1\begin{table}[tp]2\centering3\footnotesize4\begin{threeparttable}5\caption{Realized correlations between level (L), slope (S) and curvature (C) factor changes by sampling interval and sub-period, 2011--2026 (H4)}6\label{tab:correlations}7\setlength{\tabcolsep}{2pt}8\begin{tabular}{lrrr@{\hspace{4pt}}rrr@{\hspace{4pt}}rrr@{\hspace{4pt}}rrr}9\toprule10Period & \multicolumn{3}{c}{1 minute} & \multicolumn{3}{c}{5 minutes} & \multicolumn{3}{c}{30 minutes} & \multicolumn{3}{c}{Daily (CMT)} \\ \cmidrule(lr){2-4}\cmidrule(lr){5-7}\cmidrule(lr){8-10}\cmidrule(lr){11-13} & L,S & L,C & S,C & L,S & L,C & S,C & L,S & L,C & S,C & L,S & L,C & S,C \\11\midrule122011--2026 & $-$0.93 & $-$0.26 & 0.11 & $-$0.97 & $-$0.14 & 0.05 & $-$0.98 & $-$0.11 & 0.03 & $-$0.89 & $-$0.13 & $-$0.11 \\132011--2015 & $-$0.95 & $-$0.54 & 0.37 & $-$0.98 & $-$0.46 & 0.32 & $-$0.99 & $-$0.42 & 0.30 & $-$0.97 & $-$0.44 & 0.28 \\142016--2019 & $-$0.97 & $-$0.25 & 0.07 & $-$0.99 & $-$0.08 & $-$0.06 & $-$0.99 & $-$0.01 & $-$0.11 & $-$0.87 & 0.03 & $-$0.31 \\152020--2021 & $-$0.93 & $-$0.39 & 0.12 & $-$0.97 & $-$0.44 & 0.28 & $-$0.99 & $-$0.48 & 0.37 & $-$0.91 & $-$0.51 & 0.31 \\162022--2026 & $-$0.88 & $-$0.06 & $-$0.04 & $-$0.95 & 0.08 & $-$0.13 & $-$0.98 & 0.12 & $-$0.16 & $-$0.82 & 0.04 & $-$0.30 \\17\midrule18Tick-noise benchmark & $-$0.84 & $-$0.76 & 0.57 & -- & -- & -- & -- & -- & -- & -- & -- & -- \\19\midrule \multicolumn{13}{l}{\textit{Mean daily realized variance of the factor changes (bp$^2$ per day)}} \\20 & L & S & C & L & S & C & L & S & C & L & S & C \\212011--2026 & 40 & 46 & 351 & 35 & 38 & 231 & 33 & 35 & 194 & 38 & 53 & 282 \\22\bottomrule23\end{tabular}24\begin{tablenotes}[flushleft]\footnotesize25\item \textit{Notes:} Entries are correlations of intraday factor changes aggregated to 1-, 5- and 30-minute intervals, computed as the ratio of the summed realized covariance to the square root of the product of the summed realized variances over all yield days of the period; L,S, L,C and S,C denote the level--slope, level--curvature and slope--curvature pairs. ``Daily (CMT)'' is the correlation of the daily CMT-based factor changes over the same period (Table~\ref{tab:dailyfactors} reports the 2008--2026 values). ``Tick-noise benchmark'' is the correlation that independent price-rounding noise (one tick per contract, variance tick$^2/6$ for a difference) would induce through the mimicking weights, averaged over quarters; the associated noise standard deviation per minute is 0.09, 0.11 and 0.40 bp for level, slope and curvature. The last block reports, with its own L/S/C column labels, the mean daily realized variance of each factor by sampling frequency (2011--2026) and, in the daily columns, the variance of the daily CMT factor changes (2008--2026). Entries are sample moments without standard errors.26\end{tablenotes}27\end{threeparttable}28\end{table}29