Credit Cards: On the Optimal Design of Disclosure Regulation

Designing effective disclosure regulation is difficult.  I mentioned some of these difficulties in my previous post.  Here I want to focus on the type of information that should be disclosed, and I want to draw a distinction between two categories of information: (1) product attribute information, e.g., the late fee is $35; and (2) use-pattern information, e.g., there is a 30% chance that you will pay late and trigger a late fee.  Both types of information are necessary for the efficient operation of markets.  Specifically, use pattern information is crucial.  A consumer who knows that the overlimit fee charged by issuer A is much higher than the overlimit fee charged by issuer B might still get a card from issuer A if he optimistically thinks that he will never exceed his credit limit.

However, disclosure regulation in general and TILA in particular have largely focused on product attribute information.  This focus may be explained by the belief that consumers have better information about how they will use the product.  But it is not at all clear that consumers have better information than issuers about use patterns. Duncan McDonald, former general counsel of Citigroup’s Europe and North America card businesses, observed that: "No other industry in the world knows consumers and their transaction behavior better than the bank card industry. It has turned the analysis of consumers into a science rivaling the studies of DNA …. The mathematics of virtually everything consumers do is stored, updated, categorized, churned, scored, tested, valued, and compared from every possible angle in hundreds of the most powerful computers and by among the most creative minds anywhere. In the past 10 years alone, the transactions of 200 million Americans have been reviewed in trillions of different ways to minimize bank card risks." (Duncan A. MacDonald, “Viewpoint: Card Industry Questions Congress Needs to Ask,” American Banker, Mar. 23, 2007, at 10)

If this is the case, why not require disclosure of use-pattern information in addition to product attribute information?

Congress has taken a few small steps in this direction.  The TILA amendments in the Bankruptcy Abuse Prevention and Consumer Protection Act require disclosure of some use-pattern information on the time it would take to pay-off a specified balance if the consumer makes only the minimum payment each month.  But surely more could be done.  The cost of slow repayment can be more effectively disclosed.  And other use-pattern information can also be disclosed.  For example, as Ronald Mann suggested, issuers could be required to disclose, through merchants, when a certain purchase would take the consumer over her credit limit, triggering an overlimit fee.  Such a disclosure could help the consumer avoid inadvertently exceeding her credit limit, perhaps by switching to another card or to another payment system.

(Speaking of disclosure, I should mention that I am currently working on an article that stresses the importance of use-pattern information and the need to pay more attention to use-pattern information in designing disclosure regulation.)

Comments

[Katie Porter]

I think the distinction that you draw between "product attribute" and "use pattern" disclosure is quite useful. I completely agree that the latter is underused in current disclosure schemes, and I see potential applications beyond credit cards. In the payday lending area, for example, I suspect that most people know that payday lenders charge a high APR. What they do not "know," is how very, very likely it is that they will end up rolling over the loan, which has the effect of multiplying the initial APR by a large factor. I've also often wondered whether "use disclosure" should have a role in consumer bankruptcy counseling. Should attorneys have to tell clients that the statistical chances of a Chapter 13 case leading to discharge are only 1 in 3? The current bankruptcy disclosure about chapter choice is the "product attribute" type that describes the law as written, but doesn't tell consumers anything about actual outcomes from bankruptcy.

[Jim Hawkins]

Is there any evidence about whether consumers consider pattern of use information? Since consumers have an optimism bias, I wonder if they'd dismiss statistics about, for instance, the high number of people who pay late fees.

[bailey]

This site is FANTASTIC!

[Adam Levitin]

I think the focus on "use pattern" in addition to "product attribute" informatin is quite useful, but "use pattern" information has its own problems.

First is generating a statistic that is sufficiently fine-tuned to the individual so as to be meaningful. Knowing that say 30% of consumers will pay late is too generic. A consumer needs to know her/his own personal risk profile, and I suspect that is quite expensive to generate (although perhaps it just involves putting a FICO score into an algorithm).

Second, use pattern information is a lot like historical securities prices or baseball statistics or the outcomes in a series of coin tosses--it is historical and not predictive as to the next instance. And then there is the framing of any use pattern--is it a 30% chance you will pay late or a 70% chance you will pay on time? (Or as a cautious card company might frame it, "at least a 70% chance you will pay on time")

[Oren Bar-Gill]

A few quick responses:
1. Optimism is indeed a problem. It will reduce the efficacy of average-use information (although not necessarily render such information completely ineffective). The solution is individual-use information. Sellers/lenders who maintain long-term relationships with consumers, e.g., credit card issuers, posses a lot of individual-use information.
2. Historical information can be predictive: If my credit card issuer tells me that I have paid late at least 4 times a year over the past 5 years, triggering hundreds of dollars worth of late fees, I can expect to pay late, and incur late fees, next year as well. And, perhaps, this disclosure will make me seek out a credit card with lower late fees.
3. Framing is always a problem – with use-pattern disclosures and with product attribute disclosures.

[Angie Littwin]

I agree with your point that "use pattern" disclosure has the potential to help consumers make decisions that are more in line with their long-term preferences. But I also worry about many consumers overcoming the effects these disclosures, even the ones about their own behavior, with optimism bias.

One possibility that has not been mentioned is that of the disclosures being designed to debias consumers. They could draw on specific debiasing techniques, such as vivid, concrete narratives. This is more plausible in a public-service campaign than on credit card bills themselves, but the principle is the same.

I'm also not sure that Ronald Mann's point about disclosing at point-of-sale when a consumer is about to go over her credit limit is actually a past use pattern disclosure. It's more of a warning about the present transaction in which the consumer is about to engage. The focus on the present limits the consumer's ability to think, "oh, I'm not going to make that mistake again," and is why I think it's likely to be highly effective.

[Michael Guttentag]

To anticipate and reject a challenge to your proposal, it may be useful to look at Securities and Exchange Commission disclosure regulations. A common complaint in that domain is that too much disclosure works to the competitive disadvantage of the company required to make the disclosure, and creates a disincentive to invest in information gathering activities. In fact, these disclosure costs typically create positive externalities, so there social welfare implications are far from clear. Moreover, a “gather and disclose” rule does not create a marginal disincentive to gather information.