Testing for differences in high-frequency network connectedness from variance decompositions

We develop a Bayesian Wald-type test for detecting temporal differences in connectedness from variance decompositions, and establish the asymptotic normality of impulse response functions and variance decomposition matrices at fixed local time points. The resulting test statistic is asymptotically $\chi^2$ distributed under the null hypothesis and performs well in Monte Carlo simulations. Applying our methodology to high-frequency options data on sectoral exchange-traded funds (ETFs) from January 2013 to June 2023, we identify and quantify monetary policy surprises following Federal Open Market Committee (FOMC) announcements. Our empirical results indicate that total connectedness among sectoral ETF implied volatilities increases significantly following FOMC announcements, with these elevated connectedness levels persisting for approximately one hour on days with monetary policy surprises. Our findings indicate more than one-fifth of FOMC announcements surprise financial markets.