SUMMARY The treatment of financial services has long been viewed as one of the more technical, and difficult, areas in value added taxation. Financial intermediaries add value by reducing transaction costs for clients. While in principle this value added should be taxed under a comprehensive VAT, this has proven to be difficult to do in practice because of measurement issues. The predominant approach adopted in most countries — albeit with several variations on the theme — has thus been to exempt most financial services from VAT. This was the approach adopted in Canada at the initiation of the GST just over twenty years ago. While this approach is far from perfect, and introduces several distortions into the economy, it has by and large been concluded that it is the most practical approach to dealing with financial services under a VAT. Two decades of legal wrangling and Ottawa's habit of retroactively legislating changes to the GST as it relates to financial services have served to muddy the waters in Canada. Recent changes have significantly altered the scope for exemption and resulted in an uneven playing field across financial services. This paper argues that the best solution for Canada is to stick with the exemption approach, but to go back to basics with an eye for reducing existing distortions and restoring a semblance of neutrality. Specifically, the paper calls for a reset of the "arranging for" exemption for financial services; the creation of a new GST-recovery system for financial services; a new structure for taxing imported supplies; and a limit to retroactive legislative amendments and minimum requirements for future amendments. The authors also argue that consideration should be given to zero-rating "business to business" financial transactions so as to remove the GST embedded in transactions between financial institutions and businesses.
I. INTRODUCTION
January 1, 2011 marked the 20th anniversary of the introduction of the GST in Canada. Five provinces are now fully harmonized into the federal taxing statute (New Brunswick, Newfoundland & Labrador, Nova Scotia, Ontario and British Columbia), one of those (British Columbia) having recently resolved through a plebiscite to reverse that move, and one province (Quebec) has been promised $2.2 billion in subsidy payments by the federal government for harmonizing, notwithstanding it will continue to operate a provincial near-clone of the tax in a provincial statute. While some progress has been made, only when the GST is fully harmonized with a uniform rate and base will Canada reap the full rewards of introducing the tax.
To be so far from achieving that goal after 20 years is a sad story of missed opportunity, and a future of continuing and unnecessary administrative and compliance costs. Moreover, after 20 years, despite numerous tweaks, the GST/HST is beginning to show its age. In particular, while the GST was most certainly not a perfect VAT at its inception, there is reason to believe that it has drifted even farther from some notion of an ideal VAT in the intervening 20-year period.1 This, we think, is particularly true in the treatment of financial services.
The purpose of this paper is to provide an analysis of the GST as it relates to financial services in Canada. The treatment of financial services is perhaps one of the more complicated and controversial areas in the analysis of VAT. To lay the groundwork for the discussion we begin with a discussion of first principles, primarily from the theoretical economics literature, on the subject of VAT applied to financial services. The underlying question that motivates this discussion is: if we could apply the GST to financial services, should we? We begin with a general discussion of first principles as they apply to the taxation of all goods and services, and then refine the discussion to the financial sector.
We then turn to a brief discussion of the nature of financial services and the underlying difficulties that are associated with taxing them. We rotate the question posed in the previous paragraph, asking: if we should apply the GST to financial services, could we?
This is followed by a description and analysis of the current system in Canada as it applies to the financial sector. The treatment of financial services under the GST is characterized by the exemption of many activities from the tax. The numerous legislative changes that have occurred over the last 20 years, in particular over 2009 and 2010, involved significant changes in the scope for exemption.
We then address the broad question of whether or not the financial sector is advantaged by the current exemption-based system. We argue that the answer to this question is not straightforward, and depends upon a number of key factors.
This is followed by a summary of where we stand on a number of the issues, and by our list of recommendations. Our emphasis is on suggesting practical options for improving the working of the exemption-based approach to taxing financial transactions, as it currently exists in Canada. The approach is grounded by a familiarity with the current state of affairs, and concerns about the ability to implement options that deviate significantly from the basic exemption approach. We end with a short concluding section.
In terms of what we do not cover, this paper is not intended to be a cookbook of the VAT as it applies to financial services. In particular, it does not include an extensive discussion of the approaches to taxing financial services taken in other countries around the world, or of various (largely theoretical) alternative approaches to taxing financial institutions under a VAT. While we do discuss these matters where appropriate, we do not deal with them in a systematic way. These issues have been well discussed and documented elsewhere.2
II. IF WE COULD APPLY THE GST TO FINANCIAL SERVICES, SHOULD WE?
General Discussion
We begin with a general discussion of the VAT from a theoretical perspective. While we take a very pragmatic approach in the rest of the paper — understanding that tax policy falls under the heading of the art of the possible — we think that these theoretical considerations are important not only to fix ideas, but to establish a prism through which we can filter the ensuing discussion. Following the general discussion, we refine the dialogue to consider the implications for financial services in particular. For the most part practical issues associated with applying VAT to financial intermediation will be ignored in this section (but will be addressed in more detail below).
To address the question posed in the title to this section we begin with a discussion of first principles underlying the VAT. Virtually by definition, a fundamental first principle underlying a comprehensive VAT is that the final consumption of all goods and services should be taxed at a uniform rate. The key words here are "final," "all" and "uniform." This principle is fairly straightforward and is in fact the defining feature of a pure VAT.3
However, the notion that a consumption tax should be levied on all goods and services at a uniform rate is in fact antithetical to much of optimal tax theory in economics. There are two seminal results, or first principles, in optimal tax theory that guide the thinking of economists in this regard. The first concerns which goods should be taxed, and the second concerns the optimal configuration of tax rates on the goods that are taxed.
With regard to the former — which goods should be taxed — one of the most important guiding principles in the theory of tax policy is that, in general, taxes should only be levied on final consumption (purchases by final consumers), and not on intermediate transactions or inputs (purchases by businesses). This insight is formally contained in the seminal production efficiency theorem developed by Diamond and Mirrlees.4The production efficiency theorem states that if there are no restrictions on the tax instruments that can be deployed by the government, including the use of taxes on pure economic profits, distortionary taxes should not be imposed on intermediate transactions. This theorem follows from the simple fact that any distortionary tax levied on intermediate inputs will lower aggregate output by introducing production inefficiencies (distortions in the use of intermediate inputs); eliminating these inefficiencies, and using other tax instruments to capture a share of the resulting increase in aggregate output, in order to generate the required tax revenue, is desirable. It turns out that this idea is perfectly consistent with the idea behind a comprehensive VAT mentioned above, which in principle does not tax intermediate inputs, but only final consumption goods. We will have reason to refer to the production efficiency theorem several times in the ensuing discussion.
The second issue — the optimal configuration of the rates imposed on final consumption goods that are subject to tax — falls under the purview of the so-called Ramsey Rules of optimal commodity taxation. In their simplest form these rules abstract from distributional considerations, focusing on efficiency, and describe the characteristics of a commodity tax system that minimizes the cost of economic distortions associated with raising a given amount of revenue. These distortions arise because of changes in relative prices associated with taxes.
There are two popular interpretations of the Ramsey Rules.5One falls under the heading of the Inverse Elasticity Rule, which says that under certain conditions higher tax rates should be levied on goods with lower demand elasticities.6Low elasticity means that the demand for a good is not very responsive, or sensitive, to changes in its price. So, the Inverse Elasticity Rule says that in order to minimize the costs associated with distortions caused by the commodity tax system, we should levy higher tax rates on goods whose demands are not very responsive to price changes.
The second interpretation, due to Corlett and Hague,7and not surprisingly referred to as the Corlett-Hague Rule, is more general and says that to minimize the costs of distortions caused by the tax system, goods that are more complementary with the consumption of leisure (which is generally viewed as being non-taxable) should be taxed at higher rates. The phrase "complementary with the consumption of leisure" perhaps requires some interpretation. First, what do we mean by leisure? Leisure is simply time not devoted to paid work. Economists treat leisure as a good just like any other good, the price of which is the forgone wages that could have been earned by working. So when economists refer to consuming leisure, they mean spending time doing anything other than earning a wage. Leisure is a non-taxable good simply because it cannot be properly measured due to the absence of a market based transaction; we cannot measure it so we cannot tax it. What about complementary? Loosely speaking, a good is complementary with leisure if it tends to be consumed jointly with leisure.8Thus, and again speaking loosely, the Corlett-Hague Rule says that because leisure cannot be taxed directly, we can tax it indirectly by imposing higher taxes on goods that are consumed jointly with leisure. This will reduce the incentives for individuals to substitute away from taxed goods towards the non-taxed consumption of leisure, and reduce the distortions caused by the commodity tax system.
The important lesson of both the Inverse Elasticity Rule and the Corlett-Hague Rule is that, in general, since goods differ in their degree of complementarity with leisure and in their demand elasticities, optimal commodity tax considerations emphatically do not proscribe uniform tax rates on all goods and services.
Finally, it should be noted that the introduction of distributional considerations into the mix can argue for even more deviation from uniformity, as goods consumed disproportionately by individuals with high marginal social valuations should be taxed at a lower rate.9
These two sets of first principles — the idea of a comprehensive and uniform GST on all final consumption goods, and the non-uniform taxation associated with optimal commodity tax theory — would appear to be in direct contradiction to each other. How can we square the circle? There are several ways to approach this.
First, one can think of a uniform and comprehensive GST as equivalent to a proportional tax on labour income. In this case, one can imagine a uniform GST combined with a set of individualized commodity taxes (or subsidies) to accommodate optimal commodity tax considerations. The GST may then be viewed at the average rate imposed on commodities, and in this sense should be levied on a comprehensive rate at a uniform rate. While this view is widely held,10it really just avoids the uniformity question by defining it away. We are, and should be, concerned with the properties of the overall commodity tax regime.
Second, implementing a commodity tax regime according to the principles of the Ramsey Rules would be administratively costly and difficult to implement in practice. The informational requirements and the administrative and compliance effort required to implement an optimal commodity tax regime along these lines are extremely high. While this may be viewed as more of a practical than a theoretical consideration, and as such should be relegated to the discussion of practical issues to be addressed below, it bears mentioning here because research that explicitly incorporates administrative costs into a theoretical optimal commodity tax framework shows that these costs tend to push the optimal commodity system towards uniformity, at least across broad classes of goods.11
Third, it turns out that in the special case where consumer preferences take a particular form, and the government has other instruments to achieve distributional objectives (such as a nonlinear income tax), the optimal commodity tax system in the Ramsey sense does in fact call for a comprehensive tax at uniform rates.12While this result neatly squares the circle with respect to our two sets of principles — it calls for a comprehensive and uniform GST on final consumption even accounting for optimal tax considerations — it rests on a restrictive set of assumptions regarding individual preferences that we know are not likely to be true in practice.13
So where does this leave us? As the above discussion suggests, it is not a straightforward matter to write down a set of robust first principles that may be used to address the question of what should and should not in principle be taxed under a GST, and at what rates. Depending upon how the question is posed and what assumptions are made, in principle the GST should be levied on a comprehensive base at uniform rates, or at differentiated rates according to their demand elasticities and/or the complementarity of the goods with leisure, and possibly differentiated according to the consumption patterns of different groups in society.
What Does this Mean for Financial Services?
Turning the question explicitly to financial services, when addressing the question of whether financial services should (if they could) be taxed, it is important to distinguish between the purchase of financial services by businesses and consumers. For purchases by businesses — so called business-to-business (B2B) transactions — the literature provides an answer that is relatively clear: purchases of financial services by businesses should not be subject to GST. For purchases by consumers — business-to-consumer (B2C) transactions — the answer is somewhat less clear.
Kenneth McKenzie (The School of Public Policy and Department of Economics, University of Calgary) also co authored this.
Originally published in The School of Public Policy Volume 5"Issue 29"September 2012
To read the rest of this article please click here.
Footnotes
1 For more on this, see Smart, M. (2012), "Departures from Neutrality in Canada's Goods and Services Tax," SPP Research Papers, Volume 5, Issue 5, February. The School of Public Policy, University of Calgary.
2 See European Commission (1997). Value-Added Tax: A Study of Methods of Taxing Financial and Insurance Services (Brussels: European Commission).
3 See, for example, Poddar, S. (2003), "Consumption Taxes: The Role of the Value-Added Tax," in Patrick Honohan (Ed.), Taxation of Financial Intermediation: Theory and Practice for Emerging Economies, Washington, DC: The World Bank.
4 Diamond, Peter and James Mirlees (1971), "Optimal Taxation and Public Production I: Production Efficiency," The American Economic Review 61(1), 8-27.
5 For a more in-depth discussion of the Ramsey Rules at the undergraduate level see Rosen, H., B. G. Dahlby, R. Smith, J.F. Wen and T. Snodden (2008), Public Finance in Canada, 3rd edition. McGraw-Hill. For a more technical treatment see Auerbach, A. and J. Hines Jr (2002), "Taxation and Economic Efficiency," Chapter 21 in Handbook of Public Economics, Volume 3, A. Auerbach and M. Feldstein eds., Elsevier Science.
6 The inverse elasticity rule arises under the extreme simplifying assumption that the cross-price elasticity between all taxable goods is zero.
7 Corlett, W.J. and D.C. Hague (1953), "Complementarity and the Excess Burden of Taxation," Review of Economic Studies (21), 21-30.
8 More formally, a good is complementary with leisure if the compensated demand for leisure declines as the price of the good increases.
9 Diamond ((1975), "A Many-person Ramsey Tax Rule," Journal of Public Economics, 335-43) first added equity objectives to the Ramsey analysis. Individuals, or groups of individuals, can have a high marginal social valuation for two reasons. The first is because the underlying "social welfare function" governing the choices of government places a higher weight on some individuals. The second is that the assumptions that economists typically make regarding individual preferences suggests that their own individual marginal valuation of income declines as their income increases. Thus, the contribution that an extra dollar of income makes to the well-being of a low-income individual is higher than that of a higher-income individual. These two factors point to a higher marginal valuation placed on income received by lower-income individuals, and in favour of levying lower tax rates on goods consumed by those individuals.
10 See Poddar (2003) op. cit.
11 See Kopczuk, W. and J. Slemrod (2006), "Putting Firms into Optimal Tax Theory," American Economic Review (Papers and Proceedings) 96(2): 130-34.
12 A. Akinson and J. Stiglitz ((1976) "The Design of Tax Structure: Direct versus Indirect Taxation," Journal of Public Economics 6, 55-75) show that if the income tax is optimally designed to achieve distributional objectives, and preferences are weakly separable between consumption goods and leisure, the optimal commodity tax rate is uniform across all goods. As discussed above, this can alternatively be viewed as eliminating the VAT altogether, and augmenting the non-linear income tax with a linear proportional tax on (non-capital) income. Louis Kaplow ((2006) "On the Undesirability of Commodity Taxation Even When Income Taxation is Not Optimal," Journal of Public Economics 90, 1235-1250) extends this result to show that it holds even when the non-linear income tax is not optimally designed. In particular, he shows that moving to a uniform consumption tax is optimal under any income tax system, so long as the income tax is adjusted so as to maintain "distributional neutrality." This result also rests upon the assumption of the weak separability of preferences. Finally, Angus Deaton ((1981) "Optimal Taxes and the Structure of Preferences," Econometrica 49, 1245-1260) abstracting from distributional considerations, employs yet another separability concept. He shows that uniform commodity taxes will be efficient if preferences between goods and leisure are quasi-separable.
13 Martin Browning and Costas Meghir ((1991) "The Effects of Male and Female Labour Supply on Commodity Demands," Econometrica 59(4), 925-951) test for the separability of commodity demands from labour and reject weak separability.
The content of this article is intended to provide a general guide to the subject matter. Specialist advice should be sought about your specific circumstances.
