Welcome to the Factor Zoo: Where Mutual Fund Alpha Hides
Abstract: This paper revisits the "no-positive-alpha'' view of active mutual funds. I estimate alphas using both the Fama--French model and LASSO-selected benchmarks drawn from 213 tradable factors. LASSO allows the selected factors to vary across funds and over time, producing a strategy-relative performance measure based on each fund’s revealed exposures. The resulting benchmarks are parsimonious (with a median of seven factors) and economically interpretable. LASSO delivers better in-sample and out-of-sample performance and near-zero alpha for passive funds. The conclusions differ sharply: mean net active alpha is +9 b.p. per month under LASSO, significant and persistent, versus a significant -13 b.p. under the Fama-French model. The cross-sectional median of the LASSO-based active alpha exhibits a downward trend, consistent with the rise of passive investing.
Presented at: the 1st Georgia Ph.D. Symposium on Macroeconomics and Finance; 2026 Goizueta Doctoral Conference (Best Paper Award); Goizueta Brownbag
Another Moment for CAPM
Abstract: The non-Gaussian CAPM states that beta and coskewness are the only two systematic risks: beta captures second-order risk, while coskewness captures third-order risk. However, the empirical performance of coskewness is known to be weak. Using a return-moment expansion, I decompose coskewness into a sum of correlations between sample stock moments and sample market variance. If investors care less about higher-order risks, these components should receive unequal prices. In Fama--MacBeth regressions of individual stocks excluding microcaps, only the lowest-order component, the variance--variance correlation, earns a significant premium; higher-order cross-moment correlations do not. Thus, third-order risk is priced, but in a more nuanced way. A strategy combining beta and the variance–variance correlation earns about 9% annually, unexplained by Fama--French factors.
Presented at: 2026 FMA Annual Meeting (Best Paper Award Semifinalist); Goizueta Brownbag
Mini Flash Crashes and Predictability in High-Frequency Markets
with Tarun Chordia, Donald Lee, and Andrei Kostyrka
Abstract: Mini flash crashes (MFCs) are brief, V-shaped price drops that reverse rapidly. Using one-minute data for nearly 800 large-cap U.S. stocks from 2015-2024, we estimate nonparametric models with a broad set of predictors. MFCs account for about 80% of downside extreme price movements, exhibit a pronounced intraday U-shape, and are predictable one minute ahead. Predictive information is concentrated in recent stock-level volatility and price variation, while lagged liquidity, order flow, and aggregate conditions contribute substantially less. Deeper MFCs do not systematically recover more slowly, and contemporaneous co-occurrences with market and sector ETFs are common. Together with their limited association with identifiable firm-specific news, the results characterize MFCs as predominantly transitory extreme movements arising in predictable high-volatility states.
Presented at: Goizueta Brownbag
The Good, the Bad, the Asymmetric: Evidence from a Conditional Density Model
with Andrei Kostyrka
Abstract: We propose a novel univariate conditional density model and decompose asset returns into a sum of copula-connected unobserved "good" and "bad" shocks. The novelty of this approach comes from two factors: we explicitly model correlation between unobserved shocks and allow for the presence of copula-connected discrete jumps. The proposed framework is very flexible and subsumes other models, such as "bad environments, good environments". Our model shows certain hidden characteristics of returns, explains investors’ behaviour in greater detail, and yields better forecasts of risk measures. The in-sample and out-of-sample performance of our model is better than that of 40 popular GARCH variants. A Monte-Carlo simulation shows that the proposed model recovers the structural parameters of the unobserved dynamics. We estimate the model on S&P 500 data and find that time-dependent non-negative covariance between "good" and "bad" shocks with a leverage-like effect is an important component of total variance. Asymmetric reaction to shocks is present almost in all characteristics of returns. Conditional return distribution seems to be very time-dependent with skewness both in the centre and tails. We conclude that continuous shocks are more important than discrete jumps at least at daily frequency.
Presented at: the 13th International Conference on Computational and Financial Econometrics*; the 5th Econometric Research in Finance Workshop; the 6th International Symposium in Computational Economics and Finance; 2021 Asian Meeting of Econometric Society*; EEA-ESEM 2021 Congress*; brownbags/seminars at Emory Goizueta, the University of Luxembourg*, the HSE University, the Central Economic Mathematical Institute, and the Centre for Econometrics and Business Analytics.
*by coauthor
Bad Tails, Bad Times** [Work-in-progress]
with Justin Raymond Eloriaga
Abstract: This paper studies whether the shape of the option-implied distribution of equity returns contains information about future macroeconomic fluctuations beyond that contained in implied variance. Using risk-neutral variance, skewness, and kurtosis of market returns, we show that option-implied higher moments have state-dependent predictive power for real activity and macroeconomic volatility. The evidence suggests that tail risk matter primarily during recessions, whereas variance is more closely associated with aggregate scale uncertainty and nominal dynamics. To rationalize these patterns, we develop a financial-accelerator model in which risk-neutral tail fragility raises external finance premia when borrower net worth is low. The model links option-implied higher moments to macroeconomic outcomes through priced default losses and provides a mechanism through which the shape of risk becomes macroeconomically relevant in financially fragile states.
Presented at: Goizueta Brownbag
**The views and opinions expressed in this paper do not necessarily represent the views of the IMF, the IMF Executive Board, or IMF Management.