Applying Regression Discontinuity Design to Social Policy: An Evaluation of the Welfare-to-Work Program in South Korea

Abstract

Social policy studies focusing on poverty reduction attempt to measure poverty reductions rates and poverty gaps, but they do not provide criteria to determine whether a given social policy is a success or failure. In this study, we suggest using regression discontinuity design to establish evaluation criteria and validate estimation results in social programs. Using the dataset from the Korean Welfare Panel, first we conduct, first, a difference-in-differences comparison between welfare recipients under the National Basic Livelihood Security system and nonrecipients whose income falls under the minimum cost of living. Secondly, we establish the counterfactual effects of the program among nonrecipients whose income is below the minimum cost of living and among nonrecipients whose income is above the minimum cost of living. Last, we analyze treatment effects by comparing welfare recipients with income below the minimum cost of living and nonrecipients with income above the minimum cost of living using the regression distribution design method. We argue that the National Basic Livelihood Security system as a welfare-to-work program has positive effects on labor market participation, which has not been established by previous studies.