ABSTRACT
Researchers have long been interested in the determinants of within-country income redistribution. Persson and Tabellini (1994), Saint-Paul and Verdier (1993), and others, relying on the median voter hypothesis, predict greater redistribution in less equal societies. Perotti (1996), however, rejects this claim empirically, finding no increased propensity to redistribute income in countries with greater levels of inequality (as measured by the income share of the middle class). Milanovic (1999) examines redistributive effects in 17 democracies in terms of Lorenz shifts. Consistent with Perotti’s findings, the author finds the median voter hypothesis unable to explain differences in redistribution across these countries. Ravallion and Lokshin (2000) (building on the model of Bénabou and Ok (2001) and Hirshman’s (1973) theory of a ‘tunnel effect’) argue that preferences for redistribution depend not just on the median voter’s current position in the income distribution but also his/her expectations of future income mobility. Gradstein and Milanovic (2004) present evidence suggesting that democracies (specifically countries with greater voting participation by the general population) – except for the recent experiences in Eastern Europe – tend to favor redistribution. In conclusion, Milanovic (1999) calls for ‘a totally different mechanism to explain redistribution.’
