Employment Uncertainty and the Incidence of Job Loss in Transition Economies: The Case for Central Europe.

• Author: Herzog, Jr., Henry W.
• Position: Statistical Data Included

Henry W. Herzog, Jr. [*]

A shared concern among Central European countries is the significant social cost imposed by market-oriented reforms. Although a growing body of literature has addressed the incidence of such reforms on unemployment distribution and duration, little is known of the actual transitions of workers from employment to unemployment and specifically the dynamics of transition-induced job loss. This study examines such transitions in the Czech and Slovak Republics, Poland, and Slovenia within econometric models whereby voluntary and involuntary employment separations are jointly determined. In this regard, employment reductions triggered by demand-driven reforms are accommodated at the industry and the occupation level by redundancies as well as voluntary quits in anticipation of redundancy. In addition, employed workers within the models consider the "costs" of job loss (likelihood of long-term unemployment) while both shirking, and thus risking dismissal, and contemplating quitting. Estimates on personal characteris tic variables in the involuntary and voluntary separation equations provide important new information on the incidence of job loss in Central European transition economies and specifically how such loss likely varies by gender, age, marital status, and completed education. Particular attention is devoted to gender differentials.

1. Introduction

   A common outcome within Central European economies is the immense social cost imposed by market-oriented reforms, as well as the skewed distribution of these costs among workers, prior workers, and their families. It is generally agreed that measures to reduce such hardship should play a critical role in the design of technical assistance projects to aid the transition process (OECD 1992). However, knowledge concerning the incidence of job loss in Central Europe is rather limited and confined for the most part to analyses of unemployment and its distribution. [1] At present, little is known concerning the involuntary movement of workers from employment to unemployment and specifically the dynamics of transition-induced labor shedding. [2]

   Contrary to popular belief, labor markets existed in Central European countries under central planning, insofar as labor supply decisions were mostly voluntary and driven by economic factors such as differentials in both wages and various forms of nonwage compensation. However, labor markets were heavily distorted in the sense that market signals provided incorrect incentives to workers and their employers. Labor demand was to a large extent independent of the financial viability of enterprises, reflecting instead the sectoral priorities established under central planning, the outcomes of bargaining between managers and bureaucrats over the allocation of state subsidies, and the social security role of enterprises in guaranteeing workers the right to employment (OECD 1992).

   As noted by Filer (1996), the economic transformation in Central Europe offers a "massive natural experiment" where truly exogenous shocks have created unusual research opportunities, a situation I exploit below. In this regard, the introduction of market-based prices, reduction of state subsidies, and redirection of trade to western economies have triggered major changes in sectoral performance, and in turn significant labor shedding. Thus, an analysis of transition-induced job loss throughout Central European countries should not only provide useful information about appropriate labor market policy there but also contribute to the growing body of literature on the more subtle process of economic restructuring and concomitant worker "displacement" in Europe, North America, and elsewhere (Hamermesh 1989; OECD 1990; Kletzer 1998).

   With this in mind, the study examines job loss early in the transition period in the Czech Republic, Slovak Republic (Slovakia), Poland, and Slovenia within a modeling framework whereby voluntary and involuntary employment separations are jointly determined. In this regard, employment reductions triggered by demand-driven reforms are accommodated at the industry and the occupation level through both redundancies as well as quits in anticipation of redundancy. In addition, decisions among employed workers within the models to shirk (and risk dismissal) on the one hand and to quit the enterprise (and enter job search) on the other--both acknowledge the cost of job search--are represented here by the likelihood of long-term unemployment. Estimates on personal characteristic variables in the voluntary and involuntary separation equations provide important new information on the incidence of job loss in Central European transition economies and specifically on how such loss likely varies by gender, age, marital s tatus, and completed education.

   The following section briefly describes the labor force microdata upon which the study is based. This section also considers mean likelihoods of involuntary and voluntary employment separations in the four Central European countries, as well as how such likelihoods vary across workers by demographic characteristics. Joint models of involuntary and voluntary separations are subsequently developed in section 3 and provide the basis for a discussion of likely feedback effects from redundancies to quits (and conversely from quits to redundancies), as well as probable interdependencies among dismissals, quits, and labor market conditions in the worker's industry and occupation. Econometric estimates of involuntary and voluntary employment separations are considered next in section 4. Examination of estimates on personal characteristic variables in these models provides important information on the incidence of job loss in Central European transition economies. Of particular interest is the role of gender in the d etermination of dismissals and redundancies (involuntary separations). Finally, conclusions from the study are summarized in section 5.

2. Data and Preliminary Evidence

   Employment separations contributed significantly to the rise in unemployment in all Central European countries starting in 1991, the outset of meaningful reforms. This trend resulted in an increasing share of displaced workers among registered job seekers and was bolstered by the phenomenon of mass layoffs. For instance, in Poland, almost one-fourth of the unemployed as of December 1991 were persons dismissed via group layoffs, up from 2% in January 1990 (OECD 1992).

   Of course, the burden of such redundancies and dismissals, as well as the uncertainty of job tenure in the immediate future, varied considerably by industry and occupation as market-driven reforms promoted significant labor shedding, particularly in agriculture, manufacturing, and construction. However, it is the incidence of this job loss on individuals that motivates this study, and it is to that end that I now turn.

   Information on employment separations, as well as the incidence of such separations, can be derived from labor force microdata recently made available by the Luxembourg Employment Study (LES). LES provides this information in a harmonized and standardized format and for a consistent set of variables across countries (see Forster, Helliesen, and Kolberg 1996). Based upon a number of considerations, to include the desire to work with a set of Central European transition economies exhibiting both high and low rates of involuntary separation (and unemployment), the labor force surveys (years) of Poland (1994), the Czech Republic (1994), the Slovak Republic (1995), and Slovenia (1994) were selected for analysis. [3]

   Observations drawn from this microdata were confined to individuals between the ages of 20 and 64 who were either employed at the survey date or without work at the survey date but had previously been employed. More specifically, the former individuals were required to be continuously employed over the 12 months immediately preceding the survey date, and the latter (separations) consist of individuals classified under International Labor Organization (ILO) definitions as either inactive or unemployed at the survey date who last worked within the previous 12 months. Among such separations, the distinction as to whether employment was terminated at the discretion of the employer or the employee was based upon responses to the question "main reason for leaving last job." Prior workers who stated that they were "dismissed or made redundant" (a single response) were assigned to involuntary separations; all others were assumed to be voluntary. [4] Finally, individuals employed (previously employed) in agriculture, forestry, or fishing, as well as those working (previously working) as skilled agricultural or fishery workers, were excluded from this population because of the unique nature of their work.

   Table 1 provides information on mean likelihoods of involuntary employment separation (number per 1000 at risk) over a retrospective period of 12 months or less in the four Central European countries under consideration. To add perspective to such involuntary rates of job loss, the overall 1994 (1995) unemployment rates within a consistently defined labor force (employed plus unemployed with prior work experience) were 12.9, 3.1, 9.9, and 7.7% in Poland, the Czech Republic, slovakia and Slovenia, respectively. With this in mind, note the significant variation in overall employment risk among the four countries in Table 1, with rates ranging from 24.3 per 1000 in the Czech Republic to 57.1 per 1000 in Poland. In this regard, the likelihoods of involuntary job loss (among the employed) and unemployment (within the labor force) are ordered the same. Of course, not all dismissals and redundancies search for reemployment, and the proportions of those who do also display considerable variation among Central Europe an countries. Although not shown in Table 1, the likelihoods of job search (and measured unemployment)...