A framework for the decomposition of poverty differences with an application to poverty differences between countries

We propose an Oaxaca-Blinder-like decomposition of poverty differences. The decomposition is based on a parametric model of the income distribution and can be used to decompose differences in poverty rates across countries or years. Poverty differences are decomposed into differences in the underlying distribution of poverty-relevant characteristics and differences in the incidence of poverty conditional on these characteristics. We illustrate our method by comparing levels and patterns of relative poverty in the USA, Great Britain and Germany during the 1990s. Our results suggest that the higher aggregate poverty rates in the USA and in Britain relative to Germany were mostly accounted for by higher poverty rates conditional on characteristics, which were partly offset by a more favourable distribution of poverty-relevant characteristics, in particular higher employment rates.This paper is part of the research programme of the TMR Network ‘Living Standards, Inequality and Taxation’. Financial support from the European Union (Contract #ERBFMRXCT980248), the UK Economic and Social Research Council, the University of Essex, and the Deutsche Forschungsgesellschaft (DFG) is gratefully acknowledged. The data used in this study were made available by Cornell University (Cross-National Equivalent File), the University of Michigan (Panel Study of Income Dynamics), the UK Data Archive (British Household Panel Survey), and the German Institute for Economic Research (German Socio-Economic Panel). Martin Biewen would like to thank the Institute for European Studies and the Department for Policy and Management at Cornell University, in particular Jonas Pontusson, Richard Burkhauser and Dean Lillard, for their hospitality and support. We are also grateful for comments by Bernd Fitzenberger, Joachim Winter, Christoph M. Schmidt, an anonymous referee and seminar participants in Essex, Heidelberg and Mannheim. Last but not least, we thank Nick Cox for providing us with updated versions of his Stata programs for drawing quantile plots.First version received: May 2003/Final version received: December: 2003