How to Design an Optimal Soft' WelfaretoWork Program
Welfare-to-work programs are government programs aimed at the poor and the jobless. Their objective is to provide income support and assist welfare recipients to obtain or prepare for employment. A range of diverse welfare-to-work programs was implemented across US states over time: some programs were more focused on assisting individual job search, others on education and training and the rest on moving the individual as soon as possible into work. Therefore, states took very heterogeneous approaches to the design of welfare-to-work programs. One natural question thus would be: "what explains this heterogeneity?". This is what Nicola Pavoni (Department of Economics), Ofer Setty (University of Tel Aviv) and Giovanni Violante (New York University) investigate in their recent paper The Design of 'Soft' Welfare-to-Work Programs, published in Review of Economic Dynamics.
In the paper, the authors model the welfare-to-work programs as contracts offered by the government to unemployed agents in an environment with moral hazard where the provision of assistance interferes with individual incentives to exert effort for finding and retaining a suitable job. In order to tackle the hidden action problem, governments use a wide range of policy instruments (e.g., job-search, assisted search, mandated work) in combination with welfare benefits, in order to minimize the costs of delivering promised utility to the participants. The authors restrict attention to a particular class of welfare-to-work programs that they call 'soft'. Soft welfare programs do not use any form of punishment or sanction on the participant. In essence, unlike many unemployment insurance programs, the generosity of the program, measured by the utility promised to the unemployed, cannot decline along the program.
A key contribution of the paper is to analyze the optimal design of welfare programs and the trade-offs among the available policy instruments. Also, the authors introduce to the standard set-up two additional ingredients, besides the job search technology. Firstly, they contemplate the presence of a secondary sector (e.g. government agencies, non-profit institutions, or community service organizations) that is less productive than the (primary) one used in market-sector jobs but that, like the latter, requires effort to yield output. Secondly, they introduce an assisted job search technology that allows the unemployed to defer their job search to an agency at a cost. This technology frees up time from search to either work or rest.
The main result of the paper is the identification of the policy that is best to implement for every combination of two crucial variables that fully identify the relevant 'state' of the worker: (i) the level of generosity of the program (summarized by the promised utility) and (ii) the skill level of the unemployed (their human capital). In sum, the generosity of the program and the skill level of the unemployed agent determine the optimal policy instrument to be implemented. The authors also provide a handy graphical representation of the policy assignment outcome. This analysis could provide a framework to rationalize for the heterogeneity of optimal policies implemented in different US states according to the desired generosity of the program and the skill distribution in the state. Restricting attention to 'soft programs' - contracts that make no use of punishments or sanctions - allows a fully analytical characterization of the optimal program.