Empirical studies in labour economics

Abstract

This dissertation addresses differences in characteristics of individuals and resulting inequalities in the labour market. This is a broad field which is why the chapters of this thesis cover quite diverse aspects of it: on a micro-level, I analyse differences of fixed-term contract workers as opposed to permanent contract workers in a descriptive analysis in chapter 2 before I turn to gender inequalities in contract duration evaluating a parental leave reform in Germany in chapter 3. On a macro-level, I analyse (co-authored with Patrick A. Puhani) differences in the labour mobility across Indian states and districts induced by labour market shocks in chapter 4. The first analysis addresses the question, which characteristics of employees are related to fixed-term employment. Fixed-term employment compared to permanent employment yields different working conditions, which are not always advantageous for employees, although an increased flexibility might be beneficial for employers. Applying machine learning techniques (LASSO, Elastic Net, and variations), I find that these models considerably increase the predictive power compared to benchmark models including control variables usually used in literature. The selection models make use of the rich information in the SOEP data and identify a variety of control variables that have been mostly neglected in literature so far. The applied variation of a LASSO regression (Double Selection LASSO) shows that this negligence leads to an omitted variable bias in the tested benchmark models and results let conclude that models should be more complex when analysing fixed-term employment to account for the broad variety of individual differences. I show that the results are robust across methods and reveal heterogeneity of the estimates by subgroups. Focussing on gender related differences in contract duration only a part of the differences can be explained by child-related leaves and the resulting lower labour market experience of women. My results for Germany indicate that it is enough to be a potential mother without any child-related career breaks to increase the probability of being employed on a fixed-term basis rather than permanently. To identify the causal effects, I apply a difference-in-differences approach using a change in the legal regulations about parental benefits in Germany in 2007 as natural experiment. I find that women of childbearing age are significantly more likely than men to be employed on a fixed-term basis after the reform. The effect is more pronounced for young women without children indicating an employer-sided discrimination of potential mothers. On a macro-level, an analysis of (labour) mobility across Indian states and districts covers geographical aspects of labour market differences. The chapter analyses how labour market shocks - measured by changes in non-employment rates, unemployment rates, and wages in fixed-effects regressions - affect regional migration. Comparing the results with those for the United States and the European Union, the most striking difference is that, in India, we do not find significant reactions to asymmetric non-employment shocks at the state level, only at the district level, whereas the estimates are statistically significant and of similar size for the state/NUTS-1 and district level in both the United States and Europe.