FINTECH THRUST SEMINAR | (Almost) Model-free Dynamic Mean Quadratic Variation Analysis of Log Returns
Supporting the below United Nations Sustainable Development Goals:支持以下聯合國可持續發展目標：支持以下联合国可持续发展目标：
(Almost) model-free Dynamic Mean Quadratic Variation Analysis of Log Returns
Abstract:
In this paper, we propose an almost model-free dynamic mean quadratic variation (MQV) asset allocation analysis for log returns, which we termed as log-MQV. It has several advantages such as time-consistent optimal investment decision, conforming to investment wisdom, and the explicit closed-form optimal investment strategies for most stochastic models in fi- nance under both complete and incomplete market settings. Through a unified framework, the proposed model can incorporate Ito diffusion models, jump risks, regime switching, and stochastic volatility features. We also illustrate that the proposed framework allows for a data-driven implementation utilizing historical time series data, and this paves the path for a fully model-free robo-advising investment strategy. Extensive numerical and empirical experiments illustrate the performance of the proposed optimal log-MQV portfolio as compared to the log-MV portfolio.
Dr. Cui is an Associate Professor of Financial Engineering at the School of Business at Stevens Institute of Technology. He holds a BS (with first class honors) in Actuarial Science from the University of Hong Kong, a master in quantitative finance, and a PhD in statistics from the University of Waterloo. His research lies in financial engineering, insurance analytics and operations research. His research has been published at leading journals including Mathematical Finance, Finance and Stochastics, SIAM Journal on Financial Mathematics, Econometric Theory, European Journal of Operational Research, Journal of Economic Theory, Journal of Financial Econometrics, Insurance: Mathematics and Economics, and INFORMs Journal on Computing. He is member of the Society of Actuaries and the Society of Financial Econometrics.