by Jeff Sovern
Norm Silber of Hosfstra has pointed out to me that the New York legislature has passed two consumer protection bills that await Governor Cuomo's signature. One, S03704, would amend New York's existing Plain Language Law to require that consumer contracts involving up to $250,000 be written "in a clear and coherent manner using words with common and every day meanings." Currently the limit is $100,000. I'm not sure how many consumer contracts involve more than $100,000 (perhaps some multi-year residential leases or expensive cars) but I also don't see a justification for having any consumer contract that is not written in plain language. Those entering into consumer contracts should not be put to the choice of not understanding their contracts or having to hire a lawyer to explain it to them. Few consumers read contracts, but maybe those who sign large ones do.
The second bill, S02302, would block consumer reporting agencies from evaluating a consumer's creditworthiness based on the members of a consumer's social network, or reporting creditworthiness information about their social network. CNN has reported that Facebook has taken out a patent to evaluate consumer creditworthiness by looking at the creditworthiness of social network friends. This practice may be dicey under the Fair Credit Reporting Act, which requires that consumer reports be issued only for permissible purposes: disclosure of information about others in your social network would not be for a permissible purpose involving them but rather for you. Maybe the information is saved by the fact that a collective score is provided about many people or by some form of anonymization, but maybe it isn't. I am told that companies are not currently using this information, but credit scores are such a black box that I'm not certain of that, and just because it's not used today doesn't mean it won't be used in the future. It seems unlikely that companies would use information about people in a social network without testing empirically whether it is predictive of defaults, but perhaps that testing is now occurring. What I wonder about is whether use of this information would help the roughly 11% of American adults who are credit invisible. If social network data would increase the number of consumers seen as worthy of credit, and if those consumers actually are worthy of credit, the bill might be a bad idea. On the other hand, if social network data would make it harder for consumers who are likely to repay loans to obtain credit, the bill would seem worth supporting. I guess it comes down to what the data show. In addition, the practice of using social network data to evaluate creditworthiness might have incentive effects: would consumers decline friend requests for fear of damaging their credit?
Interesting post, Jeff.
Re the first point, perhaps plain language is enough, but sometimes I feel that some sort of annotations would be even more helpful. Something like the instructions that come with a 1040EZ that explain what each line is for. For example, I just negotiated a solar panel installation contract. There were lots of promises made via email and in an oral presentation about the expected output of these panels. But the contract also had a merger clause. Even written in plain English, it’s not clear that the average consumer would realize that all of the promises made before signing the contract were being excluded via this one clause. But maybe it’s just that the devil is in the details on what, exactly, “plain language” means.
Re the second point, I suspect that the answer to whether social network data improve credit access depends on whether you are similarly situated to others in your peer group. For example, if you are a unusually financially prudent 19 year old, you are likely to be penalized by social network data because you will be lumped with peers who are probably not as prudent.
And I imagine that Facebook’s patent might relate to their own use of that data in the future and is not necessarily about selling it to third parties.
As for empirical testing, there are certainly companies out there that claim to be using social media data in credit underwriting decisions. And there is some early data, particularly from the Philadelphia Fed evaluating these loans.