Title: Labor Market Dynamics and Policy Evaluation: Empirical Evidence from Micro Data
Abstract: This thesis consists of three separate studies which investigate different aspects of labor market flexibility. First, by looking at changes in the stability of employment relationships in the 1980’s and the 1990’s in West Germany it is investigated whether employment relationships became more flexible. A descriptive analysis of the evolution of elapsed job duration of those currently in work is conducted based on data of the German Socio-Economic Panel (GSOEP) 1984-1999. In addition, the evolution of the completed job duration is analyzed. A competing risk Cox Proportional Hazard Rate model is estimated. It is distinguished between different reasons for ending a job and different exit states. This way, not only a secular trend can be identified but also potential reasons for the observed changes in job stability can be empirically investigated. The results show that job stability of men declined. Part of this can be attributed to an increase in layoffs and part to an increase in transitions to unemployment. However, these two developments are not significantly related to each other. For women no significant change in job stability is found. Some evidence is presented that downsizing of large firms might be responsible for part of the decline in job stability for men, whereas no significant impact of skill-biased technological change on job stability or evidence for a general weakening of the attachment between firms and their employees is found. Chapter 3 and 4 are evaluation studies for two different Active Labor Market programs, namely for training programs and job creation schemes. It is investigated whether these programs increase the flexibility on the labor market by helping the participants finding a job and/or remaining employed. Chapter 3 evaluates the effects of training programs in East Germany on the employment chances of the actual participants for the time period 1990-1999. A novel evaluation approach is developed which builds upon a dynamic employment model and thus takes account of the dynamics of the employment process. This new estimator is called conditional difference-in-differences in hazard rate estimator and is an extension of the conditional difference-in-differences estimator, which is a popular evaluation method usually applied to employment rates or earnings. The conditional difference-in-differences in hazard rate estimator assesses the treatment effects separately for the different transition rates, here the reemployment rate and the rate to remain employed. Additionally, a sensitivity analysis is conducted in order to compare the results of the conditional difference-in-differences in hazard rate estimator with a further way to model state dependency. Especially, in East Germany unemployed often do not participate only once in a program of Active Labor Market Policy but several times. Here the effect of training as the first participation in a program of Active Labor Market Policy is estimated. In addition, it is differentiated between different treatment sequences where the incremental and combined effect of multiple participations is evaluated. The results are estimated on the basis of survey data of the Labor Market Monitor of the state of Sachsen-Anhalt. With regard to the transition rates it is found that the employment effects of participation in a first training program are mostly insignificant but that there are some significantly positive effects for selected starting dates of training programs. In contrast, with respect to unconditional employment rates the results show significantly negative effects. Combined sequences of two programs with a first training program are not successful with respect to the transition rates whereas the incremental effect of the second treatment appears to have slightly positive effects on the probability to remain employed. Chapter 4 evaluates job creation schemes using the same data set as in chapter 3. Here not only the treatment-on-the-treated but also population average treatment effects are estimated. In order to account for employment dynamics, the outcomes of interest consist of changes in the reemployment rate and the rate to remain employed. The conditional difference-in-differences in hazard rate method is used for correcting for selection effects. In order to be able to estimate population average treatment effects, which vary with treatment time, a solution is proposed for setting hypothetical starting dates of the nonparticipants. Job creation schemes show zero to positive effects on the reemployment rate and often significantly positive effects on the rate to remain employed. The effects on the actual participants and the average population resemble each other and improve for programs which start later.
Publication Year: 2004
Publication Date: 2004-01-01
Language: en
Type: dissertation
Access and Citation
AI Researcher Chatbot
Get quick answers to your questions about the article from our AI researcher chatbot