Research Interests

Asset and Derivatives Pricing, Machine Learning


Option Return Predictability with Machine Learning and Big Dat, with Turan G. Bali, Heiner Beckmeyer, and Florian Weigert
Review of Financial Studies,  2023, 36(9)

Media: alpha architect, LexTech Institute

Abstract: Drawing upon more than 12 million observations over the period from 1996 to 2020, we find that allowing for nonlinearities significantly increases the out-of-sample performance of option and stock characteristics in predicting future option returns. Besides statistical significance, the nonlinear machine learning models generate economically sizeable profits in the long-short portfolios of equity options even after accounting for transaction costs. Although option-based characteristics are the most important standalone predictors, stock-based measures offer substantial incremental predictive power when considered alongside option-based characteristics. Finally, we provide compelling evidence that option return predictability is driven by informational frictions and option mispricing.

Presentations: Virtual Derivatives Workshop PhD Session (2021); Goldman Sachs STS (2021); Hull Tactical Asset Allocation (2021); BVI-CFR Workshop 2021; University of Muenster(2021); 14th Annual Hedge Fund Conference at Imperial College London 2022; SFI Research Days 2022; Federal Reserve Board (2022); Georgetown University (2022); University of Fribourg (2022); University of Liechtenstein (2022); University of Minnesota (2022); Morgan Stanley (2022)

Credit Variance Risk Premium, with Manuel Ammann
European Financial Management, 2023, 29(4); (Open Access at EFM)

Media: alpha architect

Abstract: This paper studies variance risk premiums in the credit market using a novel data set of swaptions quotes on the CDX North America Investment Grade and High Yield indices. We find that returns of credit variance swaps are negative and economically large, irrespective of the credit rating class. Shorting credit variance swaps yields   annualized Sharpe ratios well above their counterparts in other asset classes. The returns remain highly statistically significant when accounting for transaction costs and cannot be explained by established risk-factors and structural model variables. By means of corridor variance swaps, we also dissect the overall variance risk premium into receiver and payer variance risk premiums. We show that credit variance risk premiums are mainly driven by the payer corridor, which is associated with worsening macro-economic conditions. 

Presentations: AFA PhD Poster Session (2019); Finance Research Seminar, University of St.Gallen (2019); Finance Research Seminar, University of Konstanz (2019); SoFiE Financial Econometrics Summer School (2019); Paris Financial Management Conference (2019); 26th Annual Meeting of the German Finance Association Doctoral Workshop (2019)

Commodity Tail Risk, with Manuel Ammann, Marcel Prokopczuk, and Christoph Würsig
Journal of Futures Markets, 2023, 43(2)  (JFM version)

Abstract: In this study, we investigate the cross-section of option implied tail risks in commodity markets. In contrast to findings from equity markets, left and right tail risk implied by option markets are both large. Commodity specific variables exert the largest influence on tail risk, while there is no evidence of systematic commodity factors that are linked to tail risk. Additionally, we find strong links to the equity markets, but also co-movements to macroeconomic factors. Left or right tail risks are largely independent of variance risk premiums. Finally, both left and right tail risk are priced in the cross-section of commodity futures returns. 

Presentations: Finance Seminar at the Leibniz University of Hannover

 Working Papers

Machine Forecast Disagreemen, with Turan G. Bali, Bryan T. Kelly, and Jamil Rahman

Abstract: We propose a statistical model of differences in beliefs in which heterogeneous investors are represented as different machine learning model specifications. Each investor forms return forecasts from their own specific model using data inputs that are available to all investors. We measure disagreement as dispersion in forecasts across investor-models. Our measure aligns with extant measures of disagreement (e.g., analyst forecast dispersion), but is a significantly stronger predictor of future returns. We document a large, significant, and highly robust negative cross-sectional relation between belief disagreement and future returns. A decile spread portfolio that is short stocks with high forecast disagreement and long stocks with low disagreement earns a value-weighted alpha of 15% per year. A range of analyses suggest the alpha is mispricing induced by short-sale costs and limits-to-arbitrage.

Presentations: Georgetown University (2023), Yale School of Management (2023), University of Konstanz (2023)

Option Factor Momentum, with Niclas Käfer, and Tobias Wiest

Abstract: We document profitable cross-sectional and time-series momentum in a broad set of 56 option factors constructed from monthly sorts on daily delta-hedged option positions. Option factor returns are highly autocorrelated, but momentum profits of strategies with longer formation periods are mainly driven by high mean returns that persistently differ across factors. Momentum effects are the strongest in the factors’ largest principal components, consistent with findings for stock factor momentum. Finally, we find a new form of momentum in options markets: momentum in single delta-hedged option returns. Option factor momentum fully subsumes option momentum, whereas option momentum cannot explain option factor momentum. Our findings provide insights into the channels that drive option momentum and have implications for designing profitable option trading strategies.

Presentations: University of Konstanz (2023); University of St.Gallen (2023); 36th Australasian Finance and Banking Conference (2023); 1st Elsevier Finance Conference at FGV EBAPE in Rio de Janeiro (2023); Frontiers of Factor Investing (2024); FMA Europe (2024); 40th International Conference of the French Finance Association; 2nd Structured Retail Products and Derivatives Conference (2024)

Liquidity Provision to Leveraged ETFs and Equity Options Rebalancing Flows: Evidence from End-of-Day Stock Price, with Andrea Barbon, Heiner Beckmeyer, and Andrea Buraschi 

Media: Financial Times, alpha architect 

Abstract: Rebalancing of leveraged ETFs (LETFs) and delta-hedging of equity options by intermediaries are two distinct and economically significant sources of liquidity demands. We show that they induce end-of-day momentum and mean-reversion in returns. While gamma effects are persistent throughout our sample, LETFs effects have decreased over time. We empirically study these effects and their potential drivers. We find that LETF flows attract more liquidity provision and their effects on prices are shorter-lived. Intermediaries can strategically decide the timing of their delta-hedging, resulting in less predictable flows. This shows the benefits of information disclosure on market liquidity and price distortion.

Presentations: NFA Annual Meeting 2021; 5th SAFE Market Microstructure Conference (2021); 37th International Conference of the French Finance Association (2021); FMA Annual Meeting 2021; EFMA Annual Meeting 2021; FMA Conference on Derivatives and Volatility (2021); Goldman Sachs STS (2021); Bank of America Merrill Lynch (2021); Morgan Stanley (2021); Societe Generale (2022); Fidelity (2022); 13th Annual Hedge Fund Research Conference Paris 2022; MFA Annual Meeting 2022; SFS Cavalcade North America 2022; Hofstra Financial Regulations & Technology 2022; WFA Annual Meeting 2022

A Bayesian SDF for Equity Option, with Niclas Käfer, Florian Weigert,  and Tobias Wiest

Abstract: Building on Bryzgalova, Huang, & Julliard (2023), we conduct a Bayesian analysis of linear factor models for the stochastic discount factor (SDF) in the individual equity options market. In both cross-sectional and time-series out-of-sample tests, a Bayesian model averaging SDF outperforms reduced-form benchmark models in pricing option portfolios and option return anomalies. In line with results from stocks and corporate bonds, the space of factors spanning the risks and return drivers in the options market is dense. Notably, the difference between implied and realized volatility, option return momentum, and jump risk emerge as highly likely factors to be included in the SDF.

Presentations: 2nd Structured Retail Products and Derivatives Conference (2024); 17th Annual Conference Behavioural Finance Working Group

An Autoencoder Based Factor Model for Option Returns [Available upon request]