Volatility of stock price as predicted by patent data: An MGARCH perspective

This paper proposes to model stock price volatility and variations in innovation effort using a Multivariate GARCH structure designed to extract information for risk prediction. The salient feature is that the model order, alongside other parameters, is endogenously determined by the estimation procedures. Using stock prices of U.S. computer firms, it is found that the model can pick up the correlation between the two variables and aid in producing accurate Value-at-Risk estimates.