Using Google Searches of Firm Products to Detect Revenue Management and Predict Firm Performance

• In Event: Earnings Management III

Abstract

We test whether an external publicly available data source, Google Trends, that summarizes statistics about potential customer searches of firms’ products, is useful for verifying firms’ reported revenues, and for predicting firm performance. We find that an increase in firm product queries (∆SVI) via Google searches nowcasts sales growth, especially in end-user industries, and that ∆SVI predicts abnormal stock returns from fiscal quarter end to the earnings announcement. Furthermore, the deviation between ∆SVI and reported sales growth is useful for detecting revenue management. Potential upward manipulators, MUP—those in the highest quartile reported sales growth but the lowest quartile ∆SVI in each industry-quarter—increase accrued revenues and reserve lower allowance for bad debts, experience lower cash sales, are more likely to conduct M&A activities and issue new debt and equity, and have a higher frequency of SEC enforcement actions for revenue misstatements. Investors do not appropriately discount revenue surprises of MUP firms at the time of earnings announcement; there is on average a negative return reaction one quarter post-announcement, consistent with a market correction of overvaluation. These findings are robust to a battery of robustness tests.

Authors

• Peng-chia Chiu, The Chinese University of Hong Kong

• Xuan Huang, California State University, Long Beach

• Yinglei Zhang, Chinese University of Hong Kong

• Siew Hong Teoh, University of California, Irvine