May 22, 2023

Gatton Chaired Professor of Accountancy Brian Bratten is a prolific researcher whose work focuses on a variety of financial accounting and auditing topics. Normally, his financial accounting and audit projects are separate from one another. But in his recently published paper, "Fair Value Exposure, Auditor Specialization, and Banks’ Discretionary Use of the Loan Loss Provision," the two topics are explored in tandem. "This project was unique and exciting in that my co-authors and I were able to look at how banks use discretion when reporting earnings; and whether auditor industry specialization is associated with curbing this discretion," he said. 

To gather the data, Bratten and co-authors Monika Causholli and Linda A. Myers (University of Tennessee) used bank financial statement data from the Bank Holding Companies Database maintained by the Federal Reserve Bank of Chicago, and auditor data from Audit Analytics.

They found that banks differ in the tools they use to smooth reported earnings depending on the proportion of assets they hold that are marked-to-market (i.e., exposed to fair value financial reporting). "Specifically, banks with more fair value exposure use more discretion over realized gains and losses than other banks, but they use less discretion in estimating the loan loss provision," says Bratten. "However, banks with industry-specialist auditors use less discretion in smoothing earnings with the loan loss provision."

When asked if any of the results were surprising, Bratten mentioned that initially, it was not clear whether banks holding more assets at fair value would reduce the discretion over the loan loss provision. "The loan loss provision is a complex estimate that faces a lot of auditor and regulator scrutiny. While holding assets at fair value affords some discretion in the timing of the recognition of some related gains and losses, banks with more assets at fair value may also face more volatile earnings, and could actually increase the discretion used when estimating the loan loss provision to smooth out a more volatile earnings stream. Instead, we find that they trade-off these two discretionary items."

Bratten maintains that it is relatively straightforward to understand the auditor’s role in curbing manager's discretion over the provision for loan loss. Despite the complexity of the estimate, auditors spend considerable time understanding their clients’ loans and assessing the reasonableness of the loan loss estimate. But understanding the auditors’ role in the trade-off between their clients’ discretion over the loan loss estimate and the discretion in banks’ realization of gains and losses is less straightforward since the latter activity is typically considered “real earnings management” and this study is one of the first to document such an association. "Some contemporaneous work by our colleague Ben Commerford provides an understanding of why auditors may be concerned about real earnings management, based on interviews and experiments conducted with auditors. However this work was not published when my co-authors and I began working on our study," he said. 

The broader implications from this study relate to the debate about fair value accounting, and suggest that one cannot consider its effects in isolation. "Although we observe banks with more fair value exposure use more discretion when recognizing gains and losses, they use less discretion when estimating loan losses. Specialist auditors likewise recognize these trade-offs, and are associated with improved audit quality."

This research will likely be of interest to regulators, members of the banking industry, the academic community, and anyone interested in the earnings management behavior of banks and the financial reporting of complex estimates. Said Bratten, "Our results suggest that certain banks have a different and potentially less visible set of earnings management tools available to achieve desired earnings outcomes. Further, a new accounting standard now directs banks to calculate the loan loss provision using the current expected loss model. This will likely lead to different estimates of the LLP; results from our study are informative that the standard may affect the discretion of banks differently depending on banks’ characteristics."