Creating failures in the market for tax planning.

I. INTRODUCTION

Although most people recognize the necessity of taxes, few people like to pay them. Governments expend costs to collect taxes; people expend resources to avoid paying them, engaging in tax planning activities that are, for the most part, socially wasteful. The government knows that people will engage in such activities; a benevolent government aiming to maximize social welfare should take the costs of such activities into account when designing tax policy. Tax policy thus best maximizes social welfare if it limits taxpayers' costs incurred in developing and using methods to avoid or minimize taxes (1) while also limiting lost revenue. Our article develops two related normative corollaries of this insight.

In Part II we highlight some stylized facts and working assumptions in order to develop a simple model. In Part III we elaborate our first normative proposition, concerning the government's optimal attitude towards tax planning activities. We suggest that given the tax revenue constraint that governments face, tax policy should be designed keeping in mind the two interrelated costs of tax planning: the costs arising from existing tax planning methods and the costs associated with the taxpayers' search for additional ones. In this article we identify an interesting tradeoff between these two costs.

Our article relates to the existing literature on optimal taxation. (2) In particular, it identifies an additional variable which may be relevant in the application of Samuelson's rule on public good provision. In its simplest form, the Samuelson rule (3) states that when deciding how much of a public good to supply, governments should supply it up to the point where the marginal cost of its provision equals the sum of the marginal benefits across all affected individuals. Scholars have since restated Samuelson's rule explicitly to include among the costs of public good provision the deadweight loss of taxation. (4) In this article we suggest that consideration of these other indirect costs of public good provision may lead to a more lenient attitude towards tax planning activities.

In Part IV we elaborate upon the second normative corollary of our article, concerning the use of governmentally-created market failures to combat tax planning activities. The literature typically discusses traditional legal approaches: increasing expected sanctions by increasing the severity of penalties and/or the likelihood of detection and enforcement. In this article we consider a different conceptual approach. Tax planning can be discouraged by creating failures in the market for tax planning activities. In the presence of market failures, the incentives for tax planning may be undermined. Paradigmatically, legal intervention is aimed at correcting market failures; here, we contemplate the possibility that market failures may be purposely created and utilized as instruments of tax policy. (5) We consider the conventional categories of market failure--externalities, asymmetric information, public goods, and monopoly (6)--in light of such a possibility. We consider ways the law might create market failures, and discuss the extent to which any such failure can be used as an instrument to disrupt markets for tax planning.

In Part V we contextualize the analysis, looking at practical applications of our theory of optimal governmental action. We first present two examples in which the government decided to allow particular tax planning methods; we consider the examples in light of our argument that some tax planning ought to be allowed. We next explore how the government might create failures in the market for tax planning activities, thereby reducing such activities. We proceed with a critical appraisal of our theoretical framework for feasibility and desirability. In this regard, there have been various developments and proposals along some of the lines we suggest, and we appraise the reactions thereto.

In a companion piece, we set forth a formal economic model to identify the optimal amount of tax planning the government should permit, and the features and effects of our hypothetical governmentally-created market failures. (7)

II. SETTING THE STAGE: OUR ASSUMPTIONS

We begin by noting the obvious: taxes are important to society, but people do not like to pay them. (8) Governments invest resources to collect tax revenue, and people expend resources to find ways to reduce their tax burden. Despite government's best efforts, it is unable to prevent people from developing and using methods to reduce their taxes: legislators are unable to write tax laws that do not have "loopholes."

Let us first consider how taxpayers would search for tax planning methods that would exploit loopholes. The greater their expected return, the more effort they would expend. Specifically, the more people believe they can save by finding a new tax planning method (or earn by selling the method to others), the harder they will search. (9)

What can we say about people's expectations of the return to be had? We begin by assuming that tax planning yields diminishing marginal returns. Effort spent on looking for tax planning methods is rewarded, but at a decreasing rate. (10) The benefit available from identifying a second tax planning method will generally be lower than the benefit from the first one. Thus, the amount of money that people can save increases in the number of available tax planning methods, but at a decreasing rate. The loopholes the methods exploit can overlap, such that there is some redundancy in the second loophole. Later searches thus may yield methods that would shelter income or gains already partially or wholly sheltered. Taxpayers may be able to carry forward tax losses that exceed a particular year's taxable income; however, all else being equal, the less tax they pay, the more likely they are to be scrutinized by the government, potentially resulting in the disallowance of some of those losses. Beyond a certain point, then, the use of any additional tax planning methods may expose the taxpayer to increased chances of detection. It therefore follows that the more tax planning methods currently are allowed, the less effort taxpayers will exert looking for new ones.

Given that taxpayers will behave in this manner, how should the government proceed in designing tax policy? First, we note that tax planning efforts are for the most part unproductive, socially wasteful activities. When people engage in tax planning to reduce their tax burdens, they are not creating new wealth for society. They are simply putting (or keeping) money in their own pockets that would have gone to somebody else (specifically, the government). In economic terms, tax planning is a form of rent seeking behavior. Thus, a benevolent government should attempt to design government policy to take into account both the revenue raised and the wasteful efforts that people will expend to avoid taxes. Hereinafter, we shall refer to taxpayers' search and tax avoidance costs as "dissipation costs" and to the social deadweight loss from government's losses in tax revenue as "tax revenue losses."

Second, we assume that the government is trying to minimize the sum of dissipation costs and tax revenue losses. We also assume that governments design tax policy with imperfect foresight but with rational expectations. Even though, in any given year, there may be newly-discovered techniques that the government did not anticipate (imperfect foresight), the government fully anticipates that some tax planning techniques will be devised and that some tax revenue will unavoidably be lost through the use of those techniques (rational expectations). The government anticipates losing some revenue to tax planning--it just does not know the exact method by which the revenue loss will occur. The amount the government anticipates losing on account of tax planning activities is included in the government's policy plan. In the short term, the government cannot simply change its tax regime to make up the revenue it loses to tax planning activities. Thus, the government's own cost-benefit calculations limit the amount of revenue it can feasibly raise under any current tax regime.

Finally, we assume that the marginal social cost of a dollar of lost tax revenue is increasing. That is, we assume that the social costs of lost revenue are increasing at an increasing rate. Under these assumptions, it is not cost-effective for the government to seek to eliminate tax planning activities entirely. At a certain point, the cost of the additional search effort that taxpayers will exert outweighs the additional revenues the government might obtain. It is therefore optimal for the government to allow some amount of tax planning.

III. TAX POLICY AND OPTIMAL GOVERNMENTAL ACTION: ENDOGENIZING "ILLEGITIMATE" TAX PLANNING

The above description of a basic model allows us to consider the following scenario. The government enacts a tax policy. Somebody immediately develops a tax planning method to reduce her tax burden. The government has to decide what to do with respect to the tax planning method. Should it close the loophole the method exploits, declaring the method illegitimate, or should it allow use of the method in the future? (11)