RDP 2012-05: Payment System Design and Participant Operational Disruptions 2. Literature Review

In recent years there has been a sharp increase in payments settled in hybrid RTGS systems. In 1999, a sample of 22 countries found that 3 per cent of the total value settled in large-value payments systems was settled in RTGS systems that incorporated hybrid features; by 2005 this share had grown to roughly 32 per cent (Bech et al 2008).^[2] At the same time, hybrid systems have received increased attention in the payments literature, for example: McAndrews and Trundle (2001) and CPSS (2005) provide detailed expositions of hybrid systems; Johnson, McAndrews and Soramäki (2005) and Ercevik and Jackson (2009) use simulation analysis to quantify the impact of introducing hybrid features on liquidity demand and settlement delays; while Martin and McAndrews (2008) and Galbiati and Soramäki (2010) use theoretical models to analyse the impact of hybrid features on participants' incentives.

A separate stream of the payments literature has focused on analysing operational risk in RTGS systems through simulation studies. This literature generally follows the methodology established by Bedford, Millard and Yang (2005) to analyse the systemic effects of simulated operational disruptions that prevent a participant (or multiple participants) from submitting payments. Bedford et al simulate operational disruptions in the UK RTGS system, Clearing House Automated Payment System (CHAPS), while Schmitz and Puhr (2007), Andersen and Madsen (2009), Glaser and Haene (2009) and Lublóy and Tanai (2009) perform similar analyses on Austrian, Danish, Swiss and Hungarian RTGS systems, respectively. Results from the simulation analysis vary across studies, with differences largely explained by the level of liquidity in the system under consideration and the size of the participant experiencing the disruption, as well as assumptions about non-stricken participants' reaction to the disruption.

Ledrut (2007) investigates the mitigating effect of participants' reactions by simulating a participant disruption in the Dutch RTGS system. Participants are assumed to react by stopping payments to the stricken participant after a pre-determined time has elapsed or once their exposure to the stricken participant has reached a certain threshold. Ledrut concludes that more timely participant reactions can significantly reduce the systemic consequences of participant-level operational disruptions. Merrouche and Schanz (2009) also investigate counterparties' reactions to a participant's operational disruption. Based on an econometric model of CHAPS, they find that payment flows to stricken participants tend to decrease until around one hour into the disruption, but increase slightly afterwards, presumably as the cost of violating contract obligations or market practices by delaying payment increases.

Footnote

This is based on a study covering 13 countries that were members of the Bank of International Settlements' Committee on Payment and Settlement Systems (CPSS) as at 2005 and 9 non-CPSS euro area countries. [2]