Can an Accounting-Based Duration Model Effectively Measure Interest Rate Sensitivity?

Abstract

I examine whether a duration transformation of financial accounting information adequately captures interest rate sensitivity in financial institutions. First, I calculate an ex-ante measure of interest rate sensitivity in complex banking institutions based upon a model developed by bank supervisors at the Federal Reserve Board of Governors; the duration measure is a combination of financial accounting information and duration proxies. Second, I test for an association between this ex-ante measure of interest rate sensitivity and both accounting and capital market ex-post measures of interest rate sensitivity. The results show a predictable relationship between ex-ante modeled interest rate sensitivity and banks’ ex-post accounting performance and stock returns given observed changes in market interest rates. I conclude that an interest rate sensitivity model based upon a combination of financial accounting information and duration proxies can be used to predict interest rate sensitivity in complex banks, including publicly-traded bank holding companies. The issue is important because interest rate risk management is integral to financial institutions, and its reliable measurement using publicly-available accounting data is relevant to bank managers, researchers, market participants, and regulators.