Research

The Effect of High-Tech Clusters on the Productivity of Top Inventors: Comment (pdf, conditionally accepted Jan. 2026, American Economic Review)

I show that Moretti (2021) has two major problems. The event study uses an incorrect model, and the IV regressions are based on a coding error. Correcting both problems gives null results, so the baseline elasticity may not be causal.
I also document eight other problems in the appendix, including unreproducible code; the citation results using log(y+0.00001); and a coding error in the results for heterogeneity by cluster size.

Infrastructure costs: A replication analysis of Brooks and Liscow (2023) (I4R discussion paper), November 2025.

The authors find that the cost per mile of the US interstate highway tripled from the 1960s to the 80s. I find that the results are robust.

How many jobs can be done remotely? A reanalysis of Dingel and Neiman (2020) (I4R discussion paper), May 2025.

The authors report that 37% of jobs in the US can be performed at home. But small changes to the classification algorithm give estimates of 25% and 59%.

Comment on “Supply Constraints do not Explain House Price and Quantity Growth Across U.S. Cities” (pdf), May 2025.

I show that the equations in Louie et al. (2025) are incorrect, because they do not use exogenous variation in housing demand.

Can we detect the effects of racial violence on patenting? Reanalyzing Cook (2014). Published in Econ Journal Watch, Oct. 2024.

Spillover effects of cash transfers: Reanalysis of Egger et al. (2022) (pdf, code), September 2024.

As part of GiveWell’s evaluation of GiveDirectly, I reanalyzed Egger et al. (2022) on the general equilibrium effects of a cash transfer program in Kenya. I find mixed evidence that spillovers accrue to richer households. The main findings are robust.

Do gasoline price shocks during adolescence reduce driving as an adult? A replication exercise (pdf, code), June 2024.

I reanalyze the results in Severen and Bethem (2022). I find supporting evidence for the claim that gas price shocks reduce adult driving, with different effects for price increases and decreases.

Administrative procedures as tax enforcement tools, with Jeff Hicks and Wei Cui, Economics Letters, April 2024.

Why do Chinese firms report income just above a threshold that doubles their tax rate?

Do subway openings reduce air pollution? A replication exercise (pdf, I4R discussion paper, code), March 2024.

I reanalyze the results in Gendron-Carrier et al. (2022). I find that the main result (new subways in high-pollution cities reduce pollution) is robust, while the effects for low-pollution cities are mixed.

Replicating the literature on prefecture-level meritocratic promotion in China. Published in Research and Politics, Feb. 2024.

A popular idea is that Chinese mayors are promoted based on GDP growth. I show that the evidence in the literature is not robust.
September 2023 version: html, pdf, code
2020 version

Can we detect the long-run effects of the measles vaccine? Replicating Atwood and Pearlman (2023) (pdf, code), February 2024.

I run the same analysis from my comment on Atwood (2022) in the context of the Mexico measles vaccine, and again find that the results are explained by trends.

The Long-Term Effects of Measles Vaccination on Earnings and Employment: Comment (html, pdf, I4R discussion paper, code), January 2024.

I run an event study and find that the results in Atwood (2022) are explained by trends, instead of a treatment effect of the vaccine.
Extended Commentary on Atwood (2022), (pdf, code): additional results on the the event study, heterogeneous effects, plotting the raw data, redoing the disease event study, and more.

Does meritocratic promotion explain China’s growth? (pdf, code), 2020.

Using my own data, I find no evidence that prefecture leaders are promoted based on GDP growth.