Difference-in-Differences with a Continuous Treatment

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Callaway, Brantly, Goodman-Bacon, Andrew, Sant'Anna, Pedro H. C.
Format: Preprint
Published: 2021
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866909979220377600
author Callaway, Brantly
Goodman-Bacon, Andrew
Sant'Anna, Pedro H. C.
author_facet Callaway, Brantly
Goodman-Bacon, Andrew
Sant'Anna, Pedro H. C.
contents This paper analyzes difference-in-differences designs with a continuous treatment. We show that treatment-on-the-treated-type parameters are identified under a parallel trends assumption analogous to the binary treatment case. However, comparing these parameters across treatments is challenging because parallel trends does not rule out selection bias. We discuss alternative, typically stronger, assumptions that eliminate selection bias. We further show that popular two-way fixed effects estimands admit multiple interpretations, depending on the underlying causal building block, all having important limitations as meaningful summaries of treatment effects. Finally, we introduce alternative estimation procedures that avoid these drawbacks and demonstrate them in an empirical application.
format Preprint
id arxiv_https___arxiv_org_abs_2107_02637
institution arXiv
publishDate 2021
record_format arxiv
spellingShingle Difference-in-Differences with a Continuous Treatment
Callaway, Brantly
Goodman-Bacon, Andrew
Sant'Anna, Pedro H. C.
Econometrics
This paper analyzes difference-in-differences designs with a continuous treatment. We show that treatment-on-the-treated-type parameters are identified under a parallel trends assumption analogous to the binary treatment case. However, comparing these parameters across treatments is challenging because parallel trends does not rule out selection bias. We discuss alternative, typically stronger, assumptions that eliminate selection bias. We further show that popular two-way fixed effects estimands admit multiple interpretations, depending on the underlying causal building block, all having important limitations as meaningful summaries of treatment effects. Finally, we introduce alternative estimation procedures that avoid these drawbacks and demonstrate them in an empirical application.
title Difference-in-Differences with a Continuous Treatment
topic Econometrics
url https://arxiv.org/abs/2107.02637