The U.S. Treasury department initiated a study of the rapidly growing marketplace lending sector. In a request for information, the public is invited to comment on 14 questions. Marketplace lending has experienced rapid growth, with new lenders originating over $12 billion. While the origination volume is tiny in comparison to lending by traditional banks, the growth rate is exponential. If the historic growth rate continues, these lenders have the potential to challenge the highly profitable consumer and small business lending franchises of leading banks. The platforms have enjoyed limited regulatory scrutiny, which looks set to change.
Marketplace lenders have largely been viewed as consumer champions. And from the tone of Treasury Department questions, it looks like the regulators want this industry to grow and to succeed. In the document, it is written that “online marketplace lending has filled a need by often delivering lower costs and faster decision times than traditional lenders.” The lower cost base, better ability to integrate new data into risk decisioning, lower interest rates and speed have helped these new entrants win over borrowers, investors and customers. In addition, the lenders have been exceptionally transparent. Borrowers can see if they are approved and check their interest rate without hurting their scores. And the interest rates are very low. MagnifyMoney (my website) conducts pricing surveys daily, and personal loan interest rates are now as low as 4.04%. That would have been impossible just a few years ago.
Potential investors are given an almost overwhelming amount of data, at loan level, prior to investment. The entire business model could not look more different than traditional banking. Despite the friendly tone, the questions in the document reveal where regulators may decide to focus. I think there are three areas where regulators may concentrate:
- How should marketplace lending be defined and segmented?
- How reliable are the credit risk models being used to make lending decisions
- Should marketplace lenders be forced to put more “skin in the game?”
What Is A Marketplace Lender?
For people who have worked in banking for a while (like myself), a lot of these marketplace lenders resemble nonbank lenders of previous lending cycles. In the 1990s, some very smart people set up credit card businesses in Delaware. Businesses like MBNA, First USA and Juniper had a simple model. They kept costs low. They invested heavily in data and analytics. And they sold credit card loans to investors. And ultimately they were sold to large banks like Bank of America BAC +0.00%, JP Morgan Chase and Barclays . Nonbank lenders have been connecting investors and borrowers for a long time. And small, entrepreneurial organizations usually execute cheaper, faster and often with more advanced analytical tools than banks.
If you look in the mortgage sector, non-bank lenders are now originating more volume than traditional banks. Businesses like Quicken and LoanDepot connect borrowers with investors, and often provide a much better customer experience than traditional banks.
And some banks look a lot like marketplace lenders. DRB Student Loans has rapidly created a very large student loan business. A small, traditional bank in Connecticut has built a large, national student loan refinancing business that is competing with players like SoFi and CommonBond.
The Treasury Department is deciding how to classify and categorize the new industry. People are using the word “marketplace lending” broadly. But I think regulators are looking for clearer segment definitions, and regulations may be different depending upon how each business is structured. Small nuances in the business models of different marketplace lenders could end up having large regulatory impacts as the rules are ultimately written.
In every introductory risk management course, you are taught that there are credit risk cycles. Right now, we are in a particularly benign environment. Credit losses are at historic lows. The new data being used in new credit models have not made it through a credit cycle. It is easy to look good now, but how robust are these models?
I have been to marketplace lending conferences where the 2008 crisis has been oversimplified. I hear leaders of marketplace lenders saying that loans originated before 2008 were “bad.” And loans originated now are “good.” I wish it was that easy. New underwriting models will only be validated when they are put through a full credit cycle.
I have a lot of respect for the new risk variables being used in lending models. The credit card industry exploded after data scientists leveraged bureau data to create highly predictive models. But mainstream models still have blind spots. Lenders like SoFi have demonstrated that a recent Stanford graduate working at Google is a probably a great credit risk, even if she doesn’t have a credit card. Lenders like StreetShares are leveraging the abundance of small business data that is now available from online bank accounts, online accounting software and social media footprints to create new underwriting models. These new tools for segmenting and underwriting risk will be game-changing for large swathes of the population. But true validation only comes through a downturn, and not all new models will work.