The economic effects of unemployment insurance.

• Author: Meyer, Bruce D.

Unemployment Insurance (UI) programs in the United States provide payments to 2 million unemployed workers in an average week at an annual cost of about $25 billion. These programs smooth the income of laid-off workers, but have many other effects as well. For example, they induce the unemployed to spend more time out of work, they encourage firms to lay off their workers temporarily in slack times, and they subsidize industries with frequent layoffs. In a series of papers, many of them jointly with Patricia M. Anderson, Lawrence F. Katz, or Dan T. Rosenbaum, I have examined these and other issues. This research quantifies many of the concepts that bear on the design of UI programs and tests some of the basic predictions of economic theory.

We use the detailed records on firms and workers that state agencies employ to administer their UI programs, including the earnings measures that determine a person's eligibility and level of benefits, and the amount and dates of UI benefit receipt. In many cases, these records also contain key firm characteristics, including the firm's tax rate. Unlike survey data, these administrative records accurately identify the incentives that a program provides to a given individual or firm.

This research on UI also makes use of the large differences across state programs and the frequent changes in these programs. A common methodological problem in analyses of social insurance programs such as UI is that an individual's eligibility and benefits usually are determined by the individual's work history and state of residence. Unfortunately, both of these determinants are likely to have their own influence on unemployment or wages, for example. Convincing estimates of the effects of UI require variation in program incentives that is not attributable to causes that might influence the outcomes under study. In these papers, I control carefully for a worker's earnings history. Several papers use a "natural experiment" approach, examining outcomes such as an individual's length of unemployment before and after changes in state UI laws.[1]

UI has effects on both workers and firms. The theoretical frameworks that have been used most often to analyze worker actions are search and labor supply models. In search models, an unemployed individual seeks out job offers and considers whether or not they meet his requirements. The existence of UI reduces the incentives for an individual to search hard and to take the first job offered. This reduced search effort implies that the length of unemployment is prolonged. However, the prolonged search may lead to a better new job than the worker would find otherwise. These models also suggest that unemployed individuals will intensify their search and become less choosy around the time that unemployment benefits run out. The other main theoretical approach, labor supply theory, assumes that workers select their length of time out of work to balance the income they receive from work and the alternative uses of their time. Similarly, this approach predicts a longer time out of work if UI is more generous, and an increased rate of job finding around the time that benefits run out.

In three papers, I focus on the effects of UI program characteristics on the time recipients spend...