Do US commercial banks use FAS 157 to manage earnings?

Abstract

Purpose The purpose of this paper is to examine whether commercial banks manage earnings through their use of Level 3 valuation under The Statement of Financial Accounting Standards No. 157 (FAS 157), Fair Value Measurements. To achieve this goal, the authors investigate the association between banks' performance and size with the percentage of Level 3 inputs for valuing assets and liabilities. The FAS 157 requires or permits US listed firms to adopt one of the three levels of inputs to measure and report fair values of their assets and liabilities. Level 1 input is based on quoted prices in the active markets; Level 2 is from the observable markets other than quoted prices; while Level 3 is based on unobservable and firm‐generated inputs. Design/methodology/approach The authors extract information from all the 10Q‐reports submitted by the US commercial banks (SEC Code 6021) for the first two quarters of 2008. Since Level 3 input is subject to the highest level of managerial discretion, this is expected to be an avenue for earnings management. Multivariate regression was applied to test whether the banks' performance and size are associated with a higher level of Level 3 input. Findings Out of 329 quarterly bank observations, the authors find large banks and poor performing banks with lower returns on assets, lower cash flows, and higher amounts of provision for loan losses, use more Level 3 inputs on valuing their assets and liabilities. Research limitations/implications The results suggest that US commercial banks use FAS157 as an avenue for earnings management. The results have implications for the accounting standard setters, banking sector and policy decision makers. Originality/value This is the first paper providing evidence that banks use fair value measurements to manipulate earnings.

Cite this work

No public copy is stored for this record, so only the metadata is cited.

Stored only when both papers are in the corpus — 24 citations exist upstream.

Papers that cite the same references — the relational fallback while embeddings are still backfilling.