Those of us that do not bank with RBS, or any of their subsidiaries, are probably breathing a sigh of relief that the recent computer glitch did not happen at our bank. After all, how many of us could have coped with the ostensible freezing of our bank balances for the time it eventually took to fix?
But, unbeknown to many of us, and beyond the obvious immediate ground-zero of no anticipated money in one’s Nat West Account, there are potential side-effects that could still cause our own domino to be in line for toppling, even when RBS are not our bank – invisible, radiation-style effects brought about by IT systems automated to collect data for credit rating reports, that could diminish one’s apparent financial health in the longer-term?
We are all have credit-ratings, whether we need them or not. They are essential in today’s fast-moving financial markets to enable potential lenders to prejudge the ability of a borrower to fulfil their obligations. Gone are the days of the wise old bank manager running his ruler over your personal finances, then interviewing you personally to see whether you are a good-enough egg to trust with his institution’s money. He has long since been ‘downsized’ and his function ‘outsourced’ to a networked computer programme that can be operated by a service-desk trainee, the ink on their five GCSE’s barely dry, who can take details over the phone, input them (hopefully accurately) into ‘The System’ and get an automated decision in seconds. If ‘Computer Says No’, then that’s the end of it.
‘The System’ is heavily reliant on credit-scoring, and as a result credit reference agencies have mushroomed into our lives, so perniciously that we are now regularly encouraged by pundits to monitor our own credit-rating regularly – as if we don’t have enough to do already. Except for a possible initial free-trial, that service will attract a monthly fee, which we pay to maintain yet another on-line account that will tell us all we need to know in nano-seconds when we need to know it which, in reality for the vast majority of us, is very rarely.
What has this to do with the RBS scenario you may ask; after all, it wasn’t your fault that the bank couldn’t update your balance, was it? Well the sensible answer to that is no, their recent computer glitch should have nothing to do with creditworthiness – except that RBS, and every other bank in the world, are the suppliers of the data that forms the bedrock of those credit reference agency scoring systems. And in just the same manner that a lender’s decision-making software now makes a judgement based purely on cold, hard, algorithmic processes, so a credit score is based upon the judgement of a piece of software code examining bald effects, not the experienced eye of a human-being analysing complex causes.
Most people have their major outgoings scheduled to leave their bank account by direct debit a few days after their regular salary is credited – plain and simple sound financial planning. Take the simple domino-line of an employer banking with Nat West, an employee whose account is in a different Bank altogether, and that employee’s mortgage being with a Building Society unconnected to either bank. A glitch at the employer’s bank delaying the payment by a day or two might be an annoyance, but it should not be an embarrassment. A week’s delay, however, is an entirely different matter, as the outgoing payments cannot be made if there is an insufficient balance to cover them. Nevertheless, a quick phone-call by the employee to the Building Society explaining that the problem was caused by a third-party, and therefore beyond their control, would normally be sufficient for the lender to add an annotation to the account history to avert any major problems. However, credit scoring does not work on phone calls, or understanding annotations from customer service staff. It works on things known as “Triggers”.
Triggers are negative events. They are the reaction of a piece of computer code to an entry in a report that is out of the ordinary. So the software code is not looking for something that happens correctly and regularly, but something that doesn’t happen when expected – essentially records of missed payments. When it spots one, it is scored against the individual’s record, in the entirely impersonal manner that every computer examines the ones and zeros of binary code, the only way it can efficiently process the trillions of lines of data it has to examine hourly. That was, after all, why we invented them – wasn’t it?
So taking our example above, when Nat West fails to update the balance on the employer’s account, the account shows there are insufficient funds to pay the wages scheduled for the same day, and therefore no payment is issued to the employee’s Bank. With no payment into the employee’s account, there are insufficient funds showing, so their Bank’s computer does not issue the scheduled payment to their Building Society, causing the Building Society to register an apparent default on its own computer system.
Of course, the subsequent phone call from the employee explains the circumstances, the customer service agent at the Building Society annotates their account, Nat West sorts-out its problems, the wages are paid, the mortgage payment is subsequently made and the mortgage account is no longer in arrears. So job done, another storm in a teacup blows over, and we can all relax back into normal life again.
Except that “Triggers” happen in real time, they do not await, or require, human intervention, they are just systematic observations of numeric data fields. The non-payment to the Building Society is registered on their systems as a numeric code, which is added to a system report that is sent to the credit reference agency. The subsequent payment, a week or so later is also registered on the Building Society’s system as a different numeric code, and included in the next report, but that is not what the trigger system is looking for; neither is it interested in that fact. As for the annotation made by the customer service agent, giving the reason for the failed payment, the trigger system will never see that; neither would it be able to interpret it even if it did. As far as the credit reference agency’s computer is concerned, the payment was late – which in reality it was.
Not that a single late payment should cause anybody a problem with their credit score. Credit reference agencies are not judgemental, nor are their systems totally inflexible. What they are looking for are patterns of financial delinquency that would call the creditworthiness of an individual into question. To do that, they collect data from a wide range of sources that provide goods or services on a credit-account basis. These include banks and building societies, hire-purchase companies, credit-card companies, utility companies, mobile phone companies, insurance companies – the list these days is quite extensive.
Back to our employee of the company that banks with Nat West, the person with the sound financial planning and all of their major outgoings scheduled over the next few days after payday: their mortgage, the HP payment for their car, the payment clearing their credit card, the regular payments to gas and electricity suppliers, their mobile phone bill, the standing order for the house insurance, and so on. Of course, all of those suppliers were equally sympathetic when they received the phone call explaining that the non-payment was beyond the employee’s control, annotating their accounts accordingly. All of those suppliers also provide credit reference agencies with regular automated system reports containing numeric data codes recording payments and non-payments.
So now our employee’s credit score is beginning to look less dependable – not just one late payment, but multiple failures in reports spread-out over a number of days. Once again, this should not cause financial Armageddon, except that “Triggers” do not just work as historical markers of past indiscretions. They are ever-more frequently used, by those same financial institutions that feed-in the data, as early-warning systems for potential future defaulters.
I published an article in 2007 where I pondered whether the recent introduction, back then, of these software techniques had had any influence on the creation of the sub-prime mortgage crisis in the USA. The principle of the early-warning system is that, when a credit reference agency receives a notification of a late payment on an individual’s file, they send out an alert to all other lenders, and prospective lenders, that are linked to that individual’s file by current credit agreements. These reports are issued daily, and within 24-hours of an ‘event’.
Essentially this was a move from ‘Pull’ to ‘Push’ Technology. For those unfamiliar with the terms, ‘Pull Technology’ is where you search for, and take, information when you need it – a good example of Pull Technology is an internet search engine. With ‘Push Technology’ you are sent the information whether you need it or not – spam e-mails are Push Technology.
Without going into any great detail, I considered in that article how financiers might be handling this new and extensive flow of data, and concluded that they would likely be using ever-more automated systems. The question that raised was could the introduction of such services at the credit-scoring end, and their associated automated filtering applications at the lender’s end, have been part of the scenario that triggered the wave of mortgage defaults at the time? Quite simply, were lenders being spooked, in some cases, by the sudden revelation that their existing borrowers might be experiencing financial difficulties that were not previously apparent?
Back then, “Triggers” were quite new, but they have since become embedded in the credit-rating process. Our fictional employee would have no idea that their various suppliers had automatically reported the payment failures, creating multiple ‘defaults’ registered by a credit-rating system that had, in turn, generated automated alerts to all of their other financial suppliers, driven by these “Triggers”.
There are additional automated bank reports to credit reference agencies alerting when commercial payments are withheld. Nat West’s reportage, assuming that part of their system was still working, would show that the employer had payments stopped through their account having insufficient funds to cover them, potentially affecting their commercial score. But nowhere would it show the reason for that low balance, or the connection to the next domino to fall – the employee.
That is where this entire automated system falls apart. As the credit-scoring is connected to an individual, it is easy for the ‘events’ to be added to their report, as the reportage from all of the data suppliers is linked by common data fields – the individual’s name, date of birth, address, etc. But there is no connectivity whatsoever to that individual’s employer, or to the employer’s bank account. The credit scoring system sees dominos falling, but it has no idea which finger knocked the first one over – effect with no cause – and the fall is recorded for ever. Put graphically, a credit report after these recent events might resemble a photograph of Stonehenge today, whereas beforehand it would have looked like the artist’s impression of when the monument was first built.
Every supplier handles such reportage in its own way. One would hope that, when a mark-one eyeball was applied, attached to a logical mind that could read annotations on accounts, common-sense would prevail. But financial institutions are regularly reducing staff and utilising evermore systems run by outsourced suppliers, so what happens when an alert is acted upon without the application of a sensible eye? Is the credit card stopped? Is the mortgage rate increased at the next review point? The car repossessed? All of those subsequent events would be added, inexorably and automatically, to the credit-scoring system that would, eventually, spiral the rating down to show that our fictional employee with years of sound financial planning behind them was now as credit-worthy as a Eurozone country in search of a bail-out.
Ridiculous, you say. Well maybe so, if the individual starts with a flawless record. But we are in a recession, and many are struggling to keep their heads above water, thus far successfully; in their case, such a scenario might be the straw that breaks the camel’s back. Worse still, what of the individual that has suffered financial problems in the past, but has worked hard to clear their debts and bring their credit-score back to an acceptable level, often paying monthly fees they could not really afford to several credit reference agencies to be able to monitor the results? Because of the way the system works, such an apparent relapse could devastate all of that hard work.
In which case, you may add, all they would need to do is let the credit reference agency, the one that they have paid all those monthly fees to, know the cause of the problem, and leave them to sort it out. I may have agreed with you until I recently assisted a friend to try to do just that, and found that the process was akin to crawling over broken glass. By that, I mean that the credit-rating agency will not change its records, regardless of how glaringly-obvious the error is – they consider themselves the guardians of the data, not the creators. Instead they insist that the lender that issued the automated report is contacted, and persuaded to change their record to reflect the truth. Think about it – how many times have you fought your way through an automated telephone system and actually spoken to a customer service operative who can even understand what your problem is, only to be told that they cannot find any record of your previous phone call, or of the promised annotation by their predecessor on your account?
Then I heard James Jones, Head of Consumer Affairs at Experian, one of the leading credit reference agencies, interviewed on BBC Five Live on 26th June, about a week after the RBS problem emerged. He was asked: “looking at the information you get when a payment is missed, could you tell whether it was caused by the problems at RBS?” The reply was: “not really, no. We get data in from so many organisations, every minute of every day; we don’t normally manually-review the data, so we wouldn’t normally have visibility of that.” He went on to say that: “these days we cherish* our credit ratings, they are more closely-scrutinised than ever, so it is understandable that people have concerns.”
The interviewer pressed him further: “how are the mechanics of this going to work, especially when Nat West is not your bank?” The reply was stark: “it is going to be a challenge, absolutely. So, I think, equally important, alongside the work that the industry is doing to try and tackle these issues, is what we can do as consumers. So we are encouraging people to get in touch with all of their creditors, all of the organisations they pay on a regular basis, to alert them to any issues they are having, but also at the same time to monitor their own credit record so that if adverse data is registered, they will be aware of it straight away, then they can deal with it.”
The interviewer rightly took him to task on this, pointing out that the responsibility was being put on the individual when there was no fault on their part. He then asked: “what about something more radical here? Why don’t you just scrap all of the negative credit information for that week, because you can’t know whether or not this was down to the problems at Nat West?” Mr Jones’ reply was disappointingly banal, made more so by the hint of a nervous chuckle it commenced with: “that would be a massive oversimplification of how things work, because, for a start, the data we get in is normally going to reflect the month before, so it’s not data this week, it’s not data next week, and organisations send us data at different times of the month, so that would be an incredible challenge.”
The interviewer pushed him harder on knock-on effects, exploring a similar scenario to the fictional employee in this article, highlighting that no single organisation in that chain would be able to tell whether the situation was down to the RBS system glitch, finishing “… it would seem unfair if that would harm someone’s credit rating and you don’t seem to be saying that won’t happen.” The reply was chilling: “we can’t give that cast-iron guarantee, but we will do everything we can working closely with the banks concerned to try and minimise that error, that incidence, but at the same time we have a responsibility as consumers and we have a role to play here by reviewing the data and working with our creditors to try and tackle any issues that do arise.”
Roughly translated: We have created a monster that we no longer know how to control. As a consequence, you – the hapless employee whose wages did not arrive through no fault either of your own, or your employer – are on your own here; it’s all down to you now. But you can, of course, subscribe to one of our services to not only see what problems you have, but also to get advance notice of what pain this will cause you for evermore.
And remember, this fictional employee did not even bank with RBS.
*OED definition of ‘cherish’ – foster, nurse, keep warm, hold in one’s heart, cling to
Brilliant. Just proves what a load of rubbish Credit Ratings are – and it’ll be just another bullet in the clip for banks to not lend money.