MPs round on banks following IT failures

MPs round on banks following IT failures

Following a string of high profile technical blackouts, Britain's biggest banks have been urged by the UK's influential Treasury Select Committee to invest more resources to improve their technology infrastructure and to appoint board directors with knowledge of IT.

In recent months, Royal Bank of Scotland, Barclays and HSBC have all experienced IT outages, angering customers who have been locked out of their accounts and unable to perform routine transactions or make and receive payments.

The Treasury Select Committee chairman Andrew Tyrie has written a public letter to the banks and regulators, stating: "We can’t carry on like this.

“Bank IT systems don’t appear to be up to the job. This brings with it not just conduct risk, but also systemic risk. Someone - probably the head of the PRA, needs to assume a leadership role, bring together those most involved among regulators and government agencies, and ensure that there are improvements at the banks.”

Alongside the Tyrie broadshot, the Select Committee also published correspondence with some of the UK's top bankers, including Barclays chairman John McFarlane, HSBC’s UK and Europe CEO Antonio Simoes and RBS chief executive officer Ross McEwan.

McFarlane says that Barclays is splitting some of IT systems into smaller units to prevent contagion from one problem spreading to other parts of the bank, while RBS' McEwan, says the bank is conducting a root and branch review of its IT management in an effort to shore up its crumbling IT systems.

Comments: (5)

A Finextra member
A Finextra member 25 January, 2016, 14:59Be the first to give this comment the thumbs up 0 likes

FFS the last thing that Banks need is more IT people on their Boards.  It is the IT people who have created the mess, and then made it worse by chucking money at large IT companies.  If the Treasury Select Committee want more or the same and worse, then their strategy is bang on!

A Finextra member
A Finextra member 25 January, 2016, 15:19Be the first to give this comment the thumbs up 0 likes

It's not the IT people who have created the mess, it's the "chronic lcak of investment" and the wholesale offshoring of resource, coupled with the redundancy of all the knowledge and experience Banks have, in favour of cost reductions.

Banks DO need more IT people on their boards to help them understand the error in their thinking and planning. 

Cost cutting has lead to this and all Banks are suffering.

Keith Appleyard
Keith Appleyard - available for hire - Bromley 25 January, 2016, 18:45Be the first to give this comment the thumbs up 0 likes

Certainly when I've been impacted by numerous outages at RBS, and the protracted time it takes to get a resolution, and the instances when I've had to put up with 'garbage' narrative on my Bank Statement because the original text is irretrievably lost (eg Payer name), I'd have to say its a combination of chronic under-investment by the business, coupled with piss-poor software quality.

A Finextra member
A Finextra member 23 February, 2016, 22:12Be the first to give this comment the thumbs up 0 likes

"It's not the IT people who have created the mess"  ... the Finextra member's statement to a certain extent is quite true. In most large FI's ,legacy mainframe systems do all the back-end heavy lifting, this is especially true with post-memo batch applications. In RBS's case they got caught trying to restore a previous version of CA7 during a critical over-night batch window (apparently they ran some 90,000 batch jobs per night to meet their Online window availability in the morning). My understanding is that the DCB attributes of the new release (log file) did not match the DCB attributes of the old release (log file). So having said that normally this would of been quite routine (if they matched), however, this back out procedure caused the corruption of the CA7 log file, necessating a "cold start". The CA7 log file is absolutely critical if your trying to rebuild all the application Q's (input Q, output Q, dispatch Q, waiting Q, completion Q, print Q,Jes Q etc etc) with their recovery utility routine. Now your faced with a manual intervention re-build and you have absolutely now clue when to begin without critical positional input.... remember these guys were processing feeds and updates from some 90K jobs per night..... It all comes down to "getting behind big-time " in your over-night batch .... once your Online comes up ... your toast ... now you accumulating more transactions’ and getting behind by the hour .... now your into the next night processing ect ect ... its in my opinion , not the fault of "RBS IT" staff completely as we all know that CA7 settled out of court for an un-disclosed sum .... they assumed that a "back out" restore procedure should of worked and would of/should ,of been backward compatible ....., back to the management point " legacy systems are life blood" if your outsourcing key legacy application staff and support people to the outsourcer you need to look at the consequence of "knowledge drain" verses short term cost reduction .... I have seen this happen and I have seen the reversal when management come to the realization that they have nobody who really knows their application flows or recovery scenarios ... "The answer" ... you need to encapsulate as much of this knowledge as possible, the CA7 was a prime example .... my approach for this legacy process was too automate the recovery using off-host systems that tracked scheduler positional Q's .... if you incurred a catastrophic error and had to re-build without the log file (RBS case) then at most you were into a few hours at best for recovery and automated restart .... "old legacy knowledge is still important" .... don't waste it         



A Finextra member
A Finextra member 23 February, 2016, 22:50Be the first to give this comment the thumbs up 0 likes

By the way in case your wondering just how good this processing automation is .... it manages the batch critical path completion times based on real-time and historical data without manual scheduler intervention ... changes priorities and suppresses non-production workload to maximize resources when required ...