A Measure of Financial Statement Benchmarking

Abstract

We propose a pairwise measure of financial statement benchmarking (FSB) that captures the degree of overlap in the financial statement line items reported by two firms. We validate FSB by showing its association with actual peer choices by analysts for performance benchmarking and by corporate boards for compensation benchmarking. We then test the practical implications of FSB in the context of strategic peer selection by these parties. We find that analyst (board) chosen peers with low pairwise FSB are more likely to be optimistically (opportunistically) biased and that the set of peers assembled by an analyst (board) collectively having low FSB is associated with more optimistic and less accurate earnings forecasts (higher CEO overpay). We also document that using peers with high FSB increases the explanatory power of peer-based valuation models. Finally, we demonstrate alternative applications of FSB by aggregating the pairwise measure at the firm level as well as decomposing it into finer financial statement-specific components. Our evidence collectively suggests that FSB can be a relevant tool for those who use benchmarking applications, including practitioners and academics.

Keywords: Financial statement benchmarking, Peer benchmarking, Analyst forecasts, CEO compensation, Peer-based valuation, XBRL

JEL Classification: M41