Working Papers

When do Treasuries Earn the Convenience Yield? – A Hedging Perspective, with Viral Acharya. R&R. June 2025. SSRN. Slides.  Coverage: Financial Times

We document that the convenience yield of U.S. Treasuries exhibits properties consistent with a hedging perspective of safe assets; i.e., Treasuries are valued highly if they appreciate with poor aggregate shocks. In particular, the convenience yield tends to be low when the covariance of Treasury returns with the aggregate stock market returns is high. A decomposition of the aggregate stock-bond covariance into terms corresponding to the convenience yield, the frictionless risk-free rate, and default risk reveals that the covariance between stock returns and the convenience yield itself drives the effect in a substantive capacity.

Application to recent  movements in Treasury markets:
Tariff War Shock and the Convenience Yield of US Treasuries – A Hedging Perspective, with Viral Acharya. April 2025. SSRN. Coverage: Financial Times, Financial Times 2x WSJ

The Research Behavior of Individual Investors, with Jeff Wurgler. March 2025. SSRN. Coverage: Matt Levine's Newsletter. Allnews.ch WSJ

Browser data from an approximately representative sample of individual investors offers a detailed account of their search for information, including how much time they spend on stock research, which stocks they research, what categories of information they seek, and when they gather information relative to events and trades. The median individual investor spends approximately six minutes on research per trade on traded tickers, mostly just before the trade; the mean spends around half an hour. Individual investors spend the most time reviewing price charts, followed by analyst opinions, and exhibit little interest in traditional risk statistics. Aggregate research interest is highly correlated with stock size, and salient news and earnings announcements draw more attention. Individual investors have different research styles, and those that focus on short-term information are more likely to trade more speculative stocks.  

Discounting Timing Strategies, June 2024.  SSRN. Slides.

Prior work has documented a number of timing strategies that obtain superior Sharpe ratios and alphas relative to underlying buy-and-hold portfolios. I establish a novel fact: the risk-return tradeoff of such strategies deteriorates substantially as the investment horizon lengthens, providing a rationale for the seemingly good returns. The documented effect is large: multifactor alphas are more than halved going from a one-month to a 10-year horizon, emphasizing the importance of establishing portfolio performance at different horizons. I show that such return dynamics arise in an equilibrium model with seasonalities in the volatility of price of risk and expected cash-flow shocks and present a connection with the pricing of dividend strips.

Pre-Announcement Risk.  R&R. August 2024. SSRN. AFA 2021 presentation video. Slides.

I propose and test a new explanation for the pre-FOMC announcement drift puzzle. I show that such a drift arises in a model where investors interpret a given FOMC action based on recent news. If recent news has been good, FOMC announcements are seen as signals about economic conditions; if recent news has been poor, they are seen as signals about the Fed's own policy stance. The pre-FOMC drift represents a risk premium associated with the resolution of uncertainty about announcement interpretation. Consistent with the model, I demonstrate that the market return pre-announcement predicts the interpretation of Fed action.

Precautionary Savings and the Stock-Bond Covariance. March 2025. SSRN. Short slides.

The precautionary savings motive can account for the high-frequency variation in the stock-bond covariance. An increase in the price of risk lowers risky asset prices on account of an increase in risk premia; it lowers bond yields on account of the precautionary savings component. Consequently, times when the price of risk is volatile see a more negative stock-bond covariance. I demonstrate that a model of time-varying price of risk, calibrated to fit equity moments, matches well the evidence regarding both the nominal and real stock-bond covariance. All else equal, a secular shift towards more volatile risk compensation would bring along a decrease in stock-bond covariance.

Announcement Risk Premium Reconsidered, August 2019.

Ai and Bansal (2018) claim that a non-zero risk premium earned in a tight window around macroeconomic announcements is inconsistent with expected utility preferences if aggregate consumption cannot respond to news at a high frequency. I show that the claim results from a misapplication of the Envelope Theorem. I calculate asset prices in their model and show that an announcement risk premium is consistent with expected utility preferences, even if aggregate consumption takes arbitrarily long to adjust to the news. I provide examples from well-studied settings.

Work in Progress

Monetary Policy by Committee, with Ben Matthies, Kaushik Vasudevan, and Will Yang. March 2025.

We study how groups make decisions using over forty years of meeting transcripts from the Federal Open Market Committee. We document that committee members employ distinct models of the macroeconomy, leading to persistent differences in policy preferences. On average, the committee tilts its decision towards the views of members whose models better align with recent data. These tilts induce monetary policy surprises. Guided by our findings, we construct a theoretical framework of group decision-making under model uncertainty in which committee members provide different models. The committee may outperform its best member if members are sufficiently open to each others' perspectives but must balance this benefit against the risk of excess sensitivity to recent data.

Published Work

The Retail Habitat, with Marco Sammon. Journal of Financial Economics, accepted. SSRN. Short slides. 

Retail investors trade hard-to-value stocks. We document a large and persistent spread in the stock-level intensity of retail trading, even allowing for known biases in the attribution of retail trades. Stocks with a high share of retail-initiated trades exhibit higher shares of intangible capital, longer duration cash flows, and a higher likelihood of being mispriced. Consistent with retail-favored stocks being harder to value, we document that these stocks are less sensitive to earnings news and more sensitive to retail order imbalances. Such segmentation of trading intensity arises in a model where informed investors face a trade-off between the benefits of hiding their trades within noisy retail investor order flow and the costs of producing information about the fundamentals of hard-to-value stocks.

1930: First Modern Crisis, with Gary Gorton and Tyler Muir. Financial History Review. December 2023. SSRN.

Modern financial crises are difficult to explain because they do not always involve bank runs, or the bank runs occur late. For this reason, the first year of the Great Depression, 1930, has remained a puzzle. Industrial production dropped by 20.8 percent despite no nationwide bank run. Using cross-sectional variation in external finance dependence, we demonstrate that banks' decision to not use the discount window and instead cut back lending and invest in safe assets can account for the majority of this decline. In effect, the banks ran on themselves before the crisis became evident.

Stock Market Stimulus, with Robin Greenwood and Jeff Wurgler. Review of Financial Studies. April 2023. SSRN. Slides. The Economist. WSJ.

We study the stock market effects of the arrival of the three rounds of “stimulus checks” to U.S. taxpayers and the single round of direct payments to Hong Kong citizens. The first two rounds of U.S. checks appear to have increased retail buying and share prices of retail-dominated portfolios. The Hong Kong payments increased overall turnover and share prices on the Hong Kong Stock Exchange. We cannot rule out that these price effects were permanent. The findings raise novel questions about the role of fiscal stimulus in the stock market.

Mobile Collateral versus Immobile Collateral, with Gary Gorton and Tyler Muir. Journal of Money, Credit and Banking. February 2022. SSRN.

The financial architecture prior to the recent financial crisis was a system of mobile collateral. Safe debt, whether government bonds or privately-produced bonds, i.e., asset-backed securities, could be traded, posted as collateral, and rehypothecated, moving to their highest value use. Since the financial crisis, regulatory changes to the financial architecture have aimed to make collateral immobile, most notably with the BIS liquidity coverage ratio (LCR) for banks which requires that (net) short-term (uninsured) bank debt (e.g. repo) be backed one-for-one with high-quality bonds. We evaluate this immobile capital system with reference to a previous structurally identical regime which also required that short-term bank debt be backed by Treasury debt one-for-one: the U.S. National Banking Era. The experience of the  U.S. National Banking Era suggests that the LCR is unlikely to reduce financial fragility and may increase it.

The Run on Repo and the Fed's Response, with Gary Gorton and Andrew Metrick. Journal of Financial Stability, June 2020. SSRN.

The Financial Crisis began and accelerated in short-term money markets. One such market is the multi-trillion dollar sale-and-repurchase ("repo") market, where prices show strong reactions during the crisis. The academic literature and policy community remain unsettled about the role of repo runs, because detailed data on repo quantities is not available. We provide quantity evidence of the run on repo through an examination of the collateral brought to emergency liquidity facilities of the Federal Reserve.  We show that the magnitude of repo discounts ("haircuts") on specific collateral is related to the likelihood of that collateral being brought to Fed facilities.

Other Papers

Collateral damage, with Gary Gorton. Banque de France Financial Stability Review, April 2018. En français.

A financial crisis is an event in which the holders of short‑term debt come to question the collateral backing that debt. So, the resiliency of the financial system depends on the quality of that collateral. The authors show that there is a shortage of high‑quality collateral by examining the convenience yield on short‑term debt, which summarises the supply and demand for short‑term safe debt, taking into account the availability of high‑quality collateral. They then show how the private sector has responded by issuing more (unsecured) commercial paper at shorter maturities. The results suggest that there is a shortage of safe debt now compared to the pre‑crisis period, implying that the seeds for a new shadow banking system to grow exist.

Genes under weaker stabilizing selection increase network evolvability and rapid regulatory adaptation to an environmental shift, with Pedro Bordalo and Bernardo Lemos. Journal of Evolutionary Biology, August 2016.

Regulatory networks play a central role in the modulation of gene expression, the control of cellular differentiation, and the emergence of complex phenotypes. Regulatory networks could constrain or facilitate evolutionary adaptation in gene expression levels. Here, we model the adaptation of regulatory networks and gene expression levels to a shift in the environment that alters the optimal expression level of a single gene. Our analyses show signatures of natural selection on regulatory networks that both constrain and facilitate rapid evolution of gene expression level towards new optima. The analyses are interpreted from the standpoint of neutral expectations and illustrate the challenge to making inferences about network adaptation. Furthermore, we examine the consequence of variable stabilizing selection across genes on the strength and direction of interactions in regulatory networks and in their subsequent adaptation. We observe that directional selection on a highly constrained gene previously under strong stabilizing selection was more efficient when the gene was embedded within a network of partners under relaxed stabilizing selection pressure. The observation leads to the expectation that evolutionarily resilient regulatory networks will contain optimal ratios of genes whose expression is under weak and strong stabilizing selection. Altogether, our results suggest that the variable strengths of stabilizing selection across genes within regulatory networks might itself contribute to the long-term adaptation of complex phenotypes.