Studies in Contemporary Economics Editorial Board D.Bos B. Felderer B. Gahlen H. J. Ramser K. W. Rothschild Wolfgang Franz (Ed.) Structural Unemployment With 25 Figures Springer-Verlag Berlin Heidelberg GmbH Professor Dr. Wolfgang Franz Department of Economics University of Konstanz P. O. Box 5560 D-7750 Konstanz, FRG ISBN 978-3-7908-0605-2 ISBN 978-3-642-58163-2 (eBook) DOI 10.1007/978-3-642-58163-2 This work is subject ta copyright. AII rights are reserved, whether the whole ar part ofthe material is concerned, specifically the rights of translation, reprinting, reuse of iIIustration, recitation, broadca sting, reproduction on microfilms ar in other ways, and storage in data banks. Duplication ofthis publi cation ar parts thereof is only permitted under the provisions of the German Copyright Law of Sep tember 9,1965, in its version of June 24,1985, and a copyright fee must always be paid. Violations fali under the prosecution act of the German Copyright Law. © Springer-Verlag Berlin Heidelberg 1992 Originally published by Physica-Verlag Heidelberg in 1992 The use ofregistered names, trademarks, etc. in this publication does not imply, even in the absence of a specific statement, that such names are exempt from the relevant protective laws and regulations and therefore free for general use. 710017130-543210-Printed on acid-free paper Preface High and persistent unemployment rates in Europe during the eighties gave rise to a lively discussion about the nature and causes of joblessness. Among other sources structural unemployment was blamed for the lack of response of unemployment to increasing aggregate demand. Renewed attention was thus devoted to an analysis of the magnitude and the development of structural unemployment as well to its possi ble determinants. In this literature, the Beveridge curve experienced a resurrection and, at first glance, it seemed to be an appropriate tool to analyse the aforementioned issues. However, it was soon recognized that the Beveridge curve, i.e. the relation between unemployment and vacancies, was anything but stable, thus requiring a care ful distinction between dynamic loops around a (stable?) long-run Beveridge curve and possible shifts due to, say, an increasing mismatch between labor supplied and demanded. The controversy is far from being settled at the time of this writing. This book contains a collection of hitherto unpublished papers which are devoted to a theoretical and econometric analysis of structural unemployment. The papers put considerable emphasis on the question to what extent the Beveridge curve can serve as an adequate tool for such studies. The countries under consideration are Germany and Austria. In what follows a very brief summary of each paper will be outlined. Franz and Siebeck present, at some length, a theoretical and econometric analysis of the Beveridge curve in Germany. The theory consists of three elements: the search process seen from both the employer's and the unemployed seeker's viewpoint and the matching technology governing the labor market. Major explanatory variables of possible shifts of the Beveridge curve are variables which determine the contact and contract probability, respectively, such as unemployment benefits, long-term un employment, minimum hiring standards set by the firm, costs of regional mobility, training costs, turnover rate, and the perceived wage distribution. Many of these vari ables, however, have an ambiguous effect on the location of the Beveridge curve. In the empirical part, attempts to overcome data deficiencies are reported and various specifications of the Beveridge curve are tested. The Beveridge curve turns out to be rather unstable. There may be outward shifts but the dynamics around the curve remain to be analyzed further. The authors conclude that several problems make this relationship a less straightforward tool for analysing structural unemployment and for drawing policy implications to combat this type of unemployment. Borsch-Supan's paper is concerned with the identifiability of the Beveridge curve. He starts with the notion that a shift of this relationship has often been interpreted as a secular deterioration of the equilibrating mechanisms between labor demand and VI Wolfgang Franz labor supply, hence as a secular increase of structural unemployment. Moreover, this recognition has eventually led to a shift in economic policy from Keynesian macroe conomic stabilization policy towards more microeconomically-oriented labor market policies that intend to reduce labor market frictions. This interpretation, however, is arguable. His paper pursues the question whether the unemployment-vacancy relation is indeed a stable structural macroeconomic relation that can be properly identified by standard econometric techniques. The conclusion arrived at is that neither theory, nor evidence for Germany support this view. It therefore appears hazardous, if not haphazard, to use the empirical unemployment-vacancy relation as an instrument for economic policy. The preceding papers have emphasized the role of various mismatch factors as an explanation for a possible outward shift of the Beveridge curve. It is therefore impor tant to have a closer look on mismatch unemployment. Buttler and Cramer engage in such an analysis and emphasize in some detail implications for labor market poli cies. First, an econometric analysis of the duration of vacancies and unemployment, respectively, brings the profile discrepancies between the unemployed and vacancies to the center of the discussion. The characteristics which imply a lower competitiveness, like age and health restrictions, have the greatest effect on the duration of unemploy ment according to their results. The various technical and organizational schemes for placement activities presently being used have been almost completely overwhelmed. In view of their findings labor market policies face greater difficulties than ever. The licensing of private placement agencies cannot overcome this; on the contrary, additional aide for the integration of problem cases - especially the older unemployed - is needed more than ever. With respect to unemployment Austria is in some aspects different from other European countries. It is therefore of great interst to become acquainted with the Austrian experience. In his paper Christl presents an analysis of the Beveridge curve in Austria. He concludes that in Austria, as in some other European countries, persis tence phenomena seem to account for the larger part of the rise of structural/frictional unemployment in Austria. Furthermore, indications that increased qualifications mis match may have plagued the matching process in the Austrian labor market are em phasized. On the other hand, regional and occupational mismatch and, moreover, unemployment benefits did not contribute to the rise of structural unemployment in Austria. What can be learned from these studies? Most importantly, the Beveridge curve is anything but a straightforward tool to analyse the development and causes of struc tural unemployment. Like other "simple" economic relationships such as Okun's law or the Phillips curve it suffers, firstly, from a rather complicated theoretical foundation which makes it difficult if not impossible to draw clear conclusions. Secondly, due to a lack of adequate data it is hard to obtain firm econometric estimates of the Beveridge curve, let alone of its possible shifts and dynamics. On the other hand, in light of many unsubstantiate and speculative discussions about "structural" unemployment inside and outside the economics profession, it seems to be preferable to base one's assessment of structural unemployment on something like the Beveridge curve, with considerable care, of course, taking into account various caveats some of which are pointed out in this book. VII Preface As the editor of this book, I wish to express my gratitude for enthusiastic co operation of all contributors. Moreover, my thanks go to Elisabeth Fetsch for her engagement in preparing this volume and to T. Schneeweis, T. Herz, R. Heisig and M. Redford for assistance. Fortunately, remaining errors are with the contributors. September 1991 Wolfgang Franz University of Konstanz Contents Wolfgang Franz and Karin Siebeck A Theoretical and Econometric Analysis of Structural Unemployment in Germany: Reflections on the Beveridge Curve 1 1 Introduction.................. 1 2 Theoretical Analysis . . . . . . . . . . . . . 3 2.1 Basic Concept of the Beveridge Curve 3 2.2 Theoretical Aspects of Job Matching 4 2.2.1 Overview of the Model. . . 4 2.2.2 The Firm's Decisions. . . . 5 2.2.3 The Job Seeker's Decisions 11 2.2.4 The Matching Process . . . 15 2.2.5 Derivation of the Beveridge Curve 18 3 Empirical Analysis . . . . . 21 3.1 Data Analysis. . . . . 22 3.2 Econometric Analysis 32 3.2.1 Data. 32 3.2.1.1 Mismatch Indicators: Regional and Professional Mis- match. . . . . . . . . . . . . . . . . . . . . . . . 32 3.2.1.2 Mismatch Indicators: Qualification Mismatches. 36 3.2.2 Impact of Explanatory Variables . . . . 42 3.3 Estimated Rates of Structural Unemployment. 46 4 Conclusions....................... 48 Appendix . . . . . . . . . . . . . . . . . . . . . . . . . . 49 A.l Decisions of the Firm: Second-Order Conditions and Comparative- Static Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ., 49 A.2 The Effects on the Reservation Wage . . . . . . . . . . . . . . . . . .. 53 A.3 Reactions of the Probability of a Contract, 11'2, with Changing Deter- minants of xO and wO ...•.••.....•••••..•••••••• 54 AA The Probability 11' that a Vacancy is Filled by an Unemployed Person 54 x Contents Axel H. Borsch-Supan On the Identifiability of the Relation Between the Rate of Unemploy- ment and the Vacancy Rate 59 1 Introduction............................. 59 2 A Model of Structural Unemployment . . . . . . . . . . . . . . 61 3 Panel Data for the Federal Republic of Germany, 1963 to 1986 66 4 The Role of Structural and Cyclical Variables in Explaining the Shifts of the Beveridge Curve ..... . . . . . . . . . . . . . . . . . . . . . . . . . . .. 67 5 Causal Relations Between the Shifts of the Beveridge Curve and Cyclical Variables .......... 71 6 Summary and Conclusions. . . . . . . . . . . . . . . . . . . . . . . . . . .. 74 Friedrich Buttler and Ulrich Cramer Developments and Causes of Mismatch Unemployment in West Ger- many 79 Introduction ................ . 79 2 Matching through Labor Offices. . . . . . 80 3 Beveridge-Curve and Matching Function 85 4 Analysis of the Duration of Vacancies .. 91 5 Analysis of the Duration of Unemployment 95 6 The Results in Context ......... . 101 7 Political Labor Market Recommendations . 103 Josef Christl Structural Unemployment in Austria 107 1 Introduction........................ 107 2 U IV Curve and Matching Function .......... . 108 3 Empirical Estimates of the Austrian Beveridge Curve 110 IV 4 Why has the Austrian U Curve Shifted Outwards? 116 4.1 Occupational, Regional and Qualificational Mismatch 116 4.2 Lower Search Intensity of the Unemployed. 119 4.3 Increased Selectiveness of the Firms 120 4.4 Hysteresis . . . . . . . . . . 121 4.5 The Econometric Analysis . 122 5 Conclusions............ 125 Appendix .............. . 126 A.1 The Calculation of the Corrected Vacancy Rate. 126 A.2 Calculation of the Underutilization Rate of Labour 128 A Theoretical and Econometric Analysis of Structural Unemployment in Germany: Reflections on the Beveridge Curve by Wolfgang Franz and Karin Siebeck Department of Economics, University of Konstanz, Germany· 1 Introduction Unemployment in the Federal Republic of Germany has remained at a permanent high level for years. The explanation for this continued lack of jobs has shifted both in the academia and in the public towards considerations which blame "structural factors" as the major source of this problem. The prerequisite for an economic analysis of these presumptions is a theoretical framework which is able both to capture factors which may be viewed as determining structural unemployment and to allow for an empirical • Work on this paper has been supported by a grant from the Deutsche Forschungsgemeinschaft which is gratefully acknowledged. We are indebted to M. Albert, A. Borsch-Supan, M. Keller, P. Michels, M. Rauscher, W. Scheremet, T. Schneeweis and, especially, F. Spiecker for helpful comments and able research assistance. 2 Wolfga.ng Fra.nz and Karin Siebeck test of the relevance of these determinants. The Beveridge curve, i.e., the relation between vacancies and unemployment, is often used as an analytical instrument to identify the extent and the causes of structural unemployment. In recent years several theoretical and empirical studies employing the Beveridge curve have been carried out such as Christl (1987) for Austria, Borsch-Supan (1988) and Franz (1987) for the Federal Republic of Germany and, Jackman, Layard, and Pissarides (1989), Jackman and Roper (1985) and Budd, Levine and Smith (1988) for Great Britain, to name but a few. Most if not all studies of this type wind up with the conclusion that the Beveridge curve has shifted outwards thus indicating a higher degree of malfunctioning of labor markets. In Germany, as in some other countries, this result is also obtained by estimating a disequilibrium macro-model as is shown in the study by Entorf, Franz, Konig, and Smolny (1991). The aim of this paper is to investigate whether the Beveridge curve is indeed a straightforward tool for analysing the extent and the causes of structural unem ployment. Some skepticism about this appropriateness is raised by the observation that while the studies mentioned above are quite certain about the identification of a Beveridge curve and its possible shifts, they are more speculative about the deter minants of the shifts. Prominent candidates are the rising proportion of long- term unemployed persons and various indicators for labor market maladjustments such as a regional and/or qualifications mismatch. However, while plausible, these variables are far from being unambiguous in determining the sign of the shift of the Beveridge curve. In the case of the proportion of long-term unempl~yment, for example, this can be seen clearly in the study by Budd, Levine, Smith (1988, p. 1082). In order to see why this is so and whether other explanatory factors have an ambiguous effect too, a theory of the unemployment/vacancy-relationship is required. Such a theory should also identify additional determinants (if any) for the location and curvature of the Beveridge curve. The first part of our paper attempts to make a contribution to a theoretical deriva tion of the Beveridge curve based on a microeconomic analysis of the search process seen both from the employer's viewpoint and from the perspective of an unemployed searcher. It also includes a treatment of the matching technology governing the labor market. In the empirical part we have tried to identify the functional form of the Beveridge curve for the Federal Republic of Germany. Moreover, we tested whether this relationship, if it exists, is stable .. As it turns out there are reasons which suggest an outward shift. Therefore, we elaborated on the significance of various explanatory variables which emerged from the theoretical considerations. The paper is organized as follows. In the next section we sketch the basic idea of the Beveridge curve in order to provide a guide to those readers unfamiliar with this relationship. Moreover, this results in questions which will have to be answered by the theoretical framework in subsequent considerations. Section (2.2) gives a detailed description of the theoretical model. It starts with an overview of the model (2.2.1) and goes on with a formulation of the search and hiring process seen from the firm's viewpoint (2.2.2). The search process seen from the applicant's viewpoint is out lined in subsection (2.2.3), while section (2.2.4) shows under which conditions and circumstances a match is formed. Subsection (2.2.5) derives the Beveridge curve and highlights several sources for shifts. The empirical part starts with a data analysis