Monday, March 19, 2012

Credit checks and employment

An article from the Fort Collins Coloradoan caught my eye, in part because it deals with a similar issue to the "avoid ghetto" GPS app: whether giving more information can make people worse off.

The article discusses efforts by Colorado lawmakers to prohibit businesses from running credit checks on prospective employees and using the results in hiring decisions. Obviously, the reason employers run these credit checks in the first place is because they think the people with better credit tend to be more responsible and thus better workers, as well as being less likely to steal from the till. The argument for banning such checks is that they can create an "unemployable class" of people because people can't improve their credit without income and can't get a job unless they have good credit, and also that it's possible to have bad credit even if you are financially responsible (say because of a layoff).

The readers' comments are predictable. Most of them say that it's just wrong for employers to "discriminate" based on credit checks, while there is a minority view that the government should not have a right to tell businesses how they can make hiring decisions. (I would like to point out the comment by Jeff Emmel, in the first category, which blames the problem on "too many people chasing too few jobs." If true, this seems to negate the idea that credit checks are the problem. If there are N people looking for a job and there are M openings, then at least N-M people will fail to get a job, regardless of what methods are used to screen applicants.)

Here is my analysis. A traditional economist would likely say something like the following:

If credit checks really provided no useful information about the quality of an employee, businesses would have no reason to use them. Any business that did use them anyway for whatever reason would face a competitive disadvantage in the marketplace because they would be screening out qualified employees, and a business that didn't use credit checks could gain an advantage by grabbing all the good employees that were rejected by others. Thus, if the practice really has no benefits, there is no need for a law against it; it will die out naturally. If it doesn't die out, that is evidence that it is in fact providing useful information.

Of course, exactly the same argument was made with race, gender, etc. discrimination laws, and look how that turned out. But this case is different. The difference with racial discrimination was that customers had a preference against minority employees so businesses reacted to that, but we needed to get minority employees into restaurants and stores so that people would learn to overcome their prejudice. In this case, customers neither know nor care about the credit histories of the employees with whom they interact. Also, credit checks are a relatively new thing in employment, so unlike the issue with race discrimination, we're not starting from a status quo where everyone discriminates so nobody has a chance to realize what they're missing out on.

One could argue that it really is true that people with poor credit tend to not be as good workers, but that there's a public benefit to giving those people jobs that outweighs any harm done to employers. This position is not unreasonable; after all, someone who is consistently poor does cost society money to maintain because they are more likely to need welfare and other social services. But it is important to realize that this argument only applies if there is a greater public benefit to giving a poor-credit person a job than giving whoever the replacement would be a job. And most importantly, if there really is a public benefit to giving particular people or classes of people jobs, a better way to do it is with subsidies and incentives targeted at that group and let businesses decide whether the incentive is worth the cost, rather than through the indirect way of banning credit checks.

However, I can also think of several counter-arguments to the above.

1. It may be true that the practice will die out eventually if it is not providing useful information, but that could take a significant amount of time. In the meantime, many people will still be affected.

2. Let's say that half the people who have poor credit really are financially irresponsible, but the other half have poor credit for reasons completely out of their control. Then it makes sense from the businesses' perspective to avoid the risk, but the innocent half is still caught in the crossfire. (However, the arguments in the last paragraph above still apply.)

3. Suppose that it is possible to "manipulate" your credit rating*. Then everyone will manipulate their rating in order to get a better chance, but of course all this manipulation is a zero-sum signaling game, so it's a waste of effort. (If everyone manipulates their rating up by X points, then everyone is still in the same order so it gives exactly the same information as before, but nobody ends up better off.)

4. As for the comparison to racial/gender discrimination laws, there's another consideration which cuts the other direction. One problem originally identified with racial/gender discrimination laws was that it's essentially impossible for a business to prevent itself from obtaining information about an applicant's or employee's race or gender. Thus, for instance, a business might want to fire an unproductive worker but might have a difficult/expensive time proving in court that the firing was not because of race/gender, so it might decide not to. In contrast, it is very easy for a business to avoid obtaining information about an employee's credit history, so this is a non-issue.

5. It may be possible that there is a public benefit to giving jobs to certain classes of people - which tend to overlap with the people who have lower credit scores - but the first-best solution of "subsidies and incentives targeted at that group" is impractical due to other constraints such as political feasibility. So banning credit checks could be a good "second-best" solution.


Perhaps surprisingly (but maybe not), very few of the comments in the article itself or the reader's comments directly address any of the questions above. I wonder why?

*This "manipulation" need not be anything shady or illegitimate. The only thing relevant for this analysis is that the "manipulation" has no relevant effect other than improving your credit score. Given all the advice about "how to improve your credit score" that you can find all over the place, it's very likely that this type of "manipulation" is possible.

No comments: