Review of Canadian Federal Fiscal Forecasting, Processes and Systems: 1

## Archived - Executive Summary

Archived information

Archived information is provided for reference, research or recordkeeping purposes. It is not subject to the Government of Canada Web Standards and has not been altered or updated since it was archived. Please contact us to request a format other than those available.

### 1. Introduction

This examination of the federal government’s fiscal forecasting accuracy was undertaken at the request of the Finance Minister, the Honourable Ralph Goodale, and follows by ten years the previous full review of the processes used to determine the government’s budget projections. At that time, there were concerns about the credibility of a forecasting process in which the federal government had often underestimated the size of its deficits. Since then, the budget balance has been underestimated rather than overestimated and, in each of the past seven years, the government has achieved surpluses larger than originally projected.

The primary objective of this report is to assess the basis for the persistent forecast errors of the last ten years and determine what changes to the budget process might improve the accuracy of those forecasts and ultimately improve the conduct of public policy.

The major factors examined are:

(i) The accuracy of the forecasts of key economic variables – e.g. real and nominal gross domestic product, inflation, short- and long-term interest rates, employment growth – which affect revenue and expenditures.

(ii) The sensitivities of specific revenue and expenditure streams to key economic variables, particularly where the relationship(s) might be undergoing structural changes.

(iii) The impact of the quality and timeliness of the information available for both economic and fiscal forecasts and, in particular, the influence of significant data revisions.

(iv) The influence on expenditures of factors such as rules governing departmental spending and lags in finalizing transfers to provinces.

(v) The effects of making provisions to deal with anticipated contingent liabilities of the government.

(vi) The degree to which implicit caution is incorporated in the translation from economic forecasts to fiscal projections.

The report also deals with an issue that is germane to forecast accuracy even though not directly linked to the specific processes of budget projections. Fiscal forecasting is influenced by the fiscal rules – tacit or explicit – under which budgets operate. It matters whether the rule is to ensure no deficit under any circumstances or some other target, such as balancing the budget over the economic cycle.

To complete this report under a tight time schedule, work was contracted out or requested in three main areas. The analysis of forecast accuracy was carried out jointly by the Policy and Economic Analysis Program (PEAP) of the Institute for Policy Analysis at the University of Toronto and the Center for Interuniversity Research and Analysis on Organizations (CIRANO). The analysis comparing Canada’s forecasting track record to that of the other G-7 and OECD countries was conducted, at the request of Finance Canada, by the North American Division and the Fiscal Affairs Division of Western Hemisphere Department at the International Monetary Fund. An extensive consultation was carried out of private sector forecasters and other individuals with insights into the technical and political economy issues involved. Most of the interviews were conducted by Bruce Little, formerly with The Globe and Mail, under contract with Finance Canada. The author also conducted several interviews. Most of the interviews were with Canadian experts, but the author also interviewed officials in Europe and the United States.

### 2. Qualitative Analysis of Fiscal Forecast Accuracy: Summary of Consultations

The federal government’s record of economic and fiscal forecasting has been a topic of discussion among private sector forecasters, academics and political party researchers for some time now. Some 20 experts in forecasting and budget preparation were interviewed; these were mainly private sector and university economists, other academics, staff advisors to political parties and some former senior Finance Department officials.

The summary of the consultations reflects the views of those interviewed and, in some instances, their qualitative assessment of the nature and causes of the persistent under-forecasting of budget balances. In some instances, their positions are not supported by the evidence and analysis of the quantitative examination of the issue. There are simple factual inaccuracies as well as interpretations and inferences that can be (and, in a number of cases, are) challenged by the research carried out in this study.

In broad terms, there were several areas of agreement among those interviewed. First, the consistent underestimate of future surpluses does indeed constitute a problem although some argued that it was a better problem to have than running deficits. Second, the surplus "surprises" have been primarily the result of cautious fiscal forecasting that flowed directly from the government’s determination never to run a deficit. Third, solutions are available that should be implemented.

There was also general agreement that the succession of larger-than-expected budget surpluses over the past seven years has, whether intentionally or not, hindered the public and Parliamentary debate over the main budget choices available in an era of surpluses – tax cuts, increased spending and reducing debt. It has also undermined the Finance Department’s credibility to the point where many simply do not believe its published forecast numbers.

Several people criticized the use of year-end surprises for transfers to provinces or foundations as hasty last-minute decisions. Many opposed tax rebates as bad tax policy, even though they supported permanent tax cuts. Most argued that regular surplus surprises effectively "load the dice" against long-term tax cuts and against increases in focused program spending as well, making debt reduction the default winner in any competition for the "extras" generated during the fiscal year. The combination of forecast errors and the narrow range of policy choices at the end of a fiscal year was widely seen as both a forecasting issue and an accounting issue. Forecast inaccuracies might be the cause of both the year-end and final surplus surprises, but accounting rules narrow the range of alternative uses to which the extra funds can be put.

There was broad agreement that little can be done on a technical level to improve economic forecasts, though most forecasters said data revisions by Statistics Canada frequently frustrate their efforts to predict economic growth. More important to most of those interviewed is the process of converting the economic forecast into an accurate fiscal forecast – getting right the projections of federal government revenue and spending. This, most said, is the major source of any fiscal forecast accuracy problem. However, many noted that small misses in forecasting revenue and spending can produce big errors in forecasting the surplus, and several observed that because many people do not understand this arithmetic, they expect a much greater degree of accuracy in forecasting the surplus than is realistic.

Those interviewed offered two quite different reasons for the surplus surprises. The first group, comprising a large majority of those interviewed, saw the recent track record of larger-than-forecast surpluses as the logical outcome of two forces – the government’s practice of budgetary prudence (seen mainly as a by-product of its determination to stay out of deficit) and several years in which the economy grew faster than expected. The second camp, which was very small, saw them as the result of deliberate manipulation that is specifically designed to stifle a wider public debate over how to use future surpluses – tax less, spend more or reduce debt.

Many said revenue forecasts are harder than they look because of changes in the relative importance of different revenue sources. They argued that forecasts of interest costs and program spending should be relatively straightforward, the latter because departments are given spending limits in advance. Many cited an exception to the latter as a source of forecast error – money set aside as allowances for bad debts, the possibility of losing litigation or other disputes that are already under way. Each case involves a judgment call and if each such call is prudent, the cumulative effect might turn out to be large.

As long as the government follows a no-deficit rule, several argued, there will be pressure on the forecasters to project spending levels that reflect (as one put it) "everything you could possibly imagine going wrong." This tilts the system to produce not only surpluses, but bigger-than-forecast surpluses as well.

There was a sharp divide among those interviewed on the no-deficit rule. Most said it was too rigid. If the government is determined never to run a deficit, it might feel compelled to raise taxes or reduce spending during a period of weakness and thus exacerbate an economic downturn. A few, however, felt that any deficit – however small – must be avoided because it would put the government on a "slippery slope" back to the kind of large federal deficits that characterized the 1980s and early 1990s.

There was general agreement on the importance of a fiscal anchor, which they defined as a strong fiscal policy goal that can be easily communicated to the public, but most thought the value of simple deficit avoidance as an anchor has diminished. The government’s target of a 25% debt/GDP ratio was seen as reasonable but easily achievable through a combination of decent economic growth and small federal surpluses. Almost all endorsed the continued use of the annual contingency reserve (usually $3 billion) and the allowance for economic prudence (usually$1 billion in the first fiscal year of the projection and growing thereafter). These were viewed as reasonable sums which, however, should be the upper limit on fiscal cushions in the budget forecasts.

Most of those consulted dismissed the idea of setting up in Canada a body resembling the US Congressional Budget Office to provide an arm’s-length check on the government’s fiscal forecast. But there was broad support for new institutions that might enrich the debate over budget issues by conducting research on a variety of long-term fiscal issues facing the government such as the impending retirement of the baby-boom generation. Some suggested that the House of Commons Finance Committee should acquire additional resources to help it examine the government’s forecasts and deepen its understanding of fiscal issues. Others were attracted by the idea of setting up some mechanism to study a range of medium- and longer-term fiscal issues.

### 3. Quantitative Analysis of Fiscal Forecast Accuracy

The process of getting from private sector forecasts of the economic outlook to the Public Accounts fiscal forecasts used in the annual Budget involves four basic steps: collecting the private sector forecasts; calculating the average economic forecast; using that to produce the fiscal forecast on a National Accounts basis; translating National Accounts projections into a fiscal forecast on a Public Accounts basis.

Economic forecasts are collected from private sector participants each quarter. The economic forecast values used for fiscal projections are calculated as the average of all participants’ estimates available for a specific indicator. Key variables from the average forecast are entered in the Department of Finance’s macroeconomic forecasting model, to which some judgmental adjustments are made. The detailed fiscal projections on a National Accounts basis are built from the bottom up, as each revenue and expense category is projected separately, reflecting its own unique and specific base, target population, program parameters and other explanatory variables. The conversion of these National Accounts projections to a Public Accounts fiscal forecast involves several types of adjustments. Some reflect straightforward differences between the definitions of categories in the two systems while others require some interpretive adjustments. As well, the process incorporates more up-to-date information available to the Department, including data from the Canada Revenue Agency on various revenue sources.

It is important to get a proper broad perspective on the issue of forecast accuracy before diving into detail. A budget balance, whether surplus or deficit, is the arithmetic difference between two very large numbers – revenues and expenditures. A small error in either or both can translate into a large error in the difference between the two. If, in fiscal 2003-04, revenues had been 1% higher and spending 1% lower than they actually were, the surplus would have been $3.6 billion higher than the actual reported result of$9.1 billion. If a projected surplus were smaller than that – say, $4 billion – then a comparable 1% error would almost double the surplus while the surplus would disappear if revenues were lower and spending higher by 1%. Academic and practitioner research on fiscal forecasting in Europe and the United States concludes that there exists an unavoidable level of uncertainty; accuracy, as one US analyst put it, "is largely a matter of luck." In Canada, studies analyzing budget uncertainty have suggested that the government would need to include a cushion of up to$9 billion to be certain of avoiding a deficit.

Any assessment of fiscal forecast accuracy is far from straightforward, as a review of several such recent attempts illustrates. In the analysis done for this report, the PEAP – CIRANO group makes two important adjustments that have not been done in other studies of this issue. First, the original projections and actual outcomes are adjusted for changes in accounting methods that were introduced in 2002. The second adjustment takes account of policy initiatives – revenue or expenditure – that occurred during the fiscal year. In particular, the government has frequently increased spending during the year in the face of evidence that revenues were likely to be larger (or other expenditures lower) than originally forecast. These in-year policy initiatives need to be excluded from the calculation to get an accurate picture of forecast accuracy. Their results show that the persistent under-forecasting of the year-ahead budget balance evident from conventional assessments is even more pronounced when adjustment is made for in-year policy initiatives. In no year was the balance over-forecast and, in four of the last seven fiscal years, the under-forecast exceeded $10 billion. Two additional conclusions emerge from the calculation of the difference between fiscal projections and the actual outcomes. First, in seven of the last eight years, total revenues came in above projections and dramatically so in three of these years. However, since 2001-02, the revenue projections have been close to the final outcomes and made only a modest contribution to the budget balance under-forecast in this period. Second, total program expenditures have consistently contributed to the budget balance projections coming in too low. They have been over-forecast in all but one year with the gap exceeding$1 billion in all but one year.

These observations do not explain why the budget balances have been under-forecast as they tell us nothing about how it came to pass that program spending has been consistently over-forecast and total revenues regularly and, in some cases, substantially under-forecast.

Before answering the why question, however, a comparison of Canada’s forecasting track record with that of other countries helps provide a useful perspective. The IMF study finds that Canada is an outlier compared to the ten other countries examined. Canada has more persistently and significantly under-forecast its budget balances. As well, Canada has had the most consistently under-forecasted revenues and over-forecasted expenditures. Although the IMF was unable to use the approach of PEAP – CIRANO (they did not adjust for accounting changes nor for in-year policy initiatives), their results are consistent with those of PEAP – CIRANO.

In assessing the reasons for the budget balance under-forecasts, the most obvious starting point is the accuracy of economic forecasts. There are two distinct but related questions. First, how accurate were the projections of Canada’s economic performance? Second, of the persistent differences between budget outcomes and projections, how much can be attributed to inaccuracies in economic forecasts?

The PEAP – CIRANO analysis shows that the economic forecasts have not been particularly accurate, but that the errors have not persistently been in one direction or another. So while the private sector economic forecasts used by Finance to generate its budget projections have sometimes contributed significantly to budget forecast errors, a large share of the explanation for persistent under-forecasting of budget balances lies elsewhere.

Both the PEAP – CIRANO and IMF studies found that projections of nominal GDP (important because this is the underpinning for the tax base) were affected by underestimates of base year GDP levels. The GDP levels used by forecasters as their starting points were later revised higher. So even if their forecasts of growth and inflation rates had been perfect, the size of the projected increase in GDP in a given year would be underestimated; so, by extension, would any increase in government revenues. In the IMF analysis, the underestimation of the GDP level was larger than that for any other country examined.

Using three different techniques to estimate the impact of economic forecast errors on the fiscal results, the PEAP – CIRANO study yielded the following overall conclusions:

• Economic forecast inaccuracies explain some, but not all, of the forecast differences for revenues.
• The economic forecast errors stem more from revisions to nominal GDP levels from which forecasted growth rates are projected than they do from errors in the forecasted growth rates themselves.
• A relatively small proportion of the forecast differences in program expenditures can be explained by reference to the economic forecasts.
• A substantial portion of the forecast differences for public debt charges can indeed be attributed to errors in forecasting interest rates.

Taken together, these results provide only part of the explanation for the persistent under-forecasting of budget balances, although the precise share of the fiscal projection errors cannot be quantified. Therefore, the explanation for those mis-forecasts includes other factors.

One is the timeliness of information flows – from the Canada Revenue Agency (CRA) on tax collections of all types, from Statistics Canada on current economic data and possible revisions, and from government departments on expenditures. For several major revenue sources, including the personal income tax, the corporate income tax and the goods and services tax, there are significant deficiencies in the information available even by the end of the fiscal year. The move to full accrual accounting has added to the difficulty in projecting personal income tax collections. Thus the budget estimate for the fiscal year just ending is made with critical information not yet available, which may affect the projections for the year ahead.

Certain non-tax revenues have also generated consistent forecast differences in recent years. These particular revenue under-forecasts appear to have been the result of higher-than-expected profits from several Crown corporations and a reluctance to believe that the profits at such levels would indeed continue.

While expenditures should be easier to forecast because the government simply determines what it will spend, in a number of categories the actual outcome has led to an over-forecast of expenditures. First, program spending "lapses" and provisions for "re-profiling" (allowing some funds to be carried over to the next fiscal year) have contributed to the over-estimates. Second, provisions for contingent liabilities – funds set aside for lawsuits, other court or arbitration decisions or bad loans – have, on occasion, been higher than required. Third, data lags in the final calculation of equalization payments have, in some years, been revised down long after the fiscal forecasts were produced.

Economic forecast errors, including the impact of data revisions, do not explain a large share of the revenue forecast errors. Nor is it likely that the consistent expenditure over-forecasting is due entirely to intermittent factors like departmental spending lapses, over-provisioning for contingent liabilities and after-the-fact equalization payments adjustments.

This logically leads to the proposition that, in addition to these factors, implicit caution has been added to the explicit contingency reserve and prudence incorporated into the budgets of the last decade. This likelihood was suggested by many of those we consulted and by the authors of both the PEAP – CIRANO and IMF studies. In fact, it is an almost inescapable conclusion that extra prudence – not evident in the budget documents – has been an important factor in explaining the persistent under-forecasting of budget balances since 1994.

The fact that program spending was materially over-projected in six of the last ten years strongly suggests that this is the key area in which extra caution was added. Circumstantial "evidence" on the existence of implicit caution can also be inferred from the IMF’s comment that the comparative pessimism of the Canadian forecasts is not the result of specific large forecast errors but "the accumulation of small but persistently negative errors." That is more consistent with a systematic infusion of caution than with a steady series of one-off errors, all of which happen to be in the same direction.

Why this might happen is a contentious issue. Some commentators have accused the government of deliberately "hiding" part of the surpluses to avoid pressure to spend the "surprise" amount on programs or to avoid cutting taxes. There is, however, a far less sinister and a far more plausible explanation for the addition of implicit caution. It is that the current fiscal rule of "no deficits" created incentives, for those responsible for producing the fiscal projections, to incorporate extra (implicit) prudence into their forecasts.

Although the no-deficit fiscal target is not formalized in legislation, it has been adhered to strenuously – in fact, much more so than in many countries with formal deficit limit targets. If the officials responsible for producing forecasts are faced with an unequivocal commitment on the part of the government that no deficit will be tolerated, no matter how small and irrespective of the economic circumstances, there will be inevitable behavioural consequences. When considering a range of possible outcomes for a revenue or expenditure item, the prudent civil servant will tend to pick a point estimate at the low end of the range for revenues and at the high end for expenditures. There is nothing sinister or underhanded about such behaviour. It appears, then, that a significant and recurring influence on the persistent under-forecasting of surpluses is the fiscal rule under which the federal government has operated since 1997.

### 4. Recommendations

The recommendations fall under four headings:

(1) Increased transparency in budget-related information.

(2) Improvements in data quality and analysis.

(3) Options for the fiscal rules under which budget forecasts are made.

(4) Options for changes in the structures/institutions used in the forecast process.

The first two would modify the current forecast process; the latter two would require substantial adjustments to that process.

The Need for Transparency

Transparency is widely regarded as a key element in sound budget-making. The existence of information asymmetry – actual or perceived – between the government and the public is at the foundation of credibility concerns related to transparency. Governments have access to an extensive array of data and other information and to expertise that most citizens either don’t have or can’t have in the same time frame. The credibility problem regarding the soundness of fiscal policy tends to arise when the government is underperforming relative to its targets (e.g. Canada in the early 1990s), but can also arise when governments outperform as well.

The most important goal of improving the transparency of the federal government’s fiscal forecasting procedures and information is to increase the level of trust in the budget process itself. Apart from sensitive items, full, detailed disclosure should be the norm for all budget-related documents including the Budget itself, the Economic and Fiscal Update, the Annual Financial Report of the Government of Canada and the monthly Fiscal Monitor.

To that end, there are six recommendations on transparency. The first three relate to the need for Finance to provide a detailed breakdown of the linkages between the (external) economic and (internal) National Accounts fiscal forecasts and of the reconciliation of the National Accounts to the Public Accounts fiscal projections. These include:

(i) In the Budget and the Economic and Fiscal Update, fully explore the key risks and uncertainties in the economic forecast and discuss their implications for fiscal projections.

(ii) In the same documents, provide details on the rules of thumb used to estimate the impacts on revenue and (certain) expenditure categories of key economic variables such as nominal GDP growth and short- and long-term interest rates.

(iii) In major fiscal documents, spell out the details of the reconciliation between the National Accounts and Public Accounts fiscal forecasts.

In the Budget, there should be documentation of the long-term (e.g. ten year) track record of Finance’s projections, which will allow all interested parties to better assess the government’s fiscal forecast accuracy.

In addition, Finance should, as part of every third Fiscal Monitor, provide an analysis of fiscal developments in the current year and the risks to the projected fiscal outcome. Where possible, it should include a complete update of the current year fiscal forecast.

The final recommendation is that Finance increase the frequency of its formal briefings to the House of Commons Finance Committee. In addition to the appearances related to the Budget and the Economic and Fiscal Update, there should be at least one additional briefing provided in the early summer.

Improving Accuracy & Timeliness of Data

Enhancing the accuracy and timeliness of the data used for the economic and fiscal projections can increase the accuracy of the forecasts themselves. There are three recommendations made in this area of the report. The first relates to estimates of nominal GDP made by Statistics Canada. The fiscal forecasts have been affected by persistent upward revisions to GDP growth, which caused revenue projections to be lower than would have been the case had the revisions been known earlier. The recommendation is that Statistics Canada and the Department of Finance jointly examine the causes of this pattern and options for mitigating it.

Another cause of fiscal forecast inaccuracy is the adjustment in the relationships between the economy’s performance and several major categories of revenue. It is recommended that research be undertaken to determine the factors that have caused these changes. In addition, the analysis should focus on potential future adjustments in the revenue sensitivities.

Improved monitoring of several key government operations would enhance forecast accuracy. In particular, the reasons for the consistent upside surprises to Crown corporation earnings need to be determined and in-year monitoring of their financial status should be improved. Enhanced ongoing tracking of departmental expenditures would allow lapses to be better anticipated.

Fiscal Rules

The federal budget process currently operates under not one, but two fiscal rules or targets. The first – the no-deficit rule in effect since 1997-98 – was discussed in the previous section of the report. The other rule is of more recent vintage. In the 2004 Budget, the Finance Minister announced that the government was targeting a reduction in the debt/GDP ratio to 25% from (the then) 41% within ten years. The target was reiterated in the recent 2005 Budget.

There is no uniformity in the fiscal rules under which national governments operate. In Europe, countries that adopted the euro agreed to limit annual deficits to 3% of GDP and debt to 60% of GDP and to achieve balanced budgets over the medium-term. The United Kingdom allows deficit financing only for public investment expenditures and sets a debt/GDP target of 40% over the cycle. In Sweden, the government has a surplus target of 2% of GDP over the economic cycle and specified spending limits over a rolling three-year period. In Australia, the current fiscal strategy calls for a balanced budget over the cycle and a lower net debt level. New Zealand targets an operating surplus over "a reasonable time" and a debt/GDP ratio of 20% before 2015. In the United States, the federal government operated under the Budget Enforcement Act from 1992 until 2001. The Act set upper limits for discretionary spending and required offsetting cuts for increases in mandated spending. While governments can choose a variety of fiscal rules under which to operate, experience suggests that they will achieve their established fiscal targets only if they are willing to take measures that ensure the various rules are followed.

The fiscal target under which the federal government has functioned since 1997 is that no deficit shall be incurred under any circumstances. It is a key conclusion of the analysis of forecast accuracy that the no-deficit rule has been a major cause of the persistent upside surplus surprises at the end of each fiscal year. It is recommended that the government consider adopting a different rule that is more appropriate to its fiscal circumstances and to its increased focus on medium- to long-term commitments.

The report examined the pros and cons of three options – retaining the no-deficit rule, achieving zero balance over the economic cycle and targeting a modest surplus, on average, over the cycle. The big differences between the current rule and the two alternatives are:

(i) The balance and surplus over-the-cycle targets allow for deficits to be incurred when warranted by economic circumstances such as a significant economic downturn.

(ii) While the inherent imprecision of economic (and fiscal) forecasts necessitates that caution be built into budget projections, less will be required than for a no-deficit rule.

(iii) Unexpected budget windfalls (or shortfalls) will only be apparent after several years of the cycle rather than annually as is the case with the no-deficit rule; decisions on the allocation of unanticipated surpluses will not have to be made annually.

(iv) With the no-deficit rule, there is no ambiguity about the target each year nor about whether it has been attained; attempting to achieve balance or a surplus over the cycle involves inherent uncertainty about whether the rule will be successfully followed.

(v) If the no-deficit rule is strictly adhered to, it can require adjustments in downturns that exacerbate economic weakness (i.e. it is pro-cyclical); the balance- and surplus-over-the-cycle targets imply counter-cyclical impacts from automatic stabilizers.

The report recommends that the federal government adopt the fiscal rule of achieving a surplus, on average, over the cycle. This target represents a less dramatic departure from the current rule and can be clearly linked to the long-term fiscal goal of lowering the debt/GDP ratio to 25% as the cumulative surplus would be used to reduce the nominal debt level.

If the government decides to retain the current no-deficit rule, it will need to adopt a more formal and structured process for dealing with fiscal surprises. It should establish, in advance, the contingent allocations among tax cuts, spending increases and debt reduction of any unexpected windfalls. This can be incorporated into the (year-beginning) Budget and debated in Parliament as part of the Budget deliberations.

The establishment of a debt/GDP ratio target well below the current level is an important recent initiative, especially given the demographically driven fiscal pressures that will need to be addressed in Canada. The government should consider setting the target below 25% – 20% or even 15% – to ensure that the fiscal challenges can be readily met. It will also be necessary to prepare the public for the transition to targeting a deficit, on average, over the cycle. This is an inevitable consequence of stabilizing the debt/GDP ratio at the targeted level.

Possible Institutional Changes

Institutional change would introduce potential new participants into the process, individuals or agencies whose role would be to make significant improvements in the fiscal forecasting and/or fiscal policy analysis process. An example would be the private economic forecasters engaged by the Commons Finance Committee to provide quarterly forecast updates.

It is important to be mindful of an old adage: when proposing a solution, check first that you actually have a problem that needs solving and then ensure that this proposed solution is the right one. In this case, the analysis and evidence presented in this report cast serious doubt on the proposition that the problem lies with fiscal forecast accuracy per se. That there are differences between fiscal projections and outcomes is beyond question. Nor is there any ambiguity about the unidirectional nature of the fiscal forecast errors – deficits were smaller and then surpluses larger than projected in every year since 1995-96.

If we apply the adage and the evidence to assess the Finance Committee’s recent solution, the question is: what problem is it designed to solve and will it do so effectively? It appears the Committee feels it is not receiving enough information on the in-year fiscal situation of the federal government. If that is the concern, our suggestion that Finance provide the Committee with more frequent and complete fiscal updates is, at minimum, the cheaper resolution. Since the new arrangement calls on the Department to assist the economists hired by the Committee, the Committee will be getting the same fiscal data it can get directly from the Department. If members of the Committee don’t trust the information provided currently by the Department, it is not clear why their trust would be much enhanced by having the same information filtered through the external economists.

More critically, the issue is whether any institutional changes are likely to add much value to short-term forecasting and monitoring. The analysis in this report suggests that, in the economic forecast end of the process, a reduction in the size and frequency of data revisions could have a discernible impact on forecast accuracy. Improved understanding of the sensitivities of key revenue components to changes in nominal GDP could also enhance accuracy. Both of these items are taken up in the data improvement recommendations. Resolving them does not require a new agency.

Finally, if the real culprit in the story of surprise surpluses is the predictable response of the system to a no-deficit fiscal rule, hiring outside economists to monitor and produce fiscal forecasts in conjunction with Finance will not resolve that either. Instead, what is required is a change in the fiscal target which affects the incentives driving behaviour in the forecasting process.

More generally, it is difficult to see how an institutional change that involves transferring some of the forecasting responsibility to an independent agency would make much of a difference to short-term forecasting accuracy.

Accordingly, there are two main recommendations in this segment of the report, one of which has negative elements. It is proposed that the economic and fiscal forecasting structure that has evolved over the last decade be maintained. The degree of accuracy of the budget projections will not be materially improved by creating new institutions to produce them. The potential improvements in accuracy described above can be achieved within the existing framework. This means that the hiring of four economic forecasters by the House of Commons Finance Committee to provide quarterly projections should not be continued. Nor should consideration be given to establishing an agency like the Congressional Budget Office (CBO) in the US or the Central Planning Bureau (CPB) in the Netherlands.

However, if the focus is shifted from short-term forecasting to long-term policy analysis, both the CBO and the CPB have facets of their mandate which could usefully be incorporated into the Canadian context. It is recommended that a small agency be set up with a mandate to focus on the medium- to long-term fiscal implications of structural factors like changing age demographics and productivity growth and of significant policy initiatives such as changes in tax structures. If the government were to shift to a fiscal target of balance or surplus over the cycle, it is further recommended that the agency’s mandate be expanded to monitor and report to Parliament on the tracking of the cyclically adjusted fiscal balance.

Although any new institution should work with and respond to the research priorities of the Finance Committee (and through it, of Parliament), it should also have some latitude to determine, on its own, its research agenda.

## Section 1 – Introduction

This examination of the fiscal forecasting accuracy of the Canadian federal government was undertaken at the request of the Finance Minister, the Honourable Ralph Goodale. Several factors would appear to have influenced the decision to conduct a review of this nature. It has been ten years since the federal government thoroughly reviewed the processes used in determining its budget projections. In the intervening period, a number of the proposals emanating from that review have been incorporated into the budget process. As well, some of those recommendations adopted have been modified or expanded. Hence, there is merit in reviewing the new budget forecast process to determine whether it is functioning as anticipated.

At the time of the last review, there were concerns about the credibility of a forecasting process in which the federal government had, during the early 1990s, consistently underestimated the size of (persistent) deficits. In the current situation, there have also been criticisms of the government’s fiscal forecast accuracy albeit from a completely different perspective. For the past ten years, the budget balance has been underestimated rather than overestimated and, in each of the last seven years, the government has achieved surpluses larger than it had originally projected.[1] The alleged credibility problem this has given rise to is linked to concerns that the government and Parliament are basing fiscal policy decisions on inaccurate and/or incomplete information. To the extent that improvements in forecast accuracy can ameliorate these concerns, that will be an additional benefit of the review.

However, it is the primary objective of this report to assess the basis for the persistent fiscal forecast errors of the last ten years and determine what changes might be effected in the budget process to improve the accuracy of those forecasts. While the reporting of the outcome of a series of consultations conducted as part of this review will touch upon the credibility issue, it is not a primary focus of this report. To make it so would require analytical and forensic skills beyond the capacity of the report’s author. As well, it would detract from the primary goal of the report – i.e. to offer analysis and recommendations that could ultimately improve the conduct of public policy.

The report examines a number of factors which can influence the accuracy of any fiscal projection. The major ones are:

(i) The accuracy of the forecasts of key economic variables – e.g. real and nominal GDP, inflation, short – and long-term interest rates, employment growth – which affect revenue and expenditures.

(ii) The sensitivities of specific revenue and expenditure streams to key economic variables, particularly where the relationship(s) might be undergoing structural changes.

(iii) The impact of the quality and timeliness of the information available for both economic and fiscal forecasts and, in particular, the influence of significant data revisions.

(iv) The influence on expenditures of factors such as rules governing departmental spending and lags in finalizing transfers to provinces.

(v) The effects of making provisions to deal with anticipated contingent liabilities of the government.

(vi) The degree to which implicit caution is incorporated in the translation from economic forecasts to fiscal projections.

In addition to conducting this analysis for Canada on a stand-alone-basis, the review also contrasts Canada’s fiscal forecasting practices and experiences with those of other G-7 and OECD countries. The analysis looks at relative accuracy in economic and fiscal forecasting and assesses the factors which may account for the inter-country similarities and differences in forecast outcomes.

The mandate for the report calls for, in addition to the assessment of fiscal forecast accuracy, proposals to improve said accuracy and by doing so, lead to better informed fiscal decision-making. The final section makes recommendations in four main areas: transparency, data quality, fiscal rules and institutional changes.

It is not a part of this mandate to examine or to make recommendations about the content of federal budgets. That is, the focus is on how to more accurately forecast revenues, expenditures and budget balances rather than on how to allocate them. However, the report does deal with an issue that is germane to forecast accuracy even though not directly linked to the specific processes of budget projections. It is clear from the analysis presented here that fiscal forecasting is influenced by the fiscal rules – tacit or explicit – under which budgets operate. That is, it does appear to make a difference whether the rule under which fiscal authorities are operating is no deficit under any circumstances versus, say, balance the budget over the economic cycle. This issue is explored in the report and options for future rules considered.

To complete this report under a tight time schedule, work was contracted out or requested in three main areas. The analysis of forecast accuracy and the main factors influencing it was carried out jointly by the Policy and Economic Analysis Program (PEAP) of the Institute for Policy Analysis at the University of Toronto and the Center for Interuniversity Research and Analysis on Organizations (CIRANO). The analysis comparing Canada’s forecasting track record to that of the other G-7 and OECD countries was conducted, at the request of Finance Canada, by the North American Division and the Fiscal Affairs Division of the Western Hemisphere Department at the International Monetary Fund. An extensive consultation was carried out of private sector forecasters and other individuals with insights into the technical and political economy issues of fiscal forecasting. The bulk of the interviews were conducted by Bruce Little, formerly with The Globe and Mail, under contract with Finance Canada. The author also conducted several interviews. Most of the interviews were of Canadian experts. However, the author also conducted interviews with officials in Europe and the United States primarily to supplement the comparative quantitative analyses carried out by the IMF with some informed qualitative analysis. There is no separate report of these interviews but their observations are incorporated into the text where appropriate. (See Appendix 1 for the list of individuals consulted in Canada, the US and Europe.)

Section 2 of the report is a summary of the results of the consultations and, in effect, forms the qualitative assessment of fiscal forecasting at the federal level in Canada. It provides a backdrop to the more formal quantitative analysis and assessment of forecast accuracy that follows in Section 3. The consultations also generated useful proposals for possible changes in the fiscal forecasting regime, which will be dealt with in the Recommendations section of the report.

## Section 2 – Qualitative Analysis of Fiscal Forecast Accuracy: Summary of Consultations

The federal government’s record of economic and fiscal forecasting has been a topic of discussion among private sector forecasters, academics and political party researchers for some time now. At the heart of the debate is the government’s record of budget surpluses since 1997-98, almost all of which were bigger than the government predicted in its budgets and mid-year fiscal updates.

This is not an issue of widespread concern and commentary among the general public. Rather, for those outside the government, it is confined to a few score of Canadians who devote time and expertise to analyzing federal budgets and fiscal policy in general. Since their views help frame the debate, it was decided to canvass a sample of this group for their perceptions of the forecasting issue. Interviews were conducted from mid-December 2004 to the end of January 2005 with some 20 experts in forecasting and budget preparation – mainly private sector and university economists, other academics, staff advisors to political parties and some former senior Finance Department officials.

The interviews were designed to probe views on three broad questions. First, is there a problem? Second, if so, what is the nature of the problem? Is it technical, political, or does it contain elements of both? Here, the definition of political is not related to partisan politics but rather addresses the issue of the credibility of budget forecasts. Third, to the extent that there is a problem, is it amenable to solution and, if so, what solutions are possible or desirable?

The views expressed here are solely those of the participants in the consultations. We have tried to convey their views as clearly as possible, but we have two warnings for the reader. In some cases, those interviewed made statements that were factually inaccurate; although we will report those statements, we will also correct them. In other cases, we will report statements that involve interpretations or inferences that can be challenged – and, in many cases are challenged by the research carried out for this study. In those cases, we will note that these views will be addressed further in Section 3 of the report and should not be taken at face value without reading that section. As we will see, there are a number of areas in which the conventional wisdom surrounding the fiscal forecasting issue is not entirely correct, and we do not wish to contribute to further confusion on these matters. We will also, in this section, occasionally add explanatory notes to clarify the participants’ opinions and to provide the background readers might need to understand their comments.

In broad terms, there were several areas of agreement among those interviewed. First, the consistent underestimate of future surpluses does indeed constitute a problem, though a few characterized it as an issue rather than a problem. Second, the surplus "surprises" are primarily the result of cautious fiscal forecasting that flowed directly from the government’s determination never to run a deficit. Third, solutions are available that should be implemented.

The government has reported surpluses for seven years now, almost all of which were bigger than initially projected. Although most analysts were not particularly critical in assessing the Budgets at the time they were released, the size and persistence of these surplus surprises have begun, in recent years, to raise questions about the reliability of the Finance Department’s forecasts. It is safe to say that, among many experts, there is now agreement that the succession of larger-than-expected budget surpluses has, whether intentionally or not, inhibited an informed public debate over the main budget choices available in an era of surpluses – tax cuts, increased spending and reducing debt. It has also undermined the Finance Department’s credibility to the point where some observers simply do not believe its published forecast numbers.

A few of those interviewed have carried out a systematic review of the forecast record to make their own judgments about the nature of any forecast errors. Most, however, have not. Their comments are based on their own observations over the years, which often focused on the specific aspects of the budget process that most interested them. For example, forecasters had more to say about the technical aspects of the issue than those whose main interest is budget policy itself.

### 2.1 Nature and Key Elements of the Problem

There was an overwhelming consensus that federal economic and fiscal forecasting has become a problem of some sort, though several preferred to call it an issue rather than a problem, the distinction reflecting the seriousness with which they regarded the record of forecast errors. Several said that the subject was both trivial and real at the same time – trivial in the sense that no budget surplus should be seen as a problem per se, but real in the sense that forecast errors have both undermined the credibility of the forecasts and narrowed the range of policy options that are open to debate.

A few suggested that Canada is a lucky country indeed if one of its biggest problems is a national government whose finances regularly turn out to be in stronger shape than predicted. Other G-7 countries are running deficits (often bigger than predicted), and Canada’s series of surpluses represents a major accomplishment after a quarter century of deficits that were also often bigger than forecast. If we are going to have a fiscal problem, they said, better this one than the other. This view was echoed by several European officials who noted that the generation of persistent surpluses is a pleasant problem to have and one that many OECD countries would be delighted to experience. They also pointed to Canada’s exemplary record, to date at least, in governing with surpluses when the inclination in other countries is to increase permanent spending and risk creating structural deficits.

Several of those interviewed observed that the question of fiscal forecasting accuracy is one that deeply interests only the handful of Canadians – mainly in the public policy community and the financial industry – who follow budget policy. In this view, the subject is simply too technical or too arcane to grab the attention of most Canadians. One individual opined that the average Canadian is more interested in knowing only that the government’s finances are sound and that the government is implementing whatever tax and spending policies they personally endorse.

At the same time, there was broad agreement among Canadian experts that there is a public interest issue that goes beyond any technical weaknesses in the forecasts themselves. In this view, the growing debate over the persistence of surplus surprises is evidence of a failure in public policymaking. Almost all of those interviewed argued that the tendency to under-forecast the surplus hinders the public and Parliamentary debate over fiscal policy options and even distorts the decision-making process itself. There was also a widely held view that federal accounting rules are partly to blame for the dissatisfaction over the surplus surprises. If the unanticipated additional funds appear only after the end of the fiscal year, it was argued that current accounting rules allow them to be used for only one purpose – reducing debt.

Finally, all clearly supported the idea that better forecasts are a good thing and would support changes to the forecast-making process that might produce more accurate forecasts, although they also recognized that precision in economic and fiscal forecasting is not possible.

Surplus surprises and their implications

At root, the problem is seen to lie not in the surpluses themselves but in the surprise nature of these surpluses – the unexpected or, at least, unprojected increments at fiscal year-end. Almost everyone interviewed focused on two facts. First, projected surpluses for any given fiscal year have tended to grow as each fiscal year unfolded, leaving additional room for government initiatives as the end of the year approached. Second, the final surplus reported some five months after the end of each fiscal year has tended to be larger, often substantially so, than the one predicted in the budget presented just before each fiscal year ended.

We might call these the year-end surprise and the final surprise respectively. Using 2003-04 as an example, the fiscal year opened with a 2003 budget projection of an underlying surplus of $4 billion. The fall 2003 Update reported$1.2 billion in new spending initiatives taken since the budget and the 2004 budget contained another $3.6 billion in such measures. Yet the projected surplus, despite$4.8 billion in new spending, was still reported at $1.9 billion (year-end surprise). When the books finally closed on 2003-04 in October 2004, the surplus was$9.1 billion (final surprise).

There was a general agreement that this pattern impedes sound longer-term policymaking and limits a complete public debate over the policy alternatives available. Expected surpluses can be used in one of three ways – to reduce the federal government’s debt, reduce taxes, or increase spending – but, it was argued, these options narrow over time.

Under the current legislative framework, a final surprise (that comes after the books close) can be allocated only to debt reduction, since the accounting rules permit new spending or tax measures only if they have been announced and committed before the end of the fiscal year. Any change to this practice would require legislative authority to permit the allocation of a final surplus to one or more of spending, tax relief and debt reduction.[2] In the absence of such authority, one person suggested, the appearance of bigger-than-expected surpluses after the fiscal year-end has amounted to a policy of "stealth" debt reduction.

There are more options for a year-end surprise (i.e. added surplus discovered just before March 31), though some are of a limited nature. First, it can be used to reduce debt. Second, it can be allocated to one-time program spending, but only if that spending is clearly enunciated and for which institutional arrangements exist before the fiscal year ends on March 31 (transfers to the provinces or foundations, for example). Third, it can be used to finance a year-end tax cut (or rebate) that would return to taxpayers some of the taxes they have – in most cases – already paid.

Several criticized the use of year-end surprises for transfers to provinces or foundations as hasty last-minute decisions that are often subject to little debate or discussion beforehand and that are resorted to mainly as a means of reducing an unexpectedly large surplus in a hurry. One person commented that such "makeshift solutions are not seen as democratic" in a world "that prizes transparency and where respect for politicians has hit new depths." It is worth noting, however, that year-end spending initiatives are explicitly included in the (year-ahead) Budget – extra debt reduction is implicit – and, hence, amenable to discussion in the Budget debate.

To date, the federal government has chosen not to use tax rebates to reduce surplus surprises,[3] though some provinces have done so in allocating unexpected windfalls. Conventional tax cuts normally involve the announcement of permanent reductions that will go into effect on a given date. As such, they would apply only to the future and leave unaffected the finances of a fiscal year that is about to end. In this context, there was widespread agreement that regular surplus surprises effectively "load the dice" against such long-term tax cuts most of all and against increases in focused program spending as well, though to a lesser degree. Debt reduction, it was felt, is the default winner in any competition for the "extras" generated during the fiscal year.

### 2.2 The Economic and Fiscal Forecasts: Issues and Solutions

Developing the private sector forecast

The economic forecast used by the Finance Department in its budgets is based on projections made by a group of private sector forecasters. This is sometimes referred to as the "consensus" forecast, but the term should be used with care. The forecasters consulted by Finance do not thrash out their projections and come to some sort of broad agreement on the final numbers used. Rather, they present their forecasts and Finance then calculates an average of the numerical projections. Accordingly, we will avoid the use of the term "consensus" and use the word "average" instead.

One economist summed up several issues regarding the economic forecast process quite succinctly, and his list was echoed – in whole or in part – by several other forecasters who have also participated in the Finance Department’s forecast process.

First, the government takes all comers. It surveys a group of about 20 forecasters for their views of the economic outlook, but no one has to pass either a test of previous forecast success or a test ensuring that the forecast offered is internally consistent. That said, research has shown that average economic forecasts in general, both in Canada and elsewhere, have a better track record than any individual forecaster. This is supported by the quantitative analysis presented in Section 3.

The issue of how the average outlook should be calculated was also raised. It can be a simple average, or arithmetic mean, of the values forecast for each indicator (like nominal and real gross domestic product, the consumer price index and interest rates), a "trimmed mean" (in which the highest and lowest values are tossed out and a simple average taken of the remaining predictions) or something else, like the median (in which half the forecasters are above and half below the reported number). It is not conceptually clear which method is best. However, the analysis presented in Section 3.5 indicates that the choice of an averaging method makes little, if any, difference to the ultimate accuracy of the economic forecasts.

Third, in multi-year forecasts, participation drops off the further out you go. There may be 20 forecasts for Year One, but only four or five forecasts for Year Five. Most forecasters go out only two years, so the representative nature of the average from the third year on is arguably weaker than it is for the first two years.

Aside from these concerns, there was broad agreement that little more can be done on a technical level to improve the economic forecasts per se. A common comment was that the economic forecasters are professionals who are already doing their best to make accurate forecasts simply because they owe that effort to the institutions that employ them or the clients they serve. While errors in economic forecasts do contribute to inaccuracy in fiscal projections, as discussed in Sections 3.5 and 3.6, those are not avoidable through any changes in methodology or models.

From economic to fiscal forecast

Almost all the forecasters interviewed cited a general problem they face. They said data revisions frequently frustrate their efforts to predict economic growth in general and federal revenues and spending in particular.[4] In recent years, for example, Statistics Canada’s initial estimates of nominal GDP (which is the primary input to generating the government’s revenue base) were typically too low and later revised higher. One person noted that this is common during periods of economic growth, but when the economy goes into a downswing, the first estimates are often too high and later revised lower. Hence, a forecaster may be right in predicting the rate of growth for revenue or a spending item, but if the starting point for the level of revenue or spending turns out to have been wrong, then the resulting forecast of the level a year or two later (the starting-point level plus the percentage growth in that period) will also turn out to be wrong. As well, it can influence judgments about year-ahead revenue growth. This is a problem for both private forecasters and Finance.

More important to most of those interviewed is the process of converting the economic forecast into an accurate fiscal forecast – getting right the projections of federal government revenue and spending. For budget purposes, this is left to the Finance Department, which takes the economic forecasts from the private sector forecasters and uses those projections to develop its fiscal forecast of revenue and spending. There was widespread agreement that if there is a fiscal forecast accuracy problem, this is where it lies. The translation of economic projections to fiscal forecasts involves a high degree of uncertainty, an issue addressed at length in Section 3 below.

Impossible accuracy

Many of those interviewed made one broad point about fiscal forecasts, no matter when or how often they are done – viz., small misses in forecasting revenue and spending can produce big errors in forecasting the surplus, which is the residual when spending is subtracted from revenue. This is simple arithmetic. In 2003-04, the government’s revenue amounted to $186.2 billion and total spending was$177.1 billion, leaving a surplus of $9.1 billion. If revenue had been 1% lower and total spending 1% higher, the surplus would have been$5.4 billion, or 40% lower than the actual number. Had revenue been 1% higher and spending 1% lower, the reported surplus would have been 40% higher than reported.

Several of those interviewed observed that because many people do not understand this arithmetic, they expect a much greater degree of accuracy in forecasting the surplus than is realistic. One economist said that any mistake in predicting the surplus of less than $4 billion is trivial. His reasoning is that because an error of only 1% on roughly$180 billion in revenue and spending can put each number out by $1.8-billon, then such errors, if they go in opposite directions, can generate an error in the surplus of$3.6 billion, which he rounds to $4 billion. Another suggested that if total revenues are within 2% to 3% of the forecast on either side, the result is acceptable; an outcome outside that range means the forecasting job was not done properly. Spending forecasts, he added, should be more accurate than revenue forecasts. Beyond that, there is an active debate over the reasons for the government’s habit of winding up with surpluses that are bigger than predicted. Reasons for the surplus surprises Broadly speaking, those interviewed offered two reasons for the surplus surprises. The first group, comprising a large majority of those interviewed, saw the recent track record of larger-than-forecast surpluses as little more than the logical outcome of two forces – the government’s practice of budgetary prudence (seen mainly as a by-product of its determination to stay out of deficit) and several years in which the economy grew faster than expected. Such a combination, they argued, made bigger-than-forecast surpluses inevitable. Almost all thought the government’s caution was amply justified in the late 1990s, when it had just achieved its first surpluses after a quarter-century of deficits and was still wary of any slippage. But most thought some of this caution can now be relaxed somewhat because the government has established a solid pattern of surpluses and the ratio of debt to GDP is falling rapidly and almost certain to keep falling. This issue is, of course, the central focus of Section 3 (especially 3.5-3.7) and of Section 4.3. The second camp, which was very small, saw the surprises as the result of deliberate manipulation that is specifically designed to stifle a wider public debate over how to use future surpluses – tax less, spend more or reduce debt. This group noted that in the past decade, the government has always produced better-than-forecast fiscal results – smaller deficits than forecast in the early years and bigger surpluses than forecast in the later years. To them, such a record is clear evidence that the government has consistently and consciously "low-balled" its surplus projections so that it could resist pressure for higher spending and lower taxes. "You cannot explain away 11 years of forecast errors that were always in the same direction, not when the false forecasts limited policy choices," one said. "You have to argue the possibility or the probability that this was deliberate." The prudence question Among those who believe that surplus surprises are partly the result of abundant forecasting prudence, most thought this habit was bred into the Finance Department by Paul Martin who – as Minister of Finance from 1993 to 2002 – insisted that the department should always underpromise and overdeliver. Some argued that the department has gladly complied because its record during the 1980s and early 1990s had been one of forecasting steady substantial declines in the deficit, only to wind up with deficits that rarely went below$30 billion.

This perception of the record is, however, somewhat at odds with the facts. The department’s fiscal forecasting problems in the early 1980s and early 1990s were largely related to the severity of the 1981-82 and 1990-91 recessions and the associated lags on federal revenues and expenses. The 1994 Ernst & Young study showed that Finance’s fiscal forecasts were relatively on track during the late 1980s – from 1985-86 to 1989-90. Indeed, the difference might have been larger during the period had the private sector forecasts been used.

Nevertheless, among those interviewed, there was a common view that a history of predicting declines in the deficit that did not materialize was embarrassing and demoralizing for departmental officials, who were thus only too happy to follow a strict political directive to reverse the record. One of those interviewed vividly described the department’s experience in the 1980s and early 1990s as "a searing experience that might be analogized to the Great Depression" and added that it takes a long time for the effects of such an episode to disappear from an organization’s internal culture. Another scoffed at such an explanation: "It makes it too much of a psychological drama and not enough of a policy issue." Nevertheless, many argued that since Mr. Martin had laid down such a strong line as Finance Minister, it would be difficult (especially since he is now Prime Minister) for a successor to deviate from it. Several commented that the first Finance Minister to "go soft" on the deficit – to be seen to be risking a return to deficits – will attract enormous attention, partly because public opinion polls indicate that Canadians are opposed to deficit financing. One said that minister would wear "the cloak of shame."

Many saw it as ironic that Finance, having lost credibility before 1993 for its excessive optimism and forecasts of ever-dwindling deficits, now stands charged with a loss of credibility for the opposite reason. A few said the department is just as embarrassed now by its sequence of bigger-than-forecast surpluses as it was by its earlier record of bigger-than-expected deficits. Almost all agreed, however, that the department has indeed lost credibility, especially in the eyes of Parliamentarians and the media, and that this development does not enhance the budget process.

Sources of fiscal forecast error

There are three elements to the fiscal forecast – revenue, interest payments on the federal debt and program spending. The first is strongly affected by the growth of the economy – and thus of the government’s tax base – and the second by the course of interest rates, so both are influenced by the state of the economy at large. The third involves a myriad of programs and other considerations.

(i) Forecasting revenue

The revenue forecast may be seen as relatively straightforward, though difficult. Nominal GDP is the tax base, so if the government has a good handle on the ratio of revenue to GDP, its revenue forecast should be as good as its forecast of nominal GDP. Of course, it is not that simple. Several forecasters noted that it is important to know the composition of future GDP growth. If personal income growth is expected to be strong, but profits are expected to fall, this would imply substantial revenue growth from personal income taxes, but declining revenue from corporate income taxes. If the growth were generated more by the domestic economy than exports, then revenue from the goods and services tax would be higher than it would be if domestic growth faltered, but external demand remained strong. A second point mentioned by some forecasters is that the ratio of revenue to GDP can shift around considerably and is particularly prone to changing at times when economic growth is either accelerating or decelerating. This makes it difficult to predict.

(ii) Forecasting program spending

Several of those interviewed suggested that program spending is not so much forecast as managed. Most agreed that it is relatively easier to predict major items, like old age security payments and transfers to the provinces, that have been fixed by agreement. Interest charges on the public debt are usually forecast with a high degree of accuracy, some said. Spending on programs by departments can be fixed in advance – in that they are given a specific budget to live within – but may turn out to be lower than forecast. One person said program spending often winds up being lower than forecast because that is where the obvious bias lies. Departments cannot exceed their budgeted spending, but can fall short of it for a variety of reasons. For example, a major new program might be announced and a price tag attached, but if it is slow to get off the ground, spending in the current fiscal year might be much lower than expected. If there is going to be an error in predicting spending, then it will likely result in actual spending being lower than predicted spending, and that will inflate the final surplus.

Almost all, however, pointed to another source of forecast differences. The government regularly sets aside money for bad debts; it also sets aside money to allow for the possibility of losing litigation that is already under way or to allow for additional payments in contracts that are in dispute, but in the process of negotiation. When the government settled a multi-billion pay equity dispute several years ago, for example, it emerged that the sum involved had already been counted as spent and therefore the government did not have to make any additional allowances to pay what it owed. There have also been cases in which bad debts were written off for accounting purposes, but a subsequent reversal of fortune meant the debts in fact remained good and those sums were restored to the accounts. In both such situations, the costs to the government may turn out to be less than the provision made for it. This would not be the result of a deliberate choice but because outcomes were better than expected.

As it makes its forecast in each such case, the department has to make a judgment call on how much money to set aside. Several of those interviewed said officials are likely to estimate that the liability will be in a range from A (low) to B (high), and then choose a sum that might be 80% of the way to B. This might (and probably would) be done solely in the interest of prudence, they said, not deliberately to pad the spending side of the books. But if this pattern is repeated across several dozen line items in the accounts, even if the individual sums are small, the cumulative effect might turn out to be large. To avoid that problem, some suggested that when those individual estimates are made, the government should aim for the mid-point of any likely range, rather than one end of the range.

These comments apply not only to projections of future fiscal years, but of the current fiscal year as well. One person suggested this scenario might explain some of the final surprise in the 2003-04 surplus (the $9.1 billion rather than the$1.9 billion forecast in the March 2004 budget). Even though the budget was presented near the end of the fiscal year, Finance might not have had a clear idea of how many items might turn out. "If there are 30 things you don’t know, you go with the worst-case scenario on each of the 30," one person commented.

The no-deficit rule

The high level of budget forecast caution goes back to what some call the asymmetry issue – the insistence that deficits be avoided, because the small-p political cost of a deficit, in terms of the government’s credibility, is much higher than the cost of a surprisingly large surplus. This, several people argued, creates the underlying pressure to forecast spending levels that reflect (as one put it) "everything you could possibly imagine going wrong." The system, in effect, is so tilted in favour of surpluses that bigger-than-forecast surpluses are more likely as well. A related factor, one person suggested, is the Canadian economy’s exposure to international market shocks and – after September 11, 2001 – to domestic and North American security concerns. This reliance on trade and openness to the global economy means the Canadian economic outlook is often uncertain and the possibility of volatility is always present. This, he said, interacts with the no-deficit rule to reinforce any caution in making fiscal projections.

One person suggested that the practice of abundant – even excessive – caution was necessary in the mid- to late-1990s because "everyone had to believe they were serious" about eliminating the deficit. But he said this was a transitional process and, once credibility had been earned, the government should have eased back. Several people argued one point in almost the same language: As long as the government insists on never running a deficit, then surplus surprises are inevitable. Indeed, one observed that if the $3 billion contingency reserve is widely expected to turn into a surplus at the end of the fiscal year, then anything less than a$3 billion surplus is seen as failure – a de facto deficit. Three, one person quipped, is "the new zero."

The result, most of those interviewed suggested, is that forecasts of spending made in the Update and the Budget contain implicit reserves in addition to the explicit reserves (the contingency reserve and the allowance for economic prudence) published in each document. Several said the 2004 Update took a big step towards greater transparency by publishing a much more detailed presentation of program spending than was usual in the past.

Internal information-gathering

A few of those interviewed raised concerns about the government’s internal monitoring of its own expenses, arguing that Finance Canada and the Treasury Board Secretariat do not have as clear a fix as they once did on spending and other items during the fiscal year. One said that in 2003-04, three Crown corporations (Export Development Canada, Canada Deposit Insurance Corporation and Canada Post) reported a total profit of $2.5 billion, while the March 2004 Budget projected a profit of only$515 million for that fiscal year. This would suggest that there are weaknesses in the government’s internal information-gathering system. The same person noted that when the final figures for 2003-04 were released, one reason cited for the larger-than-expected surplus was that spending lapses (money approved but not actually spent) reached a historical high.[5] This should not have come as a surprise, he and others suggested, because the government implemented a spending freeze in December 2003, and many items of planned spending never got off the ground. This latter view is, however, overstated. In fact, the spending freeze applied only to capital spending; and because under accrual accounting such costs are accrued over the useful life of capital projects, there was little effect on expenses for the year.

### 2.3 Surplus Versus Deficit

No one questioned the government’s commitment to running surpluses and its abhorrence of ever again slipping into deficit; some noted that the Prime Minister reiterated this view in his pre-Christmas (2004) interviews with several media outlets. However, there was a sharp divide among those interviewed on the appropriateness of such a policy, and these views influenced the respondents’ views on how cautious the government should be in setting budget policy and how much prudence should be built into any budget.

Most – especially the economists – were fairly sanguine about the prospect of running small deficits. Though they understood the government’s reason for insisting on a continuation of regular surpluses, they argued that too rigid an adherence to this outcome is bad policy. Fiscal policy is seen to constitute what is called an automatic stabilizer for the broader economy. If the economy were to go into a recession, tax revenues would fall and some spending (employment insurance benefits, for example) would rise automatically. The resulting decline in the surplus – even to the point of slipping into deficit – would stimulate the economy and speed its return to growth.

Many expressed the fear that if the government were determined never to run a deficit, it might feel compelled, in order to balance the budget, to raise taxes or reduce spending during such a period and thereby exacerbate the economic downturn. If the federal government were, instead, to slip into a small deficit during an economic downturn, according to this view, so be it. A deficit would be nothing more than a sign that the system is working as it should. One proponent of this view said the government should not adjust taxes or spending to avoid deficits in a downturn and that it should balance this practice by not spending surpluses when the economy is doing well.

A few of those interviewed, however, felt that any deficit – however small – must be avoided. In most cases, they put forward variations on what might be called the "slippery slope" argument. If the government were to allow even a small deficit, the pressure to run even bigger deficits would become too strong to resist and the country would return to the kind of large federal deficits that characterized the 1980s and early 1990s. Clearly, this is a more political take on the issue, though in some cases, it was put forward by people who are personally at ease with small deficits, but see political danger in them. As one person put it, Parliamentarians and the general public would quickly realize that even with small deficits, the debt/GDP ratio would keep falling, though at a slower pace than it would with surpluses.[6] Having allowed even a small deficit – unless it was clearly the result of some major unforeseen catastrophe – the government would then have trouble finding a reason, clearly explainable to the public, for a return to surpluses.

Views were divided on how financial markets would react to a deficit, though this was not generally seen as a major concern. A few people suggested the markets would react badly to any sign that the government was losing its commitment to maintaining surpluses, but a similar number said an inadvertent slip into a small deficit would not bother investors, especially if it were clear that a deficit was the result of forces beyond the government’s control and that the government remained committed to returning to a surplus. A couple of those interviewed noted that financial market commentary was very negative in late 2001, when the government reduced its usual $3 billion contingency reserve to$1.5 billion for the 2001-02 fiscal year in the wake of the September 11 terrorist attacks in the United States. They expressed some puzzlement at this reaction. The reserve is explicitly designed to cushion the budget against unexpected negative shocks; surely, they said, the uncertainty after the terrorist attacks was a good reason to use it.

Where people stood on this issue of surpluses versus deficits was an important determinant of potential solutions they offered to the broader problem of fiscal forecasts. Specifically, it influenced their views on two questions – that of a so-called fiscal anchor and that of the appropriate degree of prudence needed in budgets.

### 2.4 Fiscal Anchor and the Role of Caution

Most of those interviewed stressed the importance of a fiscal anchor, which they defined as a strong fiscal policy goal that can be easily communicated to the public. The government’s original fiscal anchor in 1993 was to reduce the deficit-to-GDP ratio to 3% by 1996-97, 2% in 1997-98 and 1% in 1998-99. The 1998 budget introduced a target of a balanced budget (backed by a $3 billion contingency reserve) beginning with the 1997-98 fiscal year. After the books moved into surplus in 1998, the anchor was less clear, many felt, but the danger of returning to deficit still appeared real enough that the commitment to deficit avoidance was enough of an anchor to serve the government’s purposes. Most thought, however, that after several years of bigger-than-expected surpluses, the value of simple deficit avoidance as an anchor has diminished. Last year, the government announced an additional medium-term anchor – a goal of reducing the debt/GDP ratio to 25% over the next decade from 41% in the 2003-04 fiscal year. Most of those interviewed regarded the government’s target of a 25% debt ratio as reasonable, though most noted that it can be easily achieved through a combination of decent economic growth and small federal surpluses. One found it strange to create a target involving a ratio in which the government controls the numerator (debt) but not the denominator (nominal GDP). Setting a debt target in absolute terms was mentioned by a few, but not recommended as such; it was seen as simply another way of targeting specific surplus targets each year. For example, a goal of reducing the debt by$15 billion over five years is just another way of planning to run annual surpluses averaging $3 billion over the next five years. Discussions of a fiscal anchor, however, were intertwined with discussions of the government’s contingency reserve and other forms of prudential forecasting. Almost all of those interviewed endorsed the use of the annual contingency reserve (usually$3 billion) and the allowance for economic prudence (usually $1 billion in the first fiscal year of the projection and growing thereafter.) These were widely seen as reasonable sums. However, most said every effort should be made to ensure that any fiscal cushions should be limited to these explicit items. At the same time, most recognized that reserves set aside for the settlement of such things as litigation, potential contract settlements and possible bad debts should be maintained and cannot be made public. There was at least one firm dissent from a person who said the contingency reserve was crucial in the early stages of deficit reduction, but is now no longer needed. There was also a general recognition that the transparency of the budget process has been improved by the use of the contingency reserve (since 1994) and the economic prudence factor (expressed as a dollar value since 1999), both of which were thought to have also created a safety zone, or cushion, to allow for unforeseen events. One person suggested that the government could minimize the chance of going into deficit by setting the contingency reserve much higher than$3 billion. Under this proposal, the government would aim for a surplus of $10 billion, but manage its affairs as if the$10 billion level was really zero. In an economic downturn, the government could allow the surplus to fall to $2 billion, and thereby provide$8 billion worth of stimulus for the economy, while during a period of strong growth, it could allow the surplus to rise to $18 billion, and thereby provide$8 billion worth of restraint. This would allow for a much faster repayment of debt, he said, freeing up funds now used to pay interest for other purposes.

Some economists found appealing the use of what they called full-employment budget targets. If full employment were defined as a 7% unemployment rate, the government could set a target of what the budget balance should be at that level, and deviations from the target could then be attributed to the strength or weakness of the economy. One acknowledged that it would be difficult to explain this to the general public. It should be added that it would even be difficult to achieve widespread agreement among economists and other analysts on what constitutes "full employment."

### 2.5 Time Frame

From 1994 through 2004, the government provided two-year forecasts in the budget, but extended its outlook period to five years in the autumn Updates. Although some said five-year forecasts are of limited use because the future rarely turns out as expected over such a period, most thought the longer-term projection is still helpful because it can cast light on issues that will emerge further down the road, issues that policymakers and the public should begin thinking about.

This view was popular with those who would advocate further tax cuts, mainly because they said tax cuts should be permanent, so the government must be reasonably confident that its financial position is strong enough that it can afford to give up that portion of its future revenue flow.

Some suggested that a longer-term time frame (something actually introduced in the 2005 Budget) would allow the government to use internal reallocations of spending to achieve its desired targets. It is easier, for example, to change spending priorities over several years by freezing one area while allowing others to grow than to make outright cuts to a lower-priority program to free up funds for more important initiatives.

Three long-term issues

Those interviewed discussed three longer-term issues for budget policy that they thought should be raised soon for public debate. One is the impending retirement of the baby-boom generation, which several suggested will have implications for federal (and provincial) revenue and spending, a subject that is little discussed in the public arena and not fully understood. One said federal revenues will be affected by the fact that baby-boom women, more than earlier generations, have spent most of their lives in the workforce and will thus have more retirement income (which will be taxable) from workplace pensions and registered retirement savings plans. Others said health care costs will rise rapidly since the elderly account for a disproportionate share of health spending.

A second issue involves the government’s target of reducing its debt to 25% of GDP. Several people mentioned that the government will face some difficult decisions once that target is reached. If it continues to run surpluses – or even just balanced budgets – then the debt ratio will fall below 25% as the economy grows. If a future government were to decide that a 25% ratio is appropriate, it would need to run deficits to hold the ratio steady at 25%. In effect, it would have to increase its debt at the same rate of growth as nominal GDP. This is not an impossibly distant prospect, several noted. Assuming moderate economic growth and small surpluses, the 25% ratio would be reached early in the next decade, a period that will begin to show up in the five-year forecasts within the next two or three years. This issue, several argued, will come as a shock to a Canadian public that has come to value surpluses, so it will require a thorough debate, one that should begin several years before decisions need to be made.

A few people looked at these two longer-terms issues together. One person in particular argued that the government will have to run deficits in the 2020s to finance the needs of baby-boom generation Canadians when many will be in their 70s. Lowering debt now and in the near future will make it easier for the government to deal with demographic pressures later on by running small deficits that will increase the overall debt level (while keeping debt/GDP ratio constant). The question of a debt target, he said, should focus on such demographic issues rather than short-term considerations. In other words, he added, policymakers should begin to shift the entire budgetary discussion; rather than worry about how to allocate current surplus surprises, they should be more concerned about how to deliver services to people in their 70s twenty years from now.

A third issue involved the longer-term outlook for the Canadian economy. One person suggested that the possibility of a major adjustment in the United States economy poses a substantial risk for Canada. This might arise through a substantial decline in the value of the US dollar or increased protectionism, both of which would make it harder for Canadian industry to generate the kind of growth it has in the past decade. This in turn would affect government revenue.

In one way or another, many suggested, the government should provide longer-term projections that highlight such issues and give them greater prominence to encourage more public awareness and discussion. An example of this can be found in the 2005 Budget, which included a concise discussion of the demographic issue.

### 2.6 Institutional Remedies

In the public debate over surplus surprises, some have suggested that Canada should develop new forecast institutions that could operate at arm’s length from Finance Canada and thus act as a check on the government’s own forecasting of the fiscal outlook. The Congressional Budget Office in the United States is often cited as a possible model for such an institution. Most of those interviewed dismissed that idea as incompatible with Canada’s Parliamentary system of government. One said it would no more work here than Question Period would work in the US congressional system with its clear separation of powers between the executive and legislative branches of government.

Still, there was broad support for institutional remedies that would enrich the debate over budget issues in Parliament and among the general public. One person noted that the CBO, in addition to making budget forecasts, also carries out research studies into a variety of fiscal issues. A comparable Canadian research group could, for example, analyze the cost of government programs, or the revenue impact of various tax changes, or a variety of long-term fiscal issues facing the government.

Some suggested that the House of Commons Finance Committee should acquire additional resources to help it examine the government’s forecasts and deepen its understanding of fiscal issues. One said the Library of Parliament should be given enough additional funds to provide dedicated specialized research support to the Finance Committee. At the time they were interviewed, most were unaware of the Committee’s recent decision to commission its own quarterly economic forecasts, a proposal that envisages the Committee’s getting assistance from Finance Canada in using those forecasts to generate fresh fiscal forecasts.

Many were attracted by the idea of setting up some mechanism to study a range of medium- and longer-term fiscal issues. Although the basic idea had appeal, there was a variety of views on the nature of the appropriate mechanism. A few argued for a body like the old Economic Council of Canada, which was disbanded in the early 1990s. Others said this would amount to an unneeded additional layer of bureaucracy. A few others leaned towards the creation of a smaller body that could develop a research agenda and commission studies by experts in universities and private research organizations. Such a body would be patterned more along the lines of Canadian Policy Research Networks, the Centre for the Study of Living Standards, the Institute for Research on Public Policy or the Canada West Foundation, which have developed networks of researchers who pursue questions of interest.

### 2.7 Conclusion

Whatever their analysis, there was a strong consensus among those interviewed that better forecasting and more transparency are necessary, but they often came to this conclusion for entirely different reasons. Among those who would prefer to see more tax cuts or more spending on new or expanded programs, there was a view that better forecasting would open the door wider to their preferred policy options. Among those who give high priority to reducing the federal debt, there was an acknowledgement that the existing system has largely led to a result they like, but that the government’s credibility has been undermined in the process, a result they do not like.

One person in the former camp went so far as to say that if the forecasts were accurate, the process of arriving at them, no matter how flawed, was largely irrelevant. One person in the latter camp said he had long recognized problems in the process, but had held his tongue because he liked the outcome. Over time, however, he had come to worry about forecast accuracy because he believes excessive prudence is undemocratic and amounts to withholding information from Parliament.

All, however, would endorse solutions that would make the process more open, improve the forecast record and allow for a more complete and informed debate on budget policy.