Systemic risk in payments clearing and settlement?

Is the recent RBS system failure a one-off, or does it highlight a bigger and broader problem? 

Robert Peston, the BBC’s Business editor, says  ‘….RBS does not believe that the issue raised by many - that all the banks have creaking ancient and complex computer systems, on to which they are constantly grafting updates - is the core weakness. That said, it is striking that in the boom years before the crash of 2007-8, none of the big banks replaced their so-called legacy systems with new modern ones - which some would see as another example of how they did too little sensible long-term investment when the sun was shining, and are now struggling to mend their roofs in less clement weather.’

Back-office systems are not islands.  National clearing and settlement doesn’t exist in isolation from the banks; the banks’ own messaging and accounting systems are an inherent component of it. Since every payment has two participants, the failure of one bank’s system may have a considerable impact on others who bank elsewhere. For example, the RBS systems are likely to be involved in 50% of UK payment transactions, on either the sender or the beneficiary side.  The economy as a whole – not just the reputation of a single bank – may be at risk. Adding more banks as direct participants to the clearings may therefore increase the risk of failure – a matter which should be considered during the ongoing HM Treasury review of the governance of the UK payments industry.

Upgrading legacy systems is no easy matter. Their complexity makes them hard to maintain and to test.  The people who knew their functional, technical and operational idiosyncrasies are fast retiring from the workforce.  Rationalising them or upgrading is a costly, risky programme akin to open heart surgery while the patient pedals in the peloton. And introducing new requirements – say, the Faster Payments Service or new AML functions – presents another level of risk.

But complex problems can often be tackled by looking at them in a different way. Component business modelling might identify functions which can be built and run once centrally, by a national utility, instead of multiple times by each bank. Examples might include anti fraud and sanctions processing, though there are regulatory challenges. But regulators are responsible for the integrity of the financial system, so perhaps they might be sympathetic.

Some years ago, a broadly comparable systemic risk was analysed by a team led Peter Allsopp, then Head of the Payment, Settlement and Clearing Systems Division at the Bank of England. The Allsopp Report analysed the potential impact of failure of several banks as a result of the unexpected foreclosure of one. The conclusion drawn was that the potential impact of a relatively infrequent event, coupled with the small probability of a knock-on impact, nevertheless resulted in an unacceptable risk to the global economy. The banking industry’s response was to establish CLS Bank. In June 2012, CLS settled an average daily value of US$5.12 trillion – which isn’t newsworthy as it has been doing so quietly and successfully over a number of years. It is a prime example of how the industry, by collaborating, can deliver value – in this case by dissipating risk in the financial system. The accounting and payment systems of the clearing banks, connected as they are through clearing and settlement, represent a similar unpredictable and infrequent risk to the workings of the national economy.

It is an over-simplification to say that payments is a non-core business and can be outsourced; nonetheless, perhaps more parts of it can be, without compromising bank revenues. Indeed, by introducing new products and services, it may be possible to help banks grow their market shares and revenues. A transitional approach, in which a separate environment is created to support new business, may be one option. At least one innovator is assembling a ‘common banking infrastructure’ designed to minimise market entry and ongoing costs for ‘challenger’ banks, credit unions and savings and loans; which might eventually provide another option. Perhaps the bigger banks might do well to invest in it, or something similar; outsourcing is only feasible if there is a proven ‘target’ to which to migrate.

A more radical approach to risk management might be to treat the entire national clearing and settlement as a national asset; by bringing all the systems (those already centralised, and those currently within the banks) under a single ownership and governance.

Another article on the BBC’s business page argues that a more holistic approach is needed, ‘without which, the RBS scenario will be the first of many and it won't just be the banking sector that suffers.’ Renovating infrastructural systems is burdened with a difficult business case. Benefits can usually only be fully realised when the heritage systems are retired. Carrying two sets of systems and processes in the meantime – as we are seeing with SEPA – adds to the cost line. But the damage to reputation and trust resulting from a major outage may end up costing more.  The £125million set aside for customer compensation by RBS (in addition to the cost of sorting the problem) is an indicator of the serious business cost of a failure event.