A central problem in raising public sector efficiency is effectively incentivizing public servants to exert effort and not resort to rent-seeking. While this issue is more salient in developing countries, it is to some degree a concern everywhere, and is particularly a concern in tax administration. While there is a substantial theoretical literature on incentive theory for bureaucrats, there are few carefully designed and evaluated empirical studies, particularly in the tax sector. This study seeks to test the impact of various wage and incentive schemes on the performance of tax officials, in collaboration with the Excise and Taxation Department in Punjab, Pakistan. The study will focus on evaluating a variety of wage and incentive schemes for tax inspectors ranging from efficiency wages, incentive mechanisms that tie wages to revenue collection, and incentive mechanisms that give rewards for accuracy and customer satisfaction in addition to revenue collected. We focus on urban property tax, both because this is an important yet relatively untapped revenue source and because independent measures can reliably assess true tax liability. The project will be implemented as a randomized experiment: individual tax units, known as circles, will be randomly assigned to one of three main schemes: an efficiency wage scheme, in which wages will be approximately doubled; the incentive scheme, in which an average of 30 percent of all tax revenues above a historical benchmark will be rebated to the tax officials as an honorarium; and an incentive plus scheme, where in addition to the revenue incentives, tax officials will be incentivized based on the accuracy of their tax assessments and customer satisfaction. These different schemes allow us to test not only whether incentives increase tax collections, but also whether by explicitly incentivizing tax collectors on multiple dimensions the government can increase collections while maintaining accuracy and customer satisfaction. To the best of our knowledge, this represents the first randomized, controlled trial on incentives for tax inspectors anywhere in the world. By measuring multiple outcomes we can potentially separate low effort from other channels, as well as disentangle the various mechanisms through which incentives can matter, such as adding new properties to the tax rolls, updating property valuations, issuing more tax bills, and greater effort at ensuring payment. While we focus on one developing country and specific interventions, the broader goal is to generate insights that are generalizable to other settings throughout the world.
