The standard poverty lines applied in empirical research tend to be problematic in terms of validity, reliability, ease of application or socio-political credibility. This paper introduces an international version of an alternative method, which originally has been developed for the Netherlands. The approach starts from a detailed expert reference budget for a single person, which is subsequently generalised to other household types and over time. The empirical analyses try to assess whether Esping-Andersen's famous distinction between social-democratic, liberal and corporatist institutional regimes is related to actual differences in the 'production of poverty' in 11 countries, as measured by the generalised budget approach. Bivariate results from the Luxembourg Income Study indicate that liberal regimes (Australia, Canada, UK, USA) attain a substantially higher degree of poverty than representatives of the corporatist type (Belgium, Germany, France). Poverty in the latter group exceeds the level reached by exponents of the social-democratic regime (Denmark, Norway, Sweden) and the hybrid Dutch system. Multi-level analysis, however, shows that much of these differences have to be attributed to the characteristics of individuals, and to the divergent level of prosperity of these countries. The 'pure' effects of the regime type on poverty all run in the direction that was expected on theoretical grounds, but are rather modest. Only the difference between the high poverty rates in the liberal group and the lower incidences in other countries turned out to be statistically significant.
poverty poverty lines budget method social security welfare regimes institutions