Theory v Data: The Substitution Hypothesis

Economists often take the position that they would be glad to help poorer families, but that such efforts would either be a waste--or would make things worse.  In the area of credit regulation, for example, the constant pitch from some academics (and echoed by the banking lobbyists) is that any constraints on credit cards, for example, will result in driving the country's most vulnerable citizens to far more dangerous lenders.  The theory, called the substitution hypothesis, has been in vogue for two decades, helping to checkmate any serious effort to rein in predatory lending practices. 

The theory sounds plausible, but a little evidence can shake up a lot of ideas.  Angie Littwn has a terrific new piece that directly challenges the conventional wisdom regarding the substitution hypothesis, showing, for example, that those without credit cards simply have less total debt than those with credit cards--not that they take on "worse" forms of credit if they don't use credit cards.  But the really unexpected zinger in Angie's paper is that, according to the families she studied, there were no credit options worse than credit cards.  In other words, in the words of these struggling families, credit cards are as bad as it gets.

Options like pawn shops and rent-to-own, which sound terrible to my middle-class ears, were rated on par with credit cards.  According to the actual experiences of the people who lived with such credit, the credit cards got low ratings because they were seen as much more dangerous and more costly than pawn shops and rent-to-own.  To be fair, Angie's study came from residents of a state that has outlawed payday lending, so there still might be a lower-than-low rating for these lenders. 

So what would low-income families do if they didn't have access to credit cards or payday lenders and they really, really needed money?  Angie's subjects gave a resourceful answer:  borrow from friends and family.  In fact, respondents gave the friends-and-family approach a high rating.  Low cost, low risk, but enough interpersonal pain to make people not to turn to it unless the need was high.

Angie's study has a small N (these are very expensive data to gather), and she is quite careful to call her findings "preliminary."  But the work has broken open some new lines of inquiry.  Most importantly, the substitution hypothesis that has been accepted as fact looks a little less steady than it did before her paper was posted.

Comments

[Ken Doran]

What lenders exist that could be the beneficiaries of such a massive shift, and that are "more dangerous" than payday lenders and the worst credit cards? Does anyone imagine that violent underworld loan sharks exist, or would appear and thrive, in such mass quantities? If not that, who else? I admit to being unfamiliar with this literature, but it doesn't sound like I am missing much.

[Bob]

Two questions:

First, how does the fact that, on average, a person without cards has less total debt than a person with cards prove the "subsitution" hypothesis is wrong? Isn't it quite possible this can be explained by either (1) less demand for debt or (2) less supply of debt, credit card or otherwise, offered to these consumers?

Second, is it immediately obvious that an outcome where people don't have cards, and have less debt is better than an outcome where they have a credit card and more debt?