IdEP Seminar "Rethinking the Welfare State" - Nezih Guner

The US spends about 2.5% of its GDP on non-medical, non-retirement means-tested transfers. These transfers, which cover a wide range of programs, provide income support for low and middle-income households. Can the U.S. economy do better? Is it possible to find welfare-improving ways to support households that are supported by a majority? Answer these questions in a general equilibrium, heterogeneous-agents model with single and married households facing idiosyncratic labor market productivity risk. All households decide how much to work and save, but married couples also decide whether one or two members should participate, taking into account of potential skill losses of females associated to non-participation. Government taxes households according to a non-linear tax schedule and provide transfers.
We find that household value the current transfers, and their potential elimination generates large aggregate welfare losses.
Yet, most households would support eliminating current transfers since losses are concentrated among a small number of households. We also find that a universal basic income program does not improve upon the current system. On the other hand, if per-person transfers are implemented together with a proportional tax, which we label as a negative income tax experiment, it is possible to find a per-person transfer level that can be better than the current system.
Providing per-person transfers to all households is costly, and reducing distortions in the tax system provides the government with additional resources to expand transfers.

The paper is joint work with Remzi Kaygusuz (Sabanci University) and Gustavo Ventura (Arizona State University)

Nezih Guner
Professor at CEMFI