Sentiment and Speculation in a Market with Heterogeneous Beliefs (with I. Martin)
[American Economic Review (2022), 112:8: 2465‒2517 ]
Slides, Online Appendix
We present a model featuring risk-averse investors with heterogeneous beliefs. Individuals who are correct in hindsight—whether through luck or judgment—get rich, so sentiment is bullish following good news and bearish following bad news. Sentiment makes extreme outcomes far more important for pricing and has asymmetric effects on left/right-skewed assets. Investors take speculative positions that can conflict with their opinions about fundamentals. Moderate investors are contrarian: they trade against excess volatility created by extremists. All investors view speculation as socially costly; but they think it is in their self-interest, and the market can collapse entirely if speculation is banned.
Trading Under Uncertainty about other Market Participants
[Financial Review (2023), 58:2: 343-367]
I present an asymmetric information model of financial markets that features rational, but uninformed, hedge fund managers who trade against informed and noise traders. Managers are uncertain not only about fundamentals, but also about the composition of informed and noise traders in the market and use prices to update their beliefs about these uncertainties. Extreme news leads to an increase in both types of uncertainty, while it decreases price informativeness. Price reacts asymmetrically to positive and negative news, with higher expected returns at times of increased uncertainty about market composition. The model generates a price-volume relationship that is consistent with established stylized facts. I then extend to a three-period model and study the dynamics of expected returns and volatility.
Managing Other People's Money: An Agency Theory of Financial Management Industry (with K. Tokis, G. Vichos and P. Mourdoukoutas)
[Journal of Financial Research (2024)]
This paper builds an active asset management model to study the interplay between the career concerns of a manager and the prevailing market conditions. We show that fund managers over-invest in market neutral strategies, as these have a reputational benefit. This benefit is smaller in bull markets, when investors expect more managers to use high beta strategies, making their performance less informative about their ability than in bear markets. We empirically verify that flows of funds that follow high beta strategies are less responsive to the fund's performance and that the flow-performance sensitivity is higher in bear markets.
The Double-Edged Sword of 2020 European Short-Selling Bans (with P. Della Corte, R. Kosowski and N.P. Rapanos)
Video
[Under review]
We study the effects of short-selling bans on liquidity and prices through the lens of a model that incorporates a regulator aiming at preventing a significant price decline. Empirically, we evaluate our model's predictions by exploiting the introduction of short-sales restrictions in some but not all European countries during the COVID-19 pandemic, while measuring the share of informed stockholders with institutional ownership. Consistent with our model, bans were detrimental to market liquidity and failed to support average returns, but were effective in limiting large price drawdowns in countries that implemented short-selling bans, especially for stocks with low institutional ownership.
Covering Trades Uncovered (N.P. Rapanos)
[2nd round R&R Journal of Empirical Finance]
Using publicly disclosed short positions in European stock markets, we study what drives short sellers’ covering decisions. Comparing holders of the same stock on the same day, we find that losses on a holder’s other short positions predict a higher probability of reducing the position below the disclosure threshold, even with portfolio composition fixed in advance and same-sector stocks excluded. Such exits are more likely after price increases and when borrowing fees are relatively high, and coincide with positive returns. Positions that are more profitable at the time of covering are followed by lower returns, but not relative to stocks with similar past returns. Our findings suggest that short covering reflects limits to arbitrage at the portfolio level as well as short sellers’ assessment of the stock.
Climate Risks, Near and Far (with R. Bidder and S. Zhang)
This paper adds a temporal dimension to news-based climate-risk measures: using the language of each article, we separate short-horizon from long-horizon climate information. Combining the resulting indices with single-stock dividend futures and analysts' dividend forecasts, we show that the horizon of the news matches the horizon of the cash flows it reprices, with short-horizon news moving near-term dividends and long-horizon news moving distant ones. In periods of intense transition-policy news, markets mark down the long end of firm value while near-term dividends hold steady in both prices and analysts' forecasts. Markets expect firms to wait rather than sacrifice payouts to adapt.
Equity financing constraints and investor sentiment (with Y. Xiao)
Equity-constrained firms earn no unconditional return premium. Conditioning on lagged investor sentiment, however, reveals two offsetting states: using the Linn and Weagley (2024) equity constraint index over 1973–2024, constrained firms underperform by 0.18% per month after high sentiment and outperform by 0.15% after low sentiment. This two-sided pattern differs from the one-sided overpricing implied by short-sale constraints and is consistent with information asymmetry amplifying belief shocks in both directions. The effect is absent for debt constraints, which originate in collateral rather than information. Offsetting states explain conflicting unconditional estimates in the prior literature.
CAPM: Does Investment Horizon Matter?
I examine how return horizon affects asset pricing under a mis-specified CAPM. Assuming i.i.d. returns and validity at short horizons, I derive the implied relation for long-horizon returns, which is nonlinear in beta. Horizon aggregation induces systematic biases in estimated betas and generates pricing errors, with high-beta assets exhibiting negative alphas and vice versa. However, these effects are second-order in market variance and quantitatively small under realistic calibrations.
Intermediaries and Noise Traders
All papers can be provided upon request.