Measuring Bank Risk
Most existing empirical studies investigating the relationship between bank competition
and financial stability at the microeconomic level focus either on credit risk alone, using
some form of credit risk measure such as non-performing loans, or resort to bank risk measures
constructed from balance sheet information, such as a z-score. Bank risk measures
constructed from balance sheet information, however, have the disadvantage that they
do not provide information on actual distress events, or even outright failures of banks.
Our measure of bank risk is comprised from the distress database collected by the Bundesbank.
This dataset contains information on bank-level distress events that range from
weaker incidences such as capital support measures by the deposit insurance schemes to
outright bank defaults (i.e., bank moratoria or takeovers in the banking market which are
classified by supervisors as “distressed mergers”).5 Hence, our measurement of bank risk
directly captures the possibility of banks exiting the market because of distress events.
According to Aspachs, Goodhart, Tsomocos, and Zicchino (2007), the probability of bank
distress events is a much more appealing bank risk statistic because, by covering all types
of risk, it provides a more exhaustive picture of risk borne by the banking system. We
consider two different concepts for defining bank distress: first we construct an indicator for broader bank distress events (Bank Distress),6 as well as one more narrowly defined
indicator for banks exiting the market in a distressed merger or in a moratorium (Bank
Default). In order to give the analyses a forward-looking perspective, we forward bank
distress and bank default events by one year with respect to the explanatory variables.