Using a data set that records banks’ ongoing requests of information from small commercial borrowers, we examine when banks use financial statements to monitor borrowers after loan origination. We find that banks request financial statements for half the loans and this variation is related to borrower credit risk, relationship length, collateral, and the provision of business tax returns, but in complex ways. The relation between borrower risk and financial statement requests has an inverted U-shape; and tax returns can be both substitutes and complements to financial statements, conditional on borrower characteristics and the degree of bank–borrower information asymmetry. Frequent financial reporting is used to monitor collateral, but only for non–real estate loans and only when the collateral is easily accessible to lenders. Collectively, our results provide novel evidence of a fundamental information demand for financial reporting in monitoring small commercial borrowers and a specific channel through which banks fulfill their role as delegated monitors.

Lending concentration features prominently in models of information acquisition by banks, but empirical evidence on its role is limited. Using bank-level loan exposures, we find banks are less likely to collect audited financial statements from firms in industries and regions in which they have more exposure. These findings are stronger in settings in which adverse selection is acute and muted when the bank lacks experience with an exposure. Our results offer novel evidence on how bank characteristics are related to the type of financial information they use and support theoretical predictions suggesting portfolio concentration reveals a bank's relative expertise.

ABSTRACT We use a proprietary data set of financial statements collected by banks to examine whether economic growth is related to the use of financial statement verification in debt financing. Exploiting the distinct economic growth and contraction patterns of the construction industry over the years 2002–2011, our estimates reveal that banks reduced their collection of unqualified audited financial statements from construction firms at nearly twice the rate of firms in other industries during the housing boom period before 2008. This reduction was most severe in the regions that experienced the most significant construction growth. These trends reversed during the subsequent housing crisis in 2008–2011 when construction activity contracted. Moreover, using bank‐ and firm‐level data, we find a strong negative (positive) relation between audited financial statements during the growth period, and subsequent loan losses (construction firm survival) during the contraction period. Collectively, our results reveal that macroeconomic fluctuations produce temporal shifts in the overall level of financial statement verification and temporal shifts in verification are related to bank loan portfolio quality and borrower performance.