Canadian Tax Journal Revue Fiscale Canadienne

■ 2018 volume 66, no 2

Canadian Tax Journal Revue fiscale canadienne

269 Interaction of the Foreign Affiliate Surplus and Safe-Income R egimes: Selected Anomalies, Issues, and Planning Considerations jim samuel

309 VAT/GST Thresholds and Small Businesses: Where To Draw the Line? yige zu

349 Policy Forum: Editors’ Introduction—The Case for a Tax Commission alan macnaughton and kevin milligan

351 Policy Forum: Why Canada Needs a Comprehensive Tax R eview fred o’riordan

363 Policy Forum: Then and Now—A Historical Perspective on the Politics of Comprehensive Tax R eform shirley tillotson

375 Policy Forum: Building a Tax Review Body That Is Fit for Purpose—Reconciling the Tradeoffs Between Independence and Impact jennifer robson

387 Policy Forum: From Independent Tax Commission to Independent Tax Authority joseph heath

401 Current Cases: (FCA) Gervais v. Canada; (FCA) Canada v. Oxford Properties Group Inc. robert korne, joy elkeslassy, and andrew stirling

421 Personal Tax Planning: Revisiting Planning for Private Company Shareholders After July 2017 robert santia

447 Planification fiscale personnelle : Réexamen de la planification pour les actionnaires de sociétés privées après juillet 2017 robert santia

475 Selected US Tax Developments: A Close Look at Some of the New US Tax Rules michael miller

491 Current Tax Reading robin boadway and kim brooks

■ v ■ canadian tax journal / revue fiscale canadienne (2018) 66:2, 269 - 307

Interaction of the Foreign Affiliate Surplus and Safe-Income Regimes: Selected Anomalies, Issues, and Planning Considerations

Jim Samuel*

Précis Le gouvernement canadien a présenté, dans son budget de 2015, des modifications substantielles à l’article 55 de la Loi de l’impôt sur le revenu. Ces modifications comprennent deux nouveaux critères de l’objet qui permettent d’établir si le paragraphe 55(2) s’applique pour requalifier un dividende intersociétés autrement libre d’impôt lorsqu’il est versé par une société qui réside au Canada à une autre société résidente en tant que gain en capital qui est assujetti à l’impôt. Étant donné la large portée de ces nouveaux critères de l’objet, et de l’incertitude qu’ils créent, il sera à présent probablement plus courant pour les sociétés de se prévaloir, lorsque c’est possible, de l’exception pour revenu protégé. Ces circonstances pourraient inclure, par exemple, le paiement d’un dividende intersociétés par une filiale canadienne détenue en propriété exclusive à sa société mère. En outre, si cette filiale détient, directement ou indirectement, une ou plusieurs sociétés étrangères affiliées, il est possible que la totalité ou une partie des surplus mis en commun d’une ou de plusieurs de ces sociétés affiliées puisse faire partie intégrante, dans certaines circonstances, du calcul d’un revenu protégé. Cet article présente une comparaison des aspects fondamentaux de ces deux régimes, en mettant particulièrement l’accent sur les anomalies, les questions et les considérations de planification les plus courantes dans le contexte de leur interaction. L’analyse présentée par l’auteur montre que, bien que les deux régimes aient un objectif de calcul similaire, l’interaction des deux peut parfois être difficile et produire des résultats imprévus.

Abstract In the 2015 budget, the Canadian government introduced sweeping amendments to section 55 of the Income Tax Act. These amendments include two new purpose tests that apply in determining whether subsection 55(2) applies to recharacterize an otherwise “tax-free” intercorporate dividend paid between two Canadian-resident corporations as

* Of KPMG Canada, Calgary (e-mail: [email protected]). I am grateful for the assistance of Gregory Bell of KPMG LLP, Ottawa/Toronto. Any errors herein remain my responsibility and any views expressed herein are my own.

269 270 n canadian tax journal / revue fiscale canadienne (2018) 66:2 a capital gain that is subject to tax. Given the broad scope of, and the uncertainty arising from, these new purpose tests, it will now likely be more common for corporations to, where possible, rely on the safe-income exception. These circumstances could include, for example, the payment of an intercompany dividend by a wholly owned Canadian subsidiary to its parent company. Furthermore, if that subsidiary owns, directly or indirectly, one or more foreign affiliates, it is possible that all or a portion of the surplus pools of one or more of those affiliates could, in certain circumstances, form an integral part of a safe-income calculation. This article provides a comparison of the fundamental aspects of these two regimes, with a particular focus on some of the more commonly encountered anomalies, issues, and planning considerations that can arise in the context of their interaction. The analysis presented by the author shows that, although the two regimes have a similar computational objective, the interplay between them can at times be uneasy and can give rise to unexpected results.

Keywords: Foreign affiliates n surplus n safe income n dividends n amendments n surplus stripping

Contents Introduction 271 Historical Overview of the Foreign Affiliate Surplus and Safe-Income Regimes 273 Foreign Affiliate Surplus Regime 273 Safe-Income Regime 276 Overview of the Section 55 Amendments, and the Computation of Safe Income 280 Overview of the Computation of Foreign Affiliate Surplus Pools 282 Conceptual Comparison of the Foreign Affiliate Surplus and Safe-Income Regimes 285 Intersection of the Foreign Affiliate Surplus and Safe-Income Regimes 291 Sandwich Structure 291 Outbound Structure 292 Selected Anomalies, Issues, and Planning Considerations 293 Use of the Book Depreciation Election To Compute Active Business Earnings 294 Treatment of Deficits of a Foreign Affiliate 295 Foreign Exchange Rate To Be Used for Inclusion of Foreign Affiliate Surplus Pools in Safe Income 295 Maximization of a Particular Foreign Affiliate’s Tax-Free Surplus Balance 296 Payment of Dividends by a Foreign Affiliate 296 Stub-Period Earnings of a Foreign Affiliate 298 Amounts That Accrue Prior to Becoming a Foreign Affiliate of the Taxpayer 299 Ownership of Shares of a Foreign Affiliate Through a Partnership 300 Retroactive Adjustments to Surplus Pools 301 Reorganizations Involving the Shares or Property of Foreign Affiliates 303 Two Peas in a Pod? 304 Conclusion 307 interaction of the foreign affiliate surplus and safe-income regimes n 271

Introduction In the 2015 budget, the Canadian government introduced sweeping amendments to section 55 of the Income Tax Act.1 After a lengthy consultation period, a modi- fied version of those amendments was enacted on June 21, 2016 as part of Bill c-15.2 The amendments, which apply to dividends received after April 20, 2015, include two new purpose tests3 that apply in determining whether subsection 55(2) applies to a particular dividend received by a corporation. Subsection 55(2) is an anti-avoidance rule that, if applicable, can recharacterize an otherwise “tax-free” intercorporate dividend paid between two Canadian-resident corporations as a capital gain that is subject to tax. As was the case under the former version of subsection 55(2), it is still possible in certain circumstances to avoid the potential application of subsection 55(2) to a particular dividend if that dividend does not exceed the “safe income” of any relevant corporation. Given the broad scope of, and the uncertainty arising from, the two new purpose tests, it will now likely be more common for corporations to rely, where possible, on the safe-income exception—even in circumstances where no disposition of shares is expected to occur. These circumstances could include, for example, the payment of an intercompany dividend by a wholly owned Canadian subsidiary to its parent company. Furthermore, if that subsidiary owns, directly or indirectly, one or more foreign affiliates,4 it is possible that all or a portion of the surplus pools5 of one or more of those affiliates could, in certain circumstances, form an integral part of a safe-income calculation. Alternatively, if a foreign affiliate is itself a shareholder of a taxable Canadian corporation6 or of a corporation resident in Canada,7 the concept of safe income could also be relevant for the purposes of determining the treatment of a

1 RSC 1985, c. 1 (5th Supp.), as amended (herein referred to as “the Act”). Canada, Department of Finance, 2015 Budget, April 21, 2015. Unless otherwise stated, statutory references in this article are to the Act. 2 Budget Implementation Act, 2016, No. 1; SC 2016, c. 7. See, for example, amended subsections 55(2) to (2.5) of the Act. 3 Paragraph 55(2.1)(b). 4 As defined in subsection 95(1). In general terms, a non-resident corporation is a foreign affiliate of a Canadian-resident taxpayer if the taxpayer’s equity percentage in the corporation is not less than 1 percent and the total equity percentage in the corporation owned by the taxpayer and each person related to the taxpayer is not less than 10 percent. 5 A corporation resident in Canada is generally required to maintain exempt surplus (or deficit), hybrid surplus (or deficit), taxable surplus (or deficit), hybrid underlying tax, and underlying foreign tax pools for each of its foreign affiliates (herein collectively referred to as “surplus pools”). 6 As defined in subsection 248(1). 7 For ease of reference and except as otherwise noted, in the text that follows a “taxable Canadian corporation” and a “corporation resident in Canada” may also be referred to as a “Canadian corporation.” 272 n canadian tax journal / revue fiscale canadienne (2018) 66:2 dividend that is received by the affiliate from such a corporation in computing the affiliate’s foreign accrual property income8 (fapi) and surplus pools. The amendments to section 55 have been discussed at length in other papers,9 as have the safe-income and foreign affiliate surplus regimes.10 Accordingly, this article will not undertake a detailed analysis of those amendments or regimes, or the legislative and policy considerations related to them. The interaction between the safe-income and foreign affiliate surplus regimes has also been considered, to some extent, in other papers.11 However, given the increasing importance of the computation of safe income in light of the recent amendments to section 55, the interaction of the two regimes has become correspondingly more important. Thus, the purpose of this article is to provide a closer

8 As defined in subsection 95(1). FAPI of a foreign affiliate includes, among other things, “income from property” as defined in subsection 95(1) (for example, dividends, interest, rents, royalties, etc., depending on the facts and circumstances); business income that is deemed under subsection 95(2) to be income from a business other than an active business (and consequently FAPI); and, very generally, taxable capital gains from the disposition of property that does not constitute “excluded property” (as defined in subsection 95(1)). In computing taxable income for a particular taxation year, a taxpayer resident in Canada is required under subsection 91(1) to include its participating percentage (as defined in subsection 95(1)) of the FAPI earned by each of its controlled foreign affiliates. 9 For a more detailed analysis of the recent amendments to section 55, see, for example, Rick McLean, “Subsection 55(2) Amendments: What’s the Purpose?” in 2016 St. John’s Tax Seminar (Toronto: Canadian Tax Foundation, 2016), 2:1-48, and Rick McLean, “Subsection 55(2): What Is the New Reality?” in Report of Proceedings of the Sixty-Seventh Tax Conference, 2015 Conference Report (Toronto: Canadian Tax Foundation, 2016), 22:1-71. 10 For a more detailed analysis of the safe-income regime, see, for example, Mark Brender, “Subsection 55(2): Then and Now,” in Report of Proceedings of the Sixty-Third Tax Conference, 2011 Conference Report (Toronto: Canadian Tax Foundation, 2012), 12:1-35; Janette Y. Pantry and Bill S. Maclagan, “Issues and Updates—Safe Income,” in 2008 British Columbia Tax Conference (Toronto: Canadian Tax Foundation, 2008), 4:1-35; Greg C. Boehmer and John M. Campbell, “Safe Income and Safe Income on Hand: An Update,” in 2005 Ontario Tax Conference (Toronto: Canadian Tax Foundation, 2005), 5:1-33; Peter K. Rogers, “Safe Income: A Review of the Calculation of ‘Safe Income’ and Tax Planning Opportunities,” in 2004 British Columbia Tax Conference (Toronto: Canadian Tax Foundation, 2004), 16:1-20; Duncan Osborne, “Practical Issues in Computing Safe Income,” in Report of Proceedings of the Fifty-Fourth Tax Conference, 2002 Conference Report (Toronto: Canadian Tax Foundation, 2003), 42:1-23; and Janette Y. Pantry, “Safe Income—How Safe Is It?” in 1999 British Columbia Tax Conference (Toronto: Canadian Tax Foundation, 1999), 8:1-25. For a more detailed analysis of the foreign affiliate surplus regime, see, for example, Firoz K. Talakshi and James Samuel, “Foreign Affiliate Surplus: What You See Isn’t Always What You Get,” in Report of Proceedings of the Sixtieth Tax Conference, 2008 Conference Report (Toronto: Canadian Tax Foundation, 2009), 28:1-34; Firoz K. Talakshi and Jim Samuel, “Foreign Affiliate Surplus—Got Enough?” (2005) 18:1Canadian Petroleum Tax Journal; and Drew Morier and Raj Juneja, “Foreign Affiliates: An Updated Primer,” inReport of Proceedings of the Sixty-Fourth Tax Conference, 2012 Conference Report (Toronto: Canadian Tax Foundation, 2013), 28:1-58. 11 See, for example, Paul Dehsi and Korinna Ferhmann, “Integration Across Borders,” International Tax Planning feature (2015) 63:4 Canadian Tax Journal 1049-72. interaction of the foreign affiliate surplus and safe-income regimes n 273 and more comprehensive comparison of the fundamental aspects of these regimes, with a particular focus on some of the more commonly encountered anomalies, issues, and planning considerations that can arise in the context of their interaction. The analysis presented in this article will show that, although the two regimes have a similar computational objective, the interplay between them can at times be uneasy and can give rise to unexpected results. To set the stage for the discussion that follows, I first provide a brief history of each of the foreign affiliate surplus and safe-income regimes before diving into overviews of certain relevant aspects of the recent amendments to section 55 and the skeletal frameworks of the two regimes. Next, I highlight some of the key similarities and differences between these regimes and subsequently focus on certain anomalies, issues, and planning considerations that can arise from their interaction. Finally, I attempt to decipher the reasons behind the stark difference in the approach taken in the drafting of each of these regimes, before turning to the question of whether the time has finally come not only to have a detailed set of mechanical rules for the computation of safe income but also to more fully harmonize the two regimes.

Historical Overview of t he Foreign Affiliate Surplus and Safe-Income Regimes In trying to understand the differences in, and the legislators’ approach to, the foreign affiliate surplus and safe-income regimes, it is informative to look to the past.

Foreign Affiliate Surplus Regime In the years leading up to 1972, the taxation of dividends received by a Canadian corporation from a foreign subsidiary was remarkably simple. That is, a dividend, regardless of the nature of the subsidiary’s underlying earnings, was generally received by the corporation entirely free of Canadian tax, by virtue of an offsetting deduction,12 provided that certain basic tests of share ownership were met. The policy rationale behind permitting such dividends to be received free of tax appears to have been primarily to achieve simplification. The ability to simply claim an offsetting deduction against a dividend that is included in computing the Canadian corporation’s income, regardless of the quantum of foreign tax that was paid in respect of the underlying earnings of the foreign subsidiary, is clearly straightforward for both taxpayer compliance and administration by the Canadian tax authorities. In designing the Canadian taxation of such dividend income in a manner that favours simplicity over complexity, it is likely that the government of the day decided to arbitrarily presuppose that the quantum of taxes paid by a foreign subsidiary in the course of earning its income available for distribution would generally approximate the Canadian corporate tax rate. The rationale implicit in such an approach was likely that the granting of foreign tax credits to a Canadian corporation to offset

12 Paragraph 28(1)(d) of the 1952 Income Tax Act, RSC 1952, c. 148. 274 n canadian tax journal / revue fiscale canadienne (2018) 66:2 the Canadian tax otherwise payable in respect of a dividend received from a foreign subsidiary could result in significant complexity for taxpayers and the Canadian tax authorities alike, with little, if any, additional Canadian income tax likely being payable at the end of the day. The cost of such simplicity was that there were large gaps in the system that could be exploited, particularly where a foreign subsidiary earned income through a low-tax jurisdiction. Furthermore, Canada’s approach to the taxation of such dividends was perceived as being out of line with, or overly favourable as compared with, the approach being taken by other countries. At the time, it was not uncommon for countries to tax a dividend received by a domestic corporation from a foreign subsidiary and grant a deduction or credit for any underlying foreign tax paid in respect of such dividend. In recommending the repeal of the exemption system that existed prior to 1972 for dividends received by a Canadian corporation from a foreign subsidiary, as part of an overall objective of promoting neutrality in the Canadian tax system, the 1967 report of the Carter commission stated:

It is apparent that within the Canadian treatment of foreign source business income may be found the two classical extremes of allowance for foreign taxation. One provides for the full, accurate and precise measurement of the foreign income and tax liability, with a precisely computed credit against Canadian tax. The other grants an exemption from tax under conditions very easily met. The United States and the United Kingdom have followed the first method both for branch income and direct investment income, and no provision comparable to section 28(1)(d) of the Income Tax Act may be found in the tax system of either country. There are some examples of exemption of foreign dividends to be found in other countries, but Canada is virtually unique in its adoption of a provision as sweeping as section 28(1)(d). Its origins and effects are therefore of considerable interest. The exemption contained in section 28(1)(d) appears to have had as its original purpose the achievement of an equitable and administratively simple alternative to the complexities of the gross-up and credit procedure. At the time of its introduction in 1949, the bulk of Canadian foreign source income originated in countries having corporation taxes as high as the Canadian, mainly the United States and the United Kingdom. The effect of this section was undoubtedly to provide directly for the virtual exemption of foreign source income from Canadian tax which was the end result of the complicated gross-up and tax credit procedure previously in force. Its origins in section 4(r) and subsections (2a) and (2b) of section 8 of the Income War Tax Act are clearly discernible, and the fact that both of these sections were repealed in 1949 on the enactment of section 27(1)(d) (now section 28(1)(d)) supports the conclusion that, initially, the provision was looked on mainly as a device for administrative simplification. At first the ownership requirement was 50 per cent or more but in 1951, following the recommendation of the Advisory Committee on Overseas Investment, the ownership test was reduced to its present 25 per cent as a means of encouraging foreign investment by Canadians. It has since remained at that level. One result of these provisions is that the Canadian taxpayer has enjoyed a much greater simplicity and ease of calculation for foreign income than his United States or interaction of the foreign affiliate surplus and safe-income regimes n 275

United Kingdom counterparts. The tax minimization possibilities of the exemption privilege, in combination with the use of foreign tax havens, have not gone unnoticed. The provision can be used to reduce Canadian tax on income generated in Canada for the benefit of Canadians. By establishing companies in jurisdictions which impose little or no tax, Canadians can reduce their Canadian tax by engaging in a series of paper transactions which exploit the provisions of tax treaties in combination with section 28(1)(d). There is also evidence that the provision has offered the possibility to use Canada itself as a tax haven for international business. Data compiled for us by the Taxation Division show that over a period of years a very substantial part of the dividends reported under this section has originated in jurisdictions imposing little or no tax, and that a very high proportion of these dividends has been received in Canada by holding companies not having a substantial Canadian economic interest but represent- ing for the most part foreign ownership. Of a total of $1,500 million received by all Canadian corporations (including those that were owned by non-residents) in the five years from 1957 to 1961, only 10 per cent came from the United States and 4 per cent from the United Kingdom. The defects of the present section 28(1)(d) are obvious, and we therefore recommend its repeal.13

Much of the analysis underlying the recommendations of the Carter report was considered by the Department of Finance in its 1969 white paper.14 The white paper also contained various proposals to reform Canada’s international tax system, and it is in those proposals that one can see the beginnings of the foreign affiliate system that is in place today.15 After the release of the Carter report and the white paper, an extensive tax reform process followed in the early and mid-1970s and ultimately culminated in a complete redesign of Canada’s approach to the taxation of income earned by, and dividends received from, a foreign subsidiary of a Canadian corporation. As part of this redesign, the regulations dealing with the computation of foreign affiliate surplus were finalized in 1976. Although there have been many amendments to these regulations over the years, the framework for the computation of foreign affiliate surplus remains largely the same as it was when the regime first came into existence. As evidenced by chapter 26 of the Carter report (from which the extract above is taken), the white paper, and the extensive tax reform process, it is clear that a great deal of time and effort was spent in designing the foreign affiliate surplus regime that exists today. Furthermore, the manner in which the legislators designed the

13 Canada, Report of the Royal Commission on Taxation, vol. 4 (Ottawa: Queen’s Printer, 1967), at 510-12 (herein referred to as “the Carter report”). 14 E.J. Benson, Proposals for Tax Reform (Ottawa: Department of Finance, 1969) (herein referred to as “the white paper”). 15 J. Scott Wilkie, Robert Raizenne, Heather I. Kerr, and Angelo Nikolakakis, “The Foreign Affiliate System in View and Review,” inTax Planning for Canada-US and International Transactions, 1993 Corporate Management Tax Conference (Toronto: Canadian Tax Foundation, 1994), 2:1-72, at 2:44. 276 n canadian tax journal / revue fiscale canadienne (2018) 66:2 regime, and its detailed framework, is such that the underlying policy objectives are quite clear. What is not so clear, as a matter of interest, is why the legislators decided to place the rules for computing foreign affiliate surplus pools in the regulations, as opposed to the Act itself. Presumably the primary reason for this structural approach was that the intricate and technical nature of the subject matter was such that it was best served by locating the surplus regime in the regulations. After all, it is commonly believed that the main difference between a provision in the Act and a provision in the regulations is that the latter, by virtue of not always requiring passage by Parliament, can be amended frequently and on a fast-track basis. Nevertheless, in recent years many substantive additions (or amendments) to regulations, particularly those that relate to the foreign affiliate surplus regime, have been passed by Parliament as part of a bill. So common has this become, and given the confusion that can arise from following two methods to pass regulations, at least one author has questioned whether the concept of regulations is outdated.16

Safe-Income Regime Whereas the government legislated a detailed framework for the computation of foreign affiliate surplus, such was not the case for safe income, even though the latter regime was enacted almost five years after the foreign affiliate surplus regime. In enacting subsection 55(2) and ultimately the concept of safe income in 1981, the government decided not to legislate a detailed set of mechanical rules for the computation of safe income. In other words, the legislators deliberately chose not to use the foreign affiliate surplus regime as a precedent model in enacting the concept of safe income. To try to understand why the legislators decided to adopt a far less prescriptive approach for the determination of safe income, and perhaps why such an approach continues to this day, it is helpful to understand the evolution of the corporate tax system in Canada and in particular the historical tug-of-war involving the taxation of capital gains and dividends. Prior to 1972, capital gains were not subject to tax. Accordingly, an individual was not taxed on a capital gain realized on a disposition of shares of a Canadian corporation. In contrast, and unlike an intercorporate dividend paid by one Canadian corporation to another, a dividend received by an individual from a Canadian corporation was generally taxable, although starting in 1949 a dividend tax credit was available to partially offset the tax otherwise payable on such a dividend. Given the preferential tax treatment of a capital gain as compared with a dividend, there was an inherent incentive in the tax system for individuals to try to convert what would otherwise be a taxable dividend into a tax-free capital gain. It was therefore not uncommon for taxpayers to engage in so-called surplus-stripping transactions that attempted to convert undistributed surplus (that is, retained earnings) of a Canadian corporation into a capital gain. Surplus-stripping transactions

16 Xiao Jin Chen, “Regulations: An Outmoded Idea?” (2013) 3:3 Canadian Tax Focus 3. interaction of the foreign affiliate surplus and safe-income regimes n 277 were structured in a variety of ways, some of which involved surplus of one corporation being extracted to another via a tax-free intercorporate dividend. To combat surplus stripping, the government introduced various specific measures that were targeted at certain types of transactions. These measures included, for example, rules that deemed a shareholder to have received a dividend where certain types of reorganizations or transactions involving the shares of a corporation occurred at a time when that corporation had undistributed surplus. Another measure was the introduction of the concept of designated surplus. In general terms, designated surplus included undistributed surplus of a corporation at the time that control of the corporation was acquired by another corporation. If a dividend was considered to have been paid out of a corporation’s designated surplus to another corporation, that dividend was not eligible for the intercorporate dividend exemption and thus was taxable to the recipient. In addition, tax was generally imposed on a dividend that was paid by a corporation out of designated surplus to a non-resident corporation or a tax-exempt person. Given their narrow application, these specific anti-surplus-stripping measures had only limited effect in curbing surplus-stripping transactions. As a result, in 1963 the government enacted a new targeted anti-avoidance rule in the form of section 138a(1) of the former Act. Although this anti-avoidance measure seemed to be more effective in combatting surplus stripping, concerns were raised by taxpayers regarding the uncertainty that arose from this vaguely worded measure. In a review of Canada’s system of corporate taxation, chapter 19 of the Carter report17 highlighted numerous significant defects and made various recommendations for reform. Those recommendations included the introduction of an integrated tax system, which would achieve full integration of personal and corporate income taxes. As part of such a tax system, and because the favourable treatment of capital gains encouraged surplus-stripping transactions, the Carter report also recommended that capital gains be taxed. After the publication of the Carter report and the subsequent release of the government’s white paper, the stage was set for significant changes to the Canadian tax system. The resulting tax reform ultimately included far-reaching changes to the taxation of shareholders and corporations, starting in 1972, particularly in respect of the taxation of dividends and capital gains. After tax reform, the general rule was that any distribution made by a corporation would be treated as a taxable dividend to the recipient. There were, however, a number of important exceptions. For example, intercorporate dividends generally continued to be received tax-free. Also, a shareholder could now access the paid-up capital of the shares of a corporation on a tax-free basis. Similarly, if a corporation followed certain procedures to treat a distribution as having been paid out of its “1971 undistributed income on hand” or its “1971 capital surplus on hand,”18 that

17 Supra note 13, at 3-98. 18 See, for example, former sections 83, 89, and 196. 278 n canadian tax journal / revue fiscale canadienne (2018) 66:2 distribution was not taxable to the recipient, but tax was imposed on the corporation in respect of a distribution that was paid out of the first of these two pools. As in the case of a distribution of paid-up capital, the price for receiving a distribution from one of these pools was that the adjusted cost base of the shares of the corporation held by the shareholder(s) was reduced by the amount of the distribution. In addition to changes to the taxation of dividends, capital gains were also now taxable. However, since the capital gains inclusion rate was only 50 percent whereas dividends received by non-corporate shareholders were fully taxable, the government also made some other significant supporting changes in a continued effort to combat surplus stripping. In particular, a decision was made not only to retain the concept of designated surplus, but also to strengthen it. Instead of taxing a corporation on the receipt of a dividend that was paid out of designated surplus, such a dividend was now deductible in computing the income of the corporate recipient but was subject to tax at a rate of 25 percent.19 As a backstop measure, the targeted anti-avoidance rule in section 138a(1) of the former Act was retained but moved to subsection 247(1). Also, section 84 was introduced to deem an amount to be a dividend, as opposed to a capital gain, in certain situations involving a redemption of shares, a reduction of share capital, or the liquidation of a corporation. After the Act was amended to tax post-1971 capital gains, taxpayers turned their attention to ways of structuring their affairs so as to minimize the amount of tax otherwise payable on capital gains. In other words, tax reform ultimately effected a change in the behaviour of taxpayers whereby the focus gradually shifted from surplus stripping (that is, the conversion of taxable dividends otherwise receivable by individuals into capital gains that were either not taxed or taxed at preferential rates) to capital gains stripping (that is, the conversion of capital gains into tax-free intercorporate dividends). Prior to 1981, there were limited legislative measures in place to discourage capital-gains-stripping transactions. The primary tool to combat these types of transactions was a broadly worded anti-avoidance rule found in a former version of section 55. That rule, which was introduced as part of the tax reform in 1971, applied where a taxpayer disposed of property in circumstances whereby the taxpayer may reasonably be considered to have artificially or unduly reduced the amount of the capital gain otherwise arising from the disposition. Given the lack of a detailed legislative framework for the application of section 55, in the late 1970s Revenue Canada, the predecessor to the Canada Revenue Agency (cra), started developing and publishing its own views as to what constituted an artificial or undue reduction of a capital gain. Some of these views or administrative practices, commonly referred to as “the Robertson guidelines,” were outlined in two papers written by J.R. Robertson, then director of the Rulings Division of Revenue Canada, and published in 1978 and 1979 by the Canadian Tax

19 The concept of designated surplus was ultimately repealed for dividends paid or received after March 31, 1977. interaction of the foreign affiliate surplus and safe-income regimes n 279

Foundation (ctf).20 It is in the Robertson guidelines that one can see the beginnings of the safe-income regime that exists today, as is evident in the following statement:

In general terms the Department considers that the capital gain should not be less than the increase in the value of the shares reasonably attributable to unrealized or untaxed appreciation in goodwill and other assets after 1971. Therefore, the above transactions would be viewed as unduly reducing the gain if the dividends exceed the amount of Opco’s retained taxed earnings.21

Owing to limited success in using existing tools to combat capital-gains-stripping transactions, in 1981 the government enacted new legislation, which included the original version of subsection 55(2), to try to more adequately counter these types of transactions. Although this legislation bore some similarity to the Robertson guidelines, there were also some significant differences.22 However, in enacting subsection 55(2) and the related provisions, the government deliberately chose not to legislate a comprehensive set of mechanical rules defining how safe income was to be calculated. In the absence of a detailed legislative framework, the calculation of safe income came to be developed largely through administrative practice. One of the earliest examples of these administrative practices was the publication of the so-called Robertson rules presented at the ctf’s 1981 annual conference.23 So much has the determination of safe income relied upon administrative practice that one author has remarked that “[o]ver the years, these administrative practices have taken on statute-like authority, being in large measure accepted by the tax community as a practical set of rules and guidelines to address issues related to the computation of safe income.”24 So why did the government decide not to legislate a detailed definition for the computation of safe income, and why does that approach continue to this day? Before attempting to answer these questions, and the question of whether it is time to reconsider the government’s decision, it is perhaps helpful to first understand the impact of the recent amendments to subsection 55(2) on the relevance of the safe- income and foreign affiliate surplus regimes, as well as some of the similarities and differences between these regimes.

20 J.R. Robertson, “Recent Developments in Federal Taxation,” in Report of Proceedings of the Thirtieth Tax Conference, 1978 Conference Report (Toronto: Canadian Tax Foundation, 1980), 52-67; and J.R. Robertson, “An Update on a Departmental Perspective on Recent Developments in Federal Taxation” (1979) 27:4 Canadian Tax Journal 428-32. 21 Robertson, “Recent Developments in Federal Taxation,” supra note 20, at 58-59. 22 See Brender, supra note 10, at 12:7-8. 23 John R. Robertson, “Capital Gains Strips: A Revenue Canada Perspective on the Provisions of Section 55,” in Report of Proceedings of the Thirty-Third Tax Conference, 1981 Conference Report (Toronto: Canadian Tax Foundation, 1982), 81-109. 24 Brender, supra note 10, abstract. 280 n canadian tax journal / revue fiscale canadienne (2018) 66:2

Overview of t he Section 55 Amendments, and the Computation of Safe Income The recent amendments to section 55 included, but were not limited to, the restructuring of subsection 55(2) into two provisions (subsections 55(2) and (2.1)) and the addition of two new purpose tests in subparagraph 55(2.1)(b)(ii). In general, where either of these two new purpose tests is satisfied and none of the exceptions in subsection 55(2) is available, subsection 55(2) will apply with the result that an otherwise “tax-free”25 intercorporate dividend26 received by a corporation resident in Canada as part of a transaction or event, or a series of transactions or events, will be treated as a capital gain. The new purpose tests are satisfied where one of the purposes of the payment or receipt of the dividend is to effect either a significant reduction in the fair market value of any share or a significant increase in the total cost of properties of the recipient of the dividend. As a result, it is no longer as clear as it might have been in the past that a corporation resident in Canada can claim an offsetting deduction under subsection 112(1) or (2) (or subsection 138(6)) in respect of a taxable dividend received by it from another corporation resident in Canada, a taxable Canadian corporation, or a non-resident corporation described in subsection 112(2), as the case may be, including a taxable dividend received from a wholly owned subsidiary of the recipient corporation. The safe-income exception is one of the potential exceptions to the application of subsection 55(2). Whether this exception is available in respect of a particular dividend will ultimately depend upon the particular facts. In this regard, the safe- income exception essentially has two requirements.27 First, the amount of the dividend cannot exceed the amount of the income earned or realized by any corporation, after 1971 and before the “safe-income determination time”28 for the transaction or event, or series of transactions or events (herein referred to as “safe income”). Second, that safe income must reasonably be considered to contribute to the capital gain that could be realized on a disposition at fair market value, immediately before the dividend, of the share on which the dividend was received.29

25 By virtue of the ability of the corporate recipient of the dividend to claim an offsetting deduction under subsection 112(1) or (2), or subsection 138(6). 26 Other than, pursuant to the carve-out in subparagraph 55(2.1)(b)(ii), a dividend that is received on a redemption, acquisition, or cancellation of a share, by the corporation that issued the share, to which subsection 84(2) or (3) applies. 27 Paragraph 55(2.1)(c). 28 As defined in subsection 55(1). 29 The concept of safe income is generally thought of as the “income earned or realized” by the corporation during the relevant holding period, as determined under subsection 55(5). In contrast, the concept of safe income on hand is generally interpreted as the portion of the income earned or realized by the corporation during the relevant period that could reasonably be considered to contribute to the capital gain that would be realized on a disposition of the particular share. As a result, certain adjustments must be made to safe income to reflect the interaction of the foreign affiliate surplus and safe-income regimes n 281

Although it is not the purpose of this article to undertake a detailed analysis of the various issues and considerations associated with the computation of safe income (or the determination of the availability of the safe-income exception), the following points are particularly relevant for the purposes of providing context for the discussion that follows:

1. Subject to certain limitations (in particular those in paragraph 55(2.1)(c) and subsection 55(5)), for the purposes of the safe-income exception, safe income potentially includes that of the dividend payer as well as that of any corporation, domestic or foreign, in respect of which the payer is a direct or indirect shareholder.30 2. Subsection 55(5) is the starting point for determining safe income. However, the computation of safe income is also based on a significant amount of juris- prudence, as well as published administrative views of the CRA. 3. Depending on the facts, the safe-income determination time may not always be clear or obvious. More specifically, the definition of “safe-income determination time” is such that it can be triggered by the payment of a dividend as part of a series of transactions. For example, the safe-income determination time is the time immediately before the payment of the first dividend that is paid as part of a series of transactions, as opposed to the time immediately before the payment of a dividend that occurs later in the series. In the context of whether annual recurring dividends could be viewed as part of a series, the cra stated the following at the Tax Executives Institute—cra liaison meeting held on November 15, 2016:

In a recent ruling, the cra took the view that regular, recurring annual dividends would not, in the circumstances of the ruling request, be part of a series of transactions. Accordingly, a ruling confirmed that the safe income determination time in respect of the first and second annual dividends will be immediately before each such dividend.31 4. Even if it can be established that there is sufficient safe income at the safe- income determination time, the safe-income exception may not always be available. For example, assuming that it is determined that one of the purpose tests in paragraph 55(2.1)(b) applies to a particular dividend, the exception is not available if

portion thereof that remains on hand; for example, safe income is reduced for taxes, dividends previously paid, etc. For the purposes of this article and except as otherwise specifically noted herein, the concepts of safe income and safe income on hand are simply referred to as “safe income.” 30 In particular, paragraph 55(2.1)(c) refers to the amount of the income earned or realized by “any” corporation. 31 CRA document no. 2016-0672321C6, November 15, 2016. 282 n canadian tax journal / revue fiscale canadienne (2018) 66:2

a. there is no capital gain on the share on which the dividend was paid, including, for instance, where there is an accrued loss on the share or its fair market value is equal to the adjusted cost base32 of that share; or b. the safe income cannot reasonably be considered to contribute to the capital gain that could otherwise be realized on a disposition at fair market value, immediately before the payment of the dividend, of the share on which the dividend was received.33 5. If a dividend received by a corporation exceeds the safe income of the dividend-paying corporation at the safe-income determination time, on the surface it appears that the entire dividend may potentially be recharacterized as a capital gain under subsection 55(2). Under the former version of paragraph 55(5)(f ), the recipient corporation was provided with the discretion to designate any portion of the dividend as a separate taxable dividend. By making such a designation, the recipient corporation could ensure that the portion of the dividend equal to safe income was not recharacterized as a capital gain under subsection 55(2). As a consequence of the amendments to section 55, paragraph 55(5)(f ) now applies automatically so that the portion of a dividend that is considered to have been paid out of safe income will be treated as a separate dividend without the need for the recipient corporation to make a designation.

Overview of t he Computation of Foreign Affiliate Surplus Pools Although it is not the purpose of this article to undertake a detailed analysis of the various issues and considerations associated with the computation of the surplus pools of a foreign affiliate, the following points are particularly relevant for the purposes of providing context for the discussion that follows:

1. A corporation resident in Canada is generally required to maintain separate exempt surplus (deficit), hybrid surplus (deficit), taxable surplus (deficit), hybrid underlying tax, and underlying foreign tax pools (herein collectively referred to as “surplus pools”) for each of its foreign affiliates. These pools are computed in accordance with part lix of the regulations. The computation involves a complex interaction of Canadian and foreign tax principles. 2. Although a detailed discussion of the components of the surplus pools of a foreign affiliate is beyond the scope of this article, the following is a non- exhaustive, layman’s summary of the primary components of a foreign affiliate’s surplus (deficit) pools:

32 As defined in subsection 248(1). 33 Paragraph 55(2.1)(c). interaction of the foreign affiliate surplus and safe-income regimes n 283

a. Exempt surplus (deficit)34 of a foreign affiliate includes, among other things, i. earnings35 from an active business where the affiliate is a resident of a designated treaty country36 and those earnings are considered to be attributable to activities occurring in such a country; ii. the non-taxable portion of any capital gains (losses) as well as the taxable portion of any capital gains (losses) arising on a disposition by the affiliate of excluded property,37 other than any capital gains (losses) that are included in computing the affiliate’s hybrid surplus (deficit);38 and iii. exempt surplus dividends received from another foreign affiliate of the corporation. b. Hybrid surplus (deficit) of a foreign affiliate includes, among other things, i. gains or losses of the affiliate from certain dispositions of partnership interests and shares of another foreign affiliate of the taxpayer; and ii. hybrid surplus dividends received from another foreign affiliate of the corporation. c. Taxable surplus (deficit)39 of a foreign affiliate includes, among other things, i. earnings from an active business where the affiliate is not a resident of a designated treaty country and/or such earnings are not considered to be attributable to activities occurring in such a country; ii. fapi; iii. the taxable portion of any capital gains (losses) arising on a disposition by the affiliate of property that is not excluded property; and iv. taxable surplus dividends received from another foreign affiliate of the corporation. d. Hybrid underlying tax40 includes the portion of any income or profits tax paid to a government of a country by the affiliate that can reasonably be regarded as having been paid in respect of an amount that is included in computing the affiliate’s hybrid surplus. e. Underlying foreign tax41 includes the portion of any income or profits tax paid to a government of a country by the affiliate that can reasonably be

34 As defined in regulation 5907(1). 35 As defined in regulation 5907(1). 36 As defined in regulations 5907(11) through (11.2). 37 As defined in subsection 95(1). 38 As defined in regulation 5907(1). 39 As defined in regulation 5907(1). 40 As defined in regulation 5907(1). 41 As defined in regulation 5907(1). 284 n canadian tax journal / revue fiscale canadienne (2018) 66:2

regarded as having been paid in respect of an amount that is included in computing the affiliate’s taxable surplus. 3. Surplus pools are relevant in characterizing a dividend paid by a foreign affiliate of a corporation resident in Canada, including the determination of a deduction that may be claimed by that corporation under subsection 113(1) in respect of such a dividend so received.42 4. Subject to the potential application of the so-called 90-day rule,43 the characterization of a dividend paid by a foreign affiliate is based on the surplus pools of that affiliate at the time of payment. For these purposes, the exempt surplus (deficit), taxable surplus (deficit), and underlying foreign tax pools of a foreign affiliate at a particular time only includes earnings (losses) and income taxes paid in respect of taxation years of a foreign affiliate that ended before that time. In contrast, amounts are added to, or deducted from, the hybrid surplus (deficit) and hybrid underlying tax of a foreign affiliate at a point in time and without regard to completed taxation years of the foreign affiliate. 5. Subject to certain potential elective mechanisms,44 the characterization of a dividend paid by a foreign affiliate45 is generally determined by an ordering rule that takes into account any deficit that the affiliate may have in any of its surplus pools.46 A dividend paid by a foreign affiliate is typically deemed to have been first paid out of the affiliate’s available exempt surplus, if any, at the time of payment.47 If the amount of the dividend exceeds the affiliate’s

42 Surplus pools are also potentially relevant in determining the Canadian tax consequences that arise on a disposition of foreign affiliate shares. Section 93 effectively permits, in certain circumstances, all or a portion of a taxable capital gain otherwise realized by a Canadian corporation or another foreign affiliate of the Canadian corporation (and in certain cases, a partnership) on a disposition of shares of a foreign affiliate to instead be treated as a dividend (which is potentially eligible for an offsetting deduction under section 113). The application of section 93 is elective where a Canadian corporation disposes of shares of a foreign affiliate, but is automatic in the case where one foreign affiliate of a Canadian corporation disposes of shares of another affiliate and the vendor affiliate otherwise recognizes a gain in respect of that disposition. 43 Regulation 5901(2)(a). 44 See, for example, subsection 90(3), regulation 5901(1.1), and regulation 5901(2)(b). 45 In general, for the purposes of the Act, an amount is deemed under subsection 90(2) to be a dividend paid or received, as the case may be, on a share of a class of the capital stock of a foreign affiliate of a taxpayer if that amount was paid pro rata on all of the shares of that class that were issued and outstanding at the time of the distribution and that distribution did not involve a redemption, acquisition, or cancellation of any of the issued and outstanding shares of the affiliate. 46 Regulation 5901(1). 47 Regulation 5901(1)(a). For this purpose, the amount of the whole dividend (as defined in regulation 5907(1)) considered to have been paid from exempt surplus is the lesser of (1) the amount of the whole dividend and (2) the amount by which the affiliate’s exempt surplus exceeds the total of the affiliate’s hybrid and taxable deficits (if any). interaction of the foreign affiliate surplus and safe-income regimes n 285

available exempt surplus, the excess is considered to have been paid out of the affiliate’s available hybrid surplus, if any, at the time of payment.48 If the amount of the dividend exceeds the total of the affiliate’s available exempt and hybrid surplus, the excess is considered to have been paid out of the affiliate’s available taxable surplus, if any, at the time of payment.49 To the extent that a dividend exceeds the total of a foreign affiliate’s available exempt, hybrid, and taxable surplus pools, the excess is deemed to have been paid out of “preacquisition” surplus.50 6. Preacquisition surplus is not defined in the Act or the regulations. It is a notional account for which no calculations are maintained. Conceptually, preacquisition surplus represents not only the shareholder’s investment in the shares of the affiliate, but also the shareholder’s accrued gain in respect of those shares to the extent that such gain is not otherwise reflected, for one reason or another, in the underlying exempt, hybrid, and/or taxable surplus pools of the affiliate.

Conceptual Comparison of t he Foreign Affiliate Surplus and Safe-Income Regimes From a purely theoretical perspective, the safe-income and foreign affiliate surplus regimes share a similar computational objective: to determine, or reflect, the earnings of a particular corporation that are available for distribution to the corporation’s shareholders. However, as might be expected, there are many potentially significant differences in the computation mechanics that apply under the two regimes. Table 1 attempts to highlight, on a conceptual basis, some of the similarities and differences between the frameworks of these regimes.

48 Regulation 5901(1)(a.1). For this purpose, the amount of the whole dividend considered to have been paid from hybrid surplus is limited to the lesser of (1) the amount by which the whole dividend exceeds the portion of the dividend that was considered to have been paid out of the exempt surplus of the affiliate and (2) the amount by which the affiliate’s hybrid surplus exceeds the affiliate’s exempt deficit and/or taxable deficit (except to the extent that such taxable deficit is not otherwise offset by the exempt surplus, if any, of the affiliate at that time). 49 Regulation 5901(1)(b). For this purpose, the amount of the whole dividend considered to have been paid from taxable surplus is limited to the lesser of (1) the amount by which the whole dividend exceeds the portion of the dividend that was considered to have been paid out of the exempt and/or hybrid surplus pools of the affiliate and (2) the amount by which the affiliate’s taxable surplus exceeds the affiliate’s exempt deficit and/or hybrid deficit (except to the extent that any such deficit is not otherwise offset by the exempt or hybrid surplus, if any, of the affiliate at that time). 50 Regulation 5901(1)(c). 286 n canadian tax journal / revue fiscale canadienne (2018) 66:2

Table 1 Comparison of the Safe-Income and Foreign Affiliate Surplus Regimes

Key attribute Safe-income regimea Foreign affiliate surplus regime

Primary purpose Exception to the anti-avoidance Characterization of a dividend rule in subsection 55(2). paid by, or potentially the conversion (under section 93) into a dividend of all or a portion of the proceeds of disposition, and ultimately a capital gain, otherwise arising on a disposition of shares of, a foreign affiliate of a corporation resident in Canada. Elective versus Generally automatic to the Generally automatic, subject to automatic application extent of safe income.b certain elective mechanisms.c to a particular dividend Restrictions on ability Safe income is computed in Except where section 93 applies to access respect of a particular share, to convert into a dividend all or a and, subject to considerations portion of the proceeds of that could arise where there is disposition, and ultimately a more than one class of issued capital gain, otherwise arising on and outstanding shares,d a disposition of the shares of a generally accrues on a pro rata foreign affiliate, the ability of a basis.e The safe-income Canadian corporate taxpayer to exception is not, however, access surplus pools of the available in respect of a affiliate for the purposes of particular dividend if (1) there characterizing a dividend paid by is no capital gain on the share that affiliate is not necessarily on which the dividend was paid, restricted to the taxpayer’s or (2) the safe income cannot proportionate “share” of those reasonably be considered to pools (as determined on the basis contribute to the capital gain of the taxpayer’s surplus that could otherwise be realized entitlement percentage or some on a disposition at fair market other basis). For the purposes of value, immediately before the characterizing a dividend that is dividend, of the share on which paid on a particular share of a the dividend was received.f foreign affiliate of a Canadian corporate taxpayer, there is no requirement that the surplus pools of the affiliate must reasonably be considered to be attributable to a gain in respect of the share. (Table 1 is continued on the next page.) interaction of the foreign affiliate surplus and safe-income regimes n 287

Table 1 Continued

Key attribute Safe-income regimea Foreign affiliate surplus regime

Consolidated versus Generally computed on a Computed on a foreign-affiliate- single-entity consolidated basis (that is, by-foreign-affiliate basis, except approach includes safe income of the where section 93 is applicable to dividend payer and potentially deem all or a portion of the that of any corporation, proceeds of disposition, and domestic or foreign, in respect ultimately a capital gain, of which the payer is a direct or otherwise arising in respect of a indirect shareholder).g In foreign affiliate share to be a contrast to the foreign affiliate dividend. surplus regime, there is no explicit ownership threshold that must be met in order for a particular corporation’s safe income to be potentially eligible for inclusion. Computation period Determined on a share-by- Subject to certain rollover (subject to comments share basis and generally provisions and/or certain changes below regarding the limited to the income earned or in the surplus entitlement inclusion of realized by the corporation and percentage of the Canadian stub-period earnings) its subsidiaries during the corporate taxpayer in respect of periodh throughout which the the foreign affiliate,j restricted to particular shareholder owned the period that begins with the the particular share (except first day of the taxation year of potentially where the share is the affiliate in which it last acquired in a rollover became a foreign affiliate of the transaction), and ending with taxpayer and that ends at a the safe-income determination particular time.k time.i Inclusion of Includes income earned or Exempt earnings and taxable stub-period earnings realized by a corporation up to earnings are added to the the safe-income determination respective surplus pools of the time.l foreign affiliate only at the end of its taxation year. As a result and unless the so-called 90-day rule in regulation 5901(2)(a) is applicable in respect of a dividend that is paid by the foreign affiliate, current-year earnings of the affiliate are generally not relevant for the purposes of characterizing a dividend.m (Table 1 is continued on the next page.) 288 n canadian tax journal / revue fiscale canadienne (2018) 66:2

Table 1 Continued

Key attribute Safe-income regimea Foreign affiliate surplus regime

Computational Income earned or realized by a Complex interaction of Canadian framework corporation is generally to be and foreign tax principles.o determined under Canadian tax principles using division B of part I of the Act.n After determining the income earned or realized by a corporation, certain adjustments (for example, reductions for taxes and dividends previously paid) are made to arrive at safe income on hand (the portion of the income earned or realized by the corporation during the relevant period that could reasonably be considered to contribute to the capital gain that would be realized on a disposition of the particular share). Treatment of Notional, or so-called The treatment of notional “notional” income phantom, income is generally income and expenses is not and expenses included in computing the safe entirely clear in all cases, but income of a corporation to the general prevailing practice is to extent that such amounts are exclude such amounts in included in computing the computing surplus pools.q corporation’s net income for tax purposes under division B of part I of the Act.p Treatment of deficits Because safe income is The deficit of one affiliate is computed on a consolidated generally not relevant in basis, the deficit (loss) of a computing the surplus pools of particular corporation is another affiliate since surplus generally relevant in computing pools are computed on a the safe income in the shares of foreign-affiliate-by-foreign- the dividend-paying affiliate basis. However, corporation.r depending on where a deficit is located in a chain of affiliates, it could be relevant in certain circumstances.s (Table 1 is continued on the next page.) interaction of the foreign affiliate surplus and safe-income regimes n 289

Table 1 Continued

Key attribute Safe-income regimea Foreign affiliate surplus regime

Currency Safe income is to be computed The surplus pools of a foreign in Canadian currency,t but affiliate of a corporation resident presumably subject to any in Canada are to be maintained functional-currency election on a consistent basis from year to that has been made by a year in the currency of the relevant corporation to country in which the affiliate is compute its Canadian tax resident, or any currency that the resultsu in a different currency.v corporation demonstrates to be reasonable in the circumstances.w Treatment of Generally deducted in Subject to certain potential permanent non- computing safe income on exceptions,y generally deducted deductible outlays hand.x in computing surplus pools. (including taxes) a For a more detailed explanation of one or more of the key attributes in the context of the safe-income regime, see, for example, Duncan Osborne, “Practical Issues in Computing Safe Income,” in Report of Proceedings of the Fifty-Fourth Tax Conference, 2002 Conference Report (Toronto: Canadian Tax Foundation, 2003), 42:1-23. b Paragraph 55(5)(f ). For dividends paid before April 21, 2015, paragraph 55(5)(f ) applied only to the extent that one or more designations were filed in respect of the dividend. c Such elective mechanisms include, for example, subsections 93(1) and 90(3), and regulations 5901(1.1) and (2)(b). d See, for example, CRA document no. 2015-0593941E5, December 3, 2015, in which the CRA commented on the allocation of safe income on hand to preferred shares that are entitled to discretionary dividends. e The Queen v. Nassau Walnut Investments Inc., 97 DTC 5051 (FCA). f Paragraph 55(2.1)(c). g Ibid. Also see D & D Livestock Ltd. v. The Queen, 2013 TCC 318, at paragraph 14, and Lamont Management Ltd. v. Canada, 2000 CanLII 17131 (FCA). h Paragraphs 55(5)(a) through (c). i See John R. Robertson, “Capital Gains Strips: A Revenue Canada Perspective on the Provisions of Section 55,” in Report of Proceedings of the Thirty-Third Tax Conference, 1981 Conference Report (Toronto: Canadian Tax Foundation, 1982), 81-109. j See, generally, regulation 5905, including the definition of “surplus entitlement percentage” in regulation 5905(13). k See the preambles of the definitions of “exempt surplus,” “hybrid surplus,” “taxable surplus,” “hybrid underlying foreign tax,” and “underlying foreign tax” in regulation 5907(1). l Paragraph 55(2.1)(c) and subsection 55(5). m It should also be noted that the 90-day rule in regulation 5901(2)(a) is not applicable for the purposes of determining any dividend that is deemed to arise under section 93. n Paragraphs 55(5)(b) and (c). Also see Canada v. Kruco Inc., 2003 FCA 284, and D & D Livestock Ltd., supra note g, at paragraph 11. o Regulation 5907. p See, for example, Kruco Inc., supra note n; Income Tax Technical News no. 33, September 16, 2005; and Income Tax Technical News no. 34, April 27, 2006. (Table 1 is concluded on the next page.) 290 n canadian tax journal / revue fiscale canadienne (2018) 66:2

Table 1 Concluded q See, for example, CRA document no. 2000-0021965, January 9, 2001, regarding stock option expense, and CRA document no. 9709005, March 2, 1998, regarding Mexican inflationary adjustments. However, the technical basis for the CRA’s views is not clear. Alternatively, in certain circumstances, regulation 5907(2) might, directly or indirectly, require notional amounts to be excluded in computing a foreign affiliate’s earnings from an active business where those earnings are required to be computed under foreign tax principles. r There is an exception to this position where there is a deficit in the shares of a foreign affiliate and the deficit does not affect the gain or value of the dividend-paying corporation. See, for example, paragraph 55(5)(d) and Canada v. Brelco Drilling Ltd., 1999 CanLII 8151 (FCA). s A detailed discussion of when a deficit of a foreign affiliate could be relevant is beyond the scope of this article. However, examples of situations that may be encountered include (1) a “blocking deficit” in a higher-tier foreign affiliate that prevents the distribution of all or a portion of the exempt (or other) surplus pool from a lower-tier affiliate to the Canadian corporate shareholder of that higher-tier affiliate; (2) certain reorganizations, such as a merger described in regulation 5905(3) or a transaction that triggers the application of the fill-the-hole rules (as described in regulation 5905(7.2)); or (3) by virtue of the consolidation mechanism in regulation 5902(1)(a), a deficit of a foreign affiliate (other than that of a foreign affiliate at the bottom of a chain of foreign affiliates) that is relevant for the purposes of a transaction that results in the application of subsection 93(1). t Subsection 261(2). u As defined in subsection 261(1). v Subsections 261(3) and (5). However, the computation of safe income in the context of a relevant functional-currency election appears to be unclear in numerous respects since section 261 (in particular, subsections 261(5) and (7)) does not make any reference to section 55. Also, where a shareholder of a particular corporation has not made a functional- currency election, but the particular corporation has, it is not entirely clear whether the CRA’s views in document no. 2016-0642111C6, May 26, 2016, could apply to require the safe income of the corporation to be computed in Canadian currency, on the basis that the computation of safe income relates to the Canadian tax results of the shareholder in respect of a particular taxable dividend. w Regulation 5907(6). x See, for example, Deuce Holdings Limited v. The Queen, 97 DTC 921 (TCC), and Income Tax Technical News no. 37, February 15, 2008. Also see CRA document no. 2016-0672321C6, November 15, 2016. y In computing a foreign affiliate’s earnings from an active business, regulation 5907(2) applies to reduce those earnings for any permanently non-deductible outlays incurred by the affiliate only if those earnings are required to be computed under subparagraph (a)(i) or (ii) of the definition of “earnings” in regulation 5907(1). interaction of the foreign affiliate surplus and safe-income regimes n 291

Intersection of t he Foreign Affiliate Surplus and Safe-Income Regimes There are two basic scenarios in which the foreign affiliate surplus and safe-income regimes potentially intersect. One involves a “sandwich structure,” in which a foreign affiliate of a corporation resident in Canada owns shares of another Can- adian corporation. The second, and by far the more common, scenario involves the typical “outbound structure,” in which a corporation resident in Canada owns shares of a foreign affiliate and that corporation is itself owned by another Canadian corporation.

Sandwich Structure An interesting dichotomy arises within the foreign affiliate regime in a situation where a foreign affiliate of a corporation resident in Canada owns shares of another Canadian corporation. The shares of such a corporation do not constitute excluded property of the affiliate. As a result, any gain arising to the affiliate on a disposition of those shares should be included in computing that affiliate’s fapi. Assuming that the shares are held on account of capital, the taxable and non-taxable portions of the capital gain should be included in computing the affiliate’s exempt and taxable surplus pools respectively. Notwithstanding that the shares of a Canadian corporation constitute a “fapi property” of the foreign affiliate, all or a portion of a taxable dividend51 that the foreign affiliate receives from the Canadian corporation could, in certain circumstances, be treated differently. The reason for this is that such a dividend does not constitute fapi of the foreign affiliate, and is included in computing the foreign affiliate’s exempt surplus, to the extent that the dividend would otherwise have been deductible under section 112 had it instead been received by the Canadian taxpayer in respect of the foreign affiliate.52 Since the treatment of such a dividend depends on its deductibility under section 112, the recent amendments to section 55 may also be relevant in determining the treatment of that dividend in computing the foreign affiliate’s fapi and surplus pools. In other words, if it were determined that one or both of the purpose tests in paragraph 55(2.1)(b) was met in respect of a particular dividend, and subsection 55(2) applied to deem the dividend to be a capital gain, it appears that the dividend would be included as a capital gain in computing the foreign affiliate’s fapi (with the taxable and non-taxable portions of that gain being included in computing the affiliate’s taxable and exempt surplus respectively) unless, or to the extent that, it could be established that the safe-income exception was applicable. As a result, the ability to wholly include such a dividend in computing the exempt surplus of a foreign affiliate is no longer as clear as it might have been in the past.

51 As defined in subsection 248(1). 52 Paragraph (c) of element A of the definition of “foreign accrual property income” in subsection 95(1) and paragraph (v) of element A of the definition of “exempt surplus” in regulation 5907(1). 292 n canadian tax journal / revue fiscale canadienne (2018) 66:2

Outbound Structure In the context of a typical outbound structure in which there is a chain of Canadian corporations located above one or more foreign affiliates, the foreign affiliate surplus and safe-income regimes are linked by paragraph 55(5)(d). The safe income of a corporation for a period ending at a time when that corporation was a foreign affiliate of another corporation is deemed, under paragraph 55(5)(d), to be the lesser of

1. the fair market value at that time of all of the issued and outstanding shares of the capital stock of that affiliate and 2. the amount that would be the “tax-free surplus balance”53 (tfsb) of that affiliate in respect of the other corporation at that time.

There are a couple of potentially important observations to be made about the mechanics for determining the safe income of a foreign affiliate. First, the requirement to determine and document the fair market value of the shares of a particular foreign affiliate at each and every safe-income determination time could be challenging or, at the very least, administratively burdensome. Second, the tfsb of a foreign affiliate is intended to represent the amount of the affiliate’s surplus pools, as determined on a stand-alone or non-consolidated basis, that could be repatriated by the affiliate to the particular corporation resident in Canada as a dividend54 if the affiliate were directly owned by that corporation.55 In other words, and building on the concepts discussed earlier in this article, the tfsb of a foreign affiliate effectively includes

1. the amount, if any, by which the affiliate’s exempt surplus in respect of the corporation exceeds the total of the affiliate’s hybrid and taxable deficits;56 2. the affiliate’s hybrid surplus in respect of the corporation, as determined after deducting any applicable exempt and/or taxable deficits of the affiliate, but

53 As defined in regulation 5905(5.5) and as if the Act were read without reference to regulation 5905(5.6). 54 For this purpose, it appears that any foreign tax (for example, withholding tax) that would otherwise be payable in respect of such a repatriation is not taken into account. In contrast, if an actual dividend had been paid and withholding tax incurred, it appears that such tax would be deducted, under general principles, in computing the amount of safe income on hand. 55 The TFSB of a foreign affiliate is, by virtue of regulation 5905(5.6), typically determined on somewhat of a “consolidated” basis to the extent that the affiliate would have otherwise received (on the basis of certain assumptions) a dividend that was paid from the “net surplus” of another foreign affiliate. However, since paragraph 55(5)(d) provides that the TFSB of a foreign affiliate is to be determined without reference to regulation 5905(5.6), solely for purposes of determining the safe income of a foreign affiliate under paragraph 55(5)(d), the TFSB of a particular foreign affiliate is effectively computed on a stand-alone basis as if that affiliate were directly owned by the particular corporation resident in Canada. 56 Regulation 5905(5.5)(a). interaction of the foreign affiliate surplus and safe-income regimes n 293

only if the entire amount of that hybrid surplus balance is fully sheltered by hybrid underlying tax (that is, if that balance were paid out as a hybrid surplus dividend received by the corporation from the affiliate, provided that such dividend would be fully deductible by the corporation under paragraph 113(1)(a.1));57 and 3. the affiliate’s taxable surplus in respect of the corporation, as determined after deducting any applicable exempt and/or hybrid deficits of the affiliate, but only to the extent that such taxable surplus, or portion thereof, is sheltered by the grossed-up amount of underlying foreign tax (that is, the taxable surplus, or portion thereof, could be received as a dividend free of Canadian tax by the Canadian corporate taxpayer).58

After determining the safe income of a foreign affiliate in accordance with paragraph 55(5)(d), it is necessary to apply the general principles, some of which are described in table 1, that are relevant for the purposes of computing safe income on hand.59 This is necessary in order to determine the portion of the income earned or realized by the foreign affiliate during the relevant period that could reasonably be considered to contribute to the capital gain that would be realized on a disposition of the particular share.

Selected Anomalies, Issues, and Planning Considerations Although the safe-income and foreign affiliate surplus regimes share a similar computational objective, they do not mesh seamlessly because of differences in the detailed computation mechanics applied to each regime. As a result, it is perhaps not surprising that the interaction of these regimes could raise planning considerations or have unexpected consequences in certain circumstances. The following is an overview of some of the considerations that can arise, from time to time, in the context of the interaction of the two regimes. Some of these considerations are more commonly known and have been at least partially addressed in other papers60 (and for the sake of completeness are included here), whereas others are perhaps less obvious or less well known.

57 Regulation 5905(5.5)(a.1). As a result, for the purposes of computing the TFSB (and thus, the safe income) of a particular affiliate, it is not possible to include either the “taxable” or the “non-taxable” portion of any capital gain that is included in computing the foreign affiliate’s hybrid surplus unless the gain is fully sheltered by sufficient hybrid underlying tax. 58 Regulation 5905(5.5)(b). 59 For example, if it were determined that a particular foreign affiliate incurred a permanently non-deductible outlay that was not, for one reason or another, required to be deducted in computing the affiliate’s surplus pools, it appears that, to the extent that such surplus is relevant for the purposes of computing safe income, such an outlay would be deducted for the purposes of computing safe income on hand (given that a similar outlay by a corporation resident in Canada is generally required to be deducted in computing safe income on hand). 60 See, for example, Dehsi and Ferhmann, supra note 11. 294 n canadian tax journal / revue fiscale canadienne (2018) 66:2

Use of the Book Depreciation Election To Compute Active Business Earnings Many foreign jurisdictions provide for accelerated rates of depreciation. In some jurisdictions, such as the United States, the claiming of depreciation expense is effectively mandatory because there is no provision for carryforward (or carryback) of the expense deduction, or a portion thereof, if it is not fully claimed in the particular taxation year. Although depreciation claims reduce the taxable income, and thus taxes payable, of the affiliate in the foreign jurisdiction, they also effectively reduce the affiliate’s earnings from an active business, potentially reducing the affiliate’s ability to pay dividends from its exempt surplus (and ultimately its tfsb). Since depreciation is a non-cash expenditure, significant depreciation deductions arising under foreign tax principles could cause a foreign affiliate to have distributable cash that exceeds its earnings. In certain situations, the making of an election under regulation 5907(2.1) may increase a foreign affiliate’s ability to pay a dividend from its exempt surplus, thus potentially increasing its tfsb for the purposes of a safe-income determination. Generally, if a regulation 5907(2.1) election is made, a foreign affiliate’s earnings from an active business carried on in a designated treaty country are computed using financial statement depreciation as opposed to foreign tax depreciation.61 If this election is made at a time when the accumulated depreciation for financial statement purposes is less than it is for tax purposes, the affiliate’s earnings (and thus its exempt surplus pool, and ultimately its tfsb) will be increased by the difference between these amounts. Of course, whether or not an immediate benefit results from making such an election in terms of an increase in the affiliate’s earnings (and thus its exempt surplus and tfsb) depends on various factors, including the age and nature of the foreign affiliate’s property. Generally, once the regulation 5907(2.1) election is made, it may not be revoked. Owing to the complexities associated with computing the ongoing adjustments required under regulation 5907(2.1), and given that the election is made with retroactive effect,62 it is normally recommended that a Canadian corporation delay

61 For greater certainty, this election is available only where the active business earnings of a foreign affiliate for a particular taxation year are required to be determined under foreign (as opposed to Canadian) tax principles. In other words, the requirements must be met for those earnings to be computed under subparagraph (a)(i) of the definition of “earnings” in regulation 5907(1). 62 If the election is not made in the first year in which the affiliate carries on the active business and in which the affiliate was a foreign affiliate of the electing corporation resident in Canada (or a non-arm’s-length Canadian corporation), pursuant to regulation 5907(2.2) the earnings of the affiliate for the year in which the election is made are adjusted to include the total of all amounts that would have been included under regulation 5907(2.1) in computing the affiliate’s earnings for each prior year if the election had been made in the first year. In other words, a regulation 5907(2.1) election has retroactive effect if it is not made in the first year in which the affiliate carried on the particular active business and was a foreign affiliate of the electing interaction of the foreign affiliate surplus and safe-income regimes n 295 making the election in respect of a particular foreign affiliate for as long as possible. However, to the extent that the affiliate’stfsb is relevant for the purposes of a particular safe-income determination time, in certain circumstances it might make sense to make this election sooner than would normally be the case if the potential increase to the affiliate’stfsb is significant.63

Treatment of Deficits of a Foreign Affiliate As noted above, the safe income of a particular foreign affiliate is defined in paragraph 55(5)(d) as the lesser of the tfsb and the fair market value of all the issued and outstanding shares of the foreign affiliate. However, it is not clear whether in determining safe income on hand of a particular foreign affiliate, the deficit of a lower-tier foreign affiliate could reduce the safe income of the particular foreign affiliate even if the fair market value of its shares exceeds itstfsb .64 If a deficit of a lower-tier foreign affiliate does indeed reduce safe income on hand, the location of the deficit affiliate in a chain of foreign affiliates does not seem to be relevant— contrary to the general framework of the foreign affiliate surplus regime, in which so-called blocking deficits in a chain of foreign affiliates are generally more relevant (and potentially more problematic) than deficits in foreign affiliates that are at the bottom of the ownership chain. Paragraph 55(5)(d) was amended in 2011. Arguably, one of the purposes of the fair market value test in paragraph 55(5)(d) was to capture deficits of lower-tier foreign affiliates, and therefore no adjustment should be required for lower-tier deficits in computing safe income on hand.

Foreign Exchange Rate To Be Used for Inclusion of Foreign Affiliate Surplus Pools in Safe Income It appears that the general practice is for the surplus pools of a foreign affiliate to be included in computing safe income using the applicable foreign exchange rate at the safe-income determination time.65

corporation. However, instead of an adjustment to the earnings of the affiliate to reflect the election for each prior year, a single cumulative adjustment is made to the affiliate’s earnings for the taxation year in which the election is made. 63 Generally, for the election to be effective beginning with a particular taxation year of a foreign affiliate, pursuant to regulation 5907(2.6) the election must be filed on or before the day that is the earliest day on or before which the Canadian corporation, or any other corporation of which the affiliate is a foreign affiliate, is required to file a tax return for its taxation year following the taxation year in which the taxation year of the affiliate ends. Depending on the facts, the CRA may possibly accept a late-filed regulation 5907(2.1) election. Regulation 600 specifically provides that the CRA has the discretion to accept a late-filed regulation 5907(2.1) election under subsection 220(3.2). 64 See, for example, Canada v. Brelco Drilling Ltd., 1999 CanLII 8151 (FCA). Also see CRA document no. 2001-0093385, October 5, 2001, and paragraph 55(5)(d). 65 CRA document no. 9414365, October 27, 1994. 296 n canadian tax journal / revue fiscale canadienne (2018) 66:2

Maximization of a Particular Foreign Affiliate’s Tax-Free Surplus Balance As mentioned earlier, for the purposes of determining safe income, the tfsb is computed on a foreign-affiliate-by-foreign-affiliate basis. Since this approach is different than the consolidated approach used when computing the tfsb for other purposes (or determining the amounts in surplus pools that are available for the purposes of a subsection 93(1) election), the location of a particular foreign affiliate in a Canadian multinational’s foreign affiliate group is perhaps of less importance from the perspective of computing safe income. Nevertheless, because of the foreign- affiliate-by-foreign-affiliate approach in determining the tfsb, in certain circumstances planning may be necessary in order to maximize the tfsb amount that is available for the purposes of determining safe income of a foreign affiliate. For example, as noted earlier, the tfsb includes hybrid surplus of a particular foreign affiliate only if such surplus is “fully sheltered” (that is, that hybrid surplus, if paid by the affiliate as a dividend to the Canadian corporation, would be fully offset by a deduction under paragraph 113(1)(a.1)). As a result, if one foreign affiliate in a chain has a hybrid surplus pool with an insufficient amount of hybrid underlying tax applicable66 whereas another affiliate has an excess of hybrid underlying tax applicable, that pool is not available for the purposes of the determination of safe income, notwithstanding that the hybrid surplus would be fully sheltered with hybrid underlying tax applicable if the surplus pools of those two affiliates were combined. Accordingly, and if circumstances permit, it might be prudent to try to “match up” or blend, prior to the safe-income determination time and through the payment of an interaffiliate dividend or other mechanism, the applicable excess hybrid underlying tax of one affiliate with a “low-taxed” hybrid surplus pool of another so as to permit the inclusion of the hybrid surplus in the tfsb computation for the purposes of determining safe income.67

Payment of Dividends by a Foreign Affiliate The payment of an exempt, hybrid, and/or taxable dividend by a foreign affiliate generally reduces the surplus pool(s) of the affiliate from which that dividend is paid.68 To the extent that such a dividend is received by another foreign affiliate of

66 As defined in regulation 5907(1). 67 However, if this dividend were determined to be part of the series of transactions or events that includes the receipt of a dividend by a corporation resident in Canada, the safe-income determination time would be the time immediately before the payment of the dividend between the foreign affiliates as opposed to the time immediately before the receipt of the dividend by the corporation resident in Canada. If this were indeed the case, such planning would not be effective from the perspective of determining the amount of safe income that is potentially available in respect of the dividend that was received by the corporation resident in Canada. 68 Since the characterization of a dividend paid by a foreign affiliate to its shareholder depends on the balances of the affiliate’s exempt, hybrid, and/or taxable surplus pools at the time that the dividend is paid, the affiliate’s earnings for the year in which the dividend is paid are generally interaction of the foreign affiliate surplus and safe-income regimes n 297 that Canadian corporation, it retains its characterization for the purposes of determining the surplus pools of that other affiliate. Notwithstanding that the payment of a dividend from the exempt, hybrid, and/or taxable surplus of the affiliate reduces the surplus pool(s) from which the dividend is considered to be paid, depending on the facts this may not necessarily result in a corresponding reduction in the affiliate’s tfsb.69 Similarly, whether or not the receipt of such a dividend increases the safe income of the recipient also depends on the facts. For example, the receipt of such a dividend by another foreign affiliate of the taxpayer could increase that affiliate’stfsb by an amount up to the dividend amount, depending on

n the surplus pool of the distributing affiliate from which the dividend was paid70 and n the surplus balance (deficit) of the recipient.71

not included in this determination. The so-called 90-day rule in regulation 5901(2)(a) is an exception to this principle. Where a foreign affiliate pays a dividend at any time in its taxation year that is more than 90 days after that year begins, the 90-day rule generally provides that the portion of the dividend that would otherwise be deemed to have been paid out of the affiliate’s preacquisition surplus is instead deemed to have been paid out of the affiliate’s exempt, hybrid, and/or taxable surplus to the extent that such portion would have been so characterized if the dividend had been paid immediately after the end of the particular year. Further, solely for the purposes of determining the surplus pools of the foreign affiliate payer, the dividend is deemed to have been paid as a separate dividend immediately following the end of the year. Despite the treatment to the payer affiliate, the 90-day rule does not change the time at which the dividend is received by the recipient. As a consequence of these deemed timing mechanics, the application of the 90-day rule is not entirely clear in the context of the determination of safe income, and in particular the TFSB of the dividend-paying foreign affiliate. If the timing mechanics of the rule are interpreted literally, it appears that there could be an increase in the safe income of the recipient at the time of a dividend payment without a corresponding reduction in the TFSB of the payer until immediately after the end of its taxation year. Such a result, if ultimately determined to be correct, would seem to be inappropriate from a policy perspective since it could result in a doubling-up of the dividend amount for the purposes of computing safe income of the group, subject to making any necessary adjustments under the general principles that apply in determining safe income on hand. 69 Whether or not, or to what extent, the TFSB of the dividend-paying foreign affiliate is reduced by a dividend generally depends on the surplus pool from which the dividend is paid. For example, as noted elsewhere in this article, only fully sheltered hybrid surplus is included in computing a foreign affiliate’s TFSB. Thus, if a foreign affiliate pays a dividend from hybrid surplus, the TFSB of that affiliate is not reduced unless its hybrid surplus was fully sheltered (that is, there was sufficient hybrid underlying tax applicable). Furthermore, the TFSB of the foreign affiliate recipient of such a dividend will not be increased unless that affiliate has sufficient excess hybrid underlying tax applicable from another source. 70 See supra note 69. 71 For example, if the recipient has an exempt deficit at the time of receipt of an exempt surplus dividend from another affiliate (and such a deficit is, for one reason or another, not relevant for the purposes of determining safe income, in accordance with the court’s rationale in Brelco 298 n canadian tax journal / revue fiscale canadienne (2018) 66:2

Given the potential for an unexpected (or possibly adverse) result that a dividend can have on the determination of the tfsb of the foreign affiliate recipient and/or payer of the dividend, consideration could be given to timely filing a regulation 5901(2)(b) election (or, alternatively, a subsection 90(3) election) to sidestep the normal surplus ordering rule and treat the distribution as a preacquisition surplus dividend (or a qualifying return of capital, in the case of a subsection 90(3) election). Generally speaking, if the recipient of the dividend is the Canadian corporation, it appears that such a dividend should increase the safe income of that corporation regardless of the surplus pool from which the dividend is considered to be paid. The reason for this is that safe income is generally determined using division b of part i of the Act, and any section 113 deduction claimed by the corporation in respect of a dividend that is received from a foreign affiliate occurs under division c of part i.72

Stub-Period Earnings of a Foreign Affiliate As noted above, one potentially significant difference between the safe-income and foreign affiliate surplus regimes involves the time of inclusion for certain amounts in the calculation of surplus and safe income. It is generally accepted that safe income includes income earned or realized up to the safe-income determination time. In contrast, the exempt and taxable earnings of a foreign affiliate (which include, among other things, a foreign affiliate’s earnings [loss] from an active business) are added to the affiliate’s exempt and taxable surplus pools, respectively, only at the end of its taxation year. Since the tfsb of a particular foreign affiliate, as well as the characterization of a dividend paid by that affiliate to its shareholder, depends on the balances of the affiliate’s exempt, hybrid, and taxable surplus pools at that time—that is, the time for determining the tfsb or the time that the dividend is paid, as the case may be—the affiliate’s earnings (loss) for the taxation year that includes that time are generally not included in this determination.

Drilling, supra note 64), such a deficit could wholly or partially offset (or absorb) the exempt surplus so distributed, whereas that exempt surplus might have been included in computing safe income had it not been distributed by the other affiliate. Similarly, only fully sheltered hybrid surplus, and the portion of taxable surplus that could, in effect, be paid to the Canadian corporate taxpayer free of Canadian tax, is included in computing a foreign affiliate’s TFSB. As a result, it is possible in certain circumstances (including, for example, the existence of a deficit in the recipient, the quantum of the recipient’s taxable surplus pool in relation to its underlying foreign tax pool, the quantum of the recipient’s hybrid surplus pool in relation to its hybrid underlying tax pool, etc.), that the dividend payment may or may not result in an increase in the TFSB of the recipient. This could be the case regardless of whether the TFSB of the distributing affiliate is correspondingly reduced by the dividend. 72 To the extent that a particular dividend is considered to have been paid to the Canadian corporation out of the preacquisition surplus of the foreign affiliate, it is not clear whether such a dividend would be excluded from the corporation’s safe income, taking into account any adjustments that are made to determine the corporation’s safe income on hand. interaction of the foreign affiliate surplus and safe-income regimes n 299

Nevertheless, the cra has previously indicated, albeit prior to the changes to paragraph 55(5)(d), that it will administratively accept the inclusion of stub-period earnings in computing the safe income of a corporation when the affiliate was a foreign affiliate of another corporation.73

Amounts That Accrue Prior to Becoming a Foreign Affiliate of the Taxpayer In computing the surplus pools of a foreign affiliate of a Canadian corporation, any capital gain (loss) realized by the affiliate in respect of a disposition of a capital property is not to include the portion of that gain (loss) that can reasonably be considered to have accrued prior to the affiliate’s becoming a foreign affiliate of that corporation.74 A similar rule applies to the extent that a foreign affiliate earns fapi in a particular taxation year that can reasonably be considered to have accrued prior to the affiliate’s becoming a foreign affiliate of the corporation.75 As a result, such amounts should not be included in computing the tfsb, and thus the safe income, of a foreign affiliate. In contrast, there is no similar rule that carves out from a foreign affiliate’s earnings from an active business (and ultimately that affiliate’s surplus pools) the portion of any recognized gains (losses) that are on account of income and can reasonably be considered to have accrued prior to the affiliate’s becoming a foreign affiliate of the taxpayer (“preacquisition earnings”).76 This could include, for example, recapture of depreciation expense that relates to the period prior to the affiliate’s becoming a foreign affiliate of the taxpayer. Thecra has previously taken the position, in respect of the former version of paragraph 55(5)(d), that the preacquisition earnings of a foreign affiliate that were included in determining the surplus pools of a foreign affiliate should be excluded in determining safe income on hand, on the basis that the inclusion of such amount would, in effect, result in double-counting of the adjusted cost base of the shares of the foreign affiliate as safe income.77

73 CRA document nos. 9523075, February 23, 1996, and 9611945, December 9, 1996. 74 Paragraphs 95(2)(f ) and (f.1) and regulation 5907(5). More specifically, paragraph 95(2)(f.1) provides that “in computing an amount described in paragraph [95(2)](f ) in respect of a property or a business, there is not to be included any portion of that amount that can reasonably be considered to have accrued, in respect of the property (including . . . any property for which the property was substituted) or business, while no person or partnership that held the property or carried on the business was a specified person or partnership in respect of the taxpayer referred to in paragraph (f ).” 75 See supra note 74. 76 See, for example, the definitions of “earnings,” “exempt earnings,” and “net earnings” in regulation 5907(1). 77 CRA document no. 2013-0499141I7, March 14, 2014. 300 n canadian tax journal / revue fiscale canadienne (2018) 66:2

Ownership of Shares of a Foreign Affiliate Through a Partnership The potential for anomalies and interpretational challenges to arise from the ownership of the shares of a foreign affiliate through a partnership is well documented.78 Not surprisingly, it appears that there is the potential for similar anomalies and interpretational challenges to arise in the context of the determination of safe income. By way of background, subsection 93.1(1) permits a Canadian corporation that is a partner of a partnership to look through the partnership to determine whether a direct or indirect subsidiary of that partnership is considered to be a foreign affiliate of that corporation. However, this lookthrough provision applies only for limited purposes—namely, the purposes of the provisions listed in subsection 93.1(1.1). Generally speaking, the effect of the combined application of subsections 93.1(1) and (1.1) is that a direct or indirect subsidiary of a partnership could be considered to be a foreign affiliate of a Canadian corporation that is a partner of a partnership for the purposes of computing the surplus pools of the affiliate vis-à-vis that corporation, and ultimately the characterization of a dividend that is made by, or deemed to have been made by, the affiliate. Except for certain limited exceptions, such a subsidiary of a partnership is generally not considered to be a foreign affiliate of the Canadian corporation for other purposes of the Act, including, for example, the computation of fapi. As mentioned, the surplus pools of a foreign affiliate are potentially included in computing safe income under paragraph 55(5)(d). However, this provision applies only if a non-resident corporation is considered to be a foreign affiliate of another corporation. Although subsection 93.1(1) permits looking through a partnership for the purposes of determining whether a direct or indirect subsidiary of the partnership is considered to be a foreign affiliate of a Canadian corporation that is a partner of a partnership, this lookthrough mechanism applies only for the purposes of the provisions that are specifically identified in subsection 93.1(1.1), and paragraph 55(5)(d) is not one of those provisions. The reason for this omission is unclear. Caution should therefore be exercised in determining the safe income of one or more foreign affiliates if there is a partnership located in the ownership chain. Nevertheless, in the absence of being able to rely on subsection 93.1(1.1) to look through a partnership for the purposes of applying paragraph 55(5)(d), in certain circumstances it still might be possible to include the income earned or realized by a subsidiary of a partnership in computing the safe income of the dividend payer that is a partner of the partnership, or possibly even a direct or indirect shareholder of that partner. The reason for this is that, as noted above, paragraph 55(2.1)(c) merely refers to the amount of the income earned or realized by any corporation. As a result, it is conceivable that safe income could include the income earned or realized by a corporation that is neither a corporation that is a resident of Canada

78 See, for example, Marc Ton-That and Melanie Huynh, “Inconsistent Treatment of Partnerships in the Foreign Affiliate Rules,” inReport of Proceedings of the Sixty-First Tax Conference, 2009 Conference Report (Toronto: Canadian Tax Foundation, 2010), 24:1-66. interaction of the foreign affiliate surplus and safe-income regimes n 301 nor a foreign affiliate of a corporation (that is, a foreign non-affiliate). Notwith- standing that paragraphs 55(5)(b), (c), and (d) only contemplate the computation of the safe income of a corporation resident in Canada and a foreign affiliate of a corporation, in Lamont Management Ltd. v. Canada,79 the Federal Court of Appeal concluded that the reference to “any corporation” in subsection 55(2) (the predecessor of paragraph 55(2.1)(c)) can include a foreign non-affiliate.

Retroactive Adjustments to Surplus Pools As noted, safe income is determined at the safe-income determination time. As a result, if the Canadian tax return originally filed by a particular corporation is subsequently amended, it might be necessary to make one or more corresponding retroactive adjustments to the safe income that is determined as at a particular safe- income determination time. Likewise, since an affiliate’s tfsb is potentially relevant in computing safe income, any retroactive adjustment that is required to be made to the surplus pools of a foreign affiliate could correspondingly affect itstfsb otherwise determined at the safe-income determination time. In some cases, such adjustments may only represent timing differences that are of no significant consequence. This might be the case, for example, if no dividends have been paid between the effective time of the retroactive adjustment and the time that the need for that adjustment becomes known. However, if one or more dividends have been paid in that intervening period, either by a foreign affiliate or by a Canadian corporation in the chain, it is possible that such adjustments could affect the quantum of safe income that was otherwise believed to have been available in respect of the dividend at the safe-income determination time. There are a few instances where it might be necessary to make a retroactive adjustment to the surplus pools of a foreign affiliate. For instance, this might be the case if a foreign affiliate is solely engaged in carrying on an active business and the income (loss) reported by the affiliate in its income tax return filed with the tax authorities in its country of residence is amended or otherwise adjusted. More specifically, the earnings (loss) of a foreign affiliate for a taxation year from an active business are, subject to certain adjustments, generally required to be computed under the income tax law of the country in which the affiliate is resident if that affiliate is required by that law to compute its income or profit.80 Another situation in which a retroactive adjustment to the surplus pools of a foreign affiliate could arise is where the affiliate is a member of a group of two or more foreign affiliates of a Canadian corporation and the tax liabilities of those members are determined on a combined or a consolidated basis. In such a situation, regulation 5907(1.1) applies in a very mechanical fashion to determine the tax liabilities of each member of the group for the purposes of computing the surplus pools

79 2000 CanLII 17131 (FCA). 80 See the definitions of “earnings” and “loss” in regulation 5907(1). 302 n canadian tax journal / revue fiscale canadienne (2018) 66:2 of those members.81 The overall effect of the adjustments to be made under regulation 5907(1.1) is to reduce (increase) the surplus pools of each member of the combined or consolidated group to reflect the amount of tax that has been borne or paid by (or refunded to) that member. In doing so, regulation 5907(1.1) requires that certain adjustments be made to the surplus pools of a foreign affiliate for a prior taxation year.82 In the absence of any tax compensatory payments being made, and assuming that the entire tax liability of the combined or consolidated group has been paid by one affiliate (“the primary affiliate”), in principle regulation 5907(1.1)(a) only adjusts the surplus pools of the primary affiliate. As the initial step in this process, regulation 5907(1.1)(a) disregards the actual income or profits tax liability (refund) of the combined or consolidated group for the particular taxation year, and then requires the stand-alone tax liabilities of each of the primary affiliate and the other members of the group (each of which is referred to as a “secondary affiliate”) to be computed on the basis that each of those affiliates had instead filed its own tax return and had no other taxation year.83 This deemed income or profits tax liability, which is wholly included in computing the surplus pools of the primary affiliate, is subsequently adjusted to reflect any current and/or prior-year losses of the primary affiliate and/or a secondary affiliate that are used in computing the tax liability of the combined or consolidated group.84 To the extent that prior-year losses of the primary affiliate and/or a secondary affiliate are used, this adjustment (increase) is reflected in the surplus pools of the primary affiliate at the end of the taxation year in which the loss arose.85 Thus, if a loss arose in one year and is used by the combined or consolidated group in the immediately following year, the surplus pools of the primary affiliate are increased at the end of the prior taxation year as opposed to the year in which the loss is used. Prior-year adjustments to surplus pools of the primary and/or secondary affiliates can also arise under regulation 5907(1.1)(b) in certain circumstances where tax compensatory payments are made between the affiliates. In this regard, if the primary affiliate compensates a secondary affiliate for the use of a loss of the secondary affiliate that arose in a prior taxation year, the surplus pools of both the primary and secondary affiliates are adjusted only as at the end of the year of the loss regardless of when the payment is made.86 Similarly, if a secondary affiliate makes a

81 Likewise, regulation 5907(1.092) is applicable where a foreign affiliate of a taxpayer is liable for income or profits tax in respect of that affiliate’s share of the income (loss) of another foreign affiliate that is fiscally transparent under the laws of a foreign country. Regulation 5907(1.092) operates in a manner that is in many ways similar to regulation 5907(1.1). 82 See supra note 81. 83 Regulations 5907(1.1)(a)(i) through (iv). 84 Regulation 5907(1.1)(a)(v). 85 Regulations 5907(1.1)(a)(v)(C) through (D). 86 Regulation 5907(1.1)(b)(ii). interaction of the foreign affiliate surplus and safe-income regimes n 303 tax compensatory payment to the primary affiliate in respect of a particular taxation year, the surplus pools of both the primary and secondary affiliates are adjusted only at the end of that taxation year.87 This is the case regardless of whether that payment occurs in the particular taxation year or in a subsequent year. On a consolidated basis, the retroactive increase of the surplus pools of one affiliate under the mechanics of regulation 5907(1.1)(b) should generally be offset against the corresponding decrease to the surplus pools of another. However, since the tfsb is, for the purposes of computing safe income, determined on a foreign-affiliate-by-foreign- affiliate basis, there could be situations where these retroactive adjustments have an impact on the overall determination of safe income at a particular safe-income determination time, especially where the adjustments involve the hybrid surplus pools of the affected affiliates.

Reorganizations Involving the Shares or Property of Foreign Affiliates There are a number of provisions that could apply to adjust a foreign affiliate’s surplus pools, as computed vis-à-vis a particular Canadian corporation, in a situation where there is an acquisition or disposition of the shares of the affiliate. For example, an adjustment to the surplus pools of a foreign affiliate may be required where there is a change in the Canadian corporation’s surplus entitlement percentage88 in respect of a foreign affiliate that arises from an acquisition or disposition of the share of a relevant foreign affiliate.89 In concept, the purpose of this adjustment is to try to ensure that the Canadian corporation’s percentage share of the surplus pools of a foreign affiliate remains the same before and after a particular acquisition or disposition. This is necessary because the surplus pools of a foreign affiliate, as computed vis-à-vis a particular Canadian corporation, are to be computed on a 100 percent basis without regard to the ownership percentage, or the surplus entitlement percentage, of the corporation in that affiliate. There are also continuity provisions that attempt to preserve the surplus pools of a foreign affiliate in respect of certain qualifying restructuring transactions, such as when a foreign affiliate is merged with another foreign affiliate90 or when a foreign affiliate with net surplus91 is dissolved into another foreign affiliate.92 Similarly, there are numerous provisions that could apply to deny or suppress the recognition of amounts that would otherwise be included in computing the surplus

87 Regulation 5907(1.1)(b)(i). 88 As defined in subsection 95(1) and regulation 5905(13). 89 Regulation 5905(1). 90 Regulation 5905(3). 91 As defined in regulation 5907(1). 92 Regulation 5905(7). 304 n canadian tax journal / revue fiscale canadienne (2018) 66:2 pools of a foreign affiliate in respect of certain intragroup transfers of shares93 or property94 of a foreign affiliate. Since all of these provisions are relevant in determining the surplus pools, and ultimately the tfsb, of a particular foreign affiliate, any proposed reorganization or transfer of the shares or property of a foreign affiliate should be carefully considered prior to undertaking the reorganization or transaction. Also, to the extent that a foreign affiliate realizes an amount, such as a capital loss, in respect of such a reorganization or transaction (or otherwise) that is suspended or not recognized for the purposes of computing the affiliate’s surplus pools, it is necessary to determine whether that amount represents an adjustment to the affiliate’s tfsb for the purposes of determining safe income on hand. Accordingly, advance planning is necessary to help minimize the risk of any adverse or unanticipated foreign affiliate surplus and/or safe-income consequences arising therefrom. For example, in contrast to the Canadian domestic context, many types of stop- loss rules are “turned off” for the purposes of computing any loss to be included in computing the surplus pools of a foreign affiliate from a disposition of excluded property.95 Likewise, it is possible in certain circumstances that a shareholder affiliate could realize, in computing its hybrid surplus (deficit), a loss on a tax-deferred dissolution of another foreign affiliate of the Canadian corporation where the requirements of subparagraph (iii) of element b of the definition of “hybrid surplus” in regulation 5907(1) are met. Depending on the facts, the inclusion of a loss in computing the surplus pools of a foreign affiliate may ultimately reduce thetfsb , and ultimately the safe income, of a relevant affiliate. There is also an anti-avoidance rule in regulation 5907(2.02) that is applicable to certain “tax-motivated” intragroup dispositions of property (other than money) that are intended to increase the exempt earnings of the transferor foreign affiliate. Where the conditions for this rule are met, amounts that would otherwise be included in computing the affiliate’s exempt earnings (and thus exempt surplus) are reclassified as taxable earnings (and thus taxable surplus).

Two Pe as in a Pod? Given the history of the foreign affiliate surplus and safe-income regimes, it not unexpected that there are some potentially significant differences and similarities between these regimes. However, there are two fundamental questions that remain unanswered: Why did the government decide not to legislate a detailed definition for the computation of safe income in the first place, and why does such an approach continue to this day?

93 See, for example, paragraphs 95(2)(c) and (f ), and regulation 5907(5). 94 See, for example, paragraph 95(2)(f ) and regulations 5907(2)(f ) and (j), (5), and (5.1). 95 Subsections 13(21.2), 40(3.6), and 93(4), and paragraphs 14(12)(a), 18(13)(a), 40(2)(e.1), (e.2), and (g), and 40(3.3)(a). Also, a rollover is not available under paragraph 95(2)(c) in respect of the transfer of one foreign affiliate to another if there is an inherent loss in the shares of the transferred affiliate. interaction of the foreign affiliate surplus and safe-income regimes n 305

Unfortunately, the answers to these questions are not entirely clear. That said, it seems likely that the legislators’ approach to subsection 55(2) and the provisions related thereto represents an attempt to strike a balance between the simplicity of a vaguely worded set of anti-avoidance rules and the certainty (and accompanying complexity) provided by a comprehensive set of mechanical rules.96 Although the rules regarding the computation of safe income may be fuzzy or vague, the underlying policy intent is clear. The fuzziness, and ultimately the lack of certainty, associated with the rules designed to prevent inappropriate capital gains stripping, and the relief that is potentially available under the safe-income exception, also implicitly act as a self-policing deterrent to taxpayers. As evidenced by the historical review presented earlier in this article, the government’s battle against stripping transactions perceived to be abusive not only has been long but also has evolved as laws have changed and taxpayers have adapted to those changes. Thus, it is entirely conceivable that the legislators may have been concerned that the adoption of a detailed legislative approach to the computation of safe income would create too much certainty for taxpayers. A risk in providing taxpayers with a detailed road map of the rules is that this gives them the opportunity to poke and prod at those rules in an endless quest to exploit gaps and loopholes, ultimately leading to the introduction of even more complex legislation designed to plug those gaps and loopholes. Also, the active role that Revenue Canada took, early on, in publishing its own administrative practices regarding the calculation of safe income arguably made the drafting of detailed legislation less of a priority for the Department of Finance. Although the foreign affiliate surplus and safe-income regimes share a similar computational objective, in respect of the determination of certain amounts that are distributable by a corporation to its shareholders, the safe-income regime has, at least historically, served a very different purpose as compared with that of the foreign affiliate surplus regime. At its heart, the safe-income regime is part of a set of anti- avoidance rules designed to prevent capital-gains-stripping transactions that the government considers to be inappropriate. In contrast, the foreign affiliate surplus regime could be viewed, at least in some respects, as being favourable to taxpayers. More specifically, under the right facts and circumstances, the regime permits a Canadian corporation to treat all or a portion of an actual or deemed dividend received from a foreign affiliate as a tax-free exempt surplus dividend as opposed to either a capital gain or a dividend that is wholly or partially taxable. Accordingly, it only stands to reason that it would be in the government’s interest to legislate detailed mechanical rules for the computation of the surplus pools of a foreign affiliate. In the absence of such detailed rules, the policing of the foreign affiliate surplus regime would likely have been much more challenging for the tax authorities. Despite the fact that the foreign affiliate surplus and safe-income regimes have historically served very different purposes, it is arguable that the recent changes to

96 See, for example, Ronald J. Farano, “Subsection 55(2)—The Moving Goalposts,” in 1993 Ontario Tax Conference (Toronto: Canadian Tax Foundation, 1993), 1A:1-46, at 1A:2-5. 306 n canadian tax journal / revue fiscale canadienne (2018) 66:2 subsection 55(2) have resulted in those purposes becoming much more closely aligned. Of course, the safe-income regime still remains a fundamental part of a targeted set of anti-avoidance rules. However, the effect of the recent changes to subsection 55(2) is that the ability of a corporation resident in Canada to claim an offsetting deduction under section 112 in respect of an intercorporate dividend is no longer as clear as it might have been in the past In effect, the default rule has become that a dividend paid by one Canadian corporation to another is now fully taxable to the recipient without the ability to claim an offsetting deduction under section 112, unless the recipient can establish that an exception applies to permit such a deduction. Depending on the facts, one of these exceptions could be that the dividend was paid out of the safe income. Likewise, a dividend paid by a foreign affiliate to a corporation resident in Canada is generally taxable to the recipient unless the recipient is a Canadian corporation and can demonstrate that the dividend was paid out of a type of surplus pool that permits all or a portion of that dividend to be deductible under section 113. In other words, regardless of whether a Canadian corporation receives a dividend from another corporation resident in Canada or a foreign affiliate, the Canadian corporation is, subject to certain exceptions, generally required to prepare calculations to support the claiming of an offsetting deduction under section 112 or 113. As a result, it is arguable that one of the purposes of the safe-income regime has now evolved into one that is more akin to that of the foreign affiliate surplus regime—that is, a calculation function that is necessary to support the quantum of an offsetting deduction against dividend income. Given that the purposes of the safe-income and foreign affiliate surplus regimes appear to have become more closely aligned, it is submitted that it is once again time to consider legislating a more comprehensive set of rules for the computation of safe income. Not only would a detailed legislative definition provide taxpayers with more certainty when attempting to claim a deduction under section 112 that relies on the safe-income exception, but it would also reduce the potential for the backlog of resource- and time-consuming disputes between taxpayers and the government that uncertainty inevitably breeds. In drafting a detailed legislative framework for the computation of safe income, the legislators could consider, at least as a starting point, looking to the foreign affiliate surplus regime as a precedent. For instance, the foreign affiliate surplus regime has, among other things, clear rules dealing with the computation of amounts to be included in surplus pools and whether, or to what extent, there is a continuity of surplus pools in the event of a reorganization and/or an ownership change. Although the foreign affiliate surplus regime is certainly complex, such complexity might be preferable to many taxpayers as compared with the uncertainty that arises from relying on the cra’s non-binding and fact-specific published administrative practices or views concerning the computation of safe income. Regardless of whether the foreign affiliate surplus regime is used as a precedent in legislating a detailed calculation of safe income, it is important that the computation of foreign affiliate surplus and safe income be harmonized. Such harmonization interaction of the foreign affiliate surplus and safe-income regimes n 307 would reduce complexity and provide more certainty, not only to taxpayers but also to the cra in administering the Act.

Conclusion The recent amendments to section 55 place a bigger magnifying glass on the computation of safe income of a corporation. Since safe income can potentially include surplus pools of a foreign affiliate, there is even more importance in not only regularly updating and maintaining both safe-income and foreign affiliate surplus pools, but also understanding the interaction of the foreign affiliate surplus and safe- income regimes. These regimes share many common aspects. In addition to certain conceptual similarities, the computation of the surplus pools of a foreign affiliate, like that of the safe income of a corporation, is often not an exact science. Furthermore, under both regimes, there are many exceedingly complicated and unexpected interpretational issues that require the use of appropriate professional judgment. This complexity is further compounded by the limited administrative and judicial interpretational guidance available in respect of these each of these regimes. There are also many notable differences that can be difficult to identify without an intricate understanding of both regimes. Since each of these regimes was designed for a specific legislative purpose, it is not surprising that there are differences in their detailed mechanics. Furthermore, the detailed mechanics of, or administrative practices involving, one regime may not always be consistent with the objectives of the other regime. Given that each of the safe-income and foreign affiliate surplus regimes is, in and of itself, a separate specialty requiring in-depth knowledge, the interaction of these regimes can be particularly challenging. To meet this challenge in the context of the recent amendments to section 55, it is prudent for tax practitioners to better understand the similarities, differences, and interactions of the two regimes. In the meantime, and in light of the closer alignment of the purposes of the foreign affiliate surplus and safe-income regimes resulting from the recent amendments to section 55, it is perhaps time to question, once again, whether the absence of a detailed legislative framework for the computation of safe income is good policy. As an integral part of such a review, it is important that consideration be given to better harmonizing the foreign affiliate surplus and safe-income regimes. canadian tax journal / revue fiscale canadienne (2018) 66:2, 309 - 47

VAT/GST Thresholds and Small Businesses: Where To Draw the Line?

Yige Zu*

Précis N’importe quelle taxe d’affaires comporte nécessairement pour les petites entreprises des frais d’observation plus élevés tandis que les administrations fiscales doivent engager des frais administratifs plus élevés pour ces petites entreprises, tout en tirant d’elles de faibles recettes fiscales nettes. En réaction à cette situation, de nombreux pays qui lèvent une taxe à valeur ajoutée (tva) adoptent un seuil d’enregistrement pour soustraire du régime de tva les petites entreprises pour lesquelles le fardeau de l’observation serait le plus onéreux, et fixe ce seuil à un niveau qui ne nuit pas beaucoup aux recettes de tva. Cependant, la distinction créée par le seuil entre les entreprises exclues du régime de tva et celles qui y sont entièrement intégrées donne lieu à des distorsions de concurrence; ces distorsions diminuent pour certaines entreprises lorsqu’elles choisissent volontairement de s’enregistrer, tandis qu’elles incitent certaines autres entreprises à adopter un comportement qui leur permettra de rester sous le seuil d’enregistrement. Qu’il soit sous la forme de contraintes commerciales, de division d’entreprise ou de sous-déclaration des revenus, ce comportement entraîne de plus amples distorsions et menace le recouvrement des recettes. Cet article examine les principaux défis et considérations dans l’établissement d’un seuil d’enregistrement de la tva et les conséquences de l’adoption de ce seuil. Deux questions liées au choix du seuil se rapportent à l’utilisation par certains pays de régimes transitoires de subvention pour réduire la discontinuité fiscale pendant que les entreprises passent au plein régime de tva, et à l’utilisation par d’autres pays de régimes spéciaux de concession pour réduire le fardeau de l’observation des petites entreprises lorsqu’elles sont assujetties à la tva. D’autres pays choisissent d’utiliser une frontière entre deux régimes différents, mais voisins, plutôt qu’un seuil d’enregistrement. Dans ce genre de régime, les entreprises qui sont au-dessus de la frontière sont assujetties au plein montant de tva, et les petites entreprises qui sont sous la frontière paient une taxe sur le chiffre d’affaires. La taxe sur le chiffre d’affaires est parfois présentée comme un régime visant la simplification, parfois l’accroissement des recettes. Il se peut que les avantages des régimes de simplification et de transition soient exagérés, et que l’on ne reconnaisse pas suffisamment les coûts et les distorsions

* Lecturer in Taxation, University of Exeter Business School; lecturer in Law, University of Exeter Law School (e-mail: [email protected]). I thank Rita de la Feria, Rick Krever, and Judith Freedman for comments on earlier drafts of this article. Thanks are also due to the UK Office of Tax Simplification for assistance with the collection of data. The final article reflects my own views.

309 310 n canadian tax journal / revue fiscale canadienne (2018) 66:2 qui en découlent. Un examen de ces régimes sera utile aux décideurs politiques qui envisagent de réformer leur régime ou d’en adopter un nouveau.

Abstract Inherent features of any business tax are the imposition of relatively higher compliance costs on small businesses and higher administrative costs faced by tax authorities in respect of these firms, allied with low net tax revenue collected from the sector.A common response in jurisdictions levying a value-added tax (vat) is the adoption of a registration threshold to remove from the formal vat system small businesses for which the compliance burden would be most onerous, with the threshold being set at a level that does not seriously undermine vat revenue. However, the distinction caused by the threshold between enterprises excluded from the vat and others fully incorporated into the tax system gives rise to competitive distortions; these distortions are ameliorated for some businesses through a voluntary registration option, while other businesses are induced to adopt behaviour that will allow them to remain below the registration threshold. Taking the form of commercial restraint, enterprise splitting, or underreporting of sales, this behaviour leads to further distortions and threatens revenue collection. This article reviews the key considerations and challenges in setting a vat registration threshold and the consequences of adopting that boundary. Two issues related to the choice of threshold concern the use of “transitioning” subsidy regimes adopted in some jurisdictions to reduce the tax discontinuity as enterprises move into the full vat system and the special concessional regimes used in some countries to reduce the compliance burden that small businesses face once they are subject to vat. Another approach found in some jurisdictions is the use of a tax regime border instead of a registration threshold, with a full vat being levied on enterprises above the border and a substitute turnover (revenue) tax being imposed on small businesses below the border. The turnover tax alternative has been variously explained as a system intended to achieve simplification or revenue-raising objectives. The benefits of the transitioning and simplification schemes may be exaggerated, while their unintended costs and distortions may be insufficiently recognized.A review of these systems can provide guidance to policy makers contemplating reform or the adoption of new regimes. Keywords: VAT n GST n voluntary n registration n small business n tax simplification

Contents Introduction 311 Balancing Revenue Needs Against Administrative and Compliance Costs 313 Threshold-Related Distortions and Inefficiencies 317 Distortion of Competition 318 Behavioural Responses to a Threshold 324 Transitioning Regimes for Small Businesses Shifting into the Full VAT 332 Simplification Regimes for Small Businesses in the VAT System 334 Less Frequent Filing and Payments 335 Cash Accounting 338 Presumptive Input Tax Entitlement Regimes 339 Alternative Regimes for Small Businesses 342 Conclusion 344 vat/gst thresholds and small businesses: where to draw the line? n 311

Introduction Two features common to most value-added tax (vat) and goods and services tax (gst)1 systems are the use of a registration threshold, usually based on annual turnover (business revenue),2 to determine when businesses are subject to the tax, and one or more small business regimes that provide specific tax rules for small businesses that have crossed the registration threshold or that sit below the threshold. The common (but not universal) support for adoption of a registration threshold does not extend to the level at which it should be set. Among members of the Organisation for Economic Co-operation and Development (oecd), in 2016 registration threshold levels ranged from £83,000 (approximately Cdn$ 148,708) in the United Kingdom to zero in Spain, Turkey, Chile, and Mexico.3 Other countries are widely scattered between the two extremes, with most of them having a registration threshold far lower than that of the United Kingdom. The threshold for Canada’s gst/hst (harmonized sales tax) system, for example, is an intermediate amount of Cdn $30,000.4 As is the case with the registration threshold, there is no unanimity on the rules for small businesses. The primary purpose of a registration threshold is to reduce administrative and compliance costs; study after study shows that the administrative costs incurred by tax authorities to apply the vat to small businesses and the compliance costs incurred by small businesses are disproportionate to the revenue that these enterprises generate. The registration threshold has the effect of omitting the smallest businesses from the formal vat system. Generally, a bright-line test is used to determine when an enterprise has reached the registration threshold, though in some cases evidence of sustained turnover above the threshold is required.5 Busi- nesses with turnovers below the registration threshold are not, however, fully outside the scope of the vat. Although they are not required to register for or remit

1 In this article, the term “VAT” is used to describe the tax broadly labelled the value added tax and the goods and services tax, as it is known in Canada and a few other anglophone jurisdictions, as well as India and Malaysia. 2 In some instances, however, there are alternative measurements of the threshold. For example, in the Netherlands, the thresholds are calculated by reference to net annual VAT due. While “turnover” is commonly used in GST and VAT laws, Canadian law prescribes a somewhat circuitous path to registration. All persons making taxable supplies in Canada are required to register, unless the person is a “small supplier.” The definition of small supplier refers to consideration received for supplies. 3 See Organisation for Economic Co-operation and Development, Consumption Tax Trends 2016: VAT/GST and Excise Rates, Trends and Policy Issues (Paris: OECD, 2016), at 89. 4 The general rule is subject to exceptions with separate thresholds for public service bodies, at $50,000, and charities or public institutions, at $250,000. In addition, taxi drivers, including Uber drivers, are required to register for and remit GST/HST regardless of their turnover. 5 For example, an exemption threshold in the French VAT allows businesses that cross the threshold to retain the exemption for up to two years provided that their turnover does not exceed €90,300 (for sales) or €34,900 (for services) for more than a year during this period. 312 n canadian tax journal / revue fiscale canadienne (2018) 66:2 vat, unregistered firms bearva t on their inputs, and this cost becomes incorporated into their selling prices. Thresholds raise two concerns. The first is competitive distortion. A threshold that creates differences in terms of tax payments and compliance costs for businesses above and below the registration borderline appears to affect the relative competitive positions of the firms.6 Many countries allow businesses with turnovers below the threshold to voluntarily register for the vat in order to mitigate this problem where unregistered businesses would be prejudiced by their exclusion from the vat system. A second and more significant concern is the possible economic and revenue cost of business behaviour aimed at keeping the enterprise below the registration threshold. The observed bunching of small businesses below the registration threshold may be the result of three types of business behaviour: dishonesty and failure to report some sales, artificial separation of a business into multiple unregistered parts, and reduction of activity to reduce sales. Underreporting of sales and business splitting lead to revenue losses, while business restraint causes economic harm of particular concern to policy makers. One way of mitigating these problems is to reduce the double shock of compliance costs and higher tax faced by businesses entering the vat system. Some jurisdictions have adopted “transitioning” rules that provide subsidies to offset the costs and tax for businesses crossing the registration threshold. While adoption of a threshold can mitigate the burden of comparatively high compliance costs faced by small businesses, it cannot eliminate the problem if some small businesses remain above the threshold. The adoption of simplified vat pro- cedural rules for small businesses to address the disproportionate compliance costs borne by these enterprises is not uncommon. Measures incorporated into simplified regimes include less frequent filing (and, often, less frequent payments) and cash basis accounting. A variation allows eligible small businesses with turnovers exceeding the threshold to use a single presumptive input tax entitlement calculation in lieu of tracking all acquisitions to determine total entitlements. Separately, instead of using a registration threshold that excludes the smallest businesses entirely from the indirect tax system, some jurisdictions adopt a turnover border that distinguishes businesses subject to the full vat and those subject to an alternative lower-rate turnover tax. Registration thresholds and small business regimes work in tandem, and the absence or presence of a simplified regime will have an impact on the fullva t registration point and vice versa. However, the interaction is complex. As explained further below, regimes that reduce compliance costs for small businesses in the formal vat system may make lower thresholds feasible, but the reduced threshold will lead to higher administrative costs with little offsetting revenue. Also affecting the threshold level is the choice of treatment of small businesses outside the formal

6 Liam P. Ebrill, Michael Keen, Jean-Paul Bodin, and Victoria J. Perry, The Modern VAT (Washington, DC: International Monetary Fund, 2001), at 119-21. vat/gst thresholds and small businesses: where to draw the line? n 313 vat. As a general rule, higher-income jurisdictions favour input taxation (that is, no recovery of input tax, leaving businesses to bear the burden of the tax in the first instance) for smaller businesses below a registration threshold. The approach taken by medium- and lower-income jurisdictions is less consistent. Some subject small businesses with turnovers below the registration threshold to input taxation, while others have substituted alternative tax borders for registration thresholds and impose a lower-rate turnover tax on businesses below this boundary. The lack of agreement on registration threshold levels does not reflect the absence of any theoretical framework for policy development in this area. Rather, it reflects the dearth of clear practical guidance in current theoretical analysis of the issue and the impact of exogenous factors on vat design. Country-specific factors and domestic political considerations play crucial roles in tax design, and there is thus no one-size solution that can address all these issues. In the case of small business regimes, policy makers must address the design issues, for the most part, without the benefit of any theoretical discussion to provide guidance or a conceptual framework on the subject. It is nevertheless possible to identify the issues that should be taken into account when policy makers consider where the registration threshold should be set, whether simplified and phasing-in regimes should be available for small businesses that have crossed the registration threshold, and whether alternative small business regimes should be adopted for businesses below the threshold at which they are required to register for the full vat. An analysis of these issues can set the stage for the development of sorely needed practical guidance.

Balancing Revenue Needs Against Administrative and Compliance Costs In principle, a neutral vat will apply to all types of supplies made by all categories of suppliers. The general principle, however, neglects the uneven distribution of administrative and compliance costs and tax revenue across different sizes of businesses. Small businesses constitute a large proportion of registered persons but contribute only a small proportion of vat revenue. The sheer number of small enterprises in the vat system means that administrative resources devoted to the group are necessarily high even as the revenue collected from them is low. At the same time, compliance costs borne by the group are disproportionally high as a percentage of turnover compared to the relative cost of compliance for larger firms. In most countries, a large share of vat revenue is collected from an exception- ally small number of registrants with the highest turnovers. In the United Kingdom, for example, in 2010-11 half of the total vat revenue was paid by only 0.4 percent of vat-registered businesses, and 10 percent of the vat registrants contributed 83 percent of the total vat revenue (see figure 1).7

7 United Kingdom, HM Revenue & Customs, “Value Added Tax Factsheet,” November 22, 2011, at section 2.2. 314 n canadian tax journal / revue fiscale canadienne (2018) 66:2

FIGURE 1 Distribution of Net Value-Added Tax (VAT) Revenue by Turnover Groups in the United Kingdom, 2010-11

100 8,210 90 204,183 80 70 60 813,616 33,714 50

Percent 40 30 22,277 20 10 1,143,780 10,622 0 746 Number of traders Net VAT receipts (£ millions) Annual turnover (£) 0-100,000 100,000-1,000,000 1,000,000-50,000,000 Over 50,000,000

Source: United Kingdom, HM Revenue & Customs, “Value Added Tax Factsheet,” November 22, 2011.

Administrative efforts, however, are primarily devoted to the large group of small businesses that generate little net tax revenue. While overall vat administrative costs are sensitive to two main factors—the complexity of the tax (the use of reduced or zero rates and exemptions) and the number of vat registrants8—those costs do not fall proportionately on small and large businesses. In the United Kingdom, more than half of the cost of administering the vat can be attributed to the administration of the smallest businesses that account for the bulk of taxable persons and a tiny fraction of vat revenue collected.9 A us study considering the implications of adopting a vat in that country estimated that removal of smaller businesses from the vat to reduce the number of registrants by more than half

8 Sijbren Cnossen, “Administrative and Compliance Costs of the VAT: A Review of the Evidence” (1994) 63:12 Tax Notes 1609-26. 9 In a study looking at the 1977-78 fiscal year in the United Kingdom, Sandford estimated that 55 percent of administrative resources were allocated to 69 percent of registered businesses (under £50,000 turnover) from which less than 5 percent of revenue was collected; see Cedric Sandford, Michael Godwin, Peter Hardwick, and Michael Butterworth, Costs and Benefits of VAT (London: Heinemann, 1981), and Cedric Sandford, “The Administrative and Compliance Costs of Taxation: Lessons from the United Kingdom” (1985) 15:3 Victoria University of Wellington Law Review 199-205. vat/gst thresholds and small businesses: where to draw the line? n 315 could reduce administrative costs by one-third while reducing revenue collection by only 3 percent.10 Compliance costs borne by registered persons may be of even greater concern than administrative costs borne by the tax authority, given that compliance costs appear to be much higher.11 There is, to be sure, a risk that measurements of compliance costs are vulnerable to overestimation and might be subject to a wider margin of error than estimates of administrative costs.12 For example, it is difficult for businesses, in particular small businesses, to separate vat compliance costs from the cost of basic record-keeping and accounting activities that would be incurred in any case in the process of running the business or meeting income tax obligations.13 Also, empirical studies on compliance costs conducted in previous decades must now be read with caution, taking into account the impact that technological advances in accounting and record keeping have had on compliance costs.14 Notwithstanding these caveats, it is clear that compliance costs fall disproportionately on small businesses.15 In one study, compliance costs as a percentage of turnover were estimated to be more than 30 times greater for small businesses than for large firms.16 These heavy and disproportionate compliance burdens on small businesses raise equity concerns,17 in particular in countries where policies tend to favour small and medium-sized enterprises (smes). In terms of both administrative and compliance costs, small businesses thus present a special case in the vat. From an efficient tax design perspective, even more important than the absolute cost of small business administration and compliance is its magnitude relative to the revenue collected from these enterprises.18 As discussed

10 United States, General Accounting Office,Value-Added Tax: Administrative Costs Vary with Complexity and Number of Businesses, GAO/GGD-93-78 (Washington, DC: General Accounting Office, May 1993). 11 Sandford, “Administrative and Compliance Costs,” supra note 9, at 201. 12 Cnossen, supra note 8, at 1610. 13 Ibid.; and United Kingdom, House of Commons Treasury Committee, The Administrative Costs of Tax Compliance, Seventh Report of Session 2003-04 (London: Stationery Office, June 2004). 14 Cnossen, supra note 8, at 1610. As early as 1993, a Canadian study showed that GST compliance costs for businesses that used computerized accounting systems were 20 percent to 40 percent lower than the costs for businesses that used manual accounting systems; see Plamondon & Associates, GST Compliance Costs for Small Business in Canada: A Study for the Department of Finance, Tax Policy (Ottawa: Department of Finance, December 1993). 15 Cedric Sandford, Michael Godwin, and Peter Hardwick, Administrative and Compliance Costs of Taxation (Bath: Fiscal Publications, 1989); Cedric Sandford and John Hasseldine, The Compliance Costs of Business Taxes in New Zealand (Wellington: Victoria University of Wellington, Institute of Policy Studies, 1992); and Plamondon & Associates, supra note 14. 16 The study looked at the 1977-78 fiscal year: see Sandford, “Administrative and Compliance Costs,” supra note 9, at 201. 17 Cnossen, supra note 8, at 1619. 18 Michelle Salvail, The Goods and Services Tax: The Government’s Administration Costs (Ottawa: Library of Parliament, Parliamentary Research Branch, Economics Division, February 1994). 316 n canadian tax journal / revue fiscale canadienne (2018) 66:2 above, revenue authorities devote significant resources to apply the vat to small businesses, and these businesses in turn incur relatively high costs to comply with the law, with little net tax revenue to show for all these outgoings. In fact, the total administrative costs incurred to collect tax from small businesses and compliance costs incurred by small businesses in respect of calculating and paying the tax may well outweigh the vat revenue generated by this group of enterprises.19 The response in most countries to the high costs and limited revenue associated with the application of the vat to small businesses has been the adoption of a registration threshold to exclude a portion of small businesses from the vat system. This approach allows revenue authorities to concentrate scarce administrative resources on larger taxpayers and relieves small businesses outside the system from the burden of vat compliance. While the outcome seems not to be fully appreciated by jurisdictions with relatively low thresholds, the revenue loss resulting from even high thresholds is unlikely to be significant. With no entitlements to input tax credits, firms below the threshold are still taxed on inputs; only their final value added escapes additional taxation. Focusing on the tradeoff between revenue and administrative and compliance costs, Keen and Mintz developed a simple theoretical rule that the optimal threshold should be set at the level where the marginal revenue gains from bringing more taxpayers into the vat equals the additional administrative and compliance costs.20 The application of the rule has proved to be more complex in practice since a host of other factors also affect the choice of threshold, leading to thresholds either above or very often below the optimal threshold to which the Keen and Mintz approach would point.21 A strict balance between revenue and collection costs also provides a case for a lower threshold for sectors with higher value-added-to-sales ratios.22 A few countries (for example, Ireland and France) apply differentiated thresholds for goods and services.23 They are, however, unlikely to be models for other countries because of the practical difficulties of distinguishing between goods and services, in particular for registered persons that provide mixed supplies.24 Differentiated thresholds increase administrative and compliance costs, compromising some of the benefit of adopting a threshold in the first place.

19 Sandford and Hasseldine, supra note 15, at 120; and Ebrill et al., supra note 6, at 117. 20 Michael Keen and Jack Mintz, “The Optimal Threshold for a Value-Added Tax” (2004) 88:3-4 Journal of Public Economics 559-76. 21 For example, most of the member states in the European Union have a threshold that is lower than the theoretically optimal threshold: see Institute for Fiscal Studies, A Retrospective Evaluation of Elements of the EU VAT System: Final Report (London: IFS, December 2011), at 83. 22 Ebrill et al., supra note 6, at 119; and Keen and Mintz, supra note 20, at 563. 23 Keen and Mintz, supra note 20, at 563. 24 alan Carter, ed., International Tax Dialogue: Key Issues and Debates in VAT, SME Taxation and the Tax Treatment of the Financial Sector (Paris: OECD, International Tax Dialogue, 2013). vat/gst thresholds and small businesses: where to draw the line? n 317

An optimal threshold that balances revenue against administrative and compliance costs is inherently transitory. In theory, the threshold should shift downward—for example, if administrative costs fall as capacity grows with experience, and compliance costs decrease with advances in technology. At the same time, inflation will cause the nominal turnovers of businesses to rise while their economic size and capacities remain constant, suggesting that a rising threshold is appropriate. The United Kingdom and Canada represent examples of opposing practice. The vat threshold in the United Kingdom has typically been increased annually in line with inflation, leaving the current threshold level in nominal terms at 16.6 times that used when the vat came into effect in 1973. In contrast, Canada has never changed its threshold since the gst was introduced in 1991. Between these extremes lie countries with ad hoc threshold lifts.25 A system of continual adjustments such as that used in the United Kingdom appears to disregard the likelihood of reductions in administrative and compliance costs over time, while the effective reduction of the threshold in real terms experienced in Canada brings ever smaller businesses facing relatively higher compliance burdens into the system. A more sensible approach would be to periodically review and adjust the threshold to balance changes in the real value of money and changes in administrative and compliance costs.

Threshold-Related Distortions and Inefficiencies A registration threshold can mitigate the relatively high compliance costs that would be faced by small businesses and the disproportionate collection costs that would be borne by revenue authorities if all enterprises were subject to the vat. Provided that the threshold is set at an appropriate level, these benefits can be achieved with a minimal cost to revenue. However, there might also be economic costs resulting from the adoption of a threshold. The omission of some small businesses from the formal vat system creates a break in the vat chain if small unregistered businesses buy from or sell to registered businesses. This break may deny revenue authorities a source of information that is useful for assessment and audit purposes. In addition, the inability of small businesses to issue tax invoices could lead to a cascading problem since unrecovered vat will become another cost of acquisition for other businesses that buy from unregistered firms. Most importantly, the differential treatment of firms above and below the threshold in terms of tax payments and compliance burdens may give rise to costly distortions to the competitive positions of registered and unregistered businesses making otherwise comparable supplies, and may induce inefficient changes in business behaviour that cause further economic harm or lead to revenue losses. Policy responses to these problems could in turn yield new problems and distortions.

25 For example, the registration threshold in New Zealand was increased by 50 percent in 2009 in response to the global financial crisis. 318 n canadian tax journal / revue fiscale canadienne (2018) 66:2

Distortion of Competition A registration threshold drives a wedge between businesses that are in the vat system and businesses that are excluded from the formal vat system and are instead subject to input taxation. It is a given that greatly different tax treatment of two groups of enterprises operating within the same market will result in competitive distortions, but identifying the precise types and levels of distortions and devising responses to mitigate the negative impact of the distortions have proved difficult. Some firms left outside the vat by a registration threshold enjoy benefits from exclusion while others are prejudiced by it. On the apparent plus side of the advantage-disadvantage scale are small businesses that sell to final consumers.26 Firms that have lower inputs relative to outputs and sell primarily to final consumers are likely to enjoy competitive advantages from being input-taxed only, with their value added escaping tax. This competitive advantage will increase as the proportion of the final price attributable to a firm’s value added rises. The claimed competitive advantages in this respect may not, however, be a real- world problem, for two reasons. First, although small firms below the threshold may enjoy tax advantages, their prices at the retail stage may still be higher than those of larger firms that are able to reduce costs by exploiting proprietary information and enjoying economies of scale.27 Second, to the extent that the economic positions of firms just above and below the threshold are largely similar, the advantages and disadvantages are only a temporary phenomenon, since business sizes are not static. On the one hand, some small businesses below the threshold may grow larger and cross the threshold. On the other hand, where firms just above the threshold are disadvantaged, their turnovers may fall below the threshold owing to a reduction in sales volumes or profit margins. While the perceived benefits of escaping the full vat may prove illusory for some enterprises, two types of small firms below the threshold may be disadvantaged relative to firms subject to the vat. The first type comprises businesses that seek to make supplies to registered businesses. Small firms below the threshold are unable to issue invoices that entitle registered purchasers to input tax credits. Regis- tered businesses will thus be reluctant to purchase from unregistered firms. The second group consists of enterprises from registered businesses with high input costs relative to taxable output sales. Examples include new businesses that incur startup costs that initially exceed turnover and businesses that make zero-rated export sales. These firms might be entitled tova t refunds if they were within the vat system. To minimize the competitive disadvantage that these firms may face, many countries, including the United Kingdom, New Zealand, and Canada, allow voluntary registration by small firms with turnovers below the threshold. The option for

26 Ebrill et al., supra note 6, at 120. 27 a similar argument was made in Ebrill et al., ibid. vat/gst thresholds and small businesses: where to draw the line? n 319 voluntary registration, which removes the compliance and administrative cost savings that flow from the registration threshold, has markedly different implications for enterprises choosing to enter the tax net and revenue agencies responsible for administering the tax. Since the former group is voluntarily incurring higher compliance costs, it must be assumed that these firms believe that they will enjoy an overall economic benefit from registration after incurring new compliance costs; otherwise, they would not have elected to enter the vat regime. In contrast, from a tax collection perspective, the result may be reduced revenue and disproportionately high administrative costs. Voluntary registrations may constitute a relatively high percentage of total registrations; in several oecd countries, well over one-third of total vat registrants are voluntary registrants.28 For example, in the United Kingdom in 2016-17, 46 percent of the businesses registered for the vat operated below the threshold.29 Tax statistics in New Zealand show even higher percentages of voluntary registration in most years, ranging from a low of 45 percent to a high of 52 percent of total registrants during the period 2009-2016.30 As a consequence of the self-selection nature of voluntary registration, this group is likely to be disproportionately popu- lated by persons claiming refunds from the revenue authority rather than those paying net tax to the government. In the United Kingdom and New Zealand, for example, a sizable proportion of voluntary registrants reported nil output sales (about 25 percent and 37 percent respectively), and in both jurisdictions, vat receipts from firms at the bottom of the turnover scale are negative.31 In 2015-16, voluntary registrants in New Zealand collectively claimed a net vat refund of nz $656.5 million (approximately Cdn $573.72 million), a figure equal to 4 percent of the net vat revenue collected. The self-selection character of voluntary registration also has significant implications for vat administrative costs. Voluntary registration exacerbates considerably the disproportionate cost of administration relative to revenue that is a feature of the imposition of vat on low-turnover enterprises. Disproportionately high administrative costs follow two aspects of voluntary registration. First, there is the increased risk of refund-related avoidance schemes and outright fraudulent claims for refunds. Refund requests must be vetted carefully, drawing costly additional administrative

28 Organisation for Economic Co-operation and Development, Taxation of SMEs: Key Issues and Policy Considerations (Paris: OECD, 2009). In 2003-4, over 34 percent of all GST/HST registrants in Canada had registered voluntarily. See Canada Revenue Agency, Compendium of GST/HST Statistics: 2004 Edition (2002 tax year) (Ottawa: CRA, 2004), at table 6. 29 United Kingdom, HM Revenue & Customs, “Value Added Tax (VAT) Factsheet 2016-17,” October 31, 2017, at section 2.7. 30 New Zealand Inland Revenue, “Number of GST Filers by Turnover Band 2007 to 2016,” October 27, 2017. 31 hM Revenue & Customs, supra note 29, at section 2.7; and New Zealand Inland Revenue, “Net GST by Turnover Band 2007 to 2016,” October 27, 2017. 320 n canadian tax journal / revue fiscale canadienne (2018) 66:2 resources.32 Second, voluntary registration opens the door to whipsaw behaviour by firms that register to claim input tax credits related to startup costs and then deregister if ongoing business activities yield turnovers below the registration threshold.33 It is not uncommon for jurisdictions to require voluntary registrants to remain registered for a minimum period of time (commonly between one and five years)34 in order to discourage whipsaw behaviour. However, the limited vat that may be collected from these firms in the compulsory registration period does little to offset the high administrative costs incurred in respect of these registrants that are responsible for negative vat remittances for much of that period. An alternative, such as that used in Canada and Australia to address the risk of whipsaw behaviour while avoiding ongoing administrative costs, is to adopt a relatively short minimum registration period but effectively recapture input tax credits on deregistration.35 As noted, in theory the optimal registration threshold would be set at the level where the marginal revenue gains from bringing more taxpayers into the vat equal the increased administrative and compliance costs resulting from the additional registrations. However, the theoretical model disregards the impact on revenue and collection costs of voluntary registration. These costs must affect the choice of threshold level. The revenue yield from voluntary registrants below the threshold is likely to be less, and quite probably substantially less, than if the registration threshold were simply lowered to bring the same number of businesses into the vat system. At the same time, the increased administrative costs incurred in respect of voluntary registrants will increase revenue needs. The task of recomputing the tradeoff between revenue and collection costs to take into account the potential impact of voluntary registration becomes a multilayered undertaking. There is, however, little evidence of careful consideration of the impact of voluntary registration on the optimal threshold level. While small businesses may seek to register voluntarily for a number of reasons, including trade with registered businesses and market benefits from appearing larger than is actually the case,36 there are few comprehensive studies of the relative weighting of different factors inducing

32 Edith Brashares, Matthew Knittel, Gerald Silverstein, and Alexander Yuskavage, “Calculating the Optimal Small Business Exemption Threshold for a U.S. VAT” (2014) 67:2 National Tax Journal 283-320. 33 OECD, supra note 3, at 75. 34 For example, the minimum registration period is one year in Canada and Australia, two years in Denmark and France, and five years in Austria and Germany. There is no minimum registration period requirement in the United Kingdom and New Zealand. See OECD, supra note 3, at 75 and 89. 35 For the Canadian rule, see the Excise Tax Act, RSC 1985, c. E-15, as amended, subsection 171(3); for the Australian rule, see A New Tax System (Goods and Services Tax) Act 1999, as amended, section 138.5. 36 United Kingdom, Office of Tax Simplification,Value Added Tax: Routes to Simplification (London: Office of Tax Simplification, November 2017), at 18. vat/gst thresholds and small businesses: where to draw the line? n 321 voluntary registration.37 It is also unclear how voluntary registration rates would change if registration thresholds were raised or lowered. Cross-country comparisons reveal no discernible relationship between threshold levels and the rate of voluntary registration.38 For example, although the threshold is significantly lower in New Zealand (in 2015-16, nz $ 60,000, equivalent to approximately Cdn $ 52,425) than in the United Kingdom (£82,000 in 2015-16, or approximately Cdn $ 146,056), the voluntary registration rate in New Zealand appears to be higher (see tables 1 and 2). At the same time, Japan, with a relatively high registration threshold (approximately Cdn $113,987), has a voluntary registration rate that is less than 1 ⁄ 10 of the uk rate.39 The comparisons must be read with extreme caution, however, given the fact that the nominal monetary value of thresholds may differ substantially from the actual purchasing power parity value. In-country comparisons also yield inconsistent results. In the United Kingdom, for example, the voluntary registration rate has remained largely static in the past 10 years, although the registration threshold has been raised annually (see table 1). At the same time, following invitations from the tax authority to deregister when registration thresholds were raised twice in each of 1977 and 1978, only 1 ⁄ 5 of the taxpayers who were newly eligible for deregistration opted to move outside the vat system.40 In contrast, in New Zealand, where there were no changes in the threshold between 2009-10 and 2015-16, the voluntary registration rate declined by about 1 percent each year (see table 2). Policy makers seeking to calibrate the registration threshold in light of the impact of voluntary registration on revenue and costs thus face a quandary. On the one hand, they may realize that voluntary registration will affect tax revenue and tax collection costs, but they have no means of ascertaining the actual effects of voluntary registration. On the other hand, they may appreciate that changes to the registration threshold can influence the rate of voluntary registration, without having any means of estimating the direction in which voluntary registrations may head or the degree to which the level might change. These challenges may go some way toward explaining why vat theorists most often ignore the question of voluntary registration when discussing an optimal registration threshold. Wherever the threshold is otherwise set, adoption of a voluntary registration option to assist businesses with turnovers below the threshold

37 a recent UK study sought to rank factors using subjective data gathered from interviews with a limited sample of small businesses, but the findings are difficult to reconcile with HMRC data based on all taxable persons. See Rebecca Klahr, Lucy Joyce, Rory Donaldson, Graham Keilloh, and Cheryl Salmon, Behaviours and Experience in Relation to VAT Registration: Final Report, HM Revenue and Customs Research Report no. 446 (London: HM Revenue and Customs, November 2017). 38 Brashares et al., supra note 32, at 287. 39 OECD, supra note 28, at 122. 40 United Kingdom, HM Customs and Excise, Review of Value Added Tax, Cmnd 7415 (London: HM Customs and Excise, 1978), at 12-13. 322 n canadian tax journal / revue fiscale canadienne (2018) 66:2 46 25 83,000 2016-17 44 25 82,000 2015-16 44 25 81,000 2014-15 44 25 79,000 2013-14 43 23 77,000 2012-13 42 24 73,000 2011-12 T) Factsheet 2016-17,” October 31, 2017. 43 28 70,000 2010-11 44 30 68,000 2009-10 dded Tax ( VA alue A dded Tax 43 30 67,000 2008-9 43 30 64,000 2007-8 . . . Value-Added Tax Thresholds and Voluntary Registration in the United Kingdom, 2007-8 to 2016-17 Registration Voluntary and Thresholds Tax Value-Added Source: United Kingdom, H M Revenue & Customs, “ V percentage of total registrants percentage of total voluntary registrants oluntary registrants as a ble 1 Ta Threshold (£) V Nil turnover registrants as a vat/gst thresholds and small businesses: where to draw the line? n 323 44.92 36.39 60,000 − 656.5 2015-16 46.12 37.11 60,000 − 518.7 2014-15 47.36 37.63 60,000 − 437.3 2013-14 48.80 38.24 60,000 − 347.3 2012-13 49.45 38.65 60,000 − 228.1 2011-12 51.00 37.70 60,000 − 450.5 2010-11 51.70 38.54 60,000 − 647.5 2009-10 . . Value-Added Tax Threshold and Voluntary Registration in New Zealand, 2009-10 to 1015-16 2009-10 in New Zealand, Registration Voluntary and Threshold Tax Value-Added ST by Turnover Band Band 2007 to 2016,” October 27, 2017 and “Net G ST by Turnover Sources: New Zealand Inland Revenue, “Number of G ST Filers by Turnover 2007 to 2016,” October 27, 2017. percentage of total registrants registrants (thousands of NZ $) . as a percentage of voluntary registrants oluntary registrants as a Threshold (NZ $) . V Net payment by voluntary Nil turnover registrants ble 2 Ta 324 n canadian tax journal / revue fiscale canadienne (2018) 66:2 removes the downside of a registration threshold for enterprises that must be in the vat system for commercial reasons.41 However, as soon as voluntary registration is contemplated, its impact on revenue and administrative costs must be factored back into the equation used to identify the optimal threshold. While exact calculations may not be possible, an effective revenue service should be able to generate sufficient information for authorities to take the probable effects of voluntary registration into account when setting the registration threshold.

Behavioural Responses to a Threshold Not surprisingly, the sharp rise in tax liability and compliance costs for firms that cross the registration turnover threshold prompts behavioural responses by firms enjoying the advantages of sitting outside the formal vat system. Coexisting with the vat registration threshold is the phenomenon of business bunching, with a large number of businesses reporting turnover just below the threshold level that would require vat registration. A registration threshold thus creates a cliff-edge: a sharp increase in the number of businesses falling into the first turnover band below the threshold compared to the number in the second band below the threshold, matched by a drop to a much smaller number of businesses in the first turnover band above the threshold.42 Figure 2 illustrates this pattern based on uk data for 2014-15. Unlike the vat, the income tax generally has no registration threshold, and income tax data in some countries provide unambiguous evidence of small businesses bunching below the vat registration threshold.43 Businesses that are registered for vat purposes but that qualify for particularly generous small business concessions or beneficial regimes within the vat are equally likely to adopt behaviours to ensure that turnovers stay below the threshold at which the concessions are withdrawn.44

41 For those who do not believe that the removal of commercial disadvantage should be a priority, the case for allowing voluntary registration may be overwhelmed by the disproportionate value of administrative costs relative to tax revenue collected from small businesses; see William J. Turnier, “Designing an Efficient Value Added Tax” (1984) 39:4 Tax Law Review 435-72, at 458-60. 42 Office of Tax Simplification, supra note 36, at 6. 43 See Li Liu and Ben Lockwood, VAT Notches, Voluntary Registration and Bunching: Theory and UK Evidence, Oxford University Centre for Business Taxation Working Paper WP 16/10 (Oxford: Oxford University Centre for Business Taxation, July 2016); and Jarkko Harju, Tuomas Matikka, and Timo Rauhanen, The Effects of Size-Based Regulation on Small Firms: Evidence from VAT Threshold, VATT Working Paper no. 75 (Helsinki: VATT Institute for Economic Research, 2016). 44 a clear example of behaviour related to concessions for registered small businesses can be found in Japan, where concessions included generous tax credits; see Kazuki Onji, “The Response of Firms to Eligibility Thresholds: Evidence from the Japanese Value-Added Tax” (2009) 93:5-6 Journal of Public Economics 766-75. The Japanese concessional regime is described in Justin Dabner, “The Japanese Consumption Tax Experience: Lessons for Australia?” (2002) 5:2 Journal of Australian Taxation 185-212; and Vicki Beyer and Koji Ishimura, “The Progress vat/gst thresholds and small businesses: where to draw the line? n 325

FIGURE 2 Number of Entities by Turnover Band in the United Kingdom, 2014-15

9,000 Threshold 8,000 7,000 6,000 5,000 4,000

No. of entities 3,000 2,000 1,000 0 Turnover band (£) 81,001-82,000 82,001-83,000 83,001-84,000 84,001-85,000 85,001-86,000 86,001-87,000 87,001-88,000 88,001-89,000 89,001-90,000 90,001-91,000 91,001-92,000 92,001-93,000 93,001-94,000 94,001-95,000 95,001-96,000 80,001-81,000 66,001-67,000 67,001-68,000 68,001-69,000 69,001-70,000 70,001-71,000 71,001-72,000 72,001-73,000 73,001-74,000 74,001-75,000 75,001-76,000 76,001-77,000 77,001-78,000 78,001-79,000 79,001-80,000

Sole proprietor Partnership Incorporated company

Source: United Kingdom, Office of Tax Simplification, Value Added Tax: Routes to Simplification (London: Office of Tax Simplification, November 2017), at 7.

Businesses seeking to stay below the registration threshold may adopt one or more of three tactics:

1. deliberately holding back expansion in order to remain input-taxed suppliers; 2. splitting enterprises with total turnover above the threshold into smaller separate entities, each of which has a turnover below the threshold; or 3. fraudulently underreporting sales where actual turnover is above the threshold.

These behaviours impose economic costs on the community or lead to lost revenue, though their impact varies. While the empirical studies successfully document the

of the Japanese National Consumption Tax” (1993) 3:2 Revenue Law Journal 115-24. More recently, the UK revenue authority uncovered a scheme to split a single enterprise into hundreds of mini-companies, each of which would be able to use a concessional presumptive input tax credit regime; see Simon Goodley, “Recruitment Advisers’ Tax Scheme Liquidated After HMRC Asks Questions,” Guardian, July 10, 2017 (https://amp.theguardian.com/ business/2017/jul/10/tax-scheme-anderson-group). 326 n canadian tax journal / revue fiscale canadienne (2018) 66:2 existence of bunching,45 findings are often ambiguous in terms of identifying the extent to which each tactic is used. A uk study attributes bunching largely to output restraint and underreporting,46 while case-law evidence in the United King- dom points to splitting as a factor as well. A Finnish study suggests that restraint is the main cause of bunching in that country,47 but the methodology employed in the study to dismiss splitting as a tactic is problematic. A later UK study based on a telephone survey of a limited pool of businesses, not surprisingly, downplayed significantly the role of underreporting and artificial splitting.48 The limited and completely subjective data set, however, makes reliance on the findings vulnerable to challenge.49 In any case, uncertainty as to the prevalence of each tactic com- pounds the difficulty of devising effective responses. Much of the emphasis in the academic literature and in policy discussions is on the first tactic, with concern that restrained production will result in “significant efficiency losses”50 from underutilization of resources and reductions in potential outputs.51 Businesses may decide to restrain outputs and remain below the registration threshold in two circumstances. First, they may wish to stay out of the formal vat if they make supplies to final consumers and much of the sale price is attributable to their value added. As long as these firms are subject to input taxation only, their value added remains untaxed, providing an important competitive advantage relative to businesses in the full vat system. These enterprises may conclude that higher sales will not yield greater profits if they have to reduce markup in order to maintain attractive pricing. Second, businesses may decide that the higher compliance costs that they would incur in the vat system would outweigh the increase in net profits from additional sales. The Finnish study that found output restraint to be the primary tactic used by enterprises to stay below the threshold suggests that concern over compliance costs, as opposed to increased tax liability, was the driving factor for this behavioural response.52 The conclusion is logical given that Finland has a relatively low registration threshold (approximately Cdn $12,918) and that compliance costs relative to turnover fall disproportionately on small businesses. Underlying the view that restraint is a primary cause of bunching is an assumption that small businesses have unlimited potential and desire for growth. However,

45 See, for example, Mesay M. Gebresilasse and Soule Sow, Firm Response to VAT Policy: Evidence from Ethiopia (New York: Columbia University, August 2016). 46 Liu and Lockwood, supra note 43, at 4. 47 harju et al., supra note 43. 48 Klahr et al., supra note 37. 49 The UK government explained its reservations about the accuracy of the data based on telephone surveys in United Kingdom, HM Treasury, VAT Registration Threshold: Call for Evidence (London: HM Treasury, 2018), at 8. 50 harju et al., supra note 43, at 32. 51 Office of Tax Simplification, supra note 36. 52 harju et al., supra note 43. vat/gst thresholds and small businesses: where to draw the line? n 327 the risk of economic harm may be exaggerated. Not all businesses that value the lower tax burden or reduced compliance costs associated with being outside the formal vat will be tempted to reduce output to stay below the threshold. Businesses that genuinely lie below the threshold fall into three groups: (1) those that are not capable of further growth; (2) those that are capable of a little growth, enough to cross the threshold but not much more; and (3) those that are capable of ongoing growth. The threshold will be an inhibition only for the second group. Behaviour modifica- tion by the first group yields no benefits, while the temporary setback of tax on value added and the increase in compliance costs will not outweigh the additional profits from continuing expansion for firms in the third group. The focus, therefore, is on small firms that would have limited prospects for future growth in a no-threshold world and that are likely to enjoy market advantages by holding back expansion to stay below the threshold. The efficiency losses caused by restrained production by these firms are unlikely to be great. An alternative to restraint for an enterprise that is concerned that increased taxation or compliance costs will offset the benefits of higher sales to final consumers is to continue to pursue a total turnover above the registration threshold but to artificially split the enterprise into smaller entities, each of which operates below the threshold. In some cases, splitting arrangements may extend to enterprises with turnovers well above the threshold. Concern over this tactic is not related to economic costs but rather to lost revenue from final consumption that should be in the vat system. Theoretical discussions of the optimal registration threshold and empirical studies of bunching tend to disregard splitting as a factor contributing to bunching.53 An exception to this observation is the Finnish study on bunching referred to earlier, which explicitly considered artificial splitting as a possible explanation for bunching behaviour but found no clear evidence that splitting was significant. However, as noted above, the methodology adopted in that study to dismiss this avoidance tactic appears to be problematic. The study examined the average number of firms owned by individuals below and above the threshold on the assumption that split entities are owned by the same person,54 an assumption that appears to discount the probability of multiple-tier ownership structures or ownership structures between related individuals where enterprises are engaging in splitting arrangements. The most convincing evidence of splitting is found in appeals documenting such behaviour uncovered by revenue authorities. uk cases reveal a number of techniques used by taxable persons who have attempted to split the output of service enterprises to yield multiple turnovers below the registration threshold. One technique is to treat individual service providers within a single operation as separate contracted businesses (for example, characterizing hairdressers in a hairdressing

53 Ravi Kanbur and Michael Keen, “Thresholds, Informality, and Partitions of Compliance” (2014) 21:4 International Tax and Public Finance 536-59; and Liu and Lockwood, supra note 43. 54 harju et al., supra note 43, at 30. 328 n canadian tax journal / revue fiscale canadienne (2018) 66:2 salon as independent contractors).55 Another is to attribute different elements of a service to different businesses and business owners (for example, treating a pub meal as separate supplies of food and drink from different related persons).56 A third technique is to seek to file separate registrations for different businesses operated by the same person (for example, attempting to register a real estate agent business separately from a land developer business).57 These types of avoidance arrangements prompted the uk government to add to its vat legislation a specific anti-avoidance provision targeting the artificial separation of business activities.58 The concern with the third tactic used to remain below the threshold, under reporting of sales, is also lost revenue. This behaviour clearly constitutes illegal tax evasion and is most likely to take place in the case of small traders making cash sales to final consumers who do not request invoices.59 False reporting is not limited to those seeking to avoid vat registration. It may also be used by registered businesses (including voluntarily registered firms) seeking to mismatch full claims for inputs while reporting a fraction of outputs,60 and by both unregistered and registered firms seeking to evade vat on sales and income tax on profits. The implications of each of the three types of behaviour for the setting of the threshold differ. One view holds that, to the extent that production restraint is a cause of bunching, a higher threshold might be desirable.61 As the threshold increases, the firms that limited their output to less than the optimal level could produce a little more.62 A contrary view holds that the preferable response to bunching attributable to restraint is a substantial reduction in the registration threshold.63 The latter view assumes that businesses tempted to bunch below the current threshold would find it impractical to hold back production in order to operate at a significantly lower threshold. The hypotheses underlying both views may fail to capture fully the implications of the proposed changes, particularly in respect of the group of businesses that find

55 Customs and Excise Commissioners v. Jane Montgomery (Hair Stylists) Ltd., [1994] STC 256 (Scot. Ct. Ex.). 56 Commissioners of Customs and Excise v. Marner and Marner, [1977] 1 BVC 1060 (VATTR Manchester). 57 Customs and Excise Commissioners v. Glassborow and Another, [1974] QB 465 (QB). 58 value Added Tax Act 1994, schedule 1, paragraph 1A (amended by Finance Act 1997, c. 16, section 31(1)). See also Alan Schenk, Victor Thuronyi, and Wei Cui, Value Added Tax: A Comparative Approach, 2d ed. (Cambridge: Cambridge University Press, 2015), at 69. 59 Dina Pomeranz, “No Taxation Without Information: Deterrence and Self-Enforcement in Value Added Tax” (2015) 105:8 American Economic Review 2539-69. 60 Liu and Lockwood, supra note 43, at 3. 61 Kanbur and Keen, supra note 53, at 551. 62 Ravi Kanbur and Michael Keen, “Reducing Informality” (2015) 52:1 Finance & Development 52-54. 63 Office of Tax Simplification, supra note 36, at 23. vat/gst thresholds and small businesses: where to draw the line? n 329 themselves just below the new threshold. A change in the threshold level does not remove the temptation to bunch, but simply shifts it to a different group. If the threshold is raised to a level inhabited by fewer enterprises, there will be a much smaller pool of potential restrainers, prima facie translating to less bunching. Re- inforcing this conclusion is the host of practical constraints that larger businesses contemplating restraint would face in terms of more substantial operating assets and a larger number of employees. Conversely, if the threshold is reduced substantially, the group of businesses that potentially might bunch grows exponentially, with the result that the absolute value of reduced output could be greater than that caused by restraint by far fewer firms at a much higher threshold. The sheer number of small firms and their dispersal through the economy may also mean that restraint could have a greater impact on the economy as a whole regardless of any output loss. In respect of the restraint issue, in theory, the optimal choice of a new threshold will turn on a balance between potentially increased production by firms with turnovers just below the previous threshold and newly suppressed production by firms with turnovers just below the new higher or lower threshold.64 In practice, however, policy makers are unable to compare actual behaviour at both points. Moreover, as noted, the efficiency losses caused by restrained production may not be large enough to warrant a specific policy response in any case. Finding an optimal threshold to discourage splitting behaviour is similarly challenging. One possibility is that a lower threshold would reduce avoidance by artificial splitting.65 The lower the threshold, the more a business would have to split to remain below the threshold. A lower threshold would raise the cost of avoidance by limiting the size of legitimate input-taxed businesses.66 However, whether it is more difficult and costly for businesses to split into multiple parts will also depend on a host of non-vat considerations, including the nature of the businesses, income tax consolidation rules, company and securities laws, and licensing rules. For example, large integrated firms with central financing and management functions would find it difficult to split finely to slip below a low threshold. Service providers that are able to characterize employees as independent contractors may be indifferent to the number of notional enterprises that they create. It is questionable how effective a lower threshold would be in reducing avoidance by artificial splitting if such activities are strongly concentrated among small service providers, as might be the case in practice in light of the uk cases noted earlier. A preferable response to splitting may be the use of a dedicated specific anti- avoidance measure to consolidate the output of associated or closely bound enterprises. This approach has been adopted in Canada as well as the United King- dom. There are, however, limits to the effectiveness of such rules since they operate

64 Kanbur and Keen, supra note 53, at 549. 65 Ebrill et al., supra note 6, at 121. 66 Ibid. 330 n canadian tax journal / revue fiscale canadienne (2018) 66:2 within the definition of associated or closely bound persons. The Canadian rule,67 for example, applies broadly to entities, but unlike other anti-avoidance rules in Canadian tax legislation that are relevant to associated persons, the output consolidation rule appears not to apply to related individuals. The uk rule may suffer from a similar shortcoming. Nor will the Canadian rule apply to unrelated persons working in the same enterprise but presenting themselves as independent contractors who have elected to operate out of the same premises. These arrangements are more likely to fall within the scope of the uk rule, though its application in these situations is not certain. A better anti-avoidance rule would explicitly overcome the doubts raised by existing models. As is the case with bunching attributable to business splitting, bunching resulting from turnover underreporting involves no constraints on business growth. Busi- nesses expand or operate to capacity and simply fail to report some outputs. Thus, the concern with respect to this tactic to remain below the threshold is lost revenue. It has been suggested that responses to underreporting by businesses above the threshold can include raising the threshold so that evaders become legitimate input- taxed unregistered enterprises.68 The argument assumes that the revenue lost to underreporters is lost in any case, so there is no revenue cost in terms of this group from raising the threshold to explicitly exclude these enterprises from the vat. It must be recognized, however, that raising the threshold to change the status of underreporters will have ancillary impacts on revenue and administrative costs. While no revenue is lost in terms of enterprises that stayed below the lower threshold by underreporting, the higher threshold will have the effect of excluding a host of honest enterprises above the original lower threshold. Any revenue loss from excluding these persons from the vat will be offset to some extent by the resulting administrative and compliance cost savings. Contemporaneously, raising the threshold will encourage a new group of enterprises above the new threshold to become underreporters. This observation may be tempered by the lower proclivity of larger enterprises to underreport. An alternative view aimed at identifying and pursuing underreporters calls for reducing the threshold so that there will be a smaller number of businesses legitim- ately outside the vat net, making it easier for authorities to detect those illegitimately below the threshold.69 However, this view is problematic in respect of the impact that a lower threshold might have on administrative costs. While the total number of businesses outside the net would be smaller, the lower threshold would apply to that pool of businesses selling primarily to final consumers and hence more likely to underreport. Administrators’ workload in terms of the number of businesses to be investigated may increase as a result. Also, there is a possibility that those who evaded in order to remain below the old threshold might still be

67 Excise Tax Act, supra note 35, subsection 148(1) and section 127. 68 Kanbur and Keen, supra note 53, at 551 and 556. 69 Office of Tax Simplification, supra note 36, at 23. vat/gst thresholds and small businesses: where to draw the line? n 331 engaging in evasion after entering the vat system. In that case, administrative resources might remain equally stretched, with responsibility for finding evaders below and above the threshold. Administrative savings may be realized, however, if authorities conclude that only limited resources should be allocated to chase the relatively small revenue lost to evasion by the larger group of low turnover enterprises. At the same time, the larger cohort of registered persons in the vat would increase administrative costs, although these would be offset to some extent by the increased revenue. The various revenue and administrative tradeoffs encountered at different turnover levels are factors that should be considered as elements of the initial threshold- setting exercise, not factors that lead to adjustments of the optimal threshold to address a problem of underreporting. Wherever the threshold is set, there will be underreporters with true turnovers above the threshold and consequent administrative costs incurred to protect the integrity of the vat. Given the uncertainty over possible business responses to higher or lower thresholds, the best way to address underreporting may be to accept this as an inevitable phenomenon whatever the vat registration threshold, and to direct resources to uncover underreporters and bring those exceeding the threshold into the vat. Persons underreporting for vat purposes are equally likely to underreport for income tax purposes. Since thresholds for income tax purposes are uncommon, enhanced enforcement of the income tax for small businesses can be used to reveal underreporters that should be registered for vat. This route is admittedly more difficult in jurisdictions whereva t and income tax are administered by separate agencies, as is the case in China and Malaysia. A registration threshold balancing revenue objectives and administrative and compliance costs is a desirable feature of an efficient and fair vat. The bunching problem that it creates, and the restraint, splitting, and underreporting tactics adopted by enterprises to fall below the threshold, yield economic harm for wider society and revenue losses for the state. A number of techniques are needed to combat these behavioural responses. Adjustment of the registration threshold has been suggested as one means of responding to the behaviour of businesses nearing the threshold. There are, however, competing views on how a registration threshold should be adjusted to address the different ways in which businesses might reduce actual or apparent turnover to avoid crossing the threshold. The contradictory conclusions along with the absence of empirical evidence on the extent of each type of behaviour make consideration of threshold responses to these issues problematic. A preferable approach is probably to look beyond threshold adjustment when considering these issues. Policy makers in jurisdictions such as the United Kingdom that have identified a bunching phenomenon but not pinpointed the behaviour that has led to bunching need to investigate further the means adopted to remain below the threshold before they can develop appropriate responses. Policy makers in jurisdictions such as Canada where the phenomenon itself has yet to be studied need first to conduct this preliminary research to ascertain the extent to which bunching is present. 332 n canadian tax journal / revue fiscale canadienne (2018) 66:2

Transitioning Regimes for Small Businesses Shifting into the Full VAT In the absence of any special rules, the consequences of crossing the registration threshold can be significant for small businesses. Compliance costs jump, and the burden of implicit tax built into the price of acquisitions is replaced by liability to remit a higher explicit tax. As noted, the increase in compliance costs and tax burden can lead to bunching behaviour by small businesses seeking to remain below, or to appear to remain below, the registration threshold. To mitigate this problem, a small number of vat jurisdictions have adopted “transitioning” regimes that sub- sidize the costs incurred by small businesses shifting into the full vat.70 The measures are intended to remove the rationales for some of the behaviour that results in bunching. They will, of course, have no impact on unregistered firms that are nearing the registration turnover threshold and have limited prospects for or interest in further growth.71 The transitioning subsidies commonly take the form of a disappearing credit provided to businesses with turnovers that climb over the registration threshold, with the credit phasing out as turnover rises.72 Variations of the transitioning regime can be found in Finland and the Netherlands. Japan had a similar regime at the time its consumption tax (as the Japanese vat is known) was introduced, but the regime was abolished after eight years. The subsidies provided in the Japanese transitioning regime were particularly generous. When it commenced, the consumption tax featured a high registration threshold of ¥30 million (approximately Cdn $341,687) and a vanishing credit for businesses with turnovers between ¥30 million and ¥60 million, starting with a full offset for consumption tax otherwise payable by enterprises with turnovers immediately above the threshold.73 As a consequence of

70 Institute for Fiscal Studies, supra note 21, at 91. Transitioning assistance is also found in some other sales tax regimes. For example, the retail sales tax (RST) in Manitoba includes a capped “commission” that businesses collecting the RST can retain. The cap has the effect of directing the commission primarily to small businesses; see Manitoba Finance, Retail Sales Tax Act Information Bulletin no. 004, “Information for Vendors,” June 2017. 71 Institute for Fiscal Studies, supra note 21, at 91. 72 Ibid. An alternative regime adopted in Mexico in 2014 provides generous offsets for compliance costs and increased tax burdens by way of a disappearing formula, which allows newly registered businesses to retain a portion of VAT collected from customers sliding from 100 percent of the tax collected in the 1st year of registration to 10 percent in the 10th year. An even more generous concession allowing very small businesses to retain all VAT collected for a decade after entering the VAT has allowed Mexico to remove the threshold entirely, although it remains to be seen whether other concessional measures for small businesses will be enough to offset the compliance costs and increased tax burden that will be incurred when small businesses emerge from the full subsidy period. See Organisation for Economic Co-operation and Development, Taxation of SMEs in OECD and G20 Countries (Paris: OECD, 2015), at 74. 73 The transition tax credit was calculated using the following formula: Transition tax credit = VAT otherwise payable × (¥60 million − annual sales)/(¥30 million). vat/gst thresholds and small businesses: where to draw the line? n 333 a high threshold and a generous offset, tax savings could be significant, and not surprisingly, there was a high takeup rate, with 93.3 percent of eligible registrants signing up for the concession when the consumption tax was adopted.74 While the concession may have had an impact on firms otherwise inclined to underreport sales,75 its most obvious impact was a loss of 88 percent of the total revenue that would have been gained if all taxpayers within this turnover range had been subject to the normal consumption tax.76 To limit the windfall gains by businesses with relatively larger turnovers and capabilities to comply with the consumption tax, Japan lowered the upper limit of the scheme from ¥60 million to ¥50 million two years after the introduction of the regime and finally abolished the concessional regime in 1997. A less generous regime was adopted in Finland in 2004. Under the Finnish system, still in effect, the vat registration threshold remains unchanged at €8,500 (approximately Cdn $12,918), but a disappearing transition tax credit in addition to the ordinary input tax credit entitlement is provided to firms with turnovers between €8,500 and €22,500.77 For businesses that have turnovers of less than €8,500 and that voluntarily register for the vat, the transitioning regime tax credit equals the total vat otherwise payable, with the result that their taxable supplies are essentially zero-rated. Because the credit is calculated by reference to the net vat payable, the transitional credit relief does not apply to businesses that have a negative vat liability. The Finnish regime had very limited impact on bunching by businesses below the threshold.78 It also attracted surprisingly little interest from registered firms that were eligible for the relief, with only 31 percent of the eligible firms applying for the transitioning credit when the concessional regime was introduced.79 The un- remarkable impact of the transitioning regime in Finland may be explained by the very limited benefit that the concession yields for eligible enterprises.80 As a result of a relatively low registration threshold and the relatively low level at which transitioning relief disappears, the average relief for eligible businesses that did not apply would have been only €617, and 10 percent of these businesses would have received less than €100 had they applied.81 The value of incentives compared to the relatively high compliance costs faced by firms entering the vat was insufficient to

74 hiromitsu Ishi, The Japanese Tax System, 3d ed. (Oxford: Oxford University Press, 2004). 75 William J. Turnier, “Accommodating to the Small Business Problem Under a VAT” (1994) 47:4 Tax Lawyer 963-86. 76 Ishi, supra note 74, at 292. 77 The tax credit is calculated using the following formula: Transition tax credit = VAT paid − [(turnover − €8,500) × VAT paid] / (€22,500 − €8,500). 78 Institute for Fiscal Studies, supra note 21, at 88-89. 79 Ibid., at 90. 80 Ibid. 81 Ibid. 334 n canadian tax journal / revue fiscale canadienne (2018) 66:2 encourage very small unregistered businesses with limited sales to expand operations and lift turnover above the registration threshold.82 The Finnish and Japanese experiences illustrate the tradeoffs encountered at the margins of transitioning regimes. If the registration threshold is low, a transitioning regime is unlikely to have a significant impact on bunching, since the value of tax relief is low relative to the compliance cost burden faced by registered businesses. If the registration threshold is high, the corresponding higher value for tax relief may reduce bunching, but the high takeup rate by eligible registered businesses above the threshold may deliver windfall gains at a high cost to revenue. The risk of windfall benefits is particularly acute in the case of voluntary registrants. By definition, these enterprises were willing to be part of the full vat system, but as a result of the transitioning regime, they will retain a portion of the output tax that they collect. It remains to be seen if transitioning regimes to reduce bunching incentives could yield better results in vat systems with thresholds that lie between the two examples described. The challenge faced by policy makers is to determine the level of relief and the withdrawal formula that will achieve an optimal balance between providing incentives that are high enough to encourage transition to a normal vat system and avoiding excessively high windfall gains to businesses that would make no deliberate effort to remain below the registration threshold.

Simplification Regimes for Small Businesses in the VAT System Both small businesses facing vat compliance costs and vat designers recognize the unfairness of the disproportionate compliance costs faced by the sector compared to the burden borne by large businesses. The response has been the adoption of a number of “simplification” systems designed to reduce the cost of compliance for small businesses that have moved into the vat system. These regimes can operate in conjunction with a vat threshold, applying to businesses in a defined band above the threshold; or, in the absence of a threshold, they can apply to all businesses with turnovers below the level at which it is agreed the unfairness has largely dissipated. Simplified regimes may have an impact on the optimal registration level if they operate in conjunction with a registration threshold. The optimal threshold balances revenue against compliance and administrative costs. The reduction of compliance costs for businesses with turnovers at the lower end of the vat system in theory makes it possible to use a lower threshold while still balancing compliance costs and revenue. However, any reduction would be limited, and perhaps negated, by the increased administrative costs that would follow if more enterprises entered the vat system. Even if the goal of reduced costs from simplified rules is not achieved, the professed outcome may provide a rationale for a lower threshold if policy makers seek to tip the balance in favour of increased revenue. Simplified regimes or methods that are commonly used to reduce compliance costs borne by small businesses fall into three groups:

82 Ibid., at 91. vat/gst thresholds and small businesses: where to draw the line? n 335

1. less frequent filing and payments; 2. optional cash- or payment-basis accounting (as opposed to accrual-basis accounting); and 3. presumptive regimes for small businesses that approximate the vat that would otherwise be paid under the normal vat system.

Many countries use more than one of these regimes. For example, the United King- dom uses all three regimes, while Canada uses two of the three—less frequent filing and payments, and a presumptive regime (known as quick method gst/hst accounting). A cross-country comparison of the regimes adopted in oecd and Group of Twenty (g20) countries is presented in table 3. Each type of regime is described in more detail in the text that follows.

Less Frequent Filing and Payments Many countries allow small businesses to file and remit less frequently than larger counterparts, often coupling the small business rules with a more frequent filing requirement for very large enterprises. The filing (and payment) periods available vary from jurisdiction to jurisdiction. The United Kingdom, Canada, and Australia use monthly, quarterly, and annual filing. The assumption that less frequent filing could significantly reduce compliance costs83 may, however, be exaggerated. Return filing often accounts for a small proportion of compliance costs. A study of compliance costs in Canada shows that the actual completion of the return accounts for only 4 percent of the total labour effort on compliance, indicating that small businesses allocate very little time to return-filing activities.84 The costs of filing returns may have been largely reduced, moreover, by the availability of online filing services. The larger costs are incurred in the process of identifying inputs and outputs to be inserted into the return, and this task requires a fixed amount of time and effort regardless of the frequency of return filing. It is likely that the total compliance costs remain relatively constant whether returns are filed once or four times a year. In a worst-case scenario, less frequent filing may actually be counterproductive in terms of its simplification goals. In the United Kingdom, for example, quarterly filing is the standard option for small and medium-sized businesses (defined by turnover ranges), while small businesses below an annual filing turnover level can elect to file annually. Evidence suggests that in some cases annual return filing raised compliance costs because the single filing prompted poorer record keeping, leaving businesses struggling to assemble records required to complete one return covering a full year of transactions.85 As a result, many small businesses that joined

83 OECD, supra note 28, at 124. 84 Plamondon & Associates, supra note 14, at 46-47. 85 United Kingdom, Office of Tax Simplification,Review of Value Added Tax: Progress Report and Call for Evidence (London: Office of Tax Simplification, February 2017), at 14. 336 n canadian tax journal / revue fiscale canadienne (2018) 66:2

Table 3 VAT Simplification Measures for Small and Medium-Sized Enterprises in OECD and G20 Countries, 2015

Calculation of VAT/GST liability Accounting, payment, filing

Simplified input tax credit Reduced Exemption Presumptive calculation Cash frequency thresholds tax schemes schemes accounting of filing Other

Argentina . . . . ✓ ✓ Australia . . . . . ✓ ✓ ✓ ✓ ✓ Austria ...... ✓ ✓ ✓ ✓ ✓ Belgium . . . . . ✓ ✓ ✓ ✓ Brazil ...... ✓ ✓ ✓ ✓ Canada . . . . . ✓ ✓ ✓ ✓ ✓ Chile ...... ✓ ✓ China ...... ✓ ✓ Czech Republic . . ✓ ✓ Denmark . . . . . ✓ ✓ Estonia . . . . . ✓ ✓ Finland . . . . . ✓ ✓ France ...... ✓ ✓ ✓ ✓ Germany . . . . . ✓ ✓ ✓ Greece ...... ✓ ✓ ✓ ✓ Hungary . . . . . ✓ ✓ ✓ ✓ Iceland ...... ✓ India ...... Ireland ...... ✓ ✓ ✓ Italy ...... ✓ ✓ ✓ ✓ Japan ...... ✓ ✓ Korea ...... ✓ ✓ ✓ Luxembourg . . . ✓ ✓ ✓ ✓ Mexico . . . . . ✓ ✓ ✓ ✓ Netherlands . . . ✓ ✓ New Zealand ...... ✓ ✓ ✓ Norway . . . . . ✓ ✓ Poland ...... ✓ ✓ ✓ ✓ Portugal . . . . . ✓ ✓ ✓ ✓ Slovak Republic . . ✓ Slovenia . . . . . ✓ ✓ South Africa . . . ✓ ✓ ✓ ✓ Spain ...... ✓ ✓ ✓ Sweden . . . . . ✓ ✓ ✓ Switzerland . . . . ✓ ✓ ✓ ✓ Turkey ...... ✓ United Kingdom . ✓ ✓ ✓ ✓ ✓ United States . . .

GST = goods and services tax; VAT = value-added tax. Source: Based on data extracted from Organisation for Economic Co-operation and Development, Taxation of SMEs in OECD and G20 Countries (Paris: OECD, 2015) and updated and corrected by the author. vat/gst thresholds and small businesses: where to draw the line? n 337 the annual accounting scheme later moved back to standard quarterly returns.86 In fact, although over 90 percent of the total registrants in 2016-17 had turnovers eligible for annual accounting, less than 1 percent of all registered taxpayers had elected to join the scheme.87 An alternative rationale for using less frequent filing is to reduce administrative costs. With a view to reducing the total number of returns handled by the tax authority,88 the Canadian law was amended to provide a default annual reporting period for small businesses below a specified turnover level, effective beginning in 1994, with the option to elect to use other reporting periods. The assignment of annual reporting as the default option may have contributed to the relatively high takeup rate of this option. By 2003-4, 36 percent of small businesses below the annual filing threshold were annual filers.89 The significant difference in the takeup rates of the annual accounting option in the United Kingdom and Canada may be attributable to the way in which the option is made available—that is, whether the default is a standard filing period with an opt-in to annual filing or the default is annual filing with an opt-out to other filing periods. Thus, theuk approach could be used by countries that are concerned about compliance costs, whereas the Canadian approach could be followed by countries that are concerned about administrative costs. How- ever, the overall benefits of the less frequent filing option are uncertain given its potential to increase compliance costs and discourage good business management. For small businesses that use less frequent filing, there may be a separate fiscal benefit if tax payments are tied to the filing of returns. Less frequent tax payments may provide a small cash flow benefit to qualifying taxpayers that could be seen as compensation for the relatively higher compliance costs faced by small businesses.90 There has been concern, however, that small businesses tend to have difficulties in meeting deferred payment obligations associated with less frequent filing.91 A compromise solution is to allow optional less frequent filing but to require qualifying small businesses to make estimated advance payments with the same frequency as is required for other businesses.92 The annual accounting scheme in the United Kingdom is an example of this approach.93

86 Ibid. 87 hM Revenue & Customs, supra note 29, at section 2.11. 88 Salvail, supra note 18. 89 Canada Revenue Agency, supra note 28, at table 2. 90 OECD, supra note 72, at 112; and Cedric Sandford, “The Administrative and Compliance Costs of the United Kingdom’s Value-Added Tax” (1990) 38:1 Canadian Tax Journal 1-20. 91 alan A. Tait, Value Added Tax: International Practice and Problems (Washington, DC: International Monetary Fund, 1988), at 138-39. 92 Turnier, supra note 75, at 984. 93 For details, see United Kingdom, “VAT Annual Accounting Scheme” (www.gov.uk/vat-annual -accounting-scheme/overview). 338 n canadian tax journal / revue fiscale canadienne (2018) 66:2

Cash Accounting vat systems are generally accrual based, meaning that the vat is paid (or deducted) when invoices are issued (or received). In many countries, small businesses have the option to use a cash accounting method, accounting for vat on the basis of payments received or made. While cash accounting is often advocated as a means of reducing compliance costs,94 the overriding purpose is more likely to provide cash flow benefits to eligible businesses. In particular, businesses that collect the payments from their customers long after the invoices are issued may benefit greatly, since the output tax is not due until payments are received. Cash accounting thus avoids vat being paid on bad debts. The common practice in most countries is to set a threshold based on turnover, with businesses below the threshold being eligible to use cash accounting. Busi- nesses with turnovers above that threshold should account for vat on an accrual basis. Concurrent cash and accrual accounting, however, creates a timing mismatch between input tax deduction and output tax liability when a cash-basis supplier makes a supply to an accrual-basis purchaser.95 The accrual-basis purchaser claims an immediate input credit while the cash-basis seller may defer the payment for a significant period of time or even indefinitely. New Zealand and Australia had the experience that some related cash- and accrual-basis taxpayers aggressively exploited the timing mismatches to obtain what were effectively interest-free loans from the government.96 The schemes were considered avoidance arrangements and were attacked by tax authorities using general anti-avoidance rules (gaars). New Zealand also subsequently adopted a specific anti-avoidance rule (saar) to address the problem.97 In the United Kingdom, where the gaar does not apply to the vat, a saar was used to target these schemes.98 New Zealand’s experience shows that inad- equately designed saars may limit avoidance schemes but do not eliminate them.99 It is quite possible that the cases uncovered by the Australian and nz authorities

94 OECD, supra note 72, at 110. 95 New Zealand Inland Revenue and New Zealand Treasury, Options for Strengthening GST Neutrality in Business-to-Business Transactions (Wellington: New Zealand Inland Revenue, Policy Advice Division, June 2008), at 29. 96 See, for example, Ch’elle Properties (NZ) Limited v. Commissioner of Inland Revenue, [2007] NZSC 73; Education Administration Ltd. v. Commissioner of Inland Revenue, [2010] NZHC 663; and VCE and Commissioner of Taxation, [2006] AATA 821. 97 New Zealand, Goods and Services Tax Act 1985, section 19D. 98 value Added Tax Regulations 1995, regulations 58(e) and (f). For an analysis of the Australian, New Zealand, and UK approaches, see Yige Zu and Richard Krever, “GST Cash and Accrual Mismatches: Avoiding the Avoidance” (2017) 46:4 Australian Tax Review 271-83. 99 In Case X25, [2006] 22 NZTC 12,303 (TRA), the parties avoided a SAAR by structuring the transaction to fall just below the trigger threshold for the SAAR. The tax authority ultimately prevailed applying the GAAR. vat/gst thresholds and small businesses: where to draw the line? n 339 represent only the tip of a cash-accrual mismatch abuse iceberg. The problem is avoided in the first place in countries such as Canada where theg st law does not provide the cash accounting option.

Presumptive Input Tax Entitlement Regimes Presumptive input tax entitlement regimes seek to simplify the calculation of vat liability by removing the need to record and total input tax on all acquisitions and instead allowing qualifying persons to substitute a single presumptive input tax entitlement. Under presumptive regimes, small businesses charge vat at regular rates on all taxable supplies of goods and services. A single flat rate is then applied to the total (vat-inclusive) turnover to determine the amount of tax to be remitted to the tax authority. This amount is a proxy for the amount of net vat that the enterprise would have remitted after deducting actual input tax credits from output tax in the ordinary vat system. Importantly, registered customers of persons using the presumptive input tax entitlement system are entitled to full input tax credits, since they have been charged full vat on their acquisitions. The presumptive input tax calculation regimes do not affect the amount of vat imposed on supplies and charged to customers. Their only role is to determine the amount of net vat remitted by qualifying small businesses while obviating the requirements that the businesses track and then total the vat included in the price of every acquisition. The presumptive input tax entitlement may vary across industries or sectors, and may reflect both the average costs incurred by enterprises in a sector and the extent to which acquisitions within the sector are likely to be exempt, zero-rated, or reduced- rate supplies.100 In addition to the presumptive input tax credit entitlement provided through retention of a proportion of output tax collected, further explicit input tax credits are allowed for acquisitions of capital assets. Notable examples of such presumptive schemes include the flat rate scheme frs( ) in the United King- dom and the quick method of gst/hst accounting in Canada. Both are optional for registered businesses below a specified turnover. An inherent problem in any presumptive regime that determines net vat to be remitted by the use of a single flat rate applied to turnover is inaccuracy in specific cases. The input tax credits notionally incorporated into the flat rate are based on averages that are by definition not accurate for most individual traders. The finer the group used to determine an average, the more accurate the calculation will be in theory. However, the finer the group, the greater the number of boundary problems that are created. For example, seeking to mitigate the presumed input tax inaccuracy problem, the United Kingdom has calculated different presumptive rates ranging from 4 percent to 14.5 percent for 54 categories of industries. The proliferation of categories creates a new level of complexity for businesses that must determine

100 OECD, supra note 72, at 107. 340 n canadian tax journal / revue fiscale canadienne (2018) 66:2 which type of business activity they conduct when registering101 and a new set of policing problems for the tax authority.102 Successful challenges to the uk tax authority’s guidance and decisions illustrate the difficulty that revenue officials have in applying the law.103 Enterprises with higher than average inputs that believe that they will be unable to recover fully all input taxes under the industry-by-industry flat rate scheme can simply not opt to join the scheme. In 2016-17, only 25 percent of the taxpayers eligible to join the uk scheme were actually in it.104 Deciding whether or not to opt in is not a cost-free exercise. Eligible businesses will regularly incur internal and external costs to estimate vat liabilities under both the frs and the normal vat regime before making the decision.105 At the same time, the flat rate will provide a tax advantage to businesses in each group that have lower than average input purchases. The additional compliance costs incurred by businesses to determine whether the “simplified” scheme prejudices or enhances their recovery of input tax clearly offset some of the simplification benefits that thefrs is expected to achieve. The frs therefore functions in practice as a concessional scheme for some businesses rather than a simplification scheme as intended. The main purpose for many of these businesses to enter the scheme is actually to reduce their tax liabilities. Subsequent to the adoption of frs, the uk government ascertained that about 30 percent of the businesses that used the scheme enjoyed substantial cash advantages relative to the position that they would have faced under the ordinary vat regime.106 The government viewed this outcome as an abuse of the frs and consequently introduced a new remittance rate to remove the benefit from service providers with presumed limited input costs.107 A new 16.5 percent rate then applies to “limited cost traders” whose expenditure on goods is less than 2 percent of their

101 KPMG LLP, Administrative Burdens—HMRC Measurement Project: Report by Tax Area Part 27: Value Added Tax (London: KPMG, March 2006). 102 James Mirrlees, Stuart Adam, Tim Besley, Richard Blundell, Stephen Bond, Robert Chote, Malcolm Gammie, Paul Johnson, Gareth Myles, and James M. Poterba, Tax by Design: The Mirrlees Review (Oxford: Oxford University Press, 2011). 103 See, for example, Idess Ltd. v. Revenue & Customs, [2014] UKFTT 511 (TC); SLL Subsea Engineering Ltd. v. Revenue & Customs, [2015] UKFTT 43 (TC); and JJK Engineering Ltd. v. Revenue and Customs, [2016] UKFTT 615 (TC). 104 hM Revenue & Customs, supra note 29, at section 2.10. 105 KPMG, supra note 101, at 24. 106 United Kingdom, HM Revenue & Customs, “Explanatory Memorandum to the Value Added Tax (Amendment) Regulations 2017,” 2017 no. 295 (www.legislation.gov.uk/uksi/2017/295/ memorandum/contents). 107 United Kingdom, HM Revenue & Customs, “Tackling Aggressive Abuse of the VAT Flat Rate Scheme—Technical Note,” December 5, 2016 (www.gov.uk/government/publications/ tackling-aggressive-abuse-of-the-vat-flat-rate-scheme-technical-note/tackling-aggressive -abuse-of-the-vat-flat-rate-scheme-technical-note). vat/gst thresholds and small businesses: where to draw the line? n 341 turnover. The limited cost trader test adds further complexity for businesses that genuinely use the scheme to save compliance costs, since it requires businesses to separate records of purchases of goods and services, a distinction that does not readily exist in a modern economy. Moreover, the test can be easily avoided by those that use the scheme “abusively” to achieve tax savings.108 For example, a firm primarily supplying services may be able to enjoy a fiscal benefit if it engages in a subsidiary activity of buying goods and selling them at a loss.109 The uk experience shows that specific anti-avoidance measures may have limited value in preventing businesses from benefiting from a system that is presumptive in essence. At worst, the measures may induce unproductive and inefficient business responses. The initial design of the Canadian quick method may have avoided some of the problems experienced in the United Kingdom. Instead of calculating dozens of industry-specific rates, Canada has only two sets of rates, distinguishing between businesses that purchase goods for resale and businesses that provide services, with the former being allowed a higher notional input tax credit by way of a greater retention of output tax.110 The Canadian system thus avoids much of the complexity found in the uk system as a result of the need to identify each business’s specific industry category. However, a sharp distinction between businesses that primarily sell goods and those that primarily provide services may induce inefficient behaviour by enterprises operating near the margin, a risk noted in the United Kingdom. While the Canadian system may prevent excessive windfall benefits for pure service providers, with its broad sweep approach across all industries this system has much more scope for inaccuracy in respect of any given enterprise within each rate category. The number of rates, to a degree, reflects a government’s perception of the balance between neutrality and simplicity. The quest for greater accuracy in, and less abuse of, the presumptive regime in the United Kingdom, for example, has had an impact on its effectiveness as a simplification system. Cost savings from theuk frs were initially estimated to average £750 per business,111 but a later evaluation estimated average compliance cost savings to be only £45.112 Potential savings are

108 Office of Tax Simplification, supra note 85, at 11. 109 Neil Warren, “VAT: Limited Cost Trader Category for FRS Users,” December 20, 2016 (www.accountingweb.co.uk/tax/hmrc-policy/vat-limited-cost-trader-category-for-frs-users). 110 a further mandatory regime applies to charities that allows them to retain 40 percent of the GST/HST they collect in lieu of input tax credits. 111 United Kingdom, HM Customs and Excise, Easing the Impact of VAT: Consultation on a Flat Rate Scheme for Small Firms: HM Customs and Excise Summary of the Responses to the Consultation Document Issued in June 2001, Parliament Deposited Paper 02/956 (London: HM Customs and Excise, 2002). 112 United Kingdom, HM Revenue & Customs, “VAT Flat Rate Scheme (FRS): Impact Assessment of Changes to the Flat Rate percentages in January 2010,” December 7, 2009 (www.legislation.gov.uk/ukia/2009/327/pdfs/ukia_20090327_en.pdf ). 342 n canadian tax journal / revue fiscale canadienne (2018) 66:2 likely to fall even further with increased computerization of cash registers and accounting systems, which, as noted earlier, have greatly reduced compliance costs.113 However, these technological changes cannot entirely eliminate the need for simplification schemes. Computerized systems are useful where they can automatically record tax attributes of sales or acquisitions—for example, whether supplies are taxable or exempt, and in the former case whether they are subject to standard, reduced, or zero rates—but they offer no savings where judgments are required, such as the apportionment of input tax credits by businesses that make both taxable and exempt supplies. Simplified regimes will continue to play a cost- reduction role in these cases.114 From a tax policy perspective, the need for simplification regimes in these circumstances arguably reinforces the need to address the underlying complexity of a concession-ridden vat. Of the three techniques that have been used—less frequent filing and payments, cash-basis accounting, and presumptive input tax calculations—the first appears to entail the least risk of abuse or inaccurate and inappropriate outcomes. However, the simplification benefits of all three techniques are uncertain. Leaving political considerations aside, it is difficult to pursue so-called simplification regimes as a reform priority. If measures are necessary, less frequent filing and payments seems to be the best candidate of the trio for adoption.

Alternative Regimes for Small Businesses As noted, the primary rationale for a threshold is to reduce compliance and administrative costs. Firms with turnovers below the threshold are input-taxed, meaning that they need not remit any tax on sales, but at the same time they are not entitled to claim input tax credits for their acquisitions. Subjecting small businesses to input taxation comes at a small revenue cost to the government but shields the firms almost entirely from compliance costs. Highest-income countries have universally concluded that a registration threshold with input taxation of enterprises below the threshold is the preferable response to disproportionate compliance and administrative costs associated with very small businesses in a full vat system, although there is a wide variation in their conclusion on the level of the optimal threshold. Commonly, these jurisdictions also adopt simplified rules for small businesses above the registration threshold. A handful of middle-income countries have instead opted to remove registration thresholds entirely in a bid to extend the revenue base. In some jurisdictions, particularly

113 See also Office of Tax Simplification, supra note 85, at 11. The adoption of a simplified regime has been seen as a direct outcome of the compliance challenges faced by small businesses in the period before widespread computerization; see L. Dana, “A Goods and Services Tax (GST) and the Small Business Sector: Some Canadian Reflections” (1993) 52:4Australian Journal of Public Administration 457-64. 114 Office of Tax Simplification, supra note 85, at 11. vat/gst thresholds and small businesses: where to draw the line? n 343 developing countries, another approach to simplification for small businesses is used, with the registration threshold being replaced by a turnover border that distinguishes larger businesses subject to full vat and smaller businesses subject to an alternative turnover tax system. The latter is commonly described as a “simplified” vat system.115 There is no doubt that compared to full vat, a turnover tax imposed solely on sales without regard to input tax credits is easier from both a tax administration perspective and a taxpayer compliance viewpoint.116 It is equally clear, however, that subjecting small businesses to a turnover tax is not as simple as the alternative of removing the businesses from the vat system entirely and leaving them subject to input taxation. Not surprisingly, other rationales are offered for the application of so-called simplified regimes to these firms. A common explanation for the alternative turnover tax is to bring informal businesses into the “formal” economy, but advocates of this view offer no example of formality apart from formally paying higher taxes.117 In theory, tax authorities could pass on details of known enterprises to other authorities responsible for business licences or other attributes of the formal economy; in practice, however, particularly in developing economies, channels for the automatic exchange of data between ministries are limited. It is possible that if appropriate information channels were established, incorporation of small businesses into the formal (taxpaying) economy could provide the government with a better understanding of the overall economy, allowing it to make more informed macroeconomic decisions. Somewhat ironically, many of the jurisdictions that have embraced an alternative turnover tax as a means of reducing informality have a relatively lower administrative capacity than those jurisdictions that exclude enterprises with turnovers below the vat threshold from tax with the goal of a smaller taxpayer base. A second explanation for simplified turnover tax regimes for small businesses is to provide these businesses with basic fiscal skills as preparation for compliance should they grow sufficiently to cross the full vat border.118 The theory is that an introduction to simple turnover accounts can mature into more sophisticated record- keeping skills at a later stage. While small business turnover tax regimes appear to lack all the fundamental attributes of a vat, particularly entitlement to input tax credits on acquisitions and the provision of tax invoices on sales, the description of these regimes as simplified quasi-vat alternatives is not wholly misleading. A single low tax rate imposed on total turnover is notionally similar to the full rate applied to sales reduced by input

115 Examples of jurisdictions with a turnover tax for small businesses that is notionally incorporated into a VAT include China, Ethiopia, and West Bengal in India. 116 Richard M. Bird and Pierre-Pascal Gendron, The VAT in Developing and Transitional Countries (Cambridge, UK: Cambridge University Press, 2011), at 29. 117 The weakness of this explanation was noted in Kanbur and Keen, supra note 62, at 52. 118 Bird and Gendron, supra note 116, at 187. 344 n canadian tax journal / revue fiscale canadienne (2018) 66:2 tax credits on acquisitions. However, the analogy between a turnover tax and an actual vat is limited. Without tax invoices in hand, registered taxpayers under the normal vat system who purchase from suppliers in the turnover tax system cannot claim input tax credits, and the turnover tax becomes a cascading non-recognizable cost of acquisition for enterprises in the vat system. The cascading effect is notice- ably higher than that encountered when small businesses are simply left outside the vat but incur input taxation on acquisitions. In some ways, the turnover tax applied to small businesses below the registration threshold resembles the presumptive input tax system designed for small businesses above the threshold. Under both regimes, the amount of tax remitted to the tax authority is determined by applying a reduced rate, lower than the standard vat rate, to gross receipts. In the case of the presumptive input tax, the net remittance is presumed to reflect the application of full tax to sales and full recovery of input tax credits, so that a registered customer will receive a tax invoice evidencing payment of the full vat. The notional netting of input tax credits against output tax is mani- fested in a reduced remittance by the supplier to the tax authority, not in the tax paid by customers. In the case of the turnover tax, however, there is no presumption that the final price always includes a full vat component, and the vendor is not able to issue a tax invoice. In a revenue-neutral context, a lower-rate turnover tax on small businesses allows a government to raise the threshold at which a higher-rate vat applies.119 This is because the tax collected by way of a lower-rate turnover tax applied to gross receipts with no recognition of input tax credits is likely to exceed the revenue that would be collected if these enterprises were excluded entirely from the vat and consequently subject to input taxation. However, the tradeoff comes with an economic cost. To begin with, the non-recoverable turnover tax provides an incentive for self-supply that may result in less specialized and less efficient businesses.120 Concern over this outcome was one of the factors prompting members of the predecessor to the Euro- pean Union to replace their turnover taxes with vat systems.121 Equally importantly, the non-creditable feature of the turnover tax and the tax cascading to which this leads put small suppliers subject to the tax at a significant competitive disadvantage relative to enterprises in the full vat system when selling to registered businesses.

Conclusion Registration thresholds and small business regimes are primarily designed to reduce administrative costs borne by tax authorities and compliance costs that fall on small businesses. Each of these features of vat systems gives rise to concerns.

119 Ibid., at 120. 120 Richard Krever, “Designing and Drafting VAT Laws for Africa,” in Richard Krever, ed., VAT in Africa (Pretoria: Pretoria University Law Press, 2008), 9-28, at 10. 121 For example, France, Italy, and the Netherlands. See Henry J. Aaron, ed., The Value-Added Tax: Lessons from Europe (Washington, DC: Brookings Institution, 1981). vat/gst thresholds and small businesses: where to draw the line? n 345

The registration threshold raises concerns because of the distinction that it creates between small businesses bearing lower tax and compliance burdens and slightly larger firms with higher liabilities and compliance costs. Businesses left outside the full vat that primarily sell to final consumers will generally enjoy a competitive advantage from the lower tax burden and compliance costs. Those selling to registered enterprises will be disadvantaged, a problem that can be addressed through voluntary registration, albeit with implications for both revenue and administrative costs. These factors will affect the optimal registration threshold, particularly in countries where the voluntary registration rate is high. These reduced tax burden and compliance costs for unregistered enterprises create incentives for businesses selling primarily to final consumers to stay below the threshold through real output changes or avoidance or evasion activities. Empirical studies clearly reveal business behaviour to bunch below a registration threshold in the jurisdictions in which this research has been conducted. They do not reveal the extent to which the capped turnover that leads to business bunching is attributed to real output changes, avoidance, or evasion—causes that may vary in impact across different jurisdictions depending on the nature of incentives given to small businesses above the threshold, relative compliance costs, and administrative and enforcement capacities. Views on a possible role for the registration threshold level to address unwanted behavioural responses are ambiguous, with different observers proposing higher or lower thresholds to address the same type of behaviour in some cases. Each recommendation is based on assumptions regarding probable changes to business behaviour, a risky basis for policy development. Balancing the recommendations may thus be impossible, in particular where empirical studies do not provide evidence on the extent of each behaviour. At the same time, it seems that concern over restraint as a cause of bunching may be exaggerated. In the face of uncertainty coupled with the availability of alternative direct policies that can be used to address splitting and underreporting, the best approach for policy makers appears to be to seek a threshold that balances revenue and compliance and administrative costs without regard to bunching behaviour but taking into consideration the impacts of voluntary registration. A few countries have sought to minimize inefficiencies and distortions caused by a sharp increase in tax liability and compliance costs at the threshold by adopting a graduated phasing-in regime. Experience nevertheless suggests that it is difficult to set the correct incentives to make the regime work effectively while not providing excessive windfall benefits. Wherever the vat registration threshold is set, just above the threshold are enterprises relatively smaller than those further up the turnover scale. These smaller businesses face disproportionate compliance costs, prompting the adoption of simplified regimes to reduce those costs. If the simplified regimes truly led to reduced compliance costs, the optimal threshold level balancing costs and revenue might shift, subject to the constraint of greater administrative costs. Even if an actual reduction of costs does not materialize, as appears likely in many instances, 346 n canadian tax journal / revue fiscale canadienne (2018) 66:2 the nominal outcome of a simplified system may provide political cover for adoption of a lower threshold. Further considerations that may affect the decision to adopt a simplified regime include the risks entailed in simplified systems and questions about their fairness and behavioural consequences. Simplified regimes that allow concurrent cash- and accrual-basis accounting for vat purposes are vulnerable to avoidance schemes involving cash-basis sellers who may defer tax liability indefinitely while related accrual-basis buyers claim immediate input tax credits. Presumptive regimes intended to remove the need for small businesses to track input tax will approximate the impact of a vat in respect of only a tiny number of businesses that mimic exactly the characteristics of the models used to calculate the notional input tax entitlement built into the retention formula. For all other eligible businesses, these regimes provide either windfalls or penalties. At the same time, separate presump- tions for enterprises that provide different types of supplies may induce businesses to add particular types of supplies to or remove others from their business models. While higher-income jurisdictions have concluded that input taxation only of smaller businesses below the threshold is the optimal solution to the compliance cost and administrative cost problems, there are some jurisdictions with no registration threshold, where the vat is extended to all enterprises. A different approach adopted in some middle- and lower-income jurisdictions replaces the registration threshold with an alternative tax border, with a lower-rate turnover tax being imposed on businesses below the full vat threshold. The turnover tax leads to cascading and consequent competitive disadvantages for many businesses, while encouraging others to adopt less efficient self-supply structures. vat thresholds and small business regimes are among the most difficult policy areas in a vat, presumably because the evidence on the extent of problems associated with them and business responses to attempted solutions is so ambiguous. Subject to this caveat, however, a number of tentative conclusions can be reached. First, the adoption of a registration threshold is the most efficient measure to reduce administrative and compliance costs. A higher threshold with the option of voluntary registration is generally preferable to a lower threshold with concessional or simplified regimes for businesses above the threshold, and input taxation rather than alternative turnover taxation is preferable for enterprises with turnovers below the registration threshold. Second, where simplified rules for small businesses with turnovers above the registration threshold are necessary (for political reasons), the option for less frequent filing, possibly coupled with less frequent payments, is the least harmful concession available. Third, incentive schemes designed to facilitate the transition of small businesses into the vat appear not to be successful. Distortions and inefficiencies caused by a threshold may have to be seen as a necessary price to be paid for achieving administrative and compliance cost savings, at least before more evidence on the negative effects of a threshold is available. What are the implications of these conclusions for Canada? Prior to the adoption of the federal gst, most provinces levied retail sales taxes, often with lower thresholds. The shift to gst/hst in selected provinces allowed some small businesses to vat/gst thresholds and small businesses: where to draw the line? n 347 reduce their compliance costs and lowered overall administrative costs. In the absence of any adjustments or indexation, the Canadian registration threshold, set in 1991, has fallen in real terms for 16 years, drawing an ever-higher proportion of businesses into the gst net.122 This may not be a problem. Over the same period, administrative capacity to collect and enforce the tax has grown, and technological developments have reduced compliance costs, likely softening the impact of more firms entering the formalg st regime. Study of the relative weighting of these issues will help to determine whether the enlargement of the gst net has been appropriate. This balance must also be considered in light of current concessional regimes for small businesses. In particular, Canada’s quick method gst/hst accounting for small businesses may yield more problems than solutions to the compliance cost issue. Winding up the system and at the same time raising the registration threshold to remove more enterprises from the gst could generate similar or greater compliance savings without the current risk of concession abuse. In terms of the bunching issue, the implications for Canada will turn on empirical data that until now have not been collected. Before a full response can be developed, the distribution of businesses must be investigated to ascertain the extent of bunching and to identify the means used to remain below the registration threshold in Canada. In the absence of evidence of either the extent of or the means used to reduce turnover, simply continuing along the current path may do less harm than changes in response to unknown behaviour. A better approach, however, would be to commence the research required for more reasoned long-term reform.

122 If the real value of the threshold used in 1991 had been adjusted in line with inflation, registration in 2017 would have been required only for businesses with sales revenue of $47,334 or more. canadian tax journal / revue fiscale canadienne (2018) 66:2, 349 - 50

Policy Forum: Editors’ Introduction— Should Canada Have a Tax Commission?

Canada’s first, and to date its only, broadly focused royal commission on taxation, the Carter commission, received its mandate in 1962 and reported in late 1966— more than 50 years ago.1 The last comprehensive tax reform occurred in 1987, now 30 years in the past.2 Since then, tax changes have arrived with a narrower focus on one part of the tax system, such as sales tax reform in 1991, federal-provincial “tax on income” changes in 2000, and most recently amendments affecting shareholder-owners of private corporations, in 2017-18.3 Is it best to continue with sequential narrow reforms, or should Canada pursue a wide-ranging investigation to reset the tax system for the needs of the 21st century? This Policy Forum addresses the case for a new tax commission. We have assembled four articles by authors with diverse backgrounds. Below we introduce each of the four contributions, with some of our own added context. The forum begins with a statement of the case for a broad-based tax commission. Fred O’Riordan, now with Ernst & Young llp and formerly an assistant commissioner of appeals at the Canada Revenue Agency, argues that only a comprehensive review can properly assess the tax system for its ability to reach the goals of fairness, neutrality, efficiency, and international competitiveness. Tax systems require not just regular updating but also a periodic deep review. O’Riordan focuses mainly on international competitiveness and the nuts-and-bolts problems that affect the daily work of tax practitioners (relating, in particular, to simplicity, administrative feasibility, and certainty). He argues that these problems alone suggest the need for such a review. Also, given the changes in the economy and taxation since the last major review over 50 years ago, a systematic study seems overdue.

1 Canada, Report of the Royal Commission on Taxation, vols. 1-6 (Ottawa: Queen’s Printer, 1966-67) (“the Carter commission”). The Royal Commission on the Taxation of Annuities and Family Corporations (“the Ives commission”) reported in 1945. Other royal commissions have had important taxation aspects, such as the Rowell-Sirois commission in 1940 (Canada, Report of the Royal Commission on Dominion-Provincial Relations (Ottawa: King’s Printer, 1940)). The full text of these reports is available on the federal government publications website (http://publications.gc.ca/site/eng/publications.html); the Carter commission’s report is also in the Canadian Tax Foundation’s TaxFind database (www.ctf.ca). 2 Canada, Department of Finance, The White Paper: Tax Reform 1987 (Ottawa: Department of Finance, June 18, 1987) (http://publications.gc.ca/collections/collection_2016/fin/ F2-75-1987-2-eng.pdf ). 3 The budgets that proposed these measures, and all others from 1968 on, are available on a Department of Finance web page (www.budget.gc.ca/pdfarch/index-eng.html).

349 350 n canadian tax journal / revue fiscale canadienne (2018) 66:2

Some caution regarding the efficacy of a new tax commission is injected by Shirley Tillotson, of the Department of History at Dalhousie University. Drawing on the political history of tax commissions, Tillotson argues that tax commissions and the reforms that they generate will never be free from politics and lobbying by groups looking out for their own interests. Any resulting tax change will necessarily be shaped by this political process. In particular, we should expect that any government of the day will have its political objectives in mind when setting mandates and choosing commissioners. Tillotson closes by emphasizing the risks of large-scale reform, stemming from the difficulty of maintaining the necessary depth of public engagement in this era of fast-paced and superficial social media. Incremental reforms, in her view, are more likely to succeed. Given these natural political pressures, how should we structure a tax commission to achieve its objectives? Jennifer Robson, of Carleton University’s Faculty of Public Affairs, tackles this foundational question by comparing the advantages and disadvantages of different potential structures, ranging from internal reviews by public servants to a full-blown royal commission. She considers aspects such as access to data and expertise, mandate and accountability, and credibility. Done well, a tax commission can deepen knowledge, enunciate important principles, and provide technical solutions that can support sustainable and effective reform and embolden a government to take action. Done poorly, a tax commission’s work will be ignored. Because the demands placed on the tax system are ever-changing, Robson closes with the suggestion of a permanent body to study and analyze reforms to the tax system. The idea of a permanent body is explored in greater depth by Joseph Heath, of the School of Public Policy and Governance at the University of Toronto. The Bank of Canada operates with a large degree of independence from the government of the day because the pressures of politics intersect poorly with the highly technical field of monetary policy. Heath asks whether tax policy should not follow the same independent structure for the same reasons. He discusses the particular challenges of tax reform in the political context and grapples with the best way to ensure democratic legitimacy while capturing the benefits of technical expertise and guidance. Perhaps the government could set the main parameters of tax policy, such as the level of progressivity and the amount of tax revenue to be raised, and allow civil servants, independent of the government of the day, to work out the details in a non-political way. A new tax commission may offer the promise of constructive and lasting reform to Canada’s tax system. However, as is evident from the articles presented here, advocates for a tax commission need to be clear and realistic on what they expect such a commission to achieve, and be sure to build a structure for tax reform that is capable of delivering on that promise.

Alan Macnaughton and Kevin Milligan Editors canadian tax journal / revue fiscale canadienne (2018) 66:2, 351 - 62

Policy Forum: Why Canada Needs a Comprehensive Tax Review

Fred O’Riordan*

Précis Cet article traite de la question de savoir si le temps est venu de procéder à un examen complet de la politique fiscale au Canada. L’auteur propose tout d’abord qu’on procède à des examens périodiques de portée générale plutôt que d’élaborer une politique à la pièce dans le cadre du processus budgétaire annuel souvent politisé, et il établit quatre ensembles de principes ou d’attributs qui caractérisent un bon régime fiscal. Ceux-ci comprennent l’impartialité, l’équité et la légitimité; la neutralité et l’efficience économique; la simplicité, la faisabilité administrative et la certitude; et la compétitivité internationale. L’auteur examine ensuite les deux derniers ensembles de principes à l’aide d’un certain nombre d’indicateurs disponibles permettant d’établir comment notre régime fiscal existant se mesure à chacun de ces indicateurs. Il conclut que les résultats de cette analyse fondée sur ces deux seuls principes montrent que le régime fiscal existant a suffisamment besoin d’être amélioré pour que l’on procède à un examen complet de la politique fiscale. Il présente également quelques brefs commentaires corroborants sur les deux premiers ensembles de principes.

Abstract This article addresses the question of whether the time has come for a comprehensive tax policy review in Canada. The author begins with the proposition that periodic broad- based reviews are necessary as an alternative to piecemeal policy making done through the often politicized annual budget process, and he sets out four sets of principles or attributes that characterize a good tax system. These include fairness, equity, and legitimacy; neutrality and economic efficiency; simplicity, administrative feasibility, and certainty; and international competitiveness. The author then proceeds with an examination of the latter two sets of principles, using a number of available indicators of how well our existing tax system measures up against each of them. He concludes that evidence using these two principles alone suggests that there is sufficient scope for

* Of Ernst & Young LLP, Ottawa (e-mail: [email protected]); formerly an assistant commissioner of appeals at the Canada Revenue Agency. A portion of this article appeared in the Globe and Mail on January 9, 2018. I would like to thank the editors of this journal, Alan Macnaughton and Kevin Milligan, as well as Jack Mintz, for helpful comments on an earlier draft. Any remaining errors are solely my responsibility.

351 352 n canadian tax journal / revue fiscale canadienne (2018) 66:2 improvement in the existing system to merit a comprehensive policy review. He also provides some brief corroborative commentary on the first two sets of principles. Keywords: Tax policy n reviews n reforms n principles

Contents Introduction 352 Simplicity, Administrative Feasibility, and Certainty 353 International Competitiveness 357 Conclusions 360

Introduction Maintaining a mature tax system is a lot like maintaining a vintage house. The owner/occupant makes periodic improvements to keep the house in good running order and meet current family needs. But with the passage of time, more significant structural changes may have to be made, for various reasons. Perhaps the foundation is starting to crack, or the roof is starting to leak, or the heating and plumbing systems need an upgrade to meet newer technology and efficiency standards. Or it could be that changing family demographics make a baby’s room or a carriage-house “granny flat” a new priority. At the same time, if other properties, particularly the next-door neighbours’, start to look far more attractive than this vintage house, “keeping up with the Joneses” may become more than a cliché, since it is important to maintain the property’s value and ensure that it is not at risk, now or when the home is passed on to the next generation. If for any of these reasons major structural changes and improvements are required, the current owner/occupant engages professionals, including architects and engineers, for advice on design and construction, and consults with family members to ensure that their views and needs are taken into consideration. Like a vintage house, Canada’s tax system has been well maintained over the past 100 years, with periodic upgrades to suit the country’s changing economic and demographic circumstances. But adjustments have been made incrementally in the context of the annual budget process, and the different and sometimes conflicting priorities of successive governments have also meant that the tax system has evolved in a piecemeal fashion, too often for short-term political expediency and in the interests of the governing party, and at times in arguably contradictory or conflicting directions not necessarily consistent with the broader or longer-term needs of the country. The watershed historical policy interventions that have helped to rejuvenate the system and reconcile or resolve these conflicts have been comprehensive tax policy reviews and their reports and recommendations, most notably the Report of the Royal Commission on Taxation1 in 1966-67 (“the Carter commission”) and most

1 See Canada, Report of the Royal Commission on Taxation, vols. 1-6 (Ottawa: Queen’s Printer, 1966-67). policy forum: why canada needs a comprehensive tax review n 353 recently the Report of the Technical Committee on Business Taxation2 in 1997 (“the Mintz committee”), along with a number of broad reform exercises in the intervening years.3 While not all the recommendations of such reviews were implemented, the reviews nevertheless resulted in significant research, analysis, and findings that have enriched our knowledge and understanding of the impact and implications of tax measures, and have established important principles in reference to future tax policy decision making. Has the time come for another comprehensive tax review? The roof may not yet be leaking or the foundation crumbling, but for various reasons many informed and concerned observers think so.4 Consider the principles of a good tax system—the characteristics or attributes that reflect the best way to realize the underlying objectives of tax policy while causing the least harm. These principles provide a means by which to judge the performance of our current system and evaluate whether we have reached the point where a comprehensive policy review is advis- able. They fall fairly neatly into four groups, which include the following attributes:

1. fairness, equity, and legitimacy; 2. neutrality and economic efficiency; 3. simplicity, administrative feasibility, and certainty; and 4. international competitiveness.

Without disregarding the first two important groups of attributes, my view is that an examination of where we stand on the last two is sufficient to show clearly the need for a comprehensive policy review.

Simplicity, Administr ative Fe asibility, and Certainty Simplicity is the veritable Holy Grail of tax systems: ever sought after but ever elusive, always seemingly beyond reach. Simplicity is a relative concept that means different things to different people, but all taxpayers share a common expectation:

2 See Canada, Report of the Technical Committee on Business Taxation (Ottawa: Department of Finance, April 1998). 3 For example, at the provincial level, Ontario, Fair Taxation in a Changing World: Report of the Ontario Fair Tax Commission (Toronto: University of Toronto Press in cooperation with the Ontario Fair Tax Commission, 1993). For more recent provincial examples, see Québec, Focusing on Québec’s Future: Final Report of the Québec Taxation Review Committee (Québec: Gouvernement du Québec, March 2015) (“the Godbout committee”); and British Columbia, Commission on Tax Competitiveness, Improving British Columbia’s Business Tax Competitiveness (Victoria: Ministry of Finance, November 15, 2016). 4 Among them, the federal government’s own Advisory Council on Economic Growth. See the recommendation for an “overdue” targeted review: Canada, Advisory Council on Economic Growth, Investing in a Resilient Canadian Economy (Ottawa: Department of Finance, December 1, 2017), at 16-20. 354 n canadian tax journal / revue fiscale canadienne (2018) 66:2 that they can fulfill their tax obligations in a reasonable time and at reasonable expense, and soon thereafter obtain certainty about their tax status from tax authorities. For the majority of individual middle-income taxpayers who have deductions at source and few (or no) complicated family or financial arrangements, Canada’s current tax system could be judged sufficiently simple to meet those taxpayers’ self- assessment and voluntary compliance needs (if not their expectations). But this is largely because they are rarely or never exposed to the more complex areas of taxation. At the opposite end of the spectrum are large multinational corporations that have their own full-time tax staffs or engage tax experts in professional services firms for tax planning, preparation, and documentation. For them, compliance with increasingly complex provisions of tax legislation is much more challenging. And for many of them, self-assessment and filing a return is just the beginning of a long and protracted process of dealing with the tax authorities. But more on that later. Complexity is even an issue for Canadians at the low-income end of the spectrum. As the 2018 federal budget itself acknowledged,

the complexity of the tax system, low literacy and lack of access to available assistance are all barriers to tax filing among low-income individuals that can cause them to miss out on potential tax benefits. Indeed, according to a 2016 Prosper Canada survey of over 300 tax practitioners and experts, insufficient access to clinics and services and the high cost of commercial tax help were the most commonly cited barriers to tax filing among low-income Canadians.5

In order to address this problem, the budget proposed to double the size of the community volunteer income tax program, which provides assistance to modest- and low-income Canadians in preparing their income tax returns. The Canada Revenue Agency (cra), for its part, faces its own challenges in administering tax provisions that are becoming increasingly voluminous and complex. Indications of these difficulties may be found in the findings and recommendations of recent performance audits conducted on behalf of Parliament by the Office of the Auditor General of Canada oag( ). As an example, in a 2017 audit focusing on whether the cra’s call centres provided taxpayers with timely access to accurate information,6 the oag found that the cra gave taxpayers only limited access to its call centre services (including both the automated self-service system and call centre agents), blocking more than half of the calls received (about 29 million out of 53.5 million) because it could not handle the volume. Just as critically, the oag concluded that call agents gave inaccurate information to taxpayers 30 percent of the time.

5 Canada, Department of Finance, 2018 Budget, Budget Plan, February 27, 2018, at 275. 6 Office of the Auditor General of Canada,2017 Fall Reports of the Auditor General of Canada, report 2, Call Centres—Canada Revenue Agency (Ottawa: Office of the Auditor General of Canada, 2017) (www.oag-bvg.gc.ca/internet/English/parl_oag_201711_02_e_42667.html). policy forum: why canada needs a comprehensive tax review n 355

An oag audit of income tax objections in 20167 concluded that the cra did not process income tax objections in a timely manner, contributing to a significant backlog in inventory.8 Almost two-thirds of the objections reviewed by the cra’s Appeals divisions resulted in decisions that favoured taxpayers in full or in part.9 As a result of these favourable decisions, $6.1 billion out of a total of $11.6 billion of taxes in dispute was allowed to taxpayers, and most of the amounts claimed were allowed in full. In the same period, Appeals cancelled almost $1.1 billion in penalties and interest related to the objections. As the oag pointed out, this finding is important because each objection pro- cessed costs the cra and the taxpayer time and resources, and objections vacated in full or in part have a financial impact on the government because they reduce its expected revenues. They also tie up significant amounts of capital for large corporations, which, unlike other taxpayers, are required under the Income Tax Act to remit one-half of total tax assessed, even if they object to the assessment and final resolution of the dispute is not reached for years.10 The United States has a notoriously complicated tax code, so another indication of administrative complexity might be found by comparing the relative size of the cra and the Internal Revenue Service (irs). Other things being equal, one would expect that more complex legislation would translate into increased administrative requirements by the tax authority, a larger budget, and a bigger workforce. Given roughly 10 times the taxpayer population and size of the economy, and assuming 10 times the resulting workload, one might expect that the irs could be as much as 10 times the size of the cra, although there are clearly large economies of scale in tax administration that would narrow that size differential significantly. As of fiscal year 2015-16,11 actual irs expenditures were us $11.7 billion compared to Cdn $4.1 billion for the cra. In conducting its work, the irs utilized 77,924 full-time equivalent positions, compared to 37,977 for the cra.12 On the

7 Office of the Auditor General of Canada,2016 Fall Reports of the Auditor General of Canada, report 2, Income Tax Objections—Canada Revenue Agency (Ottawa: Office of the Auditor General of Canada, 2016) (www.oag-bvg.gc.ca/internet/English/parl_oag_201611_02_e_41831.html). 8 As of March 31, 2016, the CRA had an inventory of 171,744 objections outstanding for personal and corporate income taxes, which represented more than $18 billion of federal taxes. 9 Of the 223,739 objections resolved in the five-year period ending March 31, 2016, the CRA’s Appeals divisions dismissed 49,221 (as invalid objections) and reviewed 174,518. 10 See subsection 225.1(7) of the Income Tax Act, RSC 1985, c. 1 (5th Supp.), as amended. In some circumstances, the CRA will accept specified security in lieu of payment, but this also entails financial costs to the taxpayer. 11 October 1, 2015 to September 30, 2016 for the IRS; April 1, 2015 to March 31, 2016 for the CRA. 12 See United States, Internal Revenue Service, Internal Revenue Service Data Book, 2016 (Washington, DC: IRS, March 2017) (www.irs.gov/pub/irs-soi/16databk.pdf ); and “Spending and Human Resources,” in Canada Revenue Agency, Departmental Plan 2017-18 (Ottawa: CRA, 2017) (www.canada.ca/en/revenue-agency/corporate/about-canada-revenue-agency-cra/ departmental-plan/departmental-plan-2017-18-10.html). 356 n canadian tax journal / revue fiscale canadienne (2018) 66:2 basis of workforce size, assuming that the cra is roughly one-half the size of the irs in absolute terms, this means that in relation to processing a roughly equivalent workload, the cra is about 5 times larger than the irs. Of course, one should keep in mind that many American tax policy analysts contend that the irs is under resourced and that its funding has been declining in real-dollar terms (adjusted for inflation) for years.13 This simple comparison may instead, or also, indicate that the cra is better resourced for a more aggressive audit presence than the irs. As an example that may give some credence to this hypothesis, in resolving double-taxation cases under Canada’s network of bilateral tax treaties, of a total 526 negotiable cases resolved over the five fiscal-year period from 2010-11 to 2014-15, 459 cases or over 87 percent were initiated by cra audits, as opposed to 67 cases or just under 13 percent that were initiated by the foreign tax authority. Although the other tax authorities in these bilateral disputes are not identified by thecra , most of this caseload is with the irs.14 The cra’s resource base has been expanded even further since the time of this comparison in fiscal year 2015-16. The federal government allocated an additional $796 million to the cra over five years in the 2016 budget and $523.9 million over five years in the 2017 budget.15 In the 2018 budget, the cra received a further $515 million over five years, with an additional allocation of $41.9 million over five years and $9.3 million ongoing to the Courts Administrative Service, including “support for new front-line registry and judicial staff, most of whom are expected to support the Tax Court of Canada.”16 The Organisation for Economic Co-operation and Development’s (oecd’s) Centre for Tax Policy and Administration tracks and publishes comparative statistics for various member and non-member tax administrations. In 2007, for example, it found that the administrative costs of the cra were 1.22 percent of revenue collected compared to only 0.45 percent for the irs and that the cra’s tax administration costs were 0.212 percent of gross domestic product (gdp) compared to just 0.078 for the irs.17 Comparable statistics for more recent years are not published.

13 See for example, William G. Gale, “Steven Mnuchin Makes a Welcome Case for Boosting IRS Funding,” January 31, 2017 (www.brookings.edu/opinions/steven-mnuchin-makes-a-welcome -case-for-boosting-irs-funding/). 14 See Canada Revenue Agency, Mutual Agreement Procedure: Program Report 2014-2015 (Ottawa: CRA, 2015), at 13 (www.canada.ca/content/dam/cra-arc/migration/cra-arc/tx/nnrsdnts/cmp/ mp_rprt_2014-2015-eng.pdf ). 15 Canada, Department of Finance, 2016 Budget, Budget Plan, March 22, 2016, at 216, and 2017 Budget, Budget Plan, March 22, 2017, at 201. 16 2018 Budget Plan, supra note 5, at 69. 17 Organisation for Economic Co-operation and Development, Tax Administration in OECD and Selected Non-OECD Countries: Comparative Information Series (2008) (Paris: OECD, January 28, 2009), at 87-89, table 11, “Comparison of Aggregate Administrative Costs to Net Revenue policy forum: why canada needs a comprehensive tax review n 357

It would be wrong to conclude that these indicators in themselves prove that the cra is inefficient, or that existing legislation has become too complicated to administer effectively and efficiently, or that simplification would resolve all these problems. But they do provide evidence that administrative complexity may be contributing to increased compliance costs to taxpayers and administrative costs to government. These collectively could be characterized as a significant opportunity cost to the Canadian economy, consuming resources and talent that could have been used more productively in other ways or in other areas of the economy. Particularly troubling in this context is the high percentage of initial cra audit assessments and reassessments that are eventually overturned and reversed in whole or in part by Appeals. Related to tax simplicity is tax certainty. Simplicity enhances certainty, and certainty is of critical importance for financial and estate planning by individuals and for operational and investment decisions by businesses on behalf of their owners and shareholders. As tax provisions have become more complex, bright-line rules and tests that are simple to administer and comply with have become more difficult for tax authorities to write into law; more specific and detailed anti-avoidance rules are required, and more guidance and interpretation become necessary, as well as the exercise of more administrative discretion in decision making. In areas of tax that are imprecise or subjective, such as transfer pricing, there is no single right answer in determining tax liability, but a number of right answers within a fairly broad range. These com- plications invariably lead to more tax controversy, more domestic and international tax disputes to resolve, increased costs, and further delays before certainty is obtained. Tax disputes beyond the administrative objections stage typically take years to resolve. In a world with complex business relationships and taxable transactions, the Holy Grail of tax simplicity may ultimately remain elusive forever, but it is hard not to conclude that a comprehensive tax review with simplification as one of its objectives would be a quest worth undertaking at this time.

International Competitiveness Despite our status as a small, open economy, international tax competitiveness was not a major concern in Canada (or in many other countries, for that matter) at the time of the Carter commission in the 1960s. International trade liberalization was under way, but it had not yet gained the momentum that would be more fully realized decades later with the negotiation of the Canada-us free trade agreement (fta) and the North American free trade agreement (nafta), and more generally the phenomenon of globalization. Foreign investment in Canada, particularly American

Collections,” and at 90-91, table 12, “Revenue Body Expenditure as % of GDP (OECD Countries)” (www.oecd.org/ctp/administration/CIS-2008.pdf). The ratio for the CRA excludes its non-tax-related expenditure; non-tax-related expenditure for the IRS is zero. 358 n canadian tax journal / revue fiscale canadienne (2018) 66:2 foreign direct investment, was viewed back then with suspicion, and even open hostility in some quarters. Since then, the pendulum has swung in the opposite direction. Few Canadians now question the economic benefits of free trade and globalization. But with freer movement of capital globally, and well-integrated North American supply chains for most industries, retaining our international economic competitiveness in general, and the competitiveness of our tax system in particular, has become much more important. A company’s decision to invest is very sensitive to the rate of return on capital. Other things being equal, capital flows into jurisdictions where the rate of return is highest. Taxes imposed on businesses reduce the rate of return and affect both the amount and the location of investment undertaken. Since 2000, Canadian federal and provincial governments have gradually reduced business taxes to attract investment, primarily by implementing staged reductions in corporate tax rates, eliminating taxes on capital, and reducing taxes on business inputs. A measure of the effectiveness of this tax policy strategy is that, in spite of the rate reductions over this period, corporate tax revenues have continued to increase and the ratio of corporate taxable income to gdp has remained stable.18 Canada is not unique in regard to this policy direction. It is consistent with a global trend among many oecd and Group of Seven (g7) countries, the most notable exception being the United States.19 Until now, that is. In January 2018, the United States implemented the most sweeping package of legislative changes to its tax code in more than 25 years.20 The most notable element of the new us tax law is a reduction in both corporate and personal tax rates—the former on a permanent basis and the latter temporarily through 2025. How these changes will affect the United States, on balance, is subject to debate. Many expect that they will have a short-term stimulative effect on us economic growth, inbound investment, and job creation. Others express concern about their longer-term sustainability and effect on federal deficits, the national debt, and the after-tax distribution of personal income.

18 See Organisation for Economic Co-operation and Development, “Revenue Statistics—OECD Countries: Comparative Tables,” at chapter 4, “Countries—Tax Revenue and % of GDP by Selected Taxes” (https://stats.oecd.org/Index.aspx?DataSetCode=REV#). 19 Until the 2018 US tax reforms, the United States had the highest statutory corporate tax rate among the OECD countries. See Organisation for Economic Co-operation and Development, “Statutory Corporate Income Tax Rate” (http://stats.oecd.org/index.aspx?DataSetCode= TABLE_II1). For an analysis of the economic effects of a lower US rate, see Scott A. Hodge, “The Economic Effects of Adopting the Corporate Tax Rates of the OECD, the UK, and Canada,” Tax Foundation Fiscal Fact no. 477, August 2015 (https://files.taxfoundation.org/ legacy/docs/TaxFoundation_FF477.pdf). 20 An Act To Provide for Reconciliation Pursuant to Titles II and V of the Concurrent Resolution on the Budget for Fiscal Year 2018 (known as the Tax Cuts and Jobs Act), Pub. L. no. 115-97, enacted on December 22, 2017. policy forum: why canada needs a comprehensive tax review n 359

Less debatable is the impact that these changes will have on Canada. To the extent that they spur the us economy, Canada will share some of the benefit through increased exports, owing to the integrated nature of the two economies, provided that access to the large us market is not disrupted by stalled NAFTA renegotiations. However, there is a strong risk that these US tax reforms will have a negative overall economic impact, drawing capital and skilled labour away from Canada to the larger and more tax-competitive US market. The us federal corporate income tax rate has now fallen from 35 percent to 21 percent, compared to the Canadian federal rate of 15 percent. More importantly, the gdp-weighted-average combined federal-state corporate tax rate in the United States has fallen from 39.1 percent to 26 percent and is now slightly below the average combined Canadian federal- provincial rate of 26.7 percent—completely eliminating Canada’s competitive tax advantage. A useful measure of this erosion in Canada’s competitive position, more informative than a simple statutory-rate comparison, is a comparison of the marginal effective tax rate (metr) on new business investment between the countries. The metr includes not only the corporate tax rate but also deductions and credits associated with purchasing capital goods and other capital-related taxes paid by the corporation. It measures the extra return on investment necessary to pay these taxes and maintain the same total return or, put differently, the share of the gross- of-tax rate of return on a marginal unit of capital needed to pay the business taxes on that capital. Canada has had a lower metr than the United States since 2007.21 This has been tremendously beneficial to Canada, influencing not only where businesses choose to locate but, for those multinational firms with significant cross-border operations and activity, where to place the highest-value corporate functions and thereby report their related revenues and pay the associated proportion of corporate taxes in the lower-tax jurisdiction. us tax reform has sharply reduced the country’s aggregate metr from 34.6 percent to only 18.8 percent compared to Canada’s 20.9 percent.22 Disaggregated by industry, Canada’s wide relative tax-competitive advantage has been eliminated, and the United States now has the advantage in all but three sectors—manufacturing, oil and gas, and “other services.”23

21 Philip Bazel, Jack Mintz, and Austin Thompson, “2017 Tax Competitiveness Report: The Calm Before the Storm” (2018) 11:7 SPP Research Papers [University of Calgary School of Public Policy] 1-40, at appendix A, table A.3. 22 Ibid., at 2. 23 Statistics Canada, “North American Industry Classification System (NAICS) Canada 2012,” definition 81, “Other Services (Except Public Administration)” (http://www23.statcan.gc.ca/ imdb/p3VD.pl?Function=getVD&TVD=118464&CVD=118465&CPV=81&CST= 01012012&CLV=1&MLV=5). 360 n canadian tax journal / revue fiscale canadienne (2018) 66:2

As if this reversal in effective tax rates were not bad enough economic news, the most recent statistics for foreign direct investment in Canada are additional cause for concern. On March 1, 2018, Statistics Canada reported:

For the year 2017, direct investment in Canada amounted to $33.8 billion, the lowest level of investment since 2010 and well below the record of $126.1 billion observed in 2007. Cross-border mergers and acquisitions generated a withdrawal of funds from Canada for the first time since 2007, when these data started to be compiled.24

When the impact of the us tax reforms is factored into capital allocation decisions over the next year and beyond, direct investment in Canada could be further eroded. Labour is less mobile than capital, but us personal tax reforms and associated rate reductions may also pose a threat to Canada’s tax competitiveness, particularly when retaining and attracting the best and brightest talent is a top priority for Can- adian companies seeking to maintain their own competitiveness.25 The loss of our international tax competitiveness vis-à-vis the United States is perhaps the most compelling reason to undertake a comprehensive tax review, so that Canada can consider and act on the appropriate policy responses on an urgent basis. This loss could have quicker, broader, and more dire implications for the Canadian economy as a whole than all the shortcomings in achieving the other principles reviewed here combined. This does not necessarily mean that tax rates need to be drastically reduced in a race to the bottom, as some observers fear and caution against. On that count, Canada is still roughly in the middle of the pack as far as other Group of Twenty (g20) and oecd countries are concerned,26 and there is plenty of scope for base-broadening measures as a substitute for or complement to any rate reductions.

Conclusions There are four distinct sets of principles or attributes of a good tax system, and this examination of just two of them—simplicity, administrative feasibility, and certainty; and international competitiveness—shows clearly that there is ample room and need for improvement in Canada’s current tax system to merit a comprehensive tax policy review.

24 Statistics Canada, “Canada’s Balance of International Payments, Fourth Quarter 2017,” The Daily, March 1, 2018 (www.statcan.gc.ca/daily-quotidien/180301/dq180301a-eng.htm?HPA=1). 25 Non-tax factors also influence the location decisions of skilled workers, but there is some empirical evidence that personal income tax rates have a robust negative effect on cross-border flows of skilled workers in the OECD. See Peter Egger and Doina Maria Radulescu,The Influence of Labor Taxes on the Migration of Skilled Workers, CESifo Working Paper no. 2462 (Munich: Center for Economic Studies and Ifo Institute, November 2008) (www.econstor.eu/ bitstream/10419/26507/1/589230298.PDF). 26 See Bazel et al., supra note 21, at 28. In 2017 (prior to the US tax reforms), Canada had the 6th lowest METR among G7 countries, the 14th lowest among the G20 countries, and the 12th lowest among 34 OECD countries. policy forum: why canada needs a comprehensive tax review n 361

Before closing, I will add a few brief comments on how our tax system might also measure up against the other two sets of attributes, starting with fairness, equity, and legitimacy. How taxpayers perceive the fairness and integrity of the tax system is important because it can affect their voluntary compliance, the cornerstone of our system. If trust and confidence are eroded, behavioural changes can reduce compliance rates and significantly increase enforcement costs. High tax burdens in particular can cause adverse behavioural responses by taxpayers who perceive that they are being overtaxed. This is especially relevant now that Canada’s highest combined federal-provincial marginal rates on ordinary personal income exceed 50 percent in 6 of Canada’s 10 provinces.27 The public’s opinion may be influenced by media reports of high-wealth individuals sheltering income in offshore tax havens, or multinational corporations allegedly paying less than their fair share of business taxes, but Canadians would be far better informed by a broad empirical analysis of individual tax burdens and business tax incidence done in the context of a comprehensive policy review. In terms of the neutrality and efficiency attributes, Canada has made significant improvements in the 50 years since the Carter commission, but many inefficiencies remain and there is scope for further improvement. A plethora of tax expenditures still exist, making a veritable Swiss cheese of the tax system.28 These tax expenditures were designed to achieve once-valid social or economic policy objectives; however, such measures often entail unintended costs in terms of economic efficiency, and they are not necessarily good substitutes for direct expenditure programs, which could achieve the same objectives more transparently, with more public scrutiny and accountability. Those tax expenditures that have been analyzed have often been found ineffective in achieving their stated objectives.29 The four sets of principles are interrelated, and the pursuit of competing policy goals entails inevitable tradeoffs among them (for example, redistributing income versus maximizing economic growth). If anything, the existence and added complexity of navigating these tradeoffs provide an additional strong argument in

27 Manitoba (50.40), Ontario (53.53), Quebec (53.31), New Brunswick (53.30), Nova Scotia (54.00), and Prince Edward Island (51.37)—rates in effect in 2017. For an analysis of behavioural responses, see Office of the Parliamentary Budget Officer,The Fiscal and Distributional Impact of Changes to the Federal Personal Income Tax Regime (Ottawa: Parliamentary Budget Officer, January 21, 2016) (www.pbo-dpb.gc.ca/web/default/files/Documents/ Reports/2016/PIT/PIT_EN.pdf ). 28 Canada, Department of Finance, Report on Federal Tax Expenditures: Concepts, Estimates and Evaluations (Ottawa: Department of Finance, 2017) (www.fin.gc.ca/taxexp-depfisc/2017/ taxexp-depfisc17-eng.pdf ). The report provides fiscal cost estimates and projections for some 209 personal and corporate income tax and goods and services tax expenditures, and contains an analysis of trends for the period 1991-2015. While the value of these tax expenditures expressed as a share of tax revenues and of GDP declined sharply in the 1990s and early 2000s, this trend reversed in 2001, and both ratios have followed an upward trend since then (apart from a drop during the 2008 financial crisis). 29 Examples are the children’s fitness tax credit and the children’s arts credit: ibid., at 303-19. 362 n canadian tax journal / revue fiscale canadienne (2018) 66:2 favour of periodically tackling tax policy in a broad review context rather than only making ad hoc piecemeal changes. Finally, the federal Department of Finance’s own polling suggests that there may be a growing public mood for the government to tackle tax reform as a priority.30 In a poll done in the summer of 2017, respondents were asked on a 10-point rating scale how much of a priority each item in a list of suggested initiatives should be for the government. The top choices were “creating jobs” (favoured by 87 percent of respondents), “increasing economic growth in Canada” (86 percent), “making the richest Canadians pay their fair share in taxes” (76 percent), “making the tax code fairer” (75 percent), and “making Canada more competitive internationally” (75 percent).

30 See Bill Curry, “Canadians Say Ottawa’s Deficit Approach Is ‘Wrong’: Survey,”Globe and Mail, March 7, 2018; and Quorus Consulting Group, Summer 2017—Survey and Focus Groups on the Economy: Final Report, prepared for the Department of Finance (Ottawa: Quorus Consulting Group, November 24, 2017), at 72-73 (http://epe.lac-bac.gc.ca/100/200/301/pwgsc-tpsgc/ por-ef/finance/2018/010-17-e/report.pdf ). canadian tax journal / revue fiscale canadienne (2018) 66:2, 363 - 74

Policy Forum: Then and Now— A Historical Perspective on the Politics of Comprehensive Tax Reform

Shirley Tillotson*

Précis Dans cet article, Shirley Tillotson analyse quatre types de contingences qui déterminent les possibilités en matière de politique fiscale, premièrement par rapport aux réformes Carter-Benson de 1962 à 1971, et deuxièmement relativement aux perspectives actuelles d’une réforme fiscale complète. Comparant les deux moments, l’auteure conclut que la situation n’est pas propice à la mise en oeuvre d’une réforme fiscale complète mobilisant l’opinion publique.

Abstract In this article, Shirley Tillotson analyzes four kinds of contingencies that determine political possibility in tax matters, first in relation to “the Carter-Benson moment,” 1962-1971, and second in relation to present-day prospects for comprehensive tax reform. Comparing the two moments, she concludes that we are currently not well situated for a comprehensive tax reform process that engages public opinion. Keywords: Tax reform n reviews n royal commissions n politics n public

Contents Introduction 364 Configuration of Taxpayer Interests: Changes in Tax Law and Practice Generate New Tax Publics 365 Intersection with Other Kinds of Politics: Cold War Oppositions and Incipient Populism 367 Kinds of Expertise: Macroeconomics, Meet Micro-Politics 368 Means of Political Communication: The Expansion of Group Politics 370 Impact 372 Current Implications 372

* Of the Department of History, Dalhousie University, Halifax (e-mail: [email protected]).

363 364 n canadian tax journal / revue fiscale canadienne (2018) 66:2

Introduction Canada has had no golden age of tax politics, when public opinion was well informed, fair minded, and inclusively representative of all interests. We are not in one now. And yet, though history can make us cynically despair, it shouldn’t. History prompts humility, to be sure, and usually counsels against panic; but most usefully, it requires us to see politics and policy choices as contingent on circumstances. Different periods afford different opportunities and threats. What can’t be done in one period may be done in another. In this brief article, based on some of the research reported in my recent book,1 I suggest that we should consider in our current moment the historical context of the Carter-Benson period, 1962-1971. I provide some selected examples of how that context shaped what was then possible and what was not. On the basis of this comparison, I suggest why 2018 is probably not a propitious time to launch comprehensive tax reform, at least not by means of a reform process that will engage public opinion. Building political consensus around broad and transformative tax reform faces formidable barriers, now as then. In speaking of the “Carter-Benson” moment, I use the names of two men—the prominent and publicly minded accountant Kenneth Le Mesurier Carter and Finance Minister Edgar J. Benson—to refer to a process of comprehensive tax review in which both played a leading role. Carter was the chief of a Royal Com- mission on Taxation that was launched in 1962 and reported in 1966-67.2 Benson helmed a subsequent process of public consultation on the 1969 white paper on tax reform3 that led to a new federal income tax statute in 1971.4 My analysis of their historical moment is organized around four kinds of contin- gency that determine political possibility in tax matters:

1. Configuration of taxpayer interests. Though always broadly organized around property and poverty, interests realign as markets, demographics, and non- tax institutions change. 2. Other aspects of electoral politics. Alignments ranging from sectarianism to regionalism to feminism complicate the strategic ground for political actors, and often connect to taxpayer interests.

1 Shirley Tillotson, Give and Take: The Citizen-Taxpayer and the Rise of Canadian Democracy (Vancouver: UBC Press, 2017). 2 Canada, Report of the Royal Commission on Taxation, vols. 1-6 (Ottawa: Queen’s Printer, 1966-67) (herein referred to as “the Carter commission”). 3 E.J. Benson, Proposals for Tax Reform (Ottawa: Queen’s Printer, 1969) (herein referred to as “the white paper”). 4 Income Tax Act, SC 1970-71-72, c. 63. A related and significant change was the termination of the Dominion Succession Act and the parallel measures that gradually eliminated provincial succession duties, until they were all gone in 1985. This process is described in David G. Duff, “The Abolition of Wealth Transfer Taxes: Lessons from Canada, Australia, and New Zealand” (2005) 3:1 Pittsburgh Tax Review 71-120. policy forum: then and now—the politics of comprehensive tax reform n 365

3. Kinds of expertise. Changes in economics as a discipline and the nature of the data available shape the kinds of imaginable options for tax policy and administration. 4. Means of political communication. Not only the kind of knowledge that is available but also how that knowledge is disseminated change over time, affecting the actors and relationships of policy change.

In the Carter-Benson moment, then as now, historically specific contingencies of these types set limits on tax reform.

Configuration of Taxpayer Interests: Changes in Tax L aw and Pr actice Generate New Tax Publics To understand the Carter-Benson moment, we need to consider its past. In a period stretching from the late 1940s through the 1960s, new publics were forming in tax politics.5 The result was both wider engagement on income tax questions and increasing alienation and conflict. Among the new tax publics was a cohort of low- income wage and salary earners. In their number also figured the small-business owners who, in the 1948 Income Tax Act,6 first enjoyed a special lower corporate income tax rate. Another increasingly anxious population of income taxfilers during the 1950s were the owners of substantial investment property, for whom the non- taxation of capital gains was a valued means of countering the impact of marginal tax rates that had been reduced only modestly from their wartime levels.7 Both the lower-income and the wealthy taxfilers would shape the politics of the Carter- Benson moment. By 1962, when the Carter commission began, there were proportionately more low-income taxpayers than there had been in 1949. Slow but fairly steady inflation, spiking after 1965, had reduced the purchasing power of the basic exemption, which remained unchanged from 1949. This growing class of low-income taxpayers were told from the 1940s onward that their many small income tax contributions funded social security, though in fact specific charges such as provincial health taxes and unemployment insurance premiums played a large part. In the background, the federal sales tax made a healthy contribution to the general revenue.8 Still, paying “their share” through the new federal personal income tax (new in its mass scope of incidence in peacetime) became the foundation of claims to federally funded social

5 Tillotson, supra note 1, at 210-37. 6 SC 1948, c. 52. 7 J. Harvey Perry, Taxes, Tariffs, & Subsidies: A History of Canadian Fiscal Development, vol. 2 (Toronto: University of Toronto Press, 1955), at 411 and 692, table 34. 8 Stated as a percentage of gross national product, the federal sales tax was a close third by size among the sources for federal fiscal requirements (aside from debt). See W. Irwin Gillespie, Tax, Borrow, and Spend: Financing Federal Spending in Canada, 1867-1990 (Ottawa: Carleton University Press, 1991), at 288, table C-3. 366 n canadian tax journal / revue fiscale canadienne (2018) 66:2 services. Opposition to this exchange of tax for services found a political home only among the Créditistes, a rural and working-class populist party in Quebec. In 1962, the Créditistes were calling for the married breadwinner’s personal exemption to be raised to $5,000, which would have ended the mass income tax.9 In spite of its otherwise broad basis for consent, the federal income tax after 1948 brought many Canadians into a fraught relationship with the state as the practices of tax administration slowly evolved. Once the 1948 Act was in place, the tax authority signalled a more vigorous enforcement program. In newspaper and magazine stories from the late 1940s through the 1950s, readers were told that “tricks” used to evade income tax were unlikely to work. For small-business owners, the threat of stepped-up enforcement was especially important. Along with owners of professional practices, they would learn that this better-staffed tax authority was prepared to teach Canadians that income tax reporting and paying were serious legal obligations. Pre-war habits of paying what seemed a “reasonable” amount on a roughly estimated income, rather than an amount based on solid accounts, were no longer acceptable. After the war, the tax authority actively pursued fraud prose- cutions, and investigations reached back into the war years. One rural Cape Breton merchant learned to his distress that, contrary to his own estimates, he had in fact made a taxable profit every year for the 10 years since 1936, and the tax bill on those profits was overdue. In 1950, a Saskatchewan farmer worried that the truck he had purchased with his veteran’s allowance would be seized for an income tax debt. In 1957, one judge described a physician’s wartime tax evasions as tantamount to trea- son. Frictions like these gave politicians such as the new Progressive Conservative (pc) Party leader, John G. Diefenbaker, material that they could use to woo voters.10 For taxfilers at the other end of the wealth scale, a particularly bitter war was fought over capital gains during the post-war period. As most readers of this journal will know, the partial inclusion of capital gains in taxable income was first introduced in the 1971 revision of the statute. But in the 1950s, when the gain from selling capital assets was usually tax-free, disputes flared between property-rich taxfilers and the tax authority over whether particular transactions generated a capital gain or income taxable under the Act. A tax adviser would provide one answer, and the tax authority would dispute the accounting involved. Shock. Horror. Surely tax obligations should not be so uncertain! Commenting on a 1950 case in which a sup- posed capital gain was taxed as income, the Globe and Mail’s editors struck an almost American note of tax outrage:

The very essence of tyranny is the levying, at the whim of officials, of taxes not clearly specified or intended by the law. Our national Government, as is its wont, is seizing with enthusiasm an opportunity to employ the tyrannical method.11

9 Tillotson, supra note 1, at 210-13 and 280-82. 10 Ibid., at 222-27, 243, and 279-80. 11 “Taxation by Stealth,” editorial, Globe and Mail, May 23, 1950, quoted in Tillotson, supra note 1, at 235. policy forum: then and now—the politics of comprehensive tax reform n 367

My tone of gentle mockery here is not meant to discount the real problem of uncertainty in tax incidence, a problem that would be addressed in successive revisions of the statute. But it must be acknowledged that in the 1950s accountants and lawyers specializing in tax were beginning to actively market methods of con- verting income to capital gains without it passing through the tax collector’s hands. Small wonder, then, that inclusions in taxable income were contested. Tax advisers and their clients constituted a greatly expanded element of public opinion in the 1950s and 1960s. Their avoidance practices signalled the limited consent to steeper and higher graduation of personal income tax rates at the upper reaches of the income scale.12 All of these are now mostly ordinary troubles, but they were new ones in the 1950s. As John Stuart Mill had predicted, a nation of income taxpayers could generate a fierce political wind.13

Intersection with Other Kinds of Politics: Cold War Oppositions and Incipient Populism Those who felt the frictions of post-war tax administration found a useful set of political tools in Cold War vocabulary. This early period of the Cold War, peaking in the 1959 Cuban revolution and the us reaction to Cuba, generated an intense polar opposition, succinctly expressed as “freedom versus Communism.” Tax complaint flowed readily into this form. Militant Catholic anti-Communism helped to fuel anti-tax feeling and opposition to social spending in rural and even urban working class Quebec, providing voters for the anti-statist Créditistes. In the early 1950s, Diefenbaker himself floated small-state rhetoric about “keep[ing] the tax collectors poor.”14 Tax cutting, especially for the vote-rich ranks of lower income earners, was part of the Diefenbaker pcs’ program after their 1957 election win. By the June 1962 election, Finance Minister Donald Fleming and Prime Minister Diefenbaker were pushing stories about personal income taxes being raised by “30 to 40” percent, or 50 percent, or even 60 percent if the Liberals were elected. The choice, Diefenbaker asserted, was between free enterprise and confiscatory socialism. Canada’s Chamber of Commerce joined in with its 1962 “Operation Freedom.” It aimed to fight high taxes and excessive social security (“excessive” before medicare, the Canada Pension Plan/Quebec Pension Plan, or student loans), and to overcome Canadians’ “apathy and indifference with respect to freedom.”15 Toronto’s Board of

12 Tillotson, supra note 1, at 214-16 and 233-35. 13 John Stuart Mill, The Principles of Political Economy, 7th ed. (Oxford: Oxford University Press, 1994), at 238-39. 14 Canada, House of Commons, Debates, December 10, 1951, at 1724-25, quoted in Tillotson, supra note 1, at 248. 15 “Face Threats to Freedom, Nation Asked,” Globe and Mail, July 10, 1961, quoted in Tillotson, supra note 1, at 283. 368 n canadian tax journal / revue fiscale canadienne (2018) 66:2

Trade firmly discouraged Operation Freedom, which was drawing accusations of John Birchism from the newly formed New Democratic Party (ndp). Calling the Chamber’s campaign naïve and suited only to small towns, Toronto’s business leaders misread its appeal. In the 1962 election, the weekly press of rural English Canada vigorously supported this anti-tax anti-Communism; and in Quebec, similar small towns and not so small ones gave the Créditistes enough members of Parliament (mps), along with four western Canadian Social Credit members, to hold the balance of power in the new federal Parliament.16 This was the election in which the pcs made comprehensive review of the tax system the main plank of their platform. Diefenbaker announced it in terms that spoke artfully to different audiences, audiences whose interests were not only dispa- rate but opposed. His government, if re-elected, would appoint a royal commission that would “examine anomalies in the existing laws, consider inequitable tax burdens, close loopholes that now existed and ease hardships now caused by some tax laws.”17 Canadians who loathed rich tax dodgers and thought that exemptions were too low could have heard that message as a commitment to do something for them. Tax practitioners and their clients who worried that capital gains were not reliably protected or that compliance costs were too high would also have heard a promise in “examine anomalies” or “ease hardships.” The comprehensive review was eventually delivered by Carter’s Royal Commis- sion on Taxation, but not before the Diefenbaker government was defeated in 1963—on its own failings, to be sure, but also by the Créditiste/Social Credit decision to vote with Lester B. Pearson’s Liberals. The dark energies of populism had given Diefenbaker two big victories in 1957 and 1958, with unprecedented voter turnouts, but in 1962 those same currents put the knife into the Créditistes’ hands, and they wielded it soon after, turfing the Tories. When comprehensive tax reform was proposed in 1969, a Cold-War-inflected populist public opinion was still present and ready to pounce.

Kinds of Expertise: Macroeconomics, Meet Micro-Politics Meanwhile, among the experts, on the eve of the move toward comprehensive tax reform, economists were feeling their scientific oats. In 1961, a group of economics professors challenged the governor of the Bank of Canada, James Coyne, to a show- down on macroeconomics. They exposed splits in the policy community that had been already made public by the report in 1957 of the Royal Commission on

16 See Tillotson, supra note 1, at 248-49 and 283-85. Many small town weekly newspapers published a syndicated column of the conservative journalist Ambrose Hills (pseudonym of Walter A. Dales). See, for example, “Of Many Things. Election Opportunities,” Kingsville Reporter, May 10, 1962. 17 “Will Review Tax Laws, Dief Says. Main Tory Plank Revealed by PM in Keynote Speech,” Winnipeg Free Press, May 7, 1962, quoted in Tillotson, supra note 1, at 285. policy forum: then and now—the politics of comprehensive tax reform n 369

Canada’s Economic Prospects (“the Gordon commission”).18 Some of the divisive questions were, how serious a threat was inflation; what level of unemployment was tolerable as “frictional”; and could fiscal and monetary policy be coordinated (the former a matter of politics; the latter, central banking)?19 Regarding the last question, the Gordon commission thought yes, but acknowledged that there were difficulties and “honest differences of opinion” among government, the central bank, and other agencies.20 The commissioners acknowledged that government intervention in the business cycle was still “a new art which [was] relatively untried.”21 And they con- fessed that the federal government’s economic interventions could be undone by the junior governments: this was “a real problem for the future” and one to which, they bluntly said, they saw no solution.22 As remedies for the other points of conflict, the Gordon commission called for a beefed-up, more rapidly reporting Dominion Bureau of Statistics, more economics research, and a permanent, non-partisan council of experts to work through differences, to educate the voting public, and to give voice to a credible consensus on economic questions.23 The commission acknowledged, in other words, that on key economic issues, ones in which tax was intimately involved, voters might well be unable (or unwilling) to take expert guidance, as then constituted, as their guide. This perception of a profession divided and as yet uncertain can only have been reinforced when economics professors and the governor of the Bank of Canada were so publicly at odds. Often, in the 1960s, when finance ministers were challenged on “high taxes,” they replied that the record of post-war prosperity surely meant that the level of taxation was doing no harm.24 But as economics, that argument was pretty thin. The Gordon commission had deemed tax to be too big a topic to include in its work. But the conflicts alluded to in its report foreshadowed the troubles that later beset Carter and Benson. The Carter commission’s report read as the work of mod- ernist men of economic science in white hats, cleaning up the fiscal chaos caused by the black-hatted politicians. While Carter’s broad program and later Benson’s narrower proposals offered something for every voter, there was no unifying macro good that would carry decisive weight in all circles, as price stabilization had during the Second World War. Instead, reactions fractured Benson’s every proposal into a promise for some and a threat for others, mostly at a microeconomic level. The

18 Canada, Royal Commission on Canada’s Economic Prospects, Final Report (Ottawa: Queen’s Printer, 1957). 19 Ibid., at 421-33. 20 Ibid., at 436. 21 Ibid., at 426. 22 Ibid., at 435. 23 Ibid., at 433-36. 24 Tillotson, supra note 1, at 299. 370 n canadian tax journal / revue fiscale canadienne (2018) 66:2 promise that income tax reform could reduce the economic vulnerability of the poor was compelling to some, but a sign of state socialism to others. One sector’s essential incentive—say, the small business tax rate—was to other taxpayers both ineffective and a pointless drain on the public revenue. Tax subsidies to mining and petroleum, offering a tax haven to investors and clearly distortionary in one view of taxation, from another perspective were an attack by Ottawa on a regional industry. Most radically, the proposed methods of taxing capital gains generated intense objections, even among some who acknowledged that it was probably time to include capital gains in taxable income.25 The breadth of the Carter commission’s vision and the subsequent Benson budget proposals meant a mobilization of equally unprecedented breadth from opponents of change. Neither the scale of the reform nor the scale of the reaction could have hap- pened earlier. Modern economics had been successfully represented as connecting taxation to a vast array of economic effects and even, by means of incentives, to economic behaviour that reached into the social and moral realms. Tax reform was no longer a matter of tweaking the tax structure to manage public finance and more fairly distribute the burden by region, sector, and class. It was about all of that, but it was now also about child care, home ownership, education, retirement planning, and the role of the state, not to mention Canadian control of the economy and the redistribution of wealth for a more just society.26 There were even a few ultra- Protestants still banging the drum about favours to Catholics extended through the non-taxation of priests’ incomes. For all that macroeconomics had enhanced the policy voice of economists, the breadth of macroeconomics and microeconomics combined also exposed tax policy to an enormous array of policy interests.

Means of Po litical Communication: The Expansion of Group Politics Many of these new connections between tax policy and other areas had found or would soon find organizational weapons. In a process that Paul Pross documented in Group Politics and Public Policy,27 governments since the 1930s had encouraged Canadians to form peak organizations for different social and economic interests, to facilitate interaction with the state. Labour, manufacturing, and farmers had long been organized, but in the 1950s and 1960s, civil society mobilized on a new scale.28 For example, Canada’s previously scattered pensioners’ organizations had formed the National Pensioners’ and Senior Citizens’ Federation (np&scf). Its member- ship of over 200,000 featured prominently in Justice Minister John Turner’s

25 Ibid., at 287-88 and 294-301. 26 Ibid., at 231-32, 259-75, 296-97, 298-99, and 303-4. 27 A. Paul Pross, Group Politics and Public Policy, 2d ed. (Toronto: Oxford University Press, 1992). 28 Ibid., at 38-60 and 64-68. policy forum: then and now—the politics of comprehensive tax reform n 371 representation to Benson of its policy wishes.29 The white paper debate was an accelerant to tax organization. To the left of np&scf, Canadian Pensioners Con- cerned would be formed to advocate on tax matters, especially for low-income seniors.30 Anti-tax activism flourished. The white paper prompted a Toronto man, John Bulloch Jr., to organize the Canadian Council for Tax Fairness, soon renamed the Canadian Federation of Independent Business. The somewhat shadowy Equit- able Tax Foundation would help to pay for the publication of a high-profile tax practitioner’s objections to the white paper in a book titled The Benson Iceberg.31 The Department of Finance used all of the now-abundant means of managing public relations to engage these many publics, and the better bankrolled of the white paper’s critics did too. National television broadcasts gave Finance Minister Benson’s message a wide reach, but also amplified his critics. One especially well- funded opponent was Colin Brown, an insurance executive from London, Ontario. He ran a massive newspaper ad series with clippable coupons for readers to use in letters to Benson. He also paid for a remarkably biased opinion poll to feed his story that everyone was opposed to the Benson reforms. Brown’s lasting legacy was the founding of the National Citizens’ Coalition shortly after the 1971 budget.32 Throughout the debates of 1969 to 1971, Benson argued, with reasonable numbers, that most taxpayers would be better off under the proposals and that the income tax burden would follow “ability to pay” more closely. But those who stood to lose from the reforms found in the mechanisms for achieving those good goals alarming threats, some perhaps real, some certainly imaginary, to the pocketbook interests of the mass public. Under the leadership of people such as Bulloch and Brown, a significant electoral force was mobilized. And the Liberals’ 1971 budget showed its impact. Many small tax expenditures served to allay at least some of the popular opposition, and a compromise version of capital gains taxation was achieved, blended with the end of federal succession duties.33 The newly complex statute launched a thousand tax-practice careers and, no doubt, a half-dozen research programs in the economics of taxation. Comprehensive reform had been minced into many pieces.

29 On the disorganized state of pensioners in the early 1950s, see James Snell, The Citizen’s Wage: The State and the Elderly in Canada, 1900-1951 (Toronto: University of Toronto Press, 1996), at 181-85; and Library and Archives Canada, Department of Finance fonds, RG 19, vol. 5222, Turner to Benson, February 25, 1971, “White Paper Tax Reform Proposals from the Public Generally.” 30 Kenneth Kernaghan and Olivia Kuper, Coordination in Canadian Governments: A Case Study of Aging Policy (Toronto: Institute of Public Administration of Canada, 1983), at 33. 31 I.H. Asper, The Benson Iceberg: A Critical Analysis of the White Paper on Tax Reform in Canada (Toronto: Clarke, Irwin, 1970). See Tillotson, supra note 1, at 299, 303, and 217-18. 32 Tillotson, supra note 1, at 293-94. 33 Ibid., at 305. 372 n canadian tax journal / revue fiscale canadienne (2018) 66:2

Impact The Carter-Benson moment was the first really comprehensive review of income taxation in Canada, and many have regarded it as a tragedy, a triumph of orches- trated public opinion against rational reform. But it served one really useful purpose: It provoked an unprecedented discussion of many important questions about tax fairness and the role of government in the economy and in citizens’ personal welfare, questions that had been quietly brewing since the mass-based, peacetime federal income tax was enacted in 1948. That discussion revealed that the post-war consensus on the role of the state was neither solid nor entirely inclusive.34 For all that the Carter-Benson moment exposed important facts about Canadian political opinion, the review failed to achieve comprehensive tax reform. It did not usher in a new era of effective, simple, economically rational, and socially just taxation. The reforms it produced were barely more than incremental, and there was much left on the tax policy docket, including what to do about the manufacturers’ sales tax, a question not addressed until the late 1980s. Beyond the important effect on group politics noted above, the review process may have had a short-term effect on party politics. In the 1972 election, the Liberals were reduced to a minority government, and the ndp, describing the tax reforms as a failure, increased its seats by just over 40 percent (though the Tories harvested the largest number of new mps).

Current Implications Our present contingencies seem likely to produce similarly lively conflict around comprehensive tax reform and to limit the range of achievable reforms. Certainly we are more rather than less likely to experience the damaging role of hidden sources of big money in any contemporary tax reform project. Worries on that score have only mounted since the 1960s. In 1981, at a Conference Board of Canada event, some high-powered public relations, public administration, and public opinion experts with an interest in Canadian democracy, along with a number of politicians, discussed whether to regulate private spending on public policy campaigns. The issue was framed as “Advocacy advertising: propaganda or democratic right?” One participant, Prakash Sethi, warned that paid “speech” by big business and big government could “overwhelm public communication space” and “drive out alternative viewpoints.”35 He asked if there might be a need to “protect the public from messages that might be totally one-sided, inaccurate, deceptive, or misleading.”36 Those present solemnly agreed, in a mildly worried way, that substantial spending on advocacy advertising could make an issue important regardless of its “intrinsic merit.”37

34 Ibid., at 303-4. 35 Duncan McDowall, ed., Advocacy Advertising: Propaganda or Democratic Right? (Ottawa: Conference Board of Canada, Public Affairs Research Division, May 1982), at 65. 36 Ibid. 37 Ibid., at 99. policy forum: then and now—the politics of comprehensive tax reform n 373

With the personal computer revolution just beginning in 1981, these would-be custodians of Canadian democracy could not have known that their misgivings about the communications environment for tax politics were about to be confirmed on a new order of magnitude. Today we are all aware that the management of public opinion on policy questions is a big-budget enterprise and that some of those spending money to shape public opinion are acting out of private rather than public interests. An added wild card is the use of low-cost Internet campaigns, whose viral effects are unpredictable. Manipulative mobilizations exaggerate and distort the consequences of proposed tax reforms, and recruit to partisan purposes potential voters who generally do not, and with the information available to them probably cannot, discern the real consequences of those proposals. Tax publics in our historical moment resemble those who were emerging in the 1960s—skeptical of expertise, battered by unexplained (or difficult to understand) macroeconomic events, readily recruited on ideological lines, and then covertly (now more openly) manipulated by persuaders with deep pockets. Incremental changes might make it possible to avoid the atmosphere of overheated crisis that accompanies complex and far-reaching changes and that creates opportunities for manipulators. Comprehensive tax reform now would attract at least as wide an array of political mobilizations as the Benson white paper did. Both income taxation and sales taxation are thoroughly folded into the tissue of incentives that shape economic and social behaviour. In the age of behavioural economics, we mostly accept that “the economy” is not a realm of straightforwardly rational actors. Powerful commit- ments, whether moral or cultural, shape our perception of taxes, not only as means but also in relation to purposes, both private and public. About purposes, citizens may vigorously disagree. We argue from situated standpoints whose reconciliation entails tough emotional and political work, not just reason and evidence.38 In this respect, some voices in tax today resemble the voice of nationalism in the tax reform politics of the Carter-Benson moment: motivated by economic goals but weighty in public opinion, partly because of deep engagements around identity. Although economic nationalism (until recently) has lost its political weight, one might argue that the tax treatment of families inherits its role as a tax question that mixes culture and economics. The Carter-Benson moment highlighted a primary schism—the class politics of a welfare state—but our political schisms have since gone fractal. Inter- ests configure in complex ways. If reforms can be introduced in stages, then the work of engaging the relevant tax publics might be simpler, and less likely to produce a massive pile-on. The duration of the Carter-Benson “moment” is something else to consider in weighing the contingencies that determine what is politically possible. From mandate in 1962 to legislation in 1971, the tax reform project of the 1960s was buffeted by events, including, notably, us tax reform, inflation driven by the Vietnam War, and a Canadian constitutional crisis rising to a boil. By 1971, Finance Minister Benson could no longer point to a stable economy to defend taxation practices.

38 Tillotson, supra note 1, at 296-99 and 302. 374 n canadian tax journal / revue fiscale canadienne (2018) 66:2

Inflation worries, a point of macroeconomic controversy in 1961, were by 1969-70 a powerful factor in mobilizing opposition to full taxation of capital gains.39 Events overtook policy. How much more likely would that be now in a comprehensive tax reform process that would probably take at least as long as Carter, given the now much larger volume of research and data that commissioners would have to engage with? Perhaps, like the Macdonald commission of the mid-1980s, a vast and expen- sive research effort would fairly quickly produce, with only a bit of manufactured dissent, expert conclusions toward which the commissioners were already predis- posed.40 But that would surely be no more satisfactory as a democratic process than would prolonged and transformative deliberation by a commission whose results might be overtaken by economic or political events that altered the stakes and prompted realignments, as did the Créditiste surge of 1962-63 or the inflation of the late 1960s. And the effect of a pitched battle for 5 or 10 years over a big tax reform project risks exacerbating the problem of inflamed and distorted political communication that we face in the fragmented state of our present world. Though my analysis of the Carter-Benson moment has emphasized the risks attached to comprehensive tax reform, we may truly need such reform—a well thought out vision of improvement that considers tax policy in relation to both economic and social development and draws on solid evidence. In the framing of that vision, I must defer to others. But in comparison with the Carter-Benson moment, today the prospect of having adequate evidence seems better: we now have more economic and taxation data, and much greater data-processing capacity. The recent decision of the Canada Revenue Agency to release tax gap data to the parliamentary budget officer is promising. We have a deeper base of scholarship on tax, not only in economics but also in history, law, sociology, accountancy, politics, and philosophy. Our experts are more socially diverse, making it more likely that the expertise they develop will be responsive to the real complexity of our world. All this is hopeful. But the political work of proposing tax policy to the public has never been more difficult. Of all the factors shaping the evolution of tax politics, the recent changes in the means of political communication represent the single most distinctive aspect of our historical moment. The changing business model of journalism and the uncurated world of social media pose known dangers and worrying uncertainties. In this context, it is hard to be confident that major tax changes can be presented in ways that will be understood and will command consent. Absent those conditions, tax changes may compromise compliance, exacerbate avoidance and evasion, and ultimately undermine any real process of accountability. Incremental change, responsive to the contingencies of the moment, seems to me most likely to avoid those risks.

39 Ibid., at 300-1. 40 Gregory J. Inwood, “Of Leaps of Faith and Policy Change: The Macdonald Royal Commission,” in Gregory J. Inwood and Carolyn M. Johns, eds., Commissions of Inquiry and Policy Change: A Comparative Analysis (Toronto: University of Toronto Press, 2014), 113-29, at 122 and 127-28. canadian tax journal / revue fiscale canadienne (2018) 66:2, 375 - 86

Policy Forum: Building a Tax Review Body That Is Fit for Purpose—Reconciling the Tradeoffs Between Independence and Impact

Jennifer Robson*

Précis Dans ce court article d’opinion, j’aborde diverses options pour structurer et soutenir un examen du régime fiscal fédéral au Canada, y compris les interactions avec les mesures fiscales infranationales. Je me concentre en particulier sur les choix qui doivent être faits entre l’indépendance du gouvernement et l’influence sur l’élaboration de la politique gouvernementale. Les modèles traditionnels auxquels nous pensons généralement — les examens internes par les représentants du gouvernement, les groupes de travail et les commissions royales — comportent tous des forces et des faiblesses comme mécanismes d’examen complet, transparent, indépendant et influençant du régime fiscal. En plus de ces options, je considère des exemples internationaux, tel le rapport Mirrlees au Royaume-Uni. Enfin, je suggère de créer un organisme subventionné par l’État ayant pour mandat permanent de réaliser des analyses, de faire participer les parties prenantes, et de formuler des recommandations stratégiques. Il est peu probable que la réforme fiscale canadienne soit un exercice unique dans le temps, et il serait donc opportun de former un organisme d’examen fiscal qui se consacrera à cette tâche.

Abstract In this brief opinion piece, I discuss various options for structuring and supporting a review of the federal tax system in Canada, including interactions with subnational tax measures. I focus in particular on the tradeoffs that must be made between independence from government and influence on government policy making. The traditional models that we are accustomed to thinking of—internal reviews by government officials, task forces, and royal commissions—all have strengths and weaknesses as mechanisms for conducting a comprehensive, transparent, independent, and influential review of the tax system. In addition to these options, I consider international examples, such as the United Kingdom’s Mirrlees review. In the end, I suggest the creation of a publicly funded body with a standing mandate to conduct analysis, engage stakeholders, and make

* Of the Faculty of Public Affairs, Carleton University, Ottawa (e-mail: jennifer.robson @carleton.ca).

375 376 n canadian tax journal / revue fiscale canadienne (2018) 66:2 policy recommendations. Tax reform is unlikely to be a one-time policy need in Canada, and so we should build a tax review body fit for purpose. Keywords: Tax reform n commissions n accountability n governance n policy making n politics

Contents Introduction 376 First Principle: Form Should Follow Function 378 Second Principle: Form Should Support Policy Change 380

Introduction In thinking about the prospects for a comprehensive review of Canada’s tax system, we should not overlook the important questions of how such a review should be structured, resourced, and run. However meritorious the exercise of undertaking a fresh critique of our tax system, however expert those who carry out the task, a review is unlikely to be successful if it is not built for purpose. Regardless of how we define “success” in an ambitious tax review exercise—and there may not be consensus on the policy goals or priorities for a tax review—we have to give at least some thought to administrative questions such as accountability and mandate. Previous reviews of Canada’s tax system have often taken the form of royal commissions. Notable examples include the 1945 Ives commission on the taxation of annuities1 or, more prominently, the Carter commission from 1962 to 1966 (discussed in the preceding article by Shirley Tillotson).2 But important exercises in tax reform have taken administrative forms other than a royal commission. In recent years, federal governments have appointed a number of advisory panels and task forces to make recommendations on tax policy, though generally within a much more limited mandate. This was the case with the 1997-98 Technical Committee on Business Taxation3 and the 2006 advisory body to the minister of finance that recommended amendments to the Income Tax Act4 to create a new tax-preferred savings instrument, the registered disability savings plan (rdsp).5 Sometimes recommendations for specific tax reforms emerge from bodies appointed with a mandate to examine a policy issue where the tax system makes up a critical part of the federal policy tool- box. For example, in 2011, the government received recommendations for reform

1 Canada, Report of the Royal Commission on the Taxation of Annuities and Family Corporations (Ottawa: King’s Printer, 1945). 2 Canada, Report of the Royal Commission on Taxation, vols. 1-6 (Ottawa: Queen’s Printer, 1966-67). 3 Canada, Report of the Technical Committee on Business Taxation (Ottawa: Department of Finance, April 1998) (herein referred to as “the Mintz report”). 4 RSC 1985, c. 1 (5th Supp.), as amended. 5 Canada, A New Beginning: The Report of the Minister of Finance’s Expert Panel on Financial Security for Children with Severe Disabilities (Ottawa: Department of Finance, December 2006). policy forum: building a tax review body that is fit for purpose n 377 of the scientific research and experimental development sr ( & ed) tax credit from a task force on innovation in Canada, appointed by the minister of state (science and technology).6 Such royal commissions and task forces (or advisory panels) are external to government, though they operate with different degrees of independence and different timelines for conducting research and issuing recommendations. Finally, exercises in tax reform can also be conducted internally by officials working within the federal public service, to provide confidential advice to the government, generally through the minister of finance. Sometimes this internal work is made visible through a public report such as a white paper. This was the case in 1987, when then Finance Minister Michael Wilson released a white paper on tax reform that preceded major changes to personal, corporate, and consumption taxes in the 1988 federal budget.7 Similarly, in 1969, in response to the recommendations of the Carter commission, then Finance Minister Edgar Benson tabled a white paper on tax reform.8 However, in the latter case, many of the proposals failed to materialize as government policy. As one member of Parliament noted during the debates on the sweeping changes proposed in Benson’s white paper, Canadians “like better the devil they know than the devil unknown.”9 There are also myriad examples of tax reforms that have been conducted as internal exercises but without a public report before policy changes are made. These internal policy debates can include input from external advisers, as was the case in the 2016 federal review of tax expenditures,10 but that has been the exception rather than the rule. The historical record suggests at least two overarching and related questions. First, how far at arm’s length should a tax review body be from the government of the day? And second, if the review is to generate advice that will be adopted and acted upon by the government of the day, how close should the body conducting the review be to the institutions of government? In the remainder of this brief article, I discuss the tradeoffs between independence and impact, and offer some thoughts on the administrative questions of accountability, resourcing, and mandate for an imagined comprehensive national tax review exercise. For the purposes of this discussion, I am going to set aside the question of what a tax review should accomplish and instead focus on how a tax review might be accomplished. While there may be many readers of this journal who would agree on the need for a comprehensive and public tax review, the motivations for a review

6 Independent Panel on Federal Support to Research and Development, Innovation Canada: A Call to Action (Ottawa: Public Works and Government Services Canada, 2011). 7 Canada, Department of Finance, The White Paper: Tax Reform 1987 (Ottawa: Department of Finance, June 18, 1987). 8 e.J. Benson, Proposals for Tax Reform (Ottawa: Queen’s Printer, 1969). 9 Canada, House of Commons, Debates, June 23, 1971, at 7303 (comments by Gordon Blair). 10 I was appointed by the minister of finance to a panel of advisers tasked with providing peer review of analysis conducted by officials in the Department of Finance during the 2016 tax expenditure review. 378 n canadian tax journal / revue fiscale canadienne (2018) 66:2 are likely to be varied. For example, some may prefer a tax review aimed at creating incentives for business development and aggregate economic growth. Others may prefer a review aimed at reducing economic inequalities and enhancing redistribution or predistribution of resources. Yet others will see value in reforms to simplify tax measures and make compliance easier. There are many other possible policy objectives for tax reform, each with its own merits. But if there is disagreement on why a review is warranted, there is, I think, broader agreement on how a tax review ought to be conducted. So it is in that area of agreement—the “how”—that I will advance the rest of my brief argument.

First Principle: Form Should Follow Function The administrative structure of the body charged with reviewing Canada’s tax system should be determined by what that body is expected to do. When observers talk about a body charged with recommending comprehensive tax reform, there seem to be at least three features that they want this body to have. I believe that these design criteria can help us to determine how independent a body charged with tax reform needs to be. First, the body should deliver a package of proposals for comprehensive tax reform including personal and corporate income taxes and consumption taxes. Recent reviews have focused on a particular tax measure, such as the task force that reviewed the sr & ed credit and the expert panel that led to the creation of the rdsp.11 Instead, the mandate of the body would be to conduct a system-wide review of federal taxation in Canada, including interactions with subnational taxation. Second, the work and recommendations of a review should be transparent. In other words, the body conducting the review and making recommendations should make its analysis public, so that it can be scrutinized by other experts and debated by stakeholders. There should be opportunities for other experts to review the work of the tax reform body and opportunities for stakeholders to understand what is being proposed before any changes are implemented. Third, the package of proposals for comprehensive reform should be grounded in a clear statement of principles for optimal tax design that is independent of current policy. As Boadway has argued, principles articulated by the Carter commission, such as ability to pay and the basis of taxation (income versus consumption), continued to inform tax policy long after they were presented in the commission’s report.12 Proposals for tax reform, whether sweeping or specific, are subject to a set of assumptions about what economic activity should be taxed and who should pay.

11 While the Mintz committee issued an ambitious set of recommendations (including changes to corporate income and capital taxes, as well as payroll and excise taxes), it did not have a mandate to look at consumption or personal income taxes. See supra note 3. 12 Robin Boadway, “Policy Forum: Piecemeal Tax Reform Ideas for Canada—Lessons from Principles and Practice” (2014) 62:4 Canadian Tax Journal 1029-59. policy forum: building a tax review body that is fit for purpose n 379

Proposals are also guided by theory or shared beliefs about what constitutes a good tax system. These assumptions, theories, or beliefs should be made explicit. In addition, where longstanding assumptions are being replaced, the work of the tax review body should make it clear why new assumptions are necessary or at least preferable. For example, the final report of the United Kingdom’s Mirrlees review described criteria for evaluating tax systems, reviewed principles of optimal tax theory that would be familiar to most tax scholars, and offered certain rules of thumb for tax design.13 The three features described above argue for a body that can conduct and publish analysis, and issue detailed reports to the Canadian public. They argue for a structure that is designed to receive and make use of input from stakeholders and experts outside the government. They also argue for an advisory body that will be able to discuss the goals of the Canadian tax system independently from the policy goals of the government in power. In general, this suggests a structure located outside the federal public service and independent from the government in power. Public servants in the Department of Finance have significant expertise in tax policy and will have working relationships with some external stakeholders. How- ever, officials in the department will be limited in their ability to share their analysis and advice publicly. Public servants have access to unique forms of information, such as administrative files and tax records that cannot be freely shared for reasons of privacy and security of information, or because publication is subject to approval by the minister. To be successful, a review of the tax code must not put public servants in open conflict with the core executive they serve. To do so would put at risk the ongoing ability of public servants to work in a privileged relationship with the core executive, in which they are able to provide fearless advice and be trusted to provide loyal implementation of a government’s decisions. Asking public servants to share sensitive analysis publicly and without ministerial approval would also erode the principal mechanism of government accountability in our system—ministerial accountability to Parliament.14 In turn, if information sources cannot be readily shared, this limits opportunities for stakeholders to scrutinize analysis and methods. Departmental officials simply do not enjoy the independence needed to conduct a comprehensive, unfettered, and transparent review of the tax system. They are subject-matter experts and keepers of unparallelled Crown information, but their contribution must be made within the confines of the bargain between the government and the officials who serve it. This bargain also applies in circumstances where public servants are temporarily assigned to support a task force or commission.

13 James A. Mirrlees, Stuart Adam, Tim Besley, Richard Blundell, Stephen Bond, Robert Chote, Malcolm Gammie, Paul Johnson, Gareth Myles, and James Poterba, Tax by Design: The Mirrlees Review (Oxford: Oxford University Press, 2011). 14 In some circumstances, research papers prepared by federal employees can be published in academic journals. They cannot, however, include information not otherwise sanctioned for public release; they must be shared in advance with supervisors; and generally they cannot make or even imply recommendations for policy change. 380 n canadian tax journal / revue fiscale canadienne (2018) 66:2

Instead, the best options for a comprehensive, transparent, and independent review of Canada’s tax system seem to be housed outside the day-to-day operations of the government. In practice, external exercises in tax reform have taken several forms, ranging from a ministerial task force, to a royal commission, to an entirely independent project undertaken by a private think tank. Each of these enjoys a different degree of independence from and access to the government. In the section below, I turn to the question of impact, and consider how close to the institutions of government a body charged with tax reform might need to be in order to effect change.

Second Principle: Form Should Support Policy Change It seems reasonable to suggest that discussions promoting a comprehensive review of Canada’s tax system are starting from the premise that some important changes to the current system are necessary. We hardly need a national review of taxation if we agree that the current system is working well. But if the goal is to support policy change, to have impact, what does this imply for the structure of the body in charge of the review? The Mirrlees review in the United Kingdom presents an interesting model in that it was conducted entirely outside government. It was initiated by the Institute for Fiscal Studies (ifs), an independent and non-partisan think tank, and funded by a mix of research grants from the national academic granting council (the Economic and Social Research Council) and philanthropic funding from a charitable trust (the Nuffield Foundation).15 The ifs provided administrative support for the 2010 conference to launch the project, including the collection of conference papers,16 and for the publication and dissemination of the final collection of papers in 2011.17 In all, 63 authors contributed to the review, including many of the leading researchers in tax policy. Following the review, a parliamentary committee was prompted to launch its own inquiry on the principles of taxation. The work of the review was cited by several government officials and helped to inform the 2012 debate on the harmonization of various income-tested benefits delivered as refundable credits.18 The Mirrlees review took on the task of imagining an optimal tax system for an open economy in the 21st century. The two collections of papers, Dimensions of Tax

15 Institute for Fiscal Studies, “About the Mirrlees Review” (www.ifs.org.uk/publications/ mirrleesreview/about). 16 Stuart Adam, Timothy Besley, Richard Blundell, Stephen Bond, Robert Chote, Malcolm Gammie, Paul Johnson, Gareth Myles, and James Poterba, eds., Dimensions of Tax Design: The Mirrlees Review (Oxford: Oxford University Press, 2010). 17 Tax by Design, supra note 13. 18 University College London, Office of the Vice Provost for Research, “The Mirrlees Review: Influencing Policy and Debate on Taxation” (www.ucl.ac.uk/impact/case-study-repository/ mirrlees-review-and-benefit-reform). policy forum: building a tax review body that is fit for purpose n 381

Design and Tax by Design,19 continue to be an important reference for students, researchers, and policy makers. The review was not limited to recommending only those tax changes that would have the greatest chance of enjoying political support, but the researchers clearly recognized that their recommendations alone would not be sufficient to motivate policy action:

Government in a media-driven democracy is difficult and there is a need to work within the bounds of the politically feasible. But there is a better way to make tax policy. There are taxes that are fairer, less damaging, and simpler than those we have now. To implement them will take a government willing to be honest with the electorate, willing to understand and explain the arguments, willing to listen to and to consult experts and public alike, and willing to put long-term strategy ahead of short-term tactics.20

Some of the same features that gave the review academic credibility—access to granting council support, affiliation with a well-regarded think tank, and, most importantly, the involvement of leading scholars—also raise considerations about accountability and mandate. Think tanks and scholars are free to pursue research on tax policy, to make recommendations, and even to pursue a project as ambitious as the Mirrlees review. But without a public mandate, their work risks becoming only a reference document—albeit, perhaps, an important one—rather than a set of recommendations that exert influence over tax policy. A primarily academic exercise is also likely to suffer from difficulties in gaining the engagement and support of other stakeholders and, in particular, the general public. Academics, think tanks, and other researchers have much to contribute to policy processes, and not simply as observers and faithful critics; but they do not have the same formal accountabili- ties as elected officials or even public servants. The Mirrlees review did not seem to attract the interest of external stakeholders whose views would later play a role in parliamentary debate and legislative amendments. It appears that the Mirrlees review was able to complete its analysis removed from the political pressures familiar to ministers of finance, tax policy officials, and parliamentarians considering tax changes. For those conducting research in a more academic context, the freedom to work outside political influence is essential to the independence and academic credibility of their results. But there are downsides to working in isolation from outside interests and stakeholder lobbies, including the absence of political capital to support the recommendations. In 2017, when the uk government again tried to advance a major reform of the tax system, informed in part by the work of Mirrlees on questions of inheritance, exclusion of dividend income, taxation of non-residents, and corporate taxation, it was forced to abandon substantial parts of its agenda. Neither Mirrlees nor the government had been successful in building a public consensus on tax reforms.

19 Dimensions of Tax Design, supra note 16; Tax by Design, supra note 13. 20 Ibid., at 503. 382 n canadian tax journal / revue fiscale canadienne (2018) 66:2

Given the present all-consuming focus on Brexit, it seems unlikely that a uk government of any political stripe will spend much political capital on tax reform in the near future. The mandate to build public consensus and, in turn be held to public account, requires a public organization to lead the tax review effort. The mandate may come from government or from Parliament itself, but it cannot be self-appointed. This brings us to consider at least two forms of ad hoc public bodies have previously been used for tax reform exercises: task forces (or advisory panels) and royal commissions. The task force or advisory panel model is a poor fit for the project of comprehensive tax reform. As discussed above, recent examples show a tendency to use task forces and advisory panels for very narrowly defined policy questions that are concerned with either a particular problem in tax policy (such as the proliferation of spending measures through the tax code) or a particular problem with respect to which tax policy is a key instrument to the government (such as financial support for persons with disabilities or for research and development). Furthermore, such ad hoc bodies are appointed by a minister to report to or advise that same minister. Their mandates are generally limited in scope, as well as in duration, encouraging the appointed members to wrap up their work quickly, generally with the support of a small cadre of officials seconded from the minister’s department. The work of the panel or task force is completed when the members issue their final report; there is no mechanism to hold them to account for their analysis or advice. The government, for its part, is free to take or leave the advice given—or, more likely, to cherry-pick the parts that it finds useful as support for its preferred policy approach. In some cases, a public review of taxation can be initiated and completed within the mandate of a single government. For example, the George W. Bush administration established a task force to review the us tax code that commenced its work in January 2005 and reported in November of the same year.21 In contrast, Canada’s Carter commission was announced in 1962 by the Diefenbaker government and issued its report four years later to the Pearson government. A full formal response to the Carter recommendations did not come until 1969, when Pierre Trudeau’s government issued its policy papers on tax reform. Between the publication of the Carter report and the enactment of the Tax Reform Act in 1972,22 Canada experienced a turnover in federal finance ministers as well. On the one hand, the American example suggests that a rush to complete a review within a constrained political time frame may significantly limit the scope and ambition of the project.23 On the other hand, the Carter example suggests that one-time exercises in reform that experience multiple turnovers in government may ultimately see their policy impact diluted.

21 President’s Advisory Panel on Federal Tax Reform, Simple, Fair, and Pro-Growth: Proposals To Fix America’s Tax System (Washington, DC: US Government Printing Office, 2005). 22 SC 1970-71-72, c. 63. 23 See, for example, Scott Greenberg, “The Bush Tax Reform Panel, Ten Years Later,” November 2, 2015 (https://taxfoundation.org/bush-tax-reform-panel-ten-years-later/). policy forum: building a tax review body that is fit for purpose n 383

Mechanically, royal commissions are created by an order in council, drafted by Cabinet and signed by the governor general. Under the Inquiries Act,24 their authority comes directly from the Crown; they are able, within the terms of their mandate, to compel evidence (similarly to a court) but may not adjudicate disputes or impose sanctions on parties appearing before them.25 By law, commissioners are granted access to any public office and any government record. This access could not, for example, violate the privacy of individual Canadians, but the investigative powers of a commission would presumably facilitate access to federal tax data. A tax review body that does not have access to complete and accurate microdata, as well as administrative information, is at a significant analytical disadvantage compared to the Department of Finance. For example, the Mirrlees review had no privileged access to British tax data and was therefore limited to making broad or directional recommendations, rather than specific ones that could be tested against the current system. In addition to the power to gather information, commissioners can hire technical experts and assistants, as they feel is necessary, and may draw resources for their work from the government’s consolidated revenue fund. Parliament cannot hold a government to account for the activities or findings of a commission once it is launched, but it can ask the government to account for the purpose, mandate, resourcing, and response to a commission of inquiry. This last feature has, at times, made royal commissions an attractive mechanism for governments looking to delay decisions or avoid difficult policy choices. At first glance, these features of independence, transparency, and access to important information might seem to make a royal commission the ideal mechanism through which to complete a review of the tax system. Given the existing model of the Carter commission, it would be tempting to return to the same structure today. However, several concerns have been raised about the use of “temporary adhocra- cies.”26 For example, Doern has listed a number of common criticisms of royal commissions, including their cost and the tendency for them to be used by governments seeking to delay action.27 Osberg has argued that temporary advisory bodies have very little time to conduct any real analysis in the brief window of time between starting up, receiving public input, and issuing a report.28 Aucoin has argued that, by the early 1970s, royal commissions came to be seen as a “relatively clumsy” mechanism and, since then, governments have tended instead to keep deeper dives

24 RSC 1985, c. I-11, as amended. 25 Thomas J. Lockwood, “A History of Royal Commissions” (1967) 5:2 Osgoode Hall Law Journal 172-209. 26 Henry Mintzberg, The Structuring of Organizations (Englewood Cliffs, NJ: Prentice-Hall, 1979). 27 G. Bruce Doern, “The Role of Royal Commissions in the General Policy Process and in Federal-Provincial Relations” (1967) 10:4 Canadian Public Administration 417-33. 28 Lars Osberg, “Inquiry Without Fact-Finding” (1987) 18:3 Policy Options. 384 n canadian tax journal / revue fiscale canadienne (2018) 66:2 on policy issues in-house to the public service and core executive.29 Aucoin has also asserted that royal commissions are no more able than public servants to overcome the dominant paradigms in a field of study at a given time. While commissions of inquiry may be particularly “useful instruments for governments . . . confronted by intractable or ‘wicked’ (as opposed to ‘tame’) problems,” they cannot be a substitute for processes within government itself.30 Doern goes further and suggests that royal commissions are “not adequate to face a modern federal state’s policy problems, most of which are of a recurring nature.”31 A royal commission may be an effective and appropriate mechanism to investigate branches of government, to allow grievances and opinions be heard on a particular topic, or to gather information on which to base a one-time policy choice. They are, as Doern argues, public, ad hoc, and investigatory in their operation, and they are dismantled once the work is done.32 However, as in the case of a task force, there is no real way to hold a royal commission accountable for its analysis and advice once its work is completed. Despite their long (indeed medieval) history,33 royal commissions seem ill-suited for complex, persistent policy problems that may not be resolved by a one-time set of recommendations. Tax policy is a wicked policy problem, and it is also, most certainly, an ongoing one. Tax reform cannot, in my view, be done as a single exercise, shaped by a single report and a set of recommendations that will be expected to stand for as long as successive governments are content to avoid another review. As Doern wrote, “there is need for a permanent body of expertise related to recurring social and economic problems.”34 While some may hope that a royal commission might, for a time, rally public attention to the issue of tax reform or motivate stakeholders to participate in an inquiry, I am not so optimistic. Observers of and participants in public policy debates have no shortage of demands for their time and attention. In fact, a royal commission could easily be drowned out by any number of news cycles and competing high-profile events. And, as noted earlier, the short lifespan of a royal commission makes it difficult to get much work done in the time between setting up and shutting down. An ongoing body has at least two further advantages in advancing the goal of policy change. First, a standing external body is not very useful to a government

29 Peter Aucoin, “Royal Commissions and Task Forces as Mechanisms of Program Review” (1987) 2:2 Canadian Journal of Program Evaluation 1-10, at 4. 30 Ibid., at 5. 31 Supra note 27, at 431. 32 Ibid., at 417. 33 Lockwood, supra note 25, at 172 and 180, notes that the first royal commission in Canada was appointed in 1848 (before Confederation), but that the English tradition of an inquiry mandated by and reporting to the Crown dates back to 1080, when royal commissioners were sent by the king to assess the ownership and value of all taxable property in the kingdom, with the results being compiled in the Domesday Book. 34 Supra note 27, at 432. policy forum: building a tax review body that is fit for purpose n 385 looking to delay a decision or action. There is no end point to which the government can ask the Opposition, the media, or other critics to look before expecting the government to act in response. In short, a standing body can inform future policy choices, but it also gives no quarter to a government that is already aware of some important tax changes that it could make. Second, a standing external body becomes formally part of the policy-making system, albeit at arm’s length from the government. Over time, this allows public servants, whose confidential analysis and advice continue to provide critical input to political decision makers, to develop working relationships with—or even spend time working inside—another public body that, like the public service, is here for the long term.35 Canada has had at least one example of an advisory body that has managed the delicate balance of having proximity and influence while also enjoying political independence, the ability to engage more easily with external stakeholders, and the freedom to make public recommendations on matters of policy. From 1963 to 1993, the Economic Council of Canada (ecc) was one of a small number of arm’s-length national bodies with the public authority to collect information, conduct analysis, and make recommendations to the federal government on ongoing policy questions. Established by statute,36 the ecc was structured as a Crown corporation, accountable to Parliament through a report to a responsible minister, but independent from the direction of that minister. It was required, by statute, to produce at least one annual report and was otherwise able to conduct and publish analysis and recommendations as it saw fit. Employees in the federal public service were able to work for the ecc for a limited period before moving on (or returning) to various federal departments and agencies.37 The ecc attracted high-calibre chairs whose considerable intellectual heft contributed to the quality of the published work. Council members were generally diverse in their professional experience and expertise (but not adequately in terms of gender, ethnicity, and other forms of diversity). The ecc actively engaged with stakeholders and experts outside government, many of whom likewise spent time working at the council before returning to the private or academic sectors. It was also a prolific publisher of research and analysis. For much of its life, the ecc worked to “form a consensus on the political economy.”38

35 A permanent body would also need clear and legislated access to federal tax data. In the same way that Canada Revenue Agency makes microdata available to the Department of Finance and to Statistics Canada, rights to microdata could be made available to a new federal body that would also be subject to federal authority and oversight on the handling, use, and release of private data. The experience of the parliamentary budget officer (PBO) in obtaining access to microdata is instructive here but is at least partially explained by the PBO’s different lines of accountability compared to Crown departments and agencies. 36 economic Council of Canada Act, SC 1963, c. 11 (since repealed). 37 R.W. Phidd, “The Economic Council of Canada: Its Establishment, Structure and Role in the Canadian Policy-Making System, 1963-74” (1975) 18:3 Canadian Public Administration 428-473. 38 Ibid., at 456. 386 n canadian tax journal / revue fiscale canadienne (2018) 66:2

On the issue of comprehensive tax reform, we ought to look for similar opportunities to form consensus if we want recommendations for tax reform to be enacted into policy. Such consensus will take more public engagement than the public service is permitted. It will take more time and independence than a ministerial task force or advisory panel has available. And it will take a collaborative, cross-sectoral way of working, rather than a quasi-judicial one. Finally, decision makers and the public should be able to ask those who do the analysis and issue recommendations to defend and account for their work, and to remain formally engaged in the debates that follow. We are sorely in need of an updated look at our system of taxation—one that is as ambitious as the review conducted by the Carter commission more than 50 years ago. But even if a new exercise succeeds, we are likely to need a further update sooner than we expect. Better that we should plan ahead and design a structure that can meet ongoing needs for comprehensive, transparent, and independent analysis and advice than spend another five decades hoping and waiting. canadian tax journal / revue fiscale canadienne (2018) 66:2, 387 - 99

Policy Forum: From Independent Tax Commission to Independent Tax Authority

Joseph Heath*

Précis L’impôt est un domaine dans lequel les impératifs d’une « politique bonne » et ceux de la « bonne politique » divergent toujours. Cette situation en a amené certains à proposer que l’on dépolitise la politique fiscale en créant une administration fiscale indépendante (afi) qui serait responsable d’établir la politique fiscale générale et fonctionnerait indépendamment du gouvernement du jour. Un tel changement représenterait de toute évidence un pas vers une forme plus « technocratique » de gouvernement et susciterait de la résistance pour ces mêmes raisons. Afin de faire progresser la discussion, je commence cet article en avançant qu’une afi répond à un problème réel, soit l’exploitation par les politiciens de l’ignorance et de l’irrationalité du public en ce qui a trait aux impôts et aux taxes. Cette situation donne lieu à des inefficacités et à des inégalités qui sont intégrées dans le code des impôts, ainsi qu’à des pertes de temps et d’énergie passés à constamment détourner le débat public sur des pseudo-problèmes. Après avoir motivé la proposition, je présente ensuite les grandes lignes de la structure institutionnelle qu’une afi pourrait prendre au Canada, le genre de mandat qu’il lui serait confié et les types d’impôts et de taxes qu’elle contrôlerait. Je conclus en présentant certaines des objections les plus évidentes et en tentant d’y répondre.

Abstract Taxation is one area in which the imperatives of “good policy” and those of “good politics” consistently diverge. This has given rise to the occasional call to depoliticize tax policy, through the creation of an independent tax authority (ita) that would be responsible for setting general tax policy and would operate with substantial independence from the government of the day. Such a change would obviously represent a move in the direction of a more “technocratic” form of government and will be resisted on those grounds. In order to advance the discussion, I begin this article by arguing that an ita represents a response to a genuine problem, which is the exploitation by politicians of public ignorance and irrationality with respect to taxes. This leads to inefficiencies and inequalities that are introduced into the tax code, and also wasted time and energy through the constant diversion of public debate into arguments over pseudo- issues. Having motivated the proposal, I then outline the basic institutional structure that

* Of the School of Public Policy and Governance, University of Toronto (e-mail: joseph.heath @utoronto.ca).

387 388 n canadian tax journal / revue fiscale canadienne (2018) 66:2 an ita could have in Canada, the kind of mandate that it would be given, and the types of taxes that it would control. I conclude by presenting some of the more obvious objections and attempting a response. Keywords: Tax administration n tax policy n reviews n politics n policy makers

Contents Introduction 388 The Problem 389 The Proposal 393 Objections 397 Question 1: Wouldn’t an ITA Be Unconstitutional? 397 Question 2: What About the Efficiency/Equality Tradeoff? 398 Question 3: Wouldn’t an ITA Be Undemocratic? 399

Introduction Milton Friedman once remarked, famously, that for the government, “to spend is to tax.” Many politicians, unfortunately, strive to escape the force of this equivalency. Indeed, it has by now become a wearisome ritual in us politics that, after voting in favour of a series of spending measures, Congress tries to avoid passing the legislation required to pay for them. Things are not so bad in Canada, although the persistence of budgetary deficits, both federally and provincially, suggests an ongoing desire to wriggle free from the inevitable consequence of government spending decisions. Canadian politicians also have a lengthy track record of playing politics with tax policy. The policy platforms developed by the major parties over the past few decades have at one time or another all featured tax changes aimed at appealing to voters on the basis of false perceptions of their effects. In other words, the fact that the general public has a great deal of difficulty understanding the tax system has been seized upon by political parties across the political spectrum as a way of selling the equivalent of snake oil to the electorate. Compounding the problem, some politicians feel that, having been elected on the basis of such promises, they actually need to follow through on them. The result has been the introduction of both arbitrari- ness and inefficiency into the tax system. One bold proposal that has been made to deal with this state of affairs is the creation of an independent tax commission, which would be charged with developing recommendations for comprehensive tax reform. A bolder proposal—made by Alan Blinder in particular—is the creation of an independent tax authority (ita), which would have the authority both to develop and to implement tax policy.1 In the same way that the Bank of Canada is given substantial autonomy from the government of the day, in order to set and pursue monetary policy, the ita also would

1 Alan S. Blinder, “Is Government Too Political?” (1997) 76:6 Foreign Affairs 115-26. policy forum: an independent tax authority n 389 enjoy substantial autonomy in its decision making. On this model, Parliament would retain control over all spending decisions. The question of how to pay for those decisions, however, would be delegated to the ita, which would have the power to determine all the details of tax policy, such as which general taxes to implement and what the rates should be. The objective would be to take the specific instruments of taxation policy—such as whether to tax income or consumption, or what the corporate tax rate should be—out of the realm of partisan politics. Parlia- ment would provide to the ita only three parameters: the projected spending level, the government’s deficit tolerance, and the target level of progressivity for the tax system as a whole. The ita would then select its tax instruments, subject to those constraints. Such a change would obviously represent a move in the direction of a more “technocratic” form of government and will be resisted on those grounds. Indeed, Blinder’s proposal has received very little uptake or support in the literature. In order to advance the discussion, I will start by showing that an ita represents a response to a genuine problem, which is the exploitation by politicians of public ignorance and irrationality with respect to taxes. Taxation is an area in which the imperatives of “good policy” and those of “good politics” consistently diverge, and there is no sign of this changing any time soon. As a result, inefficiencies and inequalities get introduced into the tax code, and much time and energy are wasted through the constant diversion of public debate into arguments over pseudo-issues. Having motivated the proposal, I will then outline the basic institutional structure that an ita could have in Canada, the kind of mandate that it would be given, and the types of taxes that it would control. Following this, I will present some of the more obvious objections and attempt a response to them.

The Problem If one were to fault the general public for one thing, when it comes to comprehen- sion of tax policy, it would be for its failure to follow the money when assessing the impact of a tax.2 There are, of course, more straightforward failures, including the persistence of false beliefs about the distribution of income, the progressivity of the income tax system, and how much the wealthy pay in tax.3 The more subtle problem, however, involves the average taxpayer’s susceptibility to the “flypaper fallacy” (the idea that the person who writes the cheque to the government is necessarily the person who pays the tax), through failure to grasp the concept of tax

2 Jonathan Baron and Edward J. McCaffery, “Masking Redistribution (or Its Absence),” in Edward J. McCaffery and Joel Slemrod, eds., Behavioral Public Finance (New York: Russell Sage, 2006), 85-112, at 110. 3 Steven M. Sheffrin, “Perceptions of Fairness in the Crucible of Tax Policy,” in Joel Slemrod, ed., Ta x Progressivity and Income Inequality (Cambridge: Cambridge University Press, 1994), 309-34. 390 n canadian tax journal / revue fiscale canadienne (2018) 66:2 incidence. The same inability to follow the money makes the public susceptible to Frédéric Bastiat’s fallacy of “the seen and the unseen.”4 For instance, the economist’s traditional concerns about the effects of price distortions, as well as increased transaction costs causing deadweight losses, deal with issues that occur in the realm of the “unseen,” and are thus susceptible to being discounted or ignored by the general public. More directly, the misperception of tax incidence leads the public to favour hidden taxes over visible ones, even when their explicit policy preferences involve objectives (such as progressivity) that can more easily be achieved through the latter.5 One consequence of this cognitive failure is that it leaves the public vulnerable to certain forms of political demagoguery. The following are examples of this from recent Canadian politics:

n Perhaps the most obvious example of politicians exploiting the public’s poor understanding of tax incidence is the attempt to portray the taxation of corporations as some kind of alternative to the taxation of individuals. Popular support for increased corporate taxes is always extremely high, generating the suspicion that many people believe that a “tax on corporations” will be borne by corporations.6 Steven Sheffrin argues that the average person subscribes to an “entity” view of taxation, believing that the tax burden should be distributed over all entities, whether they be firms or individuals.7 Because of this, the suggestion that one might tax profits in the hands of the firm, or the individual, or both (and that, in order to assess the tax burden, one must look at the combined rate), is a point that eludes the average voter. n One of the peculiarities of Canadian debates on climate change policy has been the politicization of the preference for carbon taxation versus cap and trade, resulting, inter alia, in the development of a patchwork set of policies across the country. Some fraction of this is due to the New Democratic Party’s (ndp’s) repeated insistence that cap and trade is to be preferred because it imposes the tax burden on corporations (or “polluters”), while a carbon tax unfairly imposes the tax burden on consumers (or “hard-working Canadian families”). n One of the oldest chestnuts of tax demagoguery is the suggestion that tax cuts “stimulate the economy.” This is based on widespread public confusion about the difference between “consumer spending” and “aggregate demand” (or, more crudely, the perception of government as an economic black hole into

4 Frederick Bastiat, Essays on Political Economy, trans. David A. Wells (New York: Putnam, 1877), at 70-76. 5 Sheffrin, supra note 3, at 336. See also Edward J. McCaffery and Jonathan Baron, “Thinking About Tax” (2006) 12:1 Psychology, Public Policy, and Law 106-35, at 119. 6 Sheffrin, supra note 3, at 321. 7 Ibid., at 327. policy forum: an independent tax authority n 391

which money disappears).8 Obviously, to the extent that tax cuts are matched by decreased government spending, they leave aggregate demand unchanged, even if they do increase consumer spending. n A permanent feature of tax policy debates in Canada is persistent, low-level agitation to create tax exemptions for “honorific” activities or “necessary” consumption goods. High-profile instances of this include thendp platform promise in 2011 to exclude home heating costs from goods and services tax (gst). The Conservative Party also introduced “boutique” income tax credits, such as a special tax credit for money spent on children’s sports. These kinds of policies amount to the introduction of government subsidies for particular consumption goods, and yet it is doubtful that any of them would be advocated if they were described in such terms. n One of the most anomalous features of the Canadian tax system is that value- added taxes (vats)—the gst and its provincial counterparts—are visible. This has made the gst the least popular tax in Canada (unlike most European countries, in which vats are hidden and thus arouse no significant public resentment). In 2006, the Conservative government earned the ire of almost every economist in the country when it cancelled a set of scheduled income tax cuts in order to cut the gst rate. This decision was based purely on public opinion research, which showed that Canadians had difficulty remembering which political party had reduced their income taxes.9 Cutting the gst was seen as the most effective way of branding the Conservative Party as the one committed to lower taxes.

These examples are all cases in which political parties have found it strategically advantageous to advance policies that would introduce, or have introduced, inefficiencies into the tax system in return for illusory gains. The problem, however, extends much further than this. Apart from being a source of government revenue, the tax system also plays an important role in adjusting the distribution of income, not just through direct transfers, but also by providing public goods and services on terms that are less burdensome to the poor than market pricing. Unfortunately, while the general public has some relatively firm intuitions about what we can refer to as the “distributive justice” aspects of the tax system, these are also compromised by ignorance and irrationality. Unscrupulous politicians have, in turn, exploited these shortcomings in order to persuade voters to support policies that are contrary to their considered preferences. In the United States, for instance, many people believe that the income tax system is so full of loopholes that the wealthy actually

8 For discussion, see Joseph Heath, Filthy Lucre: Economics for People Who Hate Capitalism (Toronto: HarperCollins, 2008), at 82. 9 John Geddes, “Ian Brodie Offers a Candid Case Study in Politics and Policy,” Macleans, March 27, 2009 (www.macleans.ca/politics/ottawa/ian-brodie-offers-a-candid-case-study -in-politics-and-policy/). 392 n canadian tax journal / revue fiscale canadienne (2018) 66:2 pay less tax, in absolute terms, than the middle classes. One study found that a surprising 41 percent of Americans believed that the replacement of the graduated income tax with an exceptionless flat tax would result in the wealthy paying more in taxes; only 35 percent believed that the tax burden on the wealthy would decline under such a proposal.10 One can imagine circumstances in which it would be possible to correct such false beliefs. But many other more subtle errors in judgment are made when it comes to assessing the distributive justice implications of tax policy, which it is difficult to imagine being corrected. These include the following:

n The central problem with the way that people apply their distributive justice intuitions to the tax system has been variously described as an “isolation effect” or “disaggregation bias.”11 The problem, roughly stated, is that while people would like the tax system as a whole to exhibit a certain level of progressivity, they try to achieve this by having each constituent tax exhibit that same progressivity. As a result, they evaluate the “fairness” of each tax in isolation from the others, and thus tend to under- or overadjust them relative to each other, given the target level of progressivity.12 For instance, if the system contains a flat consumption tax, and the rate goes up, people will typically fail to recognize that an increase in the progressivity of the income tax system can preserve the distributive character of the system as a whole.13 n One of the most peculiar consequences of the “entity” view of taxation is that it leads many people to apply their distributive justice intuitions to corporations as well as to individuals. As a result, there is a persistent tendency to think that corporate income tax should also be progressive, with small businesses paying lower rates that large ones. Small businesses, on this view, are seen as analogous to poor people. The result is a corporate tax system that somewhat arbitrarily privileges small businesses by taxing their profits at lower rates. This has contributed to the perverse situation in which wealthy individuals across Canada create corporations—taking advantage of the “small business” tax rates—in order to shelter their own income from taxation.

10 Joel Slemrod, “The Role of Misconceptions in Support for Regressive Tax Reform” (2006) 59:1 National Tax Journal 57-75, at 64. 11 Edward J. McCaffery and Jonathan Baron, “The Humpty Dumpty Blues: Disaggregation Bias in the Evaluation of Tax Systems” (2003) 91:2 Organizational Behavior and Human Decision Processes 230-42; see also McCaffery and Baron, supra note 5. 12 Or as McCaffery and Baron, “The Humpty Dumpty Blues,” supra note 11, at 107, put it, people “generally fail to integrate parallel tax systems or tax and spending systems to form globally consistent judgments about bottom-line allocations and distributions.” 13 See Stephen Gordon, “Economic Policy Advice for the NDP, Part III: The GST,” Worthwhile Canadian Initiative blog, August 18, 2009 (http://worthwhile.typepad.com/worthwhile _canadian_initi/2009/08/economic-policy-advice-for-the-ndp-part-iii-the-gst.html). policy forum: an independent tax authority n 393

Economists often complain that the public puts too much weight on the “fairness” of the tax system, and not enough on its efficiency. Thus, the imposition of expert judgment is recommended as a way of redressing the relative neglect of efficiency considerations. As we have seen, however, the public also has considerable difficulty bringing its distributive justice intuitions to bear upon the tax system. When incidence is taken into account, even the factual question of determining the distributive impacts of the tax system quickly becomes very complicated.14 On top of this, people use a set of very limited heuristics for judging fairness, leaving them open to manipulation. As a result, there is good reason to think that bringing greater technical expertise to bear upon the tax system would permit greater realization of both its efficiencyand its distributive justice goals.

The Proposal Jean-Baptiste Colbert famously compared the art of taxation to that of plucking a goose, with the objective being to secure as many feathers as possible while produc- ing the least amount of hissing. The analogy is inapt in at least one respect. While the goose can presumably feel each one of its feathers being plucked, the complex web of invisible, indirect, and implicit taxes in a modern economy has a mixed effect, such that taxpayers are sometimes unaware of being plucked and at other times have the sensation of being plucked when in fact they are not.15 Meanwhile, elected officials, far from serving as loyal fiduciaries, helping the taxpayer to sort through the confusion, have too often succumbed to the temptation to take advantage of it—in some cases, cynically; in other cases, because they share the same confusion. Compounding the problem is the fact that public irrationality sometimes constrains politicians in return, so that beneficial reforms cannot be undertaken, entirely for reasons of political optics. This is what motivates the proposal for the creation an ita, whose central function would be to remove certain questions from the realm of partisan politics and have them instead be addressed by public officials committed to political neutrality. While the existing revenue authority, the Canada Revenue Agency, enjoys some discretion when it comes to the interpretation of Canadian tax law, its central functions are essentially administrative. An ita would enjoy a more substantial grant of delegated authority, with the power to determine both the base and the rates for all general revenue-generating taxes. It would be governed either by a board and chief executive officer appointed by the prime minister, for a term sufficiently long to establish independence from the government of the day (and in the case of a board, staggered terms). One might also institute something like the “minister’s directive” provision of the Bank of Canada Act, which allows the minister of finance to give

14 See Jonathan R. Kesselman and Ron Cheung, “Tax Incidence, Progressivity, and Inequality in Canada” (2004) 52:3 Canadian Tax Journal 709-89. 15 An example of the latter would be real estate transaction taxes, which are “paid” by the buyer, but in a competitive real estate market actually fall on the seller. 394 n canadian tax journal / revue fiscale canadienne (2018) 66:2 the governor of the Bank of Canada instructions “in specific terms and applicable for a specified period,” which the bank must then comply with.16 Thus, explicit political direction would remain always a possibility; the hope would be that a convention of non-interference would emerge, as it has with the central bank. The general principle governing delegation of power to the civil service is that it should be undertaken only when policy objectives can be clearly specified ex ante and remain relatively stable, and where administration does not involve further, significant normative judgments. Bureaucratic administration typically proves superior to political decision making when the attainment of these objectives benefits from the deployment of complex technical knowledge over a long time horizon, where there is a need for consistency over time, and where there is some desire to insulate decision making from the influence of interest groups.17 Taxation satisfies these constraints so long as the term “tax” is interpreted fairly narrowly. After all, the government imposes a dizzying array of taxes, fees, charges, fines, takings, royalties, tariffs, and surcharges, many of which are closely tied to very specific policy objectives. An ita, by contrast, would be concerned only with general revenue. We can use the term “general taxes” for payments meeting the following criteria:

(1) an exaction of money; (2) imposed and collected by a government authority; (3) on or from a private person or entity; where (4) such payment is compulsory; (5) imposed for non-punitive purposes; and (6) paid not in return for a specific service or privilege received from the government, but rather as a means to fund government operations more generally.18

In the Canadian context, this means that a federal ita would have authority over income, consumption, corporate, and inheritance taxes. It would not control customs, tariff, and excise taxes, nor would it control taxes that can be construed as Pigovian, such the usual “sin taxes,” or the revenue from a carbon-pricing regime. The chief difference is that, in the case of general taxes, the objective is to raise revenue while minimizing the distortionary impact on economic behaviour. In the latter set of cases, the objective of the tax is precisely to change economic behaviour in order to meet specific policy objectives (and only secondarily, or perhaps not at all, to raise revenue). Thus, there is good reason to want the latter sort to be under direct political control. Similarly, dedicated social insurance contributions, such as Canada

16 Bank of Canada Act, RSC 1985, c. B-2, as amended, section 14(2). 17 Alberto Alesina and Guido Tabellini, “Bureaucrats or Politicians? Part I: A Single Policy Task” (2007) 97:1 American Economic Review 169-79; Alberto Alesina and Guido Tabellini, “Bureaucrats or Politicians? Part II: Multiple Policy Tasks” (2008) 92:3-4 Journal of Public Economics 425-47; and Lars Calmfors, What Remains of the Stability Pact and What Next? (Stockholm: Swedish Institute for European Policy Studies, 2005), at 93-94. 18 Stuart P. Green, “Tax Evasion as Crime,” in Monica Bhandari, ed., Philosophical Foundations of Tax Law (Oxford: Oxford University Press, 2017), 57-78, at 60. Green limits his use of the term “tax” to payments that meet all of these criteria. For convenience, I use the term “tax” more broadly, and the term “general tax” for what Green calls a “tax.” policy forum: an independent tax authority n 395

Pension Plan or employment insurance, should remain outside the control of the ita, because they are tied to the specific policy objectives of those programs.19 With general taxation, despite the complexity and the proliferation of different instruments, there is a relatively simple overarching objective, which is to raise revenue (hence the goose-plucking metaphor). The desire to do so in a way that minimizes distortion, or deadweight losses, is what calls for the deployment of technical expertise. The details of how this should be done are extremely complicated but essentially administrative—they do not involve any other significant normative judgments per se. There is also a great deal to be said for treating all general taxation instruments as part of an integrated scheme, since they each merely contribute to the attainment of the objective. And finally, there are significant benefits to having a tax regime that is both stable over time and designed to enhance long-term prosperity.20 All of these considerations suggest that the creation of a single administrative agency with control over general taxation could have significant benefits. Perhaps the most important complication in the plan arises from the fact that the tax system cannot be organized around an exclusive preoccupation with efficiency. Taxation also raises questions of distributive justice. In the first instance, there is the question of how the fiscal burden for the state’s activities, such as the provision of health care, education, infrastructure, defence, etc. should be borne by the population. Taxes here are implicitly redistributive, in that all citizens get access to roughly the same bundle of public goods, but the contribution that they make to the funding of these goods is progressive with respect to income. Beyond this, however, there is also the possibility of using the tax system in a way that is explicitly redistributive, to effect transfers between individuals.21 Typically, however, explicit redistribution is achieved through benefit programs that are financed through taxation. Thus, anita could concern itself entirely with the implicitly redistributive character of taxes, which is to say, how the overall fiscal burden of financing the state should be distributed over individuals in the society. Here, it is not too difficult for elected officials simply to specify how progressive they would like that distribution to be. An ita

19 It is worth noting that in 2008 the government of Canada created an independent authority responsible for determining employment insurance premiums, the Canada Employment Insurance Financing Board, to balance the budget of the program over the course of the business cycle. Unfortunately, the board never achieved genuine independence, and it was abolished in 2013. As a matter of practical politics, this precedent bodes ill for the creation of an ITA. See Arthur Sweetman, “Take the Politics out of Employment Insurance,” National Post, January 19, 2011. 20 Blinder, supra note 1, puts particular emphasis on this point. 21 There is a tendency to overstate how much explicit redistribution is undertaken by the welfare state, because of confusion over the appropriate classification of social insurance programs. All insurance, whether public or private, is “redistributive” in a sense, but this is different from the sort of egalitarian redistribution undertaken by the state. Generally speaking, it is best to think of social insurance programs as public goods provided by the state, which are implicitly redistributive because individuals are not charged actuarially fair premiums. 396 n canadian tax journal / revue fiscale canadienne (2018) 66:2 would therefore have a dual mandate: to raise a specified amount of revenue as efficiently as possible, and to do so in a way that imposes the burden in accordance with a specified principle of distributive justice. What would such a principle of distributive justice look like? Consider the economist’s classic argument for progressive taxation, which takes as its central normative claim the principle that all citizens should make approximately equal sacrifice, expressed in terms of utility.22 Because money (or consumption) produces declining marginal gains in utility, it follows that the rate of taxation should increase as a function of individual income (or consumption). If this is the ideal, then it is possible to specify a single number (the elasticity of marginal utility of consumption) that will determine how much the tax burden should increase as one moves up through the income brackets, in order to maintain equal sacrifice. In principle, this can be determined empirically. It can also be specified normatively (or politically), in which case the same number can be used as a measure of inequality-aversion, to produce a prioritarian social welfare function, which will in turn recommend a more steeply progressive set of tax rates. Thus, the distributive justice principle that governs the tax system could be specified, at an abstract level, by something as simple as a single number—one that specifies the rate at which, as income rises, the sacrifice associated with the forfeiture of a given sum declines. While these end-state distributive justice objectives may be easy to specify, achieving them is extremely complex, and involves a great deal of expert judgment and technical knowledge. This is primarily because each component of the tax system makes its own contribution to the distributive impact of the system as a whole. The level at which we should care about the outcomes, from the standpoint of distributive justice, is the aggregate level, and in terms of net impact on individuals. These effects are, however, too complex for the average person to assess, and this is why people opt for the imperfect heuristics that generate the disaggregation fallacy.23 An ita would force elected officials to state clearly what their overarching conception of “tax fairness” amounts to, and then delegate to public officials the task of designing a system that actually achieves fairness according to that specification. In principle, this might also improve the quality of public debate, to the extent that it encouraged politicians to focus less on the framing of particular taxes, and instead to defend their most basic convictions about what the distribution of wealth in society should be. Finally, it should be noted that an ita could be instituted in weaker and stronger forms, depending on how much discretion is granted. It could also be structured in

22 As John Stuart Mill put it, “[a]s a government ought to make no distinction of persons or classes in the strength of their claims on it, whatever sacrifices it requires from them should be made to bear as nearly as possible with the same pressure upon all; which, it must be observed, is the mode by which least sacrifice is occasioned on the whole.”Principles of Political Economy (New York: Prometheus, 2004), at 735. 23 McCaffery and Baron, “The Humpty Dumpty Blues,” supra note 11. policy forum: an independent tax authority n 397 such a way that the level of discretion exercised could evolve over time. (By way of comparison, the growing autonomy of the Bank of Canada occurred, not so much through changes in legislation, but rather through the governor of the bank assert- ing increased authority over monetary policy, and the government refraining from retaking control.)24 A weaker version would keep existing parliamentary control over taxes in place, but give the ita some discretion to modify rates, or to make small adjustments to the base.25 The standard version would give the ita the power to determine rates (including zeroing them out, effectively abolishing a particular tax) and base. Tax expenditures could still be introduced by the government, but they would have to be implemented explicitly as spending programs. A stronger version still would give the ita the power to retain any surpluses it collects, so that it would transfer to the government only the amount requested at the beginning of the budgetary period. This would have the salutary effect of preventing the government from treating unexpected surpluses as windfall gains and engaging in reckless spending. Surpluses would of course not literally be retained by the ita, but they could be kept nominally on the books and used in future years, either to cover revenue shortfalls or to finance lowering of rates.

Objections Proposals for an ita are generally regarded as a political non-starter, on the grounds that elected officials are unlikely to want any curtailment of their authority in these matters. This is, of course, incontrovertible, although it is not so difficult to imagine various crisis scenarios that might change things. Furthermore, if the model were to be implemented and found successful in one jurisdiction, this could lead to its adoption in others. In the discussion that follows, therefore, I will abstract from everyday political obstacles and focus on more principled objections to the entire scheme.

Question 1: Wouldn’t an ITA Be Unconstitutional? Owing to certain vagaries of English history, the ability to raise taxes is one of the most jealously guarded powers of Parliament. As a constitutional convention it dates back to Magna Carta, and over time it served as the central mechanism through which the powers of the monarchy were curtailed. Creation of an ita could easily be construed as a transfer of taxation power from Parliament to the executive, and thus as a reversal of centuries-old constitutional convention. More narrowly, the obvious constitutional challenge would be that it violates the non-delegation doctrine, and hence the separation of powers.

24 Robert Bothwell, Ian Drummond, and John English, Canada Since 1945: Power, Politics, and Provincialism, rev. ed. (Toronto: University of Toronto Press, 1989), at 170. 25 This is the model that was put forward in Business Council of Australia, New Directions Task Force, Avoiding Boom/Bust: Macro-Economic Reform for a Globalised Economy (Melbourne: Business Council of Australia, 1999), at 43. 398 n canadian tax journal / revue fiscale canadienne (2018) 66:2

While in essence this is correct, the rise of the administrative state in the 20th century generated considerable ambiguity with respect to determining what constitutes “delegation.” Statutes are increasingly formulated at a high level of generality, with the task of writing specific rules being handed over to administrative agencies or departments. Thus, as a technical matter, it would not be that difficult to write a set of enabling statutes for an ita that would follow existing precedent in the delegation of rule-making authority. And if this proved controversial, there would still be the possibility of requiring major ita decisions to be ratified by Parliament, perhaps with those decisions being submitted in an omnibus package that would not be subject to amendment. It is worth keeping in mind that the constitutional conventions that would be tested by an ita are all unwritten, and as a result are not entirely inflexible. Furthermore, new conventions can always be created. The one hard constitutional constraint that does exist involves the division of powers in the realm of taxation between the federal government and the provinces (and, through delegation from the provinces, municipalities). This does not prevent the creation of an ita at the federal level, but it does complicate the mandate, since the federal agency would have to work around provincial policies. This would make it difficult to attain certain distributive justice objectives. It would also leave certain taxes in the control of elected officials—assuming that the provinces did not create their own itas. Ideally, one would want resource royalties and property taxes, as well as provincial income, consumption, and corporate taxes, to be under a unified taxation authority. Probably the most that one could hope for, in a Canadian context, would be coordination.

Question 2: What About the Efficiency/Equality Tradeoff? The ideal scenario for delegation of powers to an administrative agency is when a single objective can be specified (which in turn allows the attainment of that objective to be reduced to a technical exercise). The existence of multiple objectives, along with some measure of rivalry between them, results in a dramatic expansion of administrative discretion, along with the spectre of unelected officials making significant normative judgments when deciding how these objectives should be traded off against one another. The ita is described as having essentially a dual mandate— both to promote an efficient tax regime and to achieve certain distributive justice targets. These are far from complementary, and it is not difficult to imagine circumstances in which distributive justice objectives could be achieved only at the expense of efficiency and vice versa. Nevertheless, it is not unprecedented for an administrative agency to be given a set of complex and partially rivalrous objectives. Indeed, the phrase “dual mandate” is typically associated with the United States Federal Reserve, which has the objective of promoting both stable prices and maximum employment. The Bank of Canada has an even more complex mandate,

to regulate credit and currency in the best interests of the economic life of the nation, to control and protect the external value of the national monetary unit and to mitigate by policy forum: an independent tax authority n 399

its influence fluctuations in the general level of production, trade, prices and employment, so far as may be possible within the scope of monetary action, and generally to promote the economic and financial welfare of Canada.26

The question, therefore, is not whether a particular task can be reduced to a purely technical exercise, stripped of significant normative judgments. Rather, the question is who is likely to do a better job at balancing rival objectives—elected officials or public servants? As long as politicians see the tax code primarily as a source of wedge issues, and focus their efforts on the framing rather than the economic consequences, or on particular tax policies, it is not difficult to imagine public servants doing a better job at achieving a satisfactory resolution of the tensions that exist within this domain.

Question 3: Wouldn’t an ITA Be Undemocratic? A lot of things are undemocratic; the question is whether an ita would be objection- ably undemocratic. In Canada, we do not elect judges or Crown prosecutors; in the United States, they do. Neither arrangement is derived from first principles of democratic theory. Popular election of judges is rejected, throughout almost the entire world, because it fails to produce a sufficiently independent and impartial judiciary.27 Similarly, the question of whether the specific details of tax policy should be set by elected officials, or at what level of granularity these details should be specified, cannot be answered by consulting first principles. There is inevitably a consequentialist dimension to these questions—one has to look at the likely effects of various arrangements. In the case of taxation, however, it is helpful also to recall Freidman’s dictum, that “to spend is to tax.” There is a sense in which, under the present arrangement, Parliament gets to vote twice on the same question. It decides to spend, and then it decides how to pay for this spending. An ita would obviously have no role to play in any of the spending decisions made by Parliament, and with good reason, since making these decisions involves little other than making the complex normative judgments involved in balancing the competing demands of different programs. An ita would merely deal with the downstream consequences of these decisions. Treat- ing the choice of tax policy as a separate decision, as we currently do, encourages incoherence in government decision making. If this is not sufficient comfort, it is worth recalling that no grant of authority is irrevocable, and the autonomy of an ita need not be established all at once. The structure suggested above is one in which the agency’s independence would have clear limits, and autonomy would be allowed to evolve through convention. This would create a phase-in period in which adjustments could be made.

26 Preamble to the Bank of Canada Act, supra note 16. 27 Adam Liptak, “U.S. Voting for Judges Perplexes Other Nations,” New York Times, May 25, 2008. As one French judge put it, the option of electing judges was debated following the French Revolution: “It was thought not to be a good idea.” Cited by Liptak, ibid., at 21. canadian tax journal / revue fiscale canadienne (2018) 66:2, 421 - 45

Personal Tax Planning Co-Editors: Gabriel Baron and Maureen De Lisser*

Revisiting Planning for Private Company Shareholders After July 2017 Robert Santia**

Proposed amendments to the Income Tax Act initially announced on July 18, 2017 represent a paradigm shift in the taxation of private corporations and their Canadian- resident shareholders. These proposals give rise to tax-planning and estate-planning concerns for Canadian owner-managers or private company shareholders. In this article, the author outlines how existing planning strategies implemented by Canadian owner- managers and private company shareholders may be affected and suggests steps that such individuals can take to mitigate the consequences of the proposed changes. Keywords: Income splitting n estate planning n private corporations n owner-manager n tax planning n shareholders

* Of Ernst & Young LLP, Toronto. ** Of Aird & Berlis LLP, Toronto. I would like to thank Francesco Gucciardo of Aird & Berlis LLP for his invaluable guidance and his contributions and revisions to this article. I would also like to thank Stuart Bollefer of Aird & Berlis LLP for his comments on earlier drafts of the article. Finally, I would like to thank the editors of this journal for their helpful suggestions. Any errors or omissions remain my own. 421 canadian tax journal / revue fiscale canadienne (2018) 66:2, 447 - 74

Planification fiscale personnelle Co-rédacteurs de chronique : Gabriel Baron et Maureen De Lisser*

Réexamen de la planification pour les actionnaires de sociétés privées après juillet 2017 Robert Santia**

Les modifications proposées à la Loi de l’impôt sur le revenu, qui ont été annoncées initialement le 18 juillet 2017, représentent un changement de paradigme dans l’imposition des sociétés privées et de leurs actionnaires qui résident au Canada. Ces propositions soulèvent des inquiétudes en matière de planification fiscale et successorale pour les propriétaires exploitants canadiens et les actionnaires de sociétés privées. Dans cet article, l’auteur passe brièvement en revue l’effet que ces propositions pourront avoir sur les stratégies de planification existantes mises en oeuvre par les propriétaires exploitants canadiens et les actionnaires de sociétés privées, et propose des mesures pouvant être prises par ces particuliers pour réduire les effets des changements proposés. Mots clés : fractionnement du revenu n planification successorale n sociétés privées n propriétaire exploitant n planification fiscale n actionnaires

* De Ernst & Young LLP, Toronto. ** D’Aird & Berlis LLP, Toronto. Je remercie Francesco Gucciardo d’Aird & Berlis LLP pour ses précieux conseils, et sa contribution et ses corrections à cet article. Je remercie également Stuart Bollefer d’Aird & Berlis LLP pour ses commentaires sur les premières versions de l’article. Enfin, je remercie les rédacteurs de cette publication pour leurs suggestions utiles. Toutes erreurs ou omissions sont les miennes.

447 canadian tax journal / revue fiscale canadienne (2018) 66:2, 475 - 90

Selected US Tax Developments Co-Editors: Peter A. Glicklich* and Michael J. Miller**

A Close Look at Some of the New US Tax Rules Michael J. Miller

Amendments to the US Internal Revenue Code enacted in December 2017 include a new “forced repatriation” provision and expansion of the anti-deferral regime under subpart F. This article discusses the potentially harsh effects of these new provisions, particularly for US shareholders who are individuals. Keywords: US n shareholders n controlled foreign corporations n repatriation n distributions

* Of Davies Ward Phillips & Vineberg LLP, New York. ** Of Roberts & Holland LLP, New York and Washington, DC.

475 canadian tax journal / revue fiscale canadienne (2018) 66:2, 491 - 509

Current Tax Reading Co-Editors: Robin Boadway, Kim Brooks, Jinyan Li, and Alan Macnaughton*

Allison Christians, “Trust in the Tax System: The Problem of Lobbying,” in Bruno Peeters, Hans Gribnau, and Jo Badisco, eds., Building Trust in Taxation (Cambridge, UK: Intersentia, 2017), 151-72 (papers.ssrn.com/ sol3/papers.cfm?abstract_id=3097535) Christians confronts the claims of certain taxpayers that they are mere players in a tax game. She invokes what she calls “the full compliance defence”1: the argument that taxpayers cannot be faulted for paying little or no tax because they have com- plied with all relevant tax and accounting rules. She argues that this response masks the role that many of these taxpayers play in shaping those very rules. We have no convincing data about the magnitude of tax lobbying or about the consequences of this lobbying in countries around the world. Christians marshals the data that are available about the effectiveness of lobbying, which suggest that there is substantial return for taxpayers who spend money lobbying governments for tax reforms that benefit them. She notes that it is hard to get reliable data about how lobbying takes place because of the access that high-income or economically powerful people and corporate agents have to government officials. She offers as an illustration a report by the Financial Times that Tim Cook, Apple’s chief executive officer, personally visited the European commissioner to lobby for better tax treatment for Apple. Few taxpayers have such unregulated access to people who make tax decisions. Ultimately, she suggests a two-pronged policy response. First, she argues that lobbying should be made more transparent. There are a variety of mechanisms to support transparency, including open-source platforms and the requirement that government regulators disclose the names of lobbyists and the content of all meet- ings with them. Whatever mechanisms are used, Christians argues that transparency should be global in scope and include the ability to cross-reference data across countries.

* Robin Boadway is of the Department of Economics, Queen’s University, Kingston, Ontario. Kim Brooks is of the Schulich School of Law, Dalhousie University, Halifax. Jinyan Li is of Osgoode Hall Law School, York University, Toronto. Alan Macnaughton is of the School of Accounting and Finance, University of Waterloo. 1 At 153.

491 492 n canadian tax journal / revue fiscale canadienne (2018) 66:2

Christians acknowledges that transparency alone is unlikely to solve the challenges produced by the influence of high-income taxpayers. Therefore, she also argues that public engagement in tax policy-making processes must be diversified. She identifies the one-sidedness of political influence as particularly troubling. Ultimately, Christians’s claim is that this work matters, not only because it will ensure tax democracy but also because trust in our tax systems is a vital element of self-reporting and compliance systems. K.B.

Robert French and Philip Oreopoulos, “Applying Behavioural Economics to Public Policy in Canada” (2017) 50:3 Canadian Journal of Economics 599-635 Behavioural economists study how findings in psychology and neuroscience call into question the rational choice underpinnings of economic models. For example, small changes in the way that options are framed can have substantial effects on behaviour. This idea has implications for the delivery of public policies. French and Oreopoulos argue that interventions in policy design that cost little can lead to substantial differences in behaviour that improve outcomes for citizens. Govern- ments have recognized this fact by setting up behavioural intervention or “nudge” units, following the term coined by Thaler and Sunstein.2 Following the United Kingdom’s creation of the Behavioural Insights Team, more than 50 countries have set up agencies to apply behavioural economics findings to public policy. In this article, the authors examine such agencies in Canada, including the Innovation Hub at the Privy Council Office, Innovation Labs at the Canada Revenue Agency (cra) and Employment and Social Development Canada (esdc), and the Behavioural Insights Unit in Ontario. French and Oreopoulos begin by outlining three models used in behavioural economics as alternatives to the standard model of individual behaviour. The first of these models involves present bias, which leads to individuals putting excessive weight on outcomes that will occur in the near future. Neuroscience research has increasingly substantiated the relevance of present bias. The second model involves inattention. Too little attention is devoted to making decisions when the cost of deliberating is high and some outcomes are not salient. Some complex decisions or decisions made under stress may be made as a result of habit or heuristics that compromise long-term outcomes. Third, social identity provides a reference point that affects how we act. In each case, individuals may make choices that are against their long-term interests, which can lead to low-cost possibilities for the government to improve policy outcomes. French and Oreopoulos characterize two approaches to government intervention. One approach is to limit individuals’ choices, for example, by

2 Richard H. Thaler and Cass R. Sunstein, Nudge: Improving Decisions About Health, Wealth, and Happiness (New Haven, CT: Yale University Press, 2008). current tax reading n 493 banning certain actions (for example, drug use) or by mandating others (for example, seatbelt use). A second approach is to intervene in a choice-preserving way by encouraging beneficial actions. This approach can involve making people aware of benefits or making one option appear to be more appealing than another. Choice- preserving interventions are called “nudges.” The authors present several examples of these interventions, including examples that arise in the Canadian context. One example involves interventions to encourage saving for retirement, which may be suboptimal because of present bias. One possibility, applied in the United Kingdom, is to change the default option in employer pension plans to require employees to participate unless they choose to opt out. Another option, proposed for Canada, is to require all employees to make a choice between whether or not to enrol in a pension, thereby making the choice more significant. Another example involves encouraging greater participation in registered educational savings plans (resps). A pilot project based on a program devised by the Omega Foundation involves simplifying the application procedure for resps and Canada learning bonds to increase their use. Other examples involve health care. On the basis of evidence that more people are willing to become organ donors than have signed up to do so, some countries have legislated presumed consent, requiring individuals to opt out if they choose not to be donors. Ontario has instead used a prompted-choice system to become organ donors when people are at a Service Ontario centre. Also in Ontario, pharmacists are allowed to administer flu vaccines for the sake of customers’ convenience. Behavioural interventions have also been used in education. An example in the United States involves personal assistance in applying for student aid. In Ontario, the Ministry of Education and the Ministry of Training, Colleges and Universities have supported a program called “Life After High School,” which provides in-class assistance in applying for post-secondary education programs and financial assistance. Another intervention involved a field experiment using personal coaching to improve the classroom achievement of first-year university students. Yet another example involves easing the transition from unemployment to employment. esdc’s Job Bank matches job seekers with employers; however, the job seekers must apply, providing information about their skills, education, and experience. esdc Innovation Lab is conducting a pilot project to encourage enrol- ment using nudge statements, which are supplemented by motivational interviewing, to encourage application. Finally, tax policy is a fruitful area for intervention to affect behavioural responses by improving the salience of various tax rates, facilitating compliance, and framing tax policy. For example, various interventions that involve tax compliance include altering the wording on late-payment notices, sending personalized text message reminders, and using social norms or comparisons in tax collection letters. A cra experiment in progress involves changing the signature block on tax returns by moving it to the beginning of the form (to prompt honest reporting before filling in the form), enlarging the text, and using boldface to warn against providing inaccurate information. 494 n canadian tax journal / revue fiscale canadienne (2018) 66:2

French and Oreopoulos argue that the interventions they discuss are relatively low in cost but can result in significant improvements in behavioural outcomes. They stress, however, that behavioural economic models should complement traditional economic models in informing public policy, not replace them. R.B.

John Lester, “Policy Interventions Favouring Small Business: Rationales, Results and Recommendations” (2017) 10:11 SPP Research Papers [University of Calgary School of Public Policy] 1-51 This is a thorough and thoughtful overview and evaluation of government assistance to small and medium enterprises (smes), including tax measures, research and development (r & d) incentives, and financing. The presumption is that preferential treatment of smes should be predicated on market failure. Drawing on the existing literature, Lester begins with a detailed summary of market failures that might confront smes and possible responses by the government. Some market failures involve externalities arising from r & d, innovation, learning by doing, and agglom- eration. smes might also have to deal with adverse selection and search difficulties in the labour markets. They may face non-financial barriers, such as deterrents to entry, tax policy that treats gains and losses asymmetrically, and profit-insensitive taxes. In addition, they may face financial market imperfections, such as moral hazard and adverse selection. Lester argues that most, but not all, market failures discriminate against innovative entrepreneurs, and that there is a case to be made for the public support of innovative entrepreneurs, although not necessarily all smes. Lester then summarizes the many ways in which government policy supports smes. The main categories are programs of financing and loan guarantees, support for r & d and innovation, direct support for entrepreneurship delivered through the tax system, and other targeted programs. In total, he lists over 20 programs designed to assist smes. Roughly 80 percent of support is directed toward small business and 20 percent is directed toward entrepreneurship. Lester evaluates most of the programs by combining qualitative evaluation with rudimentary cost-benefit analyses when feasible. On the basis of these evaluations, he recommends several far-reaching reforms. Although it is impossible to do justice to these evaluations, it is useful to discuss some of the more significant programs briefly. There are three main financing programs. One is the small business deduction (sbd), which is available to all smes subject to income and asset limits. Its intent is to assist smes in financing their investments through retained earnings that are taxed at preferential rates. Since profits are fully taxed when they are taken out of a business, the sbd acts like an interest-free loan. Lester argues that while the sbd provides some financing assistance, it results in large firms being replaced by smaller, less productive ones. He estimates that the costs far outweigh the benefits, and recommends that the sbd be abolished. current tax reading n 495

A second program is the small business financing program sbfp( ), which guarantees private sector loans. Lester assumes that 56 percent of the loan guarantees go to firms that are not offered private sector loans because of credit rationing. In principle, these loans could generate a net benefit since they mitigate against a market failure. He estimates the net benefit and compares it with the program costs, which include both financing and administrative costs. The costs slightly outweigh the benefits, largely as a result of the assumption that 44 percent of the loans that are given guarantees would have been approved by lenders without a guarantee. If the number had been 33 percent, the net benefits of the program would have just offset the costs. Lester recommends that the program be reformed by making more effort to ensure that the program applies to loans that would have been rejected by the market. A third main financing policy is the Business Development Bank of Canada (bdc) financing program, which provides both loans and advice to smes. Lester again generously assumes that bdc loans correct a market failure that arises as a result of credit rationing and therefore generate some social benefit. However, the social costs of financing and administering the loans significantly outweigh the benefits. He recommends changing from a loan program to a loan guarantee program, possibly combined with the sbfp. Support for r & d by smes takes two main forms: the enhanced scientific research and experimental development (sr & ed) credit and the industrial research assistance program (irap). The enhanced sr & ed credit for smes is a tax credit based on r & d expenditures that is more generous (35 percent) than that offered to large corporations (15 percent), and it is refundable. Lester’s cost-benefit calculations reveal that while the standard credit generates a net social benefit of 12 cents per dollar of tax revenue forgone, the enhanced credit results in a net loss of 12 cents on the dollar. He recommends that the federal sr & ed credit be 15 percent for all firms, and that refundability be maintained forsme s. irap provides direct financial assistance as well as technical and business advice to small firms. Financial assistance takes the form of contribution funding, which has relatively high administration and compliance costs (about 33 percent of funding). Its subsidy of about 20 percent of costs exists in addition to any sr & ed credit. This high subsidy rate combined with high administrative and compliance costs leads to a negative net benefit of about 12 percent of program spending. Lester recommends reducing the maximum irap subsidy rate to 10 percent, and possibly transforming irap into an advice-only program. Finally, three sources of support for entrepreneurship delivered through the tax system are considered. The lifetime capital gains exemption mitigates the double taxation of capital. Lester suggests that the exemption would be most cost-effective if it were restricted to innovative entrepreneurs; since this is not feasible, he recommends leaving the exemption as it is. The allowance for business investment losses allows smes to deduct business losses against ordinary income to mitigate the asym- metric treatment of gains and losses. Lester recommends extending this allowance to unincorporated businesses. Similarly, he recommends that all small businesses be 496 n canadian tax journal / revue fiscale canadienne (2018) 66:2 allowed to roll over capital gains when proceeds from the sale of small business shares are reinvested in other small businesses. Finally, he recommends full taxation of stock options combined with the deductibility of the cost of stock options by corporations, although a case can be made for retaining the existing system in the case of young corporations. Taken together, these reforms would represent a wholesale reform of taxation and financial support for smes. Recent experience shows that such reform would be controversial, but the arguments Lester presents for his recommendations are convincing. R.B.

Kenneth J. McKenzie and Ergete Ferede, “Who Pays the Corporate Tax? Insights from the Literature and Evidence for Canadian Provinces” (2017) 10:6 SPP Research Papers [University of Calgary School of Public Policy] 1-21 The incidence of the corporate tax is highly important from a tax policy perspective. To the extent that the tax is ultimately borne by wage earners, its fairness is questionable, as is the case for integrating the personal and corporate tax systems. In this paper, the authors provide a very useful and readable overview of the corporate tax incidence in highly open economies such as Canada’s, review both the theoretical and empirical literatures, and report on their own empirical work using provincial and federal corporate tax data. McKenzie and Ferede begin with a primer on the theory of the corporate tax incidence. As they point out, the message has changed dramatically since Har- berger’s landmark paper of 1962.3 Harberger had argued that the corporate tax was likely borne by all capital owners in both the corporate and non-corporate sectors. The argument was heavily dependent on a closed economy model with predeter- mined quantities of capital and labour. Harberger’s analysis formed the basis of influential computable general equilibrium simulations summarized by Shoven and Whalley.4 Instead, if one assumes that the economy is a small open one with capital perfectly mobile internationally, the after-tax return to capital is essentially fixed, and therefore capital owners cannot bear the incidence of the corporate tax. It must be shifted elsewhere, and since output prices are determined internationally, the incidence of corporate tax is necessarily shifted to workers. This non-nuanced view of the corporate tax incidence changes when certain assumptions are relaxed. Capital that is not perfectly mobile can bear some of the tax. Capital can also bear some of the tax if a country is large enough to influence international rates of return. The

3 Arnold C. Harberger, “The Incidence of the Corporation Income Tax” (1962) 70:3 Journal of Political Economy 215-40. 4 John B. Shoven and John Whalley, “Applied General-Equilibrium Models of Taxation and International Trade: An Introduction and Survey” (1988) 22:3 Journal of Economic Literature 1007-53. current tax reading n 497 authors refer to these effects on wages that operate through changes in the capital- labour ratio as “indirect effects.” Perhaps most intriguing from a tax policy point of view, suppose that some of the returns to corporate capital represent rents or monopoly profits. In this case, one might argue that the tax on rents does not affect behaviour and therefore is borne by the owners of the corporation. However, suppose corporate rents are shared between the owners and workers who have bargaining power, for example, through a union. In this case, the incidence of a tax on rents is shared between the capital owners and labour. This is referred to as “the direct effect” of the corporate tax on wages. Taking the indirect and direct effects together, theory alone cannot resolve the question of the corporate tax incidence, and therefore one must turn to empirical evidence. McKenzie and Ferede summarize the empirical literature on corporate tax incidence. Although estimates vary from study to study, most show that labour bears at least half the burden of corporate taxation, and maybe more depending on country characteristics, such as the size of the economy, the capital intensity, the mobility of capital and labour, and the treatment of profits earned by a firm in multiple jurisdictions. The authors then report the results of their own study on the incidence of federal and provincial corporate taxes in Canada, using data for the period from 1981 to 2014. They estimate how the capital-labour ratio responds to corporate tax changes in each province, and then how the changes in the capital-labour ratio affect wages. Taken together, these estimates allow them to further estimate the effect of corporate tax changes on wages. They find that the elasticity of the capital-labour ratio with respect to the corporate rate is −23 percent and the elasticity of the wage rate with respect to the capital-labour ratio is 46 percent, leading to an overall elasticity of the wage rate with respect to the corporate tax rate of about −10 percent. This implies that an additional dollar of corporate tax revenue causes wages to fall in Alberta by about $1.50, which is sizeable, or about $2.00, in Ontario. Thus, the corporate tax has a significant negative effect on wages in Canada as a result of the indirect effect alone (no evidence is presented about the direct effect of the corporate tax on wages through wage bargaining). These results reinforce the accumulating evidence of a substantial shifting of the incidence of the corporate tax to workers. It lends further support to the argument that the withholding role of the corporate tax should be revisited. R.B.

Graeme Cooper, The Unconvincing Case for 25%, Sydney Law School Legal Studies Research Paper no. 17/90 (Sydney: University of Sydney, 2017) (papers.ssrn.com/sol3/papers.cfm?abstract_id=3059257), 28 pages

In the wake of the us changes to corporate taxation, work on the appropriate design for and rate of corporation tax seems to be urgent. Cooper examines the case for reducing Australia’s corporation tax rate to 25 percent. To be clear, he does not review the range of arguments that could be made for corporate tax rate reductions; 498 n canadian tax journal / revue fiscale canadienne (2018) 66:2 instead, he explores the ones that are being made in Australia and he finds these arguments to be wanting. Australia’s corporate tax rate was reduced from 49 percent in 1986 to 30 percent in 2001. In the 2016 elections, the government proposed a further reduction to 25 percent (to occur by 2026). This proposal became a subject of the election debate. Cooper offers a potted history of the four tax reform projects (beginning in 1998) that have advanced the view that the corporate tax rate is too high. He highlights the justifications for a rate reduction in each project. In the 1998Review of Business Taxation and the 2007-9 Henry review,5 these justifications included attracting international capital and sending a competitive signal. The 2011-12 Business Tax Working Group suggested that reductions might be supported because they would enhance profitability for capital owners, increase the product of labour (with increased wages to follow), and encourage foreign investment. Finally, in the 2015 ReThink project, concern focused primarily on the competitive disadvantage of having a high rate, relative to that in other countries. None of the reports was unequivocal in its support for rate reductions. Following these reports, in 2015 and 2016 the Treasury of Australia attempted to model the consequences of a corporate rate reduction. In each case, it suggested increasing other taxes to make up lost revenue. Cooper offers four major observations about the case for a rate reduction as presented in these various reform projects and Treasury papers. First, he underscores that in each case the proposal to reduce the corporate tax rate is coupled with a proposal to increase another tax. In this sense, the reform proposals cannot be described as “tax cuts”; instead, they should be understood as tax changes. Second, he reminds readers that there is a gap between the statutory and the effective corporate tax rates, and he urges a more nuanced approach to the debate on rate. Third, he notes that the imputation system works differently for different companies and different shareholders. As a result, arguments about the extent to which the corporate tax can serve as a backstop to prevent deferral or avoidance are often oversimplified. Finally, he queries whether the strength of commitment to reducing the rate to appeal to foreign investors is accurate. In a world where comparisons of corporate tax rates have perhaps taken on increased urgency, work such as Cooper’s, which demands policy responses that are sensitive to the nuances of corporate tax design, are most welcome. K.B.

5 Australia, Review of Business Taxation, A Tax System Redesigned—More Certain, Equitable and Durable (Canberra: Review of Business Taxation, July 1999); and Australia, Australia’s Future Tax System: Report to the Treasurer (Canberra: Commonwealth of Australia, May 2010) (herein referred to as “the Henry review”). current tax reading n 499

Michael Littlewood and Craig Elliffe, eds., Capital Gains Taxation: A Comparative Analysis of Key Issues (Cheltenham, UK: Edward Elgar, 2017), 448 pages David Duff and Ben Alarie have written in the past about the legacy of the United Kingdom’s tax system on Canadian tax law.6 The weight of that history is perhaps felt most heavily in the taxation of capital gains. Canada is not alone in carrying on the United Kingdom’s residue (given the importance of the trust concept, pun intended). The United Kingdom’s approach to capital gains taxation echoes the approaches of Australia and South Africa. The United States, New Zealand, the Netherlands, India, and China, not in- spired by the United Kingdom, have adopted different approaches to the taxation of capital gains. Each of these countries tax gains on the disposition of capital assets in ways that are idiosyncratic to their systems. The joy of comparative studies lies in part in discovering these idiosyncrasies and trying to make sense of them. Littlewood and Elliffe’s collection offers separate chapters on the approach to the taxation of capital gains in each of the countries mentioned above. The author of each chapter offers insights into the capital gains tax’s history; its revenue-raising capacity; its degree of integration with the income tax; the timing of its imposition (realization or accrual); its scope (residence- or source-based triggering events, such as immigration, emigration, or death); the use of losses; anti-avoidance provisions; administration; and the treatment of special assets (such as the family home). David Duff authored the Canadian chapter. The first part of the chapter outlines the history of the taxation of capital gains in Canada, including a detailed review of the Carter commission’s7 recommendations and the narrowing of the scope of tax as a result of the 1969 white paper.8 A noteworthy wrinkle in the white paper recommendations, never implemented, was the proposal that gains and losses on shares of widely held Canadian corporations should be taxed every five years. In part 2, Duff reviews the basic framework for capital gains tax in Canada, describing its scope, the computation of gains and losses, and timing. In part 3, Duff turns to special rules and design problems. The chapter ends with some lessons to be learned from the Canadian experience. Duff takes a position on the advisability of Canada’s policy choices, noting that Canada’s basic concepts and structural features (such as the computation rules, the loss limitation rules, the personal-use property rules, and the listed personal property rules) present good models. He cautions against adopting other aspects of Canada’s system, including our 50 percent inclusion rate and the exemptions for principal residences, lottery winnings, specified

6 See, for example, Benjamin Alarie and David G. Duff, “The Legacy of UK Tax Concepts in Canadian Income Tax Law” [2008] no. 3 British Tax Review 228-52. 7 Canada, Report of the Royal Commission on Taxation (Ottawa: Queen’s Printer, 1966). 8 E.J. Benson, Proposals for Tax Reform (Ottawa: Department of Finance, 1969) (herein referred to as “the white paper”). 500 n canadian tax journal / revue fiscale canadienne (2018) 66:2 charitable gifts, qualified farm and fishing property, and qualified small business shares. K.B. Heather Field, “A Taxonomy for Tax Loopholes,” Houston Law Review (forthcoming) (papers.ssrn.com/sol3/papers.cfm?abstract_id=2963441) Tax commentators—whether in the media, academy, profession, government, or public—often invoke the phrase “tax loophole,” generally in a derogatory way. It is immediately apparent to people who spend time thinking about tax that the phrase is employed to describe many different types of provisions. Field attempts a taxonomy of the uses of the phrase. Her objective is to facilitate our analysis of its use in a particular context with the hope of improving discourse about tax reform and to determine what we can learn about particular groups by better understanding what they perceive to be tax loopholes. To test the value of her taxonomy, she reviews the uses of the phrase tax loophole by the media (cnn, Fox, The Wall Street Journal, and The New York Times) in the 18 months leading up to the 2016 us presidential election. In part 2 of the paper, Field offers some tentative definitions of tax loophole and ultimately argues against searching for a definition. In part 3, she suggests that understanding tax loophole discourse requires understanding the normative policy objection that the phrase is designed to capture and determining who should be held responsible for resolving this problem. Field identifies five policy concerns that motivate the label tax loophole: policy concerns about reduction in the tax system’s ability to raise revenue (for example, through deferral of tax on foreign income or tax expenditures); policy concerns about fairness (for example, because one group of taxpayers receives a preference, such as a step-up in basis on death, while another does not); concerns about the distorting effect of a provision; concerns about the complexity that loopholes add to the law; and concerns about the social policy objective that underlies a particular provision. These objections coincide relatively closely with the standard criteria for evaluating technical tax rules and tax expenditures, and the fact that they form the ground for Field’s taxonomy is not surprising. Field’s unique contribution is her effort to identify the persons or entities that are blamed when the label tax loophole is placed on a provision or policy. Again, five categories are identified as being blamed at various times for creating loopholes or failing to close them: the legislative branch; the executive branch (the president, Treasury, or the Internal Revenue Service); the judiciary; taxpayers; and tax advisers. These two dimensions of the taxonomy are intended to work together. Field suggests that the analyst should clarify the precise policy problem when the label tax loophole is used and who is responsible for rectifying the problem. If speakers and listeners were more precise, we would better understand what policy reforms are needed and who might be responsible for the change (or we would be in a position to contest whether any proposed changes are necessary). current tax reading n 501

Finally, in part 4 Field applies her taxonomy to the media discussions leading up to the 2016 us presidential election. Field’s taxonomy provides her with an analytical tool to identify the underlying policy concerns (for example, fairness or economic distortion) that were reflected when the mantra tax loophole was invoked and to identify divergences in discovering who might be responsible for responding to the policy failings. Field’s purpose is not to eradicate or remove tax loopholes; rather, it is to offer us ways in which we can better understand underlying tax policy concerns, to identify the parties who might be engaged in their discussion (and possible reform), and to facilitate an understanding of larger political narratives. K.B.

Jost H. Heckemeyer and Michael Overesch, “Multinationals Profit Response to Tax Differentials: Effect Size and Shifting Channels” (2017) 50:4 Canadian Journal of Economics 965-94 There have been many studies of the effect of international corporate tax differentials on profit shifting. Most of them have estimated profit shifting indirectly by estimating how the profits of a parent firm or its subsidiary respond to tax differentials. These studies have used different techniques and databases, and have estimated widely different tax semi-elasticities of profits (the percentage change in profits resulting from a 1 percent change in international tax rate differentials). Heckemeyer and Overesch undertake a so-called meta-analysis of the different studies, which involves using characteristics of each study as dummy variables to estimate the source of differences in the studies’ estimates. This enables the authors to obtain a consensus estimate of the tax semi-elasticity and to estimate the relative importance of profit shifting by means of different mechanisms (transfer pricing versus financial planning). Their consensus estimate of the tax semi-elasticity of reported profits is 0.8, and about 80 percent of this elasticity is the result of non-financial profit shifting (transfer pricing and licensing), rather than shifting using financing mechanisms (debt shifting). The data are drawn from 27 studies of tax semi-elasticity estimates in the literature over a period of 20 years from which the authors obtain 203 observations. Estimates in these studies vary from less than 1 to over 2.5, and the estimates have been falling over time. The studies differ on the basis of whether they use total profit or profit before interest as the dependent variable. This distinction allows the authors to attribute differences in estimated elasticities to those that are the result of financial versus non-financial profit shifting. In their meta-regressions, Heckemeyer and Overesch identify four reasons why estimated tax effects might differ across studies. They include the profit variable used (before- versus after-financing profit); data sample characteristics (aggregate versus firm-level); econometric specification (subsidiary fixed effects, country fixed effects, time fixed effects); and tax rate variable (host country tax rates versus tax differentials). Each of these differences is captured by a dummy variable. 502 n canadian tax journal / revue fiscale canadienne (2018) 66:2

The consensus tax semi-elasticity when all dummy variables are included is 0.786 and is highly statistically significant. Taking into account the estimated coefficient for after-financing profitability leads to an estimate of the tax semi-elasticity of after-finance profits of 0.641, which is 82 percent of the tax semi-elasticity of total profits. The implication is that 82 percent of profit shifting is from non-financial sources, such as transfer pricing and licensing. Financial shifting accounts for only 18 percent of total shifting. These results are relevant in informing ongoing attempts to address international base erosion that arises from profit shifting. They also highlight the potential gains from international cooperation. R.B.

Laura Power and Austin Frerick, “Have Excess Returns to Corporations Been Increasing Over Time?” (2016) 69:4 National Tax Journal 831-46 The recent corporate tax reform in the United States is notable for reducing rates and moving to a territorial system. Somewhat less noticed, but arguably more important, is the change in the corporate tax base, which by allowing the expensing of investments and restricting interest deductions, moves the system toward a cash flow corporate tax. Relative to the existing corporate tax system that taxes all income earned for shareholders, a cash flow tax applies only to shareholder income above a competitive risk-free return, what Power and Frerick call “the excess returns to corporations.” This reform reduces the distortions imposed by the corporate tax, but at the expense of a smaller base and less tax revenue. In this article, Power and Frerick estimate how the excess returns to us corporations, equivalently their cash flows, have changed over time. To be more precise, corporate returns to shareholders consist of four components: marginal risk-free return, inframarginal return (economic profit), risk premium, and ex post luck. The latter three components constitute excess returns. The existing corporate tax is designed to include all four components in its base, while cash flow corporate taxes exempt the marginal risk-free return and therefore include only excess returns. In this article, the authors estimate what share of corporate returns consist of marginal risk-free, or competitive, returns. They provide these estimates for so-called c corporations—that is, corporations that are taxed separately from their individual shareholders. Previous estimates for the non-financial sector in the United States revealed that risk-free returns accounted for 32 to 37 percent of the total returns to capital; the rest were excess returns. Power and Frerick also provide these estimates, but over a longer period: 1992 to 2013. They remove all financial income from the tax base, and therefore they deal only with real cash flows (referred to as “ther base” in the Meade report9). They estimate the effects of moving from the existing system,

9 Institute for Fiscal Studies, The Structure and Reform of Direct Taxation: Report of a Committee Chaired by Professor J. E. Meade (London: Allen & Unwin, 1978). current tax reading n 503 which allows depreciation and interest deductions, to one that expenses capital investments and allows no interest deduction. Tangible, intangible, and inventory investments are all included. The authors’ results show that the marginal risk-free component of the corporate tax base is about 40 percent of the total for the first half of the time period, and falls to 25 percent for the second half. Multinational corporations have a slightly lower marginal risk-free component, and it declines more rapidly over the period (from 30 to 15 percent). Thus, excess returns are significant, and they are increasing over time. The marginal risk-free component varies across industries. For the period from 2011 to 2013, utilities, construction, and mining have risk-free returns in the order of 60 percent of total, while information and software, non-durable manufacturing, and chemicals have only about 10 to 20 percent and therefore have high excess returns. These high excess returns reflect the importance of returns to intangibles and intellectual property. No attempt is made to disaggregate excess returns by component (for example, risk premium versus rents), nor is there any attempt to estimate personal tax changes that could follow from corporate tax reform. These results also ignore any changes in behaviour that moving to a cash flow corporate tax would induce. Presumably, there would be some stimulation of corporate investment, given that a cash flow tax removes some distortions from the existing system. Overall, the results suggest that the revenue cost of moving to a cash flow tax system is modest, and might be a cost worth incurring to improve the efficiency of the corporate tax. No direct inferences may be drawn for the Canadian case, although there is no reason to believe that they would be substantially different. The size of excess returns in Canada would be worth investigating. R.B.

Vitor Gaspar, Laura Jaramillo, and Philippe Wingender, Tax Capacity and Growth: Is There a Tipping Point? IMF Working Paper no. 16/234 (Washington, DC: International Monetary Fund, 2016)

Gaspar et al. study the relationship between the tax-gross domestic product (gdp) ratio and economic growth across countries. In particular, they investigate whether there is a minimum tax-gdp ratio—that is, a tipping point beyond which gdp growth accelerates significantly in response to even small increases in tax-gdp ratios. The authors study this question using two detailed databases. One covers 139 countries from 1965 to 2011, and the other is a historical database that covers 30 advanced economies from 1800 to 1980. Strikingly, the authors estimate that there is a significant tipping point of about 12.75 percent of gdp, and that this applies across countries despite their diversity. They find that a country just above this threshold will have a gdp per capita that is 7.5 percent larger after 10 years. Their estimates take into account the endogeneity of the tax-gdp ratio both by focusing on local effects, rather than estimating global relations, and by using standard instrumental variable techniques. 504 n canadian tax journal / revue fiscale canadienne (2018) 66:2

The authors begin with a brief review of the interdependent relationship between taxation and economic development. Economic development facilitates tax collection, although tax collection falls with resource rents. The quality of governance and legal institutions improves tax capacity, as do social norms of compliance. Taxes can increase economic growth, for example, by means of infrastructure spending and by encouraging accountable and transparent government. These interdependencies can lead to multiple equilibria through feedback loops among economic agents, and these can result in tipping points. As countries approach and eventually exceed a revenue threshold, growth outcomes jump discontinuously. The stylized facts support these interdependencies. Both cross-section and historical data reveal a positive relationship between the tax-gdp ratio and real gdp per capita. As well, higher tax-gdp ratios are positively related to measures of property right protection, governance, and quality of budget institutions, and negatively related to corruption index measures. The authors then present a more detailed empirical analysis based on a regres- sion discontinuity design. The per capita rate of growth in gdp is regressed against the tax-gdp ratio relative to an endogenous threshold level, where the endogenous threshold level is obtained by estimating a structural break in the relationship between gdp per capita growth and the tax-gdp ratio. Using the contemporary database, the authors find that a tipping point occurs at a tax-gdp ratio of approximately 12.5. Countries to the left of the threshold grow at about 2 percent per year, while those above the threshold grow at 2.8 percent; crossing the tax-gdp threshold adds approximately 1 percent to annual real per capita gdp growth for the next 10 to 15 years. For the historical database, a stable threshold again occurs at a tax-gdp ratio of approximately 12.5. After 10 years of crossing the threshold, the cumulative effect on gdp is approximately 16 percent, similar to the contemporary database, and rises to 25 percent after 15 years. On the basis of these findings, the authors recommend that countries with low tax-gdp ratios aim to increase them to about 15 percent, presumably hoping that these policies will enhance social norms, governance, and accountability and lead to the desired amount of gdp growth. R.B. Christiana H.J.I. Panayi, “The Europeanization of Good Tax Governance” (2018) 36:1 Yearbook of European Law 442-95 (academic.oup.com/yel/ advance-article/doi/10.1093/yel/yex020/4803230?guestAccessKey= 722200ac-d23c-4fbc-81d2-efd462e4e9b3) Panayi set aside the common international tax topics of tax evasion, tax avoidance, cooperation, and exchange of information to focus on the emergence of debates about “good tax governance” on the international agenda. In part 2 of the article, the author investigates the concept of good tax governance, and parts 2 and 4 reveal how this concept has been employed by the European Commission. For a Canadian audience, the work in part 2 is likely to be most useful. Certainly, given the plethora of international tax initiatives, one could be forgiven for needing current tax reading n 505 a refresher on the various ongoing tax actions. In part 2, which is approximately 15 pages long, Panayi offers a taxonomy of good tax governance activities. Building from the concept of governance that is employed in corporate governance and development literatures, she organizes good tax governance activities into three broad categories: activities that support domestic revenue mobilization and capacity building, activities that support enhanced cooperation and cooperative compliance, and activities that adhere to exchange-of-information and reporting requirements. Panayi offers an overview of the international tax initiatives that fit within each of these three categories. In the category of domestic revenue mobilization and capacity building, among other international initiatives, she identifies work by the Organisation for Economic Co-operation and Development (oecd) (for example, its 2010 report Citizen-State Relations: Improving Governance Through Tax Reform),10 which offered recommendations for governments to support effective tax system development; the United Nations (un) (for example, its “Doha Declaration on Financing for Development” in 2008),11 which emphasized the importance of strengthening technical assistance and enhancing tax revenue collection; and the International Monetary Fund, oecd, un, and World Bank creation of the platform for collaboration on tax, launched in 2016, which is intended to facilitate greater cooperation on tax matters among these supranational organizations. The oecd’s initiatives serve as the focus for the activities that Panayi identifies as supporting enhanced cooperation and cooperative compliance.12 In this category of initiatives to support good tax governance, among other oecd initiatives, she identifies the creation of the forum on tax administration in 2002, the 2008 study about the role of tax intermediaries,13 the 2013 report on cooperative compliance,14 and the 2016 report on tax control frameworks.15 Finally, in the category of adherence to multilateral exchange of information and reporting,16 she focuses on “norm setting” by government or supragovernmental bodies. Again, the oecd looms large, with its emphasis on common reporting standards and country-by-country reporting.

10 Wilson Prichard, Citizen-State Relations: Improving Governance Through Tax Reform (Paris: OECD, 2010). 11 United Nations, “Doha Declaration on Financing for Development: Outcome Document of the Follow-Up International Conference on Financing for Development to Review the Implementation of the Monterrey Consensus,” A/Conf.212/L.1/Rev.1, December 9, 2008. 12 At 451. 13 Organisation for Economic Co-operation and Development, Study into the Role of Tax Intermediaries (Paris: OECD, 2008). 14 Organisation for Economic Co-operation and Development, Co-operative Compliance: A Framework—From Enhanced Relationship to Co-operative Compliance (Paris: OECD, 2013). 15 Organisation for Economic Co-operation and Development, Co-operative Tax Compliance: Building Better Tax Control Frameworks (Paris: OECD, 2016). 16 At 454. 506 n canadian tax journal / revue fiscale canadienne (2018) 66:2

Ultimately, Panayi’s review of activities undertaken in the name of good tax governance leads her to the conclusion that initiatives are moving from “soft law” to “hard law”: they are becoming legally mandated norms, enforced through legislation enacted in countries around the world. K.B.

Martin Hearson, “The UK’s Tax Treaties with Developing Countries During the 1970s,” in Peter Harris and Dominic Cogan, eds., Studies in the History of Tax Law, vol. 8 (Oxford, UK: Hart Publishing, 2017), 363-81 (eprints.lse.ac.uk/74104/) Hearson has been rapidly developing a body of scholarship that explores the politics of international business taxation through the lens of the tax treaties entered into between developed and developing countries. In this article, he draws on civil service documents to examine Britain’s tax treaty negotiations with approximately 40 developing countries throughout the 1970s. The conventional understanding is that developing countries agree to tax treaties, which limit their jurisdiction to tax, to encourage foreign investment. Hearson suggests that this idea lacks nuance. He suggests that many tax treaties between the United Kingdom and developing countries include clauses that the negotiators failed to appreciate (and therefore did not spend time negotiating), that often developing countries used whatever bargaining power they might have had to secure tax-sparing clauses, and that the parties (particularly the United Kingdom) were sometimes concerned about the consequences of deviations from the oecd model for future negotiations and as a result failed to tailor bilateral treaties to the context of the particular parties. In support of this view, Hearson offers case studies focused on treaties’ tax-sparing provisions, shipping articles, and withholding taxes. Hearson’s article includes excerpts from the notes of those negotiating tax treaties on behalf of the United Kingdom. These excerpts are revealing and sometimes funny. It is worth reading the article simply to get a taste for the attitudes and reflections of the tax treaty negotiators of that era. Hearson concludes the article with a call for the United Kingdom to revisit some of its earlier tax treaties (quite a few are still in force). Among other justifications for this stance, he notes that the conditions that resulted in these treaties (for example, the need for tax sparing) have changed dramatically in the last 40 years. K.B.

Ajay K. Mehrotra, “Fiscal Forearms: Taxation as the Lifeblood of the Modern Liberal State,” in Kimberly Morgan and Ann Orloff, eds., The Many Hands of the State: Theorizing the Complexities of Political Authority and Social Control (New York: Cambridge University Press, 2017), 284-305 Part of what makes reading Mehrotra’s work such a pleasure is his wide-ranging interest in the academic insights of other disciplines. He begins his article with current tax reading n 507

Pierre Bourdieu (which makes sense because the article is a chapter in a book that theorizes about the state from the vantage point of several disciplines). He invokes Bourdieu’s metaphor of the state: a metaphor that distinguishes the fiscal ministerial cabinets from the spending ministries. In contrast, Mehrotra posits the state as a Hindu deity with multiple arms. His point is that our concept of the state has evolved dramatically from the days when it was theorized as a monolithic entity and described in relatively straightforward ways. Mehrotra draws on his extensive work in tax history to demonstrate how tax systems in liberal democracies require constant negotiation and renegotiation because they are fundamental to our sense of how democracy should be manifest. Readers who have no time to read Mehrotra’s longer work can gain a sense of his argument in this article.17 The interaction between taxation systems and theories of the state and the influence of tax structures on concepts of citizenship are the focus of the first two sections of the article. Readers who are not particularly interested in these matters may nevertheless find the section on state delegation to the private sector of the obligation to withhold fascinating. Mehrotra details the introduction of “stoppage at the source” withholding in the United States in 1913 (when withholding was imposed on persons or organizations making payments of more than $3,000 in salary, interest, or other fixed income, such as rent and dividends). While widely accepted now, the idea that the state could impose (with democratic acceptance) on private organizations or citizens the obligation to withhold and remit taxes on behalf of other private citizens seems far from obvious in the historical account. K.B.

Eduardo Baistrocchi, ed., A Global Analysis of Tax Treaty Disputes (Cambridge, UK: Cambridge University Press, 2017), 2 vols., 1588 pages Baistrocchi’s collection is a must-have for readers interested in tax treaties. It might aptly be referred to as “the definitive encyclopedia on tax treaty case law.” Work by 42 authors spans 27 countries and provides an analysis of over 580 tax treaty cases. The two-volume collection follows the familiar International Fiscal Association format: an introduction; chapters devoted to individual countries; and conclusory chapters, each of which brings a comparative approach to an aspect of tax treaty disputes. The author of each chapter about a particular country includes his or her responses to a list of eight common questions, including the structure of the law for resolving disputes about the interpretation of tax treaties, the mechanisms for treaty implementation, the differential treatment of active and passive income, and the treatment of intangibles.

17 See, for example, Ajay Mehrotra, Making the Modern American Fiscal State: Law, Politics, and the Rise of Progressive Taxation, 1877-1929 (New York: Cambridge University Press, 2013). 508 n canadian tax journal / revue fiscale canadienne (2018) 66:2

The chapters about individual countries are lengthy and detailed. In his chapter about Canada, Arthur Cockfield offers an analysis of each of the questions posed to the contributors of these chapters. Not surprisingly, his review of the Canadian context focuses on our relationship with the United States. Canada has been a net capital exporter of foreign direct investment since 1996; in addition to the United States, the United Kingdom and Australia are Canada’s major trade partners. Not surprising to a tax audience, Barbados, the Cayman Islands, Luxembourg, and Ber- muda also find their way into our top ten foreign direct investment partners. Cockfield’s review of the major tax treaty disputes highlights the usual suspects, such as Crown Forest, grouping the cases around the core themes of the disputes. He analyzes tax treaty interpretation case law, including a discussion of how the courts have approached the meaning of terms in tax treaties and the use of extrinsic materials. He assesses the available case law on the allocation of jurisdiction between passive and active income, the elimination of double taxation, treaty shopping, non- discrimination, and the general anti-avoidance rule. Finally, he offers an overview of dispute resolution mechanisms, including mutual agreement procedures and arbitration. Cockfield provides outstanding coverage and offers a good starting place for colleagues who want a refresher in Canadian tax treaty law or who want a sophisticated sense of how treaty disputes have arisen and been resolved in Canada. The four conclusory chapters reveal the benefits of comparative tax law. Perhaps the chapter by Eduardo Baistrocchi, “Patterns of Tax Treaty Disputes: A Global Taxonomy,” is the most engaging. Baistrocchi offers a series of taxonomies in an effort to organize and analyze the tax treaty cases around the world. The reader is able to see not only global trends in tax treaty disputes but also concrete illustra- tions of the various types of disputes. The work reflected in this chapter is extraordinary. Baistrocchi has managed to distill pages and pages of reports from individual countries into a cohesive and informative whole. The book includes a table of cases and a well-developed index; perhaps most helpfully, it has a table that lists cases by the type of treaty dispute involved. For example, a reader can look up cases about the application of treaty anti-abuse provisions under subcategories such as residence tests, beneficial ownership, and implicit anti-abuse. Each of these subcategories is further subdivided—for example, “beneficial owner” includes a category for dividends (in whichPrévost Car appears), interest, royalties (in which Velcro appears), and capital gains. This review does not do justice to the rich resource that this double volume provides for tax practitioners, policy makers, and scholars with an interest in tax treaty disputes. It is a stand-out contribution in the area. K.B. current tax reading n 509

Michael Lang, Pasquale Pistone, Alexander Rust, Josef Schuch, and Claus Staringer, eds., The UN Model Convention and Its Relevance for the Global Tax Treaty Network (Amsterdam: IBFD, 2017), 325 pages Michael Lang produces edited collections just as Louis Feuillade directed films. In this volume, Lang (along with 4 co-editors) includes 19 authors writing on 11 tax treaty topics. The articles were originally drafted in response to a symposium aimed at exploring the differences between the un and the oecd model tax conventions. In each chapter, the authors are charged with exploring the degree to which the un model continues to hold sway in a range of international tax areas, including treaty interpretation, the definition of permanent establishment, the jurisdiction to tax profits from international traffic, business profits, personal services, capital gains, directors’ fees, pensions, and passive income. The book also addresses the important administrative topics of dispute resolution, information exchange, and the elimination of double taxation. The work would benefit from the addition of a few synthesizing chapters, an index, and a table of cases. Nevertheless, it is a welcome addition to the literature for those with an interest in tax treaty policy and design. K.B.