Nber Working Paper Series Minimally Altruistic Wages and Unemployment in a Matching Model

This paper presents a model in which firms recruit both unemployed and employed workers by posting vacancies. Firms act monopsonistically and set wages to retain their existing workers as well as to attract new ones. The model differs from Burdett and Mortensen (1998) in that its assumptions ensure that there is an equilibrium where all firms pay the same wage. The paper analyzes the response of this wage to exogenous changes in the marginal revenue product of labor. The paper finds parameters for which the response of wages is modest relative to the response of employment, as appears to be the case in U.S. data and shows that the insistence by workers that firms act with a minimal level of altruism can be a source of dampened wage responses. The paper also considers a setting where this minimal level of altruism is subject to fluctuations and shows that, for certain parameters, the model can explain both the standard deviations of employment and wages and the correlation between these two series over time. Julio J. Rotemberg Graduate School of Business Harvard University, Morgan Hall Soldiers Field Boston, MA 02163 and NBER [email protected] Job vacancies drop considerably in recessions, suggesting that recruitment costs are quite procyclical. As emphasized by Rotemberg (2007b) in the context of the Mortensen-Pissarides (1994) model, the procyclicality of recruitment costs implies that real wages should be much more procyclical than they actually are. The model makes this prediction because the ease with which firms can recruit workers in recessions strengthens firms’ bargaining position so that Nash bargaining between firms and workers leads to substantially lower wages. This paper departs from the Mortensen and Pissarides (1994) model in several respects. First, instead of assuming that wages are determined by a bargaining process, it assumes that firms set wages unilaterally. As in Burdett and Mortensen (1998), firms act somewhat monopsonistically in the model developed here. They realize, in particular, that reductions in their wages lead only some of their to employees to depart. In efficiency wage terms, the model of this paper is thus “turnover” based. The result of this wage-setting assumption is that, unlike in the bargaining case, employee’s reservation wages no longer matter for wage determination. Instead, wages are greatly affected by the way employees weigh wage and nonwage aspects of a job when deciding among job opportunities. Nonwage aspects of jobs, and their role in creating job satisfaction, are stressed in the managerial literature on “voluntary turnover,” and they also play a role in the model of Nagypál (2005) that tries to explain the magnitude of this turnover. This paper also extends the canonical Mortensen and Pissarides (1994) model by considering not only a specification in which firms act selfishly but also one in which firms act somewhat altruistically. This specification is motivated by Rotemberg (2007a), which explains ultimatum and dictator experiments with a model where people react with anger For a discussion of efficiency-wage models in general and their division into models based on reducing shirking, reducing turnover, improving selection, and increasing effort as a result of fairness considerations, see Katz (1986). For a discussion of the managerial literature, see Price (2001). Using exit interviews from a firm that experienced a great deal of turnover, Sutherland (2002) found that most people who left this firm for another job claimed that their reason for doing so was either higher wages or the opportunity to earn more overtime. Still, even in this case, 18 of the 48 people who left for another job reported doing so for other reasons. One difficulty with studying the sources of quits is that, as noted by Hinrich (1975), exit interviews do not provide the same answers as questionnaires on people’s intention to quit.