spb/wp7_uqo

1\begin{table}[H]2\centering3\caption{Forecast-error variance decomposition (\%), bivariate VAR(5) in ATM IV and RV.}4\label{tab:fevd}5\begin{threeparttable}6\small7\begin{tabular}{@{}rrrrr@{}}8\toprule9& \multicolumn{2}{c}{RV variance due to IV shock} & \multicolumn{2}{c}{IV variance due to RV shock} \\10\cmidrule(lr){2-3}\cmidrule(lr){4-5}11Horizon (days) & RV first & IV first & RV first & IV first \\12\midrule131 & 0.0 & 4.7 & 4.7 & 0.0 \\142 & 3.7 & 10.0 & 4.5 & 2.2 \\155 & 10.7 & 18.8 & 4.5 & 4.1 \\1610 & 15.2 & 23.9 & 4.5 & 4.8 \\1715 & 18.2 & 26.7 & 4.6 & 5.2 \\1820 & 20.2 & 28.5 & 4.6 & 5.4 \\19\bottomrule20\end{tabular}21\begin{tablenotes}[flushleft]\footnotesize22\item \textit{Notes.} Cholesky-orthogonalized forecast-error variance decomposition of the bivariate VAR(5) in standardized ATM IV and RV, averaged over the 20 tickers with the longest histories (mean $N=3{,}622$ days per ticker). ``RV first'' orders RV before IV (RV shocks may move IV contemporaneously, not the reverse); ``IV first'' is the opposite ordering \citep{sims1980macroeconomics,kilian2017structural}. The mean contemporaneous residual correlation is 0.20; mean equation $R^2$: 0.94 (IV) and 0.40 (RV).23\end{tablenotes}24\end{threeparttable}25\end{table}26