Detection of Bank Failures in Transition Economies: The Case of the Czech Republic

This paper studies bank-failure models in the context of transition economies. In order to capture the default risk of banks, data on the structure of retail deposit rates is used to improve the prognostic quality of bank-failure prediction. The Czech bank crisis of 1994?1996, during which 14 banks failed, is used to verify the suggested approach. It is shown that banking supervision did not have ? most likely given the low quality of the available accounting data ? better information with respect to foretelling bank failures than the general public did via retail interest rates. In addition, the combination of balance-sheet and interest-rate data significantly improves the quality of bank-failure prediction. Thus, the utilization of information related to interest rates can increase the efficiency of banking supervision and can provide early warning signals of bank failures.