the
22
long Canada bond yield.16 However, this practice caused many of the problems in regulatory
23
awards in the mid 1990s when the long Canada bond yield was so high due to inflation concerns,
24
government deficits and the large risk premium embedded in government bond yields, which did
25
not have a counterpart in the equity market.
16 This is also why the market risk premium is normally estimated over the long Canada bond return,
rather than over Treasury Bills returns.
54
1
The exposure of utility returns to this interest rate factor I call “gamma” to contrast it with the
2
beta which is the exposure to the market risk. Schedule 23 graphs the gammas of the gas and
3
electric and pipeline sub indexes up until 2002. These gammas are more stable than the
4
equivalent beta estimates and show that on average gammas are about 0.50. As a result I judge
5
utility stocks to have about half the exposure to the equity market as the average stock and half
6
the exposure to the bond market as the long Canada bond.
7
The two-factor model partly adjusts for the known estimation problems of the CAPM by directly
8
incorporating the risk of the long Canada bond through a term or interest rate risk premium. For
9
example, the data indicate that utilities have about half as much interest rate risk as the long
10
Canada bond and half as much risk as the stock market. If yields on long Canada bonds increase
11
and the return on the long Canada bond is only 2.0% while the stock market increases by 10%,
12
then the return from holding the utility stock will be 6% over the risk-free rate: 5% due to
13
exposure to the market factor and 1% from exposure to the interest rate factor. In Schedule 24 is
14
a graph of the utility interest sensitivity or gamma using the new TSX utility subindex. The main
15
message is that gamma is still at the 0.50 level that I estimated earlier.
16
However, incorporating interest rate risk into the risk premium model means that other
17
adjustments are necessary as well. In particular, since the interest rate or term premium is the
18
premium over Treasury Bill yields, the market risk premium must be estimated in the same way.
19
In Appendix E (Schedule E1) I show that the realised return difference between long Canada
20
bonds and Treasury Bills was 1.27% using arithmetic returns over the period 1957-2006, which
21
is also approximately the average yield difference. The market risk premium over Treasury bills
22
would therefore be on average about 1.23% higher than over long Canada bonds. Consequently
23
the 5.0% that I am using for the market risk premium over long Canada bonds should be
24
increased to about 6.23% as a risk premium over Treasury Bill yields. The utility risk premium
25
would therefore be 0.5*6.23% or 3.12% for the equity market risk premium plus 0.5*1.23% or
26
0.62% for the interest rate risk premium. The overall risk premium would then be 3.74% over the
27
long run Treasury Bill yield.
28
The long run Treasury Bill yield is simply the rate that is expected to be earned from rolling over
29
treasury bills yields for thirty years, equivalent to the long Canada bond maturity. The best
55
1
estimate for this is simply the forecast long Canada bond yield minus this 1.23% interest rate risk
2
premium. Consistent with the 5.0% forecast I estimate this at 3.77% for an overall two factor
3
required return estimate of 7.48%, which is almost the same as the simple CAPM estimate.
4
Q.
PLEASE SUMMARISE YOUR ESTIMATES.
5
A.
The risk premium testimony is based on two models: a ‘classic’ CAPM risk premium
6
model and a two-factor model. The ‘classic’ CAPM estimate is based on an historic average
7
market risk premium “adjusted” for the changing risk profile of the long Canada bond. The
8
newer two-factor model takes into account the interest rate sensitivity of utility stocks. Both
9
models have been estimated over individual firm data as well as sub-index data and over
10
extensive periods of time. As more estimation procedures and larger data sets are used, there are
11
of course more estimates. However, by examining the impact of different economic conditions,
12
as well as the risk return relationship in the US and Canada, I can be confident that the fair return
13
is bracketed by the estimates. The methods provided the following fair return estimates:
14
Classic CAPM estimate:
7.50%
15
Two-factor model estimate:
7.48%
16
I put equal weight on both estimates and judge that the required rate of return is 7.50%. A 7.5%
17
return means a real return of about 5.5% with a long-run inflation forecast of 2.0%, in the middle
18
of the Bank of Canada’s operating range. This 5.5% represents a real return only slightly less
19
than that earned by the TSE300 index as a whole since 1956. Note that in my Appendix E,
20
Schedule 1, I estimate the real return on the TSE300 since 1956 at 11.12% minus inflation of
21
4.15% (arithmetic return estimates) or a real return of 6.97%, so awarding 1.53% less for a low
22
risk company like Gaz Metro seems generous.
23
Q.
IS THIS YOUR RECOMMENDED ALLOWED RETURN?
24
A.
No, regulated firms should be allowed to recover their issue costs in the allowed return in
25
the same way that issue costs attached to debt are included in the embedded debt cost. The equity
26
issue costs are made up of a number of components including in house costs, which are passed
27
on as general administrative costs plus the costs paid the investment banker. These costs are
56
1
made up of two kinds: the out of pocket reimbursement of expenses plus the under pricing of a
2
new issue to ensure a successful offering. Overall these costs run up to 5.0% for a normal issue,
3
although they can be smaller for larger issues since there are economies of scale.
4
The conventional way of working out the extra return that is required is to use the constant
5
growth model and recognise that because of these costs the firm has to earn a higher return on its
6
net proceeds than the nominal amount of stock that it has sold. For example, assuming a stock
7
with a 4% expected dividend yield and 4% growth, the cost of equity is 8.0%, that is
d
K  g 4.0% 4.0%
P
8
9
10
However, if the firm only receives a net of 95% of the current stock price, that is, 5% issue costs
then the equity cost is
d
4 .0
K  g 
% 4.0% 8.21
P
0.95
11
12
which is 8.21% or 21 basis points more.
13
In the example, if the investor wants a fair return of 8%, the firm has to be allowed an 8.21%
14
return on the net proceeds of 95% of the issue size. In this way 8.23% on 95% of the proceeds
15
provides the 8.0% return on the amount paid by the investor. Clearly, the higher the dividend
16
yield component and the less growth, the higher the impact of the new issue costs. For example
17
if the dividend payout is 100%, then the flotation cost allowance would be 42 basis points. This
18
is because the firm, by definition, is being forced into more new issues than a firm that reinvests
19
more.17 This would be the case for Gaz Metro given its very high payout.
20
Once the tax deductibility of some of these costs is considered, a "flotation or issue cost"
21
allowance of 15 basis points is reasonable plus the out of pocket expenses. However, I normally
17
Note that with 5% issue costs, the idea is that the stock should sell at a market to book ratio of 1.053,
so that it will net out book value on any new issue. With utility market to book ratios vastly in excess of
1.052 it is difficult to rationalise any flotation cost allowance, since it is unlikely that there will ever be
any dilution.
57
1
add 50 basis points as a cushion to the direct estimates in line with this practice of many boards.
2
This is mainly to ensure that there is no dilution and stock prices are more variable than a 5%
3
floatation cost allowance would indicate. Adding 0.50% to my estimates produces a fair ROE
4
estimate of 8.00% for a 300 basis point utility risk premium over my 5.00% forecast long Canada
5
bond yield.
58
1
5.0
REASONABLENESS OF THE ESTIMATES
2
Q.
THIS ESTIMATE IS LOWER THAN THE COMPANY’S REQUESTED ROE. DO
YOU HAVE ANY CORROBORATING EVIDENCE?
3
4
A.
Yes. First it has to be pointed out that the size of the equity risk premium is usually
5
estimated from historic data and in the U.S. it has been pegged at 6.00% using the Ibbotson et al
6
data. This became very controversial when people started doing simple tests of reasonableness.
7
For example, in Schedule 25 is a simple future value chart showing how one dollar compounds
8
at 6.00%, 10.5% and 12.0%. By year thirty, an investment at 6.0% would have grown to $5.74
9
whereas an investment at 10% would have grown to $19.99 and an investment at 12% to $29.96.
10
These are staggeringly large premiums for the 10 and 12% returns that proxy for the equity
11
market versus a lower “bond” market return, which leads to the natural question of how risk
12
averse do people have to be in order to require these huge premiums. Mehra and Prescott18
13
argued that the degree of risk aversion was unreasonably high. As Siegel 19 points out, “the
14
historical (equity) return has been too high in relation to the return on risk-free assets to be
15
explained by the standard economic models of risk and return without involving unreasonably
16
high levels of risk aversion.” The high earned returns phenomenon is now known as the “Equity
17
Risk Premium Puzzle,” since people have been at a loss to understand the historic U.S. record.
18
There have been two major approaches to explaining the puzzle. First, Siegel has shown that the
19
US results are time specific. He estimates the following risk premium estimates over long bonds:
Geometric20 Arithmetic Real Return
20
21
1802-1998
3.5
4.7
3.5
22
1802-1871
2.2
3.2
4.8
23
1871-1925
2.9
4.0
3.7
18
R. Mehra and E. Prescott, “The Equity Premium Puzzle,” Journal of Monetary Economics, (March
1985)
19
Jeremy Siegel, “The Shrinking Equity Premium,” Journal of Portfolio Management, (Fall 1999).
20
The difference between arithmetic and geometric returns is discussed at length in my Appendix E.
59
1
1926-1998
5.2
6.7
2.2
2
1946-1998
6.5
7.3
1.3
3
From the above data there seems to be a U.S. market risk premium of 6.7-7.3% since 1926,
4
which is the type of data normally presented by company witnesses in rate hearings. However, as
5
the time period is lengthened, the equity risk premium drops significantly. For the longest
6
available period the equity risk premium in the U.S. is only 4.7%. This leads to the question of
7
why so much reliance is placed on US data since 1926? The answer to this is that Fisher and
8
Lorie21 of the University of Chicago started the data-base at 1926 simply to capture the huge run
9
up in stock prices prior to the Great Crash of 1929. Further their original data-base is the
10
foundation for most of the subsequent capital market data and research. If they had used all of
11
the data that was available to them at the time, subsequent US market risk premium estimates, as
12
Siegel shows, would have been much lower.
13
The final column of Siegel’s table shows the real return on Treasury Bonds (Nominal minus
14
actual inflation). Over the whole period the actual real return has been 3.5%, but over the periods
15
since 1926 and 1946 it has been only 2.2% and 1.3% respectively. This is the root of the puzzle,
16
not that equity returns have been so large but that bond returns have been so low for such a long
17
period of time. This is the theme of Appendices E & F, that the enormous increase and volatility
18
of interest rates in the post war period has lead to unreasonably low estimates of realised historic
19
bond returns. Siegel points out that the introduction of Treasury Indexed Securities or TIPS in
20
1997 in the U.S. has lead to the direct observation of the US real bond return at 4.0%, which
21
compared to the 1926-1998 actual returns indicates that the realised bond return was 1.8% less
22
than expected. This means that, but for this bias, the U.S. market risk premium should have been
23
4.9% (6.7-1.8) or essentially the long run average U.S. market risk premium.
24
It is important to note that much of the debate about the market risk premium in the US stems
25
from the fact that until 1997 they have not had an inflation indexed bond and the above bias was
26
not obvious. In contrast, this has been well known in Canada, since we have had a real return
21
L. Fisher and J. Lorie, “Rates of Return on Investments in Common Stocks,” Journal of Business, 371, 1964.
60
1
bond since 1991. In fact, many of Siegel’s arguments were previously made by me in a 1995
2
paper in the Canadian Investment Review.22 In this case, following historic US evidence amounts
3
to the “one-eyed following the blind.”
4
The second way of resolving the puzzle has been to estimate a forward looking model using the
5
discounted cash flow (DCF) model to estimate the equity return and then subtract the long bond
6
yield. In most applications the Gordon constant growth model23 is used where the equity cost is
7
the forecast dividend yield (expected dividend d1 divided by current share price P) plus the
8
expected capital gain or growth yield (g).
d
K  1 g
P
9
10
Q.
DO YOU PROVIDE A DCF ESTIMATE?
11
A.
My Appendix C presents data for all US utilities followed by Standard and Poors as well
12
as the electric and gas utilities. This data is used to estimate a DCF required rate of return that is
13
then subtracted from the US government bond yield to estimate the utility risk premium
14
appropriate for these U.S. utilities. This estimate of the utility risk premium is that it has been
15
between 1.76-2.03% over ten year US treasury bond yields and falling. This is supported by the
16
increase in the market to book ratios of these companies indicating that the market has been
17
paying higher and higher prices for the same stream of utility earnings. That is, the required rate
18
of return has fallen faster than allowed rates of return.
19
However, to be conservative, I have also estimated the utility risk premium assuming both a
20
higher return on equity and a higher retention rate than has actually been the case. These
21
adjustments serve to increase the forecast growth rate and also the utility risk premium to up to
22
2.50%. The highest of these estimates would confirm the risk premium estimates from the one
23
and two factor models, since if the risk premiums are valid for Canada, they would imply a fair
22
Laurence Booth, “Equities over Bonds, but by how much?” Canadian Investment Review, Spring 1995.
23
Developed in Appendix C.
61
1
return of 7.5% (long Canada yield forecast of 5.0% plus the 2.50% risk premium) plus the 0.50%
2
flotation cost. This is the same as my direct estimate from the CAPM and two factor model, but
3
this is purely coincidental as this US estimate needs adjusting for the yield gap between ten and
4
30 year debt yields plus I used the highest risk premium estimate.
5
We can also look at the DCF estimate for the Canadian market as a whole. The dividend yield on
6
the Canadian market is currently about 2.4% and has increased significantly over the last year
7
partly due to the inclusion of income trusts in the TSE300 index. However, traditionally the
8
dividend yield on the equity market has followed the yield on the long Canada bond down as
9
interest rates have fallen. The following chart indicates just how closely the yield on the TSX
10
Composite and that on the long Canada bond track each other.
Dividend and Bond Yields
7
18
16
6
14
5
12
4
10
3
8
6
2
4
1
2
0
0
1981M01 1984M01 1987M01 1990M01 1993M01 1996M01 1999M01 2002M01 2005M01
TSXYield
Canadas
11
12
Adjusting for the income trust effect I would forecast the dividend yield to be about 2.5%,
13
consistent with the recent profitability of Corporate Canada. Further some have argued that share
14
repurchase provides a surrogate for corporate dividend payments. This has not been as
15
significant in Canada as the US because of the income trust market, but it may be that the
16
forecast dividend yield understates the expected cash return from holding stocks by up to 0.50%.
62
1
If this is the case a maximum forecast dividend yield might be 3.0%. This leaves the critical
2
question: what is a reasonable growth estimate?
3
From the previous graph the current dividend yield on the TSX Composite (left hand scale) is
4
1.67% less than that on the long Canada bond (right hand scale). This 1.67 is the obvious break-
5
even growth rate indicating that with risk aversion equity investors must be expected share price
6
growth of at east 1.67%. For individual firms there is a huge forecasting error attached to
7
estimating growth rates, but for the market as a whole there is less error. This is because many of
8
the gains made by some firms are at the expense of other firms. Holding a diversified portfolio
9
removes this risk and leaves the investor exposed to the overall level of profits and dividends. At
10
the economy level there is then a constraint on how much of the national income (GDP) can go
11
to profits, since as the profit share increases it does so at the expense of personal incomes, which
12
in turn leads to higher wage demands.
13
In Schedule 7 I provided a graph of annual pre-tax corporate profits as a share of GDP. In
14
Schedules 26 is the dividend payout based on the earnings and dividends of the TSX Composite
15
firms where both are adjusted to their index weights. Typically dividend payouts have been
16
about 50% for these large firms with a slight downward trend, except for the undefined payouts
17
in the early 1990s and in 2002 when huge corporate losses caused the payouts to be negative,
18
that is, positive dividends paid out of negative earnings. One of the problems with the data in
19
Schedule 26 is that it is drawn from accounting statements, so that the losses in 2002 for
20
example, were not cash losses but simply the write-off of bad acquisitions made primarily by
21
Nortel and JDS Uniphase.
22
Schedule 27 graphs dividends and after tax profits as a percentage of GDP where the after tax
23
profits are those reported for tax purposes and do not reflect all the accounting games that go into
24
GAAP profits. As is to be expected, aggregate dividends are more stable than aggregate after tax
25
profits. While profits plummeted during the recessions in 1981, the early 1990s and marginally
26
in the early 2000s the effect is not nearly as pronounced as indicated by Schedule 26. In fact it is
27
quite clear that the losses in 2002 were not widespread, nor reflective of true operating earnings.
28
From Schedule 27 dividends on average are around 2.3% of GDP and after tax corporate profits
29
about 6.0%, but much more variable. Further there is no obvious upward or downward trend.
63
1
Corporate profits tend to peak at around 7-8% of GDP at the top of the economic cycle and then
2
fall back. Likewise dividends are more stable, but rarely exceed 3.0% of GDP. This pattern has
3
been disrupted lately due to the huge profits made by resource firms that are largely unrelated to
4
economic factors and driven by events outside of Canada. However, it is hard not to conclude
5
that in the long-run, dividends and after tax profits grow at about the same rate as the overall
6
economy, but that in the short run, there is considerable volatility! Given that the average real
7
Canadian growth rate since 1961 has been about 3.6%24 and the Bank of Canada’s operating
8
band for inflation centres on 2.0%, this implies long-run growth rate in dividends and earnings at
9
about 5.70% (1.02*1.036). If this is combined with the 3.0% maximum forecast dividend yield
10
the DCF equity return for the Canadian market is about 8.7%. I would judge this to be
11
marginally high due to the income trust effect in dividend yields.
12
Schedule 28 shows the dividend payout of the aggregate dividends from aggregate after tax
13
profits. Again the recessions of 1981 and the early 1990s is clearly evident, although not the
14
slowdown of the early 2000’s. However it is obvious from this aggregate data that the aggregate
15
payout is closer to 40%, implying a 60% retention rate. With the corporate ROE of about 10%
16
from Schedule 1, this would imply dividend growth of 6.0% (b*ROE), which is approximately in
17
line with the nominal GDP growth rate. This would imply a DCF equity cost for the market as a
18
whole closer to 9.0%, but again it confirms its general level. With DCF equity costs for the
19
market as a whole of 8.7-9.0% and a forecast long Canada yield of 5.00% the market risk
20
premium estimate is 3.7-4.00%, which is marginally below my direct estimate.
21
Of note are two quite recent independent estimates of the Canadian market risk premium by
22
industry professionals. The first was a recent report by TD Economics (January 2006) "rates of
23
return for the long haul," which estimated long run rates of return at cash (T. Bills) 4.40%, long
24
bonds 5.60% and common equities 7.30-7.80%. The 7.30% lower end to the range came from
25
looking at long run earnings and dividend growth in Canada and the top end from the US. This
26
recent TD estimate confirms the observation of many that Canadian risk premiums are lower
24
The Bank of Canada pegs Canada’s potential GDP growth rate as lower than the Conference Board of
Canada at 2.80%.
64
1
than in the US and that my estimate of 8.7-9.0% estimate is reasonable when compared to TD's
2
estimate of 7.30-7.80%.
3
The second was a report by Rajiv Silgardo the chief investment officer of Barclays Global
4
Investors Canada Ltd, who in a summary published in the Canadian Investment Review
5
(Summer 2003) reported the following equity market risk premiums:
6
7
Canada
US
UK
Japan Aus
Europe
8
3.75%
4.50
5.75
2.50
5.00
4.50
9
Mr. Silgado estimated the equity risk premiums by using a modified growth model, but the
10
critical points again are a lower equity market risk premium in Canada than the US and the much
11
lower level of equity market risk premiums than those used by company experts.
12
The above types of analyses are not specific to Canada. Arnott and Ryan,25 two finance
13
"professionals," that is, non-academics, estimated the real growth rate in US dividends at 1.0%
14
from 1926-1999. This is well below the real growth rate in US GDP, implying that US aggregate
15
dividends grow at a slower rate than the corresponding values for Canada. They also produced
16
the following table for international growth rates over the 1969-1999 period:
Arnot and Ryan DPS and EPS Growth Rates
17
18
19
20
21
22
Real GDP
Real EPS
Real DPS
Average
US
2.3%
1.4%
1.3%
1.3%
Canada
2.9%
-2.2%
-0.9%
-1.5%
UK
2.1%
1.3%
2.2%
1.7%
Japan
1.6%
-3.4%
-1.6%
-2.5%
23
This data shows more pessimistic growth rates than the earlier Canadian data alone, since the
24
time horizon is shorter. It is possible to make dividends grow faster than earnings by companies
25
increasing their dividend payout, which is what happened in the UK. However, across all these
26
major economies, the Arnott and Ryan data indicates that corporate profits and dividends have
25
R. Arnott and R. Ryan, “The Death of the Risk Premium,” Journal of Portfolio Management (Spring
2000).
65
1
not kept up with GDP and that the average GDP growth rate is much less than the 3.75% used
2
above for Canada.
3
Arnott and Ryan argued that the actual returns on the U.S. equity market came from a reduction
4
in the required rate of return. As the investor reduces the required rate of return, market prices
5
increase causing a change in the valuation of the same dividend or earnings stream. They show
6
that 2.0% of the U.S. real equity return came from this change in the basis of valuation and make
7
the obvious point that this cannot continue forever. They conclude
8
9
10
11
12
“More important still, our 3.2% outlook for real returns falls short of the real
return available in inflation-indexed government guaranteed bonds. For the first
time in U.S. capital markets history, the equity risk premium is probably negative,
barring some very aggressive assumptions regarding economic growth and the
share of growth that makes its way to the investor in today’s enterprises.”
13
I am not as pessimistic as Arnott and Ryan are for the US, but it is clear that a DCF model results
14
in required return estimates considerably below the actual realised equity returns earned since
15
1926.
16
Q.
DO YOU HAVE ANY ANALYSTS’ "FORWARD LOOKING" ESTIMATES?
17
A.
No. It is generally accepted that analysts’ earnings forecasts are biased high. There is
18
increasing concern that with the decline in fixed commissions, security analysts no longer get
19
paid for the quality of their research. Instead, analysts have received a share of investment
20
banking fees stemming from corporate underwritings and mergers and acquisitions. In such an
21
environment it is difficult for an analyst to be objective with their earnings forecasts or place a
22
sell order on a stock. To do so would cut the analyst's firm off from future underwritings.
23
Consequently they have effectively become part of the sales team for equities. This conflict of
24
interest has been most evident in the Internet and Technology fiascos of the late 1990s, when
25
prominent analysts issued strong buy recommendations on the way up and kept them in place on
26
the way down and got sued in the process.
27
Academics have long recognised the bias inherent in analyst forecasts. However, this bias has
28
also long been recognised in the professional investment strategy reports. The difference
29
between the strategy reports from investment banks and the analyst reports is that the strategy
66
1
reports are concerned with overall market values. Consequently, the strategy reports will offer a
2
“sell” signal on equities in general (or changes in the asset mix towards bonds) while the same
3
company’s analysts continue to recommend “hold” on the individual equities. The reason for this
4
of course is that the company with a sell recommendation on its stock will rarely do investment
5
banking business with an investment bank that has a negative analyst. On the other hand, a
6
general recommendation to lighten equities and move towards bonds doesn’t target individual
7
firms and thus does not alienate corporates and jeopodise future investment banking business.
8
For example, on September 28, 2001, Credit Suisse First Boston (CSFB) issued a substantial
9
report on whether equity markets were over or under valued in response to September 11, 2001.
10
They relied on several valuation measures, one of which was a standard DCF model. They used
11
analyst forecasts (Institutional Brokers Estimation Service or IBES) out to five years and then
12
trend earnings thereafter. Using trend earnings moderates any bias in the analyst forecasts since
13
they are not projected out to infinity as is often the case. CSFB then equated this earnings stream
14
to the current market value to determine the implied equity risk premium. Their equity risk
15
premium estimate for the U.S. market was 5.3%, but they added:
16
17
“We would remind readers that over the last ten years IBES earnings numbers have on
average been 6.0% too optimistic 12 months prior to reporting date.”
18
They then “stress tested” their estimates using more reasonable numbers and the equity risk
19
premium dropped to 3.0%-3.8%. Even at this level they warned that because of the bias in
20
analyst forecasts, “Some of our assumptions may be overly optimistic.”
21
In a later section of the same report, CSFB valued the U.S. market using the DCF model. In this
22
case they inputted their cost of equity estimate for the U.S. market and used this to discount the
23
stream of earnings generated by the consensus economic growth rate. Their estimate of the US
24
market equity discount rate was 8.5%, which was broadly consistent with their 3.0-3.8% market
25
risk premium. It is also pretty much the same as my own estimate for the Canadian market using
26
the same approach.26
26
Note in a recent report (August 7, 2005) on valuing oil sands investments RBC-DS estimated the equity
cost of these (risky) investments using a required rate of return of 9.75%
67
1
There has also been independent academic corroboration of the CSFB approach. Claus and
2
Thomas27 used IBES earnings forecasts similar to CSFB, but unlike CSFB they noted the bias in
3
the forecasts but did not reduce them, so the estimates are high.28 Their market risk premium is
4
then the estimated discount rate minus the yield on the ten-year bond. Schedule 29 provides their
5
estimates for the last ten years for the U.S. and some other countries. Note these estimates are
6
higher than would be used in a regulatory hearing for two reasons. First, in a regulatory hearing
7
the risk premium would be over the thirty-year bond yield, so these risk premiums need to be
8
reduced by the spread between the ten and thirty year bond yield (about 30 basis points). Second,
9
as mentioned the earnings growth forecasts would have to be adjusted for the analyst bias.
10
Despite these qualifications, there are two important conclusions from the Claus and Thomas
11
research. First, their average for the US of 3.40% is consistent with the CSFB stress tested
12
estimate of 3.0-3.8%. Second, the Claus and Thomas estimates for Canada are for an average
13
risk premium of 2.23%, which is 1.17% less than their US estimates. This is consistent with the
14
independent evidence that I have provided where I conclude that the US market risk premium is
15
higher than in Canada.
16
Q.
17
CAN YOU COMPARE YOUR ESTIMATE OF THE MARKET RISK PREMIUM TO
THOSE IN RECENT STUDIES?
18
A.
Yes. In Schedule 30 is a table showing my estimate of 5.0% for the Canadian market risk
19
premium together with recent studies showing alternative estimates derived by both academics
20
and non-academics. The table shows for each study whether the estimate of the market risk
21
premium is based on arithmetic or geometric return estimates and whether it is an historic or
22
forward looking estimate. In a few instances, these classifications are not applicable (n/a). In the
23
Claus and Thomas study, for example, a DCF model is employed in which the authors use IBES
24
earnings growth data to estimate the market return from which the yield on 10-year US
27 J. Claus and J. Thomas “Equity premia as low as 3%? Evidence from analyst’s earnings forecasts for
domestic and international stock markets,” Journal of Finance, October 2001.
28 They noted (page 1657) “We considered a variety of biases that may exist in the IBES forecasts but
found only the well-known optimism bias to be noteworthy.”
68
1
Treasuries is deducted to arrive at the market risk premium. Similarly, in the Fama & French
2
and Arnott & Bernstein studies, the authors also employ growth models while in the Graham &
3
Harvey study, the authors use CFO forecasts of the market risk premium one year and ten years
4
forward.
5
What is clear from Schedule 30 is that the 5.0% market risk premium estimate is quite
6
reasonable when compared to these recent studies. These estimates are based on historic realised
7
data, forward-looking methodologies, and evidence from both the US and Canada. The picture
8
that emerges is that the Canadian market risk premium is significantly less than the 5.5-6.0%
9
adopted by the National Energy Board in its Decision RH-4-2001 (pages 53-54) where it
10
reviewed its adjustment mechanism. The overwhelming evidence is that my 5.00% market risk
11
premium is a reasonable input for the determination of a fair return on equity for a low risk
12
utility.
13
Q.
DO YOU ADJUST YOUR ESTIMATES FOR THE
“INTERNATIONALISATION” OF THE WORLD’S CAPITAL MARKET?
14
15
A.
No. These issues are discussed in more detail in Appendix D. However, it is undoubtedly
16
true that investors are more aware of international investment opportunities now than say twenty
17
or thirty years ago. At that time the world was characterized by currency restrictions, investment
18
controls and very limited international investing opportunities. Since then most currencies have
19
become freely convertible, most investment restrictions have been removed and there has been
20
an increase in the coverage of international stocks among investment advisors. This latter
21
coverage has been enhanced by international collaboration between investment banks and the
22
growth of some major international investment banks. Hence, it is inevitable that investors will
23
increasingly invest in different stock markets to diversify their risk. However, this diversification
24
reduces risk and with it the risk premium. In the same way that diversification across stocks in a
25
domestic market reduces risk, then so too diversification across international markets reduces
26
risk. Consequently, the removal of pension limits on foreign investments, and the gradual
27
reduction in tax restrictions etc, should decrease the equity market risk premium in both Canada
28
and the US. I am not aware of any basis in financial theory for simply averaging the US market
69
1
experience with that in Canada on the assumption that relaxing investment restrictions will
2
increase risk premiums: except in pathological cases financial theory states the exact opposite.
3
Further it has to be pointed out that Canadian stocks have always been affected by what happens
4
in the US equity market. One obvious linkage is that the standard barometer of the US equity
5
markets, the Standard and Poors 500 index has always included Canadian stocks. In fact, it
6
wasn’t until July 10, 2002 that S&P cleaned up its S&P500 index to exclude foreign stocks and
7
make it a 100% US index. Prior to that time there had been many Canadian stocks included in
8
the Index, like Inco and Barrick, and one, Alcan, had been in the index for 65 years. Similarly
9
some Canadian stocks have at times been part of the Dow Jones index. Hence, taking the
10
performance of US indexes as representing only US stock market performance is incorrect.
11
Q.
HAVE YOU ANY COMMENTS ON THE USE OF AN ADJUSTMENT
MECHANISM?
12
13
A.
Yes. In my judgement the adoption of an automatic adjustment mechanism has turned the
14
common equity of a regulated utility into a form of floating rate, preferred share. Traditional
15
floating rate preferred shares can be described as follows:
16
17
18
19
20
‘Floating rate preferreds offer a hedge against rising interest rates. Their dividend
will adjust (according to a formula) to a change in interest rates, subject to any
stated maximum or minimum yield. The variable dividend yield is designed to
allow the preferred’s price to remain relatively stable during a fluctuating rate
environment.’
21
This description is very similar to the results of the application of an adjustment mechanism to a
22
utility’s allowed rate of return.
23
The objective of regulation is to treat investors fairly. This is accomplished by awarding a fair
24
return such that the share price should only increase by the amount of earnings retained within
25
the firm and not paid out as a dividend. If a utility paid out 100% of its earnings as a dividend,
26
the share price should approximate its book value, as long as it continues to be awarded its fair
27
return. In this case, similar to floating rate preferreds, the annual reset of the allowed return
28
allows the price to remain relatively stable during a fluctuating interest rate environment. By
29
making the annual reset a function of long Canada yields, through the adjustment mechanism,
70
1
utility shares then offer a similar hedge against rising rates, since the utility’s ROE will change
2
with the long Canada bond yield.
3
The only substantial difference between utility shares on an ROE adjustment mechanism and
4
floating rate preferred shares is that only part of the utility’s ROE is paid out as a dividend and
5
the adjustment, for example using the NEB formula, is to 75% and not 100% of a fixed income
6
yield. These differences between floating rate preferred shares and ROE adjustment mechanism
7
utility shares do not, however, negate the fact that they have much in common. One critical
8
feature is that the dividend income has favourable tax treatment. As George Lewis of RBC-
9
Dominion Securities points out,29
“The Canadian tax code, in an effort to mitigate the effects of double taxation,
taxes dividends received by individuals and corporations at a lower rate than
interest income. Since dividends are paid out of after-tax corporate earnings
(whereas interest is a tax deductible expense of companies), corporations receive
dividends free of income tax, while individuals’ dividend income is taxed at a
lower effective rate (under the dividend tax credit system) than their interest
income. This means that a given dividend yield on a common share results in a
higher after tax income than the same numerical yield (interest rate) on a fixed
income (i.e., bond) instrument.”
10
11
12
13
14
15
16
17
18
19
At the time of his analysis, George Lewis put the pre-tax equivalent yield (PTEY) at 1.37; that is
20
a 10% dividend yield was equivalent to a 13.7% bond yield. He further noted that the prices of
21
Canadian utilities tended to increase as they increased their dividend payout.
22
The tax effect is well known in capital markets. BMO- Nesbitt-Burns produces a Preferred Share
23
Quarterly that tracks the performance of the preferred share market. In the Summer 2004 issue of
24
their Preferred Share Quarterly BMO-Nesbitt Burns provided the following yields:
June 2004
25
26
Retractable Preferreds (%)
27
Dividend yield
4.01
28
Mid Canada yield
4.09
29
Chapter 11 in Joe Kan (editor) Handbook of Canadian Security Analysis, John Wiley & Sons Canada,
2001.
71
1
After tax spread (corp)
1.77
2
After tax spread (indiv)
0.63
3
4
Straight Preferreds (%)
5
Dividend yield
5.48
6
Long Canada yield
5.34
7
After tax spread (corp)
2.54
8
After tax spread (indiv)
0.98
9
10
Floating Rate Preferreds (%)
11
Dividend yield
3.42
12
BA (3 month)
2.12
13
After-tax spread (corp)
2.25
14
After-tax spread (indiv)
1.22
15
The retractable preferreds are compared to mid Canada bonds since the retraction feature
16
shortens their maturity as compared to a long bond. The traditional straight preferreds are
17
compared to long Canada bonds, while the floating rate preferreds are compared to 91-day
18
Bankers acceptances (BAs), since their dividends are usually reset quarterly.
19
The important point about the comparison is that what we observe in the capital market is a
20
yield. This is determined by both risk and taxes. Take the straight preferreds, for example, in
21
June 2004 the long Canada bond had a yield of 5.34%, while straight preferreds had a yield of
22
5.48%. Clearly the preferreds would be regarded as riskier than the long Canada bond, since the
23
corporate issuer can default. However, the yield on the preferred shares was only 0.14% higher.
24
The reason is that the dividend income gets more favourable tax treatment than the interest
25
income from the long Canada bond. The correct comparison is the after tax yield difference,
26
which BMO-Nesbitt-Burns gives as 2.54% in favour of the preferred shares for corporates and
27
0.98% for individuals, which is the correct result: that on an after tax basis the riskier preferreds
28
give a higher yield.
29
The Nesbitt-Burns data vividly indicates that risk matters in the capital market, but so too do
30
taxes. This is the third law of finance: the tax value of money. It also points to Gaz Metros’
72
1
inappropriate comparison of the yield on GMLP units with a Canada bond yield. The correct
2
comparison with an ROE adjustment mechanism is to a similar floating rate preferred share. In
3
this respect an annual adjustment mechanism would put a utility's ROE in between the quarterly
4
floating rate preferreds and the retractable, generally five year, preferreds, since the reset is
5
annual. I assume that this is why Gaz Metro compared GMLP’s yield with a 5 year Canada bond.
6
This would indicate that the true risk premium is much higher than the 3.00% that I am
7
recommending. This comparison also renders US comparables of doubtful value, since due to
8
these tax implications the utilities are predominantly "Canadian stocks," or as George Lewis of
9
RBC-Dominion Securities, stated:
10
11
12
13
“However, while the impact of institutional and foreign investors can have a
significant impact on the trading levels of utility companies, in general a typical
utility will have a greater proportion of individual and domestic shareholders than
the typical Canadian company.”
14
Hence one of the features of the adjustment mechanism is that it makes the equity return
15
analogous to a form of floating rate preferred share, which lowers investment risk. Also
16
the very fact that a formulaic adjustment is used removes some regulatory risk due to
17
delayed ROE awards as well as the possibility of a punitive award.
18
The combination of an adjustment mechanism over long Canada bond yields without explicit
19
recognition of either the tax preference for preferred shares or the higher interest rate risk of the
20
long Canada bond makes the current formulas attractive to investors and more than fair.
21
Q.
IS THERE ANY EVIDENCE THAT THE FORMULA ROES AND CURRENT
ALLOWED COMMON EQUITY RATIOS ARE HARMING UTILITIES?
22
23
A.
Not that I am aware of. In the final analysis "fair" is determined in the equity market by
24
the reaction of investors. It is a basic principle of regulation that equity investors invest money
25
up front and then rely on the Board awarding them a fair ROE. In this case if the equity investor
26
invests one dollar in regulated assets, there is an implicit contract that they will be given the
27
opportunity to earn a fair ROE, such that the dollar that is invested is still worth a dollar, that is,
28
that there is no confiscation of wealth by subsequently awarding a sub-standard ROE. This is the
29
basic meaning behind Mr Justice Lamont's definition of a fair ROE.
73
1
What this means is that once a dollar has been invested in a regulated utility, the investor has to
2
be given the opportunity to earn what he could earn in the market on other equivalent
3
investments, if he still had the dollar to invest. This process is akin to someone investing in a
4
savings account where a judge has to determine the correct savings rate each period that can be
5
withdrawn from the fund. The important implication is that if the judge (regulator) is successful
6
then the savings will always be worth their original investment. This is the meaning of the basic
7
result in finance that fair means that the market to book ratio equals one. The only thing different
8
about utilities, as compared to the savings example, is that there is some very minor business
9
risk, although as I showed earlier full cost of service pipelines like Foothills have no income risk
10
and exactly earn whatever ROE the NEB allows.
11
In Schedule 31 is a table of earned ROEs, preferred stock yields and market to book ratios for a
12
sample of ROE regulated Telcos up until 1996.30 This sort of data was previously included by
13
Professor Berkowitz and myself in estimates of risk premiums over preferred stock yields. These
14
risk premiums were then consistent with the above remarks about preferred share yields being
15
the correct tax comparison. Note that for 1970-1983 their market to book ratios were hovering
16
around 1.0 and at times were significantly below 1.0, as the combination of high inflation
17
historic test years and regulatory lag exposed these Telcos to significant risk. As interest rates
18
fell from the early 1980s highs, the market to book ratios of these utilities increased significantly
19
as allowed ROEs were not cut sufficiently to reflect these market changes. The point is that
20
observing the market to book ratio is a valid way of assessing how investors are reacting to
21
utility allowed ROEs.
22
Schedule 32 is a graph of the market to book ratios for a sample of Canadian utility holding
23
companies (UHCs). The key implication is that, except for PNG, the market to book ratios are all
24
well above 1.0. For PNG it is clear that despite the efforts of the BCUC to reduce PNG's risk, the
25
market is still sceptical of the company's long run prospects. These market to book ratios include
26
to a differing degree the impact of non-regulated operations, but there is a clear indication that
30
Source data is from my paper, The Importance of Market to Book ratios in Regulation, NRRI
Quarterly Bulletin, Winter 1997.
74
1
none of these companies have suffered a loss of financial flexibility as regulators have moved to
2
the use of adjustment mechanisms.
3
Further there is direct evidence of the value of regulated assets from sales between firms. For
4
example,
5
6
7

TCPL purchased the 50% of Foothills that it did not own at a market to book of
1.6 based on the common equity. Moreover since TCPL already owned 50% of
Foothills the number of potential buyers was limited, which reduced the price.
8
9

Aquila purchased TransAlta’s distribution and retail business at a market to book
of 1.5 based on a total rate base of $472m (premium of $238m);
10
11

Fortis purchased Aquila’s Alberta interests for a premium of $215m over a rate
base of $601mm.
12
13

AltaLink purchased TransAlta’s transmission business for a $200mm premium
over a rate base of $644m.
14

In 2005 Kinder Morgan purchased Terasen for 2.7X book value,
15
16

In 2006 Gaz Metro sold GMLP units for $16.48 when their book value was less
than half that.
17
Note that in most of these cases, the market to book ratio, based on the equity, is much greater
18
than that based on the total rate business, since the debt is normally assumed and is valued at
19
close to its book value. For example in Fortis’ purchases from Aquila it paid $1.3 billion for total
20
rate base assets of $943mm (in Alberta and BC) 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 40% book equity of $377.2 mm. The market to book ratio based on equity was therefore
24
about 1.96X. The final value depended on closing transactions, but the point is that the market to
25
book based on the common equity was well above the indicated values based on total rate base.
26
Finally to return to GMLP, the following indicates its shareholder’s equity since 2003
75
Shareholders' Equity
Retained Earnings
951,933
960,949
898,161
881,506
Total Shareholder's Equity
924,588
938,442
884,944
876,004
2,783,197
2,880,094
2,360,987
2,430,898
Total Common Equity
924,588
938,442
884,944
876,004
Average Shares
117,507
116,496
114,477
110,475
7.87
7.99
7.73
7.69
Total Liabilities & Shareholder's Eq
Book Value Per Share
1
The important point to note is that consistent with company statements, GMLP returns almost all
2
its net income to its shareholders. As a result very little is retained within the business, that is,
3
retained earnings do not increase significantly and shareholder’s capital increases due to the
4
issue of new shares. As a result the book value per share is relatively constant at $7.70-$7.90.
5
The implication is that GMLP can be valued as close to a perpetuity and the required rate of
6
return is very close to the 7.3% dividend yield, which is very close to my fair return estimate. It
7
also confirms that investors are very happy with Gaz Metro’s financial metrics, since they are
8
paying a market price in the $16-18 range for a security with a recent book value of $7.87.31
9
Overall I conclude that current adjustment mechanisms and allowed common equity ratios are
10
generous and that boards across Canada have failed to lower allowed ROEs as adjustment
11
mechanisms have lowered utility investment risk. In this way I judge this failure to adjust for the
12
effects of an automatic adjustment mechanism as being similar to the failure to adjust for the risk
13
reduction effects of the adoption of forward test years and deferral accounts. In my judgment a
14
fair ROE for Gaz Metro is 8.0% and the current adjustment mechanism is generous allowing the
15
GMLP units to sell above book value.
16
Q.
DOES GAZ METRO HAVE FINANCIAL FLEXIBILITY WITH YOUR
RECOMMENDATIONS?
17
18
A.
Yes. Gaz Metro (AR2006, P22) has $531 million in term credit facilities and $138.2
19
million in operating facilities. The distinction between the two relates to how the funds can be
20
drawn down and used and at the time of the annual report only $171.8 million had been used. A
31
Part of this happiness is no doubt due to the fact that they receive the income tax component that has
been deemed and collected in rates, but not paid.
76
1
term facility is also used as a backstop for Gaz Metro Inc’s $400 million commercial paper
2
programme, which has a DBRS rating of R1(low). These large unused balances are due to the
3
fact that Gaz Metro raised $300 million in first mortgage bond financing in the summer of 2006.
4
Unlike other gas LDCs, Gaz Metro continues to use first mortgage bonds, rather than medium
5
term notes, to raise fixed rate long-term debt, which gives it significantly more financial
6
flexibility. These bonds are rated one notch higher than Gaz Metro’s corporate credit rating due
7
to their high degree of collateralisation, in that they are effectively supported by all moveable
8
property owned by Gaz Metro Inc in Quebec.
9
These bonds are rated A by both DBRS and S&P and no gas utility in Canada has higher bond
10
ratings. DBRS confirmed these ratings as recently as April 5, 2007 after reviewing the impact of
11
the Minister of Finance’s decision to tax flow through financing vehicles. These bonds are
12
actually issued by Gaz Metro Inc and guaranteed by GMLP, where GMLP is restricted by the
13
Regie from issuing other guarantees (AR2006, P23). S&P confirmed it’s A rating in a November
14
27, 2006 report on Gaz Metro Inc and based this in part on
15

GMLP’s low risk natural gas transmission and distribution business,
16

GMI’s strong business position being enhanced by supportive regulation,
17

GMLP benefiting from performance based regulation,
18

GMLP’s monopoly-like position in natural gas distribution,
19

GMi’s average financial risk profile.
20
21
Noticeably Gaz Metro does not dispute its financial flexibility and states (AR2006, P23) that it
22
“does not have any problem in raising financing nor does it expect any problem in this regard in
23
the future.”
24
Q.
DOES THIS CONCLUDE YOUR TESTIMONY?
25
A.
Yes.
77
SCHEDULE 1
MACROECONOMIC DATA
GDP
UNEMP T BILL
GROWTH RATE YIELD
LONG EXCHANGE PROFITS
CANADAS
RATE
%GDP
AVG
ROE
1983
1984
1985
1986
1987
1988
1989
1990
1991
1992
1993
1994
1995
1996
1997
1998
1999
2000
2001
2002
2003
2004
2005
2.72
5.81
4.78
2.42
4.25
4.97
2.62
0.19
-2.09
0.87
2.34
4.80
2.81
1.62
4.22
4.10
5.53
5.23
1.78
3.09
1.82
3.30
2.94
11.9
11.3
10.5
9.6
8.9
7.8
7.5
8.1
10.4
11.3
11.2
10.4
9.5
9.7
9.1
8.3
7.6
6.8
7.4
7.7
7.6
7.3
7.0
9.32
11.10
9.46
8.99
8.17
9.42
12.02
12.81
8.83
6.51
4.93
5.42
6.98
4.31
3.21
4.74
4.70
5.48
3.85
2.56
2.87
2.23
2.71
11.77
12.75
11.11
9.54
9.93
10.23
9.92
10.85
9.81
8.77
7.85
8.58
8.36
7.54
6.47
5.45
5.68
5.92
5.79
5.67
5.29
5.08
4.41
.811
.772
.733
.720
.754
.812
.845
.857
.873
.828
.775
.732
.729
.733
.722
.674
.673
.673
.646
.637
.714
.768
.816
8.93
10.16
10.24
8.82
10.36
10.58
9.07
6.61
4.80
4.66
5.65
8.49
9.41
9.60
9.96
9.41
11.27
12.63
11.47
11.73
11.91
13.10
13.77
9.34
10.53
10.47
9.49
11.19
12.71
10.88
5.68
2.00
0.18
3.64
7.20
8.04
8.09
9.11
9.30
10.7
11.7
8.9
6.8
11.4
12.0
N/A
2006
2.76
6.6
4.02
4.29
.864
13.75
N/A
-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
"2
00
5
SCHEDULE 2
CPI I nflat ion
14.00
12.00
10.00
8.00
6.00
4.00
2.00
0.00
CPI
SCHEDULE 3
I nt er est R at es and I nfl at i on
20.00
18.00
16.00
14.00
10.00
8.00
6.00
4.00
2.00
T.Bills
Canadas
CPI
"2
00
5"
"2
00
1"
"1
99
7"
"1
98
9"
"1
99
3"
"1
98
5"
"1
98
1"
"1
96
9"
"1
97
3"
"1
97
7"
"1
96
5"
0.00
"1
96
1"
%
12.00
SCHEDULE 4
00
4"
"2
99
8"
"1
99
2"
"1
98
6"
"1
98
0"
"1
97
4"
"1
96
8"
"1
96
2"
"1
95
6"
"1
95
0"
6.00
4.00
2.00
0.00
-2.00
-4.00
-6.00
-8.00
-10.00
"1
% of G D P
Government Net L ending
SCHEDULE 5
CANADA BOND YIELDS
Overnight money market rates
4.25
Benchmark bonds
Canada
91 day Treasury Bill yield
4.36
Canada
Six month Treasury Bills
4.54
Canada
One year Treasury Bills
4.74
Canada
Two year
4.68
Canada
Three year
4.70
Canada
Five year
4.66
Canada
Seven year
4.66
Canada
Ten year
4.64
Canada
Long term (30 year)
4.55
Canada
Real return bonds
2.15
Marketable Bond Average yields
Canada
1-3 year
4.69
Canada
3-5 year
4.68
Canada
5-10
4.65
Canada
Over tens
4.50
US
Five year Treasuries
4.94
US
Long term (30 year)
4.97
Other
Source:
Bank of Canada’s web site at http://bankofcanada.ca/en/securities.htm, for June 12-20, 2007.
SCHEDULE 6
Monetary Conditions Index
25
20
15
10
5
0
1980 J
-5
-10
-15
1983 J
1986 J
1989 J
1992 J
1995 J
1998 J
2001 J
2004 J
"1
95
0
"1 "
95
3
"1 "
95
6
"1 "
95
9
"1 "
96
2
"1 "
96
5
"1 "
96
8
"1 "
97
1
"1 "
97
4
"1 "
97
7
"1 "
98
0
"1 "
98
3
"1 "
98
6
"1 "
98
9
"1 "
99
2
"1 "
99
5
"1 "
99
8
"2 "
00
1
"2 "
00
4"
SCHEDULE 7
Pre-Tax Profits % GDP
16
14
12
10
8
6
4
2
0
Manufacture
Non-farm
2006-Q3
2005-Q4
2005-Q1
2004-Q2
2003-Q3
2002-Q4
2002-Q1
2001-Q2
2000-Q3
1999-Q4
1999-Q1
1998-Q2
1997-Q3
1996-Q4
1996-Q1
1995-Q2
1994-Q3
SCHEDULE 8
Capacity Utilisation
88
86
84
82
80
78
76
74
2007M4
2006M09
2006M02
2005M07
2004M12
2004M05
2003M10
2003M03
2002M08
2002M01
2001M06
2000M11
2000M04
1999M09
1999M02
1998M07
1997M12
1997M05
1996M10
1996M03
1995M08
1995M01
SCHEDULE 9
EXCHANGE RATE
US cents
0.95
0.9
0.85
0.8
0.75
0.7
0.65
0.6
0.55
0.5
2007M4
2007M1
2006M10
2006M07
2006M04
2006M01
2005M10
2005M07
2005M04
2005M01
2004M10
2004M07
2004M04
2004M01
2003M10
2003M07
2003M04
2003M01
2002M10
2002M07
2002M04
2002M01
2001M10
SCHEDULE 10
TSX Composite Index
15000
14000
13000
12000
11000
10000
9000
8000
7000
6000
5000
SCHEDULE 11
16.00
14.00
12.00
10.00
8.00
6.00
4.00
2.00
0.00
BBB Spread
ROE
4
20
0
2
20
0
0
20
0
8
19
9
6
19
9
4
19
9
2
19
9
0
19
9
8
19
8
6
19
8
4
19
8
2
19
8
19
8
0
b a s is p o in ts
400
350
300
250
200
150
100
50
0
P e rc e n t
Corporate ROE and BBB Spread
1 9 /0 1/ 20 0 1
AA
A
BBB
1 9 /0 1/ 20 0 7
1 9 /0 1/ 20 0 6
1 9 /0 1/ 20 0 5
1 9 /0 1/ 20 0 4
1 9 /0 1/ 20 0 3
1 9 /0 1/ 20 0 2
1 9 /0 1/ 20 0 0
1 9 /0 1/ 19 9 9
1 9 /0 1/ 19 9 8
1 9 /0 1/ 19 9 7
1 9 /0 1/ 19 9 6
1 9 /0 1/ 19 9 5
SCHEDULE 12
Canadian Spreads
350.00
300.00
250.00
200.00
150.00
100.00
50.00
0.00
SCHEDULE 13
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 14759.79 18633.87 25602.46 22207.89
2622.3 1388.06
1374.7 3216.84 3325.52 3661.45 3464.11 2394.69 4589.37 3397.35 4092.02 2915.16 5151.24
Debt
13419.72
Capital trust
Limited partnership
Trust units
Total
1993
1994
1995
9309.51 10438.31
1996
14654.6
1997
1998
2001
2002
2003
69801.1 85224.06
94424.6
19457.4 26617.04 34701.19 39223.04 39822.19
2004
2005
105544 92437.44
32373.5 54240.07 60380.84 63884.55
0
0
0
0
0
0
0
3140
1750
2100
1650
602.75
0
114.69
520.37
118.14
407.7
1172.07
690.33
376.57
211.63
516.93
636.32
1876.82
1512.09
1620
0
0
411.03
4264.3 10306.57
1822.81
1498.08
2878.83
106688.2
100997 93275.43 114213.9 120565.5
6996.72 11087.38 16743.64 17028.66 20188.42
117420 128567.3 143153.1 135594.2 149578.4
191661 213585.9 205489.5
SCHEDULE 14
Financing Activity
% of GDP
18.00
9.00
16.00
8.00
14.00
7.00
12.00
6.00
10.00
5.00
8.00
4.00
6.00
3.00
4.00
2.00
2.00
1.00
0.00
0.00
1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005
Total
Private
SCHEDULE 15
EARNED ROE vs ALLOWED
TCPL
Foothills
TCPL BC (ANG)
Allowed Actual Allowed Actual Allowed Actual Allowed
1990
13.25
13.34
14.25
14.25
13.25
13.25
13.75
1991
13.5
13.65
14.25
14.25
13.38
13.38
13.75
1992
13.25
13.43
13.83
13.83
13.43
13.43
13.75
1993
12.25
12.31
11.73
11.73
12.08
12.08
12.25
1994
11.25
11.16
11.5
11.5
12
12
12.25
1995
12.25
12.56
12.25
12.25
12.25
12.25
12.25
1996
11.25
11.83
11.25
11.25
11.25
11.25
11.25
1997
10.67
11.15
10.67
10.67
10.67
10.67
10.67
1998
10.21
10.63
10.21
10.21
10.21
10.21
10.21
1999
9.58
9.64
9.58
9.58
9.58
9.58
9.58
2000
9.9
9.99
9.9
9.9
9.9
9.9
9.9
2001
9.61
10.01
9.61
9.61
9.61
6.86
9.61
2002
9.53
9.95
9.53
9.53
9.53
9.53
9.53
2003
9.79
10.18
9.79
9.79
9.79
8.21
9.79
2004
9.56
10.18
9.56
9.56
9.56
8.51
9.56
2005
9.46
9.66
9.46
9.46
9.46
9.46
9.46
Average 10.96
11.23
11.09
11.09
11.00
10.66
11.10
ovrearn
0.27
0.00
-0.34
NEB Regulated pipelines controlled by TransCanada Corporation.
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
9.84
9.82
11.46
0.36
SCHEDULE 16
Earned vs Allowed ROEs
1990
1991
1992
1993
1994
1995
1996
1997
1998
1999
2000
2001
2002
2003
2004
2005
Over
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
9.69
9.69
9.57
11.01
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
13.14
10.66
9.46
12.17
1.16
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
Terasen
Actual
12.25
na
10.65
12.00
11.00
10.25
10.00
9.25
9.50
9.25
9.13
9.42
9.15
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
10.15
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 is from
VECC #45 and that for Union from Appendix B Schedule 10 of the pre-filed testimony of Dr. William Cannon in RP-2002-0158 updated with interrogatory
answer J2-31.
SCHEDULE 17
Earned Utility Holding Company (UHC) ROEs
CU Ltd
1993
1994
1995
1996
1997
1998
1999
2000
2001
2002
2003
2004
2005
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
12.24
1.26
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
9.03
1.42
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
13.90
2.50
11.84
10.71
10.74
9.61
9.43
7.16
8.56
9.71
12.25
12.24
12.28
11.25
12.39
1.65
GMI
PNG
19.29
19.73
19.50
19.91
18.91
19.11
17.66
17.93
17.45
18.91
18.05
18.21
16.94
0.94
12.92
13.44
11.77
13.32
13.32
10.14
10.79
9.75
7.50
5.94
7.59
6.97
8.34
2.68
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
7.45
4.64
14.01
12.86
13.20
12.33
11.25
7.04
7.42
8.44
10.89
11.93
12.80
15.49
17.56
3.03
Foothills
11.73
11.50
12.25
11.25
10.67
10.21
9.58
9.90
9.61
9.53
9.79
9.56
9.46
0.98
-0.20
-0.40
Electric
Telco
Pipes
Utility
MAY/01
SEP/99
JAN/98
MAY/96
SEP/94
JAN/93
MAY/91
SEP/89
JAN/88
MAY/86
SEP/84
JAN/83
MAY/81
SEP/79
JAN/78
MAY/76
SEP/74
JAN/73
MAY/71
SEP/69
JAN/68
SCHEDULE 18
Index Beta Estimates
1.60
1.40
1.20
1.00
0.80
0.60
0.40
0.20
0.00
SCHEDULE 19
Single and Two Factor Beta Estimates
New TSX Utility Subindex
0.8
0.6
0.4
0.2
-0.4
Beta2
Beta1
D ec-05
D ec-04
D ec-03
D ec-02
D ec-01
D ec-00
D ec-99
D ec-98
D ec-97
D ec-96
D ec-95
D ec-94
D ec-93
-0.2
D ec-92
0
SCHEDULE 20
SUB INDEX BETAS
3.5
3
2.5
2
1.5
1
0.5
Energy
Materials
Industrials
ConsDisc
ConsStap
Health
Fin
IT
Telco
Utilities
Dec-05
Dec-04
Dec-03
Dec-02
Dec-01
Dec-00
Dec-99
Dec-98
Dec-97
Dec-96
Dec-95
Dec-94
Dec-93
-0.5
Dec-92
0
-0.2
-0.4
Gas-One
Pipe-One
Gas-Two
Pipe-Two
MAY/01
SEP/99
JAN/98
MAY/96
SEP/94
JAN/93
MAY/91
SEP/89
JAN/88
MAY/86
SEP/84
JAN/83
MAY/81
SEP/79
JAN/78
MAY/76
SEP/74
JAN/73
MAY/71
SEP/69
JAN/68
SCHEDULE 21
One and Two Factor Beta Estimates
1.6
1.4
1.2
1
0.8
0.6
0.4
0.2
0
20/11/2006
20/11/2005
20/11/2004
20/11/2003
20/11/2002
20/11/2001
20/11/2000
20/11/1999
20/11/1998
20/11/1997
20/11/1996
20/11/1995
20/11/1994
20/11/1993
20/11/1992
20/11/1991
SCHEDULE 22
REAL BOND YIELD
5.5
5
4.5
4
3.5
3
2.5
2
1.5
1
-0.5
Gas-Gamma
-1
Pipe-Gamma
MAY/01
SEP/99
JAN/98
MAY/96
SEP/94
JAN/93
MAY/91
SEP/89
JAN/88
MAY/86
SEP/84
JAN/83
MAY/81
SEP/79
JAN/78
MAY/76
SEP/74
JAN/73
MAY/71
SEP/69
JAN/68
SCHEDULE 23
Gas and Pipeline Sensitivity to Interest Rate Changes
1.5
1
0.5
0
Dec-05
Dec-04
Dec-03
Dec-02
Dec-01
Dec-00
Dec-99
Dec-98
Dec-97
Dec-96
Dec-95
Dec-94
Dec-93
Dec-92
SCHEDULE 24
Utility Gamma
(Interest Rate Sensitivity)
0.9
0.8
0.7
0.6
0.5
0.4
0.3
0.2
0.1
0
SCHEDULE 25
F utur e Values
35.00
30.00
25.00
6%
20.00
10.50%
15.00
12%
10.00
5.00
28
25
22
19
16
13
10
7
4
1
0.00
SCHEDULE 26
Canadian Payout Rates
U s in g TS X Com p os it e dat a
90.00
80.00
70.00
60.00
50.00
40.00
30.00
20.00
10.00
04
20
01
20
98
19
95
19
92
19
89
19
86
19
83
19
80
19
77
19
74
19
71
19
68
19
65
19
62
19
59
19
19
56
0.00
SCHEDULE 27
Dividends and After Tax Profits % GDP
4.50
12.00
4.00
10.00
3.50
3.00
8.00
2.50
6.00
2.00
1.50
4.00
1.00
2.00
0.50
0.00
0.00
1961Q1 1965Q1 1969Q1 1973Q1 1977Q1 1981Q1 1985Q1 1989Q1 1993Q1 1997Q1 2001Q1 2005Q1
DIVS
ATPROFS
2005Q1
2003Q1
2001Q1
1999Q1
1997Q1
1995Q1
1993Q1
1991Q1
1989Q1
1987Q1
1985Q1
1983Q1
1981Q1
1979Q1
1977Q1
1975Q1
1973Q1
1971Q1
1969Q1
1967Q1
1965Q1
1963Q1
1961Q1
SCHEDULE 28
Payout based on Aggregate Profits and Dividends
1.60
1.40
1.20
1.00
0.80
0.60
0.40
0.20
0.00
SCHEDULE 29
US EQUITY MARKET RISK PREMIUM
(USING THE DCF MODEL AND ANALYSTS’ GROWTH FORECASTS)
Claus and Thomas Equity Market Risk
Premiaa
a.
US
Canada
France
UK
1989
3.57
3.08
3.64
3.17
1990
3.54
1.51
3.04
2.57
1991
3.01
0.75
2.94
2.47
1992
3.09
0.42
2.26
2.77
1993
3.65
1.69
2.31
3.29
1994
4.06
1.65
1.7
2.87
1995
3.97
2.71
2.06
3.02
1996
3.45
2.69
2.38
3.34
1997
3.23
2.28
2.28
2.53
1998
2.51
2.68
2.53
2.09
C&T
Average
3.4
2.23
2.6
2.81
J. Claus and J. Thomas, “Equity premia as low as 3.0%? Evidence from analysts’ earnings forecasts for domestic and international stock markets,”
Journal of Finance, October 2001.
SCHEDULE 30
Market Risk Premium Studies
Holding
Dimson, Marsh and
Staunton
Market Risk
Country
Period
Arith/Geom.
Historic/Prospective
Premium
Canada
1900-2000
Arithmetic
Historic
6.00%
U.S.
1900-2000
Arithmetic
Historic
7.00%
U.S.
1985-1998
Prospective
3.40%
Canada
1985-1998
n/a
Prospective
2.43%
U.S.
1951-2000
n/a
Historic
2.55-4.32%
U.S.
1926-2000
Arithmetic
Prospective
5.90%
U.S.
1802-2001
n/a
Prospective
2.40%
U.S.
2001-2011
n/a
Prospective
3.60-4.70%
a
Claus and Thomas
Fama and French
b
c
Ibbotson and Chen
d
Arnott and Bernstein
f
Graham and Harvey
e
n/a
Mean
4.34%
Booth
Canada
1924-2005
Arithmetic
Historic/Prospective
5.00%
a.
E. Dimson, P. Marsh and M. Staunton, Triumph of the Optimists: 101 Years of Global Investment
Returns, Princeton University Press, 2002.
b.
J. Claus and J. Thomas, “Equity Risk Premia as Low as Three Percent? Evidence from Analysts’
Earnings Forecasts for Domestic and International Stocks”, Journal of Finance, October 2001.
c.
E. Fama and K. French, “The Equity Risk Premium”, Journal of Finance, April 2002.
d.
R. Ibbotson and P. Cheng, “Stock Market Returns in the Long Run: Participating in the Real
Economy”, Yale International Center for Finance Working Paper No. 00-44, March 2002.
e.
R.D. Arnott and P.L. Bernstein, What Risk Premium is Normal?”, Financial Analyst Journal,
March/April 2002.
f.
J.R. Graham and C.R. Harvey, “Expectations of Equity Risk Premia, Volatility and Asymmetry
from a Corporate Finance Perspective”, Fuqua School of Business Working Paper, Duke
University, November 2001.
SCHEDULE 31
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 32
Market to Book Ratios for UHCs
3
2.5
2
1.5
1
0.5
0
1995
1996
1997
1998
1999
2000
2001
CUL
Fortis
GMI
PNG
Enbridge
TCPL
Emera
Average
2002
2003
Terasen
2004
TAU
2005