Creating failures in the market for tax planning.

Our model of governmental action allows us to endogenize the distinction between "legitimate" and "illegitimate" tax planning activities. Although all tax planning activities are undesirable for the government, the government chooses to target only some of them as illegitimate. Notwithstanding our assumption that social welfare positively depends on the government's success in collecting tax revenue, our analysis leads to the (presumably controversial) conclusion that, all else being equal, social welfare may be increased if the government allows some forms of tax planning. People will look for additional loopholes regardless of whether the existing loopholes are closed, but, all else equal, will look harder the smaller the possible tax savings are from existing loopholes. Thus, the amount of wasteful search efforts is directly affected by the government's decision.

A benevolent government therefore faces a difficult tradeoff. By declaring a given tax strategy "illegitimate," the government boosts its tax revenue, but induces taxpayers to expend resources in the search for additional tax strategies. The government should also seek to reduce taxpayers' search, not only because it is through search that taxpayers find ways to minimize their taxes, but also because the ensuing dissipation costs represent social welfare losses. Our analysis suggests that, even assuming that the government could identify and close all loopholes, it should not do so. A benevolent government optimally should use soft intervention, restricting the scope of existing loopholes to varying degrees, without closing them entirely, anticipating and minimizing the dissipation that is likely to ensue from the arms-race between taxpayers and tax authorities.

Further, since the government seeks to maximize the total revenue it collects net of collection and dissipation costs, it wants to reduce both taxpayers' use of tax planning methods and their sale of such methods to others, who will also use the methods to minimize their taxes. The appropriate tax policy should consider that the taxpayers' interest in searching for new tax planning methods is not only driven by their desire to minimize their tax burden: under some circumstances, taxpayers' search may also be motivated by their interest to gain from the sale of tax planning methods to other taxpayers. (12)

IV. MARKET FAILURES AS TAX POLICY INSTRUMENTS

Dissipation deadweight losses increase in the amount of effort that taxpayers expend to reduce their tax burden (search). Search efforts increase in the expected private benefits that accrue to the finder of a new tax planning strategy and in the potential profits that could be made from selling the tax strategy to other taxpayers. Lost revenue costs are increasing in the number of new tax planning strategies that are identified (discovery) and in the number of people that use them (dissemination). In this part, we introduce basic concepts from price theory to consider the possible use of governmental intervention in reducing the search, discovery, and dissemination of tax planning strategies.

The fundamental idea is as follows: since the incentives for search, supply, and demand in the market can be negatively affected by market failures, the legal system can reduce the search, supply, and demand for tax planning strategies though the creation of governmentally-created market failures. In the presence of market failures, the market forces that normally allocate resources and create production incentives in our economy are defective or absent. Economists have identified four general categories of market failure, including monopoly, public goods, asymmetric information, and externalities. In such cases, markets may "fail" in the sense that they cannot ensure that the good is produced and consumed in the amounts in which it would have been produced and consumed had the market not failed.

While governments are generally involved in making markets run more smoothly, in this case the government prefers the market to be less efficient. How can the government introduce market frictions?

A. Creating Monopolies: Patenting Tax Planning Strategies

Monopoly is a standard case of market failure. A monopoly exists when there is only one supplier of a good in an industry and the good has no close substitutes. Patents and copyrights (13) are common examples of legally-created monopolies. Patents and copyrights give exclusive intellectual property rights to the discoverer or creator, preventing other individuals from competing with the patent or copyright holder in the exploitation of their intellectual property rights. According to economic theory, monopolists are price-makers, not price-takers. A monopolist can set prices above marginal cost, which places a wedge between the consumer's willingness to pay for the good and the producer's cost.

An increase in price implies that fewer goods will be sold. In the general case, this is undesirable because it results in lost social surplus (deadweight loss). Because monopoly firms reduce output below the socially efficient, competitive level, legislatures often pass laws to prohibit or regulate them. In the special case under consideration, if tax planning strategies can be patented, thereby granting exclusive rights to the discoverer, the resulting reduction in the quantity of available tax planning strategies may instead be desirable. (14) A social benefit, rather than a social loss, would result from monopoly underproduction. It follows that, unlike in the typical case, a monopoly over tax planning strategies should be promoted and protected, rather than prohibited and regulated.

The idea that the creation of a monopoly can be used as a policy instrument in the context of markets for tax strategies should be qualified at this point. Even though patents may result in an ex post restriction of supply, they may exacerbate ex ante search incentives. (15) If a person (or entity) obtains a patent on a tax planning method that he created, he can license the patent to others. His ability to earn fees from such a license should increase his incentive to search in the first instance; dissipation costs should therefore increase. The effect of introducing tax patents on tax revenue losses will depend on the relative magnitude of the two effects.

Price theory allows us to provide an additional qualitative assessment of the two effects under consideration. For example, if a monopolist can charge different prices to different consumers (price discrimination), then his total output will approach that of a perfectly competitive market. In this scenario, the property protection effect (i.e., the increase in search and discovery) will dominate the monopoly effect (i.e., the monopolist's restriction of output). In other words, if tax patents are available and tax patent holders or licensees can perfectly discriminate, the outcome for the government is the worst one possible. The market supply is efficient: the tax planning method becomes fully disseminated throughout the population, which maximizes lost revenue. Further, monopoly profits are maximized, meaning that the incentive to search for new tax planning strategies is at its peak, and dissipation costs will also be maximized.

The case in which a tax patent holder cannot price discriminate is somewhat more promising. In this case, the knowledge of the tax planning strategy will not fully disseminate, so lost revenue costs will not be as high. (16) Similarly, the value of having the patent will be lower because a non-discriminating monopolist is unable to capture the entire consumer surplus. Dissipation costs will therefore be lower than they would have been in the case where the monopolist could price discriminate.

Will tax patent holders be able to price discriminate? People who develop tax planning methods may very well be able to compute how much a buyer would pay. The patent effectively prevents the buyer from reselling and it is in both the developer (patent holder) and the buyer's interest to not have information about the transactions become too widely known. While the extent to which tax patent holders will be able to price discriminate is an empirical question, theory suggests that considerable price discrimination may be possible.

Which is better: to allow patenting of tax planning methods or not to do so? The answer depends on what would happen if patents are not allowed. Consider an environment in which there are many taxpayers, each of whom is able to search for new tax planning methods. Let us assume for simplicity that there are no property rights, whether through patents or by other means. Comparing this case to the one where there are patents but patent holders cannot perfectly price discriminate, we see that the patent case provides greater incentive to search (patent holders earn profits as well as reducing their own taxes), but the effect on lost revenue is ambiguous. Without property rights, the people who reduce their taxes are the ones who discover a tax planning method. With a patent, the people who reduce their taxes are those who are willing to pay the monopolist's price. We cannot say which effect is greater. So, if dissipation costs are not high and the monopolist is particularly inefficient, such that there is a large deadweight loss, then the government might do better if patents are allowed than if people cannot have property rights over their tax planning methods.