Operational Risk and Contagion in the Austrian Large-Value Payment System ARTIS

The objective of this paper is to quantify the contagion effect of an operational incident occurring at one ARTIS participant’s site on the payment activity of the other ARTIS participants. We used model simulations to focus on operational problems occurring at one of the participants, not an operational failure of the ARTIS platform itself. The scenarios are designed according to an ex-ante estimation of potential risk concentrations based on actual data for the sample period (Schmitz et al., 2006). The main conclusion from the simulations was that the contagion effect in ARTIS is low on condition that the existing business continuity arrangements prove effective. However, this is a very restrictive assumption. Without the use of business continuity arrangements or if they turn out to be not fully effective, the contagion effect on the smooth functioning of the payment system was substantial in all three scenarios. In contrast to the most common approach described in the literature, we used actual (instead of simulated) liquidity data to study the contagion effect at the individual bank level as well as at the aggregate level of unsettled payments. A non-negligible number of banks failed to settle payments in all three scenarios. The paper also provides results on two features of large-value payment systems that have hitherto gone unstudied in the literature: the stop-sending rule and debit authorization.