Chapter 2

Box 2.1

The return of growth to the transition region in 2010 and 2011 largely failed to translate into a labour market recovery. Employment has lagged behind GDP in terms of average growth rates, which only turned positive in late 2010 (see Chart 2.1.1), and there has been no improvement in youth employment, which has continued to decrease in many countries ever since the 2008-09 crisis. Average figures, however, conceal a considerable degree of heterogeneity across the region. Some countries have seen employment rebound, while others have experienced a jobless recovery.1  This analysis shows that countries vary in the extent to which growth is associated with job creation and relates this variation to differences in labour market institutions and in the share of long-term unemployment. 

Source: Eurostat and national authorities via CEIC Data. Note: The chart shows unweighted averages for the transition region.

Economists typically expect a linear relationship between growth and changes in employment or unemployment, as encapsulated by Okun’s law.2  Estimates for individual transition countries show that a one-percentage point increase in quarterly growth is associated with a sizeable increase in employment in some economies but has no significant effect in others (see Chart 2.1.2). The speed of transmission also differs, with growth only leading to changes in employment with a lag in some countries (as in the case of Croatia, Hungary, Serbia and Slovenia). Similar analysis for youth employment reveals an even greater degree of heterogeneity. The results indicate there is a clear relationship in the Baltic states, where youth employment rose steadily prior to 2008 before falling by over 30 per cent in the wake of severe recessions, but not in Hungary and Romania, where employment among young people has been declining continuously for over a decade.

Source: Underlying data are from national authorities via CEIC Data. Note: This chart depicts the results of country-by-country regressions of employment growth on real GDP growth and lagged real GDP growth using seasonally adjusted quarterly data for the period 1999-2011. Where significant, the chart shows the cumulative effect of contemporaneous (blue) and lagged growth (red).

This variation could be due to a number of factors. Labour markets may differ in the extent to which growth generates vacancies, or in the efficiency of matching prospective workers to those vacancies.3  Some transition economies also have large informal sectors, where job creation would not register in official employment figures. In others, public-sector employment, which is likely to be less responsive to growth, continues to represent a considerable share of the total.

A further explanation is that labour market rigidities in some countries dampen the effects of the cycle, preventing large falls in employment during recessions but acting as constraints in periods of growth. An analysis of employment and labour market institutions in the transition region between 2006 and 2011 provides some evidence for this. Table 2.1.1 shows the results of a panel regression in which annual employment growth was regressed on current and lagged growth rates, different measures of labour market institutions and, critically, the interaction between institutions and growth. The institutional measures are based on methodology from the World Bank Doing Business project and capture, respectively, the difficulty of hiring, the rigidity of hours and the difficulty of redundancy (see note attached to Table 2.1.1). Across the transition region, the correlation between these measures is surprisingly low, with many countries having a mixture of flexible and rigid institutions across different categories. The coefficients on the institutional measures are insignificant, which is to be expected since their levels are unlikely to have a direct effect on changes in employment. The interaction term estimates whether institutional rigidities dampen or amplify the effect of growth on employment. It is negative and significant for the difficulty of hiring measure (see column 1), which suggests that countries where hiring workers is costly and contracts are inflexible are likely to see smaller increases in employment for a given rate of growth. In contrast, rigid regulations governing working time or redundancy do not appear to affect the relationship between growth and employment (see columns 2 and 3).

Variables Dependent variable: Annual change in employment
GDP growth 0.487*** 0.312*** 0.350*** 0.487***
  -0 -0 -0 -0
Lagged growth 0.083** 0.080** 0.089** 0.155
  -0.023 -0.03 -0.016 -0.212
Difficulty of hiring 0.778      
(DHI) -0.486      
DHI interaction with growth -0.200**      
Rigidity of hours   0.043    
(RHI)   -0.968    
RHI interaction with growth   0.023    
Difficulty of redundancy     -1.095  
(DRI)     -0.232  
DRI interaction with growth     -0.025  
Share of long term unemployed     0.021  
Long term share interaction       -0.005*
Country fixed effects (Coefficients not reported)      
Constant -2.146 -1.612 -1.687 -2.195
  -0.143 -0.275 -0.11 -0.002
Number of observations 174 174 174 187
Period 2006-2011 2006-2011 2006-2011 1999-2011

Source: Underlying data from national sources via CEIC Data and Doing Business Employing Workers database.

Note: The table shows OLS regression coefficients with p-values in parentheses. "Difficulty of hiring", "Rigidity of hours" and "Difficulty of redundancy" describe several labour market institutions which affect labour market flexibility in the respective dimension, based on a methodology described in the Doing Business 2012 annex on employing workers, adapted from Botero, Djankov, La Porta, Lopez-de-Silanes and Schleifer (2004). The institutional measures in this analysis are dummy variables based on these indices which take the value of 1 if a country scores above the median (implying rigid institutions) and 0 if it scores below it (flexible institutions).

These results imply that countries where growth does not result in job creation could benefit from structural reforms that facilitate hiring. At the same time, the lack of significance for other institutions and the low correlation between the indices, suggests a need for targeted reforms, such as restrictions on contracts and minimum wage laws,4  rather than broader labour market liberalisation. 

The cross-country variation in the relationship between growth and employment is also determined by a second factor – the extent of long-term unemployment. Individuals who remain unemployed for protracted periods of time begin to lose skills and become detached from the labour force and less attractive to employers. Compared with those who have been in recent employment, they are less likely to benefit from an increase in growth and a rise in vacancies. This dynamic also appears to affect labour markets at the aggregate level. Column 4 in Table 2.1.1 shows the results of including the proportion of the unemployed who are long-term jobless, and its interaction with growth, in the regression. The interaction term is negative and (just) statistically significant, implying that an increase in growth generates fewer jobs the greater the share of long-term unemployed among the job-seeking population. 

In light of this finding, the rise in the long-term unemployment rate across many transition countries since 2008 (see Chart 2.1.3) may be of concern. Countries where this rate remains above pre-crisis levels (or rises due to a renewed downturn) could experience a lasting crisis impact on the transmission from growth to employment. 

Source: Eurostat and national authorities via CEIC Data. Note: The chart shows the long-term unemployment rate in 2011 minus the corresponding rate in 2008. Data are not available for some countries in the SEE, EEC, CA and SEMED regions.

From the analysis of individual countries it appears that institutional differences and varying levels of long-term unemployment can account for part of the heterogeneity depicted in Chart 2.1.2. Of those countries where there was no significant relationship between growth and employment, Armenia and FYR Macedonia have a high proportion of long-term unemployed, while Moldova, Morocco, Romania and Turkey have relatively rigid institutions related to hiring.5 In practice, the two factors are likely to be related, since institutional rigidities will determine how quickly those who become unemployed are able to re-enter the workforce.

In Bosnia and Herzegovina, Bulgaria, Georgia, Moldova and Serbia, real GDP has surpassed its pre-crisis level while employment remains below 2007 levels.

Okun’s law, first documented by Arthur Okun in the 1960s, is a rule of thumb measure relating changes in the unemployment rate to changes in the growth rate.

An analysis of vacancy rates and “Beveridge curves” which indicate how vacancy rates affect unemployment rates for selected transition countries provides some support for both explanations. Vacancy rates are more responsive to GDP growth in some countries than in others, while Beveridge curves indicate that the efficiency of matching also varies. Poland, where the estimated impact of growth on employment is the highest, is also the country where growth had the greatest effect on vacancies (in regressions of vacancy rates on growth and lagged growth) and where matching in the labour market has been most efficient in recent years (based on a comparison of estimated Beveridge curves).

“Difficulty of hiring” is an index based on the ratio of minimum wage to average wage, on regulations governing whether fixed-term contracts are prohibited for permanent tasks and on the maximum duration of fixed-term contracts.

In Romania reforms to the labour code implemented in May 2011 allowed for greater flexibility in hiring, redundancy and the use of fixed term contracts. These changes are reflected in a drop in the ‘Difficulty of hiring’ index, which nevertheless remains above the median for EBRD countries (indicating relatively rigid institutions).


icon-pdfOther Reports

Annual Report 2012
pdf English
pdf French
pdf German
pdf Russian

Financial Report 2012
pdf English
pdf French
pdf German
pdf Russian

pdf Donor Report 2013

pdf Sustainability Report 2012