[scarmig]

Any metric is gameable. And, if the current algorithm is the best way to predict "credit-worthiness," publicizing its details will drive consumers toward different behavior, which makes it less powerful

Imagine Google made its own credit score, based on your searches, social contacts, etc. It turns out that someone clicking on ads for product X have a very high likelihood of repaying their debt. This information ends up publicly leaked.

What do people start doing? How does this affect the quality of Google's credit algorithm?

[yebyen]

Not any metric is gameable.

A reliable indicator of whether you will pay your bills in the future might be whether you paid your bills in the past. There is no way to game this metric, other than by paying your bills.

You just chose a metric that would be easily gamed and said that any metric is gameable.

[chimeracoder]

> There is no way to game this metric, other than by paying your bills.

Except the exact situation outlined above, in which a person generated bills that they otherwise wouldn't have needed, for the sole sake of having more bills "paid on time".

[yebyen]

So, how does one get over in gaming this metric? Take out a lot of credit accounts once your credit is good and then default on them? What's the end game where they come out ahead? See prior comment about the bank job... that's just called having and ruining your credit.

[SilasX]

That's overstating it. "Whether you've paid back all debts on time" isn't gameable.

[jacquesm]

Actually it is. Simply take out a couple of loans you don't need and pay them off at the agreed upon schedule.

[ProblemFactory]

Taking out loans that you don't need, not spending the money, and paying them back on time (with interest) should be a very positive credit score indicator.

It shows that the person is planning ahead, is careful with not overspending even if they have cash available, is organized enough to make all the monthly payments on time. To me that seems like a better signal than someone who does need the loan.

How would a person who takes out the same loan, but actually spends the money be a better credit risk?

[SilasX]

I agree, but in fairness, the way credit agencies use that information is nonetheless gameable. They ignore all of your payment diligence when it comes to bills (utilities, rent, cable, etc) but consider it a positive indicator when you get a credit card that you pay off every month.

This leads people to do something they don't want to do, and is no more informative than paying all those other bills would have been, merely to increase their score. That's a kind of gaming.

[yebyen]

I think you do have to consider too, that these companies also want to give you credit (it is their business,) and in many cases they will likely have insurance that pays when you default.

If the way they have allowed you to "game" this metric results in you signing up for some credit in order to improve your credit score, that's nothing but a net win for them. They got you some credit, that's definitely a score for them.

(I'm not saying that credit card companies would prefer all of their customers to default, they would never be able to buy insurance again... I just want to put it out there that maybe the system works exactly this way, 100% deliberately, because it benefits them too.)

[SilasX]

That doesn't change the value of "has paid back all debts owed?".

[Tyr42]

But it does affect the "has paid off x% of loans"

[SilasX]

Which wasn't the couterexample.