Thus far, we have considered the nature of payday lending and the controversy surrounding it. Now it’s time to dive into the empirical literature. Studies of the affects of payday loans cover a variety of issues, largely driven by concerns of the day. Early on, critics expressed concern that payday lenders target financially vulnerable households, unsophisticated consumers, and alienated racial and ethnic minorities. Many scholars took up the challenge of examining empirically the demographic characteristics of the typical payday borrower. In Colorado, Chessin (2005) found on the basis of loan file data collected by lenders that the average age of payday borrowers was 35.8, and average monthly income was $2370. In a nationwide survey, Elliehausen and Lawrence found that 51.5% of payday borrowers had incomes between $25,000 and $50,000; while 23% had incomes below this range, 25.4% had incomes above it. A North Carolina study by Stegman and Faris (2003) concluded that income alone was a poor predictor of payday borrowing, but that blacks were much more likely to take out a payday loan. While some argue on the basis of correlation that payday lenders target blacks as a group, no conclusive evidence has yet been offered.
Of course, who borrows is not nearly as important as what effect borrowing has on those who do so. If these loans are welfare-enhancing for the consumers who use them, who uses them is of little policy relevance. If, on the other hand, payday loans are fundamentally toxic or destructive in a way analogous to Bar-Gill and Warren’s exploding toaster, any consumer is a potential victim and a worthy object of regulatory protection. Hence, much of the most recent literature on payday lending has focused on two claims made by the industry’s harshest critics: “Are payday loans predatory?” and “Do payday loans cause financial distress?” To these, I add a third: “What are reforms likely to accomplish?” Answering these questions from current research is the goal of this post.
Are payday loans predatory?
The claims of predation, as noted previously, tend to boil down to the price of the transaction. The ultra-high price results in a cycle of debt for borrowers and exorbitant profits for lenders, many critics believe. The cycle of debt argument is difficult to test, but scholars have approached it in several ways. Bar-Gill and Warren cite a finding from Chessin’s Colorado study that 65% of Colorado payday loan volume was generated by borrowers who took out 12 or more loans in the prior year. This finding is meant in part to support the proposition that payday lenders make the majority of their revenue from borrowers who use the product as a long-term credit instrument—for which it is, of course, ill-suited. But what this descriptive statistic tells us is that the largest proportion of loan revenue is generated by a category of borrowers who take out the largest proportion of loans—hardly policy-relevant in itself.
More importantly, the proportion of borrowers who took out 12 or more loans in the Colorado study was 33%. Flannery and Samolyk, in an FDIC study, found a lower proportion, 24-30%. One-quarter to one-third of customers using payday loans in this manner is troubling. However, 28.3% in the Chessin study took out 1-3 loans, and 22% took out 4-7 loans in the prior year. Thus, more than half of payday borrowers took out 7 or fewer loans in the prior year, a decidedly less troubling finding.
To address the predation claim, we must ask whether the interest rate or the repayment term drives repeat borrowing. Some studies have attempted to isolate the effects of these maligned design features. For example, Fusaro and Cirillo (2011) test the cycle of debt argument with a field experiment wherein they give a random sample of borrowers interest-free loans. They find “that the interest-free loan provided to a random group of borrowers has no effect on the repayment of the loan [and] the multi-loan rate is not reduced by the treatment.” Carter, Skiba, and Sydnor (2013), studying variation in repayment terms (from 7 to 20 days) in a Texas study, “find that initial loan durations have little impact on continuous borrowing” and that “there is almost no response to additional time to repay a loan.” In these two studies the researchers found that repeat or continuous borrowing was not driven by the interest rate or the repayment term respectively.
What these findings suggest is that, while there is no one kind of payday borrower, a substantial subset of these borrowers are high risk credit consumers. This is unsurprising. Payday lenders typically rely on services that screen out applicants who have previously defaulted on payday loans or have loans outstanding to other payday lenders, but they do not typically run credit checks, report to credit rating agencies, and have little recourse in the event of default other than denying the borrower access to their services in the future. Hence, while payday lenders seem to attract a higher proportion of high risk credit consumers, it seems to be the characteristics of the borrowers and not the design of the product which drives repeat borrowing. As I will discuss in my fourth and final post on the subject, this adverse selection may reflect more harshly on conventional credit products and their propensity to exacerbate detrimental credit consumption, than on the fringe credit market to which many financially distressed consumers resort.
Even if the price cannot be said to cause the cycle of debt, it still might reflect outrageous profit margins. Huge growth in the industry might suggest high profit margins and low entry costs, but that is not supported empirically. While it is difficult to get cost and profit data on many lenders because they are privately held, Huckstep (2006) analyzed seven publicly traded payday lenders–which constituted 36% of nationwide payday lenders, finding that “when compared to many other well-known lending institutions, payday lenders may fall far short in terms of profitability.” This study found that “For pure payday lenders, the average profit margin was 3.57%.” By contrast, Huckstep found that the average profit margin for Starbuck’s (included because of its similar hours, location, and staffing requirements) was 9% and commercial lenders averaged about 13%.
As for the costs of entry, Flannery & Samolyk (2005) found that “new stores generate negative or low profits for a few years before becoming fully profitable.” What drives these costs? Huckstep explains that
Fixed costs are increased by maintaining longer hours of operation relative to those of typical financial institutions. Since the Flannery & Samolyk study indicates that customers are not shopping based on price, but rather on convenience, the added costs of longer hours and full-time employees are necessary to obtain, serve, and retain their customers.
These findings contradict the notion that payday lenders make exorbitant profits by exploiting their customers. Rather, the business model seems to conform to many consumers’ preferences. These findings also lend support to an industry claim that a typical store cannot operate “with fees below approximately $15 per hundred borrowed.” Moreover, Flannery and Samolyk found that “Low profitability has driven many mainstream financial institutions from the market for small short-term credit (except through credit cards).” Banks and credit unions which do provide short term credit tend to do so in the form of overdraft protection with similar implied interest rates. Pawn and title operations can charge substantially less, but must rely on collateral. Part of the explanation for such a finding lay in the economics of originating small loans. As Lehman (2006) puts it:
The manpower and paper processing costs of issuing a $200, 30-day small loan are nearly identical to issuing a $5000, unsecured 24-month loan. Whereas the former offers comparatively little revenue at the typical bank rate of interest, the latter offers significantly greater revenues relative to cost. Most conventional financial institutions, therefore, have chosen to avoid the small-loan market in favor of larger more profitable loans.
To be clear, interest rates on payday loans are very high relative to conventional credit products. But absent a cycle of debt or outrageous profits, the predation claim loses force.
Finally, claims of predation are often made in tandem with the claim of deception. However, the straightforward nature of a payday lending transaction, described in the first post of this series, makes the claim of deception difficult to countenance. As Stoianovici and Maloney put it “It is hard to make a principled argument that the customer is deceived in a payday lending contract because it is very simple in terms of the costs and structure; there are no hidden fees.” Again, the predation claim seems to boil down to the price. Flannery &Samolyk (2005) offer that: “It seems likely that a substantial portion of this criticism reflects a feeling of outrage over the high interest rates implied by payday loan fees.” Such outrage, without more, does not make these loan instruments predatory in any meaningful sense.
Do payday loans cause financial distress?
Often the predation argument is simply another way of stating the belief that payday loans cause financial distress. Both defining and measuring financial distress has proven more difficult than one may initially suppose. Establishing causation is an even more difficult proposition. Some scholars have looked to risk of bankruptcy and other measures of credit health. Skiba and Tobacman (2009) used regression analysis to causally link payday loans to increased risk of bankruptcy. This study was important because it was the first empirical work to find such a causal link to financial distress and critics continue to rely on it heavily. However, Hawkins (2011) and Caskey (2010) exposed several methodological weaknesses in this study, considerably undermining its conclusions. Moreover, Skiba and Tobacman have recently begun circulating a paper coauthored with Neil Bhutta of the Federal Reserve Board in which they “show that the long-run effect of payday borrowing on credit scores and other measures of financial well-being is close to zero.” In an extensive literature review in 2012, Skiba found that “there is little evidence that payday loans per se are unequivocally bad for borrowers or that consumers overall are better off without access to payday loans.” Yet radical critics have relied on the earlier study by Skiba and Tobacman for the opposite proposition.
For reasons already noted, a substantial subset of payday borrowers are high risk credit consumers, some of whom may be experiencing financial distress quite apart from their payday loan experience. As Caskey (2010) put it, the “big question” from a public policy perspective is “Do payday lenders, on net, exacerbate or relieve customers’ financial difficulties?” Zinman’s (2010) approach was to compare the change in payday customers’ responses in Oregon and Washington before and after Oregon’s 10% finance fee cap. Customers in both states were asked about access to short-term credit, perception of present financial situation, and future expectations. Washington maintained a fairly liberal policy toward lenders over the same period. Zinman found based on a difference-in-difference model that Oregon respondents reported much greater negative change on all three counts relative to Washington respondents, which he attributes to Oregon’s cap and the subsequent flight of payday lenders from the state.
Morgan and Strain (2009) studied the effects of changes in legal access to payday loans from 1998-2008 across all 50 states. They found that restrictive regulation of payday lending was welfare reducing for consumers on the whole. For example,
Georgians and North Carolinians do not seem better off since their states outlawed payday credit: they have bounced more checks, complained more about lenders and debt collectors, and have filed for Chapter 7 (“no asset”) bankruptcy at a higher rate. The increase in bounced checks represents a potentially huge transfer from depositors to banks and credit unions. Banning payday loans did not save Georgian households $154 million per year, as the CRL projected, it cost them millions per year in returned check fees.
Morgan and Strain use complaints to the Federal Trade Commission (FTC) as a measure of consumer welfare because the agency is charged with enforcing federal laws against abusive collection practices, which are associated with “informal bankruptcy”—i.e., consumer credit defaults outside of formal bankruptcy protection. The concept of informal bankruptcy is helpful given the potentially underinclusive nature of formal bankruptcy as a measure of financial distress. The FTC maintains a hotline by which consumers can report abusive practices by their creditors. These complaints are an indicator of financial distress because they signal that “households are sufficiently bothered to appeal to the government for protection.”
Other scholars have tested the effects of payday loans on financial survivability. Morse (2009), in a California study, tests the effects of access to payday loans on consumers’ ability to weather exogenous shocks. Specifically, she uses ZIP code level data to measure rates of foreclosure and crime in communities with and without access to payday lending. She concludes that “in times of distress, access to credit mitigates 1.22 foreclosures per 1,000 homes and discourages 2.67 larcenies per 1,000 households. The implication is that access to finance can be welfare improving at 400% APR.” Morse’s finding confirms another by Wilson et al (2010) in an experimental setting. Therein, live subjects were given budget constraints, exposed to a variety of expenditure shocks, and given access to various forms of credit in a computer simulated environment designed to approximate conditions faced by typical payday borrowers. The researchers found that, while consumption of payday loans beyond a certain threshold could be financially detrimental, “the existence of payday loans, all else fixed, increases the probability of financial survival by 31%.”
Each of these studies has limitations, but the general weight of the findings should give pause to radical critics who seek restrictive regulation of payday loans. If consumer welfare is the goal, policy makers should look to more than anecdotal tales of abuse or gut reactions to the implied APRs—salient though these factors might be—to determine what rules should apply to this fringe consumer credit product. As we shall see, some of the more moderate reforms may help the subset of consumers who use payday loans to excess or as long-term credit solutions where better long-term options exist. But as I shall argue in the remainder of this post and the next, more heavy-handed approaches, though well-intentioned, are misguided.
What the reforms are likely to do
As previously noted, the most drastic reforms would alter payday lending out of existence. As the evidence presented above has demonstrated, doing away with this maligned business model might ultimately prove welfare-reducing. Therefore, bans or effective bans are likely to help the subset of payday borrowers who use this product to their own detriment, while denying a beneficial credit option to consumers who use the product appropriately. Moreover, no public policy discussion should neglect the effect of such a ban on the interests of entrepreneurs (and those they employ) who would suffer unnecessarily the foreclosure of potentially lucrative and socially beneficial business opportunities.
Moreover, even some less radical reforms could have unfortunate unintended consequences for consumers. The effects of moderate price caps is one example. As noted previously, the cost of administering small loans can be quite high relative to the size of the loan, but the market for payday loans is also quite competitive, requiring lenders to maintain rather modest profit margins. A study prepared by researchers at the Federal Reserve Bank of Kansas analyzing the effects of price regulation on payday finance charges in Colorado between 2000 and 2006 showed some perverse consequences:
For example, early in the sample when finance charges were still loosely distributed below the price ceiling, loans originated in markets with a large number of payday firms carried lower prices. But later in the sample these competitive effects disappear and evidence consistent with strategic price behavior emerges…Lenders also begin charging higher prices at stores in largely minority neighborhoods and at stores located near military bases, arguably local markets in which demand for payday loans should be relatively inelastic.
Payday lenders in Colorado over the period of the study began behaving less like a competitive industry and more like a cartel with the regulatory cap serving as a focal point for convergence. This convergence meant that many lenders gravitated upwards over time toward the price ceiling, especially in areas where consumers appeared to have the fewest alternatives. As this example illustrates, policy makers should exercise caution in designing reforms and carefully monitor effects.
Another possible reform would be to increase disclosure requirements. This is ironic given the very straightforward nature of payday lending transactions. The Truth in Lending Act (TILA) now requires payday lenders to prominently display the implied interest rate on loans. This is meant to ensure that consumers know what they are paying on payday loans relative to more conventional credit products. However, reformers should be aware that there may be non-price features of payday loans that might induce a customer to choose a payday loan even when lower cost alternatives are available. Quick access to cash at hours which accommodate busy work schedules and the physical proximity afforded by high-density store-front locations are among these features. But the autonomy retained in such transactions is another important feature. One very recent report found that Army personnel sometimes circumvent payday loan bans and shun free financial services offered through the military because with military relief “there are a lot of strings attached.” Each day consumers at every level of disposable income pay a premium for services they value. There is no reason to conclude that payday loans are very different in this regard or that payday borrowers are severely uninformed.
Moreover, it bears notice that TILA was previously amended to require credit card companies to disclose the dollar amount implied by the APR. Prior to the amendment, consumer advocates were alarmed at the very low proportion of credit card consumers who knew what they were being charged or would ultimately owe on their credit cards. Conversely, a Federal Reserve study showed that “about 95 percent of the respondents gave an accurate report of the finance fee [on their payday loan] in dollar terms,” while a CreditCard.Com analysis revealed that only 20 percent of adults could understand their credit card agreement.“Accordingly,” says Paige Skiba, “additional information (beyond the Truth in Lending Act) would not be especially helpful to consumers.”
There are, however, other attempts to curb excessive use. While Carter, Skiba, and Sydnor (2013) find little effect on rollovers from moderately extending loan terms,maximum limits on rollovers beyond some (generous) threshold might have promise. Several states have implemented rollover limits, but, given the ability of consumers to alternate their patronage, monitoring consumer behavior could prove difficult absent a statewide database such as one finds in Florida and Oklahoma. In addition, it remains a possibility that customers determined to circumvent the borrowing limit can resort to costly substitutes not covered by the database, and many consumers may balk at the idea of the government monitoring their financial transactions so closely. Time and careful study will tell whether such efforts can prevent the subset of consumers who misuse payday loans from using such credit to their own detriment. Of course, the type of study recommended here is only possible if states retain the ability to regulate their consumer credit markets with a fair degree of independence. Uniform regulation from the new Consumer Financial Protection Bureau could impede discovery of best practices and the potential for individual state-level error could become nationwide error if (when?) regulators “get it wrong.”
Continue to Part IV here.
Chessin, P. (2005).
Elliehausen, G. & Lawrence, E.C. (2001), supra note 14. There are some important weaknesses in both sets of income data that should be noted. With respect to Chessin’s loan file data, Caskey (2010) cautions that loan files of payday lenders do not include income for other adults in the household. Rates of marriage and co-habitation appear similar for payday borrowers as the general population. Therefore, Chessin could be underestimating the household income of payday borrowers. On the other hand, with respect to Elliehausen and Lawrence’s survey data, Caskey notes that survey data absent documentation can be unreliable. Given the propensity of survey respondents to over-report their incomes, Elliehausen and Lawrence could be overestimating household income for payday borrowers.
Stegman, M.A. &Faris, R. (2003).
Bar-Gill, O., Warren, E. (2008), p.44.
Flannery, M., &Samolyk, K. (2005).Payday Lending: Do the Costs Justify the Price? Arlington, VA: Federal Deposit Insurance Corp. Center for Financial Research.
Chessin, P. (2005), pp. 412-13.
 Carter, Skiba, and Sydnor (2013) examined an administrative dataset from a payday lender in Texas from 2001-2004 containing information on over 1 million loans. They found that “The average borrower in the dataset took out six payday loans per year” and that 60% of borrowers “who obtain a new loan borrow at least one more time (i.e., an effective rollover) in succession” (p.3). They also find that a small subset of individuals (4%) rolled over their loans 10 times (p.11). Thus, a subset of borrowers clearly use this financial instrument inappropriately. The Difference a Day Makes: Measuring the Impact of Payday Loan Length on Probability of Repayment, Working Paper (2013) available: http://www.law.northwestern.edu/colloquium/law_economics/documents/Skiba%20Payday.pdf. This finding is somewhat mitigated, as Bar-Gill and Warren concede: “The cost of rolling over an existing loan is, however, substantially lower.” Bar-Gill, O., Warren, E. (2008).
Fusaro, M.A. & Cirillo, P.J. (2011), p. 27.
 Carter, Skiba, Sydnor (2013), pp. 4-5.
 Li, M., Mumford, K.J., & Tobias, J.L. (2012) studied a 2006 dataset from an online payday lender with customers in 38 states; they found that 700 of 2500 loans resulted in default, and that 55% of these were immediate defaults, i.e., the borrower did not service the debt at all. “A Bayesian Analysis of Payday Loans and their Regulation” Journal of Econometrics 171 (2): 205-216. This extreme example may be peculiar to purely online lenders (which constitute only 16% of the industry), but it is ironic given that this category of payday lender is the most maligned by critics.
Huckstep, A. (2006). Payday Lending: Do Outrageous Prices Necessarily Mean Outrageous Costs? Fordham Journal of Corporate & Financial Law, Vol. XII, 203-232, p. 227. “When including pawn operators, this figure more than doubles to 7.63%.” id.
Flannery, M., &Samolyk, K. (2005), pp. 7-8. The authors collected store-level data from a random sample of 600 stores for the years 2002, 2003, and 2004.
Huckstep, A. (2006), p. 222.
Flannery, M., &Samolyk, K. (2005), p. 4.
Melzer, B.T. & Morgan, D.P. (2012).
 Lehman, T. (2006), p. 10.
Stoianovici, P.S. & Maloney, M.T. (2008), p.4.
Flannery, M., &Samolyk, K. (2005), p.4.
 Hawkins, J. (2011) p.1367 provides a litany of examples from public policy research that invokes financial distress in substantive ways without defining the term.
Skiba, P. M., &Tobacman, J. (2009). Skiba and Tobacman widely circulated their article as a working paper beginning as early as 2007 prior to publication as a law review in 2009.
Hawkins, J. (2011), supra 21. See also Caskey, J.P. (2010). Working Paper No. 10-32, Payday Lending: New Research and The Big Question, Research Department, Federal Reserve Bank of Philadelphia.
Bhutta, N., Skiba, P., Tobacman, J. (2013) “Payday Loan Choices and Consequences” Vanderbilt University Law School, Law & Economics Working Paper 12-30.
Caskey, J.P. (2010).
Zinman, J. (2010). Working Paper No. 08-32, Restricting Consumer Credit Access: Household Survey Evidence on Effects Around the Oregon Rate Cap. Research Department, Federal Reserve Bank of Philadelphia.
 Morgan, D.P. & Strain, M.R. (2008). Payday Holiday: How Households Fare after Payday Credit Bans. Federal Reserve Bank of New York Staff Reports, p. 26.
Id at 16.
 Morse, A. (2009). Payday Lenders: Heroes or Villains? p. 25, available at http://papers.ssrn.com/sol3/papers.cfm?abstract_id=1344397.
Wilson et al. (2010) at 15. The researchers found that taking out more than 10 payday loans over the course of the 30-month period covered by the experiment could introduce financial costs that exceed the consumption-smoothing benefits of this financial product.
 Many reform advocates consider extension of repayment terms and a 36% interest rate cap synonymous with a ban on payday lending. See supra 38 and 41. See also a press release by U.S. Rep. Gabrielle Giffords announcing the introduction of a bill to cap interest rates at 36% nationwide: http://www.votesmart.org/public-statement/526382/usrep-gabrielle-giffords-acts-to-ban-payday-lending-nationwide.
 The top six payday lenders account for only 20% of the market, indicating market diffusion. Moreover, payday lenders compete with banks and credit unions which offer overdraft protection at similar implied interest rates in many states and, of course, pawn brokers.
Deyoung, R., Phillips, R.J. (2009) The Federal Reserve Bank of Kansas City Economic Research Department. Research Working Paper, available at http://kansascityfed.org/PUBLICAT/RESWKPAP/PDF/rwp09-07.pdf.
See Truth In Lending Act, Pub. L. No. 90-321, § 121, 82 Stat. 146, 152 (1968) (codified at 15 U.S.C. § 1631 (2006)).
 American Public Media’s Marketplace, http://www.marketplace.org/topics/wealth-poverty/beyond-payday-loans/why-military-personnel-fall-prey-payday-lenders. One example of “strings” came from an Army couple who had previously secured a no-interest loan from Army Emergency Relief: “’They helped pay a month’s worth of bills. They gave us one big, giant check for the groceries that you could only spend at the commissary. But it was only one check. You couldn’t get change back or nothing,’ adds his wife, Lisa. ‘I’m like I can’t have milk for a month in the refrigerator and bread for a month.’”
 See the Federal Reserve’s Report to Congress explaining “Finance Charges for Consumer Credit under the Truth in Lending Act”: The TILA’s principal cost disclosure is the ‘finance charge,’ defined as the cost of consumer credit expressed as a dollar amount…. The TILA and Regulation Z require creditors to disclose the cost of consumer credit as an annual percentage rate (APR), in addition to disclosing finance charges as a dollar amount. (Board of Governors of the Federal Reserve System, 1996) (Italics mine).
Edmiston, K. D. (2011, January). Could Restrictions on Payday Lending Hurt Consumers? Federal Reserve Bank of Kansas: Economic Review, pp. 63-93 at 65. Caskey (2010) cites a 2007 California Department of Corporations payday loan study which found that “92 percent of the respondents said that they were aware of the fees on their loans before taking them out.”
Prater, C. (2010, July 22).U.S. credit card agreements unreadable to 4 out of 5 adults. CreditCard.Com Website: http://www.creditcards.com/credit-card-news/credit-card-agreement-readability-1282.php.
Skiba, P.M., Regulation of Payday Loans: Misguided?, 69 Wash. & Lee L. Rev. 1023 (2012), p. 1044.
Carter, Skiba, Sydnor (2013).