Comments on Marketplace Lending and its Evolution: Blue Elephant's response to the Department of Treasury's request for Expanding Access to Credit Through Online Marketplace Lending. Blue Elephant Capital Management has been one of the early fund movers in the marketplace lending arena. We are not an underwriter of loans. Instead, we look out over fintech and traditional lending markets for places to deploy capital to maximize returns for our investors. Our perspective is unique in that we are not married to any particular model or lending type. We use our extensive capital markets experience to make our decisions.
While our firm is only two and half years old, we do have interesting and relevant backgrounds. Brian Weinstein spent 17 years at Blackrock, overseeing $300bn of active fixed-income assets before leaving. JP Marra spent 22 years on the buy-side at Lehman, Bank of America and Nomura, where he ran treasury, mortgage, agency and funding businesses. Ashees Jain spent a decade trading structured product, hybrids and derivatives at Lehman, Barclays and Nomura.
During our time in marketplace lending, we’ve noticed positive breakthroughs along with activity that worries us. We’ve heard arguments from parties including investors, lenders, banks and lawyers. Below, we outline the positive developments, the trends that worry us, a few areas that could use clarification, and standards that the Treasury should push for.
- Non-bank lending is necessary to ensure borrower access and rate competition in the age of the large bank. Whenever we look at a potential investment in fintech space, we ask the question, “What problem does this solve?” With prime consumer lending and secured small business lending, the answer is relatively obvious. A few decades ago, those needing to borrow money would walk into a local bank. Often times, their families, businesses and credit histories were known personally by the banker or easily tracked down through the community. A combination of technology, regulation and time has significantly changed that equation. Today, there are a smaller (and declining) number of community banks, with those institutions having been replaced by large banks. There are a number of problems with the large bank model, of which I will highlight two. First, large banks by definition will have plenty of broad consumer exposure through credit card, mortgage and other primary banking functions. Second, the costs for these banks to underwrite small balance loans are very similar to the cost of underwriting very large loans. Put these together and there is little doubt why there is a lending problem in the first place – no large institution is incentivized to make small loans at any reasonable interest rate. Unless we are about to embark on the rebirth of the community bank, data-driven non-bank lenders make a lot of sense.
- The marketplace lending model promotes diversification of risk. This is related to the large bank problem. If a bank already owns the mortgage and credit card risk of a borrower, it makes little sense for them to extend additional loans for any purpose. Any downturn in real estate or the consumer causes a significant hit to earnings (i.e. the financial crisis of 2008). Non-bank lenders need to make sure that they have a diversity of funding sources as part of their business model. At any time, there are different combinations of retail, dedicated funds (such as Blue Elephant’s Consumer Loan Fund), pension funds, endowments, large asset managers and banks committing capital to marketplace lenders. This mix will shift over time in response to opportunities in the broad capital markets. This avoids the situation where four or five large institutions are stuck holding an ever-growing basket of correlated risk.
- The combination of finance and technology, done correctly, should be a symbiotic relationship. From a borrower’s standpoint, technology allows credit decisions to be made more quickly than pure manual underwriting. On the lender’s side, there is more data transparency on the underlying credit of a potential investment. This does not replace the importance of finance – the speed to market must be limited by the quality of the underwriting. In a bigger picture sense, there are many places where lending data either does not exist, or has not been broadly distributed. As time goes on, open source lending will allow for smarter models, deeper pools of available capital and more efficient capital markets.
- Diversification of risk is not the same as the elimination of risk. Fintech has not solved the lending problem. There will always be default cycles and there will always be lenders treading in waters that are treacherous. At the moment, we see growth in a few cohorts that worry us: subprime consumer and unsecured small business lending, to name two. There simply isn’t enough data to predict how these loans will perform, especially if they are made in high volumes for the first time. Too often, in our view, all of the fintech lenders get lumped together into one category. Instead, there should be some clarification over the methodologies used to make loans and how much risk is inherent with each model.
- Complete lack of industry standards. With so many fintech lenders joining the marketplace, there is a notable lack of cohesiveness. What is expected in terms of representations and warranties? Should there be universal servicing standards? From the finance side, we think that all-in lending rates should be disclosed – there are currently a few places that make it very difficult for a borrower to know their true interest rate. If nothing else, the handful of lenders that are doing significant volume should get in a room and attempt to hammer out some of these issues. There have already been a number of securitizations (we’ve done one) – but each one has slightly different language around a number of these issues. A lot could be gained by standardizing these major practices.
- Broad regulatory questions remain unanswered. Does the “rent-a-charter” model being used to apply national lending laws make sense? Many question whether or not a lender should be forced to have “skin-in-the-game” – we think this is the wrong way to look at the world, but a decision would be helpful sooner than later. As time passes and more capital gets committed to marketplace lending, a failure to address these issues will become more and more disruptive.
Stated simply, banks are no longer willing or able to fill the borrowing needs of US consumers and small/medium sized businesses. While we shouldn’t oversell the “fintech revolution,” it is important not to dismiss it as a fad. Much like Uber and Airbnb are allowing those with cars and homes to turn their capital into assets with cash flow, marketplace lending is allowing those with capital to step in for banks and earn a return on their cash that was generally unavailable before. The capital markets are sending the clear message that banks are failing in their traditional lender role and that this gap is screaming to be filled. We must be careful to allow the markets to function within the rule of law and logical oversight – we all know what happens when markets are overregulated, but also what happens when left completely alone for too long.
Below, we address the RFI questions directly.
- There are many different models for online marketplace lending including platform lenders (also referred to as “peer-to-peer”), balance sheet lenders, and bank-affiliated lenders. In what ways should policymakers be thinking about market segmentation; and in what ways do different models raise different policy or regulatory concerns?
Risk diversification is important when it comes to the lending markets. It is easy to claim that the balance sheet model, where the originator holds the risk of loss themselves, is superior to other models. However, anyone who lived through 2008 cannot make that claim with a straight face. Our largest and most important banking institutions held plenty of loans, both commercial and consumer based, that should never have been underwritten. Platform lending is unique in that the data used to underwrite the lending is made available to a wide audience. Using that data, lenders decide where the best risk/reward is for their capital. The end result is that the risk is sold to a wide variety of investors. To say that a platform lender has no “skin-in-the-game” and is therefore reckless is unfair and inaccurate – if the lender fails to make good loans, they will cease to exist as a business, just like a balance sheet lender. The danger comes into play when a platform lender is using questionable data or relying on data that is questionably correlated to their style of lending. We see this most often in areas where lending has not really been done before in large size, specifically subprime consumer, unsecured small business lending and merchant cash advance. Regulators should consider making distinctions between lenders with significant datasets and those in uncharted waters.
- According to a survey by the National Small Business Association, 85 percent of small businesses purchase supplies online, 83 percent manage bank accounts online, 82 percent maintain their own Web site, 72 percent pay bills online, and 41 percent use tablets for their businesses.Small businesses are also increasingly using online bookkeeping and operations management tools. As such, there is now an unprecedented amount of online data available on the activities of these small businesses. What role are electronic data sources playing in enabling marketplace lending? For instance, how do they affect traditionally manual processes or evaluation of identity, fraud, and credit risk for lenders? Are there new opportunities or risks arising from these data-based processes relative to those used in traditional lending?
Technology is playing a central role in marketplace lending. The fact that a good amount of information is available electronically allows for important levels of screening to be done almost instantaneously. We’ve made thousands of loans and have had only one instance of pure fraud that we are aware of to date. Technology allows marketplace lenders to electronically confirm that the borrower is whom they claim via micro-deposits, IP address location checks and a number of other more traditional driver’s license and pay stub checks. That said, technology does not “solve” the lending problem – there will still be business cycle downturns, lenders who offer loans to the wrong types of borrowers, and investors who lose money. In many cases, loans are made based on stated, instead, of confirmed data. In newer markets where lending has not traditionally happened in large size (subprime consumer, unsecured small business, merchant cash advance), there is limited data available to properly screen the borrower. While it is great that technology can help screen for fraud, it is less clear to us that it can significantly lower credit risk in areas where lending data does not exist. There will be tremendous opportunities to build these datasets and expand the horizons of lending, but not without significant risk of loss.
- How are online marketplace lenders designing their business models and products for different borrower segments, such as:
- Small business and consumer borrowers;
- Subprime borrowers;
- Borrowers who are “unscoreable” or have no or thin files;
Depending on borrower needs (e.g., new small businesses, mature small businesses, consumers seeking to consolidate existing debt, consumers seeking to take out new credit) and other segmentations?
The first movers in the marketplace lending space generally attacked markets where plenty of data was available. The most obvious was in the debt consolidation space where the value proposition was clear and potential volumes large. With Prosper/Lending Club having a 4% cost advantage over the banks and access to the data used to underwrite traditional credit cards, their business model is actual very easy to understand. A consumer with 20% credit card debt who has been paying it back regularly is even more likely to be able to pay back a 16% marketplace loan. The problem is that there are a limited number of platforms that can compete in this space before it becomes saturated.
On the small business side, we’ve seen platforms attack a number of segments. The most robust models use a combination of manual and electronic underwriting to make larger balance ($25k-$500k) loans of one to five years to mature businesses. These loans require the borrower to provide three years of actual tax returns along with business collateral and a personal guarantee. At the other extreme, there are platforms that are willing to give out short-term loans to businesses using only stated data and with no personal guarantee.
Newer entrants have entered market segments where lending has traditionally not taken place, such as thin file, subprime consumer and unsecured small business markets. There is the potential for significant growth in these underserved markets, but it will take a great deal of time until the dataset makes the risks quantifiable in any meaningful way. Platforms operating in these less developed spaces should not be considered the same as marketplace lenders operating in more established lending markets, even though they often try to make themselves look similar.
- Is marketplace lending expanding access to credit to historically underserved market segments?
Yes, for better or for worse. For anyone starting a lending platform today, it makes the most sense to focus where lending has traditionally not been done. There is much more potential for growth lending to subprime consumers, since very few lenders currently make such loans. It would be much harder to attack the prime borrower where banks, credit card companies and marketplace lenders already compete. The problem with these “underserved” segments is that it will take a few business cycles for the models to be strong enough to truly become investable. In the meantime, we worry about those who are the first to invest in their debt.
- Describe the customer acquisition process for online marketplace lenders. What kinds of marketing channels are used to reach new customers? What kinds of partnerships do online marketplace lenders have with traditional financial institutions, community development financial institutions (CDFIs), or other types of businesses to reach new customers?
Marketplace lenders use a wide variety of customer acquisition channels. The most reliable on the consumer lending side is old-fashioned direct mail. Digital media is a growing outlet for finding customers, while some platforms use brokers to find business. More broadly, we’ll leave this question to the platforms themselves to answer.
- How are borrowers assessed for their creditworthiness and repayment ability? How accurate are these models in predicting credit risk? How does the assessment of small business borrowers differ from consumer borrowers? Does the borrower's stated use of proceeds affect underwriting for the loan?
There are many factors that go into credit assessment. The most predictive inputs are FICO and debt-to-income (DTI) ratios. While these are not perfect indicators, there are plenty of data points from traditional credit card lending to predict how borrowers with FICOs above 660 behave. As lenders move lower in FICO, it gets more and more difficult to do traditional credit assessment. The simple fact is that the only way these models can get better is to make mistakes and learn from future default cycles.
We see too many small business loans made based on consumer loan metrics. Since these do not predict the success of the business, it is unlikely that the consumer metrics will be predictive of the loan’s performance. Quality small business underwriting requires companies with multiple years of history (greater than 10 is ideal) with meaningful and consistent revenues. A great deal of manual underwriting is necessary to ensure the validity of the borrower and their business. Technology helps with the initial screen and the speed to market, but the best lending in small business space retains many of the positive attributes of traditional bank underwriting.
The stated use of proceeds does help a little, but we are skeptical of relying on this since it is impossible to verify.
- Describe whether and how marketplace lending relies on services or relationships provided by traditional lending institutions or insured depository institutions. What steps have been taken toward regulatory compliance with the new lending model by the various industry participants throughout the lending process? What issues are raised with online marketplace lending across state lines?
Marketplace lenders have tried to be very open with regulators as to their methodologies in the hopes of avoiding regulatory compliance issues down the road. We would love to see regulatory clarity in regards to the Madden v. Midland case – the uncertainty will slow down the flow of credit through the system.
- Describe how marketplace lenders manage operational practices such as loan servicing, fraud detection, credit reporting, and collections. How are these practices handled differently than by traditional lending institutions? What, if anything, do marketplace lenders outsource to third party service providers? Are there provisions for back-up services?
The biggest operational difference between traditional and marketplace lending is on the loan servicing side. Most marketplace lenders are not built to handle large servicing needs. As such, a secondary servicer often handles collections out past 30 or 60 days. Away from the best-capitalized lenders, we do worry that servicing is often overlooked. As such, we have a provision for backup servicing for loans that we make through the platforms. We do not want to rely on someone else’s arrangements for something as important as servicing in a stressed scenario.
Beyond servicing, most marketplace lenders partner with various credit bureau and technology providers on the credit reporting and fraud detection side of the business.
- What roles, if any, can the federal government play to facilitate positive innovation in lending, such as making it easier for borrowers to share their own government-held data with lenders? What are the competitive advantages and, if any, disadvantages for non-banks and banks to participate in and grow in this market segment? How can policymakers address any disadvantages for each? How might changes in the credit environment affect online marketplace lenders?
The federal government should foster innovation in marketplace lending, with the caveat that lending is, was and always will be a cyclical endeavor. All the data in the world will not change this, since every default cycle is different and economies change over time. Those investing in lending markets will incur losses at some point in a business cycle. Regulation should seek to clarify the lending rules and protect borrowers from predatory lending.
All lending platforms should be required to show a stress test scenario, detailing what economic assumptions were made along with the impact it would have on their portfolio. There are two reasons we feel this is important. First, many platforms throw around the “big data” terminology even though their models are not always robust. Providing a stress test would force lenders to enter into a discussion around how much data is truly available in their lending segment and where there might be weak spots. Second, it would remind investors that losses do happen and that they should invest accordingly.
At the moment, marketplace lenders have lower brick and mortar costs along with fewer regulatory issues than banks. We do not think policymakers should directly address this. Instead, they should ensure that data continues to be freely available for investors to be able to make informed investment decisions. As the credit environment changes, the competitive advantage could easily swing back to the banks. Having traditional banks and marketplace lenders competing should make for a more efficient market over time.
- Under the different models of marketplace lending, to what extent, if any, should platform or “peer-to-peer” lenders be required to have “skin in the game” for the loans they originate or underwrite in order to align interests with investors who have acquired debt of the marketplace lenders through the platforms? Under the different models, is there pooling of loans that raise issues of alignment with investors in the lenders' debt obligations? How would the concept of risk retention apply in a non-securitization context for the different entities in the distribution chain, including those in which there is no pooling of loans? Should this concept of “risk retention” be the same for other types of syndicated or participated loans?
We do not believe that “skin-in-the-game” models are superior to models where loans are sold in their entirety. Unlike banks that have sticky capital in the form of deposits, marketplace lenders are reliant on a broad spectrum of investors to raise their capital. This is their “skin” – if they lose the support of their investor base, they effectively fail to function as a business. Forcing a platform to hold loans will not necessarily make them a better lender.
- Marketplace lending potentially offers significant benefits and value to borrowers, but what harms might online marketplace lending also present to consumers and small businesses? What privacy considerations, cyber security threats, consumer protection concerns, and other related risks might arise out of online marketplace lending? Do existing statutory and regulatory regimes adequately address these issues in the context of online marketplace lending?
Loan availability can be both a blessing and curse. Harm can be done by platforms offering very high rates for short-term loans, especially on the small business side where usury laws do not apply. Regulators should take a hard look at any platform that does not disclose the rate being charged – these lenders are likely to harm less savvy borrowers.
As far as privacy considerations, cyber, consumer protection, etc., we see similar considerations being taken as in other lending markets.
- What factors do investors consider when: (i) Investing in notes funding loans being made through online marketplace lenders, (ii) doing business with particular entities, or (iii) determining the characteristics of the notes investors are willing to purchase? What are the operational arrangements? What are the various methods through which investors may finance online platform assets, including purchase of securities, and what are the advantages and disadvantages of using them? Who are the end investors? How prevalent is the use of financial leverage for investors? How is leverage typically obtained and deployed?
Investors should consider many variables when investing in loans originated by marketplace lenders. One of the most important considerations is avoiding moral hazard. That is, investors should be careful not to invest with platforms or loan types that cater to borrowers who have been correctly turned away from other lenders. We like to ask, “What problem does this platform solve?” If we can answer that question without too much debate, we will continue our diligence process.
As far as structure goes, investors need to understand if they own whole loans or fractional loans and whether or not they are in a bankruptcy remote vehicle. There is enough credit risk inherent in loans without taking the risk of the underwriter on as well. We will only invest in loans and platforms that give us enough information to understand the broad construct of their lending model. While data will become available in newer lending spaces over time, this will come with significant defaults and losses on the investor side. Investors therefore should not be blinded by platforms claiming to charge higher yields or shorter duration investments if those lending markets have never been tested before.
Operationally, retail investors are exposed to whatever mechanism the platform has established for buying loans on individual platforms. Platforms should disclose whether or not their structure is bankruptcy remote and what happens in case of platform default. From an institutional standpoint, investors may also buy whole or fractional loans. We prefer buying whole loans and taking full custody of those loans. In addition, we employ a backup servicer that would come into play if something impacted a platform’s ability to service our loans.
Our end investors are private investors and family offices, along with some hedge funds and more traditional institutional investors. Though marketplace lending is a new asset class, these investors expect the same institutional framework that they would get in any traditional fixed-income arrangement, with some exceptions for lower liquidity.
We fund our purchases with a combination of cash and leverage. Leverage works much like it does in other fixed-income asset classes. We have access to a warehouse line that has a floating rate and a fixed term. Beyond that, we have the ability to securitize in the broad capital markets. Our first deal, priced in March, was called BLT 2015-1.
- What is the current availability of secondary liquidity for loan assets originated in this manner? What are the advantages and disadvantages of an active secondary market? Describe the efforts to develop such a market, including any hurdles (regulatory or otherwise). Is this market likely to grow and what advantages and disadvantages might a larger securitization market, including derivatives and benchmarks, present?
There are no secondary markets for these loans. An active market could drive lending rates down, which would be positive for borrowers. On the other hand, lower yields and more liquidity would drive volatility higher, lessening the amount of capital available in the space. Over time, the market will find the proper mechanism to create liquidity, but this should not be rushed.
Developing a secondary market will be difficult. The loans are very small – it is unclear who has the incentive to trade around tens of thousands of $12,000 loans. It is much more likely that securitization will open the door for more liquidity, though even this will be limited. Most platforms do not underwrite close to enough volume to support constant securitization.
Derivatives are even further away from reality. One of the large positives of the marketplace lending space is that there is no central repository for risk. Instead of ending up on the balance sheet of the big banks, loan risk is being sold to a very wide variety of investors who have effectively decided to hold the loans to maturity. With such a diverse investor base, there is little incentive to trade a derivative to hedge the underlying basket of loans.
The partners at Blue Elephant are available to discuss any and all of these points.