84 Chapter 4 The results presented in Table 4.6 show that the association between migration and welfare state effort varies according to the social welfare programme and type of mobility. First, in line with our expectations, CEE labour migration is positively and significantly associated with the social spending subdomains of incapacity, family, and unemployment benefits. The coefficients range from an increase of 0.05 percentage points on unemployment spending to a 0.08 percentage point increase in family spending and incapacity as the percentage of CEE labour migrants in the labour force increases by one percent. The coefficients for old age and ALMP spending are not significant. For Western European labour migrants (Table 4.7), we do not find any significant results, except for family spending but this is only at the 10 percent level of significance. For other immigrants (Table 4.6 and 4.7), measured as foreign-born people, we find positive and significant results across almost all spending categories. This is consistent with Fenwick (2019). Overall, the results tend to indicate that immigration is predominantly positively associated with welfare state effort. This provides tentative evidence in favour of the compensation hypothesis as the prevailing mechanism, that is, native workers are compensated through certain social security programmes for the job insecurity labour migration. However, an important alternative explanation for this positive association could be a mechanical effect if migrants drive up social expenditures as beneficiaries of welfare state programmes. Aside from the migration indicators, the results suggest that other structural economic changes have not played a substantial role in European welfare state developments since 2004. For trade openness, we do not find any significant results, for capital openness, there is only a statistically significant and negative association with incapacity spending, and deindustrialisation has only a negative association with family spending. Regarding the political variables, the results show that left-wing governments are positively associated with expenditures on ALMP spending and unemployment benefits, which is in line with classical expectations about partisan politics. However, for the other programmes, the results do not show significant coefficients for left-wing governments. As expected, union density is positively associated with incapacity and unemployment spending. It is likely that domestic institutions aim to prevent reductions in benefit levels of particular programmes, depending on their resources or agenda, and for unions that means focusing on work-related benefits for those already in employment. GDP per capita is negatively and statistically significantly associated with predominately ALMP and unemployment spending, suggesting that as the economy grows there is relatively less spending on these particular welfare state programmes. Gaston and Rajaguru (2013) also find that GDP growth has a significant and negative association with social spending because in an economic downturn, the denominator (GDP) tends to grow more
RkJQdWJsaXNoZXIy MjY0ODMw