Friday, September 14, 2007

Do Prosper.com lenders quantify adverse selection in their pricing models?

Jenny R. studied all the loan performance data made available from Prosper.com and read other people's analysis as well. Still, she says, "there is something missing". "Something does not feel right." she says to me when explaining her interest but apprehension in lending to strangers over the internet. Her dad had told her to ask me whether I thought she should invest $25,000 through an online lending site. I first asked her a bunch of questions about her financial objectives and constraints. Turns out $25,000 is money she can afford to lose, but the time commitment required to actually lend online is not a good fit for Jenny. Anyway, I asked Jenny how she could quantify the adverse selection in the borrower population. She asked me to describe adverse selection. Some of you have asked me about peer lending before so I figured a post was deserved.

What is adverse selection?:
1. Think of a credit worthy person. May be you are a great credit - you have terrific excess cash flow and liquidity - that is what a great credit is - someone or some company that does not need credit. This person uses a credit card for purchases and does not revolve balances - at least not for long. This person has stopped opening credit card offers. He/she has no time for it. The last thing this person is going to do is go online and go through a bunch of hoops to borrow $25,000 or less at prime or prime plus. Why would they?

2. Imagine a peer lending site that has originated a really meaningful size portfolio - like say $400 million in a twelve to eighteen month period. With that number of borrowers, think about the type of people who have learned about peer lending and decided to take all the steps necessary to borrow. Stop there and pause for a moment. Are FICO, profile and "community" adequate inputs to price unsecured 3 year debt? How could and how should you distinguish between (loss adjusted) expected returns between two individuals. Let's say one of the individuals is you and you only have two facts about the other person. The first is that you know their FICO score is identical to yours. Second, you know that they are seeking to borrow money from you through an online forum. Do you think the individuals deserve identical pricing? Do you think the market should price the individuals identically?

Bigger picture:
There are very valid arguments for individuals to consider alternative investments including lending or participating in debt securities. Disintermediation is highly seductive and makes for a great pitch. No one wants to piss on the idea and neither do I. I do wonder through whether a FICO score and profile are sufficient inputs to a credit decision. I also wonder if the majority of individuals who are moonlighting as consumer credit officers understand adverse selection. If you are drawn to think of the NextCard portfolios as I am you will recall that much of NextCard's growth was from transferring balances from other lenders. As I am writing this I am thinking that in these term loans you don't even have that explicit debt replacement that credit issuers do when they send money to pay off balances on other credit accounts. With Prosper, Zopa, and LendingClub.com you have incremental debt with a person who is willing to go through all these hoops to get a term loan.

In sum, I say good luck lenders, borrowers, and good luck Chris Larsen. Benchmark is just doing what they told Calpers they would do with the money. FT's associates probably wanted to call but their limited partners (Bank of America, Wells Fargo, HSBC, etc) would not have been quite so giddy. None of these guys (ELoan, Mortgage.com, NextCard, etc.) were able to lever the internet to really change the economics of consumer credit. I am not a hater (quite the opposite), but I will bet against the idea that peer lending is going to change the entire economics of the industry. Hope I am wrong. You have my profile here and my FICO is top decile. Is your money cheaper than Bank of America's or CapitalOne's? Do Prosper.com lenders quantify adverse selection in their pricing models?

No comments: