Bounds on Treatment Effects on Transitions

46 Pages Posted: 16 Jun 2016

See all articles by Johan Vikström

Johan Vikström

Uppsala University; IFAU - Institute for Labour Market Policy Evaluation

Geert Ridder

University of Southern California

Martin Weidner

University College London - Department of Economics

Date Written: April 21, 2016

Abstract

This paper considers identification of treatment effects on conditional transition probabilities. We show that even under random assignment only the instantaneous average treatment effect is point identified. Because treated and control units drop out at different rates, randomization only ensures the comparability of treatment and controls at the time of randomization, so that long run average treatment effects are not point identified. Instead we derive informative bound on these average treatment effects. Our bounds do not impose (semi)parametric restrictions, as e.g. proportional hazards. We also explore various assumptions such as monotone treatment response, common shocks and positively correlated outcomes that tighten the bounds.

Keywords: Partial identification, duration model, randomized experiment, treatment effect

JEL Classification: C14, C41

Suggested Citation

Vikström, Johan and Ridder, Geert and Weidner, Martin, Bounds on Treatment Effects on Transitions (April 21, 2016). USC-INET Research Paper No. 16-16. Available at SSRN: https://ssrn.com/abstract=2796211 or http://dx.doi.org/10.2139/ssrn.2796211

Johan Vikström (Contact Author)

Uppsala University ( email )

Box 513
Uppsala, 751 20
Sweden

IFAU - Institute for Labour Market Policy Evaluation

Box 513
751 20 Uppsala
Sweden

Geert Ridder

University of Southern California ( email )

Kaprielian Hall
Los Angeles, CA 90089
United States
213-740-2110 (Phone)
213-740-8543 (Fax)

Martin Weidner

University College London - Department of Economics ( email )

Gower Street
London WC1E 6BT, WC1E 6BT
United Kingdom

Here is the Coronavirus
related research on SSRN

Paper statistics

Downloads
33
Abstract Views
610
PlumX Metrics