Lab Labor: What Can Labor Economists Learn from the Lab?

This chapter surveys the contributions of laboratory experiments to labor economics. We begin with a discussion of methodological issues: why (and when) is a lab experiment the best approach; how do laboratory experiments compare to field experiments; and what are the main design issues? We then summarize the substantive contributions of laboratory experiments to our understanding of principalagent interactions, social preferences, union-firm bargaining, arbitration, gender differentials, discrimination, job search, and labor markets more generally. Gary Charness Department of Economics University of California, Santa Barbara 2127 North Hall Santa Barbara, CA 93106 [email protected] Peter J. Kuhn Department of Economics University of California, Santa Barbara 2127 North Hall Santa Barbara, CA 93106 and NBER [email protected] The economics literature has witnessed an explosion of laboratory experiments in the past 20 years. Many of these experiments have focused on topics that are central to the field of labor economics, including how workers respond to various forms of compensation, and the economics of discrimination, arbitration, bargaining, and matching. In this chapter we survey the contributions of laboratory experiments to our understanding of these questions. We begin our review with a discussion of methodological issues: First, we pose the general question of why (and more importantly when) a labor economist might want to conduct a laboratory experiment: What types of questions, if any, are laboratory experiments best suited to answer? How do laboratory experiments compare to field experiments? Next, once one has decided to conduct a laboratory experiment, how should it be designed? Here we review the main methodological decisions an experimenter typically needs to make, and the advantages and disadvantages of the various choices. The second half of our review turns its attention to the substantive issues in labor economics that have been addressed using laboratory experiments. While these are wideranging, we focus our review on the set of issues that have generated probably the largest volume of experimental papers in labor economics: the effects of compensation policies on the supply of effort by workers. We do this in two parts. The first uses ‘traditional’ principal-agent theory as a theoretical lens to derive predictions regarding the effects of incentives, and tests these predictions in the lab. Many are confirmed; at the same time a number of robust ‘anomalies’, such as apparent gift exchange, also appear. The second part focuses specifically on the use of experiments and the development of new theoretical models of social preferences to understand these anomalies. Finally, we also provide brief guides to the laboratory literature on a number of other labor-related topics, including union-firm bargaining, arbitration, gender differentials, discrimination, and job search. I. Why Laboratory Experiments? Why should labor economists care about laboratory experiments? After all, there are plenty of field data available for empirical tests. In addition, there have been a number of objections to lab experiments concerning issues such as a lack of realism (external validity), demand effects, and selection effects. Indeed, the laboratory is an artificial environment. On the other hand, lab experiments have some important advantages over other approaches; we begin this section with a discussion of these advantages, then move on to critiques and responses.