Deceptive for Whom? The Implications of Behavioral Economics Driven Consumer Financial Services Policy

Deceptive for Whom? The Implications of Behavioral Economics Driven Consumer Financial Services Policy July 2016 Dr. Xiaoling Ang Principal Consultant | Washington, DC [email protected] +1 202.580.7744 Dr. Stephen G. Bronars Partner | Washington, DC [email protected] +1 202.580.7777 I. Introduction Behavioral economics interventions, which include policies such as mandated disclosure that target how consumers receive and process information, have received a fair bit of attention in consumer financial services. The Consumer Financial Protection Bureau (CFPB) has a history of engaging prominent scholars in behavioral economics: Sendhil Mullainathan, the Robert C. Waggoner Professor of Economics at Harvard, was appointed the first Assistant Director of Research in the spring of 2011;1 Eric J. Johnson, the Norman Eig Professor of Business at Columbia Business School, has been a visiting scholar since 2014;2 and Professor Richard H. Thaler, the Charles R. Walgreen Distinguished Service Professor of Behavioral Science and Economics at the University of Chicago Booth School of Business, David Laibson, the Robert I. Goldman Professor of Economics at Harvard, and Justine S. Hastings at Brown3 have served on the CFPB’s Academic Research Council. Cass R. Sunstein, the Robert Walmsley Professor and the founder and director of the Program on Behavioral Economics and Public Policy at Harvard Law School was formerly the Administrator of the White House Office of Information and Regulatory Affairs, which reviews the Paperwork Reduction Act and regulatory impact analyses conducted by the CFPB, from 2009 to 2012.4 Insights from behavioral research conducted by these scholars as well as the CFPB staff’s understanding of and contributions to the behavioral economics literature are likely to interact with the design, implementation, and enforcement of consumer financial services law. These concepts are likely to be applied in analyses of potentially unfair, deceptive, or abusive acts or practices (UDAAP).5 Behavioral economics considers how consumer and firm behavior may be affected by how information is presented and interpreted by consumers. www.edgewortheconomics.com 2 | www.edgewortheconomics.com Deceptive for Whom? The Implications of Behavioral Economics Driven Consumer Financial Services Policy Consequently behavioral economics may be especially useful in analyses of alleged deceptive practices where consumers’ interpretations of these practices are salient. These interpretations are unlikely to be uniform across consumers, which, in the language of class action litigation, may be at odds with the commonality of the alleged deceptive practices. As we discuss in more detail later in the article, variation in the way in which policy affects different consumers may have greater repercussions in consumer financial services markets than in markets for other goods and services. In fact, in its standards for CFPB rulemaking, §1022(b)(2) of the Dodd-Frank Consumer Financial Protection and Wall Street Reform Act (Dodd-Frank Act) explicitly requires that the CFPB consider “the potential benefits and costs to consumers and covered persons, including the potential reduction of access by consumers to consumer financial products or services resulting from such rule.”6 In addition, the Dodd-Frank Act requires that regulators consider the effects a policy may have on different groups of consumers.7 The potential effects of a behavioral intervention,8 in contrast with price ceilings or product bans, are not explicitly part of the regulation text or administrative or court decisions, but instead flow from adjustments made by affected parties. Economists refer to these responses and adjustments as general equilibrium effects.9 While there are various types of behavioral interventions, for the purposes of this article we’ll focus on the general equilibrium effects of informational disclosure, given the prominence of mandated disclosure as the focus of §1032 of the Dodd-Frank Act.10 A disclosure intervention may be particularly appealing to policy advisors because the letter of the intervention does not explicitly limit consumer choice, product offerings, or pricing. The absence of these direct limits on choice may make it easier for constituencies with differing points of view to support the intervention.11 There’s an active research literature on whether and in what circumstances mandated disclosure impacts real life,12 and, in full disclosure, we’ve contributed to this research.13 While we cannot do justice to the body of behavioral economics literature or even the literature on disclosure in the span of this article, by relying on a few simple examples, we hope to shed light on some issues that should be considered when analyzing proposed behavioral policies or evaluating potential deceptive acts or practices related to disclosures, correspondence with consumers, or advertising. In Section II we discuss the centrality of consumer composition in consumer financial services relative to other industries, then in Section III we present an example-driven introduction to the behavioral insights related to disclosure. Section IV consists of qualitative behavioral analyses of the general equilibrium effects of two potential consumer financial mandated disclosures in student loans and deposit account overdrafts that apply the concepts introduced in Section II and III. Section V concludes. II. The Importance of Consumer Composition in Consumer Financial Services Financial service providers, should, by design, be more concerned with the composition and behavior of borrowers after a product is selected than providers of many other consumer goods and services. To illustrate the contrast, let’s first consider a physical consumer product, like a snack cake. The snack cake manufacturer does not particularly care which customers purchase the product or what happens with the snack cakes after they’re sold. To the seller, whether Steve and Ling Ling each buy one snack cake or Steve buys two snack cakes, total revenue, costs and profits are the same. In addition, the seller does not care whether the snack cakes are eaten or thrown away. Now consider the case of a consumer financial product, the residential mortgage. The loan owner14 (initially the lender) is interested in the mortgage being used to purchase real property because the www.edgewortheconomics.com | 3 Deceptive for Whom? The Implications of Behavioral Economics Driven Consumer Financial Services Policy property is collateral for the loan. If the terms of the mortgage are not fulfilled and the loan is in default, the property can be sold to make the lender whole. Furthermore, over the life of the loan, the borrower has repeated interactions with the lender, perhaps through a contracted servicer. Unlike snack cakes where payment is received at point-of-sale, the loan owner’s revenues are received incrementally through monthly mortgage payments. A disruption in this cash flow—a delinquency or default—affects the value of the loan. This repeated interaction means that the lender has a vested interest in how much is borrowed, who is borrowing, and whether the borrower can repay the loan—this is the motivation for underwriting. The lender also cares how much of a product a particular consumer uses. Consider a customer with an excellent credit history and an income of $60,000 who takes out a thirty-year fixed rate mortgage for $300,000 at 4% for a house valued at $375,000. The fundamentals look solid: the loan-to-value ratio (LTV) is 80% and, assuming this is her only debt, the debt-to-income ratio (DTI) is 28.6%, so this is a qualified mortgage15 that does not require private mortgage insurance. Now if the borrower wants to take out another identical mortgage the LTV would be 160% and his DTI would be 57.3%. Unlike the case of one consumer buying two snack cakes, a second mortgage would probably raise red flags. The lender would probably refuse to make the loan, reduce the loan amount, or charge an interest rate much higher than 4%. This example illustrates how borrowing levels might affect the products offered to one borrower, but does not show how the actions and decisions of one borrower might affect other borrowers. One way that one borrower can affect another borrower in a loan market is through the risk pool. While it’s difficult to determine whether any individual borrower will default on a loan, through data analytics a lender can draw inferences about how a group of borrowers may perform on average. These analyses inform which loan applicants will receive loans and how much they will be charged. Suppose that a lender makes 100 loans for $100 with a single payment at the end of one year. Assume that lenders, if repayment were guaranteed, would be willing to make the loan at an annual interest rate of 10% for total interest income of $1000. If it is expected that 95% of borrowers will repay the loan and 5% of borrowers will default, then expected interest income declines to $950, and $500 in principal is lost, which lowers net revenue to $450. To receive $1000 in net revenue from this loan product, or $10 per loan, the interest rate would have to increase to 15.79%.16 What happens when the mix of borrowers changes? Suppose that borrowers who repay the loan are 100% reliable, and the remaining borrowers are simply not going to repay the loan, and the lender cannot tell whether a borrower is reliable or unreliable with available information. A change in market conditions that causes some reliable borrowers to no longer apply for a loan will make the pool of borrowers riskier. If 10% of borrowers are expected to default, lenders can only earn $10 per loan on average, if an interest rate of 22.22% is charged.17 The riskier the pool gets, the higher interest rates will need to be for lenders to be willing to make loans. But keep in mind that the “reliable” borrowers who remain in the pool, as other “reliable” borrowers exit, now pay a substantially higher interest rate even though they have not changed their own behavior or their reliability as borrowers. III. What is Behavioral Economics? The “behavioral” part comes from repeated empirical observations of people acting in ways that don’t appear to make sense when we consider (what we believe is) the full range of options available to the person, including the cost of each option. Economic reasoning applies to any situation where resources are scarce, whether budgets are measured in dollars, calories, or time. Behavioral economics is the marriage of these two concepts. 4 | www.edgewortheconomics.com Deceptive for Whom? The Implications of Behavioral Economics Driven Consumer Financial Services Policy To make this more concrete, suppose that you are trying to eat a healthy, calorie-restricted diet. You have plans to eat fresh produce, whole grains, healthy proteins and fats, and generally be mindful of what you eat. But...there’s cake in the kitchen at work. It’s run-of-the-mill supermarket birthday cake. You wouldn’t buy it instead of bananas if you were at the supermarket. But it’s there, next to the coffee, and you have a meeting that’s probably going to run over lunch. Cake it is! The cake can seem to be the best alternative at the time if you aren’t thinking of a broad range of food alternatives, but rather about what is directly in front of you. There may be several brick-and-mortar restaurants serving lunch near your office and even a few food trucks in the neighborhood, but evaluating all available food alternatives takes time and effort; each restaurant has a menu and research is required to learn which food trucks are within walking distance (even if you follow your favorites on Twitter). Filtering through all of this information can pile on pretty quickly,18 so it may make sense to focus on the limited information at hand. And cake is pretty hard to resist—we can have a completely separate discussion about the self-control19 behavioral economics literature. Behavioral economics incorporates insights from psychology to build on a substantial body of economic research. It studies how people use scarce resources, and what happens when constraints change. The difference between behavioral economics and the version of economics you may have encountered in an introductory economics class in college is that behavioral economics incorporates potential psychological or cognitive factors that may influence how people weigh tradeoffs. What does this have to do with cake? Suppose the menu for the salad restaurant next door were posted next to the cake. Posting the salad menu might cause some people to resist the temptation of the cake because they are now thinking about a salad. For others, the craving for cake will outweigh the effect of the salad menu. If we compared total birthday cake consumption on days when salad menus were posted to cake consumed on days when the menu was not visible, we might be able to calculate the average effect of the salad menu intervention on all office personnel. But we might also care about the effects on specific individuals of this intervention. Suppose the menu dissuades Steve from eating cake, but Ling Ling is not swayed by the intervention. Because Steve has resisted eating cake, more of it is available and Ling Ling may now choose to eat a second piece of cake. If an unexpected phone call causes Steve to be too busy to buy a salad for lunch, is he better off for having resisted the temptation of the cake? The only intervention in this example is the prominent placement of a piece of paper, which is essentially a disclosure. As we’ve illustrated, there can be consequences of introducing a disclosure even for people who appear to ignore the information in the disclosure. This is especially true when the availability or price of the regulated product depends on the composition of people who purchase the product. IV. Applications of Behavioral Economics to Consumer Financial Service Disclosures Behavioral interventions can affect different people in different ways and impact the mix of consumers who want a particular consumer financial product. This can affect the price of the products available as well as cause changes in which products are offered. This implies that behavioral interventions, by potentially changing risk pools, may have real effects on access to and the price of credit, despite not setting limits on pricing or product offerings. To get a handle on how this might work, we think about two hypothetical mandated disclosures: one in the student loan market and one for deposit account overdrafts.20 A. Overborrowing in the Student Loan Market? There are various factors that should enter into the www.edgewortheconomics.com | 5 Deceptive for Whom? The Implications of Behavioral Economics Driven Consumer Financial Services Policy decision to take out a student loan, including how much a student expects to earn after completing his education and how certain he is that he will complete his schooling. These decisions are tied to decisions about what to study, where to go to school, and how much to work while in school. While students who may have borrowed too much (overborrowers) may receive a lot of press,21 there is also the potential to borrow too little (underborrow).22 Underborrowing can manifest in different ways: a consumer may forgo college or training in order to work because he does not appreciate how much his lifetime income could increase with training, or because he’s not confident in his ability to succeed in school. Students may also be overly zealous about avoiding student loans and work more hours for pay while in school, for which they may suffer academically.23 On the flip side, overborrowers may overestimate their future earnings or academic success, or choose to spend more on non-academic related expenses because they assume they can pay for those expenses in the future. Both overborrowers and underborrowers would be better off if they changed their borrowing behavior, but these adjustments should be in opposite directions. Consider a campaign to inform consumers about the potential dangers of borrowing too much in student loans. In addition to potentially reducing borrowing by overborrowers, the intervention might dissuade borrowing by some consumers with a tendency to underborrow, as shown in Figure 1. In this case the intervention could exacerbate the tendency to underborrow. If prices (interest rates) don’t change, the affected overborrowers are better off and the affected underborrowers are worse off. Prices could change, however, if the mix of credit risks changes. Underborrowers may be better credit risks because they’re borrowing less relative to their expected earnings while overborrowers may not be increasing their future earnings by borrowing more. So the effect of the exit of overborrowers on expected future cash flow per borrower may be offset by the exit of underborrowers. If the effect of underborrowers exiting is stronger than the effect of overborrowers then prices (interest rates) should increase, making all borrowers that remain in the market worse off (and vice-versa). Figure 1: Student Loan – Intervention to Discourage Overborrowing, Effect on Amount Borrowed B. Deposit Account Overdraft Consider an informational disclosure that makes deposit account overdraft more salient.24 This could potentially be implemented through highlighting present bias (a tendency towards impatience) or providing specific information about the costs of overdraft. In the following example we assume the current industry-standard per-transaction fee structure of overdraft fees and a fee of $34.25 Suppose that consumers can be divided into groups based on how much attention they pay to the possibility of overdrafting (attentive vs. inattentive) and whether they have another source of liquidity to draw on such as a credit card or another bank account (unconstrained vs. constrained). This is illustrated in Figure 2(A). The consumers in the upper left quadrant pay attention, have enough funds, and never overdraft. The consumers in the upper right quadrant have enough funds but overdraft because they’re not paying 6 | www.edgewortheconomics.com Deceptive for Whom? The Implications of Behavioral Economics Driven Consumer Financial Services Policy attention—by inadvertently selecting their debit card instead of a credit card or forgetting about an online payment. Consumers in the lower right quadrant overdraft because they are out of funds, but they are not paying attention to their balance and are unaware of the first transaction that triggers an overdraft. The last group in the lower left is interesting—they are also out of funds but they are strategic in their use of overdrafts. Suppose that a consumer in this category has to make $100 of purchases—$50 at the grocery store, $20 at the pharmacy, and $30 at a big box store—but only has $10 in her checking account. She could trigger an overdraft fee on each purchase for a total of $102 in fees or could anticipate these fees and instead withdraw $100 from an ATM and incur a total of $34 in fees. Granted, consumers would be better off if fees were lower, but conditional on the fee structure and financial constraints, she has minimized overdraft fees. Figure 2(B) represents an intervention that causes more consumers to be attentive to overdraft fees. Formerly inattentive, unconstrained consumers who became attentive are made better off because they no longer pay overdraft fees, and are represented in green. Formerly inattentive, constrained consumers, in yellow, now consolidate their potential overdraft transactions into a single ATM withdrawal. If the overdraft fees stay the same, then no consumers have been harmed and some have benefited from this intervention. A key insight is that the number of overdraft transactions declines and both the composition of who overdrafts and the average magnitude of an overdraft also change. To the extent that some smaller overdrafts were cross-subsidizing other larger overdrafts, fees may adjust to reflect the new smaller pool of consumers with larger average overdraft amounts. Financially constrained consumers in the lower left quadrant and in the yellow area will struggle to reduce the number and magnitude of overdrafts, and will likely be harmed if the financial institution raises overdraft fees. Inattentive consumers on the right side of the figure do not adjust their overdrafts in response to fees. Consequently, the bank could raise the fee making all the consumers in the red areas strictly worse off. Furthermore, while some consumers in the yellow area may gain from the intervention, the gain from a reduction in overdrafts could be offset by a higher fee so that some consumers may be harmed overall. V. Conclusion Behavioral economics is a vibrant area of economic research and has the potential to provide valuable insight to policy makers. However, it is important to keep in mind that it is embedded in a broader economic framework; the general equilibrium consequences of behavioral interventions should be taken into account. These consequences are especially important in consumer financial services markets, because the composition of consumers who use a product can affect risk pools, pricing, and product offerings. We hope that in this article we have demonstrated that it can be foolhardy to treat consumers as a homogeneous group in considering the effect of either behavioral interventions or alleged deceptive practices. Behavioral economics does not assume that all consumers are identical, so some consumers may not be harmed at all by, or may even benefit from,26 an alleged deceptive act. Behavioral economics in the context of disclosures raises many practical considerations for industry participants and regulators, including whether information may be omitted in the interest of brevity and clarity, whether consumers can be expected to interpret the disclosure reasonably under the circumstances, and whether or not it comes across as misleading to certain consumers. When analyzing the impacts of behavioral consumer financial services interventions through regulation, litigation, supervisory guidance, or enforcement, policy makers should take into account not only the average direct effects of the interventions, but also how different consumers may be impacted indirectly by the www.edgewortheconomics.com | 7 Deceptive for Whom? The Implications of Behavioral Economics Driven Consumer Financial Services Policy intervention through the general equilibrium effects on prices and product offerings. These impacts can often be evaluated empirically using real world data and applying tools from economics— much of the work in behavioral economics has been quantitative. To study the effects of a practice that might be considered deceptive, we would apply econometric techniques to a financial institution’s data on customer characteristics (e.g., credit scores, collateral, Census tract) and product usage (e.g., repayment history, draws on lines of credit), and then evaluate how the mix of consumers using the product and their use of the product may be affected by the practice or act in question. We can happily spend (and have spent) hours discussing these approaches in different applications, but will practice self-control and save these topics for a later discussion. Figure 2: Overdraft Salience Intervention 8 | www.edgewortheconomics.com Deceptive for Whom? The Implications of Behavioral Economics Driven Consumer Financial Services Policy Notes 1 Press Release, CONSUMER FINANCIAL PROTECTION BUREAU, Treasury Bureau Announce Senior Leadership Hires for the Consumer Financial Protection Bureau (May 11, 2011) available at http://www.consumerfinance.gov/about-us/newsroom/ treasury-departmentannounces-senior-leadership-hires-forthe-consumer-financial-protection-bureau/. 2 COLUMBIA BUSINESS SCHOOL. Eric. J. Johnson CV (Jan. 2016) available at http://www8.gsb.columbia.edu/cbs-directory/ sites/cbsdirectory/files/faculty_cvs/Johnson%20Eric%20 -%20%28MKT%29%20-%20CV%202015.pdf (last visited May 11, 2016). 3 CONSUMER FINANCIAL PROTECTION BUREAU, Academic Research Council Members (April 2014) http://files. consumerfinance.gov/f/201404_cfpb_bios_academicresearch-council.pdf. 4 Cass R. Sunstein, HARVARD LAW SCHOOL. http://hls.harvard. edu/faculty/directory/10871/Sunstein (last visited May 11, 2016). 5 See CONSUMER FINANCIAL PROTECTION BUREAU, CFPB Supervision and Examination Manual, Version 2 at 110 (Oct. 2012) http://files.consumerfinance.gov/f/201210_cfpb_ supervision-and-examination-manual-v2.pdf. 6 12 U.S.C. § 5512(b)(2) (2010). 7 See 12 U.S.C. § 5521(b)(2)(A)(ii) for a reference to rural consumers. 8 A behavioral intervention is a policy designed to counteract consumers’ behavioral biases, such as a tendency towards instant gratification. These interventions typically take the form of targeted information provision. 9 General equilibrium effects are the net effects of a policy once all adjustments are made. To the extent that there are multiple consequences of a policy, general equilibrium analysis takes into account the interactions between all of these effects. 10 12 U.S.C. § 5532. 11 See Lauren E. Willis, When Nudges Fail: Slippery Defaults, 80 U. CHI. L. REV. 1155–1229 (2013) at 1158. 12 See Omri Ben-Shahar & Carl E. Schneider, The Failure of Mandated Discourse, 159 U. PA. L. REV. 647 (2010); Eric J. Johnson & Daniel G. Goldstein, Do defaults save lives?, 302 SCI. 1338-1339 (Nov. 21, 2003). 13 Xiaoling Ang & Alexei Alexandrov, Choice Architecture versus Price: Comparing the Effects of Changes in the U.S. Student Loan Market (June 29, 2016), http://papers.ssrn.com/ abstract=2504660. 14 By loan owner we mean any owner over the life of the loan— if a loan is sold, the seller has an interest in observable loan quality at time of sale. 15 CONSUMER FINANCIAL PROTECTION BUREAU, Basic Guide for Lenders: What is a Qualified Mortgage?, http://files. consumerfinance.gov/f/201310_cfpb_qm-guide-for-lenders. pdf (last accessed May 11, 2016). 16 This comes from 95 × $100 × 15.79% = $1500 in interest payments and 5 × $100 = $500 in lost principal. 17 This comes from 45 × $100 × 22.22% = $1000 in interest payments and 5 × $100 = $500 in lost principal, for = $10 revenue per loan. 18 As the people writing this article, and probably the people reading this article, are paid to filter information, this is not necessarily a bad thing from our point of view. 19 Self-control is the ability to forgo something enjoyable today in order to receive future benefits. Procrastination is an example of a failure of self-control. For a discussion of the impact of self-control biases on credit markets, see Paul Heidhues & Botond Kőszegi, Exploiting Naïvete about SelfControl in the Credit Market, 100 AM. ECON. REV. 2279–2303 (2010), http://www.wiwi.unibonn.de/kraehmer/Lehre/ TopicsSS14/HeidhuesKoszegi-ExploitingNaivete.pdf. 20 There are currently mandated disclosures in both of these markets: see 34 CFR 601.11 for the Self-Certification Form for private educational loans and 12 CFR 1005.17 for the A-9(A) model consent form for deposit account overdraft services. 21 See Andrew Martin & Andrew Lehren, A Generation Hobbled by the Soaring Cost of College, N.Y. TIMES, May 12, 2012, A1, http://www.nytimes.com/2012/05/13/ business/student-loans-weighing-down-a-generation-withheavydebt.html?pagewanted=all&_r=0; Break the death grip of student loans ROCHESTER DEMOCRAT & CHRON., May 16, 2016 http://www.democratandchronicle.com/story/ opinion/editorials/2016/05/16/break-death-grip-studentloans/84422212/. 22 Christopher Avery & Sarah Turner, Student Loans: Do College Students Borrow Too Much—Or Not Enough?, 26 J. ECON. PERSP. 165–192 (2012). 23 Sarah Grant, You Can’t Work Your Way Through College Anymore, BLOOMBERG (Oct 28, 2015) http://www.bloomberg. com/news/articles/2015-10-28/you-can-t-work-your-waythrough-college-anymore. $500 50 www.edgewortheconomics.com | 9 Deceptive for Whom? The Implications of Behavioral Economics Driven Consumer Financial Services Policy 24 See 80 Fed. Reg. 53503 (2015). The CFPB issued a 60 day Paperwork Reduction Act Notice on September 4, 2015 for an information request titled “Web-Based Quantitative Testing of Point of Sale/ATM (POS/ATM) Overdraft Disclosure Forms.” 25 This is consistent with the $34 median fee in 2012 at the 33 largest institutions in the market monitoring data used by the CFPB in its 2013 Study of Overdraft Programs. CONSUMER FIN. PROT. BUREAU, CFPB Study of Overdraft Programs at 52 (June 2013) http://files.consumerfinance.gov/f/201306_cfpb_ whitepaper_overdraft-practices.pdf. 26 For example, the student borrowers who benefit from lower prices because there are lower risk borrowers in the pool without the disclosure in yellow in Figure 1 or the strategic overdrafters face lower fees in the lower left quadrant in Figure 2. Disclaimer: The opinions expressed herein do not necessarily represent the views of Edgeworth Economics or any other Edgeworth consultant. This article is intended to inform readers about legal developments. Nothing in this article should be construed as legal advice or a legal opinion, and readers should not act upon the information contained in this article without seeking the advice of legal counsel. About Edgeworth Edgeworth Economics provides quantitative and economic consulting in the course of litigation and business to its clients, which include world-class law firms, Fortune 500 companies, and government agencies. Edgeworth experts apply their knowledge and experience, along with state-of-the-art computing infrastructure, to help clients efficiently manage complex issues including antitrust litigation, privacy & data security, transfer pricing, intellectual property, mergers and acquisitions, class actions, labor, and data & HR analytics. As a rapidly growing firm with a fresh approach, Edgeworth attracts leaders and teachers from across the industry including PhD economists, MBAs, statisticians, and programmers. Edgeworth has offices in Washington, DC, Pasadena, and San Francisco. www.edgewortheconomics.com