WPS6759
Policy Research Working Paper 6759
Does Growth Generate Jobs
in Eastern Europe and Central Asia?
Kaspar Richter
Bartosz Witkowski
The World Bank
Europe and Central Asia Region
Poverty Reduction and Economic Management Department
January 2014
Policy Research Working Paper 6759
Abstract
In Eastern Europe and Central Asia, the link from growth the EU since 2004, the countries of former Yugoslavia,
to jobs was tenuous in the first decade of the transition, the Countries of Independent States and Turkey. The
giving rise to the notion of jobless growth. Yet, European authors compare these findings with other regions in
countries suffered large job losses during the recent the world. The paper shows that the responsiveness of
recession, suggesting that jobs and growth are closely employment to output increased in the second decade
entwined. This study takes a new look at this issue. It of the transition. It also finds that in some instances
provides a cross-country analysis of the employment employment growth increases with reforms of labor
intensity of growth over the last decade and a half in and product markets, stronger macroeconomic policy
Eastern Europe and Central Asia, which includes the frameworks, better governance, and more economic
11 Central and Eastern European countries that joined integration and diversification.
This paper is a product of the Poverty Reduction and Economic Management Department, Europe and Central Asia
Region. It is part of a larger effort by the World Bank to provide open access to its research and make a contribution to
development policy discussions around the world. Policy Research Working Papers are also posted on the Web at http://
econ.worldbank.org. The authors may be contacted at kaspar.richter@ec.europa.eu.
The Policy Research Working Paper Series disseminates the findings of work in progress to encourage the exchange of ideas about development
issues. An objective of the series is to get the findings out quickly, even if the presentations are less than fully polished. The papers carry the
names of the authors and should be cited accordingly. The findings, interpretations, and conclusions expressed in this paper are entirely those
of the authors. They do not necessarily represent the views of the International Bank for Reconstruction and Development/World Bank and
its affiliated organizations, or those of the Executive Directors of the World Bank or the governments they represent.
Produced by the Research Support Team
Does Growth Generate Jobs in Eastern Europe and Central Asia?
1
Kaspar Richter and Bartosz Witkowski
JEL No. E24,E32, J20
Keywords: economic growth, labor markets, Eastern Europe
Sector Board: EPOL
1
This paper was written by Kaspar Richter (kaspar.richter@ec.europa.eu), while employed at the World Bank, and
Bartosz Witkowski as background paper for the 2014 World Bank report "Back to Work – Growing with Jobs in Europe
and Central Asia". The views expressed in this paper are those of the authors and should not be attributed to the
World Bank or the European Commission. We are grateful to Carolina Sanchez-Paramo and Gabriele Giudice for
comments.
Introduction
One reason why countries care about growth is that growth generates jobs, and jobs are central
for living standards and social cohesion (World Bank 2012). Indeed, the recent recession came with
large job losses in some European countries, providing stark evidence that jobs and growth are
closely entwined (Figure 1). Yet, the relationship is far from straightforward. There are other
factors than the state of the business cycle that influence labor market outcomes. Economic
regulations and structural reforms affect whether businesses create jobs as the economy expands
and destroy jobs as the economy contracts (Orlandi 2012, Bernal-Verdugo, Furceri and Guillaume
2012a and 2012b). In particular, labor market regulations can influence whether employment or
wages adjust in response to output changes. For example, the Russian labor market appears to
respond to economic shocks through adjustments in wages rather than employment (Gimpelson
and Kapeliusshnikov 2011). Likewise, regulations affect whether enterprises can avoid layoffs
during recessions with adjustments in the intensive margin of the number of hours, as in some
European countries during the global financial crisis (Boysen-Hogrefe and Groll 2010).
Furthermore, the sectoral pattern matters (Arias-Vazquez et al 2012). In the 1990s, economic
growth in the transition countries of Eastern Europe was driven in large measure by higher labor
productivity rather than more employment, as countries moved resources across and within
sectors, invested in capital and adopted new technologies (World Bank 2008). In developing
economies, employment might be fairly stable over the business cycle due to large self-
employment, on-farm and informal work (Singh, Jain-Chandra and Mohommad 2012). In natural
resource rich countries, growth that is driven by higher commodity prices might generate little
employment by itself. These factors meant that for a long time the link from growth to
employment was tenuous in countries in Europe and Central Asia (ECA), giving rise to the notion of
jobless growth (World Bank 2005).
Figure 1: Change in employment and unemployment rates from 2007 to 2010
4
Change 2007 to 2010 (%)
0 -2 2
Peers ECA Peers WE Western Europe ECA
Unemployment Rate Employment Ratio
Note: See Box 1 for country groupings.
2
Box 1: Data and Groupings
Our dataset combines a number of sources. The ILO’s 7th edition of the Key Indicators of the Labor
Market (KILM) dataset provides labor market variables; the IMF’s April 2012 World Economic
Outlook dataset gives us information on GDP; and the World Bank’s 2012 version of the World
Development Indicators is the source for natural resource rents. We take structural reform
indicators from the following data sets: the World Bank’s Worldwide Governance Indicators for the
global sample; and the EBRD transition indicators for the ECA sample; the Fraser Institute’s
economic freedom indicators; the KOF Institute’s globalization indices; and MIT’s Economic
Complexity Observatory for a measure of export diversification.
The ECA sample includes 29 countries, although data is missing for some variables and years. The
global sample comprises 168 countries, which includes all countries for which labor market and
GDP data is available. The unit of analysis is the country. In ECA, we give the same weight to Russia,
a country with over 70 million employed workers, and Estonia, a country with less than 600,000
employed workers.
We often group countries to capture broad trends. We use a mainly geographic grouping for ECA.
We contrast the ECA countries to 19 Western European countries (WE) and to 100 ECA peers (Peers
of ECA), defined as low and middle income countries outside of Europe. We also look at 21 high-
income countries (Peers of WE) outside of Europe as peers for Western Europe. Within ECA, we
distinguish between the Balkan (countries of former Yugoslavia without Croatia), the EU11
countries (the ten EU member countries of Central Europe and Croatia, which will join the EU next
year), the Countries of Independent States (CIS), and Turkey. Finally, to make sure that group
averages are not driven by outliers or missing data, we generally report the median.
As Central and Eastern Europe entered its third decade of market-based economic reforms, this
study takes a new look at link of economic growth and job growth. It provides a cross-country
analysis of the employment intensity of growth over the last decade and a half in the region of
Eastern Europe and Central Asia (ECA). It makes the following contributions:
First, we look at trends in labor market outcomes in ECA. We establish that ECA looks worse in key
dimensions to the rest of the world. We also find that within ECA, labor market outcomes tend to
be worst for the Balkan and best for Commonwealth of Independent States (CIS) countries.
Second, we calculate elasticities of economic growth to employment. Using different econometric
specifications, we find that ECA’s elasticity (around 0.085 to 0.232) is noticeably less than in
Western Europe (0.407 to 0.494) but higher than for the ECA peers and the Western Europe peers.
Within ECA, it is highest among the EU10 countries and lowest among CIS countries.
Third, we find that the employment intensity of growth increased over time in ECA. During 2002 to
2007, it was about twice as high as during 1995 to 2001. The elasticities rose further during the
global financial crisis of 2008 to 2010, in line with the trends outside of Europe. Within ECA, the
increase over time is least pronounced for the CIS.
Fourth, we control for a number of factors that could affect the employment intensity of growth.
We look at labor market and product market regulation, governance, macroeconomic policy,
3
globalization, transition and diversification. We find that in many cases structural factors do not
have a statistical significant impact on employment growth. However, we uncover some evidence
that reforms of labor and product markets, economic policy frameworks, governance and
diversification are associated with higher employment growth.
Labor Market Trends
We assess trends in labor market performance with three indicators: labor force participation
rates; employment ratios; and unemployment rates (Box 2).
Box 2: Definitions of Labor Market Variables
The labor force participation rate is a measure of the proportion of a country’s working-age
population that engages actively in the labor market, either by working or looking for work. It
provides an indication of the relative size of the supply of labor available to engage in the
production of goods and services.
The employment-to-population ratio, or employment ratio, is defined as the proportion of a
country’s working-age population that is employed. It provides information on the ability of an
economy to create employment.
The unemployment rate gives us the proportion of the labor force that does not have a job and is
actively looking for work. It is probably the best-known labor market measure. However, the
unemployment rate is viewed to be more reliable as an indicator of unutilized labor supply in
developed countries than in developing countries due to informal work and underemployment.
Source: KILM Manuscript 2012.
Median labor force participation rate is low in both ECA and Western Europe compared to their
peers (Figure 2). But while Western Europe’s labor force participation peaked in the 2000s, ECA’s
labor force participation stayed low throughout, in spite of a modest improvement in the mid-
2000s. In 2010, ECA is the only region with labor force participation rates of less than 60 percent.
Within ECA, the Balkan and Turkey fare worst and the CIS best. However, labor force participation
rates in all four subregions were no higher in 2010 than in 1995.
4
Figure 2: Labor Force Participation Rates
70
65
Labor Force Participation
Labor Force Participation
65
60
60
55
55
50
1995 2000 2005 2010
Year 1995 2000 2005 2010
Year
ECA WE
Peers_ECA Peers_WE EU10 Balkan/Turkey CIS
Source: KILM, 7th Edition Source: KILM, 7th Edition
ECA is also the worst performing region with regard to employment ratios (Figure 3). Matching the
pattern in Western Europe and in contrast to the pattern elsewhere, employment ratios in ECA
declined noticeably as a result of the global financial crisis. In 2010, just over one in two working
age persons were employed. Within ECA, the picture is again worst in the Balkan and Turkey and
best in the CIS.
Figure 3: Employment Ratios
65
60 55
Employment Ratio
60
Employment Ratio
50
55
45
50
40
1995 2000 2005 2010
Year 1995 2000 2005 2010
Year
ECA WE
Peers_ECA Peers_WE EU10 Balkan/Turkey CIS
Source: KILM, 7th Edition Source: KILM, 7th Edition
Even through ECA's labor force participation rates are low, its unemployment rates are the highest
across the four regions (Figure 4). After converging to ECA peers in the years leading up to the
global financial crisis, unemployment rates increased above 10 percent in recent years.
Unemployment rates are especially high in Balkan/Turkey. In the CIS, unemployment rates were
noticeably lower in 2010 than in the late 1990s, while in the EU10, they increased to the levels of
the mid-1990s with the global financial crisis.
5
Figure 4: Unemployment Rates
15
25 20
Unemployment Rate
Unemployment Rate
10
10 15
5
5
0
0
1995 2000 2005 2010
Year 1995 2000 2005 2010
Year
ECA WE
Peers_ECA Peers_WE EU10 Balkan/Turkey CIS
Source: KILM, 7th Edition Source: KILM, 7th Edition
Employment Elasticity of Output
We will now shift our attention to yet another labor market indicator: the employment intensity of
growth, or, alternatively, the elasticity of employment with respect to output. This indicator
measures how employment changes with economic output. For example, a value between 0 and 1
implies that output growth of one percent is associated with positive employment growth of less
than one percent. Studying the employment elasticity serves two purposes. First, by evaluating
how employment growth evolves with output growth, it can shed light on structural changes in the
economy over time. Second, employment elasticities are closely linked to labor productivity
elasticities (Kapsos 2005). Defining labor productivity as output per employed, for small changes in
output, the labor productivity elasticity of output corresponds to one minus the employment
elastiticy of output. In case the employment elasticity is positive but less than one, the labor
productivity elasticity is also postivite and less than one.
We provide econometric evidence on the relationship between output and employment using a
panel of advanced and developing economies from 1995 to 2010. As estimates can be biased due
to problems of omitted avriables, endogeneity or measurement errors, we look at a range of
estimators. First, we calculate the average elasticities from 1995 to 2010 in the simplest fashion
using bivariate OLS (Table 1, Column 1). ECA’s elasticity is 0.18, noticeably less than in Western
Europe (0.44) but higher than for peers of ECA and peers of Western Europe. Within ECA, the
elasticities are highest among EU10 countries (0.32) and lowest in the CIS (0.12). Second, we
include country-fixed effects to capture country-specific factors, such as geography, and economic
and institutional environment, along with year indicators to control for general time effects
(Column 2). The estimated elasticities are somewhat smaller but the pattern across regions is
unchanged. Third, we run robust regressions that correct for outliers (Column 3). The coefficient
for ECA declines further, but the broad pattern is the same. Finally, we include lagged employment
on the right-hand side to capture the persistence in employment levels, and run system GMM
6
estimations (Column 4). The coefficient increases for ECA, but it remains less than half the level of
Western Europe.
The macroeconomic situation changed during 1995 to 2010. The transformation to a market
economy and the Asian crisis in the late 1990s gave way to an economic upswing in the early 2000s
until the global financial crisis hit in 2008. Separating three time periods (columns 5 to 16), we find
that the employment intensity of growth ECA doubled in ECA from 1995 to 2001 to 2002 to 2007,
and increased further in 2008 to 2010.The trends in other regions are less clearcut.
In Table 2, we distinguish three subregions within ECA: EU10, the CIS and the Balkan and Turkey.
There are two consistent patterns. Employment reponds most to output in the EU10, and the least
in the CIS (with the exception of the period of 1995 to 2001 in the Balkan and Turkey). Second,
elastiticies tend to increase over time for all three subregions.The only exception is the EU10,
where the coefficients remained stable or declined from 2002 to 2007 to 2008 to 2010.
Structural Indicator Trends
Aside from economic growth, there are many factors that are likely to affect the performance of
the labor market. We focus on those factors that we can trace consistently across time and country
groups. We distinguish five sets of indicators: labor markets (regulations and informality), product
markets, economic policy, governance and diversification. The variables are typically not available
for the full period. As an illustration, we briefly look at one variable for each dimension that will
turn out to be significant for the regression analysis:
• Hiring regulations improved in ECA during 2002 and 2008, driven by improvements in the EU10
and the CIS (Figure 5), before they worsened somewhat during the global financial crisis.
• ECA has a larger shadow economy than advanced regions. Its size declined somewhat due to
improvements especially in the CIS (Figure 6).
• ECA closed the gap to the doing business frontier over time, along with its peers, thanks to
improvements in all three subregions (Figure 7).
• ECA experienced rapid economic globalization, like Western Europe (Figure 8). Both regions
became more globalized than their peers over time. All three ECA subregions globalized
economically over time.
• ECA improved somewhat corruption control over time due to the Balkan and Turkey (Figure 9).
ECA’s export diversification was overall fairly stable (Figure 10). Since it declined among the peers
of Western Europe, ECA became the second most diversified region after Western Europe. Export
diversification increased in EU11 but declined in CIS.
7
Figure 5: Hiring regulations
9 10
10
9
3 4 5 6 7 8
8
Hiring regulations
3 4 5 6 7
Hiring regulations
2
2
1
1
0
0
2000 2002 2004 2006 2008 2010
Year 2000 2002 2004 2006 2008 2010
Year
ECA WE
Peers_ECA Peers_WE EU10 Balkan/Turkey CIS
Source: WB Doing Business, Fraser Institute Source: WB Doing Business, Fraser Institute
Figure 6: Shadow economy
70
50
Shadow Economy (% of GDP)
10 20 30 40 50 60
Shadow Economy (% of GDP)
20 30 40
0
10
1999 2001 2003 2005 2007
Year 1999 2001 2003 2005 2007
Year
ECA WE
Peers_ECA Peers_WE EU10 Balkan/Turkey CIS
Source: Schneider et al (2010) Source: Schneider et al (2010)
Figure 7: Ease of doing business
80
80
Distance to Doing Business Frontier
Distance to Doing Business Frontier
70
70
60
60
50
50
40
40
2005 2006 2007 2008 2009 2010
Year 2005 2006 2007 2008 2009 2010
Year
ECA WE
Peers_ECA Peers_WE EU11 Balkan/Turkey CIS
Source: WB Doing Business Source: WB Doing Business
8
Figure 8: Economic globalization flows
100
100
Economic Globalization Flows
Economic Globalization Flows
80
80
60
60
40
40
20
20
1995 2000 2005 2010
Year 1995 2000 2005 2010
Year
ECA WE
Peers_ECA Peers_WE EU10 Balkan/Turkey CIS
Source: KOF Index of Globalization Source: KOF Index of Globalization
Figure 9: Corruption control
2
2 1
1
Corruption control
Corruption control
0
0
-1
-1
-2
-2
1996 1998 2000 2002 2004 2006 2008 2010
Year 1996 1998 2000 2002 2004 2006 2008 2010
Year
ECA WE
Peers_ECA Peers_WE EU10 Balkan/Turkey CIS
Source: WB WGI Source: WB WGI
Figure 10: Export diversification
2
2
Export diversification
1
1
Export diversification
0
0
-1
-1
-2
-2
1995 1997 1999 2001 2003 2005 2007
Year 1995 1997 1999 2001 2003 2005 2007
Year
ECA WE
Peers_ECA Peers_WE EU11 Balkan/Turkey CIS
Source: MIT Source: MIT
9
Employment Elasticity and Structural Reform
Structural factors are likely to influence the relationship between employment and output change.
For example, firms might be more readily responding to economic expansions in case workers can
be hired and layed off at low cost. More broadly, labor market institutions, product market
reforms, economic policy, governance and economic diversification could all affect the
employment elasticity of growth. There are a large number of structural variables that could be
included in the regressions, and parameter estimates could turn out to be sensitive to the selection
of variables. In the following, we use a simple approach. We include variables separately as right-
hand side regressors, along with growth and fixed year and country effects. Even with this
simplified approach, structural variables are often not significant. However, we find a number of
statistically significant relationships where structural reforms boost employment growth:
• On labor markets, employment growth increases with less hiring regulations in ECA and the
EU10, and a smaller shadow economy in the EU10 and the Balkan and Turkey (Table 3).
• On product markets, employment growth increases with a greater ease of doing business in
ECA and the CIS and less credit regulations in ECA (Table 4). With regard to EBRD transition
indicators, we find that employment growth increases with more firm restructuring in ECA and
the EU10, better competition policy in the CIS, more banking reform in the EU10, and more
infrastructure reform in the CIS (Table 5).
• On economic policy, employment growth increases with smaller size of government in ECA,
sound money in ECA, and greater economic flows across borders in ECA, the EU10 and the CIS
(Table 6).
• On governance, employment growth increases with more corruption control in ECA and the
CIS, more political stability in the CIS, better quality of regulation in ECA (but not in the EU10),
more government effectiveness in ECA and the CIS, and more voice and accountability in ECA
(Table 7).
• On economic diversification, employment growth increases with higher export diversification in
ECA. It also decreases with higher natural resource rents in ECA and in the CIS, but the
coefficients are not statistically significant (Table 8). In addition, we look at how sectoral
employment responds to overall output growth. The impact is insignificant on agricultural
employment. Industrial employment increases with economic growth in ECA and the EU10.
Service sector employment increases with economic growth in ECA, the EU10 and the CIS.
Conclusions
The sluggish response of employment growth to the post-transition recovery in ECA countries gave
rise to the notion of jobless growth in the 1990s. In this paper, we provide evidence that the
responsiveness of employment to output increased in the second decade of the transition. In other
words, employment decisions of businesses depend more on the state of the business cycle than in
the past, perhaps because structural reforms have increased the flexibility of the labor market (IMF
10
2012a and 2012b, Ball, Leigh and Loungani 2013). The changes are most pronounced for the EU10
and Balkan and Turkey, while the employment elasticity remained low in the CIS even during the
global financial crisis. In addition, we provide some evidence that employment growth increases
with reforms of labor and product markets, stronger macroeconomic policy frameworks, better
governance, and more economic integration and economic diversification (Crivelli, Furceri and
Toujas-Bernate 2012). However, in many instances, structural factors appear to have no significant
impact on employment growth. This suggests that job dynamics depend on the interplay of a
number of factors, and reforms in one area might not improve labor market outcomes due to the
interaction with other areas (Blanchard, Jaumotte and Loungani 2013; World Development Report
2012).
These findings imply two policy conclusions. First, the state of the business cycle, including the
strength of aggregate demand, matters for labor market outcomes in many ECA countries. This
suggests that weak labor market trends are foremost a result of weak growth, especially in the
short term. Second, structural reforms are likely to improve employment growth although the
impact might often be visible only over time.
11
Table 1: Regression results for employment elasticity of growth: world
Variables 1995 to 2010 1995 to 2001 2002 to 2007 2008 to 2010
(1) (2) (3) (4) (5) (6) (7) (8) (9) (10) (11) (12) (13) (14) (15)
OLS FOLS ROLS SGMM OLS FOLS ROLS SGMM OLS FOLS ROLS SGMM OLS FOLS ROLS
ECA
Output coeff. 0.181 0.120 0.085 0.232 0.074 0.075 0.058 0.062 0.140 0.124 0.060 0.128 0.325 0.221 0.201
Output t-stats. 7.2 3.7 3.6 4.5 1.9 2.0 1.9 1.1 2.5 1.4 1.4 1.3 6.8 2.4 3.9
Number of obs. 413 413 413 413 161 161 161 161 168 168 168 140 84 84 84
R squared 0.111 0.289 0.472 0.023 0.433 0.555 0.036 0.343 0.697 0.358 0.689 0.883
AB AR2 test (p value) 0.411 0.913 0.087
Hansen J statistics (p value) 0.942 0.231 0.125
Western Europe
Output coeff. 0.438 0.407 0.494 0.472 0.538 0.320 0.331 0.516 0.448 0.428 0.536 0.572 0.333 0.028
Output t-stats. 13.7 7.1 11.1 4.1 8.4 3.3 3.3 5.1 6.9 4.2 5.4 4.5 3.7 0.2
Number of obs. 285 285 285 285 114 114 114 114 114 114 114 95 57 57
R squared 0.397 0.521 0.681 0.388 0.649 0.647 0.300 0.630 0.680 0.201 0.669
AB AR2 test (p value) 0.154 0.232 0.614
Hansen J statistics (p value) 0.997 0.316 0.508
Peers of ECA
Output coeff. 0.046 0.032 0.019 0.103 0.049 0.022 -0.004 0.038 0.028 0.013 0.015 0.062 0.083 0.065 0.004
Output t-stats. 4.2 2.7 3.1 2.6 2.8 1.1 -0.5 0.7 1.6 0.6 1.8 0.6 3.2 1.9 0.9
Number of obs. 1462 1462 1462 1462 566 566 566 566 596 596 596 498 300 300 300
R squared 0.012 0.239 0.587 0.014 0.325 0.872 0.004 0.352 0.810 0.033 0.500 0.986
AB AR2 test (p value) 0.485 0.228 0.449
Hansen J statistics (p value) 0.243 0.119 0.977
Peers of Western Europe
Output coeff. 0.070 0.038 -0.008 -0.013 0.007 -0.007 -0.014 -0.010 0.432 0.284 0.266 0.036 0.591 0.087
Output t-stats. 3.3 1.8 -0.8 -0.4 0.6 -0.5 -1.3 -1.2 5.9 3.8 10.5 0.1 5.6 0.9
Number of obs. 310 310 310 310 121 121 121 121 126 126 126 105 63 63
R squared 0.033 0.523 0.538 0.003 0.556 0.685 0.219 0.710 0.951 0.340 0.914
AB AR2 test (p value) 0.52 0.17 0.326
Hansen J statistics (p value) 0.996 0.423 0.175
Note: The dependent variable is employment growth. The output coefficient is the parameter estimate of output growth.
OLS stands for cross-sectional bivariate regressions. FOLS include in addition country and year fixed-effects.
ROLS stands for robust regressions. SGMM stands for system generalized method of moments regressions.
The null hypothesis of the Arellano-Bond AR(2) test is that the first-differenced errors exhibit no second-order serial correlation.
The null hypothesis of the Hansen J-statistics is that the instruments are not correlated with the residuals.
12
Table 2: Regression results for employment elastiticy of growth: ECA
Variables 1995 to 2010 1995 to 2001 2002 to 2007 2008 to 2010
(1) (2) (3) (4) (5) (6) (7) (8) (9) (10) (11) (12) (13) (14)
OLS FOLS ROLS SGMM1 OLS FOLS ROLS SGMM1 OLS FOLS ROLS SGMM1 OLS FOLS
EU10
Output coeff. 0.324 0.240 0.251 0.369 -0.022 0.022 0.049 0.023 0.467 0.431 0.414 0.835 0.421 0.297
Output t-stats. 7.7 3.9 4.5 5.6 -0.3 0.2 0.5 0.1 3.7 2.1 2.9 5.4 6.0 1.8
Number of obs. 150 150 150 150 60 60 60 60 60 60 60 50 30 30
R squared 0.288 0.483 0.525 0.001 0.361 0.381 0.194 0.523 0.636 0.566 0.779
AB AR2 test (p value) 0.375 0.57 0.048
Hansen J statistics (p value) 1 0.782 0.749
Balkan and Turkey
Output coeff. 0.169 0.092 0.145 0.251 -0.006 -0.087 -0.030 0.151 0.264 0.382 0.425 0.895 0.597 0.687
Output t-stats. 1.6 0.7 1.5 1.5 -0.1 -0.8 -0.6 1.0 0.7 0.8 1.6 1.6 3.2 2.4
Number of obs. 94 94 94 94 31 31 31 31 42 42 42 35 21 21
R squared 0.028 0.271 0.402 0.000 0.514 0.897 0.013 0.337 0.494 0.352 0.770
AB AR2 test (p value) 0.377 0.141 0.542
Hansen J statistics (p value) 1 0.999 0.917
CIS
Output coeff. 0.124 0.082 0.063 0.091 0.110 0.083 0.039 0.055 0.145 0.081 0.049 0.114 0.176 0.126
Output t-stats. 4.5 2.6 2.9 2.2 2.4 1.8 1.6 0.7 2.6 1.2 1.7 2.0 3.3 2.2
Number of obs. 180 180 180 180 72 72 72 72 72 72 72 60 36 36
R squared 0.102 0.498 0.754 0.074 0.550 0.822 0.087 0.542 0.896 0.240 0.859
AB AR2 test (p value) 0.912 0.734 0.187
Hansen J statistics (p value) 1 0.734 0.597
Note: The dependent variable is employment growth. The output coefficient is the parameter estimate of output growth.
OLS stands for cross-sectional bivariate regressions. FOLS include in addition country and year fixed-effects.
ROLS stands for robust regressions. SGMM stands for system generalized method of moments regressions.
The null hypothesis of the Arellano-Bond AR(2) test is that the first-differenced errors exhibit no second-order serial correlation.
The null hypothesis of the Hansen J-statistics is that the instruments are not correlated with the residuals.
13
Table 3: Fixed-effects regression results for employment elastiticy of growth: labor market
Variables ECA WE PECA PWE EU10 BalTur CIS
(1) (2) (3) (4) (5) (6) (7)
Hours regulations
Output Coeff. 0.255 0.351 0.088 0.093 0.465 0.502 0.103
t-stat. 4.3 5.0 4.1 1.7 5.7 1.8 1.8
Other variable Coeff. 0.001 0.001 0.000 -0.001 0.000 0.014 0.003
t-stat. 0.5 0.5 0.3 -0.5 -0.2 1.0 1.1
N 232 196 746 136 100 63 76
R2 0.36 0.53 0.25 0.87 0.63 0.37 0.60
Hiring regulations
Output Coeff. 0.252 0.356 0.087 0.092 0.476 0.451 0.104
t-stat. 4.4 5.1 4.1 1.7 6.1 1.6 1.8
Other variable Coeff. 0.003 0.000 0.000 -0.006 0.004 0.002 0.001
t-stat. 1.7 0.0 -0.5 -1.2 2.3 0.3 0.2
N 242 196 767 134 110 63 76
R2 0.37 0.53 0.25 0.88 0.64 0.36 0.59
Size of shadow economy
Output Coeff. 0.103 0.426 0.012 0.043 0.134 -0.481 0.148
t-stat. 1.3 5.5 0.7 1.3 0.9 -1.7 2.0
Other variable Coeff. -0.001 -0.005 -0.001 -0.010 -0.010 -0.051 0.000
t-stat. -0.2 -1.7 -1.7 -10.0 -1.5 -2.7 0.1
N 223 171 842 172 89 45 89
R2 0.24 0.66 0.28 0.77 0.46 0.36 0.49
Informality
Output Coeff. 0.080 0.368 0.032 0.064 0.168 -0.143 0.104
t-stat. 1.9 5.5 2.3 2.2 2.2 -0.7 2.2
Other variable Coeff. 0.033 1.183 0.037 -0.136 -0.488 1.862 0.030
t-stat. 0.4 4.2 1.0 -1.5 -2.1 1.0 0.4
N 301 265 1213 261 134 50 117
R2 0.22 0.56 0.22 0.51 0.38 0.35 0.43
Note: The dependent variable is employment growth. The output coefficient is the
parameter estimate of output growth. All regressions also include country and year
fixed-effects.
14
Table 4: Fixed-effects regression results for employment elasticity of growth: product market
Variables ECA WE PECA PWE EU10 BalTur CIS
(1) (2) (3) (4) (5) (6) (7)
Distance to frontier
Output Coeff. 0.209 0.295 0.015 0.048 0.391 0.539 0.062
t-stat. 3.3 2.9 0.6 0.8 4.9 1.5 2.0
Other variable Coeff. 0.002 0.000 0.001 0.001 0.003 -0.005 0.001
t-stat. 1.9 0.2 2.4 0.8 1.2 -1.2 2.0
N 162 109 586 107 60 41 66
R2 0.45 0.59 0.34 0.92 0.82 0.42 0.83
Credit regulation
Output Coeff. 0.232 0.332 0.068 0.122 0.434 -0.068 0.131
t-stat. 4.2 4.6 3.5 1.6 5.5 -0.3 1.8
Other variable Coeff. 0.006 0.003 0.000 0.001 0.003 0.013 0.003
t-stat. 2.0 1.5 0.1 0.3 0.9 1.5 0.7
N 210 209 810 175 110 44 56
R2 0.37 0.53 0.24 0.61 0.62 0.38 0.51
Business regulation
Output Coeff. 0.354 0.413 0.120 0.144 0.422
t-stat. 5.5 6.3 3.6 2.9 5.3
Other variable Coeff. 0.009 0.003 0.002 -0.003 0.008
t-stat. 1.5 0.7 0.7 -0.8 1.1
N 143 187 362 99 110
R2 0.53 0.60 0.24 0.72 0.62
Note: The dependent variable is employment growth. The output coefficient is the
parameter estimate of output growth. All regressions also include country and year
fixed-effects.
15
Table 5: Fixed-effects regression results for employment elasticity of growth: ECA product market
Variables ECA EU10 BalTur CIS
(1) (2) (3) (4)
Firm restructuring
Output Coeff. 0.125 0.219 0.073 0.079
t-stat. 3.7 3.3 0.4 2.6
Other variable Coeff. 0.015 0.032 -0.024 0.010
t-stat. 1.8 1.9 -0.7 1.1
N 386 135 71 180
R2 0.30 0.51 0.29 0.50
Competition policy
Output Coeff. 0.13 0.24 0.06 0.08
t-stat. 3.78 3.40 0.32 2.72
Other variable Coeff. 0.01 0.00 -0.02 0.02
t-stat. 1.3 0.3 -0.7 2.4
N 386 135 71 180
R2 0.30 0.49 0.29 0.52
Banking reform
Output Coeff. 0.124 0.202 0.068 0.083
t-stat. 3.6 3.0 0.3 2.7
Other variable Coeff. 0.008 0.022 -0.011 0.003
t-stat. 1.3 2.1 -0.5 0.4
N 386 135 71 180
R2 0.30 0.51 0.28 0.50
Infrastructure reform
Output Coeff. 0.125 0.261 0.083 0.092
t-stat. 3.6 3.8 0.4 2.9
Other variable Coeff. 0.005 -0.018 0.005 0.018
t-stat. 0.7 -1.7 0.1 1.8
N 386 135 71 180
R2 0.29 0.50 0.28 0.51
Note: The dependent variable is employment growth.
The output coefficient is the parameter estimate of output growth.
All regressions also include country and year fixed-effects.
16
Table 6: Fixed-effects regression results for employment elasticity of growth: economic policy
Variables ECA WE PECA PWE EU10 BalTur CIS
(1) (2) (3) (4) (5) (6) (7)
Size of government
Output Coeff. 0.300 0.350 0.075 0.123 0.398 -0.146 0.133
t-stat. 4.3 5.0 3.9 1.7 4.7 -0.7 0.8
Other variable Coeff. 0.008 0.002 0.002 0.014 0.006 -0.011 0.003
t-stat. 1.9 1.0 2.5 2.5 1.2 -0.7 0.5
N 165 209 780 175 110 33 22
R2 0.46 0.52 0.26 0.63 0.62 0.57 0.76
Sound money
Output Coeff. 0.301 0.370 0.074 0.121 0.414 -0.153 0.051
t-stat. 4.4 5.3 3.8 1.6 5.4 -0.7 0.3
Other variable Coeff. 0.004 -0.005 -0.001 -0.001 0.006 0.005 0.001
t-stat. 1.8 -1.3 -1.2 -0.1 2.2 0.8 0.2
N 165 209 780 175 110 33 22
R2 0.46 0.53 0.26 0.61 0.63 0.57 0.75
Economic globalization flows
Output Coeff. 0.118 0.452 0.031 0.010 0.300 -0.047 0.090
t-stat. 3.5 6.8 2.4 0.6 5.1 -0.3 2.6
Other variable Coeff. 0.001 0.000 0.000 0.000 0.001 0.001 0.001
t-stat. 3.9 1.1 1.2 0.3 2.2 0.7 2.0
N 357 266 1316 247 126 77 154
R2 0.32 0.53 0.23 0.36 0.62 0.25 0.49
Economic globalization restrictions
Output Coeff. 0.171 0.462 0.072 0.219 0.332 -0.035 0.145
t-stat. 4.1 7.1 4.2 3.3 5.4 -0.2 2.7
Other variable Coeff. 0.000 0.000 0.000 0.000 0.000 0.000 0.000
t-stat. 0.7 -1.5 1.7 -0.3 -0.4 0.4 -0.7
N 315 266 1068 219 126 77 112
R2 0.29 0.53 0.20 0.55 0.60 0.24 0.45
Note: The dependent variable is employment growth. The output coefficient is the
parameter estimate of output growth. All regressions also include country and year
fixed-effects.
17
Table 7: Fixed-effects regression results for employment elasticity of growth: governance
Variables ECA WE PECA PWE EU10 BalTur CIS
(1) (2) (3) (4) (5) (6) (7)
Rule of law
Output Coeff. 0.140 0.398 0.028 0.170 0.387 0.232 0.069
t-stat. 3.6 6.2 2.0 4.0 6.2 0.9 1.9
Other variable Coeff. -0.001 -0.005 0.004 -0.015 -0.002 -0.042 0.009
t-stat. -0.1 -0.5 1.3 -1.1 -0.2 -0.9 1.0
N 320 228 1145 249 108 68 144
R2 0.31 0.52 0.24 0.61 0.70 0.29 0.49
Corruption control
Output Coeff. 0.149 0.396 0.029 0.144 0.387 0.217 0.078
t-stat. 3.8 6.1 2.1 3.5 6.3 0.8 2.2
Other variable Coeff. 0.019 0.004 0.001 0.020 0.004 0.031 0.019
t-stat. 2.4 0.5 0.5 2.2 0.4 0.9 2.3
N 320 228 1144 249 108 68 144
R2 0.33 0.52 0.23 0.62 0.70 0.28 0.51
Political stability
Output Coeff. 0.134 0.379 0.032 0.160 0.431 0.305 0.070
t-stat. 3.4 5.8 2.2 3.9 6.0 1.1 2.0
Other variable Coeff. 0.008 0.009 -0.001 0.002 -0.013 -0.024 0.011
t-stat. 1.2 1.5 -0.8 0.2 -1.1 -0.9 1.7
N 320 228 1146 249 108 68 144
R2 0.31 0.52 0.23 0.61 0.70 0.29 0.50
Quality of regulations
Output Coeff. 0.137 0.405 0.028 0.152 0.435 0.243 0.071
t-stat. 3.5 6.1 2.0 3.7 6.8 0.9 2.0
Other variable Coeff. 0.014 -0.004 0.004 0.025 -0.033 0.031 0.005
t-stat. 1.6 -0.5 1.7 2.1 -2.2 0.7 0.6
N 320 228 1145 249 108 68 144
R2 0.32 0.52 0.24 0.62 0.72 0.28 0.49
Government effectiveness
Output Coeff. 0.139 0.365 0.027 0.178 0.390 0.350 0.063
t-stat. 3.6 5.7 1.9 4.2 6.3 1.3 1.8
Other variable Coeff. 0.019 0.019 0.006 -0.021 -0.002 -0.080 0.024
t-stat. 2.1 2.9 1.9 -1.5 -0.1 -1.4 2.9
N 320 228 1144 249 108 68 144
R2 0.32 0.54 0.24 0.61 0.70 0.30 0.52
Voice and accountability
Output Coeff. 0.152 0.395 0.029 0.161 0.390 0.255 0.077
t-stat. 3.8 6.1 2.0 3.9 6.4 0.9 2.2
Other variable Coeff. 0.014 0.005 0.002 -0.013 -0.028 0.008 0.011
t-stat. 1.6 0.3 0.9 -0.8 -1.2 0.2 1.2
N 320 228 1146 249 108 68 144
R2 0.32 0.52 0.23 0.61 0.71 0.27 0.49
Note: The dependent variable is employment growth. The output coefficient is the
parameter estimate of output growth. All regressions also include country and year
fixed-effects.
18
Table 8: Fixed-effects regression results for employment elasticity of growth: diversification
Variables ECA WE PECA PWE EU10 BalTur CIS
(1) (2) (3) (4) (5) (6) (7)
Export diversification
Output Coeff. 0.081 0.519 0.046 0.313 0.110 -0.058 0.073
t-stat. 2.2 8.0 2.9 4.6 1.5 -0.3 2.2
Other variable Coeff. 0.014 0.012 -0.003 0.011 0.012 -0.026 0.000
t-stat. 2.0 1.6 -1.2 1.1 0.6 -0.5 0.0
N 337 208 873 195 130 64 143
R2 0.25 0.60 0.22 0.60 0.35 0.26 0.53
Natural resource rent
Output Coeff. 0.125 0.408 0.029 0.037 0.234 0.027 0.084
t-stat. 3.8 7.1 2.4 1.8 3.7 0.2 2.7
Other variable Coeff. 0.000 0.000 0.000 0.000 0.002 0.004 0.000
t-stat. -1.3 0.4 0.5 0.2 0.7 0.9 -0.8
N 413 285 1435 304 150 83 180
R2 0.29 0.52 0.23 0.52 0.49 0.26 0.50
Agricultural employment
Output Coeff. -0.007 0.282 0.959 0.569 -0.062 0.036 0.174
t-stat. 0.0 0.7 0.2 0.5 -0.3 0.1 0.8
N 283 281 425 177 143 47 93
R2 -0.02 0.00 -0.05 0.00 0.23 0.29 0.33
Industrial employment
Output Coeff. 0.328 0.899 0.762 0.512 0.624 0.034 0.057
t-stat. 2.4 6.6 4.0 3.2 6.9 0.0 0.3
N 283 281 425 177 143 47 93
R2 0.16 0.43 0.22 0.74 0.69 0.35 0.41
Service sector employment
Output Coeff. 0.649 0.153 0.166 0.241 0.292 -0.208 1.412
t-stat. 2.4 1.7 0.9 2.8 5.2 -0.6 1.9
N 283 281 425 177 143 47 93
R2 0.00 0.17 0.01 0.42 0.45 0.43 0.26
Note: The dependent variable is employment growth. The output coefficient is the
parameter estimate of output growth. All regressions also include country and year
fixed-effects.
19
References
Arias, Omar S., Carolina Sánchez-Páramo, María E. Dávalos, Indhira Santos, Erwin R. Tiongson, Carola Grun,
Natasha de Andrade Falcão, Gady Saiovici, and Cesar A. Cancho. 2014. Back to Work: Growing with Jobs in
Europe and Central Asia. Washington, DC: World Bank.
Arias-Vazquez, Francisco Javier, Jean Lee, David Newhouse and David Robalino. 2012. The Role of Sectoral
Growth Patterns in Labor Market Development. World Bank Transcript.
Ball, Laurence, Daniel Leigh and Prakash Loungani. 2013. Okun’s Law: Fit at 50? NBER Working Paper No.
18668.
Bernal-Verdugo, Lorenzo E., Davide Furceri, and Dominique Guillaume. 2012a. Labor Market Flexibility and
Unemployment: New Empirical Evidence of Static and Dynamic Effects. IMF Working Papers 12/64.
Bernal-Verdugo, Lorenzo E., Davide Furceri, and Dominique M. Guillaume. 2012b. Crises, Labor Market
Policy, and Unemployment. IMF Working Papers 12/65.
Blanchard, Olivier, Florence Jaumotte, and Prakash Loungani. 2013. Labor Market Policies and IMF Advice in
Advanced Economies During the Great Recession. IMF Staff Discussion Note SDN/13/02.
Boysen-Hogrefe, Jens and Dominik Groll. 2010. The German Labour Market Miracle. National Institute
Economic Review October 2010 vol. 214 no. 1.
Furceri, Davide, Ernesto Crivelli and Joël Toujas-Bernate. 2012. Can Policies Affect Employment Intensity of
Growth? A Cross-Country Analysis. IMF Working Papers 12/218.
Gimpelson, Vladimir and Rostislav Kapeliusshnikov. 2011. Labor Market Adjustment: Is Russia Different?
Higher School of Economics Working Paper WP3/2011/04, Series WP3.
International Monetary Fund. 2012a. Unemployment Dynamics during Recessions and Recoveries: Okun’s
Law and Beyond. World Economic Outlook April 2012, Chapter 3.
International Monetary Fund. 2012b. Resilience in Emerging Market and Developing Economies: Will it Last?
World Economic Outlook October 2012, Chapter 4.
Kapsos, Steven. 2005. The Employment Intensity of Growth: Trends and macroeconomic determinants.
Employment Strategy Papers 2005/12. International Labor Organization.
Orlandi, Fabrice. 2012. Structural Unemployment and its Determinants in the EU Countries. Economic Papers
455, May 2012.
Singh, Anoop, Sonali Jain-Chandra, and Adil Mohommad. 2012. Inclusive Growth, Institutions, and the
Underground Economy. IMF Working Papers 12/47.
World Bank 2005. Enhancing Job Opportunities. Washington, DC: World Bank.
World Bank 2008. Unleashing Productivity: Productivity Growth in Eastern Europe and the Former Soviet
Union. Washington, DC: World Bank.
World Bank 2012. Jobs. World Development Report. Washington, DC: World Bank.
20