Volatility Flocking by Cucker-Smale Mechanism in Financial Markets

摘要

We analyze empirical evidence of flocking stock volatilities according to the Cucker-Smale (C-S) mechanism. By using daily realized volatilities of stocks listed on Dow Jones Industrial Average from January, 2007 to December, 2009, we calibrate key parameters such as time-varying coupling strength, communication weight and stochastic noise in coupling of a benchmark C-S model in Bae et al. (Math Models Methods Appl Sci 25:1299-1335, 2015). Our numerical solutions show that the flocking mechanism explains average volatility dynamics better than a stochastic volatility model without the mechanism over the sample period. The model's empirical implications are found from cyclicality of Volatility Flocking Index (VFI), an aggregate measure of differences between volatilities. Results from Granger causality tests after vector autoregression estimation show that VFI helps us predict the implied volatility index, and weighted average return of S&P500 Index.