Event-Study Designs for Discrete Outcomes under Transition Independence

TL;DR

This paper introduces a novel identification method for estimating average treatment effects on the treated in panel data with discrete outcomes, addressing limitations of traditional difference-in-differences approaches.

Contribution

It proposes transition independence as an alternative to parallel trends, incorporating a latent-type Markov model to account for unobserved heterogeneity in short panel data.

Findings

ATT estimates differ significantly from conventional DiD results

The method effectively captures transition dynamics in categorical outcomes

Empirical applications demonstrate the approach's practical relevance

Abstract

We develop a new identification strategy for average treatment effects on the treated (ATT) in panel data with discrete outcomes. Standard difference-in-differences (DiD) relies on parallel trends, which is frequently violated in categorical settings due to mean reversion, out-of-bounds counterfactuals, and ill-defined trends for multi-category outcomes. We propose an alternative identification strategy with transition independence: absent treatment, transition dynamics conditional on pre-treatment outcomes are identical between control and treated groups. To capture unobserved heterogeneity, we introduce a latent-type Markov structure delivering type-specific and aggregate treatment effects from short panels. Three empirical applications yield ATT estimates substantially different from conventional DiD.