The Passive-Ownership Share Is Double What You Think It Is [Paper] (with Marco Sammon)
Journal of Financial Economics (2024) 157(1): 1-22
Each time a stock gets added to or dropped from an index, we ask: “How much money would have to be tracking that index to explain the huge spike in rebalancing volume we observe on reconstitution day?” While index funds held 16% of the US stock market in 2021, we put the overall passive ownership share at 33.5%. Our number is twice as large because it reflects both index funds as well as other passive investors, such as institutional investors with internally managed index portfolios and closet-indexing active managers.
A New Test of Risk Factor Relevance [Paper] (with Samuel Hartzmark and Abigail Sussman)
Journal of Finance (2022) 77(4): 2183-2238
For an asset-pricing model to actually explain market fluctuations, investors must care about the model’s risk factors. The standard approach is to just assume that, if a model fits the data, then investors must be a) recognizing correlations with the model’s risk factors, b) preferring assets with lower correlations, and c) adjusting their demand accordingly. We survey investors and find no support for any of these assumptions.
The Ex Ante Likelihood of Bubbles [Paper, Online Appendix]
Management Science (2022) 69(2): 1222-1244
Market participants are swimming in a sea of psychological biases and trading constraints. Yet, in spite of this, large pricing errors such as speculative bubbles are rare. Why? How often should we expect the limits of arbitrage to bind? This paper proposes a new model, which hinges on the transformative power of exponential growth, that successfully predicts where future bubble episodes are most likely to occur.
Estimating the Anomaly Base Rate [Paper] (with Andreas Neuhierl and Michael Weber)
Journal of Financial Economics (2021) 140(1): 101-126.
The academic literature contains hundreds of statistically significant cross-sectional predictors, causing many to question whether we’re using the right statistical tests. But, here’s the thing: even if a researcher does use the right tests, he will still draw the wrong conclusions if he starts out with the wrong priors. So, which priors should a researcher use? We propose a new statistical approach to answer this question.
The Sound of Many Funds Rebalancing [Paper] (with Vyacheslav Fos)
Review of Asset-Pricing Studies (2021) 11(3): 502-551.
RAPS 2022 Best Paper Award
We propose a new source for demand noise: complexity. In modern financial markets, different investors often end up trading the same asset for very different reasons. So, even if you are fully rational and each individual trading rule is very simple, an asset might still appear to realize random demand shocks because the interactions between these various simple trading rules are too computationally complex to predict.
Sparse Signals in the Cross-Section of Returns [Paper] (with Adam Clark-Joseph and Mao Ye)
Journal of Finance (2019) 74(1): 449-492.
We use the LASSO to make 1-minute-ahead return forecasts for each stock. The LASSO boosts out-of-sample predictive power by choosing predictors that trace out the consequences of unexpected news announcements. What’s more, the size of the LASSO’s best-fit penalty parameter, 
per month, shows that this success is not just the result of loading on traditional monthly predictors only faster.
Misinformed Speculators and Mispricing in the Housing Market [Paper] (with Chris Mayer)
Review of Financial Studies (2016) 29(2): 486-522.
When you buy a home, you have to tell the county clerk where to mail your property tax bill. We use this information to identify both local and out-of-town second-house buyers in 21 US cities from 2000 to 2008. We find that demand from out-of-town (but not local) buyers predicts future house-price booms. And, we introduce a novel identification strategy to show that this effect isn’t being driven by reverse causation.