Abstract

We propose a parametric approach for directly estimating the optimal portfolio weights based on a fundamental economic theory by employing deep learning techniques. The deep tangency portfolio is a combination of the market portfolio and a deep long-short factor constructed using a large number of characteristics. We apply our approach to the corporate bond market. Albeit acting as a market-hedge portfolio, the deep factor achieves a sizable market price of risk with an out-of-sample annualized Sharpe ratio of 2.08. The deep tangency portfolio outperforms those constructed from commonly used observable or latent factors with an out-of-sample annualized Sharpe ratio of 3.34. In addition, our findings provide further empirical evidence supporting the integration between bond and equity markets.

About the Speaker

Junye Li is a Li Dasan Chair Professor of Finance at the School of Management of Fudan University. He holds a PhD in Economics from Bocconi University (Milan, Italy), and a Master of Engineering in Systems Engineering from Beijing Jiaotong Univeristy.

His research interests include Empirical Asset Pricing, Financial Econometrics, Machine Learning and Financial Data Analytics, and Macro Finance. His recent research has appeared in Review of Financial Studies, Journal of Econometrics, Journal of Financial and Quantitative Analysis, Journal of Money, Credit and Banking, Journal of Business and Economic Statistics, Journal of Banking and Finance, and so on.