1Decentralisation and the Poor

I study patterns of decentralisation and pro-poor policy outputs in 67 countries in 1996. The wide variety of indicators of decentralisation clustered around fiscal, administrative, and political dimensions, and these dimensions had independent relationships with social policy. Consistent with much of the decentralisation literature, administratively decentralised countries showed greater attention to social spending. There were two surprises. First, fiscal decentralisation showed no relationship to social policy, and second, politically decentralised countries spent less on social policy. In particular, the negative relationship associated with political decentralisation suggests a need to look deeper into the impact of political decentralisation on the ability of the poor to advance demands. _____________________________________________________________________ Decentralisation biases political conflict. By redrawing the political-administrative division of the state, decentralisation changes the geographical boundaries of political contestation, alters the relative power of different actors, and changes the groups available for cross-class alliances.1 The current project seeks systematic cross-national patterns in the ways that shifting resources and power from the centre to the local level affects policy outputs. The project proceeds by examining a specific policy area, pro-poor policy, which is likely to pitch wealthy groups against poor groups. The chief hypothesis is that centralisation provides the poor with certain organisational and material advantages that they either lack at the local level or possess only in scattered jurisdictions. In contrast, the rich do not depend o