Identifying volatility signals from time-varying simultaneous stock market interaction : conference paper

In the academic literature, the economic interpretation of stock market volatility is inherently ambivalent, being considered an indicator of either information flow or uncertainty. We show in a stylized model economy that both views suggest volatility-dependent cross-market spillovers. If higher volatility in one market leads to higher (lower) reactions in another market, volatility reflects information (uncertainty). We introduce a simultaneous time-varying coefficient model, where structural ARCH-type variances serve two purposes: governing the time variation of spillovers and ensuring statistical identification. The model is applied to data of US and further stock markets. Indeed, we find strong nonlinear, volatility-dependent effects.