A Tale of Two Provinces
THIS IS A WORK IN PROGRESS
Extremely Rambling, Very Drafty Notes
Do not quote or cite under any circumstances!
Kenneth J. McKenzie
University of Calgary
Over the past several years the governments of Alberta and Ontario have lead the
country in the area of fiscal reform, particularly in the area of tax policy. Notably, they
have been among the most aggressive tax cutting in the country. Although the
approaches to tax reform followed in the two provinces share some common
characteristics, in some important and fundamental ways they are also very different. My
purpose in this paper is to compare and contrast the approaches to fiscal reform followed
in these two provinces, with an emphasis on taxation.
I begin by looking at the big picture in the two provinces. I then turn to an
analysis of specific tax policy initiatives in the areas of personal and corporate taxation.
The approach I take is to analyze the tax policy initiatives in the two provinces in terms
of their impact on labour markets and capital markets. This is, I think, a more natural
way to think about tax policy in the two provinces.
I
The Big Picture
My starting point for the comparison is the election of Ralph Klein as premier of
Alberta in 1993 and the election of Mike Harris as premier of Ontario in 1995. Even
those with a casual awareness of the political scene in Canada recognize the election of
Progressive Conservative government’s under the leadership of these two gentlemen as a
watershed in the provincial fiscal environment in this country.
What was the state of affairs in Alberta and Ontario just prior to the election of
Klein and Harris? Reference to a couple of key figures paints a fairly dismal fiscal
picture in each case. Ontario had a run of five consecutive deficits in excess of $10
billion prior to 1995, which caused the net debt/GDP ratio in the province to rise by eight
percentage points, from 22% in 1991/92 to over 30% in 1994/95. Prior to 1993 Alberta
had a run of eight consecutive deficits, ranging from $1.3 billion to over $4 billion, which
caused the net debt/GDP ratio to rise from 2% in 1986/87 to 15% by 1992/93, a
substantial thirteen percentage point deterioration. In each case this was to be the high
water mark on the net debt/GDP ratio. Under the Klein government Alberta moved into a
net asset position of over 3% of GDP by 1999/00; a remarkable eighteen percentage point
turnaround in just seven years. Harris’s government oversaw somewhat less dramatic,
but still significant five percentage point reduction in the net debt/GDP ratio to 25% by
1999/00.
How did they do it? In order to answer this question it is useful to present the
standard equation of motion for the net debt/GDP, which can be derived from the
government budget identity:
Δbt = bt+1-bt= (gt – τt – ft + (r - α)bt)/(1+α)
where b is the net debt/GDP ratio, g is government program spending as a percentage of
GDP, τ is provincial own source tax revenue as a percentage of GDP, f is federal
transfers as a percentage of GDP, r is the interest rate on government debt and α is the
growth rate in GDP. This equation makes it clear that there are three key sources of
growth for a provincial government’s debt/GDP ratio. One is the primary deficit (g-τ). A
positive value for the primary deficit indicates that program spending exceeds tax and
transfer revenue so that in this fiscal year the government has made a net addition to its
debt. A negative value for the primary deficit indicates the opposite; tax and transfer
revenue exceeds program spending so that in this fiscal year the government has made a
net reduction to its debt. The second key variable is the interest rate payable on
outstanding debt. As this interest rate increases, the annual cost of servicing the debt
increases. If nothing else changes, the annual deficit will therefore grow and debt will
accumulate more quickly. The third key variable is the rate of growth in GDP. The more
quickly the economy grows, the smaller will be the debt to GDP ratio. As equation the
equation indicates, it is the relative magnitudes of the interest rate payable on government
debt (r) and the rate of growth in GDP (α) which is key. Intuitively, if the interest rate
exceeds the rate of growth, the addition to debt resulting from servicing outstanding debt
exceeds the increase in GDP due to economic growth. Thus, the ratio of debt to GDP
rises.
I employ this simple budgetary equation in several ways. The first is to use it to
obtain an expression for what is known as the sustainable tax rate (Blanchard 1993). The
sustainable tax rate is the tax rate (tax revenues as a percentage of GDP) required to
finance a given level of program spending (as a percentage of GDP) while maintaining
the existing debt/GDP ratio. It is obtained by setting Δbt=0 in the above expression and
solving for τt, which gives
τ*t = gt – ft + (r - α)bt
This equation is intuitive. It says that in order to maintain existing program spending as a
proportion of GDP (gt) and keep the net debt/GDP ratio from changing, the provincial
government’s own source revenues as a percentage of GDP (the sustainable tax rate τ*t )
must cover program spending not financed by transfers from the national government
(the amount gt – ft) plus the excess of debt service costs over economic growth ((r-α)bt).
Thus, the higher is program spending not financed by transfers and the higher is the cost
of debt service relative to the economy’s growth rate, the higher is the sustainable tax
rate.
Given the definition of the sustainable tax rate, we can then determine whether
existing fiscal policy is sustainable or not by comparing the sustainable tax rate, τ*t, to the
actual tax rate, τt (the actual own source tax revenue to GDP ratio). If the tax gap, τ*t – τt,
is positive -- i.e., the sustainable tax rate exceeds the actual tax rate -- then the current
fiscal policy stance of the government is not sustainable in the sense that the net
debt/GDP ratio will be expected to increase over the coming year unless changes are
made to either tax or spending policy. On the other hand, if the tax gap is negative, then,
by our definition, the current fiscal policy stance is also not sustainable in the sense that
the net debt/GDP ratio will be expected to decrease over the coming year unless changes
are made to either tax or spending policy. The tax gap thus tells us what fiscal
adjustment is required in order to maintain the current debt/GDP ratio.
Figure’s 1 and 2 show the net debt/GDP ratio and the tax gap in Ontario and
Alberta respectively, from 1988/89 to 1999/00. The Klein Effect in 1993 and the Harris
Effect in 1995 are obvious. Prior to 1993 Alberta was running a substantial positive tax
gap, and as a result the debt/GDP ratio was rising accordingly. The Klein government
moved quickly to turn things around, as we see a modest negative tax gap in 1993
followed by substantial negative gaps thereafter. As indicated above, the result was an 18
percentage point turn around in the debt/GDP ratio by 1999/00. The Harris government
also quickly moved to a negative tax gap, which resulted in a 5 point reduction in
debt/GDP over the five year period.
As has been well documented, both government’s implemented substantial
spending cuts following their election. As shown in Figure 3, program spending (total
spending net of debt service) as a percentage of GDP fell in Alberta from 17.2% in
1992/93 to 12.7% in 2000. In Ontario the ratio of program spending to GDP fell from
14.8% in 1994/94 to 11.8% in 2000. The size of the cuts in Alberta is particularly
noteworthy.
On the revenue side, Alberta chose to forgo tax cuts – and indeed only
implemented significant cuts to personal taxes, and to a lesser extent corporate taxes, in
2001. The Alberta approach was instead to concentrate on paying down the debt before
giving tax relief. Ontario followed a different approach, choosing to implement tax cuts
first, primarily for low and middle income earners (more on this below). This, of course,
delayed the pay down of debt, and partly accounts for the more modest reduction of the
debt/GDP ratio. As shown in Figure 4, own source tax revenue as a percentage of GDP
has been relatively constant throughout the reign of both the Klein and Harris
governments. Indeed, there has been a modest rise in taxes as a percentage of GDP in
Ontario, despite that tax reductions. This is due to strong overall economic growth,
fueled largely by the lengthy economic expansion in the U.S. over this period. In Alberta
the fairly large increase in the revenue to GDP ratio in 199/00 was due largely to the rise
in oil prices.
Up until recently, the Klein and Harris (and other) governments have been
fortunate to implement their fiscal reforms in an environment of relatively robust
economic growth. An interesting question then concerns the extent to which the
reduction in the debt/GDP ratio can be attributed to above normal economic growth
throughout the period, as opposed to the discretionary fiscal policies of the government.
There are two ways in which growth can affect the debt/GDP ratio. First, as
indicated in the equation of motion for the debt/GDP ratio from above, the economic
growth decreases the ratio directly simply by increasing the denominator. Second, above
normal economic growth can feed through to higher taxes and lower program
expenditures through the so-called automatic stabilizing effect. Using an approach
suggested by Fortin (1998) and refined by Kneebone and Leach (2001), it is possible to
use the equation of motion for the debt/GDP ratio to decompose the change in the ratio
into its various constituent parts. We are particularly interested in the extent to which the
improvement in the debt/GDP ratios under Klein and Harris over the periods from
1993/94 and 1995/96 to 1999/00 respectively was due to good fortune as opposed to good
fiscal management.
To do this we must come up with an estimate of potential GDP for the two
provinces over the periods in question, and determine the extent to which actual GDP
exceeded potential – i.e., we need to determine the extent to which GDP growth was
above normal. There are several complicated ways of doing this. Kneebone (2001),
however, points out that running a simple Hodrick-Prescott filter through the GDP series
is a simple approach that yields an estimate of potential GDP that is very similar to that
which arises under more sophisticated approaches. Figure 5 shows actual and potential
GDP for Ontario from 1980/81 through to 1999/00. Particularly noteworthy is the fact
that Ontario GDP was actual lower than potential from 1990/91 to 1997/98. Only in the
last two years of the Harris regime has Ontario exhibited above normal growth.
Alberta’s experience throughout the Klein regime has been similar. As shown in
Figure 6, the province grew roughly at potential from 1992/93 through 1996/97, below
potential from 1996/97 to 1998/99 and above potential thereafter. Table 1 decomposes
the change in the net debt/GDP ratios in Alberta from 1993/94 to 1999/00 and in Ontario
from 1996/97 to 1999/00 into that due to discretionary fiscal policy, as measured by the
difference between cyclically adjusted program spending and own source revenues
expressed as a percentage of GDP (g-τ), and that due to the difference between the
interest rate and economic growth (r-α)b.
Table 1: Change in Net Debt/GDP in Alberta Under Klein and Ontario Under
Harris
Alberta (1993/94-1999/00)
Ontario (1996/97-1999/00)
-18.30
-5.49
Due to Fiscal Policy
-13.67
-3.03
Due to Growth (r-α)b
-4.63
-2.46
Change in Net Debt/GDP
(Δb)
(g-τ)
Of the 18.30 point improvement in net debt/GDP in Alberta over the Klein years, fully
13.67% of that, or 75%, was due to improvements in fiscal policy, as measured by the
primary deficit (g-τ). As indicated above, the bulk of this was due to expenditure
reductions. The remainder was due to economic growth and lower interest rates. Of the
5.49 percentage point turnaround in net debt/GDP in Ontario, 3.03 of that, or 55%, was
due to improvements in fiscal policy; the remainder was due to economic growth and
lower interest rates on government debt. Of course the initial size of the debt/GDP ratio
in Ontario at the beginning of the Harris reign (30% vs. 15% at the beginning of the Klein
reign) accentuates the interest rate and growth effect.
The difference in the magnitudes of the reduction in the net debt/GDP ratios in the
two provinces, and the difference in the way that those reductions were achieved,
highlights some important differences in the approach to fiscal policy followed in the two
provinces. Alberta relied largely upon expenditure reductions which were used to pay
down government debt. Economic growth (and lower interest rates) added little to the
improvement in the fiscal situation in Alberta. Ontario, on the other hand, while also
cutting spending also reduced taxes. This lowered the contribution of fiscal policy to the,
much smaller, improvement in the debt/GDP ratio. As such, Ontario relied substantially
more on economic growth to improve its fiscal situation.
Of course, all of this leaves unanswered the role that fiscal policy itself plays in
stimulating provincial growth. The approach reflected in Table 1 presumes that potential
GDP is itself unaffected by fiscal policy. In small open economies such as Alberta and
Ontario, and over the short period of time considered here, this is a reasonable
assumption. In the longer term, one might expect fiscal policy in general, and tax policy
in particular, to feed through to affect both the level and growth of potential GDP.
II
The Labour Market: Labour Supply and Migration
There are two dimensions that bear upon the issue of taxation and work
incentives. The first concerns the overall impact of taxation on the incentive to work. I
refer to this as the standard labour market efficiency dimension. The second dimension
concerns the impact of the tax system on the incentives for individuals to move across
borders - commonly referred to in the popular press as the “brain drain”, but what
economists more mundanely refer to as fiscally induced migration. While these two
issues are obviously linked, I will try and deal with them separately. I begin with the
standard labour market efficiency dimension.
Much of the discussion of tax policy at the provincial level in both Alberta and
Ontario has focused not so much on the level of taxation but on the structure of the tax
system, and in particular on the degree of marginal rate progressivity in the personal
income tax. This has important implications for the efficiency of labour markets, and
therefore for the standard of living in the two provinces. Quite simply, the more
progressive the rate structure, the greater is the disincentive to work, the more inefficient
is the tax system, and the lower the amount of labour income produced in the economy.
This feeds directly to the bottom line in terms of the standard of living.
In terms of the first dimension of taxation and work incentives - the standard
labour market efficiency dimension - the choice of the appropriate degree of progression
in the tax system boils down to the classic equity-efficiency trade-off. How much
income (standard of living) are we prepared to sacrifice in order to pursue distributional
objectives?
In terms of the second dimension of the relationship between taxation and work
incentives issue into play – fiscally induced migration – the mobility of labour across
borders and its responsiveness to differences in the fiscal environment (not just taxes)
across jurisdictions is an often debated issue, which has not been fully resolved.
Recently, most of the discussion in the popular press has focused on the international
dimension, particularly on the possible “brain drain” from Canada to the U.S. However,
it would appear obvious that labour is much more mobile intra-nationally, within Canada
and between provinces, than internationally. Indeed, Tanis Day and Stan Winer
document evidence that Canadians (particularly Anglophone Canadians) are quite
responsive to differences in net fiscal benefits (the value of government services less
taxes) across provinces.i As such, the “brain drain” is likely a much bigger factor at the
provincial level than the federal level.
The importance of the “brain drain” at the international level has been the subject
of considerable debate. Without entering that debate directly, it would seem that despite
the standard labour market efficiency implications discussed above, there is scope for at
least some degree of marginal tax rate graduation at the federal level, though perhaps less
than currently exists. The potential for fiscally induced migration suggests that at the
provincial level this scope is reduced. This insight has, I think, very important
implications for the design of personal tax systems at the provincial level. To put it
simply, it suggests that the use of the personal tax system as a redistributive mechanism
should be confined largely to the federal government, and that provincial governments
should not use the tax system to redistribute income. Thus, although the federal
government can, and arguably should, impose a progressive marginal rate structure, the
provinces should not.
It is interesting to consider the structure of the tax systems in Alberta and Ontario
in this context. As shown in Table 2, in terms of the basic rates Alberta takes a simple
approach: after the initial exemption of $12,900 all income is taxed at a single 10% rate.
The basic rate structure in Ontario is quite different. A lower initial exemption is
provided, but the marginal tax rates up to just over $61,000 in taxable income are lower
than Alberta. For taxable income in excess of roughly $61,000 the marginal tax rate is
higher, at 11.16%. (Note that all of the following tables are for single, working age
individuals).
Table 2: Basic Provincial Personal Income Tax Rates in Alberta and Ontario, 2001
Province
Alberta
Ontario
Tax Rates and Brackets
Exemption
$12,900
10.0%
All income
Exemption
$7,368
6.20%
$0 – 30,814
9.24
30,815 – 61,629
11.16
61,630 and over
But Table 2 does not tell the whole story. Unlike Alberta, Ontario levies surtaxes
on “high income” individuals. A surtax is a “tax on a tax.” Ontario levies two surtaxes,
one at 20% on basic provincial taxes greater than $3,569, and one at 36% on basic
provincial taxes greater than $4,491. These surtaxes have the impact of increasing the
top marginal rate for Ontario from the 11.16% indicated in Table 2, to over 17.4%. Table
3 shows the combined federal and provincial marginal tax rates in Alberta and Ontario
taking basic taxes and the Ontario surtaxes into account.
Table 3: Total Federal and Personal Income Tax Rates, 2001
Taxable Income
Lower Range
$0
7369
12901
30755
61510
100000
Upper Range
$7368
12900
30754
61509
99,999
Max
Taxable Income
Lower Range
$0
7369
30755
53601
61510
63201
100000
Upper Range
$7368
30754
53600
61509
63,200
99,999
Max
Alberta
Federal
0
16.0%
16.0%
22.0%
26.0%
29.0%
Ontario
Federal
0
16.0%
22.0%
22.0%
26.0%
26.0%
29.0%
Provincial
0
0
10.0%
10.0%
10.0%
10.0%
Provincial
0
6.2%
9.24%
11.088%
13.92%
17.4096%
17.4096%
Total
0
16.0%
26.0%
32.0%
36.0%
39.0%
Total
0
22.2%
31.24%
33.088%
39.92%
43.4096%
46.4096%
The biggest differences emerge in the “upper income” ranges, in excess of approximately
$63,000 in taxable income. For taxable income in excess of $100,000, the marginal tax
rates facing Ontario taxpayers are substantially higher than in Alberta, to the tune of
almost 7.5 percentage points.
While marginal tax rates are important for the overall efficiency of the tax system,
average tax rates are important both for distributional reasons and for location decisions,
which are inframarginal in nature. Under the residence principle for personal income
taxation followed in Canada individuals cannot allocate a portion of their personal
income across jurisdictions - it is an all or nothing proposition. Thus, in assessing the
“competitiveness” of the tax system in, say, Ontario vis-a-vis Alberta from this
perspective, it is the average effective tax rate that matters, not the marginal tax rate.
While the marginal rate structure obviously affects the average tax rate, so too do various
other things, most particularly the income exemptions.
Figure 7 shows the average tax rates in Alberta and Ontario for various household
income levels. These calculations are based on Statistics Canada’s SPSD/M model, and
therefore incorporate several other aspects of the tax/transfer system than the personal
income tax rates discussed above. As the figure shows, for lower income households the
average tax rate in Alberta is lower than Ontario, for middle income households (up to
about $80,000 in household income) the average tax rates are the same, while for upper
income households the average tax rates in Alberta are lower.
The discussion of taxes and labour markets to this point has focused on the
personal income tax. But corporate income taxes can also affect labour markets in
important ways. As will be discussed in more detail below, there are reasons to believe
that the long run incidence of provincial corporate income taxes lies largely with workers.
The idea is that in the long run capital is quite mobile across provinces. Thus, a decrease
in the CIT rate will eventually and over time attract capital into a province, this increase
in capital will in turn increase the demand for workers, which will increase employment
and wages.
However, what is not widely appreciated is that given the way that the Canadian
CIT allocation formula works, it turns out that workers in a province that lowers its CIT
can benefit from a reduction in the CIT rate even in the short run, even before capital
floods into the province in response to a reduction the CIT rate (or even if said capital
flows don't take place at all).
When a corporation operates in more than one province, we need a way in which
to allocate the taxable income earned by that corporation across the provinces. The way
that we do it in Canada is by way of an allocation formula. The allocation formula in
Canada is based on the corporation's share of sales and wages in each province.
Specifically, the share of corporate taxable income allocated to a province is fifty percent
of that province’s share of sales plus 50% of its share of the corporate payroll.
The implications of this allocation formula for the burden of the corporate tax at
the provincial level are not widely appreciated. As will be discussed in more detail
below, in 2000 Ontario announced its intention to lower its CIT rate on larger
corporations to 8% over five years (the general rate is currently 14% while the
manufacturing rate is 12%). In 2001, Alberta announced its intention to do the same (the
current rate in Alberta is 13.5% on all large corporations). While Ontario’s announced
rate reduction was not the only reason for Alberta’s announced CIT rate cut, it played an
important roll. In particular, the CIT cut in Alberta was justified in part because of the
fact that the reduction in the CIT rate in Ontario would act effectively like a tax on some
workers employed in Alberta. Moreover, this tax on Alberta labour would be imposed
immediately, and does not require the outflow of capital from Alberta to Ontario that
might be expected in the long run.
The reason for this is that if the CIT rate in Ontario is lower than Alberta, for
corporations operating in both provinces the cost of hiring an additional worker in
Alberta consists not only of the wages paid to that worker, but also of the higher
corporate income taxes the corporation must pay due to the increase in the average CIT
rate under the allocation formula.
The best way to see this is with a simple example. For simplicity, say that a
corporation operates in just two provinces - Alberta and Ontario. To begin, lets say that a
company operates in each province, employing 1 worker in each province at a salary of
$10 per worker (for a total corporate payroll of $20). Each worker generates sales of $50
in the province in which she works (for total sales of $100).
Case 1: Alberta CIT rate equal to Ontario CIT rate
First consider the case where both Alberta and Ontario have equal CIT rates of 50%. The
corporation's total taxable income is $80, which is simply its total sales of $100 less its
total payroll of $20. Using the allocation formula described above each province's share
this taxable income is 50%, calculated as follows:
SA=(.5 x (50/100) + (.5 x (10/20)) = .50
And similarly for Ontario.
So, taxable income allocated to Alberta is $40 (=.5x$80), and corporate income
tax paid to Alberta is $20 (=.5x$40). Similarly for Ontario. Total corporate income taxes
paid by the corporation in Alberta and Ontario is thus $40, in this case split evenly
between the two provinces. The corporation's total tax rate across both provinces is 50%
(total taxes paid divided by taxable income, or $40/$80).
Now say that the corporation decides to hire one more worker in Alberta. This
worker can also produce $50 in sales, which gives $100 in sales in Alberta. With the
existing $50 in sales in Ontario, total corporate sales are now $150. Total corporate
payroll is now $30 ($20 in Alberta and $10 in Ontario). Thus, taxable income is now
$120, which is simply total sales of $150 less total payroll of $20.
Using the allocation formula Alberta's share of this taxable income is 2/3,
calculated as follows:
SA=(.5 x (100/150)) + (.5 x (20/30))=2/3
And Ontario's share is therefore 1/3.
So, taxable income allocated to Alberta is $80 (=2/3 x $120) and taxable income
allocated to Ontario is $40 (=1/3 x $120). CIT paid in Alberta is $40 (=.5x$80) and CIT
paid in Ontario is $20 (=.5x$40), for total CIT of $60 and a total tax rate of 50% across
both provinces (=$60/$120). The cost to the corporation of hiring the worker in Alberta
is the after-tax wage rate of $5 (=$10x(1-.5)), which accounts for the fact that wages are
deductible for CIT purposes in both provinces.
Case 2: Alberta CIT rate greater than Ontario CIT rate
Now lets go through the same exercise again, but this time assume that the Ontario CIT
rate is 40% and the Alberta CIT rate is 60%. As above, in the initial situation where the
corporation employs one worker in each province, each province's share of taxable
income of $80 is 50%, or $40. Alberta CIT is $24 (=.6x$40) and Ontario CIT is $16
(=.4x$40), for a total tax bill of $40 and an average tax rate of 50%. While the
corporation's total corporate taxes, and the average CIT rate, are the same as Case 1 (I
chose these tax rates for just that reason), more CIT is paid to Alberta because it has the
higher tax rate.
Now consider hiring an additional worker in Alberta. As determined above, $80,
or 2/3, of the resulting taxable income of $120 is allocated to Alberta, and $40, or 1/3, of
the taxable income is allocated to Ontario. Alberta CIT is now $48 (=.6x$80), Ontario
CIT is $16 (=.4x$40), for a total tax bill of $64, and a combined Alberta plus Ontario tax
rate of 53 1/3% (=$64/$120).
The first thing we notice is that when Alberta has a CIT rate that is greater than
Ontario, hiring one more worker in Alberta causes the combined, or average, CIT rate
facing the corporation to increase from 50% to 53 1/3%. This is because the weight
applied to Alberta's now higher CIT rate increases, which increases the average CIT rate.
How much does it cost to hire that extra worker in Alberta in this case? The aftertax wage is now $4.67 (=$10x(1-.533)), which is lower than Case 1 because the average
CIT rate is higher, but this also means that the corporation now has to pay higher
corporate taxes on its income as well. The additional taxes due to the higher average CIT
facing the corporation amount to $4 (=.0333x$120). Thus, when the Alberta CIT rate is
greater than the Ontario CIT rate, the total cost of hiring that extra worker in Alberta is
$8.67 - $4.67 in after tax wages plus another $4 in higher CIT due to the way the
allocation formula works - instead of the $5 we calculated in Case 1 when both provinces
had the same CIT rate.
What this means is that because Ontario has a lower CIT rate than Alberta, for
corporations with operations in both provinces the cost of hiring an additional worker in
Alberta is now higher than it was before Ontario lowered its rate. This means, of course,
that these corporations will be less likely to hire that worker in Alberta and the Ontario
CIT rate cut has effectively imposed a tax on additional workers hired in Alberta. Note
well that none of the above calculations involved capital flowing out of Alberta and into
Ontario in response to the lower CIT rate in Ontario, nor any attempt on the part of the
corporation to otherwise allocate more income to the lower taxed province (Ontario).
III
The Capital Market: Savings and Investment
I turn now to the capital market, which concerns the savings decisions of individuals and
the investment decisions of firms. There are two important issues that need to be
addressed in this regard. The first concerns the structural similarity of the personal
income tax in Canada to a consumption tax. The second concerns the relationship
between savings and investment, and the implications of personal and corporate income
taxes, in a small open economy. I will deal with each in turn.
There are several aspects of the personal income tax system in Canada that act to
mimic consumption taxation. The most obvious example is the tax sheltered treatment of
RRSP/RPP contributions. Satya Poddar and Morley English estimate that as much as
75% of investment income in Canada may go untaxed, and therefore be effectively
subject to consumption tax treatment.ii This suggests that if the personal income tax in
Canada is largely in fact a personal consumption tax, then it impinges very little on the
savings decisions of taxpayers.
It turns out that this inference may not in fact be correct. Economists are well
known for their obsessiveness over the distinction between the notions of average and
marginal; in this case the distinction is quite important. The 75% figure quoted above is
an average - on average, 75 cents out of every dollar of assessed investment income is not
subject to income taxation for some reason. Yet efficiency costs and distortions should
be measured at the margin. To determine the efficiency effects of taxing the return to
savings at the personal level, the effective rate of tax on the rate of return generated by an
additional, or marginal, unit of saving must be determined. Poddar and English also
report that over half of taxable dividend and two-thirds of taxable capital gains are earned
by individuals with assessed incomes in excess of $100,000; the vast bulk of investment
income is earned by individuals with income in excess of $50,000. Yet these are also the
individuals who are most likely to have exhausted the “easy” ways of tax sheltering their
investment income. While it is possible that for some of these individuals an additional
dollar of income from savings will escape taxation altogether, it is more likely that an
incremental dollar will attract tax in some form, for example because of the exhaustion of
RRSP/RPP limits. Thus, even though the average effective tax rate on capital may be
very low, even close to zero, the marginal effective tax rate could be quite high. It is the
latter that determines the efficiency costs of taxation. Indeed, the very fact that any
investment income in Canada attracts tax at the personal level, let alone 25% of it,
suggests that many of the individuals who are saving in Canada have in fact exhausted
their ability to shelter this income.
Recent empirical support for this view comes indirectly from a paper by Michael
Veall.iii Veall investigates whether the flattening of the Canadian personal tax structure
in the 1988 tax reform had any effect on RRSP contributions. He found “no convincing
evidence that these changes affect RRSP contributions.”iv While it is important to
emphasize that he did not examine the impact of the tax changes on overall savings, just
savings in the form of RRSPs, Veall’s empirical result is consistent with the
interpretation that any increase in savings in Canada in response to reduced taxes on
investment income tends to take place in non-RRSP vehicles at the margin.
Table 4 presents calculations of the marginal effective tax rates on the real return
to savings at the personal level for interest, dividends and capital gains for the highest
income brackets in Alberta and Ontario. The calculations reflect the taxation of the
inflationary component of the return to savings, the presence of the dividend tax credit,
and the deferral effect of taxing capital gains on realization rather than upon accrual.v
The calculations show that at the margin the effective tax rate on savings under the
existing personal income tax are substantially higher in Ontario than in Alberta. This is
because of the high marginal tax rates levied on Ontario taxpayers because of the
existence of the surtaxes.
Table 4: Real Marginal Effective Tax Rates on Savings, 2001
Ontario
Alberta
Interest
65.0%
54.6%
Dividends
46.2%
33.3%
Capital Gains
25.7%
21.5%
Note: Assumes a nominal rate of return of 7%, inflation of 2%, and (for capital gains) a
holding period of 10 years.
Canada is a “small economy” because savings and investment in the country are
only a small share of the world capital market. Moreover, it is often characterized as
“open” because national savings can be used to finance investment anywhere in the
world, and domestic investment can be financed by savings from anywhere in the world there are few capital controls preventing flow of capital in to or out of Canada.vi
Canada’s place as a “small open economy” in an international context is fairly noncontroversial. A key question, on which there is more debate, is the degree of mobility of
financial capital across the country’s borders.
This is important. In a small open economy, when financial capital is mobile
there is a “disconnect” between the supply (savings) and demand (business investment)
sides of the capital market. Indeed, if financial capital is perfectly mobile, this disconnect
is complete and there is no connection between the two sides of the market from a
domestic perspective.
It is extremely important to understand the implications of this “disconnect” for
tax policy, in particular provincial tax policy. If capital is in fact perfectly mobile as
described above, it means that on the supply (savings) side of the capital market the
before-tax rate of return earned by savers in Canada is fixed from our perspective,
determined by the international financial market. On the demand (business investment)
side of the market, it means that the after corporate tax rate of return that businesses must
pay investors in order to attract their savings is also fixed, determined by the international
financial market. What this implies is that, unlike a closed economy, or an economy with
immobile capital, the savings of Canadians do not necessarily end up being invested in
Canadian companies, and even when they are an increase or decrease in the supply of
those savings has no impact on the after corporate tax rate of return that businesses must
provide to their investors. Rather, decreases or increases in Canadian savings show up
simply as a change in the proportion of domestic business investment financed by
Canadians, the remainder being financed by foreign savings. As such, while taxes
imposed on the return to savings at the personal level will distort the supply of savings by
Canadians, and generate the usual type of efficiency costs associated with these
distortions, when capital is perfectly mobile there is no impact on business investment in
Canada, and the efficiency costs are confined to the supply side of the capital market.
Similarly, changes in the taxation of business investment in Canada will distort capital
investment in the country, and generate the associated efficiency costs, but these effects
will be confined to the demand side of the capital market and will have no impact on the
saving decisions of Canadians.
Importantly, perfect capital mobility also suggests that the burden of taxes levied
on the investment (demand) side of the market, will eventually be borne completely by
less mobile factors within Canada, such as labour. This is because the after-corporate tax
rate of return on domestic investments is fixed by the international market. An increase
in the corporate tax rate on domestic capital will simply drive capital out of the country,
in order to keep the after-tax rate of return the same, which will lower the demand for less
mobile factors such as labour, which will in turn lower wages. Perfect capital mobility in
a small open economy thus has stark implications for the analysis of the taxation of
capital.
Key to this analysis is the degree of capital mobility into and out of Canada. As
mentioned above, this is a matter of some debate. John Helliwell and Ross McKitrick
provide a summary of the issues.vii They argue that if capital in Canada is perfectly
mobile then there should be no correlation between national savings and investment rates,
and the so-called “savings retention rate” should be close to zero (due to the “disconnect”
discussed above). However, empirical investigations on OECD countries, including
Canada, consistently uncover a strong positive correlation between domestic saving and
investment. Helliwell and McKitrick, for example, estimate a national savings retention
rate of around 0.60, which is significantly different than zero. This suggests that
Canadian savings do tend to manifest themselves in domestic investment. Although
Helliwell and McKitrick discuss several attempts to reconcile this “stylized fact” with the
presence of perfect capital mobility - and indeed, they discuss several theoretical models
which show that the presence of common shocks etc., can explain reconcile the two
views - the empirical results suggest that the perfect capital mobility view may not be a
completely accurate depiction of the Canadian economy. This in turn suggests that when
the country is viewed as a whole, the “disconnect” between the two sides of the capital
market may not be complete. It is likely that Canada is best described as a small open
economy with “fairly” (not perfectly) mobile capital. The implications of this
characterization have not been fully explored in the public finance literature.
What is more clear, however, is that the small open economy with perfect capital
mobility model provides a much better description of provincial economies. Helliwell
and McKitrick show that the savings retention rate, or correlation between savings and
investment, for individual provinces is statistically indistinguishable from zero. This
strongly suggests that capital crosses provincial borders in a fashion very similar to that
described by the perfect capital mobility model.
In light of the previous discussion, the implications of this for tax policy at the
provincial level are clear. In particular, it suggests that individual provinces will find it
difficult to increase “domestic” (provincial) business investment by lowering personal
taxes on saving. This does not mean that personal income taxes levied on savings are
unimportant. Rather it simply means that the efficiency gains in the capital market
arising from personal income tax cuts will be confined to supply (savings) side of the
market.
It also has important implications for the role of provincial corporate income
taxes, levied on the demand side of the capital market. In particular, it suggests that
changes to provincial income taxes designed to encourage both savings and investment
requires lower personal income taxes and lower corporate taxes.
Recent studies suggest that corporate taxes have a significant impact on business
investment decisions, particularly those of multinational corporations.viii Provincial
corporate taxes add from 9 to 17 percentage points to the overall basic income tax rate
facing Canadian corporations, and in light of the previous discussion, the provinces can
play an important role in enhancing Canada’s competitive position internationally by
reducing their corporate tax rates. This is particularly important given the increasingly
prevalent view among economists that technological advances are embodied in new
capital, and therefore that capital investment translates directly into productivity gains.
This suggests that reduced taxes on corporate capital at the provincial level may have an
even bigger impact on growth and the standard of living than first thought.ix
As indicated above, both Alberta and Ontario have announced their intention to
reduce their corporate tax rates to 8% on all non-small corporations. The current general
rate in Ontario is 13.5%, while manufacturing firms face a lower rate of 12%. Alberta
imposes a tax rate of 13.5% on all (non-small) corporations. There are, however, some
important differences in corporate tax policy in the two provinces. I will focus on two
here.
The first concerns the presence of capital taxes. Ontario levies a capital tax on
large corporations equal to 0.30% of paid up capital. There has been no indication that
the capital tax in Ontario will be eliminated in conjunction with the move to a general
corporate tax rate of 8%. Alberta does not levy a capital tax. The second concerns the
presence of corporate tax incentives for research and development (R&D). Currently
Ontario offers very generous incentives on top of already generous federal incentives.
Alberta offers no specific incentives for R&D.
When thinking about the competitiveness of corporate tax systems across
jurisdictions, most people focus on statutory corporate income tax (CIT) rates. These are
the tax rates set out in the Income Tax Act. While important, the statutory CIT rate is
only one part of the corporate tax system. Equally important is the tax base, which is
determined by the various rules that govern the rate and nature of various deductions and
write-offs against corporate revenue. There may also be tax credits associated with
certain types of investments that further reduce corporate tax liability directly. Generous
write-offs and credits can negate the impact of a high statutory tax rate. Moreover, many
jurisdictions impose other taxes on capital, such as explicit capital taxes, that are not
taken into account in a simple comparison of statutory CIT rates.
As indicated above, both of these considerations are important in a comparison of
corporate taxes in Alberta and Ontario.
One way of taking all of this into account, and therefore of comparing the overall
competitiveness of business tax regimes in various jurisdictions, is to calculate the
marginal effective tax rate (METR) on corporate capital. The idea behind the METR is
conceptually quite simple. It employs the notion of the hurdle rate of return. Investors
have many opportunities for investment, and in order to attract their savings corporations
must generate an expected rate of return that at least compensates investors for their
forgone investment opportunities - the hurdle rate of return is the minimum aftercorporate tax rate of return required to just compensate investors for their forgone
investment opportunities. For this reason, economists refer to the hurdle rate of return as
the opportunity cost of finance. All investment projects must generate a rate of return
that is at least as great as the opportunity cost of the funds used to finance the project.
Corporate taxes impinge upon the hurdle rate of return by lowering the income available
to investors. For example, say that the after-corporate tax hurdle rate of return is 5%.
This is to say that after the payment of corporate taxes, shareholders require an expected
rate of return of at least 5% in order to entice them to invest in the corporation. Now say
that after taking account of the various write-offs, deductions and credits allowed under
the CIT, paying taxes at the relevant statutory CIT rate, and paying any other taxes
imposed on the capital (such as explicit capital taxes), in order to generate a rate of return
of 5% after the payment of corporate taxes, corporations need to generate a rate of return
of 10% before the payment of corporate taxes. The METR on capital in this case is 50%,
calculated simply as (10%-5%)/10%. The METR thus measures the tax wedge driven
between the before- and after-corporate tax rate of return on marginal investment
projects, where a marginal project is simply an investment that just earns the required
hurdle rate of return after the payment of corporate taxes. In this example, the METR of
50% means that the after-tax rate of return on a marginal investment is 50% lower than
the before-tax rate of return.
A positive METR means that the tax system discourages investment, by taxing
the return to a marginal project. A negative METR means that the tax system encourages
investment, by subsidizing the return to a marginal investment. In either case the tax
system introduces distortions (or inefficiencies) into the economy by causing investment
to be either lower or higher than it would be in the absence of taxation. A METR of zero
means that the tax system is neutral with respect to investment - i.e., it neither
discourages nor encourages investment - and does not impinge upon the return to a
marginal investment. Calculating METRs allows us to assess the relative of
competitiveness of the business tax regime across jurisdictions using a single, and
sensible summary measure that accounts not only for differences in statutory CIT rates
across jurisdictions and types of investment, but also for differences in tax bases due to
write-offs and credits, and for differences in direct capital tax rates. This is particularly
important for interjurisdictional comparisons, where focusing on statutory tax rates alone
can be very misleading because of differences in the tax base as well as the presence of
other types of taxes on capital.
Table 5 shows some METR calculations for Ontario and Alberta for
manufacturing and non-manufacturing capital both currently and when the announced
reduction to an 8% CIT rate is fully phased-in. The calculations reflect both federal and
provincial corporate taxes. The rate cut calculations also take account of federal rate cuts
that will lower the federal corporate tax rate to 21% over the next couple of years.
Table 5: METRS on Capital, Alberta and Ontario
Alberta
Ontario
Manufacturing – Current
21.0%
25.3%
Manufacturing – Rate Cuts
17.3%
23.1%
Non-manufacturing – Current
29.0%
33.8%
Non-manufacturing – Rate Cuts
19.8%
25.8%
The calculations emphasize the very important role played by capital taxes in the
determination of METRs. In particular, the presence of Ontario’s capital tax adds from
four to five percentage points to the overall METR on capital in that province vis-à-vis
Alberta. Indeed, currently, despite the lower tax rate in the non-manufacturing sector in
Ontario, the METR on non-manufacturing in Alberta is lower.
The METR illustrated in Table 5 were calculated on an aggregation of capital in
the two provinces, in particular machinery and equipment, buildings and structures, land,
and inventory. As suggested above, corporate tax policy in Ontario is also characterized
by a tendency to offer special incentives for particular sectors. This is evident in the
current system in the lower tax rate levied on non-manufacturing corporations (though
this will disappear when the rate cuts are fully phased-in). It is also evident in its
treatment of R&D expenditures, which receive very generous investment allowances over
and above the federal tax credit for R&D. By way of contrast, the approach in Alberta
has been to offer a level playing field, with no sector specific tax rates and no incentives
for R&D.
The implications for this in the case of R&D are illustrated in Table 6, which
presents calculations of METR on R&D capital for Ontario and Alberta.x Two sets of
calculations are presented. The first is referred to as the METR on R&D costs. This
measures the extent to which the tax system lowers the cost of producing an incremental
unit of intangible R&D capital. The second is the METR on R&D capital. This
measures the extent to which the tax systems impinges upon the incentive to invest in a
marginal unit of intangible R&D capital.
Table 6: METRs on R&D Costs and R&D Capital: IN PROGRESS
Ontario
Alberta
METR on R&D Costs
METR on R&D Capital
i.See Tanis Day and Stan Winer, “Internal Migration and Public Policy” in Issues in the
Taxation of Individuals, A. Maslove ed.(University of Toronto Press), 1994.
ii.Satya Poddar and Morley English, “Canadian Taxation of Personal Investment
Income”, Canadian Tax Journal 47(5), 1999, 1270-1304.
iii.Michael Veall, “Did Tax Flattening Affect RRSP Contributions?”, Social and
Economic Dimensions of an Aging Population Working Paper, McMaster University,
1999.
iv.Ibid, page 11.
v.See McKenzie, supra footnote 18.
vi.An obvious exception is the 20% foreign content limitation for RRSPs.
vii.John Helliwell and Ross McKitrick, “Comparing Capital Mobility Across Provincial
and National Borders”, Canadian Journal of Economics 32(5), November 1999, 11641173.
viii.See Robert Chirinko and Andrew Meyer, “The User Cost of Capital and Investment
Spending: Implications for Canadian Firms”, in P. Halpern, ed., Financing Growth in
Canada (University of Calgary Press), 1997; Michael Wasylenko, “Taxation and
Economic Development: The State of the Economic Literature”, New England Economic
Review, March/April 1997, 49-52; Rosanne Altshuler and Jason Cummins, “Tax Policy
and Dynamic Demand for Domestic and Foreign Capital by Multinational Corporations”,
Technical Committee on Business Taxation Working Paper 97-4 (Department of Finance
Canada); and Jason Cummins, Kevin Hassett and Glenn Hubbard, “Tax Reforms and
Investment: A Cross-Country Comparison”, Journal of Public Economics 62(1-2), 23773.
ix.See Jason Cummins, “Taxation and the Sources of Growth: Estimates from United
States Multinational Corporations”, NBER Working Paper No. W6533, April 1998.
x
The approach to calculating R&D METRs in the table is based on McKenzie, K.J.
(2000), “Measuring Tax Incentives for R&D”, University of Calgary, mimeo.