Employment dynamics and heterogeneous labor markets [Elektronische Ressource] : an empirical analysis with linked employer-employee and company data / Susanne Steffes

Employment Dynamics and
Heterogeneous Labor Markets: An
Empirical Analysis with Linked
Employer-Employee and Company Data
Inaugural-Dissertation zur Erlangung der Doktorwurde˜ der Wirtschafts- und
Verhaltenswissenschaftlichen Fakult˜at der Albert-Ludwigs-Universit˜at Freiburg
Susanne Stefies
geboren in Darmstadt
vorgelegt am: 8. Juli 2010Dekan: Prof. Dr. Dieter Tscheulin
Erstgutachter: Prof. Bernd Fitzenberger, Ph.D.
Zweitgutachter: Prof. Dr. Dr. h. c. Bernd Schauenberg
Promotionsbeschluss: 13. Oktober 2010In memory of my MomAcknowledgements
First of all, I want to thank my supervisor Bernd Fitzenberger for his great
support. Our numerous and intensive discussions gave me motivation and his
comments had a huge impact on my work. His lectures and seminars, which I
joined during my studies at the University of Mannheim, already sparked my
interests in empirical labor market research. After I had written my diploma
thesis with him, I was happy for the possibility to start my dissertation under
hissupervision. IamalsothankfultomysecondsupervisorBerndSchauenberg.
I want to thank Bernhard Boockmann for his cooperation and support. Ad-
ditionally,IamthankfultoAnthonyStrittmatter. Thetimeswithmycolleagues
at the ZEW were always productive and fruitful. Especially, Melanie Arntz and
Nicole Gurtzgen˜ gave helpful comments to my papers.
I wrote three chapters of this dissertation with data of the Research Data
Centre (FDZ) of the Federal Employment Agency at the Institute for Employ-
ment Research. I am grateful to Stefan Bender, Peter Jacobebbinghaus and
Alexandra Schmucker for their support with the data. Financial support by the
Deutsche Forschungsgemeinschaft (FR 715/5-2) is gratefully acknowledged.
My parents, parents and sisters in law as well as my brother and sister gave
me a lot of energy and provided a lot of their free time (looking after my child).
WithoutthatIcouldnothavewrittenthisdissertation. Iwanttothankthemfor
this. Most of my thanks go to my husband Christoph and my son Maximilian.
They had some hard times with me, but they always pushed me up again.Contents
0 Introduction iv
0.1 General Introduction . . . . . . . . . . . . . . . . . . . . . . . . . iv
0.2 Own Contribution . . . . . . . . . . . . . . . . . . . . . . . . . . viii
0.3 Summary of the Chapters . . . . . . . . . . . . . . . . . . . . . . x
0.3.1 Chapter 1 . . . . . . . . . . . . . . . . . . . . . . . . . . . x
0.3.2 Chapter 2 . . . . . . . . . . . . . . . . . . . . . . . . . . . xi
0.3.3 Chapter 3 . . . . . . . . . . . . . . . . . . . . . . . . . . . xii
0.3.4 Chapter 4 . . . . . . . . . . . . . . . . . . . . . . . . . . . xiii
1 Workers, Firms or Institutions 1
1.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
1.2 Theoretical Background . . . . . . . . . . . . . . . . . . . . . . . 5
1.3 The Role of Institutions for Job Durations . . . . . . . . . . . . . 6
1.4 Data and Descriptives . . . . . . . . . . . . . . . . . . . . . . . . 8
1.4.1 Data on Job Durations. . . . . . . . . . . . . . . . . . . . 8
1.4.2 Sample Deflnition and Descriptive Statistics . . . . . . . . 10
1.5 Estimation Technique and Independent Variables . . . . . . . . . 12
1.6 Empirical Results . . . . . . . . . . . . . . . . . . . . . . . . . . . 14
1.6.1 Estimates for Individual and Job-Speciflc Characteristics 14
1.6.2 for Firm and Institutional . . . 15
1.6.3 The Efiects for Difierent Groups of Workers . . . . . . . . 17
1.6.4 Competing Risks . . . . . . . . . . . . . . . . . . . . . . . 18
1.7 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19
1.8 Tables . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21
1.9 Figures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27
2 Seniority and Job Stability 29
2.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31
2.2 Hypotheses on the Determinants of Employment Durations . . . 33
2.2.1 Job Matching . . . . . . . . . . . . . . . . . . . . . . . . . 33
2.2.2 Speciflc Human Capital . . . . . . . . . . . . . . . . . . . 35
iCONTENTS
2.2.3 Search Theory and the Role of Wages . . . . . . . . . . . 36
2.2.4 The Role of Institutions . . . . . . . . . . . . . . . . . . . 37
2.3 Data and Descriptive Statistics . . . . . . . . . . . . . . . . . . . 38
2.4 Estimation Technique . . . . . . . . . . . . . . . . . . . . . . . . 41
2.5 Results . . . . . . . . . . . . . . . . . . . . . . . . . . 44
2.5.1 The Impact of Firm Characteristics . . . . . . . . . . . . 44
2.5.2 The impact of individual characteristics . . . . . . . . . . 46
2.5.3 The Impact of Entry Wages . . . . . . . . . . . . . . . . . 48
2.6 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 48
2.7 Figures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51
2.8 Tables . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 59
3 Employment Dynamics and Cyclical Shocks 63
3.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 65
3.2 Theoretical Considerations and Hypotheses . . . . . . . . . . . . 69
3.2.1 The Basic Theoretical Model . . . . . . . . . . . . . . . . 69
3.2.2 The Impact of Worker Heterogeneity . . . . . . . . . . . . 73
3.2.3 The Impact of Firmy . . . . . . . . . . . . . 76
3.3 Data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 77
3.3.1 Deflnition of Dependent Variables . . . . . . . . . . . . . 80
3.4 Empirical model . . . . . . . . . . . . . . . . . . . . . . . . . . . 82
3.4.1 Econometric approach . . . . . . . . . . . . . . . . . . . . 82
3.4.2 Explanatory Variables and Speciflcations . . . . . . . . . 83
3.5 Results. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 89
3.5.1 Developments of Job and Worker Flow Rates Over Time. 89
3.5.2 The Impact of Shocks on Job Flows . . . . . . . . . . . . 91
3.5.3 The Impact of Shocks on Worker Flows . . . . . . . . . . 93
3.5.4 The Impact of Firm Characteristics on Job and Worker
Flows . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 96
3.6 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 99
3.7 Tables . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 102
3.8 Figures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 110
ii3.9 Appendix . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 114
3.9.1 Value Functions . . . . . . . . . . . . . . . . . . . . . . . 114
3.9.2 Tables . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 115
4 Employment Behavior After Maternity Leave 119
4.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 121
4.2 Theoretical Considerations and Hypotheses . . . . . . . . . . . . 124
4.3 Institutional Settings . . . . . . . . . . . . . . . . . . . . . . . . . 127
4.4 Data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 129
4.4.1 General Information . . . . . . . . . . . . . . . . . . . . . 129
4.4.2 Deflnition of Maternity Leave Spells . . . . . . . . . . . . 130
4.4.3 of Covariates . . . . . . . . . . . . . . . . . . . 131
4.5 Econometric Approach . . . . . . . . . . . . . . . . . . . . . . . . 134
4.5.1 Competing Risks Model . . . . . . . . . . . . . . . . . . . 135
4.5.2 Cumulative Incidence . . . . . . . . . . . . . . . . . . . . 136
4.6 Empirical Results . . . . . . . . . . . . . . . . . . . . . . . . . . . 137
4.6.1 Transition Probabilities from Maternity Leave. . . . . . . 137
4.6.2 Employment Stability After the Return to the Company . 143
4.6.3 Pre-birth Career and Fertility Decision . . . . . . . . . . . 146
4.7 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 147
4.8 Tables . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 149
4.9 Figures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 155
4.10 Appendix . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 161
4.10.1 Data preparation . . . . . . . . . . . . . . . . . . . . . . . 161
4.10.2 Tables . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 166
4.10.3 Figures . . . . . . . . . . . . . . . . . . . . . . . . . . . . 170
5 References 172
Bibliography of the Chapters 185
iii0 Introduction
0.1 General Introduction
In the last decades, the German unemployment rate was high and persistent,
compared with other industrial countries like the US or the UK (see flgure
1). Economists and politicians argue that low exibility in the labor market is
one reason for this persistence. In fact, compared with the US or the United
Kingdom, the German labor market is highly regulated (see OECD, 2003).
Restrictive job protection laws, as well as powerful unions and works councils,
increase costs of employment reduction and wage rigidities, which both could
result in low turnover rates and respectively high unemployment rates.
Figure 1: Unemployment rates in Germany, UK and US
Source: OECD Employment Outlook 2009, 2007, 2004, 2002. Note: Own illustrations.
As a consequence, in the early 2000s Germany experienced a remarkable
change in labor market policies resulting in - besides other things - more exi-
1bility. However, during this period, there was only few empirical evidence on
the exibility in the German labor market and thus, it was hardly possible, to
1For instance the possibilities for flxed-te