Experience needed to assess risk, not automated systems

The tragedy of the market is highlighted by my dark-haired colleague (see previous letter). Here is a skilled risk-taker who is unemployed while big lenders blithely persist with the credit scoring approach.

Don’t get me wrong, I believe scoring works with the lowest of risks and this is undoubtedly highlighted by the tightening of criteria by large banks.

But what about perceived higher risk cases? What happens to the ones that lack a full credit history, or the ones with small defaults or minor income stretches?

There are some good cases to be approved but a machine can’t discriminate between individuals – it lumps them all into one pile.

It takes a lot of experience to weed out the bad or fraudulent cases. This requires intelligence, not automation.

If anyone needs a good example of the effectiveness of credit scoring, remember that GMAC Financial Services was once the champion of electronic underwriting.

Then note the news last week that it is leaving the European market and has arrears on some of its securitised books.

Smaller building societies have had to be niche lenders for years because they couldn’t compete on price yet they are declaring some of the best results. I wonder why.