Automated credit decisions – negotiating the risk tightrope

Written by Craig Evans, Head of Business Development at Graydon UK
The risks of human error in credit decisions might be too great to take – but eradicating risk completely can lead to the commercial doldrums.
In December, a Bank of England report revealed that two of the seven largest UK banks do not currently have sufficient capital adequacy to survive a new financial shock akin to the crisis of 2007/08.

So, welcome to the season of jollity and good will.

But, as BBC Business Editor Kamal Ahmed was one of many to point out, this was far from the entire picture. Lending risks for the banks are intensifying daily, particularly in the buy-to-let and commercial property markets. As he also said, however, risks have to be taken.

“Managing risk is a balancing act,” he wrote. “Be too risk-averse and there is the danger of creating the ‘stability of the graveyard’.”

Falling risk exposure?
You don’t have to be in banking to recognise the truth of that statement. In fact, it might be even more meaningful if you’re not. For maybe, even despite buy-to-let and commercial-property concerns, today’s banks are actually becoming less exposed to risk on a daily basis than almost any other business.

After all, Lloyds and RBS, both deemed “too big to fail”, received a combined boost of £65bn in public money following the last crisis. How many other sectors received such support?

Add to that the overwhelming global rush to introduce stricter legislation and compliance procedures, and banking is potentially emerging as one of the world’s least risk-laden industries.

If this turns out to be true, it’s pretty ironic – from an industry that at one time dealt with living, breathing risk at every stage of the value chain, it’s rapidly becoming one where certainty and security are systemically ingrained.

The challenge for banks, in fact, might increasingly become one of identifying the high-return opportunities that they are actually able to consider.

By way of contrast, the day-to-day reality for the rest of us is that every business, from the smallest SME to the largest multinational, is taking a very real risk every time it makes a credit decision about a customer, a supplier or a partner.

Risk and reward
I don’t agree with the sentiment that there is “no gain without pain”, but I do recognise that in business there must be a very real relationship between risk and reward. However, while in the past it was often those adventurous organisations that were most willing to take a risk that emerged ahead of the pack, tomorrow’s winners will increasingly be those that best manage the risks they face.

Right now, good risk management increasingly means keeping up with the fast-changing and ever-improving automated systems that exist to support credit decision-making. In my view, in our data-driven world, there are almost no circumstances in which human input has the leading role to play in reaching the right credit decision.

Red alert

Don’t get me wrong. I prefer people to computers – they’re (sometimes at least) creative, intuitive and fun. But while these qualities are definite assets in a social or strategic setting, they can also be extremely dangerous if they’re used as the sole basis for making credit decisions.

When I say extremely dangerous, I mean they can be quite literally worse than useless. Not only might they not help to reach the right decision – they might also positively help in reaching the wrong one. And I haven’t even mentioned the dangers of human error, an ever-present black hole into which data can all-too-easily fall, never to emerge.

So, when a cost-effective credit analysis solution exists that can crunch and interpret data in vast quantities, to extreme levels of accuracy, at enormous speed and in a way that is fully traceable and auditable, what conceivable reason might there be for not using it to replace all human input? Add to all that its potential for reducing compliance costs and for use as a stress-testing tool, capable of enhancing and refining credit strategies, and surely you’re on to a winner.

After all, the argument might go, even if you’re not using such a system, the stark reality is that pretty soon your competitors will be (and that’s assuming they’re not already doing so).

It’s hard to argue against that view, and I am probably about 95% in favour of it. But not 100%.

For while I don’t believe that in the big-data age human beings can any longer be solely responsible for making credit decisions, nor do I think that they should be entirely divorced from the process. It all comes back to those three qualities of being human – creativity, intuition and fun.

In many such decisions, these are the qualities that enable people to see potential that machines are blind to: to look beyond the bald data and appreciate something of the personality and ambition behind the numbers.

And perhaps that’s how we can continue to take the carefully considered risks that prevent our businesses from gaining Kamel Ahmed’s “stability of the graveyard”.