ESSAYS IN EXPERIMENTAL ECONOMICS WITH IMPLICATIONS FOR ECONOMIC DEVELOPMENT

Abstract

In an exploration of the joint concerns of economic development, namely efficiency and equality, I employ experimental methods to consider several issues regarding entrepreneurship and regulation with particular applications in developing countries.

Entrepreneurship programs in developing countries may not take hold in rural populations if people there tend to shy away from competitive and uncertain economic opportunities, thus contributing to the systematic underdevelopment of rural areas. In a field experiment conducted among potential entrepreneurs in rural and urban Ghana, we found that rural subjects were 20 percent less likely than their urban counterparts to select an all-or-nothing tournament compensation scheme over a piece rate wage to per- form a simple matching task. The difference between the rural and urban tournament choice was driven by subjects who believed their own performance was the best within their group; urban subjects were twice as likely as their rural counterparts to believe that they had scored in first place and were thus more likely to select the tournament compensation.

To examine behavior in a tax setting, we develop a simple tax evasion model as a signaling game between a taxpayer and an auditor that includes a non-strategic, always compliant taxpayer. In addition to the taxpayer's income report to the auditor, he has the option to send a costly message, a donation to charity that may serve as an indirect signal to the auditor of the taxpayer's ethical type. In the case where the taxpayer has misreported his income and is audited, he must pay unpaid taxes and a penalty. We establish a Perfect Bayesian equilibrium where taxpayers will use the charitable donation to signal honesty, thereby reducing the probability of audit. Auditors will optimally audit reports without charity donations more frequently than those with donations. To test our theoretical predictions, we use a two-sided signaling experiment where the taxpayer voluntarily reports his income to determine his tax liability and can make an observable and verifiable charity donation. Our aggregate experimental results indicate players employ mixed strategies in line with theoretical predictions.