don’t trust anybody over 30: youth unemployment and okun’s ... · oliver hutengs*, georg...

Bank i Kredyt 45(1) , 2014, 1–16

Don’t trust anybody over 30: youth unemployment and Okun’s law in CEE countries

Oliver Hutengs*, Georg Stadtmann#

Submitted: 12 March 2013. Accepted: 8 August 2013.

AbstractIn recent years youth unemployment rates across Europe soared, causing the European Commission to take actions through initiatives to counter this development. This article examines youth unemployment development in selected CEE countries and compares them to the EU-15. We use Okun’s law and estimate age and country specific Okun coefficients for five different age cohorts. Our results show that young people display much higher Okun coefficients than their older peers, thus confirming that young people are more prone to macroeconomic shocks. This result might be a justification for additional governmental intervention and active labour market policies favouring young people.

Keywords: Okun’s law, labour market, youth unemployment

JEL: E24, F50, C23

* Europa-Universität Viadrina, Lehrstuhl für Volkswirtschaftslehre, Wirtschaftswissenschaftliche Fakultät, insb. Makroökonomik.

# University of Southern Denmark, Department of Business and Economics; Europa-Universität Viadrina, Lehrstuhl für Volkswirtschaftslehre, insb. Makroökonomik; e-mail: [email protected].

O. Hutengs, G. Stadtmann2

1. Introduction

In 2009 youth unemployment has been identified by the European Commission (EC) as one of the most urgent problems to solve. As a consequence, the EC outlined a strategy to address major problems concerning labour market participation of young people, including education and training deficits as well as lower employment chances (EC 2009). Recently, the European Commission (EC 2012) proposed further actions. Each member state is supposed to guarantee young people an offer of employment or continued education (i.e. apprenticeship or traineeship) supported by financial means of the European Union.

In light of this recent debate, the paper examines the unemployment development in selected central and eastern European countries (CEE countries) and the EU-151 countries as an aggregate. Using Okun’s law (Okun 1962) we estimate age-cohort specific Okun coefficients showing that young people are significantly more exposed to business-cycle fluctuations than older ones.

The remainder of the paper is structured as follows: Section 2 reviews some important aspects of the literature. Section 3 provides a general description of the dataset used, as well as descriptive statistics regarding youth versus total unemployment. Section 4 discusses the regression approach and results. Section 5 concludes.

2. Literature review

In economic theory the link between unemployment rate and GDP growth can be explained from a demand side perspective, that is, changes in aggregate demand cause firms to adapt their respective production goals. This will affect demand for labour and ultimately the unemployment rate (Sögner, Stiassny 2002). Following this argument, a negative GDP shock reduces demand for labour and ceteris paribus increases unemployment. According to O’Higgins (1997), this argumentation also holds for young people. It is true that the level of the youth unemployment rate heavily depends on country specific youth labour market conditions such as the size of skill mismatches and school-to- -work transition problems (Dietrich 2012). Nonetheless, changes in the youth unemployment rate are predominantly driven by business-cycle fluctuations. We will show later that young people indeed react more strongly to business-cycle fluctuations than adults. O’Higgins (1997) argues that the reason for this behaviour lies in company’s lower opportunity cost of firing young workers, i.e. they possess fewer firm-related skills and employment protection through legislation compared to adults. Any argument that only considers structural (i.e. skill mismatches) or frictional youth unemployment rests solely on the assumption that the national labour market follows a neoclassical point of view where labour markets are always cleared (Sögner, Stiassny 2002). Real world labour markets are not exclusively characterised by structural and frictional unemployment but also by cyclical unemployment.

O’Higgins (2003) discusses the argument that youth unemployment is of shorter duration and thus less problematic. Nevertheless, he believes that high youth unemployment rates are due to young people more often quitting their jobs and looking for better and more suitable occupations (O’Higgins 1997). Ryan (2001) provides a similar argument, i.e. young people show higher flows into and out

1 EU-15 includes all countries which were members of the European Union before the eastern enlargement in May 2004. For a country specific analysis of major EU-15 countries see Hutengs and Stadtmann (2013).

Don’t trust anybody over 30... 3

of unemployment leading to less long-term unemployment among young people. According to column 1, Table 1, about one third of the youth unemployment in CEE countries in the year 2012 can be categorized as long-term unemployment, i.e. more than 12 months. In Slovakia it is even more than half. There is a noticeable difference in long-term unemployment between young and adult people. Adults are significantly more affected by long-term unemployment than young people (column 3). Nevertheless, we argue that the difference itself does not make youth unemployment less problematic. In fact, unemployment in young years has profound negative effects on human capital accumulation leading to lower earnings in the future (Mroz, Savage 2006). The European Economic Advisory Group (2013, p. 86) makes a similar argument. In addition, Bell and Blanchflower (2011) argue that youth unemployment today will incur even higher social costs in the future, negatively impacting well-being and job satisfaction. From this point of view, combating youth unemployment is important despite lower long-term unemployment rates.

3. Dataset and descriptive statistics

In the following chapter we discuss our data set as well as recent unemployment developments in CEE countries. The dataset used to estimate Okun’s law consists of annual GDP data published in the Annual Macro-Economic Database (AMECO) of the European Commission (EC 2013) and the unemployment rate for various age cohorts provided by the Organisation for Economic Co-operation and Development (OECD 2013). The joint dataset uses the earliest available entries for each country and ends in 2011.2 The dataset contains two main variables for each country: real GDP measured in prices of the year 2005 and the unemployment rate. The latter is based on International Labour Organisation (ILO) standards ensuring comparability among different countries.

A first glance on the youth unemployment trends, i.e. people between the age of 15 and 24 years, among CEE countries reveals a rather diverse picture.

1. According to Figure 1, youth unemployment rates in the Czech Republic and Hungary reached their maximum (close to 20%) in recent years, whereas Estonia shows a decreasing rate after surpassing its all time high of over 30% in 2010. In general, all CEE countries experience relatively high youth unemployment rates recently, whereas EU-15 youth unemployment is lower. Furthermore, all CEE countries show a relatively high fluctuation over the whole period.

2. One distinctive pattern is the decrease in unemployment figures in most CEE countries after the accession to the EU in 2004. The decrease may not only be attributed to high GDP growth but also to high emigration from CEE countries (Baas, Brücker 2011). The United Kingdom and Ireland had opened up their respective national labour markets right from the start of the accession date, thus enabling workers from CEE countries to move and work there without major barriers (Barrell, Fitzgerald, Riley 2010). The recently observed increases in unemployment figures can clearly be attributed to the financial and economic crisis of 2007, slowing and partially reversing GDP growth tremendously.

3. If we assume that a change in the unemployment rate is directly related to the change in GDP, we can infer from Figure 1 that Poland and Slovakia stand out by having seemingly identical business-

2 Selection of countries is determined by data availability. Countries used and dataset starting dates with number of observations per cohort in parenthesis: Czech Republic − 1994 (18), Estonia − 1994 (18), Hungary − 1993 (19), Poland − 1993 (19), Slovakia − 1995 (17) and EU-15 − 1981 (31).


-cycles whereas the Hungarian cycle does not seem to follow the Polish/Slovak pattern. Different business-cycles among CEE countries could pose a serious problem when these countries were about to join the Eurozone, because the likelihood of the occurrence of asymmetric shocks is relatively high. The business-cycle literature tends to focus more on differences between CEE countries and the Eurozone rather than on differences among CEE countries itself. Though, there is not much empirical support in the literature regarding our specific observation. Nonetheless, a recent study (Gächter, Riedl, Ritzberger-Grünwald 2013) shows that heterogeneity among CEE countries business-cycles is more pronounced as among Eurozone countries. Furthermore, they find a decoupling of CEE countries and Eurozone business-cycles starting with the recent financial crisis. Interestingly, Furceri and Zdzienicka (2011) show that the currently existing flexible exchange rate systems have helped to mitigate negative effects brought by the financial crisis in the short-run. Blanchard, Amighini, Giavazzi (2011) show that Poland depreciated the zloty by over 15% from 2007 to 2009. In fact, Poland was the only EU country without negative GDP growth during the financial crisis.3 Interestingly, Poland’s unemployment rate is on the rise despite prolonged positive GDP growth during the crisis and has yet to show signs of reversal. The spike in Polish youth and adult unemployment seems to be due to migration back into Poland. Polakowski (2012) argues that a part of Polish emigrants turned home because of deteriorating labour markets abroad leading to higher unemployment in Poland.

In Table 2 we characterize the difference between youth unemployment and total unemployment over time. In general, youth unemployment rates are higher than total unemployment rates across all countries. This is to be expected, as young people often show a lack of required skills and working experience, leading to fewer employment chances for this cohort (OECD 2010). Moreover, labour market institutions directly influence the level and difference between both rates. The European Economic Advisory Group (2013, p. 86) argues that a minimum wage adversely affects young people by artificially lowering pay differences between younger and older workers, thus directly decreasing young people’s employment chances during a recession.

CEE countries like Poland and Slovakia show a very large gap (over 14 percentage points) between both rates. In contrast to this, the Czech Republic, Estonia and Hungary tend to have smaller gaps on average (below 10 percentage points), which are even lower than in the EU-15. The level of the unemployment gap between youth and adult unemployment can be explained by different institutional arrangements on the respective national labour markets. Summary statistics show that these three countries’ average unemployment figures resemble those of the EU-15. Poland and Slovakia substantially differ from the others because all the unemployment indicators show higher values. In contrast, the Czech Republic shows the lowest unemployment figures, having even lower values than the EU-15 across all indicators.

All CEE countries with the exception of Hungary show a specific pattern regarding the size of this gap. The gap starts to increase in the late 1990s reaching its maximum in the first years of the 2000s and beginning to level off shortly before EU accession in May 2004. The gap stopped closing after the outbreak of the financial crisis in 2007, which led to a decrease in GDP growth rates. Since then, all CEE countries experienced increases in the total unemployment rate mirrored by even sharper

3 Poland seems to be very special in this regard. Blanchard, Amighini, Giavazzi (2011) recognize this event in their macro-economics textbook as a study example while comparing the Polish and Hungarian performance during the crisis. Poland achieved positive GDP growth in that time through a joint application of expansionary fiscal (tax cuts) and monetary policy (leading to a depreciated zloty).


rises in youth unemployment. Estonia was hit very hard by the crisis showing huge increases in both unemployment rates hitting all-time highs in the year 2010. The increase in Poland was much lower than observed elsewhere favoured by positive albeit low GDP growth in the crisis years. The big exception Hungary shows an adverse pattern, i.e. decreasing unemployment rates in the 1990s and increasing ones afterwards showing a decoupling of their growth trend from the other CEE countries’ economies.

In general, the emerging pattern indicates that youth unemployment rates increase much faster than total unemployment rates in times of economic downturns. The weaker EU-15 response to the crisis compared to CEE countries is mainly due to its composition. On the one hand, there are countries like Spain and Greece with similarly strong rises in unemployment figures. On the other hand, Germany and Austria show little change in unemployment figures during the recent crisis (Hutengs, Stadtmann 2013).

4. Regression analysis

In this chapter, we introduce our regression approach and shed light on the question how business- -cycle fluctuations affect the unemployment rate of different age cohorts. There exist many different versions of Okun’s law, the original ones proposed by Okun (1962), i.e. gap and difference version, and derivations developed over time, i.e. dynamic versions (see, for example, Knotek 2007). This paper focuses on the well-known difference version which highlights the effect of the business-cycle on the change in the unemployment rate. The difference version written as a linear regression model is given by:

ttt GDPu ε

ε

α

α

β

β

∆

i,tu∆

i,tu∆

,tu∆

=

=

+

+ +

+

tGDP

tGDP

,

tiiiii DD ,

ε ti ,

,

(1)

where ttt GDPu ε

ε

α

α

β

β

∆

i,tu∆

i,tu∆

,tu∆

=

=

+

+ +

+

tGDP

tGDP

,

tiiiii DD ,

ε ti ,

,

represents the change in unemployment rate,

ttt GDPu ε

ε

α

α

β

β

∆

i,tu∆

i,tu∆

,tu∆

=

=

+

+ +

+

tGDP

tGDP

,

tiiiii DD ,

ε ti ,

,

symbolizes the discrete GDP growth rate and ttt GDPu ε

ε

α

α

β

β

∆

i,tu∆

i,tu∆

,tu∆

=

=

+

+ +

+

tGDP

tGDP

,

tiiiii DD ,

ε ti ,

,

is an assumed white noise error term. The parameter β is called the “Okun coefficient” and is expected to show a negative sign.

Due to the transition from planned to market economies earliest reliable data start in the mid 1990s. Thus, there are only a limited number of observations available for single OLS estimates. A balanced panel for each country is constructed to circumvent these limitations. The panel includes the yearly changes in the unemployment rate and the GDP growth rate for five different age cohorts. Rather than estimating each beta coefficient for each age cohort and each country separately as suggested by equation (1), we estimate the following panel least squares dummy variable model (LSDV) for each country:

ttt GDPu ε

ε

α

α

β

β

∆

i,tu∆

i,tu∆

,tu∆

=

=

+

+ +

+

tGDP

tGDP

,

tiiiii DD ,

ε ti ,

, (2)

where

ttt GDPu ε

ε

α

α

β

β

∆

i,tu∆

i,tu∆

,tu∆

=

=

+

+ +

+

tGDP

tGDP

,

tiiiii DD ,

ε ti ,

, represents the change in unemployment rate for cohort i at time t, Di is a dummy variable accounting for the different age cohorts and

ttt GDPu ε

ε

α

α

β

β

∆

i,tu∆

i,tu∆

,tu∆

=

=

+

+ +

+

tGDP

tGDP

,

tiiiii DD ,

ε ti ,

, is an assumed white noise error term. Thus, βi captures the different cohort specific Okun coefficients.


Common to panel analysis are the presence of heteroscedasticity, serial correlation and cross- -sectional dependencies. These effects may lead to inefficient estimates with biased standard errors and thus misleading results. Heteroscedasticity and cross-sectional dependencies are found in all country panels. Serial correlation also exists in almost all countries with the exception of Estonia and Slovakia. Considering these econometric problems, all results are estimated with panel corrected standard errors allowing inference on statistical significance.4

The regression results of equation (2) are summarized in Table 3. The Okun coefficients are negative across all countries and cohorts. Therefore, the postulated negative correlation between GDP growth and change in unemployment rate can be confirmed: a negative GDP growth leads to an increase in unemployment rate for any cohort. The strength of the effect differs not only across countries but also across age cohorts. The former is to be expected as there are different labour market regimes and levels of economic development present in each country. All countries show their highest absolute Okun coefficients among their respective youth cohort, which are significantly larger than the ones for the 55−64 years cohort. This clearly indicates that young people are more exposed to fluctuations and suffer most in the crisis.

Tables 4 to 9 show Wald-Tests for equality of coefficients which confirm that the Okun coefficients for the youth cohort (15−24) differ significantly from those of all older cohorts.5 This holds for all countries. Some countries such as Czech Republic and EU-15 have Okun coefficients for the 25−34 years old cohort that differ significantly from the three oldest cohorts, whereas such a result cannot be confirmed for Slovakia. Tests for equality of coefficients usually do not provide significant results while comparing the Okun coefficients of the three oldest cohorts among themselves. This is to be expected due to similar sizes of coefficients. Furthermore, these cohorts are mostly well established on the labour market in terms of skills and experience.

Our results indicate that young people are more prone to GDP fluctuations. We have already elaborated on some reasoning in the literature review, i.e. firm’s low opportunity cost for firing young people. Further reasons may be incorporated in the institutional set-up of national labour markets. Older people are usually fairly protected by employment protection laws. As a consequence, they are the last to lose their jobs in an economic downturn. Young people tend to be more on temporary contracts and thus are more prone to layoffs in crisis times. Nonetheless, they will usually profit more in upswings. Janicko (2012, p. 5) supports this view for the Czech Republic. Furthermore, he shows that older people who were threatened by unemployment were allowed to retire early. This is consistent with our findings, that older people’s unemployment rate is only slightly affected by GDP fluctuations. In addition, fewer job creation in low growth periods and recessions probably puts young people on a disadvantage as they must compete with more experienced and skilled adults for fewer openings (Unt 2012).

Figure 2 plots the Okun coefficient for the different age cohorts for each country with 95% confidence intervals. The coefficients clearly diminish with increasing age in all countries, suggesting that the accumulation of skills, experience and employment protection leads to a more secure employment environment with increasing age.

The strongest increase in the Okun coefficient is observed from the 15−24 years cohort to the subsequent cohort (25−34 years). After the initial increase the coefficient increases slower over

4 Estimation results are obtained through linear regression with Prais-Winsten standard errors.5 Tables show empirical F-values and their corresponding significance level.


the following age cohorts, indicating that young people are indeed most vulnerable during economic downturns. The change between the two youngest cohorts is most distinct in Poland and Slovakia indicating a huge vulnerability of the youngest cohort to business-cycle fluctuations.

It has been argued that young people tend to choose additional education rather than unemployment in crisis times. If true, this would lead to a drop in participation rates for young people. As people in education are neither unemployed nor part of the labour force, the youth unemployment rate must ceteris paribus statistically decrease. Thus, in the case of lower participation rates, we would probably underestimate the effect of GDP fluctuations on the unemployment rate because unemployment rates are lower than what they would be without a decrease in participation. Unt (2012) does not provide support for the education argument, i.e. the Estonian youth does not study more because of the recent crisis. Dietrich (2012) recently finds a slight decrease in an aggregated European youth participation rate. Evidence for Poland shows an increase in the youth participation rate during the recent crisis which is in line with the reverse in migration.

Until now the analysis has been constrained to a sample consisting of male and female people combined. We conduct a sensitivity analysis for our results by separating the dataset by gender. Table 10 shows the regression results for the male subgroup which are very similar to the results of the main sample, i.e. the Okun coefficients are negative across all countries and become smaller (in absolute values) with increasing age. It has been argued that young and old women should share some similarities regarding their employment level and respective participation rates, i.e. falling employment should coincide with decreasing participation rates.6 The effect would be a constant unemployment rate for these two particular female age groups. Our regression results for women in Table 11 reflect this behaviour only partially. Hungary and Poland are the two countries where the Okun coefficients for both younger and older women are either insignificant or not statistically different from one another, i.e. we do not have any changes in unemployment due to changes in GDP. Both countries also show the worst model fit among all countries suggesting that GDP alone is not a good predictor for changes in the unemployment rate. Nevertheless, the results for female unemployment in the Czech Republic, Estonia, Slovakia as well as the EU-15 aggregate are in line with our previous findings.

5. Conclusions

In this paper we examined the unemployment development of several CEE countries and estimated age-cohort specific Okun coefficients. The main results can be summarized as follows:

1. The Okun coefficient in CEE countries for a subgroup consisting of 15−24 years old people is larger (in absolute values) than for a control group of older people, showing that young people are more dependent on good economic conditions to gain access to the labour market.

2. The Okun coefficient in CEE countries is decreasing (in absolute values) with progressing age across countries, whereas the fall is most distinctive when advancing from the youth group (15−24 years) to the second cohort (25−34 years old).

3. These results also hold for a subgroup of males and with the exception of Poland and Hungary also for a subgroup of females.

6 We thank an anonymous referee for bringing this specific problem to our attention.


Youth unemployment is a serious problem in many CEE countries. The highest rates of youth unemployment are observed in Estonia which is facing severe economic problems due to the financial crisis. Nevertheless, Estonia as well as Slovakia already show signs of a decreasing unemployment rates recently. In contrast, Poland shows the highest average unemployment rates of all countries. Furthermore, the Polish labour market does not show a trend reversal in unemployment yet. In general, the youth cohort depends much more on economic growth than other cohorts.

The proposed actions by the European Commission (2009) do address major labour market problems of young people, i.e. education and training. Nonetheless, these actions take time to have an effect on youth unemployment as they can be classified as mid to long term policy tools. These tools are indeed important as they can contribute to future reductions of youth unemployment levels across Europe. Even the most recently proposed specific actions plan from the European Commission (EC 2012) focuses on school-to-work transition and mobility improvements, thus completely ignoring the bad economic conditions prevalent in Europe. As our results suggest, young people are more negatively affected by business-cycle fluctuations. Thus, any policy not considering GDP growth will probably fail to alleviate Europe’s youth unemployment problem. Support for such conclusion comes from the OECD (2010, p. 147) which argues that policies must include means to positive GDP growth and qualification provisions in order to enable young people to compete on the labour market.

References

Baas T., Brücker H. (2011), EU eastern enlargement: the benefits from integration and free labour movement, Journal for Institutional Comparisons, 9(2), 44–51.Barrell R., Fitzgerald J., Riley R. (2010), EU enlargement and migration: assessing the macroeconomic impacts, Journal of Common Market Studies, 48(2), 373–395.Bell D.N.F., Blanchflower D.G. (2011), Young people and the Great Recession, Oxford Review of Economic Policy, 27(2), 241–267.Blanchard O., Amighini A., Giavazzi F. (2011), Macroeconomics: a European perspective, Prentice Hall, London.Dietrich H. (2012), Youth unemployment in Europe − theoretical considerations and empirical findings, Friedrich Ebert Stiftung, International Policy Analysis, Berlin, http://library.fes.de/pdf-files/id/ ipa/09227.pdf.EC (2009), European Union’s strategy for youth – investing and empowering, European Commission, http://ec.europa.eu/youth/documents/eu_youth_strategy.pdf.EC (2012), Moving youth into employment, European Commission, Communication from the Commission to the European Parliament, the Council, the European Economic and Social Committee and the Committee of the Regions, http://ec.europa.eu/europe2020/pdf/nd/2013eccomm_en.pdf.EC (2013), Domestic products statistics, European Commission, Annual Macroeconomic Database (AMECO), accessed on 14 January 2013, http://ec.europa.eu/economy_finance/db_indicators/ ameco/.European Economic Advisory Group (2013), The EEAG Report on the European Economy, CESifo, Munich.Furceri D., Zdzienicka A. (2011), The real effect of financial crises in the European transition economies, Economics of Transition, 19(1), 1–25.


Gächter M., Riedl A., Ritzberger-Grünwald D. (2013), Business cycle convergence or decoupling? Economic adjustment in CESEE during the crisis, Bank of Finland, BOFIT Discussion Papers, 3/13, http://www. suomenpankki.fi/bofit_en/tutkimus/tutkimusjulkaisut/dp/Documents/2013/dp0313.pdf.Hutengs O., Stadtmann G. (2013), Age effects in Okun’s law within the Eurozone, Applied Economics Letters, 20(9), 821–825.Janicko P. (2012), Youth employment in the Czech Republic and the standpoint of the Czech-Moravian confederation of trade unions (CMKOS), Friedrich Ebert Stiftung, International Dialogue, Berlin, http://library.fes.de/pdf-files/id/09471.pdf.Knotek E.S. (2007), How useful is Okun’s law?, Economic Review, Federal Reserve Bank of Kansas City, fourth quarter, 73–103.Mroz T.A., Savage T.H. (2006), The long-term effects of youth unemployment, The Journal of Human Resources, 41(2), 259–293.OECD (2010), Off to a good start? Jobs for youth, OECD Publishing, http://www.oecd.org/employment/ emp/offtoagoodstartjobsforyouth.thm.OECD (2013), Labour force statistics: population and labour force, OECD Employment and Labour Market Statistics (database), http: //www.oecd.org/std/labour-starts/.O’Higgins N. (1997), The challenge of youth unemployment, International Social Security Review, 50(4), 63–93.O’Higgins N. (2003), Trends in the youth labour market in developing and transition countries, World Bank Social Protection Discussion Paper Series, 0321, http://siteresources.worldbank.org/ SOCIALPROTECTION/Resources/SP-Discussion-papers/Labor-Market-DP/0321.pdf.Okun A.M. (1962), Potential GNP: its measurement and significance, in: M.N. Baily, A.M. Okun (eds.), The battle against unemployment and inflation, Norton, New York.Polakowski M. (2012), Youth unemployment in Poland, manuscript available at: http://library.fes.de/pdf- files/id/09477.pdf.Ryan P. (2001), The school-to-work transition: a cross-national perspective, Journal of Economic Literature, 39(1), 34–92.Sögner L., Stiassny A. (2002), An analysis on the structural stability of Okun’s law – a cross-country study, Applied Economics, 34(14), 1775–1787.Unt M. (2012), Boom and bust effects on youth unemployment in Estonia, manuscript available at: http://library.fes.de/pdf-files/id/09473.pdf.


Appendix

Table 1Long-term unemployment as percentage of total unemployment

Country Youth 2012a Youth 2003−2012b Adult 2012c Adult 2003−2012d

Czech Republic 33.4 32.2 45.9 49.9Estonia 29.8 30.4 60.6 50.6Hungary 31.2 34.5 48.0 47.8Poland 30.3 33.2 43.4 48.5Slovakia 56.3 54.7 70.0 71.7EU-15 31.5 25.7 47.4 44.7

a Youth long-term unemployment as % of total youth unemployment in 2012.b Youth long-term unemployment as % of total youth unemployment, yearly average between 2003−2012.c Adult long-term unemployment as % of total adult unemployment in 2012.d Adult long-term unemployment as % of total adult unemployment, yearly average between 2003−2012.Source: European Labour Force Survey.

Table 2Summary statistics

CountryAverage

youth unemploymenta

Average total

unemploymentb

Δ columns (1) and (2)c

Max Δ youth and total

unemploymentd

Min Δ youth and total

unemploymente

1 2 3 4 5

Czech Republic 14.11 6.43 7.67 12.09 3.30

Estonia 15.64 8.81 6.83 15.17 0.99

Hungary 18.15 8.42 9.73 16.47 5.53

Poland 30.42 13.53 16.89 23.95 10.17

Slovakia 29.09 14.52 14.57 19.76 9.09

EU-15 18.33 9.02 9.31 12.37 6.70

a Average unemployment rate of the youngest age cohort in %.b Average unemployment rate of the total population in %.c Difference between column (1) and (2) in percentage points.d Highest value of the difference between the young and the total unemployment rate within one country over time

in percentage points.e Lowest value of the difference between the young and the total unemployment rate within one country over time

in percentage points.


Table 3Regression results: Okun coefficients for different age cohorts and standard errors

Country 15−24 25−34 35−44 45−54 55−64 R2 N

Czech Republic -0.723***(0.171)

-0.318***(0.071)

-0.190***(0.055)

-0.145*(0.061)

-0.105*(0.053) 0.47 90

Estonia -0.605***(0.143)

-0.364***(0.052)

-0.283***(0.075)

-0.331***(0.056)

-0.233***(0.080) 0.52 90

Hungary -0.575***(0.168)

-0.228**(0.073)

-0.189**(0.064)

-0.147**(0.048)

-0.179*(0.088) 0.35 95

Poland -1.324***(0.353)

-0.637**(0.201)

-0.442**(0.141)

-0.454**(0.164)

-0.308*(0.133) 0.38 95

Slovakia -1.097***(0.168)

-0.392***(0.078)

-0.353***(0.076)

-0.339**(0.082)

-0.339**(0.113) 0.63 85

EU-15 -0.697***(0.080)

-0.380***(0.053)

-0.232***(0.044)

-0.203***(0.037)

-0.199***(0.057) 0.61 155

Notes:N – number of observations; standard errors in parentheses;significance at *** 1% level, ** 5% level, * 10% level.

Table 4Czech Republic: Wald test for equality of coefficients

β25−34 β35−44 β45−54 β55−64

β15−24 10.75*** 15.47*** 20.05*** 18.29***

β25−34 16.48*** 17.49*** 15.63***

β35−44 1.66 3.16*

β45−54 0.51

Note: significance at *** 1% level, ** 5% level, * 10% level.


Table 5Estonia: Wald test for equality of coefficients

β25−34 β35−44 β45−54 β55−64

β15−24 5.35** 9.37*** 7.85*** 13.87***

β25−34 2.82* 0.76 5.13**

β35−44 0.69 0.60

β45−54 3.13*


Table 6Hungary: Wald test for equality of coefficients

β25−34 β35−44 β45−54 β55−64

β15−24 6.38** 7.62*** 10.14*** 7.64***

β25−34 0.80 2.71* 0.32

β35−44 1.07 0.02β45−54 0.26


Table 7Poland: Wald test for equality of coefficients

β25−34 β35−44 β45−54 β55−64

β15−24 8.46*** 10.02*** 8.30*** 12.17***

β25−341.92 3.27* 5.99**

β35−440.01 1.11

β45−541.59



Table 8Slovakia: Wald test for equality of coefficients

β25−34 β35−44 β45−54 β55−64

β15−24 27.32*** 29.67*** 25.92*** 22.11***

β25−34 0.36 0.56 0.25

β35−44 0.09 0.03β45−54 0.00


Table 9EU-15: Wald test for equality of coefficients

β25−34 β35−44 β45−54 β55−64

β15−24 34.75*** 68.51*** 68.48*** 50.78***

β25−34 54.96*** 47.31*** 10.89***

β35−44 2.98* 0.49β45−54 0.01


Table 10Regression results: Okun coefficients and standard errors − men

Country 15−24 25−34 35−44 45−54 55−64 R2 N

Czech Republic -0.706***(0.186)

-0.254***(0.061)

-0.169***(0.050)

-0.148*(0.063)

-0.116*(0.058) 0.42 90

Estonia -0.666***(0.140)

-0.436***(0.062)

-0.410***(0.092)

-0.454***(0.091)

-0.248**(0.080) 0.56 90

Hungary -0.794***(0.198)

-0.242**(0.082)

-0.249***(0.070)

-0.210**(0.070)

-0.187*(0.084) 0.41 95

Poland -1.750***(0.326)

-0.756***(0.212)

-0.443***(0.120)

-0.433*(0.172)

-0.381*(0.151) 0.50 95

Slovakia -1.215***(0.198)

-0.444***(0.097)

-0.404***(0.079)

-0.407***(0.085)

-0.365**(0.124) 0.62 85

EU-15 -0.870***(0.090)

-0.487***(0.055)

-0.305***(0.041)

-0.250***(0.035)

-0.246***(0.060) 0.69 155



Table 11Regression results: Okun coefficients and standard errors − women

Country 15−24 25−34 35−44 45−54 55−64 R2 N

Czech Republic

-0.751***(0.168)

-0.394***(0.093)

-0.217**(0.074)

-0.152*(0.068)

-0.075(0.059) 0.47 90

Estonia -0.524**(0.071)

-0.271***(0.051)

-0.150*(0.067)

-0.227***(0.054)

-0.269***(0.080) 0.38 90

Hungary -0.268(0.162)

-0.224**(0.072)

-0.131*(0.063)

-0.081(0.050)

-0.170(0.119) 0.16 95

Poland -0.872*(0.416)

-0.594*(0.265)

-0.461*(0.192)

-0.512**(0.188)

-0.280(0.148) 0.21 95

Slovakia -0.939***(0.161)

-0.323***(0.069)

-0.299**(0.097)

-0.268**(0.095)

-0.268(0.209) 0.44 85

EU-15 -0.514***(0.092)

-0.258***(0.057)

-0.150**(0.057)

-0.153**(0.051)

-0.133*(0.060) 0.38 155



Figure 1Youth and total unemployment rates for CEE countries

0

10

20

30

40

50 Youth

Une

mpl

oym

ent r

ate

Une

mpl

oym

ent r

ate

Une

mpl

oym

ent r

ate

Une

mpl

oym

ent r

ate

Une

mpl

oym

ent r

ate

Une

mpl

oym

ent r

ate

Total

% %

% %

1980

1982

1984

1986

1988

1990

1992

1994

1996

1998

2000

2002

2004

2006

2008

2010

Czech Republic Estonia

Hungary Poland

Slovakia EU-15

0

10

20

30

40

50

1980

1982

1984

1986

1988

1990

1992

1994

1996

1998

2000

2002

2004

2006

2008

2010

YouthTotal

YouthTotal

%

YouthTotal

YouthTotal

%

YouthTotal

05

101520253035404550

1980

1982

1984

1986

1988

1990

1992

1994

1996

1998

2000

2002

2004

2006

2008

2010

05

101520253035404550

1980

1982

1984

1986

1988

1990

1992

1994

1996

1998

2000

2002

2004

2006

2008

2010

0

5

10

15

20

25

30

35

40

45

50

1980

1982

1984

1986

1988

1990

1992

1994

1996

1998

2000

2002

2004

2006

2008

2010

0

5

10

15

20

25

30

35

40

45

50

1980

1982

1984

1986

1988

1990

1992

1994

1996

1998

2000

2002

2004

2006

2008

2010


Figure 2Okun Coefficients and confidence intervals over age cohorts

Note: the scale on the vertical axis on the “Poland” graph differs from the other ones.

-1.5

-1.0

-0.5

0

15-24 25-34 35-44Age cohorts

Czech Republic Estonia

Hungary

Slovakia EU-15

Poland

Age cohorts45-54 55-64

15-24 25-34 35-44 45-54 55-64

Oku

n’s c

oeffi

cien

t

Oku

n’s c

oeffi

cien

t

Oku

n’s c

oeffi

cien

t

Oku

n’s c

oeffi

cien

t

Oku

n’s c

oeffi

cien

t

Oku

n’s c

oeffi

cien

t

-1.0

-1.5

-0.5

0

15-24 25-34 35-44 45-54 55-64

-1.0

-0.5

-1.5

0

-2.0

-1.5

-1.0

-0.5

0

15-24 25-34 35-44 45-54 55-64Age cohorts

Age cohorts Age cohorts

Age cohorts

-1.5

-1.0

-0.5

0

15-24 25-34 35-44 45-54 55-64

-0.5

-1.5

-1.0

0

15-24 25-34 35-44 45-54 55-64