Herding on Noise: The Case of Johnson Redbook's Weekly Retail Sales Data

Abstract

Recent models of herding suggest that speculators may rationally trade on information unrelated to fundamentals when their trading horizons are short. This study provides an empirical example where this appears to be the case. Johnson Redbook's weekly retail sales figures predicted bond returns for a short time after a significant number of bond traders began purchasing and trading on the data. The significant relationship between the data and bond returns disappeared just after the Wall Street Journal started to report it. Meanwhile, there was little or no change in the relationship between the data and retailers' stock returns, perhaps because the data have long been followed by retail stock analysts, Johnson Redbook's original investor clientele.