Labour Market Structure and Reemployment rates: Unemployment dynamics in West Germany and the United States

Unemployment dynamics differ markedly between the United States and Germany: over the 1980s and 1990s, spells of unemployment have been considerably shorter, and sectoral mobility rates were consistently higher among displaced workers in the United States than those experienced by German workers. In the light of earlier research on Germany’s strongly skilland occupation-based labor market structures, the current paper addresses whether long spell durations among German workers reflect constraining effects of tight labor market structures inhibiting processes of adjustment to structural changes in the economy. While event history analyses of longitudinal micro data from the Survey of Income and Program Participation and the IAB Employment Register File show many cross-national similarities in unemployment processes, they also consistently establish stronger sectoral effects on reemployment rates in the German labor market. Interestingly, while U.S.-German differences in labor market dynamics are very well suited to explain the observed cross-national differences in unemployment duration, labor market structures do not provide an adequate explanation of country differences in post-unemployment occupational outcomes. Rather than facing unduly constraining market structures, unemployed workers in Germany in fact appear to respond more easily to changing economic structures than common for U.S. workers. Labor market structure and reemployment rates 3 UNEMPLOYMENT AND LABOR REALLOCATION Unemployment is often seen as being closely tied to processes of structural change in the labor market. To a certain extent, workers become unemployed simply because they have been „in the wrong place, at the wrong time“ (Leonard 1987), and so became exposed to the consequences of job destruction in particular industries, firms, or occupations. Workers’ mobility reactions to adverse events like job loss and subsequent unemployment have been the focus of interest of a large number of studies in both labor economics and social stratification which have addressed the issues of unemployment incidence (DiPrete 1993; DiPrete et al. 1997; DiPrete and Nonnemaker 1997; Bender et al. 2000; Layte et al. 2000), unemployment duration (Pedersen and Westergård-Nielsen 1993; Katz and Meyer 1990; Meyer 1990; Hunt 1995), as well as earnings losses (DiPrete and McManus 2000; Burda and Mertens 2001) and patterns of sectoral mobility subsequent to unemployment spells (Thomas 1996; Fallick 1993; Carrington 1993). Obviously, the latter aspect is of particular importance as far as the relation between structural change and employment careers is concerned: As job opportunities shift between sectors and occupations, prolonged unemployment might result out of workers’ inability to respond appropriately to changing economic conditions. At any rate, the analyses of DiPrete and colleagues (DiPrete 1993; DiPrete et al. 1997; DiPrete and Nonnemaker 1997; but cf. also Harrison 1988; Haveman and Cohen 1994; Idson and Valletta 1996; Thomas 1996; Bender et al. 2000; Burgess et al. 2000) have recently emphasized the effects of industrial, occupational and company expansion and decline on job dynamics and found ample empirical evidence for the effects of changing opportunities on patterns of job mobility and worker reallocation across industries, occupations and firms. Against that background, the present paper compares unemployment processes among displaced workers in the United States and West Germany. These two economies exhibit markedly different labor market structures, both in terms of aggregate labor market dynamics and the segmentation of labor market processes. Over the 1980s and 1990s, the U.S. labor market has been considerably more dynamic than the West German one; in addition, labor market pressures associated with sectoral decline and structural shifts in the labor force have been more apparent in Germany (Schettkat 1992; DiPrete et al. 1997). On the other hand, the strongly skilland occupation-based organization of the German labor market (Mayer and Carroll 1987; Marsden 1990) might have impeded processes of structural adjustment in the labor force to a larger extent than has been the case in the more flexible U.S. labor market structure. The stylized facts, at any rate, seem to support this interpretation: on average, unemployment spells have been considerably shorter in the United States, and as will be shown below, sectoral mobility rates have also been substantially higher among displaced workers there. This is also well reflected in the recent assessment by DiPrete et al. (1997:325) Labor market structure and reemployment rates 4 stating that "Germany ... with its strong credential-based occupational structure ... has restricted the viability of midcareer shifts as a tool of structural adjustment." In the following, the paper provides an empirical analysis of such adjustment processes at the level of individual workers, focusing on the role of both individual skills and features of employment sectors in these processes. Given important U.S.-German differences in labor market structures, the cross-national comparison provides the opportunity to directly test for the effects of these alleged structural sources of U.S.-German differences in unemployment dynamics. The underlying theoretical background is discussed in more detail in the following section of the paper, while Section 3 below will present details on the databases and statistical methodology used in the empirical analyses. Section 4 then presents some basic descriptive findings on the structure of unemployment dynamics in the two countries, and Section 5 will discuss the outcomes of the main multivariate analyses. Section 6 finally provides a summary of the results and the overall conclusions from the study. UNEMPLOYMENT PROCESSES, SPECIFIC SKILLS, AND OPPORTUNITY STRUCTURES From a dynamic perspective, observable labor market dynamics are the outcomes of two-sided search and matching processes between employers intending to fill job vacancies and job searchers looking for employment or a new job (Sørensen and Kalleberg 1981; Logan 1996; Mortensen and Pissarides 1999 provide an overview on economic matching models). Hence, observable labor market transitions result from the joint preferences of employers for workers with particular skill or other characteristics and workers’ preferences for particular job features. The rate of new job match formation is thus determined by the rate at which employers provide job opportunities to particular workers, and the rate at which workers consider received job offers an improvement over their current state of affairs (Mortensen 1988). With respect to unemployed workers, this implies that outflow rates into new jobs will be governed by both opportunities arising from employers’ job offers and worker choices among actual jobs offered. Given that unemployed workers are likely to search for reemployment across multiple labor market sectors, the total hazard rate results from the product of job offer arrival rates λ in sector k and the probability that workers will accept employment offered in sector k, summed across the whole labor market. In algebraic terms,