Abstract
While results are starting to emerge, not much is known yet about the dynamics of the labor markets of the former Eastern economies, especially in the context of the current financial crisis. Arguably, this is mainly due to paucity of (panel) data. By examining labor market transitions and earnings growth and their correlates using a recent panel data set for Serbia, this paper combines both of these issues. Estimation of gross transition probabilities reveals that females are disadvantaged in the Serbian labor market in terms of moving out of the two undesirable states, unemployment and economic inactivity, relative to males during the first year of the financial crisis—though males are harder hit than females in terms of the levels of unemployment. In terms of earnings growth, the picture is reversed, with females experiencing smaller earnings decreases than males (though, owing to the gender earnings gap, from a much lower base). Examining the determinants of employment, unemployment, and inactivity transitions and of earnings growth reveal substantial gender differences related to individual, job, and firm characteristics. The overall results therefore hint at both males and females being hit in terms of employment and earnings, though in different ways. Finally, the paper discusses policy implications and provides suggestions for further research.
This is a preview of subscription content, log in via an institution to check access.
Similar content being viewed by others
Notes
Which specifically denotes the flows among the main labor force status flows studied here (typically the three states of employment, unemployment, and inactivity—though one could further split some of these states; for example by distinguishing between types of employment (e.g., full-time or part time—or formal or informal sector employment) or type of inactivity (retirement or other inactivity) as opposed to, say, job or industry flows.
This section draws heavily upon Babović (2008).
Due to rationing and barriers to entry into employment there might not be much of a choice between this and unemployment. There still is a choice between being in the labor force and being economically inactive, however.
A list of all the variables used in these analyses as well as their definitions is given in Table 8, Appendix 1. The definition of variables is discussed in more detail in the remainder of this section.
Details available upon request.
Due to a few extreme values in both tails of the distribution, we censor earnings growth at −35 and +40 %, respectively.
Due to convergence problems for the male subsample in the inactivity transition analysis, the region fixed effects had to be excluded from this estimation for that one subgroup analysis.
For individuals without salary information, salaries are imputed as the mean salaries of all other workers. To account for imputed salaries, a dummy for imputation status is added.
Since these may be systematically different from other employees, as well as to account for the censoring (the zero salaries imputed to these workers).
Again, for individuals without household income information, per capita incomes are imputed as the mean per capita salaries of all other households. To account for imputed incomes, a dummy is added for imputation status.
And the associated dummy variable for the individuals with imputed per capita household income.
A potential issue when using OLS, in addition to the endogeneity concerns raised below, is the potential for selectivity of employees, so that workers with observed wages are not a representative sample of the general population. One way of addressing this is Heckman’s selection model, where the selection into employment is explicitly modeled in addition to the earnings structure. While exclusion restrictions here are frequently questionable, we did experiment with a selection model for the earnings growth analysis using marital status and the presence of children in the household as identifying variables in the selection equation—with quite similar results to those obtained from OLS. We therefore use OLS here–since OLS is both simpler and more robust.
Still leaving, of course, potential endogeneity issues related to omitted variables and/or measurement error. If we had available any potential instrumental variables (IV)—i.e. variables which are correlated with the potential endogenous variables without affecting the dependent variables directly, an IV/Two-Stage Least Squares approach could be employed. Such variables—such as distance to educational institutions to instrument for educational attainment (Card, 1995; 2001)—are not available in the present dataset, unfortunately, and so these methods cannot be employed here.
Again, in addition to being of interest in and by itself due to the impact of workers’ livelihoods from the drop in earnings following the financial crisis, the drop in earnings (negative earnings growth) is one of the potential underlying causes of the labor market transitions in Serbia during the crisis.
References
Arandarenko M, Avlijas S (2010) Behind the veil of statistics: wage trends before and after the crisis in Serbia. Study commissioned by ILO, Belgrade
Babović M (2008) The position of women on the labour market in Serbia. Gender Equality Council, Government of the Republic of Serbia and United Nations Development Programme, Belgrade
Becker GS (1964) Human capital. University of Chicago Press, Chicago
Bellman L, Estrin S, Lehmann H, Wadsworth J (1995) The Eastern German labor market in transition: gross flow estimates from panel data. J Comp Econ 20:139–170
Blau FD (1996) Where are we in the economics of gender? The gender pay gap. NBER working paper no 5664, National Bureau of Economic Research, Cambridge
Card D (1995) Using geographic variation in college proximity to estimate the return to schooling. In: Christofides LN, Grant EK, Swidinsky R (eds) Aspects of labor market behaviour: essays in honour of John Vanderkamp. University of Toronto Press, Toronto
Card D (2001) Estimating the return to schooling: progress on some persistent econometric problems. Econometrica 69(5):1127–1160
Frazis HJ, Ilg RE (2009) Trends in labor force flows during recent recessions. Mon Labor Rev 132(4):3–18
Gomes P (2012) Labour market flows: facts from the United Kingdom. Labour Econ 19(2):165–175
Heckman J (1981) Statistical models for discrete panel data. In: Manski C, McFadden D (eds) Structural analysis of discrete data with econometric applications. MIT Press, London
Huber PJ (1967) The behavior of maximum likelihood estimates under nonstandard conditions. In: Le LM, Cam J, Jerzy Neyman (eds) Proceedings of the fifth Berkeley symposium on mathematical statistics and probability, vol 1. University of California Press, Berkeley
IMF (2009) World economic outlook. International Monetary Fund, Washington, DC
Lauerova JS, Terrel K (2002) Explaining gender differences in unemployment with micro data on flows in post-communist economies. IZA Discussion Paper No. 600
Marston ST (1976) Employment instability and high unemployment rates. Brookings Pap Econ Act 1:169–203
Matković G, Boško M, Marina P (2010) Impact of the financial crisis on the labor market and living conditions outcomes, mimeo, center for liberal-democratic studies
Mincer J (1974) Schooling, experience and earnings. National Bureau of Economic Research, New York
Ribar D (1992) Child care and the labor supply of married women: reduced form evidence. J Hum Resour 27(1):134–165
Tasci HM, Tansel A (2005) Unemployment and transitions in the Turkish labor market: evidence from Individual level data. IZA discussion paper no. 1663
Toikka R (1976) A markovian model of labor market decisions by workers. Am Econ Rev 66(5):821–834
White H (1980) A heteroskedasticity-consistent covariance matrix estimator and a direct test for heteroskedasticity. Econometrica 48(4):817–830
Acknowledgments
We thank Alicia Adsera, Benu Bidani, Hank Farber, Caterina Laderchi, David Lee, Josefina Posadas, Sarosh Sattar, seminar participants at Princeton University (Industrial Relations Section Graduate Labor Lunch Seminar), Washington and Lee University (Economics Department Lunch Seminar), and the World Bank; and participants at the World Bank-DFID Poverty and Social Inclusion in the Western Balkans conference, Brussels, and the 6th IZA-WB Employment and Development Conference, Mexico City, for helpful comments and suggestions. Comments and suggestions from Editor-in-Chief George Hondroyiannis and an anonymous Referee helped greatly improve this paper. Remaining errors and omissions are our own. The first draft of this paper was written while the first author was a Departmental Guest at Princeton University’s Economics Department; the hospitality and support of the Department is gratefully acknowledged. The data were kindly provided by the Statistical Office of the Republic of Serbia (RSO). The assistance of Vladan Božanić, RSO, to help us understand the data better is highly appreciated. The findings and interpretations are those of the authors and should not be attributed to the World Bank, any affiliated institutions, or to the Statistical Office of the Republic of Serbia.
Author information
Authors and Affiliations
Corresponding author
Appendices
Appendix 1
See Table 8.
Appendix 2: Descriptive statistics for estimation samples
Appendix 3: Results from regression models
Rights and permissions
About this article
Cite this article
Blunch, NH., Sulla, V. World gone wrong: the financial crisis, labor market transitions and earnings in Serbia. Econ Change Restruct 47, 187–226 (2014). https://doi.org/10.1007/s10644-013-9147-6
Received:
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s10644-013-9147-6