Employer preferences and the determination of hours of work

This dissertation considers the hypothesis that hours of work are determined jointly by both employers and employees. I refer to it as the employer interest hypothesis since it assumes that firms have an interest in employee hours even conditional on the wage. The hypothesis stands in contrast to the neoclassical labor supply theory in which hours are determined solely by employees given a wage offer by firms. I follow two approaches to gather evidence on this issue. The first involves comparing estimates of labor supply parameters obtained under each hypothesis. This procedure is used in Chapters One and Two to test the predictions of the employer interest model regarding the differences in these estimates. In Chapter One, I consider the prediction that the intertemporal labor supply elasticity is reflected in an individual's responses of hours to wage changes between jobs, as employers may constrain hours choices within jobs. In Chapter Two, I propose using as a demand instrument a proxy measure for a factor affecting the employer's demand for hours of work that can be constructed from data solely on employees. This is the average PMSA commuting time, which is hypothesized to measure the costs of organizing and operating a typical business in the PMSA according to a hedonic pricing model of PMSA attributes. Results show that the pure supply responses, purged of the confounding demand variation, are positive and often significantly larger than the responses that do not take labor demand into account. This finding is consistent with the employer interest hypothesis, but is difficult to reconcile with the neoclassical labor supply model. The second method evaluates the claim that fixed costs generate a reason for why employers care about the hours of their employees, and is investigated in Chapter Three. I sample measures of fixed employment costs directly to test the predictions of a model of labor demand regarding the effects of fixed costs on firms' desired numbers of employees and employee hours. Results show that, to the extent that these costs are determined by the firm's industry and capital intensity, they tend to significantly reduce turnover and vacancies and to increase employee hours as predicted. Overall, results of both strategies support the importance of employer interest in determining hours of work.