Hidden heterogeneity: The distribution of treatment effects of a welfare reform experiment

Crawford School of Public Policy
Photo by Pxhere

Event details

ACDE Seminar

Date & time

Tuesday 22 February 2022


Online via Zoom


Mathias Sinning, Australian National University

This paper studies the effect of Connecticut’s Jobs First welfare reform experiment on the income distribution. We propose a novel approach to estimate various parameters that depend on the joint distribution of potential outcomes, including quantiles of the distribution of treatment effects, the fraction of ‘winners’ and ‘losers’, and average and total gains of the intervention. We also derive a generalized version of quantile treatment effects, which do not rely on a rank invariance assumption. We use baseline data to estimate the rank correlation between potential treatment and control outcomes, and to obtain permutation-specific distributions of treatment effects. We achieve point identification by averaging across permutation-specific distributions within control state outcomes. Our preliminary findings indicate that Jobs First reduced earnings of about 10 per cent of the treated, while increasing transfers of about 70 per cent and incomes of almost all members of the treatment group immediately after its implementation.

Updated:  9 August 2022/Responsible Officer:  Crawford Engagement/Page Contact:  CAP Web Team