We learned on Thursday 18th October 2018, at the launch event of Pay.uk, that any renewal of the UK's retail payments infrastructure will have to wait until after the renewal of the Bank of England's payments infrastructure. This is by far the first instance of the Bank of England standing in the way of innovation, new competitors, new services and so on.

It has been a consistent official policy, since the financial crisis, to open up the UK’s payment systems, not least as a way of reducing the concentration of market share of consumer and small business accounts in the five big UK banks. This policy crystallised no later than 2012, and resulted in the creation of the Payment Systems Regulator in 2014.

The Bank of England (“BoE”) has been a cheerleader for this policy, despite being the main roadblock towards its implementation.

The number of Payment Service Providers (“PSPs”) operating in the UK has expanded dramatically in response to the perceived opportunity:

• 440 banks (or “credit institutions”)
• 1,150 non-banks: 400 eMoney Institutions and Authorised Payment Institutions, and 750 Small Payment Institutions
• 1,590 PSPs in all

A PSP has to have a Settlement Account at the BoE in order to be a direct member of any UK payment system, and up until this year no non-bank PSP was permitted by the BoE to have one. Now two non-bank PSPs have one: ipagoo (an eMoney Institution) and Transferwise (a Payment Institution). This policy change has so far benefitted 2 out of a universe of 1,150 non-bank PSPs. That is 0.17%.

A number of new banks – often referred to as “Challenger Banks” - have joined one or more of the UK’s payment systems. Names such as Monzo, Starling and BFC have been well publicised. However, even this programme has so far only benefitted 14 banks that were not already a direct member of one of the systems. That is just on 3%.

Overall the widening of access has benefitted 16 out of 1,150 institutions: 1.4%.

The main reason for this is the IT queuing system at the BoE, resulting in very few slots being available for new Settlement Accounts:

• 13 slots were available in 2018, and all are taken;
• 19 slots are available in 2019 and most are taken.

For the BoE, opening a Settlement Account is an IT change that needs to go through a full testing cycle. This policy was introduced after the CHAPS outage in 2014 to ensure that BoE’s systems did not fall over again, these systems being of “systemic importance” to the UK’s financial and economic system. One IT change can be done per week. As there are IT freezes over the summer and over Christmas/New Year, and as there are 13 meetings of the Monetary Policy Committee that might result in a Base Rate change (also an IT change), this leaves only a small number of slots available for new accounts.

This indicates a comically antiquated IT infrastructure at the BoE. The drag that this would create on the widening of access was apparently not realised until after the CHAPS outage. Now CHAPS has been taken inhouse by the BoE from the CHAPS Clearing Company Ltd and there is a belated “RTGS Renewal Programme”, which may come onstream some time after 2021.

The roll-out timing of the replacement RTGS will also dictate the timing of any renewal of retail payment systems being undertaken by the newly re-branded Pay.uk: migration to their New Payments Architecture can only commence after the RTGS renewal is complete and in stable production – let’s say in 2023.

The CHAPS outage in 2014 was caused by the failure of a look-up routine regarding a table. A new participant was introduced, namely Danske Bank, separately from Northern Bank that it had acquired, although now only Danske Bank is a CHAPS participant.

Danske’s name was added to a table, at which point the bank at the bottom – UBS AG (London branch) – fell off and all traffic to and from UBS failed and then the whole system failed.

The way the outage was managed raised other important points that have been swept under the carpet, turning on the revelation that CHAPS contained two processes:

• A genuine RTGS process for systemically-important payments;
• A Net Settlement process for non-systemically-important payments;

All of the former were processed during the day of outage, but the latter were not, and were sometimes for mortgage completions: consumers spent the night in their cars waiting for their purchases to complete, after they had vacated the properties they were selling.

Consumers, through their solicitors, had paid £30-50 for a CHAPS payment and it was sold to them as an RTGS, first-class-post payment. Then the BoE processed it as a net settlement, second-class-post payment, because they did not deem it important.

The splitting of the market’s CHAPS volumes into two flows – full RTGS mode and Net Settlement – smacks of another limitation imposed by archaic technology: that the system was simply not capable of processing the number of CHAPS payments submitted by the market in full RTGS mode.

It also exemplifies a further, questionable BoE policy – to drive non-systemically important payments off CHAPS, and preferably onto Faster Payments.

Since 2014 the BoE’s policy has been well served by the raising of the Faster Payments System Limit to £250,000, to which it naturally gave its required “non-objection”. Now we see the result: the rocketing of both Faster Payments volumes and Authorised Push Payment Fraud on the back of Faster Payments. The Payment Systems Regulator, in a statement on 28th September 2018, made clear the correlation of this type of fraud with the Faster Payments system, and the BoE policy has fed right into that.

The bottom line is that the BoE has been running an antiquated IT complex for many years, and that the pinchpoint that this would impose on mainstream official policies was not realised (and may not even now have been fully realised) until the CHAPS outage of 2014. The situation will not be resolved until 2021 at the earliest. The industry response to the problem – increasing the Faster Payments system limit - has exacerbated Authorised Push Payment Fraud.

The BoE is the most critical single-point-of-failure in the UK financial system but it cannot do the basics, such as process all of its payments in real-time or open an account without involving the IT department.

[First published in Computer Business Review and sourced from Project Carlton research]