When analyzing the population dynamics, we checked its correlation with average unemployment rates as reported by Eurostat. There is a positive correlation as one would expect (higher unemployment is associated with faster decline in population), however, it is rather weak as one can see in the chart below. In particular, unemployment in Poland and Slovakia has been persistently high during our period of analysis but population numbers have remained stable. At the same time, Romania has had relatively low unemployment, but it experienced the highest population decline among the new EU member states.
Removing Romania raises the R square value to 0.373. At the same time, if we include Romania and remove Cyprus, which is an outlier in terms of its population dynamics, having seen its population increase by more than a fifth during last decade, the R squared value falls to 0.0474 and the correlation falls apart altogether. The correlation is weak and sensitive to the inclusion or exclusion of particular countries.
For comparison, we reproduce below the chart of the correlation between relative income levels to EU-15 and population dynamics from our post of April 29. We would also like to stress that this correlation stays robust even after single countries are removed from the data set.
From the above, we can conclude that in the new EU member states during last decade longer term unemployment levels have had a considerably smaller impact on population decline than relative income levels vis-a-vis the EU-15. Thus, if those new EU member states that are still losing population would like to stop this negative trend, they should focus their attention on achieving further convergence with EU-15 income levels, rather than reducing unemployment per se. In other words, creating new low-wage jobs is unlikely to stop the negative population dynamics, in particular, if wages are substantially below those paid in more affluent countries where workers can migrate easily because of the free movement of people within the EU. For emigration to stop , the average income level needs to go up.
Now, in anticipation of further comments and questions we would like to stress a few additional points. First, we are not claiming that this (lack of) correlation between the unemployment level and negative population dynamics is universal. For example, long-term migration from Mexico to the U.S. has been driven by much higher incomes and wages in the U.S. However, during the Great Depression of the 1930s and the recent Great Recession, there was net migration of workers back to Mexico. Korkiasaari and Söderling (2003) mention widespread unemployment in Finland as a key reason for massive emigration of Finns to Sweden in the 1960s. Second, we are also not saying that unemployment peaks such as the one experienced by Poland before joining the EU in 2004 or those experienced by the Baltic States in 2009-2010 have no effect on emigration. In fact, there is anecdotal evidence from both of those episodes to suggest the opposite. However, differences in average unemployment levels in the new EU member states appear to matter little for longer term population dynamics. The poorer EU members, such as Bulgaria, Romania and Latvia should have expected to experience net emigration for any given level of unemployment in those countries.
This conclusion has an important corollary for the continuing debate on whether the Baltic States were right to follow the path of internal devaluation in 2008-2009 instead of devaluing their currencies. It is highly likely that choosing to devalue the currencies would have resulted in more jobs being preserved. However, it would undoubtedly also have resulted in a more dramatic fall in relative incomes vis-a-vis the EU-15 than in the internal devaluation case as there would have been much less pressure to improve productivity. If in the longer term low relative incomes are a more important driver of emigration than joblessness, having kept the pegs intact in 2009 might actually now help to stop net emigration sooner than in the alternative scenario.
Reference
Korkiasaari, Jouni and Ismo Söderling. 2003. „Finnish Emigration and Immigration after World War II”, www.migrationinstitute.fi
