individuals, personal tax, consumption, tax-free savings accounts, income, capital income, canadian tax journal, capital gains, tax revenues, consumption tax, tax-deferred savings, individual, tax base, Alarie, marginal tax rates, tax rate schedule, upfront tax deduction, savings account, savings plans, tax-free savings account, investment returns, Benjamin Alarie, method, investment return, federal benefit, clawbacks, tax burdens, tfsa contribution limits, Spender, goods and services tax, Policy Forum, identical individuals, tax administration, tax avoidance, domestic savings, Horizontal Equity, income tax, tax incentives, registered pension plans, income-based, differential returns, privately held, real rate of return, individual retirement accounts, individual investor, savings accounts, Tax Collections Alarie, constant dollars, personal tax system, capital income taxes, labour income, Simpler method, marginal tax rate, asset tests, annual return
canadian tax journal / revue fiscale canadienne (2009) vol. 57, no 3, 533- 6 2 Policy Forum: Tax-Free Savings Accounts in a Consumption-Based Personal Tax Jonathan R. Kesselman* Abstract This study investigates shifting the personal tax system further toward a consumption base using the tax-prepaid method, as embodied in the tax-free savings account (tfsa). The study critically examines the negative assessment of the tax-prepaid method and the tfsa developed in Benjamin Alarie's companion article. Key findings are that a tax-prepaid savings scheme has unique and valuable attributes to add to the existing provisions for tax-deferred savings through registered retirement savings plans (rrsps) and registered pension plans (rpps); that the tax-prepaid and tax-deferred schemes can operate jointly in ways that provide complementary benefits; and that Alarie's concerns about failure to tax supernormal returns, the timing of tax revenue
s, behavioural aspects of savings, and political economy
considerations do not undercut the attractions of the tax-prepaid method or its tfsa format. Some of the design features of the tfsa--such as the ability to recontribute withdrawals and immunity from benefit clawbacks--might usefully be transplanted into rrsp and rpp provisions. However, that would not diminish the unique advantages of the tfsa relative to the tax-deferred schemes: greater economic efficiency and operational simplicity, and less loss of vertical equity. This assessment of the tfsa is set within the context of the benefits of a consumptionbased tax as well the reasons for not moving the personal tax base fully to consumption. Additionally, the study assesses alternative policies to the tfsa for shifting the personal tax toward consumption and refinements that could be made to the tfsa. The alternative policies canvassed here include increased rrsp/rpp contribution limits, tax-free capital gains
rollovers, a capital income exclusion, and a shift toward more reliance on indirect taxes like the goods and services tax; none is found to be superior to the introduction of the tfsa. Nevertheless, the tfsa is still in its infancy and could benefit from several refinements: specifying limits on immunity from federal benefit and tax-related clawbacks; making similar provision for partial immunity from clawbacks in provincial means-tested benefit programs; and allowing higher annual tfsa contribution limits for older cohorts over an extended transition period. In short, the tax-free savings account is an important and useful innovation for Canadian tax policy and is certain to attract continuing attention from both tax planners and tax analysts. Keywords: Consumption taxes n savings plans n TFSA n prepaid n deferred income taxes * Of the Graduate Public Policy Program, and holder of the Canada Research Chair in Public Finance, Simon Fraser University, Vancouver (e-mail: [email protected]
534 n canadian tax journal / revue fiscale canadienne Contents Introduction Consumption Versus Income Base for Personal Tax Savings, Efficiency, and Growth Horizontal Equity Administrative and Compliance Simplicity Vertical Equity Why Not Go All the Way? Tax-Deferred Versus Tax-Prepaid Methods Characteristics of the Two Methods RRSP Versus TFSA Design Features Supernormal Investment Returns Distribution of Taxes and Equity Concept Impact on Aggregate Tax Collections Behavioural Factors and Saving Policy Alternatives to TFSAs Increased RRSP/RPP Contribution Limits Tax-Free Rollover of Capital Gains Limited Exclusion of Capital Income Shift the Direct-Indirect Tax Mix Potential Reforms to TFSAs Limited Immunity from Federal Benefit Clawbacks Immunity from Provincial Benefit Clawbacks Differential Contribution Limits by Age Bringing TFSA Features to Tax-Deferred Plans Political Economy of TFSAs Conclusion
(2009) vol. 57, no 3 534 535 537 539 540 540 541 543 544 548 548 550 552 553 554 554 555 556 556 557 557 558 559 560 560 561
Introduction In the preceding Policy Forum article, Benjamin Alarie argues at length that the recently introduced tax-free savings account (tfsa) is a deficient method of reforming Canada's tax provisions for savings. He concludes that modifying and expanding the traditional tax-deferred1 schemes (registered retirement savings plans [rrsps] and registered pension plans [rpps]) would have been superior to introduction of the tfsa. Alarie's analysis focuses on the differential properties--both economic and political economy--between using tax-deferred and tax-prepaid methods of providing tax incentives for savings. He considers four economic aspects that could differ under the two methods: (1) the existence of supernormal returns on investments and the implications for interpersonal equity and taxation revenues; (2) potential 1 Alarie uses the term "tax-postpaid" for these schemes, following the usage in some of the taxlaw literature; I prefer to use the more traditional term "tax-deferred." The tax-deferred method has also been called a registered accounts method or a cash flow consumption tax. The taxprepaid method has also been called a yield-exemption consumption tax, a wage tax, and a labour income
tax-free savings accounts in a consumption-based personal tax n 535 differences in savings incentives attributable to various behavioural factors; (3) variations in tax rates over an individual's lifetime and over time; and (4) the timing of collection of tax revenues. Alarie argues that political economy considerations also favour the tax-deferred method. Here I deconstruct Alarie's analysis and conclude that a tax-prepaid scheme such as the tfsa carries important benefits for Canada's personal tax system. My study proceeds in the following fashion. First, I review the arguments for shifting the direct personal tax further toward a consumption base;2 both economic and pragmatic dimensions are explored. I also indicate reasons for policy to stop short of a fully consumption-based tax. Next, I examine the key alternative ways of implementing a consumption base in a direct tax: tax-prepaid accounts and tax-deferred accounts, as well as their specific policy manifestations of tfsa and rrsp, respectively. In two separate sections, I assess the taxation of supernormal returns and saving responses to tax provisions, issues that play a central role in Alarie's evaluation.3 I then consider directions that Canadian tax policy could have pursued instead of the tfsa and assess those alternatives, followed by aspects of tfsa design and implementation that merit further refinement. Finally, I offer a brief discussion of political economy aspects, followed by a concluding comment. Consumption Versus Income Ba s e f o r P e r s o n a l Ta x The personal "income" tax (pit) is nominally a tax on income, but the Canadian pit is in fact much closer to a tax on consumption for at least 95 percent of taxpayers. Only the top-income, top-wealth individuals bear personal tax on most of their capital-type income in addition to their labour earnings. Prominent features of the Canadian pit that exclude tax on returns to capital are the provisions for taxdeferred savings (rrsps and rpps), tax-prepaid savings (tfsas), and tax exclusion on returns to owner-occupied housing (both the flow of services and capital gains upon disposition). Other features that similarly incline the pit toward a consumption base are the half-inclusion and deferral of tax on realized capital gains, the $750,000 exemption on capital gains from business and farm assets, and the non-taxation of 2 Alarie takes an agnostic position on the benefits of shifting the personal tax base further toward consumption; he focuses on the relative merits of the two methods of implementing a consumption base. 3 I do not address in detail here the other two issues cited by Alarie--tax rate variation and timing of tax revenues--since they have been carefully assessed in previous studies: see Jonathan Kesselman and Finn Poschmann, A New Option for Retirement Savings: Tax-Prepaid Savings Plans, C.D. Howe Institute Commentary no. 149 (Toronto: C.D. Howe Institute, 2001); and Jonathan R. Kesselman and Finn Poschmann, "Expanding the Recognition of Personal Savings in the Canadian Tax System" (2001) vol. 49, no. 1 Canadian Tax Journal 40-101. Alarie attributes the foundation for Canada's TFSA to the arguments presented in those studies; also see Gordon Pape, Tax-Free Savings Accounts: A Guide to TFSAs and How They Can Make You Rich (Toronto: Penguin Canada, 2009), 11.
536 n canadian tax journal / revue fiscale canadienne
(2009) vol. 57, no 3
in-kind returns to ownership of consumer durables such as automobiles, appliances, and furniture. Only individuals at high earnings with savings that exceed the contribution limits to tax-deferred savings schemes and/or with large business assets face substantial tax on their capital incomes.4 The connection between income and consumption tax bases is straightforward. In the traditional definition (going back to Haig and Simons),5 income equals consumption plus savings; the portion of income not currently spent on consumption is savings. Hence, consumption can be measured as total income (from labour and capital sources) minus a deduction for savings. That relationship is the basis for taxdeferred savings schemes, which allow a deduction for savings up to a specified limit. Individuals who are not constrained by the contribution limit and undertake essentially all saving in tax-deferred schemes are thus taxed on their consumption.6 Similarly, exempting the returns to savings (interest, dividends, capital gains, and the returns on housing or consumer durables) works much like exempting tax on the savings that give rise to those returns;7 that is the basis for the tax-prepaid method of implementing a consumption-based personal tax. Thus, a consumption tax can be viewed as a tax on labour income that excludes capital income. The tfsa and the tax treatment of home ownership and consumer durable goods typify the tax-prepaid method; the deferral and half-taxation of capital gains on other assets also partially employ the tax-prepaid approach. The advantages of using consumption rather than income as the personal tax base are often encapsulated as promoting savings as a means for improved efficiency and growth.8 However, many supporters of consumption-based personal taxation
4 I ignore for now the many lower earners for whom the existing tax-deferred savings provisions are not an attractive option on account of low current marginal effective tax rates and anticipated higher METRs later in their lives; this is another group for whom a tax-prepaid savings option is attractive. 5 See R.M. Haig, "The Concept of Income--Economic and Legal Aspects," in R.A. Musgrave and C.S. Shoup, eds., Readings in the Economics of Taxation (Homewood, IL: Irwin, 1959); and Henry C. Simons, Personal Income Taxation: The Definition of Income as a Problem of Fiscal Policy (Chicago: University of Chicago Press, 1938). 6 With tax-deferred savings schemes, withdrawals from the account are added to the individual's taxable base, since funds are typically taken to finance contemporaneous consumption. 7 As explained later, the tax-deferred and tax-prepaid methods are fully equivalent only under simple assumptions; they may also differ in their effects on different generations of individuals during a transition period of shifting the tax base further toward consumption. See Lawrence H. Summers, "Capital Taxation and Accumulation in a Life Cycle Growth Model" (1981) vol. 71, no. 4 The American Economic Review 533-44; and Alan J. Auerbach and Laurence J. Kotlikoff, Dynamic Fiscal Policy (New York: Cambridge University Press, 1987). 8 Alarie's arguments against the tax-prepaid savings method centre on its allegedly weak saving incentives and why those incentives are weaker than with the tax-deferred method; he equates the impact on aggregate savings with the effects on economic efficiency and growth. For example, he begins one passage, "Assuming that the goal of the introduction of TFSAs is improving incentives to save. . . .": Benjamin Alarie, "Policy Forum: Assessing Tax-Free Savings Accounts--Promises and Pressures," in this issue (hereinafter cited as "Alarie"), at 510.
tax-free savings accounts in a consumption-based personal tax n 537 anticipate more multifaceted benefits.9 Even if shifting the personal tax base further toward consumption has only modest effects on aggregate savings, it could still promote economic efficiency and growth. Moreover, choice of the consumption base is strongly supported by horizontal equity over the lifetimes of individuals with different saving propensities. While moving toward a consumption base could compromise the vertical equity of the personal tax, this impact can be partially offset by appropriate changes in the tax rate schedule. Additional advantages of a consumption base stem from the simplifications of administration and compliance, and the reduced need to measure capital incomes. In this section, I examine all of these issues as well as reasons for not "going all the way" to a consumption base for personal taxes.10 Savings, Efficiency, and Growth The most common argument for a consumption tax base is that it would remove the distortion to savings by eliminating tax on capital incomes arising under an incomebased tax. That change would raise the after-tax return on capital, thus increasing incentives to save. The resulting increase in aggregate savings would raise the 9 Kesselman and Poschmann in A New Option for Retirement Savings and "Expanding the Recognition of Personal Savings," supra note 3, assign relatively little weight to the impact on aggregate savings in their analysis of the potential benefits from shifting the base toward consumption and, in particular, from favouring addition of the tax-prepaid method. They clearly acknowledge the mixed empirical evidence about the savings effects and proceed to elaborate the other benefits, such as improved efficiency and growth, horizontal equity, and operational simplicity vis-а-vis an income tax base. This multifaceted case for a consumption base goes back at least as far as the landmark report of the US Treasury--US Treasury Department, Blueprints for Basic Tax Reform (Washington, DC
: US Government Printing Office
, 1977) (the Bradford report)--and extends to George R. Zodrow, "Should Capital Income Be Subject to Consumption-Based Taxation?" in Henry J. Aaron, Leonard E. Burman, and C. Eugene Steuerle, eds., Taxing Capital Income (Washington, DC: Urban Institute, 2007), 49-81, at 67. 10 Large bodies of previous research have been devoted to the income-versus-consumption tax base issue. For samples of Economic Studies
, see Joseph A. Pechman, ed., What Should Be Taxed, Income or Expenditure? (Washington, DC: Brookings Institution, 1980); Auerbach and Kotlikoff, supra note 7; Henry J. Aaron and William G. Gale, eds., Economic Effects of Fundamental Tax Reform (Washington, DC: Brookings Institution, 1996); David F. Bradford, Fundamental Issues in Consumption Taxation (Washington, DC: AEI Press, 1996); David Altig, Alan J. Auerbach, Laurence J. Kotlikoff, Kent A. Smetters, and Jan Walliser, "Simulating Fundamental Tax Reform in the United States" (2001) vol. 91, no. 3 The American Economic Review 574-95; and Aaron, Burman, and Steuerle, supra note 9. For samples of legal studies, see Michael J. Graetz, "Implementing a Progressive Consumption Tax" (1979) vol. 92, no. 8 Harvard Law Review 1575-1661; Michael J. Graetz, 100 Million Unnecessary Returns: A Simple, Fair, and Competitive Tax Plan for the United States (New Haven, CN: Yale University Press, 2008); Edward J. McCaffery, "A New Understanding of Tax" (2005) vol. 103, no. 5 Michigan Law Review 807-938; Joseph Bankman and David A. Weisbach, "The Superiority of an Ideal Consumption Tax over an Ideal Income Tax" (2006) vol. 58, no. 5 Stanford Law Review 1413-56; Joseph Bankman and David Weisbach, "Reply: Consumption Taxation Is Still Superior to Income Taxation" (2007) vol. 60, no. 3 Stanford Law Review 789-802; and Daniel Shaviro, "Beyond the Pro-Consumption Tax Consensus" (2007) vol. 60, no. 3 Stanford Law Review 745-88. The last three studies show a remarkable overlap of legal and economic analysis.
538 n canadian tax journal / revue fiscale canadienne
(2009) vol. 57, no 3
economy's long-run capital stock, which in turn would enhance productivity and growth in both real wages and output. Yet the degree to which reduced taxation of capital incomes actually increases aggregate savings has been studied intensively with respect to existing tax-deferred (and to a much lesser extent tax-prepaid) savings schemes.11 Most estimates suggest that the net effect on savings, while positive, is not large. These findings are consistent with the two offsetting effects in standard economic analysis: a substitution effect whereby the increased net return to savings makes future consumption more attractive, thus raising savings; and an income effect whereby the increased income implicit in a tax decrease on capital income stimulates more current consumption, thus reducing savings.12 Alarie is properly skeptical about any large increase in savings from a further shift toward a consumption tax base, and he argues that the tax-prepaid method would be even less conducive to saving than the tax-deferred approach (analysis that I assess later).13 Yet substantial increases in aggregate savings are not critical in the case for shifting further toward a consumption base or for use of the tax-prepaid method, since those reforms yield important other benefits, as detailed below. Moreover, even if a further shift of the personal tax base toward consumption-- through either tax-deferred or tax-prepaid provisions--had zero impact on aggregate savings, such moves could still promote efficiency and growth of the economy.14 First, efficiency gains are based on the substitution effect alone, and typical estimates find a non-trivial substitution effect in saving responses to changes in net-of-tax rates of return.15 In other words, individuals will be able to allocate their lifetime economic
11 See Alarie and the studies cited in Kesselman and Poschmann, A New Option for Retirement Savings and "Expanding the Recognition of Personal Savings," supra note 3. 12 This issue is analogous to the impact of reduced tax rates on labour earnings, whereby a substitution effect increases desired work hours but is offset by an income effect that increases desired hours of leisure; efficiency loss arises even if the two effects are exactly offsetting with no change in labour supply. Put simply, the notion is that the same amount of tax revenue could be extracted from the individual by a non-distorting, lump-sum tax that left her at a higher utility level (with more leisure or more future consumption, respectively). 13 Note that Kesselman and Poschmann also were cautious about claims of large saving responses and did not rely on such claims in their analysis and findings support
ing the introduction of tax-prepaid savings schemes. 14 See Richard Kneller, Michael F. Bleaney, and Norman Gemmell, "Fiscal Policy and Growth: Evidence from OECD Countries" (1999) vol. 74, no. 2 Journal of Public Economics 171-90, finding cross-country empirical evidence that income taxes reduce ECONOMIC GROWTH
significantly, whereas consumption-based taxes do not retard growth. 15 For example, assume a 5 percent annual pre-tax real rate of return, a 40 percent marginal tax rate, and an individual contemplating saving now for retirement in 40 years. Without tax, each unit of consumption forgone today will yield just over 7 units of consumption at retirement. With an annual tax on capital income (and assuming no inflation), the unit forgone today would yield just 3.25 units at retirement, implying an effective tax rate of 54 percent on the choice to save for future consumption. If inflation is just 2 percent annually, this raises the effective tax rate on future consumption to 66 percent, so that future consumption (the return to savings) becomes nearly three times as costly as it would be without the tax.
tax-free savings accounts in a consumption-based personal tax n 539 resources over the various stages of their lives more efficiently, thus raising their lifetime utility levels (the ultimate gauge of efficiency). Second, the existing pit distortions across types of savings, most particularly favouring home ownership over savings invested in productive business capital (either directly or through financial markets), would be moderated by a shift toward a consumption tax base. The economy's aggregate savings would be more efficiently allocated, with less tax-induced overinvestment in housing and more investment in business capital, thus raising productivity and economic growth. Third, reduced taxation of capital incomes would attenuate flows of domestic savings into unreported foreign capital holdings, retaining more capital for the benefit of the domestic economy. Horizontal Equity Horizontal equity is the notion that individuals with similar characteristics and economic resources should pay similar taxes. But income and consumption tax bases can yield very different outcomes for otherwise identical individuals who differ only in their saving rates and consumption patterns.16 Consider two workers toiling at identical jobs throughout their lives, both earning the same salaries each year. "Spender" spends all of every paycheque by the next payday, whereas "Saver" saves a part of each paycheque. Spender therefore accumulates no savings, never receives any capital income, and enters retirement with no assets. In contrast, Saver earns capital income that grows every year and enters retirement with substantial assets. If one regards the two individuals' identical lifetime labour earnings and opportunities to consume as making them similar in ability to pay, then horizontal equity requires that they bear the same total tax burdens over their lives. On this view, horizontal equity would be satisfied by a consumption-based tax but not by an income-based tax. The tax base could be either labour earnings in each year (the tax-prepaid method) or actual consumption in each year (the tax-deferred method). Under the latter method, Saver pays less tax than Spender during their working years but correspondingly more taxes (compounded with interest) during retirement. Alternatively, if one believes that individuals' tax-paying abilities should be judged on an annual rather than a lifetime basis, then an income-based tax may be deemed more horizontally equitable. In this case, Saver is seen as having greater ability to pay taxes each year because Saver receives growing sums of capital income, in addition to annual labour earnings that are the same as Spender's. What this view ignores is that Saver's additional consumption is enjoyed later in life than Spender's consumption. This added consumption simply reflects the return on savings; society has the benefit of those resources for the intervening years via additional productive capital, 16 A similar example of the differential income tax burdens on three brothers, each with an identical inheritance but ranging from thrifty to spendthrift, was cited by Irving Fisher, the pioneering advocate of consumption-based taxation: Irving Fisher, The Nature of Capital and Income (New York: Macmillan, 1906), 250-53. See Simons, supra note 5, for the classic statement favouring income as the personal tax base; and Pechman, supra note 10, for more recent arguments on the two sides.
540 n canadian tax journal / revue fiscale canadienne
(2009) vol. 57, no 3
which boosts output, jobs, and real wages. On a discounted basis, Spender's and Saver's lifetime total consumption levels are equal under a consumption-based tax, while an income-based tax reduces Saver's lifetime total consumption below that of Spender. An income-based tax penalizes individuals for deferring their consumption through saving, whereas a consumption-based tax provides neutral incentives for people to time their spending as they prefer.
Administrative and Compliance Simplicity One of the most difficult aspects of implementing an income-based tax pertains to the measurement of capital incomes.17 Many forms of return to capital are not readily measured on an annual basis. For example, most capital gains are taxed only on realization, when the asset is sold, since annual valuations for some kinds of assets would be problematic. Tax provisions attempt to prevent some kinds of tax deferral, such as requiring accrual accounting for returns on strip bonds. Other tax provisions explicitly permit deferral in taxing the asset gains while allowing immediate claims for related expenses, such as flowthrough shares. A great deal of tax planning for higher-wealth individuals is oriented around ways of deferring tax on investment and business returns and of achieving returns as preferentially taxed capital gains. Moreover, difficulties in monitoring cost basis for some assets result in extensive evasion of tax on capital gains.18 In contrast, a tax base of consumption--whether implemented by the tax-deferred or the tax-prepaid method--avoids almost all of these difficulties, since capital income is excluded from the base. This carries major gains in simplicity for both tax administration and taxpayer compliance.19 Additionally, elimin ating the associated capital market distortions augments economic efficiency.
Vertical Equity One common concern about shifting the pit base further toward consumption is that it would reduce the effective progressivity of the tax system and thus compromise vertical equity. Because individuals act to buffer income variations (both year-to-year and lifetime) by saving, dissaving, and borrowing to smooth their consumption, any given progressive rate schedule would be inherently less progressive with a consumption base than with an income base. In the first instance, this matter can be
17 My discussion here focuses on problems related to measuring capital income on an accrual basis, but additional problems arise with respect to measuring capital income in real terms. Neither problem arises with a full consumption base. 18 See James M. Poterba, "Tax Evasion and Capital Gains Taxation" (1987) vol. 77, no. 2 The American Economic Review 234-39; and Joseph M. Dodge and Jay A. Soled, "Inflated Tax Basis and the Quarter-Trillion-Dollar Revenue Question" (2005) vol. 106, no. 4 Tax Notes 453-62. 19 For example, equity investments within an RRSP do not require any tracking of cost bases or sale proceeds on individual asset transactions; only amounts contributed to and withdrawn from the account need to be tracked. This also eliminates any opportunities for false or merely erroneous reporting of cost bases and erroneous or non-reporting of sale proceeds, which can be achieved with non-registered assets.
tax-free savings accounts in a consumption-based personal tax n 541 addressed by increasing the progressivity of the rate schedule when shifting the tax base further toward consumption.20 For any income class, the average tax burden would then be maintained, though of course below-average savers in that class would bear more of the burden and above-average savers would bear less. Still, one must remember that common measures of tax progressivity use an annual time frame; more consistent with a consumption perspective of individuals' well-being would be a lifetime measure of progressivity.21 Even using an annual time frame, adjusting the tax rate schedule does not resolve all issues of equity in shifting to a more consumption-based personal tax, since each income class contains both younger and older cohorts. Those near the end of their working lives or in retirement have little or no labour income and a relatively large share of capital income (much of it in tax-deferred savings schemes). Thus, shifting toward consumption using the tax-deferred method would treat older cohorts relatively harshly compared with younger cohorts,22 and conversely for a shift using the tax-prepaid method.23 Eventually, this issue will vanish, as more cohorts are born and live their entire lives under the more consumption-based tax regime, but the transition period is very long. Moreover, for very high-wealth individuals currently holding most of their assets in non-registered forms (outside rrsps and rpps), an unconstrained move to a consumption base would convey very large windfall tax savings. Since high-wealth individuals now pay a disproportionately large share of total pit revenues, that raises important issues about how and how far the personal tax system should be shifted toward consumption. Why Not Go All the Way? As just noted, an unconstrained immediate move to a full consumption tax base would entail a sharp reduction in both vertical equity and tax revenues. The increase in upper-bracket tax rates needed to neutralize these impacts might be excessive and induce new forms of tax avoidance and evasion. These higher rates would also place much higher burdens on labour earnings at higher incomes and thus create substantial intergenerational shifts in the tax burden for many years after the reform. Such a change would pose little incentive to incremental saving relative to the incentives for large wealth-holders simply to shift their non-registered assets into 20 This change in the rate schedule increases the marginal tax rates on labour income somewhat, but the associated base change reduces the tax rate on capital income more sharply (to zero). Since tax distortions cause greater inefficiencies in capital markets than in labour market
s, economic efficiency registers a net gain. 21 See James B. Davies, France St-Hilaire, and John Whalley, "Some Calculations of Lifetime Tax Incidence" (1984) vol. 74, no. 4 The American Economic Review 633-49. 22 That is, older cohorts would face higher tax rates on their withdrawals and disbursements out of tax-deferred accounts, and they would have little opportunity to contribute further to those accounts. 23 That is, older cohorts would have more non-registered assets to be shifted from taxable status to tax-free status.
542 n canadian tax journal / revue fiscale canadienne
(2009) vol. 57, no 3
tax-recognized accounts (either tax-deferred or tax-prepaid). With unlimited access to tax-prepaid accounts, all revenues currently collected from capital incomes on non-registered assets would vanish. With unlimited access to tax-deferred accounts, all of those revenues would similarly disappear, and additionally the tax deductions for the new contributions would wipe out the revenues currently collected on the labour earnings of higher-wealth individuals. In order to control the cited revenue and distributional impacts, moves of the personal tax toward a more consumption-based system need to be limited. A primary consideration is to prevent persons with large pre-existing non-registered wealth holdings from shifting their assets into non-taxable forms. Any of three methods can be employed: (1) imposing an annual dollar ceiling on contributions to tax-recognized savings (as in tfsas); (2) imposing an annual contribution limit related to the individual's annual labour earnings; and (3) imposing an annual contribution limit that is both a given percentage of annual labour earnings and also constrained to an annual dollar maximum (as in rrsps and rpps). Each of these forms of limitation means that tax-recognized savings will be related to the age and/or earnings of the individual, reflecting the life-cycle nature of most savings. Applying an annual dollar ceiling on contributions (with or without a linkage to earnings) also limits the shift to a full consumption tax base to individuals with earnings below a specified level, so that the wealthy would still operate in part under an income-based system. The development of tax policy proposals for a personal consumption-based tax reveals increasing concern over the issues described above. An early and well-known proposal was the flat tax of Robert Hall and Alvin Rabushka, which embodied both tax-prepaid and tax-deferred methods without limits at the individual level and packaged them with a cash flow tax at the corporate level.24 The extreme distributional impacts of the Hall-Rabushka flat tax, exacerbated by collapsing the tax schedule to a single rate, have been well documented.25 Official us studies of reforming the personal tax by shifting toward consumption began with a 1977 Treasury study (the Bradford report) and led to the more recent report of the President's Advisory Panel on Federal Tax Reform in 2005.26 Parallel influential research has been
24 Robert E. Hall and Alvin Rabushka, Low Tax, Simple Tax, Flat Tax (New York: McGraw-Hill, 1983). 25 For example, see Robert Eisner, "The Proposed Sales and Wages Tax: Fair, Flat, or Foolish?" in Robert E. Hall, Alvin Rabushka, and Dick Armey, Fairness and Efficiency in the Flat Tax (Washington, DC: AEI Press, 1996), 42-95. Also see David F. Bradford, The X Tax in the World Economy: Going Global with a Simple, Progressive Tax (Washington, DC: AEI Press, 2004), proposing a variant of the Hall-Rabushka tax (the "X" tax) that retained marginal rate progressivity on labour earnings and taxed cash flows from capital at the top personal marginal rate. 26 US Treasury Department, supra note 9; and United States, President's Advisory Panel on Federal Tax Reform, Simple, Fair, and Pro-Growth: Proposals To Fix America's Tax System, Report of the President's Advisory Panel on Federal Tax Reform (Washington, DC: US Government Printing Office, 2005). While the panel's two alternative reform proposals (the simplified income tax and the growth and investment tax) both moved further toward a consumption base, neither completely abandoned capital income taxation.
tax-free savings accounts in a consumption-based personal tax n 543 undertaken in Britain by the Institute for Fiscal Studies
, resulting in the Meade report in 1978 and the forthcoming Mirrlees report.27 Counterpart Canadian tax policy research was produced by the Economic Council of Canada in 1987 and more recent studies by academic tax economists.28 These studies typically conclude that movement toward a consumption-based personal tax is desirable, but with clear limits on this move for higher earners or wealth-holders. One example is the Meade report's support for a personal consumption tax base combined with a form of wealth taxation. Another is the proposal of the tax-law academic Michael Graetz for consumption taxation via a value-added tax (similar to Canada's goods and services tax [gst]), with an income tax retained only for households with incomes above us$100,000.29 Other hybrid reform proposals include recommendations by the us president's advisory panel and the British Mirrlees report tilting toward consumption-based taxation but retaining elements of income taxation for higher earners. Ta x- D e f e r r e d V e r s u s Ta xP r e pai d M e t h o d s I next offer a more detailed analysis of the distinctions between the tax-deferred and tax-prepaid methods of implementing a consumption-based personal tax.30 Alarie's essential argument is that the tax-deferred method is superior and that some key features of the tfsa could equally well have been implemented within the 27 J.E. Meade, chair, The Structure and Reform of Direct Taxation, Report of a Committee of the Institute for Fiscal Studies (London: Allen & Unwin, 1978); and Institute for Fiscal Studies and James Mirrlees, eds., Tax by Design: The Mirrlees Review (Oxford: Oxford University Press) (forthcoming). The key underlying research on the tax base for the Mirrlees report is a study by James Banks and Peter Diamond, "The Base for Direct Taxation," in Institute for Fiscal Studies and James Mirrlees, eds., Dimensions of Tax Design: The Mirrlees Review (Oxford: Oxford University Press) (forthcoming), which finds that some taxation of capital income is warranted in an optimal tax framework. 28 See Economic Council of Canada, The Taxation of Savings and Investment (Ottawa: Supply and Services, 1987), and the companion study by James B. Davies and France St-Hilaire, Reforming Capital Income Taxation in Canada: Efficiency and Distributional Effects of Alternative Options (Ottawa: Economic Council of Canada, 1987). Also see the following, more recent, studies supporting movement toward a consumption tax base: Jonathan R. Kesselman, "Base Reforms and Rate Cuts for a Revitalized Personal Tax" (1999) vol. 47, no. 2 Canadian Tax Journal 210-41; Jonathan R. Kesselman, "Tax Design for a Northern Tiger" (2004) vol. 10, no. 1 IRPP Choices 1-46; Bev Dahlby, "Restructuring the Canadian Tax System by Changing the Mix of Direct and Indirect Taxes," in Herbert G. Grubel, ed., Tax Reform in Canada: Our Path to Greater Prosperity (Vancouver, BC: Fraser Institute, 2003), 77-108; and Jack M. Mintz, "Taxation with the Least Pain--A New Tax Structure for Canada," in Grubel, ibid., 41-51. 29 Graetz, 100 Million Unnecessary Returns, supra note 10. 30 For earlier coverage of some of the same ground, see Graetz, "Implementing a Progressive Consumption Tax," supra note 10; Kesselman, "Tax Design for a Northern Tiger," supra note 28; Kesselman and Poschmann, A New Option for Retirement Savings and "Expanding the Recognition of Personal Savings," supra note 3; and McCaffery, supra note 10. Also see Alarie's analysis (in particular, the section "Assessing the TFSA Reform").
544 n canadian tax journal / revue fiscale canadienne
(2009) vol. 57, no 3
pre-existing tax-deferred schemes. Hence, he argues that the tfsa should not be assessed on the basis of features that could have been instituted within rrsps and rpps without the need to launch a new savings scheme. For these reasons, my analysis begins with discussion of generic differentiating characteristics of the two consumption-base methods and then addresses specific design features of the rrsp and tfsa schemes. I focus on the rrsp as a parallel to the tfsa, since contributions to both schemes are elective by the individual; however, employer-based rpps share most of the rrsp design features.
Characteristics of the Two Methods Table 1 summarizes key differentiating characteristics of the tax-deferred and taxprepaid methods. Both methods achieve the kind of horizontal equity described earlier--equal tax treatment of individuals with the same level and timing of labour earnings and thus equal opportunities to consume over their lives. Another kind of horizontal equity arises for individuals who undertake the same level of saving at the same time but achieve different rates of return. One version of horizontal equity for this situation is ex post equity, where those with equal realized returns from savings bear equal taxes and where those with differing rates of realized returns bear corres pondingly different taxes. Ex post equity is achieved by the tax-deferred method, which imposes tax on supernormal returns but implicitly subsidizes savers with subnormal returns;31 as with any consumption-based tax, the normal rate of return to capital is untaxed.32 The other version of horizontal equity is ex ante equity, where two individuals with the same savings and thus the same opportunities to invest are taxed identically regardless of their actual realized rates of return. Ex ante equity is achieved by the tax-prepaid method, which taxes neither normal nor supernormal returns, but also does not provide a subsidy for subnormal returns. These equity notions are explored in a later section addressing Alarie's concern over taxation of supernormal returns. Two characteristics of the tax-deferred and tax-prepaid methods are best considered jointly: vertical equity and lifetime tax averaging. With an unchanged tax rate schedule, shifting toward a consumption base under the tax-deferred method reduces the effective progressivity of taxes on labour earnings, viewed in an annual framework. Individuals choose to make their tax-deductible contributions to taxdeferred savings accounts in years when they face higher marginal tax rates and to make their taxable withdrawals in years of lower marginal tax rates. Those actions reduce the effective progressivity of personal taxes on labour earnings, but they simultaneously facilitate lifetime averaging of taxes. Since a consumption tax base is
31 This implicit subsidy is explained by the upfront tax deduction having a present value greater than the final tax on the withdrawal, since the discount rate exceeds the rate of investment return. 32 However, as explained below, even the normal rate of return will bear tax under the taxdeferred method for savers facing higher marginal tax rates in the future.
tax-free savings accounts in a consumption-based personal tax n 545
Table 1 Characteristics of Tax-Deferred and Tax-Prepaid Methods
Allows deduction for net savings, taxes net dissavings
Horizontal equity (differential returns)
Ex post equity: supernormal returns taxed, subnormal returns subsidized, normal returns not taxed
Taxes labour income, exempts capital income
Ex ante equity: neither normal nor supernormal returns taxed, subnormal returns not subsidized
Vertical equity (with unchanged rate schedule)
Reduces effective progressivity of annual tax rate schedule on labour earnings
Maintains effective progressivity of annual tax rate schedule on labour earnings
Lifetime tax averaging Facilitated
Advantageous for those who expect to have a lower MTR* at time of withdrawal
Unfavourable for those who expect to have a higher MTR at time of withdrawal
Advantageous for those who expect to have a higher MTR at time of withdrawal
Unfavourable for those who expect to have a lower MTR at time of withdrawal
Disrupted when MTRs differ across periods for individual
Maintained regardless of MTR patterns
Simplicity for taxpayer
Adds complexity: need to anticipate future MTRs; discourages small deposits and withdrawals
Simpler method: future MTRs irrelevant; encourages flexible usage
Increases future tax revenues, matching revenues to needs
High near-term revenue cost (relative to income base or taxprepaid)
Little near-term revenue cost (relative to income base or taxdeferred)
Reduces future tax revenues, poorer matching to needs
* MTR = marginal tax rate. Source: Extended and adapted from Jonathan R. Kesselman, "Tax Design for a Northern Tiger" (2004) vol. 10, no. 1 IRPP Choices 1-46, at 13.
546 n canadian tax journal / revue fiscale canadienne
(2009) vol. 57, no 3
associated with a lifetime view of individual well-being, this kind of averaging can be regarded favourably; if the resultant incidence of lifetime tax burdens is deemed insufficiently progressive, the remedy is to steepen the rate schedule.33 The taxprepaid method maintains effective progressivity of annual taxes on labour earnings, since no upfront tax deduction is allowed for savings, but it does not assist lifetime tax averaging. Even if one prefers using the tax-prepaid method for other reasons, combining it with the tax-deferred method to facilitate lifetime tax averaging could be desirable. Another three characteristics are also best considered together: the individual taxpayer's position, economic efficiency, and taxpayer simplicity. For taxpayers who expect to be facing a lower marginal effective tax rate (metr--including the impacts of income-tested benefit provisions and incremental taxes) at the time of withdrawing funds than their metr at the time of initial saving, the tax-deferred method is favourable. In that case, the tax-deferred method presents a subsidy to savings even when the asset yields only a normal rate of return. Relative to the tax-deferred method, the tax-prepaid method is unfavourable to the individual in that case. Conversely, the tax-deferred method is relatively adverse and the tax-prepaid method beneficial for individuals who anticipate a higher metr at the time of withdrawal than at the point of initial saving. In that case, the tax-deferred method imposes a tax burden on even the normal return to capital, contrary to the precepts of a consumptionbased tax. The variation in metrs between the point of saving and withdrawal of funds for consumption has differing impacts on economic efficiency and taxpayer simplicity under the two methods. The principal distortions of capital income taxes are to bias the timing of individuals' consumption choices and to place a tax "wedge" between the returns generated by investments in the real economy and the returns received by suppliers of savings. Eliminating the taxation of capital income removes these distortions. metrs that vary over time for individual savers prevent a consumption tax implemented by the tax-deferred method from eliminating these distortions; for some individuals, the normal return to savings remains taxed, while for others it is subsidized. In contrast, the tax-prepaid method leaves the full return to savings in the hands of the savers, thus eliminating all distortions and restoring full economic efficiency regardless of variations in individual metrs. That aspect also means that individuals do not have to make forecasts about their future metr in deciding whether and when to contribute to a tax-prepaid savings account, since such contributions are always attractive for anyone who is taxable. In contrast, with tax-deferred savings, individuals need to make such forecasts, with only taxpayers currently in the top
33 Alternatively, one could consider a system of lifetime taxation in which the taxes paid by an individual in any given year also hinged on the individual's income or consumption levels in all previous years: see William S. Vickrey, Agenda for Progressive Taxation (New York: Ronald Press, 1947). Also see Shaviro, supra note 10, for a discussion of similarities between lifetime taxation of income and annual taxation of consumption.
tax-free savings accounts in a consumption-based personal tax n 547 marginal brackets having a no-brainer decision about contributing (since they will at worst be at the same tax rate when withdrawing and bearing tax).34 Impacts of the two alternative methods on the timing of individual tax payments are of interest to the treasury, which may not be indifferent between receiving funds in the current period or in future years. From the treasury's position, revenue collections on savings are delayed under the tax-deferred method and, comparatively, accelerated under the tax-prepaid method. Therefore, the tax-deferred method has a relatively high current revenue cost versus the tax-prepaid method's minimal in itial but growing revenue cost. In discounted terms, total tax revenues generated under the two methods should be roughly equivalent.35 Alarie deems the differential timing of tax revenues to weigh in favour of the tax-deferred method on political economy grounds; I address this issue later. Clearly, each of the two methods has areas of superiority over the other, with some judgmental aspects in the rankings. But the two methods can work well concurrently to reinforce the effective operation of a consumption-based personal tax. If all individuals could fully anticipate their lifetime pattern of earnings and could borrow and invest without constraints, then most would smooth their consumption patterns to roughly equalize their metrs across years.36 Equalized metrs would, in turn, restore economic efficiency to the tax-deferred method. Yet, lack of perfect foresight about future earnings and investment returns as well as borrowing constraints make this outcome more of a textbook example than a reality for most people. Typically, young adults
begin their working lives at relatively low earnings, so that contributing to a tax-deferred plan while facing a low metr does not make sense. Providing a tax-prepaid option allows people to begin saving early and enjoy tax-free compounding of investment returns; then, when they enter higher-earning years with higher metrs, they can shift the tax-prepaid savings into a tax-deferred account and enjoy the associated tax savings. Thus, tax-prepaid and tax-deferred savings act as complementary methods for implementing a consumption-based tax--a point that has long been recognized by tax analysts.37 34 Conceivably, some top-tax-bracket individuals could face still higher METRs in retirement owing to public pension benefit clawbacks, but few of them are likely to have sufficiently low lifetime savings to be eligible for such public benefits. 35 The tax-deferred method's tax-averaging feature should push its revenues below those of the tax-prepaid method, but this should be roughly offset by the progressivity of the tax rate schedule combined with the dispersion of investment rates of return and thus the varying size of taxable withdrawals. 36 Of course, typical life-cycle patterns of family formation, with children being raised and then leaving the family, and subsequent retirement needs would lead to a consumption pattern that was not fully uniform. 37 For example, see US Treasury Department, supra note 9; Meade, supra note 27; Peter Mieszkowski, "The Advisability and Feasibility of an Expenditure Tax System," in Henry J. Aaron and Michael J. Boskin, eds., The Economics of Taxation (Washington, DC: Brookings Institution, 1980), 179-201; Michael J. Daly, "The Role of Registered Retirement Savings
548 n canadian tax journal / revue fiscale canadienne
(2009) vol. 57, no 3
RRSP Versus TFSA Design Features Table 2 summarizes key design features of the rrsp and the tfsa. While these schemes typify Canada's current provisions for tax-deferred and tax-prepaid savings, respectively,38 most of the specific design features are not inherent to either the taxdeferred or the tax-prepaid method. As Alarie notes, some design features of the tfsa could alternatively have been implemented within the rrsp. Still, the key differentiating design features of the two methods (deductibility or not of contributions and taxability or not of withdrawals) are clearly not interchangeable across the two consumption-base methods (without interchanging their names!). The discussion above shows that those differences are key to the two methods' differential characteristics, particularly the tax-prepaid method's superior vertical equity, economic efficiency, and taxpayer simplicity. Hence, even introducing many of the particular features of the tfsa into the rrsp--such as the ability to recontribute withdrawn funds or the exemption of withdrawals from benefit clawbacks--would not undercut the multiple reasons for introducing a tax-prepaid scheme to shift the personal tax system further toward a consumption base. Later I consider whether some of the design features of the tfsa might usefully be introduced into existing tax-deferred savings schemes as well.
Supernormal Investment Returns A recurrent issue raised in Alarie's critique of the tax-prepaid method is its failure to tax supernormal investment returns. This alleged deficiency, which other tax analysts have also cited,39 includes two distinct dimensions. First is the failure to tax individuals receiving supernormal returns at suitably high rates--that is, taxing them
Plans in a Life-Cycle Model" (1981) vol. 14, no. 3 Canadian Journal of Economics 409-21; and Davies and St-Hilaire, supra note 28. Also see Michael J. Daly and Fadle Naqib, "Designing a Non-Distortionary Personal Tax System for Canada" (1985) vol. 18, no. 2-3 Economics Letters 209-12, where the authors conclude (at 211), "[O]ur model shows that [a tax-prepaid] . . . option is a prerequisite for a non-distortionary graduated personal tax system, and that individuals should, as far as possible, be allowed to decide for themselves the degree to which their assets are prepaid or not." 38 One interesting difference is that the TFSA opens up income-splitting opportunities for individuals who have no RRSP contribution room and therefore no chance to split income by using spousal RRSPs. The TFSA thus improves horizontal equity, according to my earlier analysis of investment income splitting: see Jonathan R. Kesselman, "Income Splitting and Joint Taxation of Couples: What's Fair?" (2008) vol. 14, no. 1 IRPP Choices 1-54. Also see Pape, supra note 3, at 75-79, for practical advice on how to use TFSAs for interspousal income splitting. 39 As early as the Bradford report, which endorsed the joint use of tax-deferred and tax-prepaid methods, it was suggested that all "speculative" investments that could generate supernormal returns be restricted to tax-deferred accounts and that tax-prepaid accounts be restricted to the holding of fixed-income assets: US Treasury Department, supra note 9, at 129. For a similar view, see Mieszkowski, supra note 37, at 194. Concern over the failure of tax-prepaid accounts to strike supernormal returns has been a recurrent theme in legal (but not economic) analyses; see, for example, Graetz, "Implementing a Progressive Consumption Tax," supra note 10, and McCaffery, supra note 10.
Table 2 Comparison of Design Features, Registered Retirement Saving Plans (RRSPs) and Tax-Free Savings Accounts (TFSAs)
RRSP (tax-deferred method)
TFSA (tax-prepaid method)
tax-free savings accounts in a consumption-based personal tax n 549
Contributions Deductibility . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Amount linked to earnings . . . . . . . . . . . . . . . . . . . . . . . . Amount linked to other schemes . . . . . . . . . . . . . . . . . . . Ceiling amounts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Ceiling amounts indexed . . . . . . . . . . . . . . . . . . . . . . . . . Eligible source(s) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Carryforward of unused room . . . . . . . . . . . . . . . . . . . . . Upper age limit . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Spousal contributions . . . . . . . . . . . . . . . . . . . . . . . . . . . . Contributions for non-spouse . . . . . . . . . . . . . . . . . . . . . Excess contributions penalized . . . . . . . . . . . . . . . . . . . . Investment returns . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Withdrawals Taxability . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Affect federal benefit clawbacks . . . . . . . . . . . . . . . . . . . . Affect provincial benefit clawbacks . . . . . . . . . . . . . . . . . Can be recontributed . . . . . . . . . . . . . . . . . . . . . . . . . . . . Mandatory at specified age . . . . . . . . . . . . . . . . . . . . . . . . Attribution on early withdrawal of spousal contribution . . . . . . . . . . . . . . . . . . . . . . . . . . . Accounts and balances Holding of small-business shares . . . . . . . . . . . . . . . . . . . Ineligible assets penalized . . . . . . . . . . . . . . . . . . . . . . . . Usable as loan collateral . . . . . . . . . . . . . . . . . . . . . . . . . . Creditor-protected . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Can be self-directed . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Can be offered by employers . . . . . . . . . . . . . . . . . . . . . .
Deductible (not taxed)* Yes (18% to ceiling amount) Yes (linked to RPP) $21,000 (2009); $22,000 (2010) To average wage (from 2011) Earned income only Yes Yes (to age 71) Part of contributor's limit Prohibited Yes Tax-deferred* Taxable* Yes Yes Not allowed Yes Yes (if within 2 calendar years following year of contribution) Yes, but limited for specified shareholders Yes (above $2,000) No Yes Yes Yes
Non-deductible (taxed)* No No $5,000 (2009) To CPI in $500 increments No restrictions Yes No Additional to contributor's limit Allowed for age 18-plus Yes Non-taxable* Non-taxable* No To be determined Allowed (in following year) No No Yes, but tighter restrictions than for RRSPs Yes Yes No Yes Yes
* Indicates design features that are inherent to this method of implementing a consumption-based tax; all other design features are elective, and some could equally well be applied to the alternative method.
550 n canadian tax journal / revue fiscale canadienne
(2009) vol. 57, no 3
on their expected returns rather than their realized returns--thus possibly compromising some notion of horizontal equity. Second is the asserted cost to total tax revenues from not tapping these supernormal returns. Because the tax-deferred method taxes all withdrawals from registered accounts, it captures the cumulative supernormal returns and thus, presumably, avoids both deficiencies. However, once one recognizes the parallel presence of subnormal investment returns and investigates the sources of supernormal returns, these critiques of the tax-prepaid method lose their force.
Distribution of Taxes and Equity Concept The very notion of a "supernormal" return on investments derives from the variability of these returns. If all investment returns are not uniform at a certainty rate, then various investment choices must generate a distribution of ex post returns. A "normal" rate of return will be simply the average return, with rates below that average labelled "subnormal" and those above average labelled "supernormal."40 As Alarie states, individuals could achieve supernormal returns through above-average investing skill, effort, and/or good luck. It follows that subnormal returns could also arise through below-average skill, effort, and/or misfortune. In assessing the "inequity" charge against the tax-prepaid method, it is useful to distinguish among the various factors that can affect individual investment outcomes and yield supernormal returns: n enjoying pure luck; n engaging in high-risk investment strategies; n exercising skill and effort (a labour-type activity); and n exploiting pure economic rents. At any given time, some individuals will be the beneficiaries of pure luck in achieving high rates of return on their investments. These high returns are the joint product of good fortune and high-risk investment strategies, since low-risk investments have little dispersion of returns and therefore cannot generate even the occasional high return.41 Here I consider only the combination of luck and high-risk strategies without any special investor skill or effort, since the latter factor warrants separate analysis. Alarie asserts repeatedly that allowing high returns resulting from pure luck to go untaxed is inequitable. He states that "[a] lucky investor is just not the same as an unlucky investor ex post, even if at one point in the past they were in the same position,"42 and he quotes Graetz:
40 I ignore the finer points of whether the appropriate average should be computed as the arithmetic mean or the geometric mean and whether the rates of return on various investments should be risk-adjusted. 41 Moreover, the high-risk investment strategies that occasionally yield very high returns almost invariably yield much lower or subnormal returns in the future, since they typically depend on high leverage and/or great concentration of assets by type, industry, or company. 42 Alarie, at 518.
tax-free savings accounts in a consumption-based personal tax n 551 Horizontal equity, the most widely accepted notion of fairness in taxation, requires that persons in similar circumstances pay similar amounts of tax. . . . [T]he notion that similar circumstances should be evaluated ex ante in present value terms seems quite a radical departure. . . . [I]t seems clear that horizontal equity must be an ex post concept.43 However, rejecting ex ante measures of horizontal equity is little more than a personal value statement and not a logically derived proposition. That position is strongly contradicted by longstanding Canadian tax provisions that have wide public acceptance--namely, non-taxation of winnings from legal gambling, lotteries, capital gains on homes, and receipt of inheritances and gifts. The prospective highend returns from risky investments on the $5,000 of annual tfsa contributions pale beside winning a $50 million lottery jackpot on a $5 ticket or profiting by hundreds of thousands of dollars, or more, on the sale of the family home. Alarie also expresses concern that a tax-prepaid consumption tax method will allow individuals to "disguis[e] returns to labour as returns to capital,"44 thereby getting tax-free treatment on the part of a supernormal return that is attributable to labour efforts rather than capital. He focuses particularly on individuals who purportedly have unusual skills in finding high-return investment opportunities. One can dispense with concerns about skills and efforts devoted to business enterprises to create high measured returns on capital, part of which is actually a return to labour; tfsas permit holding of small-business shares only under highly restrictive conditions.45 As for individuals with unusual acumen in financial investments, evidence from professionally managed funds is particularly telling. Many studies find that actively managed mutual funds do not systematically outperform passive or index funds; even the most experienced investment pros do not on average add meaningful value relative to simply buying a broad index of equities.46 Moreover, the few funds with strong long-term outperformance have suffered mightily in the recent market collapse; this 43 Graetz, "Implementing a Progressive Consumption Tax," supra note 10, at 1600-1. 44 Alarie, at 517. 45 As Alarie concedes, "[t]o the extent that interests in closely held corporations are not eligible investments, the scope to earn alpha or supernormal returns in TFSAs is substantially attenuated" (Alarie, at 528). However, RRSPs permit holdings of small-business corporate shares on less restrictive terms than those for TFSAs. On this point and for a description of TFSA penalty taxes, see Ken Griffin and Louis Provenzano, "TFSA Penalties and Tax" (2009) vol. 17, no. 1 Canadian Tax Highlights 6-7. 46 Extensive research by financial economists rejects the hypothesis that active management can add net value ("alpha") relative to purchasing a broad index of equities either directly or through a passive fund, index fund, or exchange-traded fund; and it cannot reject the hypothesis that the few investment managers who significantly outperform the market have done so by luck rather than skill. See Eugene F. Fama and Kenneth R. French, "Luck Versus Skill in the Cross Section of Mutual Fund Alpha Estimates" Journal of Finance
(forthcoming); Laurent Barras, Olivier Scaillet, and Russ Wermers, "False Discoveries in Mutual Fund Performance: Measuring Luck in Estimated Alphas" Journal of Finance (forthcoming), and sources cited therein.
552 n canadian tax journal / revue fiscale canadienne
(2009) vol. 57, no 3
reflects the maxim that "past performance is no guarantee of future performance."47 For an individual investor to capitalize on outperforming funds requires foresight, not simply identifying the winners after the fact. It is unproven whether many individuals have such skills; often outcomes that are credited as skill are later seen in their truer light as luck--and therefore not replicable. Possibly some investors can outperform both the market and fund managers in small-cap and micro-cap equities, because those firms are underfollowed by analysts and do not provide sufficient liquidity for funds to handle. Yet, even if a few investors can achieve systematic high returns, this needs to be considered in the context of the $5,000 annual limit on tfsa contributions and relative to the lifetime capital gains exclusion of $750,000 available to incorporated small-business and farm owners. Finally, investments that capture the economic rents stemming from monopolistic types of activities can generate high returns. But before concluding that some individuals can systematically exploit these rents and receive them free of tax in a tax-prepaid account, further analysis is required. In particular, it is necessary to distinguish between rent-generating assets held in a publicly traded firm versus those held in a private firm or individually. If a publicly traded firm develops a patent, industrial process, trademark, resource property, business location, or another asset that generates monopoly-type profits, its share price will be rapidly bid up to the point that any further investors earn no more than a normal return on their shares. Only early investors in the shares will reap a supernormal return, and that only for the limited period while the share prices rise to reflect the monopoly returns. Hence, this situation requires the special skill and effort assessed previously, which found that few individuals, even among investment professionals, possess these abilities in an enduring way. If the economic rent-generating activity is privately held, such as a patent or business, one must inquire how that entity arose, how competitive was the process for acquiring or creating it, and what were the associated costs. Clearly, labour is a major contributor to such privately held high-return assets, including private businesses, but such asset holdings are excluded from tax-prepaid savings plans like the tfsa.
Impact on Aggregate Tax Collections Alarie asserts that "the government is better off with a [tax-deferred] account . . . since it offers the government a share of the supernormal returns earned by the taxpayer."48 Of course, with tax-prepaid accounts, the government receives no share of supernormal investment returns. What Alarie's analysis neglects is that the taxdeferred method also compels the government to share in the results of subnormal
47 For example, two of the strongest long-term outperforming funds on the Canadian scene have been the Sprott Canadian Equity Fund and the Resolute Performance Fund (and its predecessor Resolute Growth Fund); their maximal declines in the 2008-9 period exceeded 50 percent and 75 percent, respectively. 48 Alarie, at 515-16.
tax-free savings accounts in a consumption-based personal tax n 553 returns, which create revenue losses. The tax-deferred method gives an implicit subsidy on the initial savings equal to the taxpayer's metr; it later collects revenues on the funds withdrawn, including their cumulative investment returns. If the investment yields a return at the same rate as that used for discounting, and assuming that metrs are equal at the two points, then the discounted impact on tax revenues is zero. That result is a direct implication of a consumption-based tax imposing zero effective tax on capital income. But for those investments with subnormal returns, tax revenues suffer a loss in discounted terms; the government gets back less in present value terms than what it had forgone on the initial savings. With investment returns dispersed across subnormal, normal, and supernormal, total discounted tax revenues should be roughly the same for the tax-deferred and tax-prepaid methods.49 B e h avi o u r a l Fa c t o r s a n d Savi n g Individual saving responses may hinge on behavioural traits that go beyond pure rational calculation. Alarie cites behaviours that he believes will predispose individuals to save more, and for longer periods, with tax-deferred than with tax-prepaid savings options.50 His basis for these behaviours is theoretical and experimental economic research, but little direct empirical evidence that bears on the immediate topic is available. He describes the fact that a saver using a tax-deferred scheme will have to pay tax on any withdrawals as a "commitment devices against overconsumption,"51 which should promote longer-term holding of savings. Yet, a person who is normally at high income will be induced to withdraw tax-deferred savings in periods of temporary low income with their low marginal tax rates. And other scenarios can be envisioned that yield varied outcomes. Will the individual choose to save in the first place, and will this choice differ with the provision of a tax-deferred versus a tax-prepaid scheme? With a tax-deferred scheme, individuals at low and moderate incomes need to forecast their future metrs both during the remainder of their working lives and in retirement before deciding whether and how to save; with a taxprepaid scheme, saving is advantageous regardless of the individual's future metr. While one can spin many scenarios that might favour saving under one consumptionbase method or the other, this is all speculative until we have further experience with these schemes. Evidence to date does not support the notion that tax-prepaid schemes are disadvantageous in encouraging savings.52 The United States has permitted 49 Only if investors who have a choice of the two plans shift their investments with supernormal returns into the tax-prepaid plans might discounted net tax revenues be adversely affected; this requires that investors have foreknowledge of which investments will yield supernormal returns, a point rejected earlier. 50 The behaviours cited are hyperbolic discounting, loss aversion, and the endowment effect. 51 Alarie, at 526. 52 Alarie notes, "The tentative lesson from [studies in other countries] is that there is considerable reshuffling of taxable assets to capitalize on the tax advantages associated with tax-preferred savings accounts (both prepaid and postpaid), with modest new saving behaviour among those with low and middle incomes": ibid., at 509-10.
554 n canadian tax journal / revue fiscale canadienne
(2009) vol. 57, no 3
most users of its tax-deferred savings plans (individual retirement accounts, or iras) to convert their funds to tax-prepaid savings plans (so-called Roth iras). These "Roth conversions" have proved to be very popular, demonstrating that many individuals prefer to pay their tax on savings up front and not have worries over unknown future tax liabilities.53 In the United Kingdom, tax-prepaid individual savings accounts (isas) have also become very popular since their introduction in 1999 and have attracted many new savers among those with low and moderate incomes.54 About 37 percent of adult Britons now have an isa, a number rising at 6 percent annually, and total isa holdings have mushroomed to Ј270 billion. The average isa contribution approaches Ј2,500 out of an annual maximum of Ј7,200 (Cdn $13,500). Clearly, experience to date suggests that tax-prepaid savings plans like the tfsa will be widely used across the income spectrum to encourage saving, and there is no empirical evidence to support the contention that tax-prepaid plans will prove less effective than tax-deferred plans.
P o l i c y A lt e r n at iv e s t o TFS A s The government could have pursued tax policies other than the introduction of the tfsa that would also have moved the personal tax further toward a consumption base. Prominent examples worth exploring include increased rrsp/rpp contribution limits, tax-free rollovers of reinvested capital gains, a limited exclusion of capital income, and a shift in the tax mix toward greater reliance on the gst with less reliance on the personal tax. Here I briefly consider the merits of each alternative relative to the tfsa. I reserve for later discussion reforms to the tax-deferred schemes that would introduce some of the attractive design features of the tfsa, since these would not augment the consumption aspect of the tax base. Increased RRSP/RPP Contribution Limits If the tax-deferred method is inherently superior to the tax-prepaid method, as Alarie contends, an attractive alternative to the tfsa might have been to increase the rrsp and rpp contribution limits. This increased access could be pursued in any of three ways: (1) raising the percentage of earnings from 18 percent; (2) raising the maximum annual dollar ceiling;55 and (3) allowing a fixed annual contribution independent of
53 The popularity of the Roth conversion is attested to by the large number of results--"about 193,000"--yielded by a Google search for that term on June 21, 2009; these included several online Roth conversion calculators. 54 Statistics on the ISA are from Benjamin Tal, "The New Tax-Free Savings Account: How Popular Will It Be?" Consumer Watch Canada (Toronto: CIBC World Markets, 2008). 55 See, for example, Jack M. Mintz, The 2006 Tax Competitiveness Report: Proposals for Pro-Growth Tax Reform, C.D. Howe Institute Commentary no. 239 (Toronto: C.D. Howe Institute, 2006), proposing an increase in the contribution limit to 25 percent of earned income and an increase in the dollar limit to $32,000 by 2010.
tax-free savings accounts in a consumption-based personal tax n 555 earnings. Raising the allowable percentage of earnings makes little sense in that very few individuals currently exhaust their cumulative allowable contribution room, other than those at high earnings who are constrained by the dollar limit.56 A percentage higher than 18 would facilitate mainly intergenerational transfers for the few individuals wishing to save at higher rates than needed to sustain their accustomed living standards during retirement.57 A significant increase in the dollar limit would assist savings only for higher earners, since the current $21,000 annual limit equates to $116,700 of earnings at an 18 percent savings rate. To extend full consumptiontax treatment to still higher earners, and to remain competitive with tax provisions for savings in major competitor countries, raising the limit to $32,000 (which equates to earnings of $177,800 at 18 percent savings) might be desirable if the revenue and distributional impacts were deemed acceptable.58 Allowing additional rrsp contributions unrelated to earnings and to a specified limit, mimicking the tfsa, could also be attractive to provide access for retirees and to encourage inter vivos gifts that distribute wealth earlier than bequests.59 Tax-Free Rollover of Capital Gains The introduction of the tfsa was construed by some observers as a replacement for the Conservative Party's earlier campaign pledge to permit tax-free rollovers of reinvested capital gains.60 The tfsa permits not just deferral of tax on realized capital gains on assets held in the account but a full exemption from tax. In several important 56 See the evidence on this point in Ernest B. Akyeampong, "Saving for Retirement: RRSPs and RPPs" (1999) vol. 11, no. 2 Perspectives on Labour and Income 21-27. 57 Note that the 18 percent limit applies to gross earnings; if the typical worker pays about 25 percent of gross earnings in PIT and payroll taxes, the 18 percent limit equates to 24 percent of take-home pay, a very high savings rate. 58 Of course, both the revenue and distributional impacts could be offset by raising the upperbracket tax rates, but this would still have the intergenerational impacts cited earlier (adverse to the older generation). 59 Alarie deems this feature of the TFSA undesirable, characterizing it as an avenue for tax avoidance. He states that the ability to contribute inter vivos gifts "can yield dramatic intergenerational wealth transfers, entirely free of tax" (Alarie, at 527), forgetting the $5,000 annual limit on TFSA contributions. This figure pales beside gifts of cash that are invested in dividend-paying shares: a donee with no other income can receive $66,000 of annual dividends from Canadian corporations free of federal PIT; at a 4 percent dividend rate, that corresponds to the receipt of a $1.65 million gift. This figure is for 2007; corresponding provincial figures for tax-free dividends range from $14,000 (Newfoundland and Labrador) to $50,000 (Ontario, excluding Ontario health premiums) to $162,000 (British Columbia
): see ScotiaMcLeod, "Tax Free Dividends: One of Canada's Best Kept Secrets" (online: http://www.scotiabank.com/cda/ content/0,1608,CID9162_LIDen,00.html). 60 For the view that the TFSA was a possible replacement for the rollover pledge, see Heather L. Evans and Marielle Domercq, "Policy Forum: Tax-Free Savings Accounts--A Practitioner's Perspective" (2008) vol. 56, no. 3 Canadian Tax Journal 708-18, at 709.
556 n canadian tax journal / revue fiscale canadienne
(2009) vol. 57, no 3
respects, the tfsa is far superior to permitting tax-free rollovers of capital gains.61 First, it avoids the extreme distributional tilt of a capital gains rollover; more than half of all capital gains are realized by the top 0.5 percent of Canadian taxpayers. Second, an unrestricted rollover provision would provide mainly windfall tax savings on capital gains that had accrued prior to the date of implementation; it would provide relatively little incentive for new saving or risk-taking. Third, a rollover provision would have a very large revenue cost both initially and on an ongoing basis, unlike the very small initial revenue cost of the tfsa. Finally, a capital gains rollover would further distort investor incentives to structure portfolios to generate returns in the form of gains; in contrast, the tfsa provides neutral treatment of capital returns, be they gains, interest, dividends, or trust distributions.
Limited Exclusion of Capital Income Another proposed alternative to the tfsa was to introduce an annual exemption for a fixed amount of investment income earned outside an rrsp, such as $15,000 per annum.62 This proposal implicitly had a much larger scale than the tfsa; $15,000 tax-free investment income equates to $300,000 of capital at a 5 percent interest rate. At an annual dollar limit of $5,000, an individual would need decades of tfsa contributions plus cumulative returns to achieve the same $300,000 figure. Hence, this alternative would carry far larger revenue costs than the tfsa from the outset. Moreover, it would constitute a large windfall to holders of existing non-registered assets--mostly the wealthy--and provide relatively little incentive to undertake additional saving. The year-by-year nature of the exemption would encourage the holding of low-risk, low-variability assets, thus attenuating the investor's growth of savings. Because the tfsa constrains the annual contribution rather than the annual income that is free of tax, it does not share this bias. Overall, even if rescaled for comparability, a limited exclusion of capital income appears inferior to the tfsa.
Shift the Direct-Indirect Tax Mix Yet another alternative for orienting the tax system more toward consumption would be to shift the tax mix away from direct personal taxes (which have some capital income taxation) and toward indirect consumption taxes.63 Typically, these proposals
61 See Jon Kesselman, "Tory Tax Policies: Populism or Plutocracy?" Vancouver Sun, February 6, 2006. 62 For this proposal, see Christopher Ragan, "TFSAs Will Be Ineffective, Burdensome," Financial Post, March 11, 2008; and for a critique, see Jon Kesselman and Finn Poschmann, "TFSAs: Answering the Critics," Financial Post, March 18, 2008. Canada offered a much smaller annual exemption for investment income in the early 1980s. 63 For proposals of this nature and related analysis, see Dahlby, supra note 28; Kenneth J. McKenzie, "Replacing the Alberta Personal Income Tax with a sales tax
: Not Heresy But Good Economic Sense" (Calgary: Canada West Foundation, 2000); and Mintz, supra note 28. Alarie also notes this option in a footnote, but dismisses it as politically irrelevant: see Alarie, note 33.
tax-free savings accounts in a consumption-based personal tax n 557 would raise the gst rate and trim pit rates. While this approach may have some attractions, it also suffers from several disadvantages relative to reforms (such as the tfsa) that directly shift the personal tax base further toward consumption.64 For most taxpayers over a broad middle range of incomes, the pit is already sufficiently close to a consumption base that little would be gained from such a shift. Trimming pit rates dissipates much of the tax relief on labour income rather than savings, so it is less effective than more explicit pit base reforms. Furthermore, raising the gst rate would carry its own drawbacks: increasing the distortionary biases among taxed, tax-exempt, and zero-rated goods; increasing evasion of both the gst and income taxes; and raising Canada's total sales tax rates above those in the United States, thereby inhibiting any future moves toward a customs union with open borders and losing the associated economic gains. Raising the gst rate also carries adverse distributional impacts that need to be offset by compensatory provisions. P o t e n t ia l R e f o r m s t o TFS A s tfsas debuted at the start of 2009, but it is not too early to contemplate reforms to refine their operation and to avoid future problems. In particular, the Federal Government
's commitment that tfsa holdings will not be considered in any federal benefit programs is likely to cause future problems that should be addressed sooner rather than later. Other policy changes could address how provincial benefit programs will treat tfsa holdings and whether tfsa contribution limits should rise with age. Additionally, some attractive features of the tfsa that differ from the design of tax-deferred schemes (see table 2) might usefully be introduced into rrsps and rpps. As I have demonstrated earlier, transplanting any tfsa design features into the taxdeferred schemes does not undermine the strong reasons for maintaining a separate tax-prepaid scheme. Limited Immunity from Federal Benefit Clawbacks The federal government has made an explicit commitment that with the tfsa "there will be no clawbacks . . . for federal income-tested benefits."65 This position avoids the strong disincentives to saving that arise for low and moderate earners under the tax-deferred schemes.66 However, it fails to anticipate the size that individual tfsas could reach after many years of contribution and reinvested returns. 64 For an early statement of these issues, see Jon Kesselman, "Cutting the GST: Good Politics and Good Economics," Financial Post, January 10, 2006. 65 Canada, Department of Finance, 2008 Budget, Budget Speech, February 26, 2008, 7. The budget plan clarifies that TFSA withdrawals and investment returns will also not affect the individual's income-tested tax credits and transfer benefits: 2008 Budget, Budget Plan, 274. 66 With tax-deferred savings schemes, all withdrawals--including both the initial savings and cumulative investment returns--are taxable and enter the income tests of benefit programs. See Richard Shillington, New Poverty Traps: Means-Testing and Modest-Income Seniors, C.D. Howe Institute Backgrounder no. 65 (Toronto: C.D. Howe Institute, 2003).
558 n canadian tax journal / revue fiscale canadienne
(2009) vol. 57, no 3
Table 3 shows terminal values for a tfsa where an individual and a couple have invested the maximum annual $5,000 and $10,000 sums, respectively, for 40 years under varying assumptions about the real rate of return. All figures given in the table are in real terms, meaning dollars of purchasing power in the initial year. At a real annual return of 5 percent on a balanced portfolio of stocks and bonds, the individual accumulates a final account balance of $630,000 and the couple has $1,270,000. An individual who starts saving later but saves the same total of $200,000, spread evenly at $8,000 per annum over 25 years, would achieve a lower tfsa terminal value ($374,000 versus $630,000 at a 5 percent real rate of return) but still a substantial sum. Much larger terminal values arise for higher assumed rates of investment return, which could occur for portfolios weighted toward growth equities. The illustrative tfsa terminal values are so large that they should not, and undoubtedly will not, be ignored by future governments in their application of benefit clawbacks. In particular, one cannot realistically expect that future governments will pay the counterpart of today's guaranteed income supplement (gis) to individuals holding such large liquid wealth and earning significant investment incomes. Rather than waiting until some distant future year for the government to announce a policy change that reneges on the original no-clawback commitment, it would be prefer able to put appropriate rules in place now. A desirable policy would disregard some threshold amount of tfsa holdings or investment income to retain positive saving incentives for low and moderate earners. If financial institutions reported only annual tfsa balances but not annual investment income, then an appropriate rule would be based on the tfsa balance. For example, federal benefit clawbacks could ignore the first $100,000 (indexed for inflation) of a tfsa balance and impute an income on any balances exceeding that figure; a rate of 4 or 5 percent could be applied in imputing such income, which would enter into the calculation of benefit clawbacks.67 The same or related policies could be applied for the other major federal income-related tax and transfer provisions, such as the old age security (oas) clawback, gst credits, and national child benefits.
Immunity from Provincial Benefit Clawbacks Even if tfsa holdings are immune from federal benefit clawbacks, subjecting them to the asset tests in provincial income assistance programs could still undermine saving incentives for many low and moderate earners. Income assistance provisions vary by province, but typically rrsp holdings are included with other liquid assets in asset tests on the order of $500 to $2,000 for singles and couples, respectively. Locked-in rrsps and rpps are excluded from provincial asset tests, but given their
67 Any individual who found the official imputation rate too high (higher than his rate of return on his chosen portfolio) could simply withdraw TFSA funds exceeding the specified threshold and have the associated investment return enter his net income for tax and benefit calculation purposes.
tax-free savings accounts in a consumption-based personal tax n 559
Table 3 Tax-Free Savings Account Accumulations After 40 Years of Contributions
Terminal value of fund (constant dollars)a
Real rate of return (%)
5 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
630,000 1,070,000 1,840,000
1,270,000 2,130,000 3,680,000
a Terminal value: T =
(1 + r)n+1 -(1 + r) , where X is the annual contribution (constant
real $), r is the real rate of return, and n is the number of years of contributions (made at the
start of each year). b Contributes $5,000 (constant dollars) at the beginning of each year. c Contributes $10,000 (constant dollars) at the beginning of each year.
liquid nature, tfsas will likely be subjected to the asset tests.68 Analysts have proposed that a specified amount of tfsa assets be exempted from the provincial asset tests;69 if this figure were set well below the threshold proposed for exemption from federal benefit clawbacks, many lower and moderate earners who felt at risk of occasional welfare dependence could be deterred from undertaking much retirement saving. Analogous issues need to be resolved for a wide range of other income-tested (and sometimes asset-tested) provincial benefit programs such as day-care and housing subsidies. Differential Contribution Limits by Age Implementing the tfsa with an annual contribution limit of $5,000 for all individuals discriminates against those who are older, since they will have fewer years remaining in which to contribute. This problem is a transitory one that will vanish when all persons now over the age of 18 (the minimum for having a tfsa) have died--clearly many years hence. To redress that inequity, the tfsa could permit larger annual contributions for older persons and gradually phase out the differentiated limits over time. For example, individuals now aged 40-50 could have an $8,000 limit; those 50-65, a $12,000 limit; and so forth. The us tax-prepaid Roth ira recognizes this issue in small degree by providing a $6,000 annual contribution limit for those aged 50-plus versus the $5,000 limit for younger persons. Of course, allowing older persons higher limits increases the scope for asset shifting versus new saving. Economic analyses based on optimal taxation theory have also suggested the desirability of
68 Only Quebec allows a substantial exemption in its asset test and is therefore likely to allow non-trivial amounts of TFSA holdings. 69 See John Stapleton and Richard Shillington, "No Strings Attached: How the Tax-Free Savings Account Can Help Lower-Income Canadians Get Ahead," e-brief (C.D. Howe Institute, September 30, 2008) (online: http://www.cdhowe.org/pdf/ebrief_64.pdf ), suggesting an exemption on the order of $5,000 to $25,000 for TFSAs.
560 n canadian tax journal / revue fiscale canadienne
(2009) vol. 57, no 3
age-dependent allowances for tax-recognized savings, though the exact pattern is not unambiguous.70
Bringing TFSA Features to Tax-Deferred Plans The tfsa's design features differ from those of the existing tax-deferred plans, including some aspects that are particularly attractive. Two tfsa features in particular, also cited by Alarie as candidates, could be considered for transplantation in rrsps: the ability to recontribute amounts withdrawn from a plan and the immunity of withdrawals from benefit clawbacks. A recontribution feature would be easy to implement in the tax-deferred plans and would add flexibility to individual saving strategies; it would be most useful to those suffering periodic sharp declines in their earnings and assist them in both smoothing their incomes and averaging their taxes. Yet the data show that relatively few people, other than top earners, have exhausted their cumulative rrsp contribution room, so that a right to recontribute might not be widely utilized.71 Full or partial immunity of rrsp withdrawals from benefit clawbacks might also appear attractive. However, in contrast to tfsas, the original contributions to rrsps do not bear tax, so that would accord a windfall to persons who had been at high incomes (and therefore gained a large tax benefit) at the time of their contributions but whose income subsequently fell.
P o l i t i c a l E co n o m y o f TFS A s One common selling point of tax-prepaid schemes like the tfsa is their relatively low revenue cost at the outset as compared with the enhancement of tax-deferred schemes. Alarie construes this point as a liability of the tfsa in political economy terms. He contends that "the political expediency of tfsas comes at both efficiency and distributional costs vis-а-vis [tax-deferred] plans."72 Yet, by facilitating the political acceptance of greater or faster movement of the personal tax system toward a consumption base, the tax-prepaid method can achieve more of the associated gains in economic efficiency. In addition, as explained earlier, the tax-prepaid method is inherently more economically efficient (except in the unrealistic case where all taxpayers face a constant metr over their lifetime).73 As for the comparative distributional
70 See the analysis and sources cited in the study by Banks and Diamond, supra note 27, which also supports age-dependent tax rates on labour earnings. 71 See Akyeampong, supra note 56. 72 Alarie, at 529. 73 Alarie refers to "evidence that tax-prepaid savings plans are associated with greater excess burden than tax-postpaid plans" (ibid., at 530) and cites Kesselman and Poschmann, "Expanding the Recognition of Personal Savings," supra note 3. However, this alleged superiority of cash flow consumption taxation stems from its lump-sum burden imposed on older cohorts during the transition, which finances a lower, less distorting tax rate on labour earnings (see Auerbach and Kotlikoff, supra note 7). If the older cohorts are compensated for their losses, this efficiency advantage vanishes.
tax-free savings accounts in a consumption-based personal tax n 561 effects of introducing or expanding a tax-prepaid scheme versus expanding contribution limits on an existing tax-deferred scheme, the results are also clearly favourable to the tax-prepaid scheme. With tax prepayment, the saver achieves future exemption of tax on investment returns but (in contrast to tax-deferred savings) does not obtain an upfront tax benefit proportional to his tax rate, and therefore to his income. Expanding either type of scheme does tilt in favour of individuals with higher incomes and wealth, who can more readily use the extra contribution room. But expanding the rrsp contribution limit is uniquely favourable to high earners because of its linkage to earnings. Alarie raises a further concern about the political economy of tax-prepaid schemes. He believes that they raise "difficult to countenance distributional consequences" that will lead future governments to renege (albeit, "in all likelihood only partly") on their promises of no further taxation on those funds.74 Yet my previous discussion of horizontal equity noted the widespread acceptance of other tax provisions that allow full exemption of much more common and larger returns from gambling, lotteries, and home appreciation. There is little reason to think that the public would view the occasional tfsa holder who had achieved high returns any differently; in fact, the public has little means of learning about such outcomes. So long as clear specified limits on the degree of immunity of tfsas from benefit clawbacks are implemented, one should not anticipate any greater likelihood of government reneging on promises related to tax-prepaid plans than those in respect of tax-deferred plans. The popularity of Roth conversions in the United States, noted earlier, testifies to the public's lack of concern over greater risks of broken promises for tax-prepaid schemes. And broken promises regarding tax exemption are not unprecedented for tax-deferred schemes; Denmark, Italy, and Sweden have imposed tax on investment earnings within tax-deferred accounts, and several other countries have imposed partial tax.75 Conclusion Alarie begins his article by stating that "the policy case in support of tfsas is underwhelming," and his conclusions on tfsas are no more sanguine. This view contrasts sharply with the finance minister's description of tfsas as "the single most important personal savings vehicle since the introduction of the rrsp" in the 1950s,76 a sentiment echoed numerous times by tax, financial, and economic analysts and writers.77 Objective analysis supports the finding that a tax-prepaid savings scheme 74 Alarie, at 530. 75 See Kwang-Yeol Yoo and Alain de Serres, "Tax Treatment of Private Pension Savings in OECD Countries" (2004) vol. 39, no. 2 OECD Economic Studies 73-110, at 79. Australia and New Zealand
, with tax-prepaid schemes, also impose tax on fund investment earnings. 76 2008 Budget Speech, supra note 65, at 6. 77 A search of the Proquest database for the term "tax-free savings account" for the period from the 2008 budget date to June 21, 2009 produced 699 distinct citations (mostly in newspapers).
562 n canadian tax journal / revue fiscale canadienne
(2009) vol. 57, no 3
like the tfsa can play an important role in shifting Canada's personal tax system further toward consumption. And further, albeit limited, moves away from capital income taxation and toward consumption taxation are desirable for a variety reasons. Contrary to common belief, the case for consumption taxation--and for the tax-prepaid method--does not hinge critically upon the aggregate savings response. The benefits are far more wide-ranging and include enhanced economic efficiency and growth, horizontal equity, and simplicity for both tax administration and tax compliance. Moreover, the tax-prepaid method poses less sacrifice of vertical equity than the tax-deferred method of implementing a consumption base. Some specific design features of the tfsa are not inherent to a tax-prepaid scheme, and if deemed desirable, these features could easily be transplanted into the tax-deferred rpps and rrsps. Chief among those features are a right to recontribute withdrawals and immunity from benefit clawbacks on some withdrawals. But key inherent characteristics of a tax-prepaid scheme cannot be replicated in a tax-deferred scheme; such salutary attributes relate to economic efficiency, vertical equity, and ease for taxpayer planning and compliance. Nevertheless, the two methods work to best effect when they are offered jointly. Tax-deferred schemes offer the advantage of facilitating income averaging and, for those analysts concerned about the issue, capturing most supernormal returns on investment through high contribution limits; lower contribution limits can then be imposed in the companion tax-prepaid scheme. However, non-taxation of supernormal returns under the tax-prepaid method does not cause revenue loss relative to the tax-deferred method, and any associated in equities that may be deemed to arise pale beside other longstanding tax provisions for lotteries, gambling, home ownership, and small-business shares. Consideration of both saving behaviours and political economy factors yields no conclusive results favouring one method over the other. Still in its infancy, the tfsa could benefit from refinements to specify limits on immunity from federal benefit and tax-related clawbacks as well as clawbacks in provincial means-tested benefit programs. Allowing higher annual tfsa contribution limits for older cohorts could also be justified during a long transition period. None of the policy alternatives to the tfsa canvassed here--increased rrsp/rpp contribution limits, tax-free capital gain rollovers, a capital income exclusion, and a shift toward more reliance on indirect taxes like the gst--appears superior to the choice of implementing the tfsa. In short, the tax-free savings account is an important and useful innovation for Canadian tax policy and is certain to attract continuing attention from both tax planners and tax analysts.