Testing the Stationary Search Model

Abstract

This paper uses tight parametric assumptions to model individual unemployment histories and confronts the resulting structural form with data from the Federal Republic of Germany. It thereby provides favorable evidence for the usefulness of models that deviate from the familiar reduced form approach in an attempt to give more economic structure to the analysis of individual unemployment spell durations. It follows the work of Kiefer and Neumann (1979), Lancaster and Chesher (1984), Lancaster (1985), and, more recently, Jones (1988). The model allows derivation of the same log-linear simultaneous equation representation for elapsed durations and reservation wages of currently unemployed individuals and for completed durations and accepted wages of re-employed individuals.

This paper is a revised version of the first chapter of my doctoral dissertation at Princeton University. Financial support by the Industrial Relations Section. Princeton University, and by the Alfred P. Sloan Foundation for its Doctoral Dissertation Fellowship is gratefully acknowledged. Preliminary versions of this paper have been presented at the Princeton Labor Lunch series and at the EEA meeting in Lisbon, 1990. I thank Qrley Ashenfelter, Angus Deaton, Richard Quandt, Klaus F. Zimmermann and in particular David Card for many helpful suggestions and advice. The analyzed data set was provided by the Zentralarchiv für Empirische Sozialforschung. Köln.