Research

Working papers

2025

1. arXiv

   Better Understanding Triple Differences Estimators

   Marcelo Ortiz-Villavicencio, and Pedro H. C. Sant’Anna

   2025

   Abs arXiv PDF

   Triple Differences (DDD) designs are widely used in empirical work to relax parallel trends assumptions in Difference-in-Differences (DiD) settings. This paper shows that common DDD implementations – such as taking the difference between two DiDs or applying three-way fixed effects regressions – are generally invalid when identification requires conditioning on covariates. In staggered adoption settings, the common DiD practice of pooling all not-yet-treated units as a comparison group introduces additional bias, even when covariates are not required for identification. These insights challenge conventional empirical strategies and underscore the need for estimators tailored specifically to DDD structures. We develop regression adjustment, inverse probability weighting, and doubly robust estimators that remain valid under covariate-adjusted DDD parallel trends. For staggered designs, we show how to correctly leverage multiple comparison groups to get more informative inference. Simulations highlight substantial bias reductions and precision gains relative to standard approaches, offering a new framework for credible DDD estimation in empirical research.

Publications

2024

1. 

   Heterogenous treatment effects of a voluntary Inclusionary Zoning program on housing prices

   Marcelo Ortiz-Villavicencio, Gonzalo E. Sánchez, and Mario A. Fernández

   Housing Studies, 2024

   Abs DOI

   Housing affordability remains a social and economic issue in New Zealand. Affordability policies such as Inclusionary Zoning have been promoted to streamline land delivery and boost the housing supply. This paper takes the Special Housing Areas (SHA) programme implemented in Auckland, New Zealand, as a case of a voluntary IZ, and explores the effects on the distribution of housing prices. We rely on a Difference-in-Difference framework as the general identification strategy to estimate quantile treatment effects. We use about 175 thousand sales transactions between September 2011 and September 2016. Considering all transactions, the results show that the SHAs program failed to reduce housing prices across the distribution; and, even in some cases, prices increased. We also find that, for new dwellings, the program decreased prices at the lower end of the distribution while increasing them at the upper end. We argue that the mechanism driving these results is the construction of larger dwellings in the upper part of the distribution. As the SHAs did not involve a mandatory mechanism for affordability purposes, we argue that non-affordable dwellings were built and delivered in the first place with the added price premium of a streamlined delivery. Therefore, as an example of a voluntary IZ program, the SHAs may have stimulated supply, but numerous implications arise as improvements in affordability are negligible or questionable.