Ducks and Economics

This Monday I attended a legislative forum on payday loans held by Missouri House representative Mary Still (D, MO-25). Also on the panel were Representatives Stephen Webber (D, MO-23), Chris Kelly (D, MO-24), John Burnett (D, MO-40) and Charlie Norr (D, MO-137). Roughly 18 people were scheduled to testify though unfortunately I was only able to be there for the first 5-7 or so witnesses. Here is the Columbia Tribune covering; here is coverage from the Columbia Missourian.

The big concern with payday lending is that we want to prevent people from overborrowing. It is a simple scenario: people in financial difficulty borrow money and can’t repay the loan on time, so the loan rolls over with additional fees or interest, aggravating an already difficult financial situation. As an outcome, this is undesirable, particularly when the victims are vulnerable to exogenous economic shocks and politically weak, like single mothers, or people with terminal illnesses living on a very limited income.

So what are the problems with payday lending? My biggest problem with both sides of the debate is a lack of data. Or rather, a lack of clear data. Payday loan providers tend to show off their own data that claiming that people who take out these loans are generally financially stable and use them as bridge loans. But data from organizations dealing with complaints (473 in Missouri in 2008) about payday lenders show that people making complaints are generally low-income and not financially stable. Some complaints also highlight questionably ethical or downright illegal conduct by payday loan companies in collecting debts, from vicious harassment to threats of jail or violence (this is particularly undesirable when it impacts the poor and politically weak). But we don’t know the true size or magnitude of these problems since neither sides of the market have incentives to accurately self-report (we can safely assume some proportion of consumer complaints are efforts to game the system and not legitimate complaints, though the Better Business Bureau does distinguish between legitimate complaints and illegitimate complaints).

Payday loan companies might be truly reporting the consumer data that they have. But there is no guarantee that their data is correct. Generally the only vetting of clients happens through an employment check, and clients may have incentives to provide deceptive or misleading information to get loans. So that data set may have serious problems. But data from organizations like the Better Business Bureau also faces some selection bias: it is likely that their survey data comes from people looking to use these legal structures to force a renegotiation of their loan terms or have registered complaints about the conduct of loan companies (I was unable to find a methodological note on their 2007 survey of 3,700 borrowers, which is important given that the industry did 2.8 million loans in Missouri in 2008). So BBB data may not include input from a representative sample of customers. I also don’t trust a lot of the data on industry profits, which this study by Tobacman and Skiba puts at around 10.1% (here is a Columbia Tribune article that provides an unattributed statistic of 6.6% for the 5 biggest payday lending companies). The problem with profit data from payday lenders is that a large number of transactions happen in cash, which leads me to believe that  there is some level of revenue that companies have the option of keeping off their books.

Here more data is useful in determining the size of the market and the impact of regulation. But we don’t need it to claim that the existence of claims alleging serious misconduct by businesses merits attention. What we do need is a cost benefit analysis that examines the incentives the market faces. Specifically, if regulation is enacted (and it is fairly well demonstrable that even basic regulations curtail payday lending operations substantially) where does the market re-locate? A possibility is that some of this business becomes part of the black or grey market. Is it worse to be gouged by regulated providers (companies with business licenses) or gouged by unregulated providers? Do markets with unregulated providers have better ability to vet clients, make better loans, and have fewer negative outcomes, since they aren’t constricted by the law? Do those markets have better social outcomes, despite the absence of legal protections for either party? Or do black markets end up looking up like the Mafia? This line of questioning is to me unanswered, but there are valuable lessons for the market in the economic literature.

Jon Zinman of Dartmouth has a forthcoming paper in the Journal of Banking and Finance that examines credit market substitutions that consumers make in the absence of access to credit. It finds that the market looks different when interest caps are in effect: payday loan providers start charging larger service fees where they still exist and households shift to credit cards and bank overdrafts to find liquidity. Zinman finds particularly Oregon interest rate caps left households worse off when their options for liquidity were curtailed by the exit of payday loan providers.

So there are two differing theories on expensive loans. First there is the idea that access to expensive credit leaves a percentage of users (we don’t know what percentage) worse off, because they aggravate the financial distress of people already struggling. There are several models in the economic and behavioral literature that support this. I point particularly to Rabin and O’Donohue in 2006, along with Carrell and Zinman 2008 (payday loans make airmen in the USAF worse off), and Campbell et al 2008 (where payday loans exist people are more likely to have their bank accounts involuntarily closed).

On the other side of the debate are arguments that restricting access to credit is bad. Here is Karlan and Zinman defending usury in the Wall Street Journal. Here is Deyoung also in the WSJ, looking at firm-level microdata from Colorado and concluding that regulation increases the cost of credit which leaves consumers worse off. Here is Adair Morse, who examines data from disaster financiers during the San Francisco earthquake and finds out that even expensive credit is key to maintaining the pre-disaster trends in human well-being (using human welfare indicators like the number of births, deaths, foreclosures, and substance abuse). And here is Greg Elliehausen who finds that consumers act rationally in calculating the costs of high-interest loans (my thought: the fact that consumers make rational calculations doesn’t mean that calculations are right or that outcomes are optimal, ex: the subprime loan market).

The problem that we face is that both sides of the debate are true in some fashion. Studies on both sides are hobbled by selection biases at some level. Morse, particularly, notes that her study faces a huge selection bias: people applying for disaster financing in San Fransisco are not representative of people applying for payday loans in conditions of economic instability, nor are the data on outcomes generally capable of catching people who are left worse off. Interesting aside: Morse references the Italian immigrant who later founded the Bank of America after events left him in a position to be a monopoly supplier of loans after the earthquake.

Perhaps regulation should focus on improving the vetting process, to select out the people who are likely to be worse off if they obtained a high interest loan. That avenue isn’t attractive to me because as a libertarian I distrust paternalistic government, but if we can establish that some level of regulation minimizes or eliminates undesirable social outcomes then it is better than the status quo. I do caution against regulations that drive this market underground because that hurts our ability to be conscious of its existence and as social engineers the limits on what we know about populations is very meaningful in terms of what we can or can’t conceptualize in terms of solutions.

I think that the lesson from the economic literature is that populations are not homogenous, generally, and that there are sectors of the market that would benefit from regulation, specifically the slice of consumers who would be left worse off in the case of default. It is also a very pertinent question of social policy as to how we can aid or assist this sector of the population without incentivizing free-riders. The difficulty lies in tranching the payday loan market without closing it.