Flexible Employment, Risk and the Welfare State

markdownabstract__Abstract__ In most developed countries, the welfare state provides some form of protection for its citizens. The provision of social arrangements differs substantially between countries but, on the basis of institutional characteristics such as the level of spending on welfare programmes, entitlements and labour market participation, countries can be clustered into several ideal-typical welfare state regimes (Esping-Andersen, 1990; 1999); social democratic, corporatist and liberal. This typology is based on differences in how responsibilities are distributed between the state, market and the family. The social security system is central to public welfare, covering social risks such as unemployment, sickness, disability or old age. The social security system depends on public support in society. Public support is important for the legitimacy of social security systems and, in this way, is an important factor in the expansion or decline of the welfare state (Brooks and Manza, 2007; 2006; Hobolt and Klemmemsen, 2005; Burstein, 1998; Page and Shapiro, 1983). Of course, one must not overlook the relevance of all kinds of macro level factors, such as the globalization of financial markets, the costs associated with welfare reform (such as the ‘sunk costs’ in existing public arrangements), sudden shocks from outside the political system and the influence of the media, all of which may also affect the development of the welfare state (see, for example, John, 2006; Pierson, 1998; Pierson, 1994; Pfau-Effinger, 2005), but electoral pressure means that how people perceive social security practices is a very significant factor for social policy. In accordance with democratic theory (cf. Manza and Cook, 2002), it is commonly accepted that social policy preferences impact on levels of social protection, while power relations between social actors (such as the strength of left-wing political parties) may moderate social policy preferences in shaping actual policy outputs (see, for example, Korpi, 2006; Korpi and Palme, 2003)