We examine empirically the relation of factor-income inequality and government's redistributive preferences to the extent of redistribution. In the challenging task of measuring taste for redistribution, we utilize the inverse-optimum approach. Our income-inequality and redistribution variables are constructed from the Luxembourg Income Study database, and for our redistributive-preference measure we have collected data from various sources. In addition to traditional linear specifications we use flexible methods. We study 14 advanced countries for approximately four decades and find that factor-income inequality and government's redistributive preferences are associated with the extent of redistribution as suggested by the numerical results of the Mirrlees model.