Tax Structure and Tax Compliance

Abstract

A model of individual tax compliance behavior, including evasion and avoidance, is developed and estimated. The model recognizes the importance of marginal income tax rates, payroll tax contributions and benefits, and the probabil- ity of detection and the penalty on unpaid taxes. Share equations for avoidance, evasion and reported income are estimated using individual-level data. The estimation results indicate that the tax base rises with higher benefits for payroll tax contributions and falls with higher marginal tax rates; the base also falls with more severe penalties and more certain detection of evasion as individuals substitute towards avoid- ance income. T HE methods by which individuals reduce their tax liabilities take a variety of legal and illegal forms, all of which are influenced at least in part by incentives created by the tax structure. These methods can be broadly classified as avoid- ance and evasion. Tax avoidance is any legal activity that lowers taxes, such as worker substitu- tion between wage and nonwage compensation. Tax evasion is the reduction in tax liabilities by illegal means, such as underreporting income on tax returns. Despite extensive-but separate literatures on avoidance and evasion, we know very little about the tax base response to changes in tax structure. It is the purpose of this paper to examine the role that the tax structure plays in compliance. There are several reasons for the persistence of the compliance puzzle. Most prominent is the absence of detailed individual data that would allow a full empirical specification of all factors affecting compliance. This difficulty is most evi- dent when searching for individual data on the evasion-compliance decision. Even when avail- able, data have never allowed in the same work the construction of both tax and audit variables.1 Clearly, data availability is even more problem- atic when looking at the choice among evasion, avoidance, and reported income. In addition, pre- vious work has not looked at the avoidance- evasion-compliance decision as a joint process, even though these decisions are made simultane- ously.2 Further, both strands of the literature have ignored another factor that may affect the compliance decision: the benefits that accrue from participation in payroll programs. If benefits are tied only to taxable income, then their presence gives individuals an incentive to pay taxes. In short, there has been no empirical work that analyzes the effects of tax rates, probabilities, penalties, and payroll benefits on avoidance and evasion choices of individuals. In this paper we provide such an analysis. We first develop a theory of individual choice among the three types of compensation. We then esti- mate the resulting share demand equations using a unique data set, which has detailed information on the compensation paid to roughly one-quarter of the labor force in Jamaica in 1983. From these data we are able to derive measures of reported taxable income, evasion income, and avoidance income for individual workers in the formal sec- tor. We are also able to construct measures of the marginal income tax rate, marginal payroll taxes and benefits, the probability of detection, and the penalty on evasion for individual workers. We are therefore able to estimate for the first time the responses of workers to the full range of tax structure parameters. Section I presents the theoretical model of worker compensation choice and the empirical specification of the model. Section II discusses the Jamaican tax system. Data and variable con- struction are discussed in section III. Estimation results are presented in section IV. The final section summarizes the main results.