The recent rise in inequality in the distribution of disposable income in many, although not all, countries has led to a search for explanations, particularly since for much of the postwar period falling inequality has been the norm. In OECD countries, the cause has typically been identified as rising wage dispersion, coupled with persistent unemployment in Europe, but changes in the government budget can also be important. This paper is concerned with the role of the government budget, particularly taxes and transfers, in explaining the evolution of the distribution of disposable income. Do differences in welfare states across countries explain the differing evolution of the final (post-transfer, post-tax) distribution? Have active policy changes contributed to offsetting rising market inequality, or have they engendered rising final inequality? To the extent that changes in demographic structure, such as the ageing of the population, have intensified budgetary pressures, have governments been forced to cut back on the generosity of their welfare states?The first section of the paper reviews the statistical evidence available from official and other sources about the redistributive impact of the government budget, taking five OECD countries where there is a time-series of studies covering the 1980s and the 1990s. These findings need to be interpreted in the light of an analytical framework, and this is the subject of the second section. I set out a simple framework within which we can explore the distributional implications of different responses to changes in economic conditions and the different elements - on both supply and demand sides - influencing the choice of response. The actual policy changes observed in the five European countries and the United States are summarized in the third section, where I take unemployment benefits and personal income taxation as two case studies. What changes have been made in policy structures and parameters over the 1980s and 1990s? What (if anything) was said in advance about their likely redistributive impact? What can be concluded about the actual impact?