Intra-EU Labour Mobility and the Welfare State 75 For welfare state spending we disaggregate it into five subdomains: old age, incapacity, family, active labour market policies, and unemployment.11 Each category is measured as a percentage of gross domestic product (GDP). The data on replacement rates comes from two different sources. First, we use a net minimum income replacement rate from the Social Assistance and Minimum Income Levels and Replacement Rates Dataset (Wang & van Vliet, 2016a; 2016b). The indicator is based on data from the ‘Social Assistance and Minimum Income Protection Dataset’ (Nelson, 2013) and elaborates on the concept of net income replacement rates as used in the Unemployment Replacement Rates Dataset (van Vliet & Caminada, 2012). The replacement rate is an average of three household types (single person households, lone parent households with two children, and two parent households with two children). The second replacement rate is a net unemployment benefit replacement rate, developed from the OECD’s Tax-Benefit Models (OECD, 2017a, 2018) and updated data from the OECD (2024a). We use the initial period of unemployment and average across single person households and lone parent households with two children. Both replacement rates use the net average wage (AW) for the denominator. The sample time frame in which this study takes place (2004 – 2017), means that the AW is more appropriate as the methodology has been updated to represent a more modern average worker. For almost all countries, the unemployment replacement rate is greater than the social assistance replacement rate. The trends for replacement rates are more heterogeneous between countries than spending. Switzerland has the highest unemployment replacement rate, but the lowest social assistance replacement rate, while the United Kingdom has low levels for both, and Denmark has relatively high levels for both, Sweden has seen a considerable decrease in both replacement rates. Finally, Greece does not provide social assistance. Table 4.1 provides some descriptive statistics for the seven dependent variables. 11 Due to the re-categorisation of Denmark’s unemployment spending from public to voluntary private by the OECD meaning public spending on unemployment appears as 0, we used spending data on unemployment from the ‘Comparative Political Data Set 19602015’ (Armingeon et al. 2017) and the Nordic Health and Welfare Statistics (NHWSTAT) (2024) for Denmark.
RkJQdWJsaXNoZXIy MjY0ODMw