Nudging from Debt: The Role of Behavioral Economics in Regulation

Click here to download this article as a PDF.

As Harvard academics like Cass Sunstein and Elizabeth Warren have joined the regulatory ranks of the Obama Administration, “behavioral” economics has come under increased attention. Behavioral economics – the idea that individual behavior departs from strictly rational action in routine and predictable ways – has its roots in the Nobel prize-winning experimental and academic work of Daniel Kahneman and Amos Tversky. [1]

In more recent years, what started as a departure from dominant notions of the rational actor has been extended and popularized so that it now is a routine aspect of economic analysis of the regulatory state. Dan Ariely, for example, has worked to integrate several distinct “behavioral” biases in a general framework of predictable irrationality. [2] Meanwhile, the work of Sunstein, Obama’s regulatory czar, along with Richard Thaler, has made the term “nudge” a byword in policymaking circles. [3]

The central advantage of behavioral economics for those seeking to justify regulatory intervention is the premise that individuals make mistakes all the time: It is much easier to justify intervention to protect individuals from “errors” that behavioral analysis says happen all the time than it is to protect individuals from the adverse effects of decisions thought to be fully rational. So, policymakers trying to rein in consumer lenders naturally turn to behavioral finance as a way to justify constraints in consumer finance markets. Several of the most important parts of the Credit CARD Act of 2009 rest directly on the supposition that cardholder decisions about spending and repayment reflect systematic misapprehension of the likely patterns of future behavior. [4]  For example, provisions that invalidate onerous contract terms make sense if consumers are not likely to understand and “price” those terms when they acquire a credit card or use it for payment transactions. [5]

The work of Elizabeth Warren and her co-author, Oren Bar-Gill, gives us every reason to expect that the Consumer Financial Protection Bureau (CFPB) will rely heavily on behavioral analysis to support regulation. Bar-Gill’s aptly titled early work on credit cards (“Seduction by Plastic”) argues that because the profitability of credit cards is enhanced by the behavioral deficits of cardholders, credit card lenders are driven to compete as to who can most aggressively exploit those deficits. The result, Bar-Gill argues, is a race to the most exploitatitive, which easily would justify regulatory intervention. [6] His more recent work with Elizabeth Warren generalizes the concern to the entire consumer credit industry and argues that the problems are so endemic as to justify the creation of an agency with broad authority to regulate consumer credit products. [7]

I have long harbored a suspicion of the facile utility of behavioral analysis for regulatory design and was an early skeptic of the CFPB, at least as it was originally designed. [8] My general concern comes not from skepticism about departures from rationality, as I am sure that consumer use of financial products falls far short of the perfection of the rational actor. Rather, my point is a typical “second best” critique: The departures from rationality are so unpredictable and contextually specific that intervention designed to remedy one departure without accounting for the others has little chance of a beneficial result. [9] So much of the literature focuses on a single behavioral deficit that can be regularized in an experimental setting and investigates the likely policy consequences of the assumption that the deficit dominates all other concerns for consumers faced with a consumption decision.

In reality, however, the choices consumers face are far too complex to be captured in a single model of a right/wrong decision. I can make the point most aptly by considering the decisional terrain that confronts a potential payday loan customer. From the perspective of regulatory advocates, like Bar-Gill and Warren, the behavioral weakness is obvious: The transaction seems attractive only because the customer, suffering from an optimistic bias, falsely believes that it will repay the loan after two weeks. In fact, this regulator knows the customer is likely to roll the loan over several cycles before finally repaying it in full. [10] To remedy that bias, a regulator of an earlier generation might have banned the transactions entirely on the paternalistic premise that no rational borrower would use the product. [11] A gentler, modern regulator might establish a limit on rollovers, forcing borrowers to repay after a few cycles rather than continue to borrow. Even less forcefully, a regulator might require some strong affirmative action to overcome a “nudge” against further borrowing. In recent years, states like Florida, Oklahoma and New Mexico have adopted increasingly rigorous restrictions to limit repetitive rollover borrowing from payday lenders. Thaler and Sunstein trumpet this type of intervention as more benign and less intrusive, because its effects on consumer choice are limited and indirect. Like earlier writers who advocated “asymmetric paternalism,” [12] they argue that this brand of “libertarian paternalism” is to be preferred, because it allows those borrowers that strongly prefer the product disfavored by regulators to exercise their preferences. [13]

The question is whether this course is well considered. The goal of the regulation, assuming the nudge is strong enough to have any effect at all, [14] will be some constraint on the payday lending market. The core question, then, is whether the posited behavioral deficit – the optimistic bias about rollover behavior – justifies an intervention to constrain use of the product. Let me offer three complications of the behavioral story that suggest hesitation might be the better course here.

The first is the simplest: What is the empirical evidence to support the supposition of undue optimism? To be sure, there is considerable experimental evidence to support the general idea of an optimistic bias. [15] However, little of that evidence involves actual market transactions. To the extent we know anything about the actual expectations of potential payday borrowers, the work of Marianne Bertrand and Adair Morse suggests that on average they have a pretty accurate assessment of the likely rollover cycle. Thus, among the questions Bertrand and Morse asked payday loan borrowers in their forthcoming survey was their estimate of the average time it takes a payday loan borrower to repay a loan. The mean response that Bertrand and Morse report (5-6 weeks) is quite close to the correct answer based on existing research about payday loan business practices. [16] Although more direct examination of the question would be valuable, this suggests that the blithe assumption of a serious misestimation of likely rollover behavior is unfounded.

There are two distinct reasons why a bias towards optimism might not be dominating payday borrowing decisions. One is that borrowers in that context are simply better informed and more cunning than Bar-Gill and Warren suppose. The second is that any bias towards optimism is swamped by other complicating behavioral deficits that also inform the borrowing decision. [17] One intriguing candidate for a competing behavioral deficit is that borrowers use payday loans because of a quasi-rational commitment to abstain from using other lower-priced forms of credit. The empirical evidence here comes from Sumit Agarwal, Paige Skiba and Jeremy Tobacman, who show that payday borrowers often use the produce even when they have available unused credit. [18] At first glance, this seems a strong indicator of financial illiteracy: Interest rates even on high-cost credit cards typically are below 40 percent per annum, and interest rates on typical payday loan products are on the order of 400 percent per annum.

Yet what if the borrower in question has an internal pre-commitment against further credit card borrowing? Such a pre-commitment might rest on a concern about the future costs of excessive credit card use. Because credit card borrowing has the potential to grow to high levels that might burden a cardholder for decades into the future, [19] a pre-commitment against borrowing beyond a level perceived to be manageable is a rational strategy to avoid the long-term costs of high credit card debt. Because payday loans – however high the interest rate – do not have the potential to grow to a large amount, their long-term risks are much less significant. [20] So my basic point is if the regulator pushes that borrower out of the payday market, the regulator might well be pushing the borrower into borrowing in a more open-ended (and potentially riskier) credit card transaction.

Conversely, what should we think if the regulation deters the customer from borrowing at all? We know that the relatively young are the dominant demographic segment among payday loan customers. [21] If we consider the likely problem of debt aversion, [22] we might take the point that borrowing by the young is substantially below the level that rational actor theory suggests. [23] Taking that problem seriously, we would worry that regulation that lowers the overall amount of borrowing by the young would not help them but would rather exacerbate an existing general tendency to borrow at an infra-rational level. It well might be that few payday loan customers invest the funds in conventional investments that earn a return that exceeds the interest rates they pay the lenders. It too just as well might be the case that payday loans that stave off critical defaults to other creditors (car lenders, utility providers, health care providers, etc.) produce a return to the borrower that exceeds the interest rates payday lenders charge. Yet that is just my point: A reflexive response to a perceived optimistic bias that ignores the polycentric decisional terrain that borrowers confront is no more likely to improve their situations than it is to harm them.

My portrait of a stereotypical payday loan customer is of course just that – a stereotype that will not describe all or even most of the market. But my narrative differs from reality primarily because it is too simple, not because it is too complex. The lure of economic analysis is that it simplifies what is multifaceted and offers a straight and general answer to what seems unmanageable in its florid heterogeneity. The purpose of this brief sketch is to suggest that the problem here is not occasional anecdotal departures from a regularity of excessive optimism. The problem here is that the general truth is in fact not simple and perhaps not even excessively optimistic. Until we know more about how and why consumers decide to borrow in this market, we cannot sensibly intervene to stop them from doing so.

—–

Endnotes

[1] Kahneman and Tversky gained fame for their pathbreaking experiments showing systematic departures from rationality. For a useful collection, see Daniel Kahneman and Amos Tversky, Judgment Under Uncertainty: Heuristics and Biases (1982).

[2] Ariely’s work builds on the large literature that stems from Kahnemann and Tversky’s work. In recent years, he has articulated a framework for organizing all of the “quasi-rational” effects of that literature. See Dan Ariely, Predictably Irrational: The Hidden Forces that Shape Our Decisions (2008).

 [3] See Richard H. Thaler & Cass R. Sunstein, Nudge: Improving Decisions About Health, Wealth, and Happiness (2008).

[4] See Ronald J. Mann, Charging Ahead: The Growth and Regulation of Payment Card Markets (2006).

[5] For example, CARD Act §§ 102(j) (prohibiting “double-cycle” billing clauses), 102(k) (requiring cardholders to “opt in” to over-the-limit fees), 104 (mandatory rules for application of card payments), 106 (mandatory rules about due dates on credit card accounts), 171 (limiting circumstances in which issuers can increase interest rates).

[6] 98 Nw. U. L. Rev. 1373 (2004).
 
[7] Making Credit Safer, 157 U. Pa. L. Rev. 1 (2008).
 
[8] Ronald J. Mann, Unsafe at Any Price?, 157 PENNumbra 167 (2009).

[9] See R. Lipsey & K. Lancaster, The General Theory of Second Best, 24 Rev. Econ. Stud. 11 (1956); Thomas S. Ulen, Courts Legislatures, and the General Theory of Second Best in Law and Economics, 73 Chi.-Kent L. Rev. 189 (1998).

[10] Bar-Gill and Warren suggest that this specific mistake dominates the decision-making process of a large share of payday loan customers. 157 U. Pa. L. Rev. at 44, 55-56. For similar contentions, see National Consumer Law Center, Stopping the Payday Loan Trap: Alternatives that Work, Ones That Don’t (2010).

[11] Usury limits operate on that premise, barring loan transactions with high interest rates even for borrowers that find them attractive. E.g., 49 U.S.C. § 987 (prohibiting payday and certain other loans to military personnel with an interest rate greater than 36 percent).

[12] Colin Camerer et al., Regulation for Conservatives: Behavioral Economics and the Case for “Asymmetric Paternalism,” 151 U. Pa. L. Rev. 1211 (2003).

[13] Thaler & Sunstein, supra.

[14] It is not at all clear that a nudge will be effective. For one, some borrowers are driven to payday lenders by the liquidity constraints that are a commonplace of the market economy: The need to pay a utility bill to avoid interrupted service or to pay a car lender to avoid repossession. More generally, the counterpoised social impulses that shape consumption patterns are likely to be powerful: Households are simultaneously deterred from borrowing by a felt stigma associated with debt and driven to spend to maintain social position. In either case, if the motivation for the transaction is substantial, no mere nudge will deter the transaction. If any generalization is appropriate, it might be that the nudge will be most effective on the better off group of payday lending customers, for whom other liquidity opportunities are more likely. The effect, then, would be to increase the relative prevalence of the less well off among payday lending borrowers.

[15] David A. Armor & Shelley E. Taylor, When Predictions Fail: The Dilemma of Unrealistic Optimism, in Heuristics and Biases: The Psychology of Intuitive Judgment (Thomas Gilovich et al. eds. 2002) (reviewing literature about optimistic bias).

[16] Marianne Bertrand & Adair Morse, Information Disclosure, Cognitive Biases and Payday Borrowing, Journal of Finance (forthcoming).

[17] Kahneman himself emphasizes that financial decisions will depend not only on monetary considerations but also on factors related to general social identity. Daniel Kahneman, Chapter Maps of Bounded Rationality: A Perspective on Intuitive Judgment and Choice, in The Nobel Prizes 2002 (2003); see Brooke Harrington, Pop Finance: Investment Clubs and the New Investor Populism (2008).

[18] Sumit Agarwal, Paige Skiba & Jeremy Tobacman, Payday Loans and Credit Cards: New Liquidity and Credit Scoring Puzzles, American Economic Review Papers and Proceedings (May 2009).

[19] See Ronald J. Mann, Bankruptcy Reform and the “Sweat Box” of Credit Card Debt, 2007 Ill. L. Rev. 375.

[20] See Ronald J. Mann and Jim Hawkins, Just Until Payday, 54 UCLA L. Rev. 1 (2007).

[21] Data from the Federal Reserve Board’s Survey of Consumer Finance suggest that as of 2007 the median age of payday loan customers was about 36.

[22] See Erica Field, Educational Debt Burden and Career Choice: Evidence from a Financial Aid Experiment at NYU Law School, 1 American Economic Journal Applied Economics 1 (2009); Peter Tufano, Consumer Finance, 1 Annual Review of Financial Economics 227 (2009).

[23] Ian Ayres & Barry Nalebuff, Lifecycle Investing: A New, Safe, and Audacious Way to Improve the Performance of Your Retirement Portfolio (2010).

Click here to download this article as a PDF.