Optimal Income Redistribution

Abstract

We study whether redesigning the social security and income tax-transfer systems can deliver significant welfare gains. Our rich quantitative model features both realistic inequality over the lifecycle and the key distortion channels. We find that the welfare of the currently alive is non-concave in pension progressivity, so that gains can be obtained by making pensions regressive, which reduces distortions but raises inequality, or by making them more progressive, which reduces inequality at the cost of higher distortions. The regressive reform delivers larger welfare gains for current generations, and in both reforms the gains come at the expense of future generations. We show that the non-concavity arises from the transition to the new policy. Adding income tax progressivity as a second instrument barely raises the gains of the currently alive at the regressive optimum, but it reduces the losses imposed on future generations, and this is what makes the progressive reform competitive. Finally, raising pension generosity delivers even larger gains to the currently alive, again at the expense of future generations.