The Benefits of Transaction-Level Data: The Case of Nielsen Scanner Data

Abstract

This study uses Nielsen scanner data to illustrate the benefits of transaction-level data. Specifically, we explore whether granular consumer purchases contain incremental value-relevant information about the corresponding manufacturers. Using weekly consumer purchases data generated by point-of-sale systems from 2006 to 2018, which capture around $2 trillion of U.S. retail sales, we construct a measure of aggregated consumer purchases at the firm-quarter level, and find that it strongly predicts manufacturer revenues. In addition, analyst forecasts of revenues have predictable errors and revisions, which implies that analysts do not fully incorporate the information in consumer purchases in a timely manner. Exploring investment implications, we find that hedge portfolios that buy (sell) stocks of firms with high (low) abnormal consumer purchases generate annualized returns on the magnitude of 14% to 19% depending on specification. This return predictability holds after controlling for risk factors and firm characteristics, and is robust across time. Finally, about 36% of the quarterly hedge returns are concentrated over the three-day window around earnings announcements, which suggests that the returns result from the correction of biased expectations rather than risk. These findings suggest that consumer purchases convey useful insights into firm fundamentals, shedding light on the benefits of using transactional data for market participants.