Structural estimation of search equilibrium models with wage posting [Elektronische Ressource] / by Andrey Launov

Andrey Launov

Sujets

• Management

Informations

• Publié par: ludwig-maximilians-universitat_munchen
• Publié le: 01 janvier 2006
• Nombre de lectures: 25
• Langue: English
• Poids de l'ouvrage: 1 Mo

Extrait

Structural Estimation of Search
Equilibrium Models with Wage Posting
by
Andrey Launov
Submitted to the Department of Economics
in partial fulflllment of the requirements for the degree of
Doctor oeconomiae publicae (Dr. oec. publ.)
at the
Ludwig Maximilians University, Munich
2005
Thesis Supervisor: Prof. Stephan Klasen, Ph.D.
Thesis Co-Supervisor: Prof. Dr. Joachim Winter
Final Committee Consultation: January 27, 2006Acknowledgements
I would like to thank all my classmates and colleagues at the Institute for Advanced
Studies (Vienna), University of Munich and University of G˜ottingen for their numer-
ous comments and suggestions that helped improving this work considerably. I’m
particularly indebted to Andrea Weber, Joachim Wolfi and Christian Holzner whose
contribution to this research is di–cult to overestimate.
MyveryspecialthanksaretoStephanKlasenforhisexcellentsupervision,hispatience
and permanent support and encouragement.Contents
Preface i
1 The Model With Continuous Productivity Dispersion 1
1.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
1.2 Main Theoretical Results . . . . . . . . . . . . . . . . . . . . . . . . . . 4
1.3 The Data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6
1.4 Estimation Methodology . . . . . . . . . . . . . . . . . . . . . . . . . . 10
1.4.1 The Likelihood Function . . . . . . . . . . . . . . . . . . . . . . 11
1.4.2 The Problem of Top-Coded Wages . . . . . . . . . . . . . . . . 13
1.5 Estimation Results and Discussion . . . . . . . . . . . . . . . . . . . . 15
1.5.1 Pareto Parameters . . . . . . . . . . . . . . . . . . . . . . . . . 16
1.5.2 Structural Parameters . . . . . . . . . . . . . . . . . . . . . . . 17
1.5.3 Sensitivity Analysis . . . . . . . . . . . . . . . . . . . . . . . . . 20
1.6 Concluding Remarks . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22
2 Discrete vs. Continuous Productivity Dispersion 24
2.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24
2.2 Unemployment Rates in West Germany . . . . . . . . . . . . . . . . . . 27
2.3 Necessary Theoretical Results . . . . . . . . . . . . . . . . . . . . . . . 32
2.4 The Data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36
2.4.1 Classiflcation of Workers in the Stock Samples . . . . . . . . . . 36
2.4.2 Unemployment and Job Durations, Exit States . . . . . . . . . 37
2.4.3 Wages and Beneflts . . . . . . . . . . . . . . . . . . . . . . . . . 39
2.5 Structural Econometric Model . . . . . . . . . . . . . . . . . . . . . . . 41
2.5.1 The Likelihood Function . . . . . . . . . . . . . . . . . . . . . . 41
2.5.2 Nonparametric Estimation and Its Limitations . . . . . . . . . . 422.5.3 Parametric Estimation of the Model . . . . . . . . . . . . . . . 44
2.6 Estimation Results and Discussion . . . . . . . . . . . . . . . . . . . . 47
2.6.1 Preliminary Discussion . . . . . . . . . . . . . . . . . . . . . . . 47
2.6.2 Overall Fit of the Model . . . . . . . . . . . . . . . . . . . . . . 49
2.6.3 Estimation Results by Skills . . . . . . . . . . . . . . . . . . . . 50
2.6.4 by Age . . . . . . . . . . . . . . . . . . . . . 53
2.7 Summary and Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . 55
3 Extension to Non-Linear Technology with Increasing Returns 57
3.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 57
3.2 Theory . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 59
3.2.1 Framework . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 59
3.2.2 Workers’ Search Strategy . . . . . . . . . . . . . . . . . . . . . . 60
3.2.3 Employers’ Wage Posting . . . . . . . . . . . . . . . . . . . . . 61
3.2.4 Properties of the Equilibrium Ofier and Earnings Distributions . 64
3.3 Econometric Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 66
3.3.1 The Likelihood Function . . . . . . . . . . . . . . . . . . . . . . 66
3.3.2 Homogeneous Firms . . . . . . . . . . . . . . . . . . . . . . . . 67
3.3.3 Heterogeneous Firms . . . . . . . . . . . . . . . . . . . . . . . . 69
3.3.4 Speciflcation Check . . . . . . . . . . . . . . . . . . . . . . . . . 72
3.4 Empirical Application . . . . . . . . . . . . . . . . . . . . . . . . . . . 73
3.4.1 The Data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 73
3.4.2 Estimation Results: Fit of the Model . . . . . . . . . . . . . . . 76
3.4.3 Social Returns to Education . . . . . . . . 79
3.5 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 81
A : Appendix to Chapter 1 83
B : Appendix to Chapter 2 87
C : Appendix to Chapter 3 93
Bibliography 99List of Tables
1.1 Duration Data for the Austrian Labour Market . . . . . . . . . . . . . 7
1.2 Estimated Parameters for the Pareto Tails . . . . . . . . . . . . . . . . 16
1.3 Frictional Parameters . . . . . . . . . . . . . . . . . . . . . . 19
1.4 Goodness of Fit Tests for the Wage Ofier Distribution . . . . . . . . . . 21
2.1 Entitlement Length of Unemployment Insurance Beneflt . . . . . . . . 31
2.2 Descriptive Statistics of Event History Data for the Two Stock Samples 38
2.3 Test Results for Search Intensities . . . . . . . . . . . . . . . . . . . . . 50
3.1 Descriptive Statistics of Event History Data . . . . . . . . . . . . . . . 74
3.2e of Earnings Data . . . . . . . . . . . . . . . . . . 76
B.1 Estimated Model for the Whole Economy. . . . . . . . . . . . . . . . . 88
B.2 Model for Difierent Qualiflcation Groups . . . . . . . . . . . 89
B.3 Estimated Model fort Age Groups . . . . . . . . . . . . . . . . 90
C.1 Estimation Results: Homogeneous Firms . . . . . . . . . . . . . . . . . 95
C.2 3-Point Employer Heterogeneity . . . . . . . . . . 97List of Figures
1.1 Earnings Density 1988 . . . . . . . . . . . . . . . . . . . . . . . . . . . 9
1.2 Density 1994 . . . . . . . . . . . . . . . . . . . . . . . . . . . 9
2.1 Relative Unemployment Rates by Skill Groups . . . . . . . . . . . . . . 28
2.2eyment Rates by Age . . . . . . . . . . . . . . 28
2.3 Kernel Estimates of Earnings Densities in 1986 and 1995 . . . . . . . . 40
A.1 Wage Ofier and Earnings Densities for the Two Samples . . . . . . . . 84
A.2 Wage Ofier and Distributions for the Two Samples . . . . . . 84
A.3 Expected Job Duration (years) . . . . . . . . . . . . . . . . . . . . . . 85
A.4 Index of Monopsony Power . . . . . . . . . . . . . . . . . . . . . . . . . 85
A.5 Proflt Ratio Plot . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 86
B.1 Estimated Theoretical Ofier and Earnings Distributions for the whole
Economy: Sample 1986 . . . . . . . . . . . . . . . . . . . . . . . . . . . 91
B.2 Estimated Theoretical Ofier and Earnings Distributions for the whole
Economy: Sample 1995 . . . . . . . . . . . . . . . . . . . . . . . . . . . 91
B.3 Estimated Theoretical Wage Ofier Densities for the High Skilled Group 92
C.1 Kernel Estimates of Earnings Densities . . . . . . . . . . . . . . . . . . 94
C.2 Kernel of Labour Cost Densities . . . . . . . . . . . . . . . . 94
C.3 Aggregate Wage Ofier Densities: Homogeneous Firms . . . . . . . . . . 96
C.4 Theoretical Earnings Firms. . . . . . . . . . . 96
C.5 Aggregate Wage Ofier Densities: 3-Point Employer Heterogeneity . . . 98
C.6 Theoretical Earnings 3-Point Employery . . . . 98Preface
The search equilibrium environment is designed to model an equilibrium in the labour
market in the presence of search frictions. Following Pissarides (2002), under search
friction we understand any time delay in getting a job by unemployed worker or,
similarly, fllling an open vacancy by a flrm. Historically, search equilibrium models
originate from one-sided job search models that described exclusively the behaviour
of the labour supply side under incomplete information about ofiered wages. Early
contributions to the economics of job search include McCall (1970) and Gronau (1971)
amongmanyothers. Inrelativelysimilarenvironmentstheyderiveanoptimalstopping
rule for an unemployed agent who has a limited access to job ofiers, drawing only
one such ofier per period of time. Later on this framework was extended to include
the possibilities of quitting and dismissals . Other in uential contributions to the
development of one-sided job search models set the following benchmarks. Jovanovic
(1979) introduces job matching with noise and uses this to explain age-earnings proflle
and thenegativecorrelationbetweenquitsandtenure. Mortensen (1977) studies time-
limited unemployment insurance and its efiects on exit from unemployment. Van
den Berg (1990) generalizes a job search model to allow non-stationary wage ofier
distributionanddurationdependenceinbothbenefltpaymentsandarrivalratesofjob
ofiers.
None of the papers on one-sided job search, however, was able to reproduce the
empiricallyobservedwagedispersion. Moreover,agentsbehaviourrepresentedbythese
modelsmaynotbeconsistentwithequilibriumbehaviour,sincethelabourdemandside
was completely ignored. These two facts brought to life extensions of one-sided models
that resulted in the creation of a search equilibrium framework.
There exist two main ways of modelling search equilibrium on the labour market,
which mostly depend on the view of the nature of search frictions and the nature of
equilibrium wage setting. The flrst approach is to view search frictions as incomplete
iinformati