Essays on labor market dynamics

Abstract: This thesis consists of five essays on labor market dynamics.Essay I examines job tenure in Sweden over the period from the late 1960s and up to the end of the century. We investigate job tenure over time and how it varies across different labor force categories. On average, men have stayed with the same employer for about ten years and women for seven and a half years. While this average is rather unchanged for men during the period, it has increased for women. We find that the average actual tenure for full-time workers has remained stable over time, but has increased for part-time workers. Information on actual job tenure is used to derive the distribution of eventual tenure, i.e., the sum of actual tenure and projected remaining time in the same firm. Some aspects of job tenure in 1991 are also compared to other OECD countries. Essay II examines the determinants of quits, layoffs and job duration by means of survival analysis applied to Swedish household survey data from the mid-1980s and mid-1990s. We find marked age and tenure effects, especially during the 1980s. The estimated models are used to characterize waiting times until a worker quits or is laid off. We find a decrease in the incidence of overall job separations between the two observation periods. Moreover, when analyzing separations by reason, we find evidence of a shift from quits towards involuntary job loss during the early 1990s.Essay III (with Kenneth Carling and Bertil Holmlund) investigates the effects of unemployment insurance on the duration of unemployment spells. In June 1995, the Swedish parliament decided to cut the replacement rate in unemployment insurance from 80 percent to 75 percent, a change that took effect on January 1, 1996. This paper examines how this change affected job finding rates among unemployed insured individuals. To identify the effect of the policy we exploit a quasi-experimental feature of the benefit cut: only a fraction of the unemployed was affected by the reduction in replacement rates. We compare the evolution of job finding rates before and after the reform among those affected and those not affected. Our estimates suggest that the reform caused an increase in the transition rate of roughly 10 percent. There is also evidence of anticipatory behavior among the unemployed; the effects of the reform seem to operate several months before its actual implementation in January 1996.Essay IV (with Bertil Holmlund) presents an aggregate "flow portrait" of the Swedish labor market by using time series data on labor force transitions between employment, unemployment and out-of-the-labor force. We are particularly concerned with the cyclical and dynamic characteristics of the flows and the implied dynamics of the stocks. Transition rate models are estimated and used to simulate the labor market responses to an adverse macroeconomic shock. We find, inter alia, that an adverse shock triggers an increase in the level of inflows to unemployment that is followed by a rise in the level of outflows, i.e., both inflows and outflows are markedly countercyclical. The rise in nonparticipation observed during the 1990s is partly a cyclical phenomenon but there is in addition a trend rise in exit rates to nonparticipation from both employment and unemployment. Analogously, the substantial increase in temporary (fixed-term) employment observed over the decade is in part a result of the recession but there also seem to be more structural forces involved that may survive under favorable labor market conditions. Essay V (with Fredrik Andersson) studies the determinants of plant closure in Swedish Manufacturing using linked employer-employee data. From our theoretical framework we derive and empirically test hypothesis regarding the linkages between the probability of plant failure and: 1) industry specific characteristics of production and product demand; 2) local labor market conditions and 3) plant-specific sources of heterogeneity, including the importance of insider mechanisms in wage determination, plant specific human capital, selection mechanisms and technology vintage effects. Our results suggest that all these factors are important determinants of plant closures.