Using Google Searches of Firm Products to Detect Revenue Management

Abstract

We introduce a novel Big Data analytics model to detect upward revenue misreporting.
The model uses freely available Google searches of firm products to provide external
entity business state (EBS) evidence. The veracity of the reported numbers is enhanced
when auditors can obtain external EBS evidence congruent with the reported numbers. The
Google search volume index (SVI) of firm products is a good candidate for such EBS
evidence because it nowcasts (i.e. predicts present) firm sales and is independent of
management control. A large discrepancy such as a high sales growth together with a large
decline in the SVI suggests possible manipulation upwards of revenues. We find that an
indicator variable, MUP, of a firm in the top sales growth quartile and bottom ΔSVI
quartile in each industry-quarter predicts revenue misstatements incrementally to the
F_Score, Discretionary-Revenues model, two alternative upward revenue manipulation
identifiers, and analyst and media coverages. MUP predictability is stronger in end-user
industries and in interim quarters relative to the fourth quarter. We also find corroborating
evidence that MUP firms have lower sales growth persistence, larger increases in accounts
receivables, and lower allowances for bad debts, consistent with their lower revenue
quality.

Keywords: Big Data Analytics, Audit Risk Assessment, Revenue Fraud Detection, Financial Reporting Quality, Auditing Theory of Evidentiary Triangulation

JEL Classification: M41, G14