Essays in Labor Economics

This dissertation consists of three essays on labor and information economics. The first presents an evaluation of a Turkish payroll tax subsidy which supported formal employment of women and young men. The subsidy's impact is identified using data from the Turkish Household Labor Force Survey. I follow a difference in difference approach comparing the outcomes of young men to older men, and a difference in difference in difference approach for the impact on female labor force participation, employment, and formal employment, as distinct from female- specific trends. Results show an increase in formal male employment, but no impact on wages or women.;The second essay considers communication games in which speakers may present hard evidence to prove their claims. Hard evidence is modeled as state-contingent restrictions on the messages speakers can send. Following Lipman and Seppi (1995), evidence is characterized by partial provability: some claims of the world can be proved, some cannot. In contrast to previous results, this essay finds that with no knowledge of the message space, robust inference is impossible, although an algorithm exists by which the listener can arrive at a non-degenerate inference whose support includes the true state. The essay also provides sufficient conditions on the listeners knowledge for robust inference.;The final essay studies economic outcomes for Spanish-speaking immigrants from less common sending nations who live in the same area as a larger immigrant population. I hypothesize that the identification of enclave effects in this community of fringe immigrants will depend on national fractionalization of that population. Using the Herfindahl Index as a measure of national concentration of the minority immigrant labor market, data from the American Community Survey show that minority immigrant population size matters only when the population is highly concentrated. This result cannot be explained by endogenous residential selection, and is interpreted within a network quality framework. All immigrant population effects are weakly negative, while the native-born Spanish-speaking population has a positive effect on immigrants incomes. The identification of this positive impact is weakened by residential sorting on unobservable skill and motivation.