Relationship between wages , labour productivity and unemployment rate in new EU member countries

e main aim of this article is to nd out the extent to which relative labour productivity and relative unemployment rate changes determine relative wage changes. We use average annual macro-data for the period 2002-2013 for Poland and other 5 new EU members: Estonia, Hungary, Slovak, Czech Republic and Slovenia. Using Poland as benchmark, rst we examine the correlation between wage, productivity and unemployment rate changes in countries in question. en, using panel data model we assess the elasticities of the relative wage changes with regard to relative productivity and unemployment changes. We found out that the trajectory of wage, productivity and unemployment rate development in new EU member countries is diversi ed. We con rmed a strong relationship between wage and productivity ratio changes in Poland related to Czech Republic, Estonia and Hungary. Moreover, an increase of productivity in Poland in comparison to Czech Republic is greater than an increase of wages in Poland in comparison to Czech Republic. e same relation occurs in Slovak and Slovak Republic. At the same time the productivity in Poland in relation to Hungary and Estonia has been growing slower than the wages in Poland in comparison to Hungary and Estonia. e correlations between wage and unemployment rate ratios are of smaller signi cance.


INTRODUCTION
Wage-setting mechanism is closely connected both with labour market and consumer goods market thereat plays the major role in the entire economy.In neoclassical approach higher labour productivity is re ected fully in higher wages.By reviewing relevant literature it can be noticed that a major part of wage determination analysis are based on the Philips curve (Phillips, 1958) or the wage curve (Blach ower & Oswald, 1994).Hence, in most studies on macro level wages are explained by unemployment.
In this article we propose quite modi ed approach: using aggregate macro-level data, we attempt to determine the relations between ratios of wages, labour productivity (hereafter, productivity refers to labour productivity) and unemployment rate.us, the wage ratio is de ned as a ratio of wage level in Poland to wage level in other country, productivity ratio means a ratio of labour productivity in Poland to labour productivity in other country and unemployment ratio denotes a ratio of unemployment rate in Poland to unemployment rate in other country.
e article is a continuation of previous own research on the wage determination in Poland and in Germany in the years 1997-2012 (Nikulin, 2013).e analysis concerned two countries with di erent rates of technological development.Estimation results indicated that an increase in productivity ratio by 1% (the ratio of productivity in Poland to productivity in Germany rose by 1%) caused a grow in wage ratio by 0,8% (the ratio of wages in Poland to wages in Germany rose by 0,8%), by other variables unchanged.Now, we try to examine the relations in countries which entered the European Union at the same time (in 2004) to point out a trajectory of some labour market indicators development.us, the main aim of this article is to look into the relations between wage, productivity and unemployment ratio in countries under consideration and to nd out the extent to which relative productivity and relative unemployment rate changes determine relative wage changes.We conduct an analysis for 6 new EU member countries: Estonia, Hungary, Slovak, Czech Republic, Slovenia and Poland, for the years 2002-2013.e structure of the article is as follows.Section 2 discusses previous empirical research on wage curves.In section 3 we present our data and methodology.Section 4 provides an empirical analysis of wage, productivity and unemployment rate in Poland in relation to other countries.We conduct a development of wage, productivity and unemployment ratio co-relation, then we use the Spearman's rank correlation coefcient to assess the correlation and nally we estimate an econometric model to explain the wage ratio variability.Section 5 concludes.

LITERATURE REVIEW
In the literature there is some debate about wage curves.Most of the considerations are based on conception provided rst by Phillips (Phillips, 1958) and their extensions (Samuelson & Solow, 1960) and (Tobin, 1972).e wage model (1) used by Welfe (Welfe, 1997) (Welfe, 1997).Note, that there is a several other attempts to assess the determinants of wage level or wage rate in Poland, see e.g.(Welfe, 1996), (Welfe, 2000), (Osiewalski & Welfe, 1998), (Kwiatkowski et al., 1999), (Welfe, Kelm, & Majsterek, 2002), (Ossowski, 2013), (Nikulin, 2013).Proposed empirical models enable measuring the elasticity of wages with respect to price level, labour productivity and unemployment rate.Moreover, some researchers provided empirical wage models using real wage level, e.g.(Welfe, Karp, & Kębłowski, 2006).Blanchard and Katz (Blanchard & Katz, 1999) estimated the real wage as the function of unemployment, given the reservation wage and labour productivity.ey considered the wage curve proposed by Blanch ower and Oswald (Blach ower & Oswald, 1994) and examined some OECD counties and US economy.Moreover, there is also a wide range of analysis of the wage determination in new EU member countries.D'Adamo (D'Adamo, 2014) examined wages in the public sector and the private traded and non-traded sector in ten transition countries which are members of the European Union.Rusinova et al. (Rusinova, Lipatov, & Heinz, 2015) found the evidence for a reaction of wage growth to unemployment and productivity growth in 19 EU countries.Wage-setting analysis for Poland, Hungary, Czech Republic, Slovakia were provided e.g. by Basu et al. (Basu, Estrin, & Svejnar, 2004) and (Estrin, Svejnar, & Basu, 1997).ey obtained a negative coe cient of unemployment in the wage equation.However, other authors also suggest the negative unemployment elasticity of pay.Iara and Traistaru (Iara & Traistaru, 2003) estimated wage curve for Poland, Hungary, Romania and Bulgaria using regional data from 1991-1999.ey retrieved that average wages were negatively associated with regional unemployment rate in Poland (the unemployment elasticity of pay was around -0,06).Similar results received Du y and Walsh (Du y & Walsh, 2001).Blanch ower (Blanch ower, 2001) examined the wage curve for 23 transitions countries (i.a.Poland, Slovakia, Hungary, Czech Republic, Estonia, Slovenia).He found the wage elasticity with respect to unemployment rate from -0,3 to -0,1.
Also note that our empirical wage equations are only partially in line with the mentioned real wage models.We contribute to the literature on labour market in European countries by applying ratios of wage, productivity and unemployment rate in Poland to other countries instead of using their levels.We believe that our approach allow to observe the changes in wages, productivity and unemployment rate in Poland in comparison to changes in the other countries, which entered the European Union at the same time.e main result of our study is a comparison of productivity and wage changes in Poland in relation to analogous changes in other new EU member countries.

DATA AND METHODOLOGY
All data were collected from OECD statistics (data.oecd.org).According to the OECD methodology the average wage is obtained by dividing the national-accounts-based total wage bill by the average number of employees in the total economy, which is then multiplied by the ratio of the average usual weekly hours per full-time employee to the average usually weekly hours for all employees.is indicator is measured in USD constant prices using 2012 base year and Purchasing Power Parities (PPPs) for private consumption of the same year, what enables an international comparison (OECD, 2015a).e average labour productivity instead is calculated on dividing the level of GDP by hour worked.e indicator is measured in USD constant prices (using 2005 as base year) and PPP of 2005 (for more information about the measure of labour input see e.g.(OECD, 2001)).Following the OECD methodology (OECD, 2015b), we use the harmonised unemployment rate (HUR), to calculate the extent of unemployment.
We use quantitative methods.First, we observe the development of wage and productivity ratio in the analysed period.en we assess the correlation between relative wages, productivity and unemployment rate in Poland in relation to other countries using Spearman's rank correlation coe cient.Finally, we examine the elasticities of wage ratio with respect to productivity ratio and unemployment ratio in all counties in question.We apply the random coe cient model for panel data.

Development of the wage-productivity co-relation
is part of the article we start with the explanation of tree types of ratios we've built.e rst one, called wage(earnings) ratio(E t ), means the ratio of average annual wage in Poland in time t to average annual wage in country i, at the same time t: Comparing the wage and productivity ratio related to Poland and Estonia, we can observe that the ratio altered signi cantly in the years 2002-2013.In 2002 wages in Poland were 167% of wages in Estonia, whereas in 2013 only 120%.Whereas, the ratio of productivity in Poland and productivity in Estonia showed minimal uctuation.In 2002 productivity in Poland is 106% of productivity in Estonia, while in 2013 -101%.It is important to note that the gap between wage ratio (E) and productivity ratio (LP) decreased signi cantly in the analysed period, what results mainly from the continually growth of the wage level in Estonia.
As can be seen from the Figure 1C the wage ratio as well as the productivity ratio were volatile in the analysed period.From 2002 to 2007 both ratios have been falling, what indicate, that both the wages and productivity have been rising faster in Hungary than in Poland.On the contrary, after 2007 an upward trend may be observed, what mean that both the wages and productivity have increased rapidly in Poland than in Hungary.Comparing year 2002 with 2013 we can observe that the wage ratio grew from 1,05 to 1,08.us, despite of signi cant uctuations in analysed period, the nal level of wage ratio was nearing to the initial one.However, the growth of productivity in Poland was more signi cant than in Hungary at the same time.Figure 1D shows that the productivity ratio in Poland to Slovak Republic was unchanged in the years 2002-2013.In particular, in 2002 the productivity in Poland constituted 78% of the productivity in Slovak Republic, whereas in 2013 the ratio was 75%.On the contrary, the wage ratio showed a slightly downward trend, particularly from 2002-2007, when the ratio fell from 1,27 in 2002 to 1,07 in 2007 (wages in Poland in 2007 constituted 107% of wages in Slovak Republic).After 2007 the relation between productivity level in these two countries was rather stable.On the other hand, for the whole period 2002-2013 the wages in Slovak Republic have been increasing faster than in Poland, therefore the gap between wages in these two countries has decreased.
Within the Figure 1E is contained that the wage ratio between Poland and Slovenia maintained the steady rate.In 2002 the average wage level in Poland constituted 74% of average wage level in Slovenia, whereas in 2013 it was 71%.Moreover, the uctuations of productivity ratio were also slight in the whole period.In 2013 in comparison to 2002 the productivity ratio was greater only by 7 pp.erefore, in 2013 productivity in Poland constituted 65% of productivity in Slovenia.

Correlation between wages, labour productivity and unemployment rate
In previous step we have examined the developing of wage and productivity ratio in 5 countries in the years 2002-2013.
Table 1 Correlation between wage and productivity ratio (r E, LP ) and between wage and unemployment ratio (r E, UR ) in new EU member counties in the years 2002-2013.
In the next part of the article we assess the correlation between wage ratio and labour productivity ratio in countries under consideration.Moreover, we investigate also the correlation between wages and unemployment rate.Because of small dataset, we use one of the nonparametric test, Spearman's rank correlation coe cient, to indicate the relation between ratios we analyse.e results of the estimation are presented in Table 1.
It can be seen from the Table 1 that only in Hungary and Slovak Republic there is a statistically signi cant relation between wage ratio and unemployment ratio (we use an alpha level of 0,10 for all statistical tests).In Hungary the correlation is inverse and rather moderate what signi es that an increase in wage ratio corresponds to a decrease in unemployment ratio, and conversely.While, in Slovak Republic an increase in wage ratio corresponds to an increase in unemployment ratio.In other countries it is not possible to determine the relations between wage and unemployment ratio, because of their statistical insigni cance.In case of correlation between wage and productivity ratio, the statistically signi cant relation occurs in all countries under consideration.In Czech Republic, Estonia and Hungary the relation is positive and strong, what implicates the fact, that changes in wage ratio are strongly connected with changes in productivity ratio.It is important to note, that an increase in the ratio of average wage in Poland to average wage in Czech Republic corresponds closely to an increase in the ratio of productivity in Poland to productivity in Czech Republic.e same relationship emerges in Estonia and Hungary, whereas in Slovak Republic and Slovenia the strength of this relationship is weaker.On the whole, in each analysed country the changes in wage ratio are signi cant connected with changes in productivity ratio.

Econometric modelling
Besides the identi cation of correlation between analysed ratios it seemed to be recommended to use econometric wage determination models to indicate the elasticities of the wage ratio with regard to productivity and unemployment ratio.e analysed data are cross -sectional time series, so we've decided to use a panel data model.Given that our entitles (countries) are heterogeneous, we estimate a random coe cient regression using the generalized least squares (GLS) method.We've examined the structural di erences across ve countries in question using a Chow test and have rejected the H 0 hypothesis about poolability.We nd it appropriate to treat the elasticities of wage ratio with respect to productivity ratio and unemploymen t ratio as random variables di ering from country to country.Random coe cient regression (RCR) model treats both intercept and slope coe cients as random variables (Swamy, 1970).e model proposed by Swamy (Swamy, 1970) is as follows: where i=1..N denotes countries, y i is a vector of observations for ith country, X i is a matrix of nonstochastic covariates, and i is a vector of parameters speci c to country i. e error term vector i is distributed with mean zero and variance ii I.Moreover, each country -speci c i is related to common parameter vector : where E( i ) = 0, E( i i ') = ∑, E( i ' j ) = 0 for j≠I, and E( i ' j ) = 0 for all i and j.
In this case we consider the following model: where: LE it -natural logarithm of wage ratio in time t and country i; LLP it -natural logarithm of productivity ratio in time t and country i, LUR it -natural logarithm of unemployment ratio in time t and country i.
After estimation of the model ( 6) we get country-speci c best linear predictors.e elasticities of the wage ratio with regard to productivity and unemployment ratio are reported in Table 2 Testing the joint signi cance of the slope parameters with the use of the Wald chi2 test, we can state that all the coe cients in the model are statistically signi cant.Based on estimated model we can make the following conclusions: -in case of Czech Republic: if the ratio of productivity in Poland to productivity in Czech Republic increases by 1 %, we can expect the ratio of wage in Poland to wage in Czech Republic to increase by an average of 0,76%.Moreover, an increase in ratio of unemployment rates in Poland and in Czech Republic by 1% causes an average increase in wage ratio by 0,11%; -in case of Estonia: if the ratio of productivity in Poland to productivity in Estonia increases by 1 %, we can expect the ratio of wage in Poland to wage in Estonia to increase by an average of 2,67%.Moreover, an increase in ratio of unemployment rates in Poland and in Estonia by 1% causes an average increase in wage ratio by 0,07%; -in case of Hungary: an increase in ratio of productivity in Poland and in Hungary by 1% causes an average increase in wage ratio by 1,24%.e in uence of unemployment ratio on wage ratio is statistically insigni cant; -in case of Slovak Republic: if the ratio of productivity in Poland to productivity in Slovak Republic increases by 1 %, we can expect the ratio of wage in Poland to wage in Slovak Republic to increase by an average of 0,84%.Moreover, an increase in ratio of unemployment rates in Poland and in Slovak Republic by 1% causes an average increase in wage ratio by 0,14%; -in case of Slovenia: if the ratio of productivity in Poland to productivity in Slovenia increases by 1 %, we can expect the ratio of wage in Poland to wage in Estonia to increase by an average of 0,8%.Moreover, an increase in ratio of unemployment rates in Poland and in Slovenia by 1% causes an average increase in wage ratio by 0,09%.
Table 2 e elasticities of the wage ratio with regard to productivity and unemployment ratio analysed countries.

Coeffi cient Standard error z p>z
Source: own elaboration.

CONCLUSIONS
Our analysis demonstrates that there is no correlation of statistical signi cance between wage and unemployment ratio in most countries in question (with an except to Hungary where there is a moderate adverse correlation).It denotes, that the ratio of wages in Poland and those of the other countries are changing differently from the ratios of unemployment rate.At the same time, the changes in ratio of wages in Poland to wages in the other countries are more connected to the changes in ratio of productivity.In the case of Czech Republic, Estonia and Hungary we found a strong positive correlation between wage and productivity ratio, what means, that the changes in wages in Poland in comparison to wages in Estonia (and to wages in Hungary and Czech Republic) correspond to changes in productivity in these countries.In Slovak Republic and Slovenia the correlation is of smaller signi cance.It can be concluded that wages do not adjust thoroughly to productivity movements.Using panel data model we found that the productivity in Poland in relation to Czech Republic, Slovak Republic and Slovenia has been growing faster than the wages in Poland in comparison to given countries.It is important to note, that the level of productivity in Poland is lower than in Czech Republic, Slovak Republic and Slovenia (in 2013 average labour productivity in Poland consisted ca.80% of average labour productivity in Czech Republic and in Slovak Republic and only 65% of average labour productivity in Slovenia), whereas the wage levels in Poland, Czech Republic and Slovak Republic are similar.e lower productivity level in Poland could be a reason for greater dynamic of productivity increase in Poland in analysed period.Conversely, in comparison to Estonia and Hungary, the productivity in Poland has been growing slower than the wages in Poland in relation to wages in Estonia and Hungary.
Summarizing we can point out that the trajectory of wage, productivity and unemployment rate in new EU member countries is diversi ed.Our remarks are consistent with previous research on the diversity of new EU member countries, see e.g.(Szymczak & Gawrycka, 2008).On the basis of our dataset and methodology we compared the dynamic of relations between wages, productivity and unemployment rate in Poland in comparison to other new EU members.We believe that our analysis could be an incentive to further research in wage determination on macro-level.

Figure 1 .
Figure 1. e wage (E) and productivity (LP) ratios in the years 2002-2013.Source: own elaboration using the OECD databases OECD (2015), Average wages (indicator).doi: 10.1787/ cc3e1387-en (Accessed on 05 February 2015), and Level of GDP per capita and productivity: http://stats.oecd.org/Index.aspx?DataSetCode=PDBI_I4 is in the line with the wage curve proposed by Phillips: t denotes average nominal wage, p t denotes average price level calculated with the use of consumer price index, z t -labour productivity, ur t -unemployment rate, t -random coe cient.Model (1) was used to estimate the wage rate in Poland in the period AWP t denotes average annual wage in Poland in time t, AWC it denotes average annual wage in other country i at the same time t.en, we can describe the productivity ratio (LP t ) and unemployment ratio (UR t ) in the same way: Czech Republic, Estonia, Hungary, Slovak Republic and Slovenia).Within the Figures 1a-1e is contained that the di erences in wage level between Poland and other countries are larger than the di erences in labour productivity.In particular, in 2002 average wage in Poland amounted to 123% of average wage in Czech Republic.Simultaneously, the ratio of productivity in Poland to productivity in Czech Republic valued at 0,74, what indicate, that the productivity in Poland equalled 74% the productivity in Czech Republic.In 2013 the wage level in Poland is 111% the wage level in Czech Republic and the productivity in Poland amounted to 79% of productivity in Czech Republic We can conclude, that predominantly wages in Poland between 2002-2013 have been rising slower than in Czech Republic, whereas the productivity in Poland have been increasing faster than in Czech Republic.
where: ALPP t denotes average labour productivity in Poland in the time t, ALPC it denotes average labour productivity in other country i in the time t.where: URP t denotes unemployment rate in Poland in the time t, URC it denotes unemployment rate in other country i in the time t.e Figure1presents wage (E) and productivity (LP) ratios development in the time 2002-2013 in 5 countries (