This paper investigates a practical issue in the use of econometric models in bank
surveillance: the number of models that are relevant for surveillance. That is, does a
model estimated to predict which banks are most likely to fail in future periods also
provide accurate predictions of which banks are most likely to be downgraded to problem
bank status? Or do supervisors need a different model to predict which banks they are
most likely to rate as problem banks in future periods? Ability to predict downgrades of
supervisory ratings, rather than failures, is especially relevant for surveillance during a
period in which there are few failures, like most of the 1990s. We compare the
performance of two econometric models in predicting which banks will have their
supervisory ratings downgraded in future periods: the SEER risk rank model, which was
estimated to predict bank failures, and another model that is estimated to predict
downgrades of supervisory ratings
This paper investigates a practical issue in the use of econometric models in banksurveillance: the number of models that are relevant for surveillance. That is, does amodel estimated to predict which banks are most likely to fail in future periods alsoprovide accurate predictions of which banks are most likely to be downgraded to problembank status? Or do supervisors need a different model to predict which banks they aremost likely to rate as problem banks in future periods? Ability to predict downgrades ofsupervisory ratings, rather than failures, is especially relevant for surveillance during aperiod in which there are few failures, like most of the 1990s. We compare theperformance of two econometric models in predicting which banks will have theirsupervisory ratings downgraded in future periods: the SEER risk rank model, which wasestimated to predict bank failures, and another model that is estimated to predictdowngrades of supervisory ratings
การแปล กรุณารอสักครู่..