1\begin{table}[htbp]\centering2\caption{Within-product cross-banner log price gaps (max$-$min), tiers A+B, storefront channel}\label{tab:dispersion}3\small4\begin{adjustbox}{max width=\textwidth}5\begin{tabular}{lrrrrrrr}6\toprule7Sample & Group-weeks & Median & P25 & P75 & Mean & Share $\geq$ 5\% & Share $\geq$ 10\% \\8\midrule9All group-weeks & 30\,042 & 0.154 & 0.059 & 0.279 & 0.189 & 0.785 & 0.633 \\10Same company (two banners) & 27\,939 & 0.151 & 0.058 & 0.268 & 0.186 & 0.782 & 0.627 \\11Different companies & 2\,103 & 0.183 & 0.080 & 0.337 & 0.237 & 0.825 & 0.701 \\12\bottomrule13\end{tabular}14\end{adjustbox}15\begin{minipage}{\textwidth}\footnotesize\vspace{2pt}16Weekly posted price = duration-weighted modal price. Residual s.d. of the product-week/banner two-way fixed-effects model: 0.115 (N = 67,732). AR(1) of banner-pair gaps: 0.663 (s.e. 0.014); placebo with banner labels permuted within group-week: 0.004 (s.d. 0.011). Share of banner pairs ($\ge$ 4 weeks) whose sign never flips: 0.46.17\end{minipage}18\end{table}19