Xun (Michael) Gong

Erasmus University Rotterdam, Netherlands.

I am a PhD student of the Econometric Institute, Erasmus University Rotterdam. My primary research areas include financial time series modelling, volatility modelling, and artificial neural network. In my PhD career, I extensively investiage the predictability of the implied volailty surface using traditional time series models, dynamic factor models, and nueral network models.

Contact me:

Email: gong@ese.eur.nl

Address: ET-06, Burgemeester Oudlaan 50, Rotterdam

Postal Code: 3062 PA

Publications

Published - Forecasting Crashes: Correlated Fund Flows and Skewness in Stock Returns

Journal of Financial Econometrics, Volume 15, Issue 1, Winter 2017, Pages 36–61,

This article uses the correlation of money flow among mutual funds to forecast the skewness of stock returns. We develop the Flow Driven Skewness measure and show that it is significantly related to future skewness of stock returns. Stocks with higher correlation between their mutual fund owners’ money flow are therefore more “crash prone.” The relation between Flow Driven Skewness and future firm-level skewness is especially important for the largest and the smallest firms in the sample, and remains true for all levels of skewness. The findings are robust to alternative drivers of skewness in stock returns, as well as the choice of calculation, empirical methodology, and sample period.

Working paper - Forecasting of the Implied Volatility Surface Using Put-Call Parity

The (implied) volatility surface is the collection of option-implied volatilities for different strike prices and maturities. Existing literature documents that the volatility surface can be modelled by a limited number of factors using simple regression techniques, and that these factors are persistent. However, regression techniques leave substantial serial correlation in the residuals. We propose an autoregressive model and an "equilibrium correction" style model which uses the information of the deviation from put call parity to directly exploit the serial correlation. We apply the models to S&P500 index options and options of 95 stocks, and show that the new models improve the existing model with a 40% decrease of both in-sample RMSE and out-of-sample RMSFE. The economic significance evaluation shows that the new models can generate higher Sharpe ratios than existing models.

Working paper - Exploring Alternative Factor Structures of the Implied Volatility Surface

The implied volatility surface is referred to as the volatility implied by option prices for different maturities and strikes. We investigate the factor structure of the implied volatility surface of equity options, and propose alternative factor formulations in the spirit of Christoffersen,Fournier, and Jacobs (2017) where we characterize the shape and the dynamics of the equity implied volatility surface in terms of those based on more actively traded S&P 500 index options. We estimate the model using the Kalman Filter and document strong explanatory power of the index options for the equity options, and find that incorporating the index option information can improve the forecasting performance particularly for illiquid options. We examine the economic performance by simulating delta hedge option portfolio based on the forecasts, and we find that the alternative factor structure carries higher option return and higher Sharpe ratio.

Working paper - Implied Volatility Surface of S&P 500 via Deep Learning

Understanding the characterization and dynamics of the implied volatility surface, a collection of option implied volatilities with different strikes and maturities, is important, since option-implied volatility is commonly viewed as a forward-looking measure for the future volatility. Existing literature typically models the shape of IVS by linear restricted regression of Principal Component Analysis, and the dynamics is characterized by Vector-autoregression. We extend these standard analysis by introducing (deep) neural network models. Our findings strongly prefer the neural network models in both IVS characterization and dynamics modeling. In particular, the deep Autoencoder combined with the Gated Recurrent Unit can reduce the forecasting error by -14% and improve the hit rate by 2.5%. In economic significance evaluation, both deep Autoencoder and Gated Recurrent Unit aggregated portfolios generate higher return, higher Sharpe ratio, and higher abnormal return than standard models.

Presentation

International Association for Applied Econometrics (IAAE)

Montreal, Canada
Topic: Forecasting Implied Volatility Surface Using Deviation From Put-Call Parity
2018 June

Netherlands Econometric Study Group (NESG)

Amsterdam, Netherlands
Topic: Forecasting Implied Volatility Surface Using Deviation From Put-Call Parity
2018 March

The Society for Financial Econometrics (SoFiE) Summer School

Chicago, United States
Topic: Factor Structure between the Equity and Index Options
2017 July

Education

Erasmus University Rotterdam

PhD candidate
Specialized in Financial Econometrics
2016 - present

Tinbergen Institute

Research Master
Specialized in Econometrics
2014 - 2016

Tilburg University

Master, Cum Laude
Specialized in Finance
2013 - 2014

Erasmus University Rotterdam

Master
Specialized in Financial Economics
2012 - 2013

Chongqing University

Bachelor
Specialized in Material Chemistry
2007 - 2011

Experience

China Construction Bank

Finance Intern

Provide application information to customers who want to apply for commercial loan. Arrange and classify customers’ application materials of commercial loan. Work with a senior analyst to investigate the credit level of customers.

Fed 2011 – May 2011

Chongqing University Student Union

President

Scheme, organize, and coordinate student activities. Find external fund for students activities. Cooperate with companies to provide internship opportunities for students.

Sep 2007 – May 2011

Skills

Languages, Operating Systems & Tools
  • Python
  • C/C++
  • git
  • linux
  • bash
  • MySQL
Softwares
  • Pandas
  • Numpy
  • Tensorflow
  • Keras
  • SAS
  • Matlab
  • Stata
Platform Development & Administration
  • git
  • Wordpress
Cloud
  • Google Cloud