Friday, February 3, 2012
Governments rely largely on tax revenues. Traditional economic models of tax compliance typically assume that it is in the government’s best interest to maximize the amount of tax due that it actually collects. In these models, enforcement is costly and imperfect, and the resulting “probabilistic” enforcement scheme is assumed to decrease social welfare by reducing overall tax revenues. More recent models suggest that the negative effects of imperfect enforcement may be less pernicious than previously believed. Yet, like the early models, the newer models often assume that perfect enforcement remains the ideal, and turn upon taxpayer risk-aversion or apply only in narrow circumstances. We show that even if perfect enforcement were costless, it would not always be socially optimal. Specifically, when uniform tax laws are suboptimal, and the legislative or regulatory costs of differentiated lawmaking are high, the enforcement agency can create de facto changes in the substantive law by raising or lowering (i.e., “measuring”) its overall level of enforcement in specific contexts. Importantly, by fostering a suitable level of noncompliance, measured enforcement can increase overall social welfare by reducing deadweight losses, while maintaining or increasing tax revenues in a wide variety of circumstances. This result does not depend on the risk profile of the taxpayer. Moreover, the benefits from measured enforcement cannot be easily replicated by traditional modifications to the tax laws, such as adjustments to the rate schedule or base of a given tax. As such, measured tax enforcement potentially offers substantial benefits that cannot be readily obtained otherwise.