Disrupting the market for tax planning.

I. INTRODUCTION

It is a pleasure to comment on Philip Curry, Claire Hill, and Francesco Parisi's article (1) on creating market failures for tax planning. It is rare that I get to comment on an article in which I agree with the basic conclusions and, more importantly, in which I agree with the methodology for reaching those conclusions. Their article provides a number of important and interesting insights and should provide the basis for much future work. I offer some comments and suggestions here.

Although their article is not organized exactly this way, I will break their conclusions into three parts. First, they analyze how to optimally set the line between legitimate and illegitimate tax planning. Second, they discuss whether patents for tax strategies are desirable. Finally, they argue that the government should exploit market failures to reduce tax planning.

II. LEGITIMATE V. ILLEGITIMATE SHELTERS

In the first part of their article, Curry, Hill, and Parisi discuss how to determine the optimal line between allowable tax planning and disallowed tax shelters, in their language, legitimate and illegitimate tax planning. Although they assume that all tax planning is socially unproductive and wasteful, they conclude that it would not be optimal to eliminate all tax planning. (2) The reason is that as the government increases the strength of its attacks on tax planning, individuals will incur additional search costs (dissipation costs) as the easy and obvious shelters become illegal. The government needs to balance the additional revenue from reducing tax planning with the additional search costs incurred when the line is moved to reduce the scope of legitimate tax planning. There will be some optimal line, which is likely to stop short of eliminating all tax planning.

There are two aspects of the analysis which are not typical and worth highlighting. First, they treat as the choice variable the line between legitimate and illegitimate tax planning. In the economics literature, this line is usually assumed and the question typically asked is about the taxpayers' decision in light of audit or detection policies. Thus, a typical economics paper will start off by noting that there is a distinction between avoidance and evasion (mapping onto legitimate and illegitimate) and then discuss the report/nonreport decision in light of some penalty structure. Curry, Hill, and Parisi instead treat this line as the choice variable, an approach that I think is valuable as this line is central to policy debates. Second, many legal analyses discuss the best way to draw this line but do not explicitly state the criteria they are using. Curry, Hill, and Parisi want to maximize welfare taking into account both government revenue and taxpayer costs. That is Curry, Hill, and Parisi use an economics approach to answer a question normally addressed only by lawyers. This is, in my view, the right way to approach tax law policy.

Their basic intuition seems to me to be correct, although there are a number of subtleties that are worth further exploration. Their original draft, presented at the Future of Tax Shelters conference, started to model the issue and I think that completing the model should be a priority. Consider some of the complexities. First, the government does not necessarily want to maximize total revenue less dissipation costs, as they suggest. Instead, we usually assume that the government has a fixed revenue constraint, with the money to be spent, say, on a fixed set of public goods. With a fixed revenue constraint and a fixed "tax dissipation function" (how taxpayers react to loophole closing), the answer would be predetermined, so such an approach would not work as a way to think about tax shelters. To make the argument work with a fixed revenue constraint, we have to posit an alternative source of funds that can be used instead of closing tax loopholes. The goal in such a case would be to set the marginal cost of closing tax loopholes equal to the marginal cost of the next source of funds. We might want to use tools such as Slemrod and Yitzhaki's marginal efficiency cost of funds formulation (3) or Slemrod and Kopczuk's optimal elasticity of taxable income formulation (4) to think about how to make such a trade-off. The way that I have conceptualized the issue in my writing is to think of loopholes as gaps in the tax base. (5) We want to balance all of the various methods of broadening the base as well as the overall rate structure. Addressing tax planning is part of this overall set of decisions and is no different, in a general sense, from broadening the base in other ways.

Second, Curry, Hill, and Parisi model taxpayer's costs as search costs. They argue that as the government closes off the obvious loopholes, taxpayers incur additional search costs to find new ones. (6) Although correct as part of the story, the costs of sheltering are likely to also include structuring or other similar costs. Thus, as the government shuts down the easy to find and use shelters, taxpayers must spend more to find new ones and also more to implement the new ones. For example, newer shelters may involve more complex changes to capital structures, risky investments, etc. Modeling the problem as search costs leads one to think about information flows (hence the discussion of patents in their article). Thinking about other costs of finding and entering into tax shelters suggests alternative avenues for investigation. Thus, the government's strategies will also involve such items as setting the degree of economic substance or risk that a transaction must have to be legitimate. The government should optimize along all these boundaries by setting the marginal benefit equal for each of them (and equal to the marginal cost of funds for other sources of revenue).

Third, later in the article the authors focus on market structure. (7) This focus might be incorporated into the study of the optimal strength of anti-shelter rules. For example, there might be economies of scale in shelter provision. If so, strategies that focus on different aspects of shelters might not be additive--they might be more than additive. Thus, if sheltering becomes expensive enough, shelter providers might shut down. Understanding the market structure for tax planning will likely be important to understanding the optimal line between legitimate and illegitimate structures.

Fourth, we need some way of operationalizing the approach. We need to get a sense of the optimal strength of attacks on tax shelters. We need to know the tax dissipation function--how taxpayers will react to various approaches. I have suggested previously that the Slemrod and Yitzhaki's marginal cost of funds might be a reasonable approach to the issue, although I am sure there are others. The assumption behind the Slemrod and Yitzhaki formula is that on the margin, all tax reduction strategies have the same cost. (8) This allows us to use the reduction in tax revenues for a given change in tax rates--the elasticity of taxable income--as a measure of the dissipation and all other costs of a tax rule. Measuring elasticities of taxable income under various tax regimes, although not easy, should be feasible.

III. PATENTS

Patents on tax strategies usually create a strong gut reaction in many that they are entirely inappropriate. Some react negatively to the idea of offering legal protection to something, like a tax shelter, that we should not want to encourage in the first place. Others, often practitioners, react to the "anticommons" effect--that patents on basic tax planning ideas will make it difficult to give everyday tax advice. Imagine a patent on some run-of-the-mill tax strategy. To take a random example, imagine a patent on various techniques used to qualify for a tax-free reorganization or liquidation. This is the kind of advice given everyday by tax lawyers. If these kinds of techniques were patented, it would be perilous to give tax advice for fear of violating a patent. Moreover, assembling the necessary licenses for a complicated tax strategy may be expensive because of the mere costs of the royalties, the costs of finding all of the patent holders, and the potential for hold-outs. Hence, the reaction that the patenting of tax advice is contrary to the basic structure of the industry today. Note that the two gut reactions point in opposite directions: patents will unduly encourage tax planning and patents will make tax planning impossible. It is possible that both hold true: everyday tax planning may be more difficult while exotic tax shelters may become more rewarding.

It is not clear that tax patents are going away. The original business method patent involved taxes. Neither the patent office nor the courts have given any indication of a reversal of course. We need to move beyond gut reactions to understanding their effect. Curry, Hill, and Parisi take a refreshingly objective approach to the issue. They want to know whether patents increase or decrease tax planning (which, recall, they assume is socially wasteful). (9) Patents increase the incentive to invest in information by creating property protection for an idea. But at the same time, by creating a monopoly, they potentially reduce dissemination. We cannot say, ex ante, which effect dominates. In the normal patent case, where the idea is socially useful, we want to increase the total supply, which means that we hope that the property protection and the incentives it creates for finding ideas outweigh the costs of giving a temporary monopoly once the idea is found. In the tax case, we want the opposite.