«PAYDAY LENDERS: HEROES OR VILLAINS? Adair Morse* Ross School of Business University of Michigan Abstract I study the effect that the availability of ...»
Draft December, 2006
Preliminary and Incomplete
PAYDAY LENDERS: HEROES OR VILLAINS?
Ross School of Business
University of Michigan
I study the effect that the availability of exceptionally high-interest consumer loans
(payday loans) has on individual welfare by using natural disasters as an exogenous
shock to communities’ financial condition. Utilizing a propensity score matched, triple
difference approach, I find that communities with payday lenders show greater resiliency to natural disasters. For each of the welfare measures considered – foreclosures, births, deaths, and alcohol and drug treatment, – the estimates suggest that payday lending enhances the welfare of communities. I discuss whether this effect is limited to individuals facing personal disasters or applies in general.
* I would like to thank Michael Barr, Alexander Dyck, Fred Feinberg, E. Han Kim, Amiyatosh Purnanandam, Amit Seru, Tyler Shumway, and Luigi Zingales for their helpful comments.
1 The 1906 San Francisco earthquake sparked hundreds of fires, leaving nearly 300,000 of the city’s 410,000 residents homeless. Leading the recovery was A. P. Giannini, a smalltime banker who profited by providing distress finance (sitting on a wharf with a bag of gold) to the citizens of San Francisco. He emerged as the heroic founder of the Bank of America. If Giannini could be considered a hero for offering distress finance, why are distress lenders today considered villains?
There is little debate that access to finance enhances value for firms. Financial institutions of different shapes and sizes have always played an integral role in corporate finance by affording such access. There is not, however, a similar consensus whether all types of financial institutions, such as high-interest rate lenders, provide a benefit to households. If individuals suffer from time-inconsistent preferences as in Laibson (1997), financial institutions will cater to this bias (Campbell, 2006), and access to finance can make them worse off.
In this paper, I study the welfare effects of access to finance offered by a particular financial institution, payday lending. Payday loans are short-term, small dollar advances that sustain individuals to the next payday. The fees charged in payday lending annualize to an implied rate of 400%. Do these 400% loans contribute to individuals’ resiliency to personal distress? I measure the outcome with foreclosures, death, drug and alcohol abuse and births.
With 20% of U.S. residents financially constrained,2 the importance of knowing the welfare implications of payday lending is likely to be both timely and large. Fifteen percent of U.S. residents have borrowed from payday lenders in a market that now provides $40 billion in loans each year (Bair, 2005; Fannie Mae, 2002).3 Despite (or because of) the growing demand, State and Federal authorities are working towards regulating and curbing the supply of payday lending. So far, fourteen States have banned payday lending outright, and most other States now regulate the fee structure and/or the process of revolving loans.
From one perspective, payday lenders should help distressed individuals to bridge financial shortfalls without incurring the greater expense of delinquency or default on 2 See Hall and Mishkin (1982), Hubbard and Judd (1986), Zeldes (1989), Jappelli (1990), Calem and Mester (1995), and Gross and Souleles (2002).
3 As a point of comparison, venture capitalists invested $21.7 billion in 2005 according to VentureXpert.
1 obligations. Acting in a vacuum of options for distress finance, payday loans should enable individuals to smooth liquidity shocks without incurring the larger costs of bouncing checks, paying late fees or facing service suspensions, evictions or foreclosures.
As such, one view of payday lending is that it should be welfare-enhancing.
An opposite, more prevalent, perspective is that payday lending destroys welfare.
The availability of cash from payday loans may tempt individuals to over-consume. An individual who is likely to fall to temptation would prefer that a self-control mechanism be set up before the temptation arises. In this case, if payday lending were banned, the temptation next period to over-consume with payday cash would be removed as per the models of Gul and Pesendorfer (2001; 2004) and O’Donoghue and Rabin (2006).4 In this view, payday lending can be welfare-destroying.5 To answer whether payday lending improves or destroys welfare, I use natural disasters as a community-level natural experiment. Natural disasters cause personal distress for at least some members of a community. The noise of having individuals in a community unaffected by the disaster biases my tests against finding any effects. I perform the analysis at the zip code level, focusing on the State of California during 1996-2005. I use positive and negative welfare measures to capture both the resiliency of lifestyle patterns during distress and the permanent consequences to distress. My positive welfare measure is births, and my negative measures are foreclosures, deaths, and alcohol and drug abuse.
The difficulty in measuring how payday lending impacts welfare is in disentangling the payday lending effect from correlated community economic circumstances that determine welfare outcomes. To overcome the endogeneities, I use a propensity score matched, triple difference (difference-in-difference-in-differences) framework. The role of the propensity score matching is to align communities on the 4 A large and growing literature documents the short-term patterns of consumption-savings behavior (e.g., Zeldes, 1989; Attanasio and Browning, 1995; Carroll, 1997; Thaler; 1994; Laibson, Repetto and Tobacman, 2005).
5 Consumer lobby groups argue that by expressly entrapping individuals who have little financial depth in a spiral of debt obligations, payday lenders permanently alter the well-being of borrowers (USAToday, August 31, 2006; Graves and Peterson (2005); Center for Responsible Lending (2004); Consumer Federation of America (2004); and Chin (2004)). The argument here is that welfare is viewed from an ex ante “long term” perspective as in O’Donoghue and Rabin (2006), not at any moment in time. Temptation consumption may of course bring high present utility at a given moment.
2 likelihood that residents are financially constrained prior to the natural experiment of disasters. I generate propensity scores by estimating the probability that an individual in the Survey of Consumer Finances (SCF) is financially constrained as a function of socioeconomic characteristics and then projecting the estimates onto Census socioeconomic data observed at the zip code level.
The role of the triple differencings is to overcome other possible endogeneities.
Differencing around natural disasters provides a set of two counterfactuals – the counterfactual of what a community hit by a disaster would have looked like if it did not have access to payday lending and the counterfactual of what a community with payday lending would have looked like if it had not been hit by a disaster. Natural disasters allow me to observe a situation in which demand for distress loans is not met by an endogenous supply. In addition, natural disasters provide an economically representative benchmark of what communities with payday lending that are subsequently hit by disasters would have looked like in the ex post period if a disaster had not happened.
Because my welfare variables are count variables, I estimate both a triple difference linear model and a triple interaction Poisson model. I measure welfare over two year windows before and after the disaster.
The results indicate that payday lenders offer a valuable service to communities by providing credit in a very incomplete market. Natural disasters induce an increase in foreclosures, but the existence of payday lenders significantly offsets this increase.
Communities with payday lenders are able to sustain their pre-disaster birth rates; other disaster-struck communities see a drop in births under the economic distress following disasters. Drug and alcohol treatments and deaths both fall in periods following disasters, consistent with prior research. The decrease in drug treatments is magnified in communities with payday lenders, and the drop in deaths only occurs for areas with access to payday lending.
Are banks substitutes for payday lenders? Because the role of financial institutions for natural disaster recovery is intrinsically important in its own right, I re-run all of the estimations using bank density in place of the existence of payday lenders. In only two of the sixteen specifications, do banks provide a valuable service of being a lender to individuals in distress in a similar way that payday lenders do. Finance is 3 valuable for community resiliency, but for the most part, the value does not come from mainstream banking.
Implications to my results must be put in context of the experimental design.
Individuals that use payday loans after natural disasters look just like people who use payday loans when faced with personal disasters, such a car breakdowns and health expenses. Since personal disasters happen all the time, my results can be extended to much of payday borrowing. However, I do not know exactly what ‘much’ means. There are likely to be payday borrowers who do not face disasters at all; rather they habitually over-consume and use payday cash to smooth cash cycles. Skias and Tobacman (2006) provide evidence consistent with the use of payday lending in such settings. Of course, the habitual over-consumers are those tempted by payday cash and thus those most likely to have negative welfare impacts. My results must be interpreted that payday lenders are providing a valuable service to communities, but do not speak to the net benefits distilling to those habitually falling to temptation. This suggests an agenda for future research which I discuss more fully in the conclusion.
The remainder of the paper proceeds as follows. Section I offers an overview of the market for payday loans. Section II develops the competing hypotheses of whether payday lending is welfare improving or diminishing. Section III outlines the triple differencing empirical methodology. Section IV describes the data sources and summary statistics. Section V presents the intermediate propensity score matching results, and Section VI presents the main empirical results showing the effect of payday stores on welfare. Section VII concludes.
I. Consumer Finance Institutions Consumers have a variety of options for their borrowing needs. The typical consumer predominantly borrows from three – banks, mortgage institutions and credit cards. A number of individuals, however, are restricted in their access to credit at these institutions and resort to borrowing from high interest lenders. These additional financial institutions are only sparsely studied in the finance literature, despite the fact that payday lending alone provides the economy with over $40 billion in loans per year.
4 The market for high-interest consumer loans divides generally into three segments. Credit cards provide the bulk of the liquidity for high-interest lending with rates up to 29.99%. (Of course, add-on fees charged by credit cards make the effective interest rates higher (e.g., Gabaix and Laibson, 2005; Massoud, Saunders and Scholnick, 2006)). The Supreme Court’s Marquette ruling of 1978 took away the de facto power of State usury laws, making it possible for credit cards to offer higher interest credit, but they have not as a general rule crossed the threshold of 30%.6 To our knowledge there is no decisive study of why this is so, but the threat of greater regulation undoubtedly concerns them (Knittel and Stango, 2003).7 From 30% to 400%, specialty markets offer collateralized loans (e.g., title lenders and pawn brokers), black market loans, and some new online instruments for the brave.
These markets are very narrow and do not provide many options for consumers constrained at their debt capacity or those with poor or no credit histories.8 As a result, it is unsurprising that prior research finds that approximately 20% of U.S. consumers are credit constrained (Hall and Mishkin, 1982; Hubbard and Judd, 1986; Zeldes, 1989;
Jappelli, 1990; Gross and Souleles, 2002).
All of this implies that for most individuals who have already maxed out their credit limits or who have poor or no prior credit history, the only option is to borrow at 400% APR from a payday lender.
How does payday lending work? An individual visits a payday loan store with a recent paycheck and checkbook. The typical loan given is approximately $300 with a fee of $50. In such a case, the borrower would write a check for $350, post-dating it to his payday 10-14 days hence. The payday lender verifies employment and bank information, but does not run a formal credit check. (Un-banked and unemployed individuals do not qualify for payday loans; thus, the notion that payday stores lend to the poor-of-the-poor 6 Marquette National Bank v. First of Omaha Services Corp. (1978) allowed credit cards to apply the State interest rate law of the corporate headquarters. Soon thereafter at the prompting of Citibank (Frontline, August 24, 2004), South Dakota lifted its usury ceiling, and all credit card companies could easily re-locate to South Dakota (or thereafter Delaware) and not be bound by usury limits.
7 Since the early history of the United States, 36% has been the cap that States have applied as the high limit on defining usury (USA Today, August 31, 2006). By charging less than 30%, credit cards may be avoiding the perception that they are approaching the height of usury.
8 For a more comprehensive survey of the market of pawn shops, title loans stores and informal lenders, see Caskey (1994; 2005), Bolton and Rosenthal (2005) and Barr (2004). Historically, usury laws’ prohibitions have forced high-interest lending into the black market.