We revisit several theoretical explanations for countercyclical retail pricing. Using a comprehensive data set on retail scanner and consumer panel data, we find evidence that retailers do not widely engage in loss leader pricing, but rather conduct it in a very specific and targeted way.