The Manipulation of Basel Risk-Weights. Evidence from 2007-10

In this paper, we analyse a novel panel data set to compare the relevance of alternative measures of capitalisation for bank failure during the 2007-10 crisis, and to search for evidence of manipulated Basel risk-weights.  Compared with the unweighted leverage ratio, we find the risk-weighted asset ratio to be a superior predictor of bank failure when banks operate under the Basel II regime, provided that the risk of a crisis is low.  When the risk of a crisis is high, the unweighted leverage ratio is the more reliable predictor.  However, when banks do not operate under Basel II rules, both ratios perform comparably, independent of the risk of a crisis.  Furthermore, we find a strong decline in the risk-weighted asset ratio leading up to the crisis.  Several empirical findings indicate that this decline is driven by the strategic use of internal risk models under the Basel II advanced approaches.  Evidence of manipulation is stronger in less competitive banking systems, in banks with low initial levels of Tier 1 capital and in banks that adopted Basel II rules early.  We find tangible common equity and Tier 1 ratios to be better predictors of bank distress than broader measures of capital, and identify market-based measures of capitalisation as poor indicators.  We find no relationship between the probability of a bank being selected into a public recapitalisation plan and regulatory measures of capital.