Using Quantile Regression to Estimate Capital Buffer Requirements for Japanese Banks

This paper investigates the impact of extreme fluctuations in bank asset values on the capital adequacy and default probabilities (PD) of Japanese Banks. We apply quantile regression analysis to the Merton structural credit model to measure how capital adequacy and PDs fluctuate over a 10 year period incorporating the Global Financial Crisis (GFC). Quantile regressions allow modelling of the extreme quantiles of a distribution, as opposed to focussing on the mean, which allows measurement of capital and PDs at the most extreme points of an economic downturn. Understanding extreme risk is essential, as it is during these extreme circumstances when banks are most likely to fail. We find highly significant variances in bank capital adequacy and default probabilities between quantiles, and show how these variances can assist banks and regulators in calculating capital buffers which will sustain banks through volatile times. Quantile regression has been successfully applied to the measurement of extreme market (share price) risk by a number of studies, and this paper develops unique and innovative techniques for extending this approach to structural credit risk modelling.