A simple model for pseudo-nonstationarity in operational risk loss data due to interest rate dependency and reporting threshold

Gerrit Arlt, Frank Neumann and Udo Milkau

Abstract

ABSTRACT

Nonstationarity in operational risk loss data time series is a known effect, but has so far rarely been analyzed in detail. Taking transaction banking as a segregated object of study, a simple model presented in this paper, based on interest rate dependency and reporting threshold, can provide an explication for this pseudononstationarity in the last decade. Especially from the perspective of a business line, the separation of this effect can improve the monitoring of measures to reduce operational risk in an active management approach.