Preface

This is the second edition of Tax Expenditures: Notes to the Estimates/ Projections, the companion document to the main report, Tax Expenditures and Evaluations. This document sets out the approach used in developing the estimates and projections contained in the main report. It also provides a description and presents the objective of each tax expenditure.

Since neither the approach taken to developing the estimates and projections nor the description and objectives of most of the tax expenditures are likely to change from year to year, this document is produced less frequently than the main report, which is published annually.

Two presentational changes have been made to this edition of the Notes. First, the estimates and projections for the business income tax and goods and services tax measures are now grouped by functional category (e.g. small business), as they were already in the first edition for personal income tax measures. Second, subheadings have been added to the "Memorandum Items" to identify various categories of measures that are part of the benchmark tax system (e.g. recognition of expenses incurred to earn income).

The main report continues to provide estimates and projections for all tax expenditures. It also contains results of evaluations or research related to specific tax expenditures or tax issues.

Chapter 1

Framework and Methodology

The principal function of the tax system is to raise the revenues necessary to fund government expenditures. How much revenue is raised is determined by tax bases and tax rates. It is also a function of a range of measures—special tax rates, exemptions, deductions, rebates, deferrals and credits—that affect the level and distribution of tax. These measures are sometimes called "tax expenditures" because they have an impact on government revenue (i.e. they have a cost) and they reflect policy choices of the Government.

In order to define tax expenditures, it is necessary to establish a "benchmark" tax structure that applies the relevant tax rates to a broadly defined tax base—e.g. personal income, business income or consumption. Tax expenditures are then defined as deviations from this benchmark. Reasonable differences of opinion exist about what should be considered a benchmark tax system and hence about what should be considered a tax expenditure. For example, a deduction for expenses incurred in earning income is generally presented as part of the benchmark and thus not as a tax expenditure. But in some cases the deduction may confer some personal benefit, making its classification ambiguous.

This report takes a broad approach and includes estimates of the forgone revenue associated with all but the most fundamental structural elements of the tax system, such as the progressive personal income tax rate structure. This includes not only measures that may reasonably be regarded as tax expenditures but also other measures that may be considered part of the benchmark tax system. The latter are listed separately under "Memorandum Items." For instance, the dividend tax credit is listed under this heading because its purpose is to reduce or eliminate the double taxation of income earned by corporations and distributed to individuals through dividends. Also included under this heading are measures for which there may be some debate over whether they should be considered tax expenditures or where data limitations do not permit a separation of the tax expenditure and benchmark components of the measure. This approach provides information on a full range of measures.

The remainder of this chapter discusses the tax expenditure concept in order to facilitate understanding of the quantitative estimates. It also discusses the calculation and interpretation of the costs of tax expenditures, including key assumptions used in the analysis.

Simplified descriptions of each tax expenditure as well as information on data sources and methodology used in constructing the estimates are presented in Chapter 2 (personal income tax), Chapter 3 (corporate income tax) and Chapter 4 (goods and services tax[GST]/harmonized sales tax [HST]).^[1] When there is some debate about the appropriate classification of a measure as a tax expenditure or as part of the benchmark, this is noted in the description.

Benchmark for Tax Expenditures in the Personal and Corporate Income Tax Systems

The benchmark for the personal and corporate income tax systems is defined by considering the existing tax rates and brackets, the unit of taxation, the time frame of taxation, and the treatment of inflation for calculating income. In addition, the benchmark includes measures designed to reduce or eliminate double taxation, to recognize expenses incurred to earn income, and to improve the fairness of the income tax system that generally uses a year as the time frame for taxation. Finally, the constitutional immunity from taxation of Canada or any province is recognized as part of the benchmark system for income taxation.

The following provides a more detailed discussion of the features of the benchmark for both the personal and corporate income tax systems.

(1) Tax Rates and Income Brackets

For the personal income tax system, the existing rate structure, adjusted for inflation, is taken to be part of the benchmark system. The basic personal credit is also presented as part of this structure since it is universal in its application and can be viewed as providing a zero rate of tax up to an initial level of income. The cost of this credit is, however, included as a memorandum item.

With respect to the corporate income tax system, the benchmark is the basic federal corporate tax rate including the surtax and the provincial abatement. Provisions that alter this tax rate for certain types of activities or corporations are regarded as tax expenditures. These include the low tax rate for small business and the low rate for credit unions. The federal capital tax, levied at the existing rate, is considered to be part of the benchmark tax system.

(2) Tax Unit

Personal income taxes in Canada are based on individual income. Consequently, the individual is taken as the benchmark tax unit for the purposes of identifying tax expenditures in this report. This choice leads to the classification of the various provisions related to dependants, such as the spouse or common-law partner credit, as tax expenditures.

For corporate income tax, the single corporation in its entirety is adopted as the benchmark tax unit. This approach is the most prevalent in the corporate income tax system. For example, income from one part of a business can be offset by other business losses within the same corporation, whereas losses by one corporation may not generally be used against the income of another corporation in the group.

Other possible choices for the corporate tax unit are the establishment or activity unit within a corporation and the consolidated group of related corporations. The present income tax system contains elements of these approaches:

• Consistent with the activity unit basis, certain deductions, such as accelerated capital cost allowances on mining assets, may only be taken against income from the related project.

• Consistent with the consolidated group basis, rollover provisions allow corporate groups to reorganize their corporate structures without triggering capital gains or recaptured depreciation. These features are also included in the benchmark tax system.

(3) Taxation Period

The benchmark taxation period for the personal income tax system in this document is the calendar year. Accordingly, any measures that provide deferrals of taxable income to a subsequent year are considered to be tax expenditures. For example, farmers are permitted to defer the receipt of income from the sale of grain through the use of special cash purchase tickets, and this is listed as a tax expenditure.

The benchmark taxation period for the corporate income tax system is the corporation’s fiscal year. As with the personal income tax system, deferrals are considered to be tax expenditures.

A strict application of the annual taxation period would imply that measures that provide for the carry-forward and carry-back of losses to other years would be tax expenditures. However, the relatively cyclical nature of business and investment income suggests that such income should be viewed over a number of years. Consequently, carry-overs of losses are presented as part of the benchmark tax system in this report and their estimated cost is presented in the memorandum item sections.

(4) Treatment of Inflation

Although the personal income tax brackets and the major credits and exemptions have been indexed since 2000, individuals report nominal income in determining their tax liability each year, as do corporations. Nominal income is therefore taken to be the appropriate basis for the benchmark income tax system.

(5) Avoidance of Double Taxation

Conceptual difficulties arise in deciding whether certain provisions that reduce or eliminate double taxation should be presented as tax expenditures.

For example, regarding the personal and corporate income tax systems as completely separate would suggest that the dividend tax credit is a tax expenditure. However, the credit is an essential feature of the overall (i.e. both corporate and personal) income tax structure and serves to eliminate or reduce double taxation. In its absence, income earned through corporations would be taxed twice, once in the corporation and once at the personal level. For this reason, the dividend tax credit is not presented as a tax expenditure.

Similarly, the non-taxation of intercorporate dividends is designed to ensure that income is taxed only once at the corporate level and that the corporate income tax system is neutral across organizational structures. For example, consider a single corporation that currently operates as a number of divisions. Now suppose the corporation reorganizes into a holding company with wholly-owned subsidiaries instead of divisions. The profits from the subsidiaries flow to the holding company through intercorporate dividends. If these dividends were subject to taxation at both the subsidiary and the holding company levels, double taxation would occur. Consequently, the exemption of intercorporate dividends is not considered a tax expenditure.

Information on some of the measures that provide relief from "double taxation" is provided in the appropriate memorandum sections of this report.

(6) Recognition of Expenses Incurred to Earn Income

Tax provisions that provide for the deduction of current costs incurred to earn income are treated as part of the benchmark system. For example, the deduction for child care expenses and the moving expense deduction are presented not as tax expenditures but as memorandum items.

(7) Government Immunity From Taxation

Section 125 of the Constitution Act, 1867, provides that "no land or property belonging to Canada or any province shall be liable to taxation." This means that neither the federal nor the provincial governments (nor their Crown agents) are liable to taxation by the other. Accordingly, constitutional immunity from taxation is recognized as part of the benchmark system for income taxation.

Features of the GST/HST Benchmark^[2]

The benchmark system used to analyze the GST/HST is a broadly based, multi-stage value-added tax collected according to the destination principle and using a tax credit mechanism to relieve the tax in the case of business inputs. The following provides a more detailed discussion of the features of the GST/HST benchmark.

(1) Multi-Stage System

The main structural elements of a multi-stage consumption tax are taken to be part of the benchmark. Under the multi-stage system, tax is applied to the sales of goods and services at all stages of the production and marketing chain. At each stage, however, businesses are generally able to claim tax credits to recover the tax they paid on their business inputs. In this way, the tax system has the effect of applying the tax only to the value added by each business. Since the only tax that is not refunded is the tax collected on sales to final consumers, the tax rests ultimately on final consumption.

(2) Destination-Based

The benchmark system applies tax only to goods and services consumed in Canada. Accordingly, the tax applies to imports as well as domestically produced goods and services. Exports are not subject to the tax.

(3) Single Tax Rate

The benchmark system has only one tax rate. This rate corresponds to the statutory rate of 7 per cent. As a result, GST/HST provisions that depart from this single rate are considered to be tax expenditures.

(4) Taxation Period

The benchmark taxation period is the calendar year.

(5) Government Immunity From Taxation

Section 125 of the Constitution Act, 1867, provides that "no land or property belonging to Canada or any province shall be liable to taxation." This means that neither the federal nor the provincial governments (or their Crown agents) are liable to taxation by the other. This is recognized as part of the benchmark system for the GST/HST.

The benchmark also recognizes that the federal and provincial governments have taken steps to simplify the operation of the tax for transactions involving government sectors.

• The federal government decided to apply the GST/HST to purchases by federal departments and Crown corporations in order to keep the tax as simple as possible for vendors. As a result, the GST/HST and the benchmark system treat federal Crown corporations in the same manner as any other business entity.

• By virtue of section 125, provincial governments and Crown agents are not liable to pay the GST/HST on their purchases. However, the federal government and most provinces have entered into reciprocal tax agreements. These agreements specify situations in which each level of government agrees to pay the sales taxes of the other, and generally this involves applying tax to purchases made by Crown corporations. As a result, provincial Crown corporations are treated like any other business entity in the benchmark system.

Unlike provincial governments, municipalities are required to pay tax. In addition, most universities, public colleges, schools and public hospitals are required to pay tax. The benchmark system distinguishes between two different situations when treating these sectors. The first situation is where these sectors provide public services that are completely funded through taxes or government transfers. In this situation, the benchmark system treats these sectors as final consumers—that is, they pay tax on their purchases and may not claim input tax credits (e.g. a public hospital is treated as the final consumer of the medical and other supplies that it purchases to provide health care services covered under provincial health insurance plans.) The second situation is where these sectors sell goods and services to consumers and businesses. In this situation, the benchmark system treats these sectors just like any other business that applies tax to its sales and claims input tax credits for the tax paid on associated inputs (e.g. a public hospital that charges for certain treatments not covered under provincial health insurance plans is treated as a business for the purpose of these treatments).

Types of GST/HST Tax Expenditures

Comparing the actual structure of the GST/HST to the benchmark system, it is possible to identify four main types of tax expenditure:

• zero-rated goods and services;
• tax-exempt goods and services;
• tax rebates; and
• the GST/HST credit.

(1) Zero-Rated Goods and Services

Under the GST/HST, certain categories of goods and services are taxed at a "zero" rate, rather than at the general tax rate of 7 per cent. Vendors do not charge GST/HST on their sales of zero-rated goods and services (whether these sales are to other businesses or to final consumers). However, vendors are entitled to claim input tax credits to recover the full amount of GST/HST they paid on inputs used to produce zero-rated products. As a result, zero-rated goods and services are tax-free.

One category of zero-rated sales is basic groceries—i.e. foods intended to be prepared and consumed at home. Other categories of zero-rated goods include prescription drugs, medical devices and most agricultural and fish products.

(2) Tax-Exempt Goods and Services

Some types of goods and services are exempt under the GST/HST. This means that the GST/HST is not applied to these sales. Unlike zero-rated goods and services, however, vendors of exempt goods and services are not entitled to claim input tax credits to recover the GST/HST they paid on their inputs.

Examples of tax-exempt goods and services include long-term residential rents, most health and dental care services, day-care services, most sales by charities, most domestic financial services, municipal transit and legal aid services.

(3) Tax Rebates

Certain sectors are eligible for rebates on all or a portion of the GST/HST paid on inputs used in the provision of tax-exempt services. For example, there are rebates for schools, universities, public colleges, public hospitals and municipalities. These rebates were implemented when the GST was introduced to ensure that these institutions would not bear a greater tax burden on their purchases under the GST than they would have under the manufacturers’ sales tax, which the GST replaced. Effective February 2004, municipal governments were granted a full rebate on their GST/HST payments. These rebates are treated as tax expenditures because, under the benchmark system, these institutions are considered to be final consumers.

Other examples of tax rebates treated as tax expenditures include the rebates for charities, substantially government-funded non-profit organizations, newly built housing, new residential rental property and book purchases made by qualifying institutions.

Also, foreign visitors to Canada are able to claim a rebate for the GST/HST they pay on hotel accommodation and on goods they take home. Only the rebate for hotel accommodation is considered to be a tax expenditure, however, because goods taken home by foreign visitors are effectively exports, which are not taxable under the benchmark system.

(4) GST/HST Credit

To ensure that the GST/HST system is fair, a GST/HST credit is provided through the personal income tax system to single individuals and families with low and moderate incomes. The credit is paid by cheque four times a year in equal instalments. The total amount of the credit depends on family size and income and is calculated annually based on information provided in personal income tax returns.

Calculation and Interpretation of the Estimates

The estimates indicate the annual cash-flow impact—not time discounted—to the Government of each measure, and not their long-run or steady-state revenue cost, subject to the following limitations:

• all measures are evaluated independently; and
• all other factors remain unchanged.

These methodological distinctions are important and have implications for the interpretation of the estimates. These concepts are discussed in further detail below.

Independent Estimates

The estimate of the cost of each tax expenditure is undertaken separately, assuming that all other tax provisions remain unchanged. An important implication of this is that the estimates cannot be meaningfully aggregated to determine the total cost of a particular group of tax expenditures or of all tax expenditures combined.

As explained in more detail in the following paragraphs, this restriction arises from the fact that:

• the income tax rate structure is progressive; and
• tax measures interact with one another.

Progressive Income Tax Rates

The combined effect of claiming a number of income tax exemptions and deductions may be to move an individual to a lower tax bracket than would have applied had none of the tax measures existed. To the extent that this occurs, aggregation of the individual estimates may under-represent the "true" cost to the federal government of maintaining all of them. For example, consider a taxpayer whose taxable income was $1,000 below the level at which he or she would move from the 16-per-cent into the 22-per-cent tax bracket. Imagine that this taxpayer arrives at this level of taxable income by using two tax deductions of $1,000 each (e.g. the deduction for home relocation loans and for registered retirement savings plan [RRSP] contributions). Eliminating either deduction by itself would increase taxable income by $1,000 and the taxpayer’s federal tax liability by $160. Eliminating both measures simultaneously, however, would not raise the tax liability by $160 + $160, but rather by $160 + $220.

Aggregating the individual estimates for these two items would provide a misleading impression of the revenue impact of eliminating both of them. Therefore, the personal income tax expenditure estimates in this document cannot be meaningfully aggregated to determine the total cost of a particular group of tax expenditures or of all tax expenditures combined.

While there is only one statutory tax rate for corporations, the low tax rate for small business creates a de facto progressive tax rate schedule for some corporations. In this way, the above argument is valid for the corporate income tax system as well, although the effect is not as large as for personal income taxes.

Interaction of Tax Measures

As noted above, the estimates are computed one at a time, assuming all other provisions remain unchanged. Given that tax provisions sometimes interact, the total cost of a group of tax expenditures calculated individually may differ from the dollar value of calculating the cost of the same group of tax expenditures concurrently. This is because adding the independently estimated costs of the tax provisions would result in double counting and so would not provide an accurate measure of the revenue that would be generated by simultaneously altering a group of measures.

For example, consider the non-taxation of veterans’ allowances, which reduces the recipient’s net income. Many measures, such as the medical expense tax credit, are calculated on the basis of net income. Thus, the reported estimate for the non-taxation of veterans’ allowances represents not only the direct impact on government receipts of not taxing the allowances, but also the indirect impact of the change on the cost of other tax measures (such as the medical expense tax credit) that depend on net income.

Since estimates for GST/HST tax expenditures are made using the same methodological approach as for income taxes, they too cannot be aggregated because they may interact. The following discussion of hospital rebates and zero-rating of prescription drugs illustrates the differences between independent and concurrent estimates for these two provisions.

• Eliminating hospital rebates: If hospital rebates were eliminated, hospitals would no longer be able to recover 83 per cent of the GST/HST they pay on their purchases.^[3] However, they could continue to purchase prescription drugs on a tax-free basis because these drugs are zero-rated. The estimate for hospital rebates recognizes that the rebate would not have been claimed in respect of zero-rated prescription drugs.

• Eliminating the zero-rating of prescription drugs: If prescription drugs were taxed at the GST/HST rate of 7 per cent, then hospitals would pay the tax on their drug purchases but recover 83 per cent of the tax through the rebate system. Therefore, the estimate for the zero-rating of prescription drugs is calculated as net of the expected increase in the payment of hospital rebates.

• Eliminating the two measures concurrently has a revenue impact greater than the sum of the independent estimates because the GST/HST would be payable on prescription drugs and hospitals would be unable to claim a rebate for these purchases.

Federal-Provincial Interaction

The federal and provincial income tax and sales tax systems interact with each other to various degrees. As a result, changes to tax measures in the federal system may have consequences for provincial tax revenues. In this publication, however, any such provincial effects are not taken into account—that is, the tax expenditure estimates address federal revenue only.

All Other Factors Remain Unchanged

The estimates represent the amount by which federal tax revenues are reduced due to the existence of each preference, assuming that all other factors remain unchanged.

In order to evaluate the extent of the revenue reduction, the approach taken here is to recalculate federal revenues assuming the measure in question has been eliminated. The difference between this recalculated amount and actual revenues provides the quantitative estimate of the cost of the tax expenditure.

The assumption that all other things remain unchanged means that no allowance is made for: (i) behavioural responses by taxpayers; (ii) consequential government policy changes; or (iii) changes in tax collections due to altered levels of aggregate economic activity that might result from the elimination of a particular tax measure (further detail is provided below). Incorporating these factors would add a large subjective element to the calculations.

(1) Absence of Behavioural Responses

In many instances, the removal of a tax expenditure would cause taxpayers to change their behaviour to minimize the amount of extra tax they would have to pay, perhaps by making greater use of other tax measures. Therefore, the omission of behavioural responses in the estimating methodology generates cost estimates that may exceed the revenue increases that would have resulted if a particular provision had been eliminated.

The effects of this assumption can be illustrated for the GST/HST by considering the housing rebate. Homeowners are eligible for a rebate of the GST/HST they pay on the purchase of new houses. If this rebate were eliminated, the price of new houses would increase relative to the price of used houses. This, in turn, might reduce the demand for new houses while increasing the demand for used houses (which are tax-exempt). Since the dynamics of the housing market are not taken into account, the revenues obtained by eliminating the housing rebate could actually be lower than the indicated estimate.

(2) Consequential Government Policy Changes

The estimates ignore transitional provisions that might accompany the elimination of a particular measure and take no account of other consequential changes in government policy. For example, if the Government were to eliminate a particular tax deferral, it could require the deferred amount to be brought into income immediately.

Alternatively, it might prohibit new deferrals but allow existing amounts to continue to be deferred, perhaps for a specified period of time. The estimates do not provide for any such transitional relief.

Similarly, the estimates make no allowance for consequential government policy changes. For example, if capital gains on owner-occupied housing were made taxable under the personal income tax system, an argument could be made that the cost of maintenance should be deductible in the same way as other investment expenses.

The estimates do not take into account the potential impact of a particular tax provision on the overall level of economic activity and thus aggregate tax revenues. For example, although eliminating the low tax rate for small business could generate a significant amount of revenue for the Government, the level of activity in the small business sector could decline. That in turn could cause job losses, a reduction in taxable income and, hence, a reduction in the amount of tax revenue collected. Furthermore, the estimates do not include speculation on how the Government might use the additional funds available to it and the possible impacts this could have on other tax revenues.

Certain tax measures defer income taxes from the current taxation year to a later one—for example, by accelerating deductions or by deferring income inclusions. Estimating the cost of tax deferrals presents a number of methodological difficulties since, even though the tax is not currently received, it may be collected at some point in the future. It is therefore necessary to derive estimates of the cost to the government of providing such a tax deferral while at the same time ensuring comparability with the other tax expenditure estimates.

Income tax deferrals are estimated on a nominal cash-flow basis—that is, the cost is the forgone tax revenue associated with the net deferral in the year. The estimates thus computed provide a picture of the ongoing cost of maintaining a particular tax provision in a mature tax system.

On a nominal cash-flow basis, deferred income taxes from current-year activities represent a positive tax expenditure while income taxes on previous-year activities for which the deferral has been completed are a negative tax expenditure. Thus, if the level of activity in question were constant from year to year, in a steady state the two amounts would cancel each other out and the tax expenditure would be zero. An increase over time in the level of activity would tend to produce a positive tax expenditure, while a decrease would tend to produce a negative tax expenditure.

While the cash-flow basis of measurement suggests that, in a steady state, there is no overall cost to the government from deferrals, there is indeed a cost to the government and a benefit to the taxpayer because of the time value of money. Because of the time value of money, a reduction in tax of a given amount today more than offsets a tax increase of the same nominal amount in a future period. This can be demonstrated with a calculation of the value of the implicit interest-free loan that is provided to the taxpayer when taxes are deferred to a later year. For example, if a taxpayer is able to defer $100 in income tax for one year, and the discount rate is eight per cent, then the present value of the future obligation is $92.59 and the taxpayer has received a benefit of $7.41 in today’s dollars. There is an equivalent implicit interest cost to the government.

Unlike the cash-flow basis, under this approach in a steady state a tax deferral would result in a positive tax expenditure. With the exception of tax-assisted retirement savings plans and some illustrations of the impact of accelerated write-offs for capital expenses, this publication does not generally provide present-value estimates of tax expenditures.

Developing Historical Estimates

The majority of the personal income tax estimates were computed with a personal income tax model. This model simulates changes to the personal income tax system using the statistical sample of tax returns collected by the Canada Revenue Agency (CRA) for its annual publication Tax Statistics on Individuals. The model estimates the revenue impact of possible tax changes by recomputing taxes payable on the basis of adjusted values for all relevant income components, deductions and credits. For example, the removal of the moving expense deduction would result in a change not only in net income but also in all of the credits, such as the medical expense tax credit, whose values depend on net income. For those tax expenditures whose costs could not be estimated using this model alone, supplementary data were acquired from a variety of sources. Details on data sources and the methodologies used for estimating the cost of specific personal income tax measures are provided in Chapter 2.

A corporate income tax model was used to measure most of the corporate tax expenditures. As with the personal income tax model, it is based on a statistical sample of tax returns collected by the CRA and is able to recompute taxes payable on the basis of adjusted tax provisions. This recomputation of taxes takes into account the availability of unused tax credits, tax reductions, deductions and losses that would be used by corporations to minimize their tax liability. Where costs could not be estimated using this model alone, supplementary data acquired from a variety of sources were used. Details on these sources are provided in Chapter 3.

The costs of the majority of the GST/HST tax expenditures were estimated using a Sales Tax Model based on Statistics Canada’s Input-Output Tables and the National Income and Expenditure Accounts. Where estimates could not be derived using this model, supplementary data from a variety of sources were used. Details on both the data sources and methodologies are provided in Chapter 4.

Developing Future Projections

As with the historical estimates, the projections represent the estimated amount of forgone federal revenue from a tax expenditure, assuming that all measures are evaluated independently, and that all other factors remain unchanged. The projections do, however, take into consideration the impact of announced tax changes.

In contrast to the estimates of tax expenditures for the historical period, when values of the tax expenditures can generally be obtained from tax statistics or other historical data, projections of tax expenditures must rely on estimated relationships between tax expenditures and explanatory economic variables. Using these relationships, the values for the explanatory variables are projected into the future and so permit an estimation of the future expected values of tax expenditures. Key explanatory variables are generally those reflecting the state of the economy.

Projections for the explanatory variables are based on either the most recent budget forecasts (e.g. gross domestic product [GDP], population, employment, corporate profits, inflation and consumer spending) or on past trends in the tax expenditure. Where projected tax expenditures were not obtained using these approaches, information on the alternative methodology is provided in Chapter 2 for personal income tax, Chapter 3 for corporate income tax and Chapter 4 for GST/HST tax expenditures.

Any projections are inherently subject to forecast error and quite substantial errors at times. Analysts familiar with forecasts prepared for the Canadian economy, or for any other economy, recognize that forecasting is not a science. Future values for key explanatory variables are based on best judgements, and unchanged policies are assumed for the forecast period. Furthermore, the relationships between variables that are being explained and those that provide the explanation may not be robust and could quickly change over time. For all these reasons, the projected values of tax expenditures should be presented as "best efforts" that do not have any greater degree of reliability than the variables that explain them. For example, if the level of GDP influences the revenue impact of a tax expenditure, one would not expect the projected level of tax expenditure to materialize if the expected level of GDP did not occur. Even if the expected level of GDP did materialize, the level of the tax expenditure might not if, in the future, the relationship between the tax expenditure and GDP turns out to be different from that estimated on average in the past. Therefore, in general, one should expect that the degree of reliability of the projected tax expenditures should be less than that of the underlying explanatory variables.

- Table of Contents - Next -