Bank risk proxies and the crisis of 2007/09: a comparison

ABSTRACT

The global financial crisis has again shown that it is important to understand the emergence and measurement of risks in the banking sector. However, there is no consensus in the literature which risk proxy works best at the level of the individual bank. A commonly used measure in applied work is the Z-score, which might suffer from calculation issues given poor data quality. Motivated by the variety of bank risk proxies, our analysis reveals that nonperforming assets are a well-suited complement to the Z-score in studies of bank risk.

KEYWORDS:

• Banking
• financial institutions
• risk proxies
• Z-score

JEL CLASSIFICATION:

• G21
• G28
• G32

Acknowledgment

We thank one anonymous referee for constructive feedback. All errors are our own.

Disclosure statement

No potential conflict of interest was reported by the authors.

Notes

¹ Some only use nonperforming loans.

² FDIC item roa.

³ FDIC item eqv.

⁴ FDIC items: p3asset, p9asset, naasset and ore.

⁵ FDIC item asset.

⁶ Allowances is equivalent. The FDIC item lnatres.

⁷ FDIC item elnatr.

⁸ We winsorize one massive outlier value (−12 018) to the next value of the distribution.

⁹ Hasan and Wall (2004) also find that the nonperforming loans ratio is a good proxy for LLR.