Content area
Full Text
(ProQuest: ... denotes non-US-ASCII text omitted.)
Articles
Introduction
In The Three Worlds of Welfare Capitalism, Esping-Andersen devised a typology of welfare states by constructing three indices from seven indicators of social stratification, each of which defines a distinct welfare regime: conservative, liberal and social democratic.1Five of these indicators measure institutional characteristics of social insurance (corporatism, etatism, private pension and health expenditure, and universalism). In addition, the indicator of benefit equality measures the ratio of post-tax maximum to basic benefits of social insurance, and the indicator of poor relief measures the extent to which welfare outside of social insurance relies on means tests. The indicators are based on 1980 data for eighteen Organisation for Economic Co-operation and Development (OECD) countries. Esping-Andersen showed that most of these countries score highly on only one of the three welfare regime dimensions. This clustering suggests the existence of three 'worlds of welfare capitalism'. The typology of three types of welfare regimes (liberal, conservative and socialist) has been very influential.2The classification is used extensively in comparative welfare research to characterize actual welfare states, but it is also frequently employed in explanatory models.3Yet, surprisingly enough, the validity of this model has never been tested in a systematic manner. This article does exactly that: it shows that Esping-Andersen's classification of countries on the basis of these seven indicators is invalid, and proposes and tests an alternative model.
We are not the first to criticize Esping-Andersen's measures. Scruggs and Allan thoroughly re-examined the data underlying six of the seven stratification indicators and replicated Esping-Andersen's indexing and scoring methods for 1980-1 and 1996-2002.4They constructed two new datasets, which diverge from the original measurements on several of the indicators. On the basis of their new data, Scruggs and Allan conclude that no clear-cut typology of welfare states emerges according to the 1980 data, and that over time, the country scores on the three regime dimensions are quite unstable. Hicks and Kenworthy have argued that since the socialist and liberal dimensions are negatively correlated, they are in effect opposite poles of a single underlying dimension.5Their argument that the three worlds of welfare can be represented by two dimensions has actually been endorsed by Esping-Andersen.6Scruggs...