An (Overdue) Review of Canada’s Fiscal Stabilization Programm
Le Programme de stabilisation fiscale du gouvernement fédéral a pour vocation d’aider les provinces frappées par une baisse majeure et imprévue de leurs revenus. Le programme est d’autant plus nécessaire que les provinces sont désormais souvent soumises à des chocs économiques qui leurs sont propres. Cependant, la portée du programme n’a cessé d’être réduite depuis sa création en 1967 de sorte qu’il offre aujourd’hui un soutien quasi négligeable. Les critiques du programme déplorent surtout le plafonnement de ses paiements à 60 dollars par habitant, mais une réflexion en profondeur s’impose. Le plafond devrait être éliminé et le gouvernement devrait s’engager dans une réforme plus fondamentale, par exemple en établissant les paiements de stabilisation en fonction de la capacité fiscale ou de la performance économique des provinces plutôt que de leurs revenus.
Ensuring governments have sufficient revenues to carry out their responsibilities is one of the central challenges of fiscal federalism in Canada. Persistent differences in economic strength mean some provinces have an easier time raising revenue than others. The equalization program helps bridge those gaps, and is regularly updated to reflect changing circumstances. For temporary shocks, such as recession, however, the one program explicitly designed to provide assistance — the fiscal stabilization program — has atrophied to near insignificance. Fiscal stabilization gives partial protection to provinces whose revenues drop suddenly from one year to the next. If their revenues fall enough, the federal government will provide additional fiscal transfers to those provinces. All provinces have received stabilization payments at some time or another since the program was introduced in 1967, but changes over time have severely limited the level of support it now offers.
Consider recent revenue drops in Alberta, Saskatchewan and Newfoundland and Labrador. Alberta’s annual own-source revenues, for example, fell nearly $9.2 billion (21 percent) between fiscal years 2014-15 and 2016-17, largely because low oil prices reduced both resource revenues and the income tax base. In response, the stabilization program made two payments totalling $500 million, but these covered only a fraction of Alberta’s loss. Saskatchewan’s own-source revenues dropped by $1.2 billion (10 percent) over the same period, and the province received only $20 million in stabilization. Pressure for reform of the stabilization program has thus been growing, especially since the December 2019 meeting of provincial and territorial premiers, and federal officials are now exploring options.
What, if anything, should be done? The current program is limited for two main reasons: first, a ceiling of $60 per capita on payments; and, second, a deductible of 5 percent for nonresource revenues and 50 percent for resource revenues. Both of these features have a material effect on the size of the program. For perspective, payments to Alberta in fiscal years 2015-16 and 2016-17 would have totalled $2.75 billion were it not for the cap, or nearly $7 billion had the original 1967 formula still been in place. This is not to say that either payment would have been optimal or that the past formula suits today’s fiscal and economic realities, but it illustrates how limited provincial revenue stabilization is today.
Should we remove the cap on payments? Should we reduce the deductibles? Does the program, in fact, need radical reform? If so, what are some alternatives? I answer these questions with reference to the latest data and analysis. I also explore Canada’s rich history and experience with federal-provincial transfers in general, and with stabilization in particular. This history matters. The cap on payments, for example, appears due to no explicit analysis, but was merely an afterthought following the experience of a single province for a single fiscal year. Today, all aspects of the program need to be given a hard look. This paper explores a wide variety of reform options, along with the inevitable trade-offs of any policy change. To appreciate the relative merits of the options, we must start with some first principles.
SHOULD CANADA HAVE A STABILIZATION PROGRAM?
There are various arguments for and against fiscal stabilization. First, a central argument in its favour is risk pooling. Smoothing out economic shocks across individuals can increase overall welfare — after all, people are generally risk-averse, and therefore willing to pay to avoid wild swings between good times and bad. Pooling risk across governments also has benefits because volatile fiscal policy can harm the overall economy. Adjusting to economic and fiscal shocks requires higher taxes or lower spending or higher debt to bridge good times and bad. The federal government, by its sheer size and fiscal capacity, is able to better absorb such shocks. The borrowing rate on federal debt is typically a full percentage point lower than that on provincial debt, for example. The mere existence of the fiscal stabilization program can also serve to lower provincial borrowing rates and make it easier for provinces to access credit. Alternatively, if one does not want to absorb shocks using government debt, regional shocks require smaller changes in federal tax and spending measures than would be required of a single province. This matters economically. Taxes create distortions, and large taxes create especially large ones. Holding tax rates steady is therefore more efficient than increasing them in bad times and decreasing them in good — to say nothing of the potentially negative macroeconomic effects of such countercyclical rate changes.
Second, equity considerations provide arguments both for and against stabilization. Canadians might wish to bear collectively the burden of adverse shocks that are beyond their control, regardless of where they live. Many of the costs of natural disasters are pooled nationally, for example, and stabilization provides a mechanism to pool the costs of regional recessions — indeed, this was its initial primary motivation. What about stabilization payments to higher-income provinces, however? Alberta has significant fiscal capacity to smooth its own finances over the business cycle, and, in a very important sense, chooses to expose itself to revenue shocks by relying heavily on oil and gas royalties to fund government operations. It could, for example, adopt a stable and efficient source of revenues — such as a sales tax — and save resource revenues in a fund. And lower-income provinces, as we will see shortly, already have access to implicit stabilization through the equalization program.
Regardless of one’s position on these points, this is a conversation that is increasingly important today. Provinces have always differed in their level of economic strength, but now they also differ widely in their year-to-year growth. Figure 1 illustrates how provincial growth rates of gross domestic product (GDP) have increasingly differed from one another since 1961. Specifically, I measure the difference between each province’s growth rate and the national average, expressed as a share of that average. This relative mean deviation, as it is known, provides an intuitive comparison over time. As the figure shows, the typical difference between provincial GDP growth rates and the national average was fairly low in the 1960s and 1970s, at roughly one-quarter of the average. That is, for each 1 percent of national GDP growth, provincial growth was an average of 0.25 percent higher or lower. Today, the typical difference approaches 0.65 percent per point. This is large.
Provincial economies are therefore increasingly subject to their own unique shocks. Stabilization is a means of transferring payments from provinces experiencing positive shocks to those experiencing negative ones. There are trade-offs, to be sure. Stabilization policy can encourage provinces to take risks when the consequences are offloaded to others. Avoiding such moral hazard is central to insurance design, and should be for stabilization, too. Resource revenues, in particular, present a challenge. Alberta, Saskatchewan and Newfoundland and Labrador rely on resource revenues to fund public services. These revenues, however, are highly volatile, so transferring the consequences of their decline to other provinces might discourage resource-rich provinces from adopting more resilient budgets. Alberta, for example, could adopt a stable and efficient source of revenues — such as a sales tax — and save resource revenues in a fund.
With all these considerations in mind, I explore in this paper various options to reform Canada’s stabilization program, ranging from simple tweaks to wholesale changes. This analysis is not exhaustive, and readers should view it in light of other work, including most recently that by Bev Dahlby. Fundamentally, I emphasize that reforms should strive not to insulate provinces from the consequences of their risky budget decisions, and should focus more on exogenous economic shocks beyond the provinces’ control. Stabilization policy should also not be viewed in isolation from other transfer programs. Recent work by James Feehan for the IRPP on equalization reform is valuable to consider. Alberta is a recurring example throughout this paper, but I make no argument for or against larger payments to that province for their own sake. Rather, Alberta provides a useful case study because of the large recession it recently underwent, and because pressure to reform the stabilization program is driven primarily by the Alberta government.
Before turning to the modern program, I begin where all robust federal policy analysis should: in the past. In a country as large and complex as Canada, understanding the origins of current policy design can reveal the economic, social and political pressures that reforms must withstand.
THE PAST: THE ORIGINS OF STABILIZATION
Canada’s history of providing revenue guarantees to provincial governments is long. In this section, I’ll briefly explore the period during and immediately following the Second World War, before turning to the modern programs that began in 1967.
Early revenue guarantees: 1941-66
Every federal-provincial fiscal arrangement since the 1941 Wartime Tax Rental Agreement has featured some form of provincial revenue guarantee. In that 1941 agreement, provinces ceded to the federal government all income tax and succession duties. In exchange, the federal government provided fixed annual payments equivalent either to what those taxes raised for the provinces in 1941 or to enough to service their debt. Even if their incomes fell, provincial revenues were protected. Guarantees were also extended to other provincial revenue sources. Gasoline tax revenues were at particular risk because of wartime rationing, a federal gas tax increase and other measures. The federal government therefore committed to topping up provinces whose gas tax revenues fell below their 1940 level. It was a wartime expedient.
After the Second World War, stabilization took a more deliberate and formulaic approach. Both the 1947 and 1952 arrangements featured guaranteed minimum payments to provinces that refrained from raising their own income taxes. This was, in part, an incentive for provinces to accept federal dominance in that field. But it was not to last. With the war now over, provinces (especially Quebec) proved unwilling to cede their income tax room permanently. There were delays and disputes, and pressure as a result of the Korean War to maintain strong central finances, but by 1957 Canada was firmly on the road to decentralizing once again.
That year marked the start of equalization and provincial revenue stabilization as one would recognize them today. The policy took a simple form at first: to stabilize the yield on “standard taxes” — taxes on personal income, corporate income and succession duties at certain rates determined by the federal government — and equalization payments. If such revenues fell by more than 5 percent relative to the average of the previous two fiscal years, then the federal government would make up the difference. In those early postwar years, however, provinces had less fiscal autonomy than they have today. Stabilization, therefore, initially was protection against federal policy changes and recession. But as provinces gained more independence and finances decentralized further, policy had to keep up.
Developing the current stabilization program: 1967 to the present
Beginning in 1967, fiscal stabilization grew to cover almost all provincial revenues — the program as it is today. The original formula was simple. If provincial revenues, adjusted for any tax changes, fell by more than 5 percent, the federal government would cover losses beyond the initial 5 percent drop. Using an intuitive analogy, this is like insurance for provincial revenues with a 5 percent deductible — although provinces do not pay a premium for coverage, so the insurance analogy is imperfect. Mathematically, the payment to province i at time t would be
where is adjusted revenues for the current fiscal year (to eliminate the effect of any tax changes compared to the previous fiscal year), and Rit–1 is the previous year’s revenues. Following this initial formulation, these notable developments occurred:
- in 1972 the 5 percent deductible was eliminated;
- in 1977 a 50 percent deductible was added for natural resource revenues;
- in 1987 a $60 per capita limit on stabilization payments was implemented; and
- in 1995, the 5 percent deductible on nonresource revenues was reinstated.
The formula today is more complex than when it started. If resource revenues increase while total revenues decline by more than 5 percent, then the 1967 formula applies. If resource revenues decline by less than 50 percent while nonresource revenues decline by more than 5 percent, the stabilization payment is again determined by the 1967 formula, but based only on nonresource revenues. Finally, if resource revenues decline by more than 50 percent, then payments are the greater of zero or
where are nonresource and resource revenues (adjusted for tax policy changes), respectively. And in all cases, payments are limited to no more than $60 per capita. This limit is a central concern around stabilization today, but its history is not well known.
Why a $60 per capita limit?
Despite its importance today, parliamentarians paid almost no attention to the $60 per capita cap when it was implemented in 1987. Tom Hockin, Minister of State for Finance in Prime Minister Brian Mulroney’s government, explained when introducing the relevant bill in Parliament: “Some adjustments are being made to improve and clarify the [fiscal stabilization] program in light of difficulties which have arisen in its administration, but in administration only. Should any future stabilization payments be very large, any assistance in excess of a base amount will also take the form of an interest-free loan.” To the minister, the $60 per capita ceiling was merely a demarcation between a “base amount” (a grant) and an interest-free loan. But why $60?
In all House of Commons and Senate floor and committee transcripts, only once was the cap explained. That one instance is worth quoting at length. On April 13, 1987, during a meeting of the Senate Special Committee on National Finance, Jim Lynn, director general of the Federal-Provincial Relations and Social Policy Branch, explained to a senator that, “previously, the stabilization program was a straight grant program. The province became eligible and calculated what the decline in their revenues was from one year to another.” The senator asked, “You are referring to this $60 amount?” Lynn replied, “Yes…the $60 per capita grant and everything above that will be an interest-free loan repayable. The $60 was based on the B.C. experience.” The senator replied, “Okay. I am satisfied with that.” And with that, everyone moved on.
In short, based on these statements by the relevant minister and the responsible public servant, the $60 per capita limit appeared to be an administrative arrangement to divide stabilization payments between a grant (limited to $60 per capita) and a loan (any excess). But, in practice, the legislation provides the minister of finance with full discretion to extend loans or not, and no loan payment has ever been made. Worse, there was no objective justification for the $60 ceiling beyond that it was the amount British Columbia received in fiscal year 1982-83. Easing the cap should top the list of reforms.
Since the BC payment, all provinces have received stabilization payments at some time or another. In all, there have been 18 stabilization payments totalling nearly $2.6 billion (see table 1). In 4 instances, payments were capped by the $60 per capita limit. The largest payment (per capita) was $419.2 million to Alberta for its revenue decline in fiscal year 1986-87. Despite Alberta’s initially applying for stabilization after the $60 cap was implemented in 1987, the cap was binding only on payments related to fiscal years beginning after March 31, 1987. Since Alberta’s claim was with respect to 1986-87, it was not subject to the cap. And with the province’s population of 2.4 million, this payment was equivalent to $174 per capita. Adjusted for inflation, that is over $350 today — nearly six times the current $60 limit. Although the next dozen payments under the program — all for fiscal years 1990-91 through 1993-94 — were subject to the cap, it was binding only in two cases — Nova Scotia in 1991-92 and Prince Edward Island in 1992-93. Stabilization payments would not be made again for more than two decades.
THE PRESENT: STABILIZATION POLICY IN PRACTICE
Today, assistance to provinces that are experiencing a temporary economic shock is provided both explicitly through stabilization and implicitly through equalization.
Explicit support: Fiscal stabilization payments
Alberta’s recent recession was severe, and revenues declined substantially: between fiscal years 2014-15 and 2016-17, resource revenues declined by $5.9 billion while nonresource own-source revenues declined by $3.3 billion. Changes in revenues are one thing, but stabilization payments are meant to buffer provinces from exogenous events, not from policy changes. For example, Alberta introduced a carbon tax on January 1, 2017, three months before the end of fiscal year 2016-17. According to the public accounts for that year, those three months generated $250 million in carbon tax revenues. So, while revenues from general and miscellaneous sales taxes, amusement taxes, carbon taxes, and so on were nearly $190 million higher in 2016-17 than in 2015-16, they would have been $60 million lower had it not been for the $20 per tonne carbon tax. The government of Jim Prentice also increased gasoline and diesel taxes from 9 cents to 13 cents per litre, effective March 27, 2015, while the tax on propane went up from 6.5 cents to 9.4 cents per litre. Taxes on tobacco and liquor increased, too. Later, the New Democratic Party government elected in May 2015 made notable increases in personal and corporate income taxes. And finally, though on a more minor note, Alberta’s personal income tax system was indexed (until the 2019 budget), which meant that the value of credits and brackets grew with inflation. This indexation reduced revenues relative to a static income tax system, and for the purposes of stabilization the forgone revenues were added back. It is tough to estimate precisely the total effect of all these policy changes, but Finance Canada officials estimate the net effect added $1.09 billion to Alberta’s revenues in 2015-16 and $1.17 billion in 2016-17. The total drop in the province’s revenues as a result of exogenous economic developments was therefore larger than actual revenue declines alone suggest.
Tax policy changes are not the only factor to consider. Not all provincial revenues are included in determining stabilization payments — property taxes, for example, are not, and neither is investment income. And even some provincial income tax revenues are excluded because of an accounting fiction created by historic tax point transfers when the federal government lowered its tax rates as provinces simultaneously increased theirs. For the purposes of stabilization (but for almost nothing else) the federal government considers this accounting fiction a transfer, not provincial revenue.
So, with all these factors considered, table 2 reports Alberta’s relevant revenue changes and adjustments for fiscal years 2015-16 and 2016-17. Adjusted revenues declined by $7.2 billion in 2015-16 due mostly to a decline in resource revenues. Since only drops above 50 percent are covered, the deductible that year left a stabilization amount of $1.6 billion. But the $60 per capita limit lowered this to $248 million. In 2016-17, the drop in total provincial revenues was due entirely to a decline in nonresource revenues of $2.2 billion. After a smaller 5 percent deductible, the resulting stabilization payment would have been nearly $1.2 billion, but the cap on payments meant that only $251 million was paid. Overall, between 2015-16 and 2016-17, total stabilization payments to Alberta were $2.25 billion lower than they otherwise would have been.
Implicit support: Stabilization through equalization
Other policies also provide some fiscal stabilization. In particular, equalization not only addresses persistent differences in fiscal capacity between provinces; it also implicitly compensates for certain temporary economic shocks. Specifically, equalization pays provinces with below-average fiscal capacity to bring them up to the national average, and is based on a rolling three-year average (with 25/25/50 weights) with a two-year lag. That is, payments in fiscal year 2020-21 will be based on data for 2016-17, 2017-18 and 2018-19. Temporary shocks, even if they last only one or two years, are therefore eventually reflected in equalization payments to recipient provinces.
To illustrate this clearly, consider a simple example involving a two-year-long recession that lowers provincial revenues by 10 percent. Presume that no policy changes occur and, therefore, that the drop in revenues entirely reflects smaller tax bases — that is, lower incomes due to the recession. Table 3 illustrates two small provinces, one rich and the other poor, to abstract from the effect of the provincial shock on the national average.
As table 3 shows, equalization compensates low-income provinces for exogenous economic shocks well — in fact, within five years of a recession’s end, they are fully compensated. High-income provinces, however, are insured only by the current stabilization policy, as they (rightly) cannot access equalization on account of their high incomes. To be clear, this is an illustration. Countless other factors affect equalization payments, so even low-income provinces will not see the kind of clean compensation for a temporary recession illustrated here. But the program nonetheless provides them at least some (and perhaps full) compensation for recessions.
THE FUTURE: A MENU OF POLICY OPTIONS
Stabilization policy is, at its core, something like an insurance arrangement. Although the provinces pay no premiums, principles of insurance design are useful guides. As in any insurance contract, behaviour changes in potentially risky ways if the consequences are borne by others. If they are insured against revenue declines, for example, provinces might be more likely to adopt policies that exacerbate such declines. That response is not only inefficient but also potentially unjust. Such moral hazard concerns are central to optimal insurance design, but they must be weighed against other important factors.
Federal-provincial transfer arrangements, moreover, have their own unique principles that have long guided policy. First, fiscal resources available to each order of government should be sufficient to discharge their constitutional responsibilities. Second, each order of government should be accountable to its own electors for its taxing and spending decisions. And third, policy should be uniform in its application across provinces. Applying these principles, however, is not straightforward. If a province’s revenues collapse, then federal support helps it undertake its responsibilities (the first principle). But some volatility might be a choice — for example, Alberta’s reliance on resource revenues — so insuring fully against such choices offloads the consequences onto other governments, and therefore violates the second principle.
In this section I explore a number of reform options with these principles in mind. I start with minor changes to the current formula before proceeding to more radical reforms.
Short-term changes: Adjusting the current formula
Tweaks to the current formula would be easiest to make. For example, the ceiling on payments could be eased or deductibles lowered. Although simple, each would have large implications: eliminating the cap completely would have resulted in a payment of $2.75 billion to Alberta for fiscal years 2015-16 and 2016-17. Of course, a ceiling might be desirable to limit risk to federal finances. Equalization payments are capped, for example, but at a level that increases with national GDP, which ensures long-run sustainability relative to the federal government’s ability to pay. As well, stabilization payments could be indexed, perhaps to economic growth, to provincial revenues or to inflation. If the $60 per capita cap had increased along with national nominal GDP per capita, for example, it would be roughly $170 today. If it had increased along with total covered provincial revenues, it would also be $170 today. And if it had merely kept pace with inflation, it would be $120 today. Table 4 illustrates the effect of each alternative cap. With a cap indexed to inflation, total payments to Alberta for 2015-16 and 2016-17 would have been $1 billion. If indexed to economic growth, they would have totalled $1.42 billion. And without any cap at all, they would have been $2.75 billion.
What about different deductibles? A variety of formulas has been used in the past, as described above. The original 1967 formula had a 5 percent deductible for all revenues. This would have resulted in nearly $7 billion in payments to Alberta for fiscal years 2015-16 and 2016-17. By 1972, the formula featured no deductibles at all, which would have meant a combined $9.38 billion in payments. The 1977 formula introduced a 50 percent deductible for resource revenues, and kept no deductible for nonresource revenues; this would have produced $4.66 billion in total payments. And in 1995, the 5 percent deductible for nonresource revenues was reinstated. If the program featured only this 5 percent nonresource revenue deductible and a 50 percent resource revenue deductible, payments would have totalled $2.75 billion — the “no cap” case in table 4.
What is the appropriate deductible? The lower the deductible, the higher the stabilization payments and the more frequently they are made. But low inflation, low population growth, demographic change and so on mean that baseline growth rates today are lower than in the past. From 1972 (when the 5 percent threshold was eliminated) to 1991 (when the Bank of Canada’s inflation targeting began), annual increases in consumer prices averaged 7.2 percent. In the years since 1991, inflation has averaged 1.8 percent annually. This matters. With slower price increases, provincial revenues are much more likely to decline. To show this, figure 2 illustrates the distribution of provincial GDP growth rates over the past six decades.
Even during the severe 1980-82 recession, only one province experienced a negative nominal GDP growth rate. And for all years between 1961 and 1995, there was roughly a 3 percent chance that any given province was experiencing a decline in nominal GDP. Since 1995, that probability has increased to 8 percent. Interestingly, the frequency of declines in provincial nominal GDP greater than 5 percent since 1995 is the same as that of declines in excess of 0 percent in the quarter-century before. The 5 percent threshold reintroduced in the 1995 federal budget therefore created roughly the same probability of the relevant triggering event as did the prior 0 percent threshold.
A look at the distribution of growth rates can tell us what the threshold should be. Table 5 reports some rough estimates of the threshold required to trigger a qualifying event once every certain number of years, based on post-1995 data. For example, if policy-makers are comfortable insuring against shocks that occur only once every 20 years, then the threshold should be set at 2 percent.
For resource revenues, a drop of 50 percent or more is also a rare event. Among the provinces with meaningful resource revenues, only in 1986 and 2015 were declines that large. That amounts to less than 5 percent of observations over the 1982-2018 period for the three oil-producing provinces. A 50 percent natural resource revenue decline is therefore as rare an event as a 2 percent decline in nonresource revenues.
But since deductibles also ensure that provinces bear some consequences of revenue declines, deciding on the appropriate level involves more than just estimating how frequently various revenue drops occur. Having a high deductible for resource revenues is also a means of encouraging responsible provincial budgeting and insulating the federal government from the effect of provincial policy choices. This is the primary motivation for the substantially higher deductible for resource revenues. And it is an eminently sensible feature of the current program.
Easing the cap or lowering deductibles would be a quick and easy adjustment to the stabilization program, but either change would leave its functioning largely unchanged. Two more fundamental reforms, however, would significantly alter the nature of stabilization payments and, therefore, would take more time to implement.
Deeper changes, option 1: Stabilizing fiscal capacity
Stabilization attempts to insure provincial revenues against shocks beyond the provinces’ control, so the program has always been adjusted for changing tax rates and structures. But some taxes are more volatile than others, and the program does not currently consider the different composition of revenue sources across provinces. Instead, the program could insure what a province would raise if its tax rates and structures were the national average. This is the same as the concept of “fiscal capacity” used in equalization. Such a “representative tax system” approach to stabilization has never been attempted, but it is worthy of consideration. Formally, such a program would pay
where fit is the per capita fiscal capacity in province i and Sit is the per capita stabilization payment. Figure 3a shows provincial fiscal capacity using the default equalization treatment (which includes 50 percent of resource revenues). The figure also shows that, when fiscal capacity declines, it does so typically by less than 5 percent (figure 3b).
For a sense of how such a formula would operate today, table 6 reports counterfactual stabilization payments to Alberta for fiscal years 2015-16 and 2016-17 based on this fiscal capacity proposal, using different revenues, resource inclusion rates and other design details. The current stabilization program excludes property taxes, but including them would dampen measured volatility and therefore reduce stabilization payments. Using the same definition of fiscal capacity as in the equalization formula, which includes property taxes, I find that total payments to Alberta in 2015-16 and 2016-17 would have been between $2.7 billion and $9.5 billion, depending on the deductible, the resource revenue inclusion rate and whether a moving average is used.
A two-year moving average is a particularly interesting option to consider. As noted, equalization already provides some implicit stabilization to receiving provinces, but with a two-year lag. A two-year average for stabilization would bridge the two programs. To see the effect, recall the example of the high- and low-income provinces, and consider a stabilization program with no deductible, a two-year moving average and with payments included in equalization’s measure of fiscal capacity.
As illustrated in table 7, such an arrangement would fully compensate both high- and low-income provinces for very short negative shocks — those lasting no more than a year — and fully insure low-income provinces against shocks regardless of how long they lasted, while transitioning seamlessly between transitory and permanent shocks. For high-income provinces, no deductible would provide full insurance against brief shocks and gradually declining insurance against longer-lasting shocks. That is, persistent shocks are better dealt with by equalization, while this moving-average approach would gradually phase out stabilization payments to provide time for the province to adapt gradually and efficiently.
Countless variations of this approach exist — higher thresholds for resource revenues could be used, for example. But, as with any policy choice, there would be trade-offs. Because fiscal capacity is a relative measure that uses the national average tax rate applied to each province’s own tax bases, policy changes in large provinces could result in lower measured fiscal capacity even if a province experiences no negative shock. Conceivably, fiscal capacity estimates could be adjusted to reflect the national average tax rate in prior fiscal years instead. But a simple alternative would be to link stabilization to changes in the overall economic conditions of a province, an option I turn to next.
Deeper changes, option 2: A macro-based formula
The original goal of stabilization was to buffer provinces against macroeconomic shocks. To quote federal Finance Minister Mitchell Sharp in 1967, “[t]he provinces have argued that in periods of recession when revenues are falling, the costs of provincial responsibilities continue to rise…Stabilization is [therefore] to be provided on a continuing basis as a protection to the provinces against sharp declines in revenue.” This is a sensible goal. And stabilization payments could be based directly on macroeconomic measures such as provincial GDP. This was difficult in the past, when high-quality provincial economic accounts did not readily exist. Today they do. To be sure, it would be very similar to basing payments on fiscal capacity, but with the added benefit of simplicity. Such a formula would also avoid many moral-hazard concerns with insuring revenues directly. Consider the following:
where γ = Rit −1 / Git −1 is the revenue share of GDP. This is effectively the same formula as prevailed in 1967, but with adjusted revenue in the current fiscal year given by γ × Git, rather than the item-by-item estimate compiled by Finance Canada. This parameter could also reflect a representative tax system by setting it equal to some national average Rt −1 / Gt −1. This has been roughly 12 percent in recent years (excluding property taxes). In this case, the above expression may be written simply as
where gt is national GDP growth, Sit is the province’s share of national GDP and Rt−1 is total provincial revenues subject to stabilization during the last period.
In practice, this macro approach to stabilization would rarely result in payments to non-oil-producing provinces. I estimate that, since 1981, this formula would have paid out only eight times, and only to Alberta, Saskatchewan and Newfoundland and Labrador. It would have paid $3.83 billion in fiscal year 2015-16 and $424 million in 2016-17. Dampening the federal government’s exposure to provincial resource revenue shocks could be achieved by basing the formula on the same set of provincial revenue sources used in the equalization formula — say, by including property taxes and only 50 percent of resource revenues. Whatever the details, a macro-based stabilization formula could be considered, just as such an approach to equalization has been contemplated numerous times in the past.
Other considerations in reforming the stabilization program
A number of other important considerations need to be explored in making changes to the stabilization program. I consider four here: federal affordability, access to federal borrowing, retroactive payments and the treatment of resource revenues.
Ensuring sustainable finances, especially over the long run, is critically important. Merely boosting the fiscal stabilization program without making other changes elsewhere might create unnecessary risks for the federal government. But affordability concerns should not be overstated. The federal government is in a very strong position. Despite recent deficits, the long-term trajectory of the federal budget is sound. The Parliamentary Budget Office’s 2018 fiscal sustainability report estimates that Ottawa could immediately and permanently lower taxes or increase spending by 1.4 percent of GDP ($29 billion per year). There is thus ample room for increases in federal transfers to provinces that are in a much more precarious fiscal position.
There could be savings, particularly if stabilization and equalization were integrated. The equalization program, for example, currently operates with a fixed pool of total payments to go around. If the formula determines smaller payments are warranted, the remaining dollars — called “adjustment payments” — are distributed across recipient provinces on an equal per capita basis. Between 2018 and 2020, nearly $2.8 billion in adjustment payments was paid. Depending on how inequality in provincial fiscal capacity evolves, such payments might be with us for some time, so eliminating the fixed pool could yield savings to help fund an expanded stabilization program. To be sure, this change would create visible winners and losers, as well as more volatility in equalization payments, which would make it more difficult for recipient provinces to plan. But stabilization could play a role here, too, as it would be a function not only of provincial own-source revenues but also of equalization payments. If equalization payments were to fall rapidly in some year — perhaps because of a recession in a large province — then stabilization payments would cushion the blow.
More flexible debt allowances
Government debt bridges bad times and good, and federal debt faces lower interest rates than provincial debt — typically, federal long-term debt yields are a full percentage point below those of provincial debt. Suppose provinces could borrow through the federal government (outside of emergency situations). Even in today’s low-rate environment, such an arrangement would be valuable. At a rate of, say, 3 percent, providing the excess of the fiscal year 2015-16 and 2016-17 stabilization payments to Alberta above the $60 per capita cap as an interest-free, five-year loan — as initially intended by the 1987 stabilization program changes — would have been worth nearly $340 million to the province. Of course, retroactively applying an interest-free loan might not be a prudent path, but at the very least the federal government would be wise to clarify the conditions under which it would provide loans through the program. If designed well, with reasonable terms and repayment schedules, a federal window for provincial borrowing would not necessarily increase either federal risk or bond yields, but it would give provinces greater flexibility to smooth out shocks.
Premiers are not merely asking for stabilization reform; they want changes retroactive to 2015. There is nothing unusual about such a request, and there is a long history of retroactive payments to settle federal-provincial disputes over transfers. Perhaps the first instance involved the 1873 increase in the Constitution’s debt allowance to settle a dispute between Ontario and Quebec. In 1884, this increase was made retroactive to 1867, and federal transfers to all provinces were consequently increased. Alberta’s history also features an interesting example when, between 1938 and 1940, the province did not receive fiscal need subsidies because it had defaulted on its debt in 1936. In 1945, after the debt was reorganized and the default dispute settled, the federal government retroactively gave Alberta the missing subsidy payments.
There are modern examples, too, and sometimes there have even been retroactive reductions in transfer payments. Ontario, for example, qualified for equalization payments in the late 1970s under the formula that prevailed at the time. But the federal government, concerned with expenditure pressures and the general optics of making equalization payments to Ontario, changed the regulations in 1979 to delay a payment. Then, in February 1981, it retroactively changed the legislation and removed Ontario’s eligibility from 1977 onward.
The time between when a payment accrues and when it is paid can also be lengthy. One notable example concerns a stabilization payment to Quebec. In September 1993, Quebec submitted a $282.5-million stabilization claim for fiscal year 1991-92. The federal government rejected it, which sparked a long legal dispute eventually culminating in a Quebec victory in the Federal Court in 2007 and in the Federal Court of Appeal in 2008. The federal government finally paid $103.4 million in March 2015 to settle the two-decade-old dispute. To be sure, this was not a retroactive policy change but the result of a long-drawn-out legal dispute between the two governments.
At the end of the day, adjusting past policy retroactively is a political question, but it is a tool governments have used at times to resolve political tensions or achieve other objectives. Today, if one accepts that the $60 cap implemented in 1987 was ill-advised, or at the very least not based on rigorous analysis, then retroactively eliminating it and recalculating past payments — not just to Alberta, but also to Nova Scotia for fiscal year 1991-92 and Prince Edward Island for 1992-93 — would be a justifiable option.
Completely excluding resource revenues from stabilization would dramatically lower the federal government’s exposure to provincial shocks. This is not a new idea, but it is a contentious one.
On the one hand, removing resource revenues from stabilization would avoid rewarding risky decisions. Alberta, Saskatchewan and Newfoundland and Labrador have the most unstable revenues by far: the average swing in Alberta’s primary budget (its standard deviation) is over 2 percent of GDP — roughly double that of other large provinces. This is entirely due to highly volatile resource revenues. The difference between resource revenues projected at the beginning of a fiscal year, for example, and the final realized amount at the end is typically 1.3 percent of GDP. Were it not for this unambiguously risky revenue source, Alberta’s primary balances would be no more volatile than those of British Columbia, Ontario or Quebec. One might argue, therefore, that such easily avoidable risks should not be borne by other Canadians through the federal stabilization program. In addition to volatility, there are other arguments to exclude resource revenues, which are, after all, an asset sale rather than revenues in the typical sense.
On the other hand, Finance Minister Mitchell Sharp had a simple yet powerful reply to both arguments: “In dealing with natural resource revenues the federal government has simply accepted the practice of the provinces…All provinces treat their natural resource revenues as current revenues…It is not so much the source of the revenue which is important for financing provincial services but the amount of revenue at the disposal of provincial governments.” His was an argument to respect provincial autonomy and remain agnostic as to their choices. Whatever view one finds convincing, the treatment of resource revenues has always been a challenge in the design of fiscal transfers.
There is much to consider when it comes to reforming fiscal stabilization in Canada. This program, like all other transfer programs, must balance competing fiscal, economic and political pressures. As those pressures evolve, so should the transfer formula. In this paper, I have detailed the main goal of stabilization and how the program has evolved over time. Today, by effectively insuring only a minuscule share of declines in provincial revenues, stabilization provides no material protection against such declines. In my view, this should change, and a number of concrete reforms could improve the program.
First, in the short term, the cap on payments should be eliminated — after all, the cap was implemented with little thought and based on no rigorous analysis. In addition, all payments since 1987 should be recalculated to reflect an uncapped program. This would result in retroactive payments to Nova Scotia, Prince Edward Island and, more recently and significantly, Alberta. But if the federal government preferred, for affordability reasons, not to uncap the program fully, then easing the cap to reflect what the amount would have been had it been indexed to GDP per capita or to inflation would still result in increased payments to those provinces.
Second, given the volatile nature of resource revenues, and the voluntary acceptance of this fiscal risk by some provinces, the federal government should not increase its coverage of provincial resource revenues. The 50 percent deductible for resource revenues is a high bar that should remain in place, absent more fundamental reforms to the program. Alberta and other provinces need to come to terms with the volatility and riskiness that result from their budget decisions. To cover reasonably expected losses through the stabilization program would be to dampen the incentive to save more revenues and to introduce more stable revenue sources such as, in Alberta’s case, a sales tax.
Finally, in addition to these short-term considerations, the federal government should initiate a comprehensive review of the stabilization program and examine more fundamental reforms. In particular, the program should align more closely with the equalization program. To integrate the two, stabilization could be made a function of fiscal capacity rather than of revenues, all deductibles could be eliminated and a two-year moving average could be used to determine the amount of the payment. The equalization program already provides significant stabilization to recipient provinces, but with a lag. Aligning the two programs would provide more coherence and extend the stabilization properties of equalization to higher-income provinces. Stabilization would act as a rapid response to temporary shocks, and equalization as a response to persistent shocks. There would also be administrative advantages from such a change. Currently, provinces must apply for stabilization, as the federal government does not have the detailed information it needs to calculate payments. With a program based on fiscal capacity, the calculations could be done for all provinces every year, and payments made without any application, although this would require more rapid data collection and preliminary estimates. Adverse incentives could be reduced further by excluding resource revenues completely and having a 5 percent deductible. But even in this case, Alberta’s two payments for 2015-16 and 2016-17 would have totalled nearly $2.9 billion.
Alternatively, to both lower dramatically the adverse incentives that transfer programs create for provincial governments and greatly simplify stabilization, the program could be made a function of economic strength, rather than of fiscal capacity. The two concepts are highly correlated already, but GDP is more difficult for provincial governments to influence and significantly simpler. The principal goal of stabilization is to buffer provinces against economic shocks, after all, so a macro-based formula that based payments on changes in nominal GDP, rather than on adjusted provincial revenues, would be a simple and transparent means of doing this. This type of macro-based approach has been proposed many times in the past for equalization, and there is a strong case for reforming the fiscal stabilization program along these lines as well. In particular, there is broad public misunderstanding about transfer programs such as equalization; moving to a simple and transparent macro-based formula could help.
Whatever one thinks of the various options, one thing is clear: there is no meaningful fiscal stabilization in Canada today. And although any reform would come with trade-offs, there is growing pressure on the federal government to support provinces going through tough times more effectively. It has been nearly a quarter-century since stabilization was last reviewed and substantively changed. It is time to look at it again.
 For context, one can use per capita GDP to compare growth rate differences with level differences across provinces. Since 2000, the share of Canada’s economy that would need to be redistributed each year to fully offset differences in per capita growth rates has averaged roughly 0.8 percent of national GDP. The level differences in GDP per capita, meanwhile, would require 6.4 percent of GDP to be redistributed to achieve equality in levels. So, while differences in levels of economic activity remain a primary concern, differences in annual growth rates — according to this measure — are just over one-tenth as large, which is historically high. The underlying cause of this growth divergence is an important topic for further study.
 B. Dahlby, “Reforming the Federal Fiscal Stabilization Program,” SPP Briefing Papers 12, no. 18 (Calgary: School of Public Policy, 2019).
 J. Feehan, Canada’s Equalization Program: Political Debates and Opportunities for Reform, IRPP Insight 30 (Montreal: Institute for Research on Public Policy, 2020).
 For more historical detail, see D.B. Perry, Financing the Canadian Federation, 1867 to 1995: Setting the Stage for Change, Canadian Tax Paper 102 (Toronto: Canadian Tax Foundation, 1997).
 Canada, Parliament, House of Commons, Debates, 33rd Parliament, 2nd Session, vol. 4 (March 19, 1987), 4330.
 Canada, Parliament, Senate Committee on National Finance, April 13, 1987.
 Alberta originally claimed $539.3 million, but recalculations and arbitration settled the amount at $419.2 million, with the final of three payments made on July 22, 1991.
 This is an odd relic of previous arrangements — specifically, the Federal-Provincial Fiscal Arrangements Act, part II, section 6(4)(b)(i) — that probably should be repealed.
 The 2015-16 payment was initially $251 million, but a revision to Alberta’s population data lowered this to $248 million. The $3 million overpayment was forgiven by the federal government in 2019.
 The Alberta government reported the effect of the cap over these same fiscal years to have been $2.4 billion. The difference reflects data revisions to corporate income tax revenues and forestry revenues that were made after the federal government’s final determination for 2015-16.
 Today the ceiling is 0.8 percent of revenues, compared with 2.3 percent in 1987. I calculated both figures using per capita provincial revenues from the equalization worksheets (specifically S-Table 6), excluding property taxes and resource revenues.
 See footnote 10 for a reconciliation between this and the Alberta government’s estimate.
 Current data lags would necessitate certain assumptions and estimates to ensure timely payment during periods of provincial fiscal stress, but this would not be an insurmountable challenge.
 Canada, Parliament, House of Commons, Debates, 27th Parliament, 1st session, vol. 13 (March 2, 1967), 13688.
 In the data, a 1 percent change in a province’s nominal GDP is associated with a precisely estimated 1 percent change in total fiscal capacity, with a 95 percent confidence interval of 0.90 percent to 1.09 percent.
 For more on this history, see T. Tombe, “Final and Unalterable — But Up for Negotiation: Federal-Provincial Transfers in Canada,” Canadian Tax Journal 66, no. 4 (2019): 1-53.
 The federal government is currently phasing in personal income tax reductions through an increase in the basic personal amount. This change represents approximately one-fifth of the fiscal space identified by the Parliamentary Budget Office.
 The total equalization payments fixed in part I, section 3.4 (5) of the Federal-Provincial Fiscal Arrangements Act is indexed to Canada’s GDP growth, so varies over time.
 The retroactive adjustment took the form of increasing the provincial debt allowances. For Ontario and Quebec, this amounted to $5.4 million, and resulted in annual increased payments of $270,000 in perpetuity. It is still paid today. This reflected the value of interest implicitly charged between 1867 and 1873 on the $10.5 million awarded the provinces in 1873. Interestingly, the calculation that arrived at this adjustment was incorrect, as it did not properly account for compound interest when carrying interest forward to 1884. See Provincial Subsidies Act, R.S., c. P-26, s. 6 and s. 8.
 This is known as the “personal income override.” Provinces with above-average per capita personal income levels were made ineligible to receive equalization regardless of the formula-determined payment. This change affected only Ontario. (See SOR/79-279 (P.C. 1979-821) and 29-30 Eliz. II, c. 46 (1981).) The long delay, and need for retroactive changes, was due to the fall of the Pierre Trudeau government in 1979, followed shortly by the fall of Joe Clark’s in 1980. There was an original attempt to implement the override with Bill C-26 in December 1978.
 Quebec (Attorney General) v. Canada, 2007 FC 826; Canada v. Quebec (Attorney General), 2008 FCA 201.
 Canada, Parliament, House of Commons, Debates, 27th Parliament, 1st session, vol. 13 (March 9, 1967), 13841.
 The stabilization payment to Newfoundland and Labrador in 2015-16 was initially at the $60 per capita limit, but subsequent data revisions resulted in a final payment of only $15 per capita.
About This Paper
This paper was published as part of the Canada’s Changing Federal Community program, under the direction of F. Leslie Seidle and Charles Breton. The manuscript was copy-edited by Barry Norris, proofreading was by Zofia Laubitz, editorial coordination was by Francesca Worrall, production was by Chantal Létourneau and art direction was by Anne Tremblay.
Trevor Tombe is associate professor of economics and a research fellow at the School of Public Policy, University of Calgary. His recent research focuses on economic and fiscal integration, especially internal trade and fiscal federalism. His work has been published in the American Economic Review, the Journal of Monetary Economics, the Review of Economic Dynamics and the Canadian Journal of Economics, among other journals, and he contributes actively to public policy debates through policy publications and in the media. He has an MA and a PhD in economics from the University of Toronto.
To cite this document:
Tombe, Trevor. 2020. An (Overdue) Review of Canada’s Fiscal Stabilization Program. IRPP Insight 31. Montreal: Institute for Research on Public Policy. DOI: https://doi.org/10.26070/wqe9-8225
I gratefully acknowledge data provided by the governments of Canada and Alberta, without which much of this analysis would not have been possible. I would also like to thank Daniel Béland, Charles Breton, Bev Dahlby, André Lecours, Mark Parsons, Leslie Seidle, France St-Hilaire, Stephen Tkachyk, and several dedicated public servants in both Finance Alberta and Finance Canada for sharing their expertise and insight. All errors and opinions are my own.
The opinions expressed in this study are those of the author and do not necessarily reflect the views of the IRPP or its Board of Directors.
IRPP Insight is an occasional publication consisting of concise policy analyses or critiques on timely topics by experts in the field.
Cover photo: Shutterstock.com
ISSN 3392-7748 (Online)
Canada’s Outdated Fiscal Stabilization Program Needs Fundamental Reform
Montreal — To provide provinces adequate protection against short-term economic shocks, key changes need to be made to the country’s fiscal stabilization program, says a new study from the Institute for Research on Public Policy.
“There is effectively no fiscal stabilization in Canada today,” argues the study’s author, Trevor Tombe, associate professor of economics and research fellow at the University of Calgary’s School of Public Policy. With growing pressure on the federal government to better support provinces suffering from economic downturns, Tombe calls for concrete and timely changes to Canada’s stabilization program: “It has been nearly a quarter-century since stabilization was last reviewed and substantively reformed. It is time to look at it again.”
In the short term, changes need to be made to the $60 per capita cap on stabilization payments, which was implemented in 1987 in an arbitrary fashion. This cap should either be eliminated or be changed to what it would have been if it had been indexed to GDP or to inflation over time. Once the cap issue is resolved, stabilization payments should be adjusted retroactively. This would give way to more substantive stabilization payments to provinces that are still feeling the brunt of sudden and drastic drops in revenues.
In the longer-term, however, a more fundamental reform should be considered. Tombe proposes two options: The first would be to more closely integrate the fiscal stabilization and equalization programs by using similar indicators of provincial fiscal capacity and smoothing mechanisms. Among other benefits, this would extend the stabilization properties of equalization to higher-income provinces. The second would be to calculate fiscal stabilization payments on the basis of declines in provinces’ economic performance, rather than in their revenues – a simpler and more transparent approach, which could lessen broad public misunderstanding about transfer programs.
Tombe concludes that, although the short-term changes are needed to ensure that fiscal stabilization remains relevant, they should not forestall a more comprehensive review of the program.
An (Overdue) Review of Canada’s Fiscal Stabilization Program, by Trevor Tombe, can be downloaded from the Institute’s website (irpp.org).
The Institute for Research on Public Policy is an independent, national, bilingual, not-for-profit organization based in Montreal. To receive updates from the IRPP, please subscribe to our e‑mail list.
Media contact: Ricardo Montrose tel. 514-985-2461 ext. 282, firstname.lastname@example.org