The CFPB has issued a new report entitled “Online Payday Loan Payments,” summarizing data on returns of ACH payments made by bank customers to repay certain online payday loans.  The latest report is the third report issued by the CFPB in connection with its payday loan rulemaking.  (The previous reports were issued in April 2013 and March 2014.)  In prepared remarks on the report, CFPB Director Cordray promises to “consider this data further as we continue to prepare new regulations to address issues with small-dollar lending.”  The Bureau indicates that it still expects to issue its long-awaited proposed rule later this spring.

The Bureau’s press release cites three principal findings of the CFPB study.  According to the CFPB:

1. Half of online borrowers are charged an average of $185 in bank penalties.
2. One third of online borrowers hit with a bank penalty wind up losing their account.
3. Repeated debit attempts typically fail to collect money from the consumer.

While not referenced in the press release, the report includes a finding that the submission of multiple payment requests on the same day is a fairly common practice, with 18% of online payday payment requests occurring on the same day as another payment request.  (This can be due to a number of different factual scenarios: a lender splitting the amount due into separate payment requests, re-presenting a previously failed payment request at the same time as a regularly scheduled request, submitting payment requests for separate loans on the same day or submitting a payment request for a previously incurred fee on the same day as a request for a scheduled payment.)  The CFPB found that, when multiple payment requests are submitted on the same day, all payment requests succeed 76% of the time, all fail due to insufficient funds 21% of the time, and one payment fails and another one succeeds 3% of the time.  These assertions lead us to expect that the Bureau may propose new proposed restrictions on multiple same-day submissions of payment requests.

We anticipate that the Bureau will use its report and these findings to support tight restrictions on ACH re-submissions, perhaps tighter than the restrictions originally contemplated by the Bureau.  However, each of the findings trumpeted in the press release overstates the true severity of the issue.

The first finding disregards the fact that half of online borrowers did not experience a single bounced payment during the 18-month study period.  (The average penalties incurred by the entire cohort of payday loan borrowers therefore was $97 rather than $185.)  It also ignores another salient fact that is inconsistent with the negative impression created by the press release: 94% of the ACH attempts in the dataset were successful.  This statistic calls into question the need to require advance notice of the initial submission of a payment request, which is something that the CFPB previously announced its intention to do with respect to loans covered by its contemplated rule.

The second finding seems to attribute the account loss to the ACH practices of online lenders.  However, the CFPB report itself properly declines to ascribe a causal connection here.  According to the report: “There is the potential for a number of confounding factors that may explain differences across these groups in addition to any effect of online borrowing or failed payments.” (emphasis added)  Moreover, the report notes that the data merely suggests that “the loan played a role in the closure of the account, or that [the] payment attempt failed because the account was already headed towards closure, or both.” (emphasis added)  While the CFPB compares the rate at which banks closed the accounts of customers who bounced online ACH payments on payday loans (36%) with the rate at which they did so for customers who made ACH payments without problem (6%), it does not compare (or at least report on) the rate at which banks closed the accounts of customers with similar credit profiles to the rate at which they closed the accounts of customers who experienced a bounced ACH on an online payday loan.  The failure to do so is perplexing since the CFPB had access to the control data in the same dataset it used for the report.

The third finding is based on data suggesting that the first re-submission is unsuccessful 70% of the time and subsequent re-submissions are unsuccessful, in order, 73%, 83% and 85% of the time, respectively.  These numbers indicate, however, that an online lender prepared to re-submit three times to collect a payment might succeed in doing so nearly 58% of the time (1 – [.70 x .73 x .83]).  Each re-submission may be less likely than not to result in collection but a series of re-submissions is more likely than not to be successful.

Not only does the press release go beyond the actual findings of the study, the value of the study is limited by methodological issues associated with it.  The new report is based on consumer checking accounts obtained by the CFPB from a subset of several large depository institutions that offered deposit advance products during a sample period spanning 18 months in 2011 and 2012.  It covered borrowers who qualified for a deposit advance at some point during the study period and excluded all lenders known to have storefronts even if those lenders also made online payday loans.

The methodological problems associated with the study include the following:

1. The data is stale. The business model in widespread use by online lenders during the 2011-2012 sample period – four to five years ago – is no longer prevalent.  Online lenders have overwhelmingly transitioned to installment loan models where each payment is a fraction of the total balance due, instead of the single payment due at maturity model used previously.  If the CFPB had studied data related to the current online payday installment lending model, the return rate undoubtedly would have been much lower.  Moreover, re-submissions of the nature described in the paper are proscribed both by the current NACHA rules and the best practices guidelines of the Online Lenders Alliance, the trade group for online lenders.
2. The CFPB limited the borrowers included in the study to consumers who at some point during the study period qualified for deposit advances.  Even with this limitation, however, it nevertheless is likely that the consumers studied were disproportionately suffering from credit problems relative to online payday borrowers generally.  Otherwise, why would these borrowers obtain payday loans rather than deposit advances, which, before banks were forced by regulatory pressure to discontinue offering the deposit advance product, typically were made at interest rates far lower than those charged in connection with payday loans?  Moreover, the CFPB never explains why it used data from deposit advance banks rather than data from other banks that have provided account-level data to it in the past (for example, banks that provided information for the CFPB’s overdraft study) and it never addresses the confounding effect of this choice.
3. The report is not necessarily representative of borrower experience with lenders who have a storefront presence.  The collections model used by storefront lenders is markedly different than the one used by online lenders.  Storefront lenders rely upon personal contact with borrowers (not automated re-submissions of payment requests) and on encouraging borrowers to return to the store to make the loan payments in cash.

While the findings are open to question, we expect that the CFPB will assert that they support tightened restrictions on the collection of payday loan payments.  We also fear that the Bureau will assert that the report somehow rationalizes the adoption of other, more fundamental regulatory restrictions under the rule that it ultimately will be proposing “later this spring.”  As we have commented previously, the CFPB has not undertaken the cost-benefit analysis required for a proper finding of “unfair” or “abusive” conduct, as required to justify the type of broad-based and restrictive rulemaking it is contemplating.