The incidence of persistent risk of poverty across the EU

The EU-SILC longitudinal data, which cover the same sample of the population over the four income years 2005-2008, can give a more reliable indication of change in the relative number of people at risk of poverty than the full set of annual data, which covers a varying sample of the population each year. Even though each sample is intended to be representative of the population as a whole, it is inevitable that there will be some (even if only small) variation in the tendency for the people concerned to have income below the poverty threshold. By allowing for sampling errors (see section on Poverty trends across the EU) we can take account of these variations; but following the same group of people avoids the problem altogether. On the other hand, some uncertainty still remains over how representative the group in question is of the population as a whole, since it is much smaller in number than the group covered by the full sample (only 25% of the latter).

Over the EU as a whole – or, more precisely, over the 24 Member States for which there are data – the average proportion of the population at risk of poverty declined in 2005 and then remained virtually unchanged until 2007 (Table 1). The pattern, however, differs across the EU, the proportion over the period falling in some countries (the Czech Republic, Belgium, Malta and Poland, especially) and rising in others (Bulgaria, Cyprus and Sweden, in particular).

The proportion of people at persistent risk of poverty in the 2008 income year averages just over 9% over the countries taken together, implying that just over three in every five people who were at risk in 2008 had income below 60% of the median for at least two out of the previous three years. 

The proportion of the population at persistent risk varies from just over 17% in Latvia to only 5% in Slovakia and Sweden and just under 4% in the Czech Republic. The proportion at persistent risk is also relatively large in Greece (over 14%) and Italy (over 13%), and is relatively small in France, Malta, the Netherlands, Austria and Finland (under 7%).

While there is a relationship between the relative number of people who persistently have income below 60% of the median and the risk of poverty in 2008, it is by no means systematic. In particular, Italy, where the proportion of the population with income below the poverty threshold is well below the proportion in Spain or Bulgaria, has a much larger proportion of people with income persistently this low. In Italy, therefore, 72% of those at risk of poverty in 2008 (around three in every four) were persistently at risk - a figure second only in the EU to Luxembourg (75%) - implying that it is more difficult than elsewhere for people in Italy to increase their income above 60% of the median once it falls below this. At the same time, there was also a decline in the proportion at risk over the period, which contributes to the relatively large number at risk who were also at risk in earlier years.

The proportion of the population at persistent risk of poverty relative to the proportion at risk in 2008 was also comparatively large in Cyprus, Slovenia, Estonia and the Netherlands (the former in each case representing around 70% or more of the latter) - though in the Netherlands, in particular, the number concerned was relatively small (only just over 6% of the population). In all of these countries, apart from the Netherlands, the proportion of people at risk of poverty either rose or remained much the same over the period, implying that it is especially difficult for those at risk of poverty to raise their income above the poverty threshold.

Conversely, the proportion at persistent risk of poverty was comparatively small, relative to the proportion at risk in 2008, in Slovakia, Malta and the UK, where only around half of those at risk of poverty in 2008 were persistently at risk over the period. In these countries, therefore, the implication is that it is easier than elsewhere for those with income below the poverty threshold to increase it to above this level.

See Tables