Marketplace Lending’s Rough Time in Court
The first and most directly applicable case is Madden v. Midland Funding LLC. In Madden the Second Circuit held that Midland Funding, a debt collection company that had purchased loans from a national bank, was not covered by the interest export rules found in the National Bank Act (NBA) and therefore not entitled to its preemption protection against state laws that cap interest rates. Under the NBA a nationally chartered bank is entitled to charge across the country the level of interest permitted in its home state, without regard for a particular state’s limits on interest. While there are some cases where non-bank entities have been able to take advantage of these provisions, they usually involve a case where the non-bank entity is acting as an agent or subsidiary of the bank or where denying preemption would interfere with the bank’s being able to exercise its powers under the NBA. In this case, the Second Circuit found that because Midland purchased and owned the loans outright, it was not operating on behalf of the bank. Applying state interest rate law (in this case New York’s) would not interfere with a national bank. Midland was not entitled to preemptive protection and was subject to New York law.
This matters because most marketplace lenders are not banks and do not originate the loans themselves. Instead they work with a national bank in a state without interest rate limits (such as WebBank in Utah) to arrange borrowers and to buy loans from the bank to service. These loans are either held on the marketplace lender’s books (where the lender gets the interest and may securitize fractional interests) or are bought outright by an institutional investor, many of which are not banks either. If non-bank loan buyers, including the marketplace lenders themselves, are not entitled to NBA interest rate preemption, then marketplace lenders and their non-national bank customers will have to comply with state laws limiting interest rates on a state by state basis and may be unable to profitably price risky loans.
If this decision stands and spreads (the Second Circuit, while incredibly influential, covers only New York, Connecticut and Vermont), it could undermine the business model used by many marketplace lenders. It should be noted that the court left open the possibility that non-bank lenders can get around this issue via contract, where the borrower agrees to have its loan governed by a specific state’s law, but that issue is still undecided. Alternatively, this decision may accelerate and expand the trend of marketplace lenders partnering more tightly with banks, so they could qualify as agents and enjoy NBA preemption. In the meantime Midland has asked for an en banc hearing, and it may end up before the Supreme Court. Its impact on marketplace lenders, while potentially great, is still unclear.
While Madden could be a very direct threat to marketplace lenders, a more subtle one may also be developing in the wake of the Supreme Court’s recent Texas Dept. of Housing and Community Affairs v. Inclusive Communities Project, Inc. In Inclusive Communities Project the Court found that the Fair Housing Act (FHA) creates a cause of action for disparate impact discrimination even though the statute does not explicitly say so. While the ruling does not directly affect most marketplace lenders, except possibly those that offer mortgages, it may have significant implications down the road.
Because the Equal Credit Opportunity Act (ECOA), which governs consumer credit, also lacks an explicit provision for a disparate impact cause of action, but Inclusive Communities Project indicates it very well might, which would mean marketplace lenders will have to avoid not only intentional discrimination but also making decisions that have the impact of discriminating. This opens the possibility of a new set of challenges for them.
To understand why, let’s first define disparate impact. Disparate impact occurs when decisions that are not in and of themselves discriminatory have a discriminatory effect. For example, while making credit decisions based on a prohibited criteria like race or sex (discriminatory intent) is an obvious violation of the law, a credit decision based on facially neutral criteria (like education level) that correlates to a prohibited category and results in credit being extended (or not) in a way that affects people based on the prohibited category can give rise to liability on the part of the lender.
Preventing disparate impact is challenging for lenders because unlike discriminatory intent, which is knowable before a lending decision is made, the existence of a disparate impact issue is unknowable until after the loans are made, and realistically it takes quite a few loans being made before there is enough statistical evidence to detect a disparate impact. Further complicating the picture for marketplace lenders is the fact that one of their major advantages – the use of algorithms tied to large and unique data sets – may make them more vulnerable, since they are personally responsible for the data they use (since they choose what to use and how to use it), rather than being able to fall back on a defense that they are just using a standard credit score. Additionally, for platforms that use machine learning, where the machine changes the data and algorithm to price risk better, there is a possibility that disparate impact can sneak in without humans noticing it and taking corrective actions. (Computers are not racist or sexist, but they also lack the context and sensitivity to anticipate issues with neutral data.)
Things are not all bad for market place lenders, however. Justice Kennedy’s having found a cause of action for disparate impact in the FHA indicates a similar cause of action could exist in the ECOA. But this is not completely certain and will require litigation under the ECOA to reach a final determination. Further, even if a cause of action does exist, the court provided a relatively narrow scope of what would constitute disparate impact. First, mere statistical differences alone are not sufficient to give rise to liability. Rather the plaintiff must be able to point to a specific policy or policies of the defendant that led to the disparate impact. Second, the court makes clear that defendants may use criteria that lead to a disparate impact if they are necessary to achieve a valid aim. A policy must be an “artificial, arbitrary, and unnecessary barrier[s]” to give rise to a claim. This may be a saving grace for lenders, since presumably the data they are using are tightly linked to their legitimate aim of proper risk pricing. There will likely need to be further litigation to determine what exactly that means.
Both Madden and Inclusive Communities Project present real challenges to marketplace lenders. Remaining to be seen is whether the industry, which prides itself on flexibility and innovation, can meet the challenge, or whether the cases constrain activity to the point where marketplace lending is no longer competitive. But given the industry’s growth and activity in traditionally underserved markets these questions merit the attention of industry, consumer advocates, and policy makers alike.