Does leaving home make you poor ? Evidence from 13 European countries

This paper examines the extent to which the relationship between leaving home and entry into poverty among young people is causal: that is, how far poverty entry is the result of leaving home, rather than arising from heterogeneity or selection. Using propensity score matching, we estimate the effect of home-leaving on entry into poverty and deprivation, with data from the European Community Household Panel. We find that leaving home does have a causal effect on poverty entry, particularly in Scandinavian countries; cross-national differences are partly, but not fully, explained by differences in destinations on leaving home. NON-TECHNICAL SUMMARY Previous work in the area of youth poverty has found a strong link between young people’s living arrangements and the incidence of poverty. Young people who have left home are more likely to be poor than those who remain living with their parents, and of all the events likely to trigger entry into poverty, the home-leaving event is the most important. Research across the 15 countries of the preenlargement European Union has found that the relationship between home-leaving and poverty is rather modest in Southern European countries (where home-leaving occurs relatively late) but very strong in the Scandinavian countries (where it occurs much earlier). However, it is not clear from previous research whether the observed relationship between homeleaving and poverty is causal (that is, that leaving home causes poverty) or whether it arises as the result of selection. For example, if young people with a certain set of characteristics (such as low educational levels), which pre-disposed them to becoming poor, were more likely to leave home early than other youngsters, we would observe a relationship between home-leaving and poverty, without the home-leaving event actually causing the poverty. This paper uses a statistical estimation technique known as propensity score matching to examine these issues of causality. Under this technique, individuals who are observed to leave home in a particular year are matched with individuals who are identical, or almost identical, on a wide range of characteristics (sex, age, employment status, educational levels, income, family structure, and so on) except that these matched individuals did not leave home in that year. The difference between the two matched samples gives an estimate of the degree to which home-leaving “causes” poverty. Once we control for selection, we find that our estimates of the effect of leaving home on poverty are actually higher than estimates which do not take account of selection: in other words, young people whose characteristics mean they are less likely to experience poverty are more likely to leave home. Differences between countries remain very pronounced, with the effect of leaving home in Spain and Portugal being very small, while it is extremely large in Denmark and (especially) Finland. In Spain and Portugal, young people who leave home are only around 5 percentage points more likely than those who remain at home to enter poverty, while this difference rises to 32 percentage points in Denmark and 55 percentage points in Finland. As a check on the robustness of our estimates, we repeat this analysis examining entry not into poverty, but into two different measures of material deprivation, and find a similar pattern. One possible explanation for the big differences between countries in the effect of leaving home is that in the Southern European countries, young people tend to leave home in order to get married, while in the Scandinavian countries it is much more common for young people to leave home to live alone – and that couples are less likely to be poor than single-person households. However, if we look at entry into poverty among young people who leave home to form a couple, we see a similar pattern: poverty entry rates are very high in Finland, and much lower in Southern European countries. Thus, differences in poverty entry rates may be partially explained by cross-national differences in destinations on leaving home, but they cannot be fully explained in this way.