Regional labour market disparities
From Statistics Explained
- Data from November 2012 and March 2008. Most recent data: Further Eurostat information, Main tables and Database.
The data on dispersion of unemployment rates in NUTS 2 regions are from November 2012. More detailed data presenting unemployment dispersion rates on NUTS 3 level and employment dispersion rates were calculated for the period 2002 - 2006, these will be updated in October 2013.
Main statistical findings
Dispersion of regional unemployment rates in 2011
In 2011, the dispersion of regional unemployment rates increased in most Member States compared with the previous year. Figure 1 presents the dispersion of regional unemployment rates in 2011 and 2010. For the EU-27, the dispersion of unemployment rates increased from 50.8 % in 2010 to 56.7 % in 2011. This overall development was reflected at national level, where disparities in regional unemployment rates rose in most Member States. Decreasing regional disparities were observed in the Czech Republic, Spain, France and Portugal, where dispersion declined from 2010 to 2011.
The highest dispersion of unemployment rates was recorded in Belgium (59.6 %), followed by Italy (43.0 %) and Germany (42.3 %). The geographic pattern of the disparities in these countries can also be seen in Map 1. In contrast, Denmark (7.3 %), Greece (10.3 %) and Sweden (11.3 %) had the lowest disparities in regional unemployment rates in 2011. However, low dispersion should not be interpreted as a positive sign for labour markets per se. In the case of Greece, which has quite small regional disparities in unemployment rates, all NUTS 2 regions recorded high unemployment rates over 14 % in 2011. This shows that dispersion only indicates the disparities between regions and not the overall level of unemployment.
Differences in dispersion rates in 2002 - 2006
The estimates for 2006 show that regional disparities in employment and unemployment were narrowing over the period 2002 - 2006. Notwithstanding these general labour market improvements, almost 20 % of the active population of the European Union still lived in underperforming regions as regards unemployment.
Regions with high unemployment at that time (see latest situation) were mainly located in north-eastern Europe: in Poland, eastern Germany and eastern Slovakia (Map 2). The French overseas departments, Extremadura in Spain and Italy's southern regions also suffered from high unemployment rates. Map 2 shows some variability in unemployment rates within EU countries: some Member States had different levels of regional unemployment, while others showed a quite similar distribution of their regional unemployment rates.
In 2006, the dispersion of regional unemployment rates at NUTS level 2 in the EU was 45.6 % and, at NUTS level 3, it was slightly higher at 50.2 %. These estimates show that the distribution of regional unemployment rates shrank over the period 2002 - 2006, with drops of more than 17 percentage points, for both NUTS level 2 and 3.
Similar results were seen for employment rates but with a lower magnitude. The dispersion of employment rates at NUTS level 2 decreased in 2002 - 2006 by 1.7 percentage points and stood at 11.4 % in 2006. At NUTS level 3, the dispersion was slightly higher, at 14.7 %, only 0.3 percentage points below the level in 2002.
Countries showed quite different levels of dispersion (Figures 3 and 4). The highest dispersion of unemployment rates was observed in Italy and in Belgium. These countries even showed dispersions of unemployment rates higher than the EU average, meaning that the distribution of unemployment rates was relatively wider in these two countries than in EU as a whole. In Belgium, the lowest unemployment rates were in Vlaams-Brabant and West-Vlaanderen, both 4.2 %, and the highest was in Région de Bruxelles-Capitale/Brussel Hoofdstedelijk Gewest, at 17.6 %, more than four times the lowest Belgian regional unemployment rate. The Italian NUTS 2 region of Sicilia had an unemployment rate in 2006 of 13.5 %, more than five times higher than Italy’s lowest regional unemployment rate, which was 2.6 % in Provincia Autonoma Bolzano/Bozen.
Countries that also showed relatively high dispersions of unemployment rates were the Czech Republic, Slovakia, Germany and Austria, but in this last country, the high dispersion is influenced by the country’s relatively low unemployment level, which stood at 4.7 %.
Bulgaria and Romania had dispersions that were not too high compared to the other EU countries at NUTS level 2, but a different picture was revealed at NUTS level 3. These marked differences between the NUTS level 2 and NUTS level 3 dispersions mean that there was a large variability between NUTS level 3 regions belonging to one NUTS level 2 region. For example, the NUTS level 2 Bulgarian region of Yugoiztochen, with an unemployment rate of 8.1 %, comprises NUTS level 3 regions with rates ranging from 4.5 % in Stara Zagora to 17.1 % in Sliven.
On the other hand, countries like Sweden, Poland, Greece, the Netherlands and Portugal all had low dispersions of unemployment rates. The dispersion of unemployment rates for Latvia, Lithuania and Ireland, which can only be measured at NUTS level 3, was also relatively small when compared to the other EU countries.
Generally, over the period 2002 - 2006, there was a reduction in the dispersion of unemployment rates (Table 1). It could have been expected that there would be a correlation between the levels of dispersion and how they have reduced over time, but that was not the case. Some countries with high dispersion showed significant reductions, like Germany or Italy, while others, also with high levels of dispersion, even saw increases over the period 2002 - 2006, like Belgium or Slovakia.
The dispersion of employment rates affects countries in a similar manner to the dispersion of unemployment rates, as the countries with high dispersions of unemployment tend to have high dispersions of employment.
There were three exceptions: Austria, Germany and the Czech Republic. These countries had a relatively high dispersion of unemployment but, at the same time, a relatively low dispersion of employment. While in Austria this seems to be related to the low unemployment level, in Germany and the Czech Republic, differences in regional unemployment affected these countries more intensively than differences in employment.
Over the period 2002 - 2006, the Czech Republic, Spain, Poland, Finland, the United Kingdom, Ireland and Sweden showed a clear decreasing trend in their dispersion of employment rates, at both level 2 and 3 of NUTS. On the other hand, Belgium and Slovakia, increased their dispersion over this period of time.
Four countries showed different evolutions of their dispersion of employment rates, depending on the NUTS level. Italy, Germany and Portugal narrowed the distributions of their regional employment rates if NUTS level 2 is considered. But, while this NUTS level 2 is converging to the same level of employment, “sub-NUTS level 2” regions are diverging at the same time. Greece showed the opposite, since its employment rates at NUTS level 3 were converging within its NUTS level 2 regions, but at NUTS level 2, there was no convergence.
A region is deemed “underperforming” if its employment rate is relatively low compared to the national employment rate (below 90 % of the national figure) or if its unemployment rate is relatively high compared to the national rate (above 150 % of the national figure).
This indicator can also be computed relative to the EU average, drawing attention to regions having trouble absorbing people into their labour markets.
In 2006, at NUTS level 2, there were 51 underperforming regions (UPRs) as regards employment out of 255 regions where data was available. These regions, where employment rates were relatively low compared to the EU average, were home to 20.6 % of the working population aged 15 to 64. Regarding unemployment, at NUTS level 2, there were 43 UPRs out of 261 where data was available, and 16.4 % of the active population was living in those regions.
Over the period 2002 - 2006, the UPR index stayed more or less the same: a decrease of 1 region, from 52 UPRs in 2002, regarding employment and a decrease of 3 regions, from 46 UPRs in 2002, regarding unemployment. This small reduction also led to a marginal decrease n in the population affected by underperformance: there was 0.2 percentage points of working population less living in UPRs as regards employment.
Regarding unemployment, the decrease in active population living in underperforming regions was 1.5 percentage points at NUTS level 2 (Figures 5 and 6). At NUTS level 2, 10 countries in 2006 had UPRs as regards unemployment (Figure 7) and eight countries had none. The underperforming region indicator cannot be computed for eight countries, since these countries comprise just one or two regions. Data for Denmark is still not available in the new NUTS code version.
Population can be affected to a greater or lesser extent, depending on the region considered to be underperforming. While in Spain, Greece and France, the active population affected is 2.5 % or less, in Austria, Belgium, the Czech Republic and Italy, this percentage exceeds 20 %. Italy, with 27.5 %, recorded the highest percentage. This is due to a clear division between the northern and southern Italian regions. The northern regions had lower unemployment rates, ranging from 3.0 % in Piemonte to 7.5 % in Lazio, while rates in the southern regions were significantly higher, ranging from 10.0 % in Molise to 13.5 % in Sicilia.
Different results were reported in employment as well. Only eight countries had underperforming regions in 2006 (Figure 8). Italy and Belgium had a significant percentage of people living in regions with relatively low levels of employment. So the underperformance of these countries is seen in both employment and unemployment.
But Austria, with a significant percentage of active population living in underperforming regions, didn’t underperform in employment: it had no underperforming regions at all as regards that criterion. Hungary was the opposite. It did not underperform in unemployment, meaning that there were no regions with a relatively high unemployment rate compared to the other Hungarian regions, but 2 regions, Észak-Magyarország and Észak-Alföld, comprising 27.4 % of the working population, had employment rates that were relatively low.
Data sources and availability
The LFS target population is made up of all people in private households aged 15 and over. The survey's definitions are based on the definitions and recommendations of the International Labour Organization. In addition, harmonization is achieved through the Member States’ adherence to common principles of questionnaire design.
Population covers people aged 15 and over, living in private households (population living in collective households, i.e. residential homes, boarding houses, hospitals, religious institutions, workers’ hostels, etc., are not included). This comprises all people living in the households surveyed during the reference week. The definition also includes people absent from the households for short periods (but having retained a link with the private household in question) owing to studies, holidays, illness, business trips, etc. People on obligatory military service are not included.
An employed person is a person aged 15 and over (16 and over in ES and UK, 15 to 74 in DK, EE, HU, LV, SE and FI) who, during the reference week, worked at least one hour for pay or profit, or was temporarily absent from such work. Family workers are included.
The unemployed are people aged 15-74 (16 to 74 in ES and UK) who were (all three conditions must be fulfilled simultaneously):
- without work during the reference week;
- available for work at the time (i.e. were available for paid employment or self-employment before the end of the two weeks following the reference week);
- actively seeking work (i.e. had taken specific steps in the four-week period ending with the reference week to seek paid employment or selfemployment), or who found a job to start within a period of at most three months.
The dispersion of the unemployment rates is expressed by the coefficient of variation of regional unemployment rates.
The coefficient of variation is the ratio between the weighted standard deviation of the regional unemployment rates (NUTS level 2 or 3), compared to the national unemployment rate, and the national unemployment rate. For the EU, the regional unemployment rates are compared with the Union's unemployment rates. This coefficient of variation is multiplied by 100 for expression as a percentage.
This indicator measures the spread of regional unemployment rates as regards the national or EU unemployment rate. If all the regional unemployment rates of a country are equal, the dispersion is zero. Big differences between regional unemployment rates within a country imply a fairly wide dispersion of unemployment rates.
An underperforming region is a region that has a relatively low employment rate or a relatively high unemployment rate. The thresholds considered are below 90 % of the national employment rate and above 150 % of the national unemployment rate. To compute the EU aggregate, the thresholds are below 90 % of the EU employment rate and above 150 % of the EU unemployment rate.
All regional results concern NUTS level 2 and NUTS level 3 regions. Down to NUTS level 2, the regional labour market data provided by Eurostat are derived from the LFS. Down to NUTS level 3, LFS data when available or registered employment and unemployment data when LFS is not available are used to break down the NUTS level 2 figures into NUTS level 3, which are used to compute the regional disparity indicators.
Since the 1st quarter of 2004, the samples for the Austrian, Italian and Maltese Labour Force Surveys have been spread over all weeks of the quarter. At the same time, the sampling and weighting procedures in Greece have been revised in order to improve coverage. The 2003 data for Cyprus refer to the 2nd quarter, while the 2004 data represent annual averages. Consequently, 2004 and 2003 data are not fully comparable for these countries.
Employment and unemployment levels are important indicators of both economic and social health. Countries with high levels of employment - and low levels of unemployment - tend to be performing well economically. In addition, employment is an important determiner of individual well-being, self-esteem and social harmony and equity. Dispersion measures the spread of regional rates within the same country or in the case of the EU-27 across all regions. An increase in dispersion indicates larger disparities, while smaller dispersion shows better cohesion between regions.
- Impact of the economic crisis on regional labour markets
- Labour markets at regional level
- Unemployment and beyond
- Unemployment statistics
- Unemployment statistics at regional level
Further Eurostat information
- Regional Statistics Illustrated - select statistical domain 'Labour market' (top right)
- European Regional and Urban Statistics - Reference guide 2008 (available in English, French and German)
- Large differences in regional labour markets show asymmetric impact of the economic crisis - Statistics in Focus 54/2012
- Regional labour market: higher unemployment rates and increasing disparities in 2010 - Statistics in Focus 60/2011
- Regional labour market statistics (t_reg_lmk)
- Employment rate of the age group 15-64, by NUTS 2 regions (tgs00007)
- Employment rate of the group 55-64 years, by NUTS 2 regions (tgs00054)
- Dispersion of regional employment rates, by sex (tsdec440)
- Unemployment rate, by NUTS 2 regions (tgs00010)
- Long-term unemployment rate (12 months and more), by NUTS 2 regions (tgs00053)
- Regional labour market statistics (reg_lmk)
- Regional economically active population - LFS annual series (lfst_r_lfpop)
- Regional employment - LFS annual series (lfst_r_lfemp)
- Regional unemployment - LFS annual series (lfst_r_lfu)
- Regional socio-demographic labour force statistics - LFS annual series (lfst_r_lfsd)
- Regional labour market disparities - LFS series and LFS adjusted series (lfst_r_lmd)