Volatility Co-Movement

Abstract:     
I estimate a GARCH-based volatility factor model that incorporates market volatility and information from high-frequency data. I find that index and stock volatility co-move more after the stock becomes part of SP500. This effect is characteristic to higher frequencies (i.e. hourly) and it is beyond what is predicted by an increase in return comovement. One proposed hypothesis consistent with the findings argues that volatility comovement is induced by 'trading mechanism noise' such as noise generated during index arbitrage operations. Additional behavioral hypotheses may be supported by my results. Moreover, volatility has more uniform intra-day distribution after the addition.

Keywords: Volatility, factor model, comovement, multiplicative error model, index addition, event study

JEL Classifications: G10, G12, G14

Working Paper Series