Harris is a managing director at Bain Capital Ventures.
Credit, private or public, is of greatest consequence to every country. Of this, it might be called the invigorating principle. -Alexander Hamilton
So which is it? Is lending a good business, where moats and profit pools are durable and one can build billions in equity value … or is it an inevitable cash bonfire when the cycle turns? Is credit a critical lubricant for the economy, the more the better, or a creeping contagion to be limited, particularly in the “shadow” banking system? As someone who co-led the seed round in one of the first alternative lending companies (OnDeck), and went on to serve as Chairman, but who has not invested in a single lender since, I stand somewhat astraddle this issue. After watching closely for ten years, I remain convinced that the leaders of Alt Lending 1.0 (specifically Lending Club, SoFi and OnDeck), if they can avoid getting acquired due to depressed stock prices, will build enduring platforms. I am most worried about the wave of “me, too” players that came next, Alt Lending 2.0. Most of them built their business plans on the now shifting sands of temporarily high valuations and cheap capital, and have insufficient differentiation to carve out a durable niche. I am most excited about the Alt Lending 3.0 players, who have had the ability to learn the appropriate lessons of the past decade and can both pick their spots and choose a more capital efficient path.
If you look at the question with an eye towards history, what is perhaps most surprising is that venture capitalists are thinking about lending at all. In the past, non-bank lending was either about disreputable segments or scale finance companies, and in either case raised equity with the promise of distributing yield based on high ROE, not primarily creating high multiples through exits. What changed? In a phrase, “marketplace lending”.
To explain what that is, I’ll quote myself … I was asked by a banker in 2014, before any of the lending IPOs, to explain Lending Club. Here’s what I wrote back to him (edited for brevity and clarity):
I’m quite familiar with them. They are the pioneer in an innovative new funding model for specialty finance companies, whereby they originate loans and then sell them off entirely, without keeping any of the risk or needing to tie up any equity capital in their assets. They receive a ~4% origination fee and an ongoing 1% annual servicing fee on the assets. They benefit from a number of unsustainable trends, including:
I believe many or all of these trends will reverse in the next 2-3 years, but Lending Club has built a substantial business and diversified away from their historical base of unsecured consumer loans, so I believe they will end up being a survivor. To the extent to which their innovation (marketplace funding of loans) is durable, it will have created a new species of competitor for banks, pure play loan originators who can compete based on marketing prowess, underwriting skill, proprietary product creation and other areas where FIs have historically been weak.
Marketplace lending is the “technology” that creates the biggest risk for FIs on the asset side, in the same way that prepaid is the “technology” that creates the biggest risk on the liability side (per this essay):
One of the good things about being an obstinate cuss with no capacity for new ideas is that it creates efficiencies like this one … I don’t have to change a word of this, even two years later.
It is still possible, though I believe increasingly unlikely, that marketplace lending will be a durable innovation. As a society, we decided that mortgage originators should hold some risk in the product they create after they nearly destroyed the world in 2007; should other types of credit be treated differently? The specific mechanism used, whereby a bank technically originates the asset on behalf of the marketplace lender for compliance reasons, has been challenged by the courts. A more straight-forward risk to the model is demand for these assets through cycles. In normal yield and risk environments, asset owners have required that originators be aligned with them regarding asset quality. If rates rise and credit quality deteriorates, this demand will likely re-emerge.
That said, I don’t think it matters whether marketplace lending ends up being a viable long-term funding strategy or not. What Lending Club, OnDeck and others proved was that they could effectively compete with banks and build share in their respective categories. The unique way that Lending Club funded its assets was a feature, not the entire product. As an example, when I backed OnDeck originally, it was before marketplace lending (or “peer to peer” lending, as it was originally called) was in vogue, and the assumption was that we would have to follow the traditional asset funding path. That path involves funding initial loans using 100% equity, until the assets are proven out, then layering in leverage over time. This generally starts with a 50% loan-to-value advance rate, eventually climbing to 80-90% over time as/if the assets perform. Ultimately, you can build a highly redundant funding machine, as OnDeck has done, with essentially equal parts balance sheet funding (using high advance rate leverage and an SPV), rated securitizations and whole loan sales (the prosaic, old school term for what now goes by the dazzling nom de guerre marketplace lending).
As an amusing aside, I’ve now been involved in two public panel discussions and multiple private conversations with Lending Club investors around funding models. It always goes the same way … they say that there is no more resilient model in capitalism than a marketplace model, hence the ineffable beauty of Lending Club. I then ask how a marketplace could be more resilient than a marketplace PLUS a securitization conduit PLUS committed multi-year balance sheet funding (the SoFi, Avant and OnDeck model, among others). Following that I am branded an apostate and asked to relinquish my Fintech Fan Club ™ membership card. In point of fact, I consider no fewer than 6 current or past Lending Club board participants close personal friends, and the visionary investments they made in that company cemented their reputation as great investors. I just think the reason they chose well is the classic venture capital formulation of “great founder, big market”, rather than any notion that Lending Club discovered the credit equivalent of cold fusion. The fact that Lending Club was able to scale without using equity in its loans because of the marketplace model was a brilliant strategy, but far from the most important aspect of the success of the company.
The things that matters about how you fund your assets are a)cost, b)liquidity and c)risk … as an originator, you need to manage all three. The marketplace-only model focuses exclusively on risk, and emphasizes to the point of fetish the idea that the originator bears no ongoing exposure to the assets once sold. Of course that’s not true: If all of your assets go pear-shaped, you’re almost as screwed as if you owned them. Risk is not the only thing to focus on.
Venture Capital in Alternative Lending
The debate around marketplace lending is a bit of a sideshow. The fundamental questions are:
I do believe that non-bank lenders will take share from banks, though I think they have to pick their spots carefully. I think an economic model can be built that works through cycles, though there will be some white knuckle moments along the way. Marketplace lending, or whole loan sales, can be a useful part of a diversified funding scheme, but is not mission-critical to creating venture returns.
To understand why and where banks are vulnerable, we need to decompose lending into its critical parts: Customer Acquisition, Credit, Funding, Servicing/Customer Relationship Management.
Lenders have to attract customers, figure out if they want to lend to them, fund the assets that they create and then service and collect from the borrowers over time. Despite much rhetoric to the contrary, banks are not particularly terrible at any of these skills, and they benefit from enormous economies of scale across the entire value chain. In order to compete, non-bank lenders need to pick one or multiples of those aspects of the chain, and differentiate in a way that is hard for banks to follow.
The hardest way to compete with banks is relative to funding, which makes it all the more peculiar that the alternative lending sector has come to be branded by its asset funding methodology, marketplace lending. The average cost of funds for a bank, according to the FDIC, is 0.06%, assuming they fund their loans using deposits. OnDeck’s funding costs for its assets averages 5.3%. Lending Club has paid a median return to its asset purchasers of 7.4%. Any start-up lender, particularly in unsecured or novel credit categories, will start out with rates in the mid-teens, if they can get leverage at all. No one can fund cheaper than a bank.
While we’re on costs for a moment, another common anti-bank argument is around inefficiency in terms of marketing and servicing. Lending Club estimates that the big banks suffer from operating expenses as a percent of loan balances on the order of 5-7%, while it has marketing and servicing costs of around half that amount. While true, I’ll point out that most of the bank expenses they highlight are fixed expenses like branches and compliance, which makes that expense burden irrelevant to the profitability of the marginal loan. In addition, what banks overspend in these areas they make up for in IT and G&A … for example, banks on average spend 7% of revenue on tech, compared to 12% at Lending Club and a whopping 24% at OnDeck. When combined with the enormous and irreducible funding cost disadvantage, we can safely say that neither funding nor cost structure represent areas of long term competitive advantage for the insurgents. That leaves us with Customer Acquisition, Credit and Servicing/CRM, plus combinations of the above. The good news here is that there is opportunity in each.
Most start-up originators focus on the opportunity to innovate in credit decisioning. The headline appeal of new data sources and new data science seem too good to be true as a way to compete with stodgy old banks. And, in fact, they are indeed too good to be true. Building new credit models is hard, inherently time-consuming and expensive. Fundamentally, you can’t figure out how to distinguish between the “goods” and the “bads” without sufficient “bads”, and “bads” cost you money. Credit losses can also take some time to develop, depending on the category, and assets can behave quite differently in different economic cycles. Until you have a model that is proven in a highly durable way, your cost of capital will be very high.
But wait, it gets harder. Sexy new data sources can be tricky, particularly in consumer lending. Not only can you not intentionally use criteria that the regulators view as prejudicial, you also can’t leverage variables that seem innocuous but end up being correlated with things like race and marital status. Further, the new data itself isn’t interesting; what’s interesting is how it affects credit quality. You can have all of the hypotheses you want about how data should rank risk, but no one cares until you demonstrate it.
The recipe for success here starts with picking a truly underserved segment. Then figure out some new methods for sifting the gold from what everyone else sees as sand; this will end up being a combination of data sources, data science and hard won credit observations. At least some of your credit observations have to be non-obvious. One good test is to see if FICO scores help to sort risk; if they do, pick another category … banks understand FICO scores. It is important to choose a segment with relatively small dollar amounts and relatively short duration. Credit models are judged on unit performance more than dollar performance, and small dollar/short duration lending allows for many more “bads” on far fewer dollars. It is also a bulwark against bank competition. With larger loans, banks can and will invest human resources to underwrite loans; with smaller loans, they don’t have that option. Finally, while starting in an out of favor category is probably essential, it can’t remain true forever. You have to have a game plan for moving mainstream once you’ve built your competitive advantage in underwriting.
There are relatively few examples of companies that have gotten all of these elements right, but there are some. Oportun provides credit to otherwise highly disadvantaged Hispanic borrowers. What a bank sees as a 620 FICO, Oportun sees as an employed but challenged immigrant with a spotless record of sending remittances home. Sub-prime and near-prime lenders like Avant and LendUp have proven out extremely advanced consumer credit models in unattractive segments, and are working quickly to leverage those models into more mainstream businesses. In small business, OnDeck chose a large segment that had historically been attractive but hard to serve. Where a bank sees a generic SMB with a business owner FICO score of 650, OnDeck sees a plumber with far higher efficiency than comparable businesses in its zip code, and a glowing set of reviews on Angie’s List. This hard work comes at a cost; OnDeck’s R&D expenditures, noted above, are characteristic of the successful companies in this category. You have to have real IP to win here.
These companies are building defensible moats against bank competitors for a variety of fundamental reasons. The one area where the funding cost advantage that banks have doesn’t work for them is in areas where new credit models are required. Regulators insist that banks keep disproportionate amounts of equity capital in reserve when they try new things, and while their debt capital (ie, deposits) is cheap, bank equity capital is quite dear. In addition, these segments generally start out with high APR models. They can support high interest rates, as by definition these customers don’t have good alternatives, and as discussed, they require high interest rates to compensate for the risk and elevated cost of capital. Over time, APRs can come down dramatically, but these segments are generally dismissed in the early days as predatory or niche.
Customer Acquisition Advantage
Developing an advantage in Customer Acquisition is harder than it might appear. At first blush, it might seem that all it takes is to pick an unloved segment and get to know its idiosyncrasies and tendencies well enough to create a focused channel and tailored dialogue. In practice, progressive and thoughtful traditional lenders like Capital One have mined most of the segments large enough to build a business around. The only way to build a durable competitive moat based on Customer Acquisition is to become integrated into a channel that is proprietary, contextual and data-rich.
Integration is an old lending concept that has been reinvented in the modern age of networked computing and hosted software. Decades-old companies like Synchrony and Alliance Data were “integrated” at the point of sale in the form of paper applications for store-branded credit cards; now we have modern versions like Greensky, Bread and Affirm. These companies don’t leverage their channel to find a new or underserved segment; instead, they gain access to well understood, prime consumer credit via partnerships with companies hoping to sell things more efficiently.
It has historically been critical that these integrations be proprietary. Most, or all, of these companies insist that they be the only credit option in that channel (or at least have exclusivity by credit tier), and in fact, in most categories the credit providers get paid by the channel to lend, as well as by the borrower. It continues to surprise me that merchants have allowed this to continue in the age of marketplace lending. If I ran a retailer, I would not only refuse to subsidize point of sale lending with a merchant discount, I would run a real-time auction between lenders for each consumer seeking credit to make a purchase. This is likely not feasible today, but certainly technically possible over time.
On the B2B side, good examples of this are starting to emerge in the reinvention of commercial sales finance. Companies like Noesis and Wunder are providing tools to vendors of HVAC, solar and other high ROI systems to both prove the value of this equipment to business customers and finance the sale at the same time. For years, Bluetarp has been replacing the historical role of vendors in providing trade finance to the buyers of construction materials; now they as well as new entrants like Behalf are pushing this methodology into other verticals.
The home run is when the credit offer is not only presented in context, but also can be structured with the benefit of relevant data. Fundbox lends to small businesses by advancing them payment on receivables, in what is akin to a traditional factoring transaction (though without claim to the asset). They integrate with online accounting systems like Quickbooks, which not only provide them access to customers, but do so in a way that allows them to learn a great deal about the creditworthiness and demand for capital of their prospective customers. Bluestem Brands has a business called Paycheck Direct, in which they integrate with payroll systems to provide access to an e-commerce catalog on credit to low FICO employees; by having access to the payroll system for both underwriting and collections, they have a massive advantage over payday lenders.
One company that is synthesizing all of these elements is AvidXchange. Avid’s core business is Accounts Payable automation and commercial payments; they integrate and control the payment process of B2B customers. Based on that data, they know when and if suppliers are going to get paid, and in fact are responsible for making those payments. As such, they are in a unique position to make highly attractive credit offers to otherwise cash-starved vendors, leveraging their native integrations, proprietary access, ability to make contextual offers and data-rich software footprint. They can provide capital to the least creditworthy vendors simply based on the strength of the buyer and Avid’s insight into the likelihood of payment. This kind of position allows more than just sifting for gold (differentiating the creditworthy from the non-creditworthy), it actually creates the opportunity for the more alchemical process of turning fundamentally credit-challenged customers into good risk.
A promising but infrequently used vector in the lending space is to capitalize on a servicing, or more accurately, customer relationship-oriented strategy. The hero company here is SoFi, who focuses on the HENRY segment (High Earning, Not Rich Yet). They started with student loans, as a low risk asset class that had historically been underwritten in an unsophisticated fashion. With its narrow spreads and government subsidies, student lending in and of itself was unlikely to be a sustainable niche, particularly in a rising rate environment. To the credit of the SoFi team, they knew this from the outset, and treated student lending as their gateway product; they have now launched a nearly full suite of consumer financial products into their loyal base of student lending customers. The company generated that loyalty through laser focus, and by going above and beyond for their customers. It is hard to imagine a traditional bank lender providing job placement services for their borrowers, much less networking dinners and dating services.
Where Do We Go From Here?
While I’ve come to be known as a skeptic on marketplace lending per se, I see a great deal of opportunity in alternative lending more broadly. That said, things are going to get harder before they get easier. Valuations are down significantly, and the market volatility combined with a slowing economy will produce challenges with both default rates and access to liquidity. Of the alternative lending 1.0 players, Lending Club will be a survivor, and a long term leader. While it doesn’t have many of the distinctions I note above as important, it has scale, a brand in the capital markets for producing high quality assets, and an unbelievable management team. In fact, Lending Club may be the best example of managerial execution in a commodity market since Amazon, and like Amazon, if they continue on their current path they will figure out a way to create enormous value. OnDeck, in a sense, has an easier road ahead. They are very far along an experience curve with regard to the new world of small business lending, and it’s hard to imagine someone catching them. I worry about the 2.0 originators, which tend to be only slightly differentiated fast followers of either Lending Club or OnDeck; it will be very hard to be anything other than the leader in this shaky market environment.
I’m quite excited about Alt Lending 3.0, made up of earlier stage originators whose strategies line up with the precepts above, and who have gotten the memo about appropriate valuations and capital efficiency. For lending high flyers, as for the rest of the “unicorns”, the notion that high growth and “disruption” justify valuations detached from estimates of future cash production has gone away. Companies seeking funding in the lending space today generally have less of the Gold Rush mentality of the Alt Lending 2.0 players, who sought to take advantage of an enthusiastic investing community; instead, they are motivated by legitimately unmet needs and the prospect of meaningful moats. For the first time in a decade, I’m feeling like it’s a great time to be starting a lending company.