Graduation Year

Fall 2010

Document Type

Open Access Senior Thesis

Degree Name

Bachelor of Arts



Reader 1

Ananda Ganguly

Reader 2

Fan Yu


Over the Past two and a half years banks have failed at the fastest pace since the Great Depression. These rapidly mounting bank failures have rekindled a debate surrounding the use of fair value accounting, with many arguing that fair value has exacerbated the severity of the recent financial crisis through asset devaluation and the forced sale of assets in an effort to meet capital requirements. This paper seeks to test if an entity’s exposure to fair value which includes assets available-for-sale, trading assets, and loans held-for-sale as a percent of total assets increases the probability of bank failure through testing different prediction models of bank failure that use ratios generated from publicly available Call Report data. Two models are generated from these ratios, one to determine the significance of an entity’s fair value exposure in predicting risk of failure, and the other to determine if a better model can be generated in the absence of the Fair Value Exposure/Total Assets ratio. The first model shows that Fair Value Exposure/Total Assets is a statistically significant ratio, and that the model employing Fair Value Exposure/Total Assets has greater bank failure predictive power than the second model that excludes this ratio. Contrary to expectations, the study determines that greater fair value exposure actually decreases a bank’s risk of failure, rather than increases it. A number of possibilities as to why this may be are presented in the conclusion of the paper.