Prosper Lending Revisited: Will Returns Drop As Defaults Increase Over Time?

Months ago I did a review of Prosper, a website which allows you to earn interest by lending money directly to others (Prosper takes a small cut). After looking at the mechanics of their system as well as their own historical loan data, the main conclusions from my initial review were:

  1. You should avoid loans from those with poor credit rated E and HR like the plague, as they have negative annual returns ranging from -10% to -30% annually. Prosper lenders as a whole priced this subprime market very poorly.
  2. If you stick exclusively to the borrowers with the best credit score (rated AA and A), manage your cash carefully, and the default rates don?t keep rising, you may achieve average net returns of about 8% annually.

A key part of that last sentence is if the default rates don?t keep rising. Sure the initial interest rate may be a snazzy 10-12%, but these loans are all three years in length, and my theory was that as time goes on more and more people will default on these loans. Or maybe some will vary between being late and becoming current again, so that the return stays pretty constant. Now that there is more history in their database, I decided to run some number to test this theory out.

As in my original review, I took all the loans that originated in the first half of 2006. Then, I looked at the ROI (average annual return after taking into account defaults and Prosper fees) as observed on different dates ranging from October 2006 to June 2007. Here are the results:

altext

Summary
It would appear that there is indeed a gradual “decay” of annual returns, with the rate of decay increasing as you drop into the lower credit grades. Although this is not conclusive evidence, it is something to consider if you are expecting a certain level of performance. If I were to lend on Prosper, I would stick exclusively to the AA-rated loans.

Even with AA-rated loans, right now we are less than halfway done with these loans. If these trends continue, by the end of the 3-year term, I expect the average net return to be about 7.6-7.8% annually. Again, this is an average value, and one would still need enough money spread across a number of loans to protect from individual loan risk. 7.7% actually isn’t bad, and is almost enough to make me commit some money to this if the spread above FDIC-insured equivalents remains high enough. I’m currently remaining on the sidelines until I see how the marketplace responds to any federal interest rate hikes.

Comments

  1. Danny at Money Socket says:

    Excellent analysis Jonathan, looks like AA is the best route. Better to get a stable and good return than a potentially higher but volatile return. I am considering starting prosper as well by the end of the year. Keep us posted on your findings.

  2. That’s a great post, Jonathan – and a great graph to go with it. Its interesting how much difference in risk moving down to B rating makes, but doesn’t increase the initial loan rate by much.

  3. Jon – there was not much borrower credit info provided for lenders to analyze until the second half of 2006. It may be better to analyze the second half of 2006.

  4. Sucker.com says:

    IMHO, you need a philosophical commitment to micro-loans to be interested in Prosper. A return of 7.6% is just not enough to compensate me for: relatively high risk of default, risk of bad pricing, risk of potentially arbitrary intermediary (Propser), and risk of non-diversity (cf a bond fund).

    It sounds like playing a financial services game like the big boys, but without their deep pockets, risk analysis, and regulatory structures.

  5. After the sub-prime mortgage loans fiasco, I can’t get myself to trust putting my money into this type of program.

  6. I would stay away from Prosper. I made 15 loans and have already had 5 go into bankrupcy. And the interesting thing was that the bankrucpy occurred within the first 3 months of the loan. Looks like they knew they were going to file and then took out the money knowing they would never have to pay it back.

    Out of the 10 remaining, 5 are late > 2 months and 2 are late

  7. anon – I sure will, once a few more months pass. Right now I don’t think there is enough history.

    Steve – Out of curiosity, what were the credit grades of your loans?

  8. It seems to me you are taking a substantial amount of interest rate risk with three year loan terms. I can put my money in a money market fund and earn 5%+ right now. If interest rates continue to climb locking my money up for 3 years with the expectation of only receiving 2% more back even for AA rated borrowers isn’t appealing.

  9. I have to say that it does seem like the lower the grade goes the worse the return in my experience. That said, you are still taking the averages, which is not necessarily indicative of what your performance might be. For instance, you could set up a trigger to bid on AA loans that where the lender is willing to pay 16% and has no prior deliquencies. These types of loans are rare, but some very good loans can be caught each day through this method. If you think about it, even if you can get a D-loan that’s better than average through this means, you could get 7-8%.

    I know it may seem like stock picking, but it’s really much more by the numbers as loans prices are influenced by outside news. Over time there should be a direct correlation of the numbers from the lender and the return from the investment.

  10. Steve: 3-year bonds aren’t yielding anywhere close to 7-8%. If you continuously reinvest in Prosper it will act like a bond fund with a average duration of 1.5 years and a yield of 7%, which is much better than anything else.

    However you have to subtract out the delay between bidding and funding a loan.

  11. Andreas Ronneseth says:

    I invested in prosper starting around October 2006. I put in $2,500, and did $50 loans to spread my risk. So far, I’ve had $97 default on me, another $92 about to default. I’ve made $249, so that leaves me with $60 profit.

    I have the fewest AA, a little more A, a little more B, and the most C & D. I don’t touch E.

    So, with the time I’ve spent on prosper, I’m thinking I’ve made about $0.02 / hour :)

    Really, I look at this as a fun thing to do, and I’ll see in the long run how it will pan out. To get all my money lent out, I probably took some undue risk, and now that I only need to make 1-2 loans per month, I’m a little more picky.

    That’s my 2 cents…

  12. Lazy – Perhaps some screening will improve the returns. Hopefully as more data become available for dates with more loans that can be teased out.

    Andy – That’s an interesting perspective to compare with a short-term bond, and is kind of how I was thinking of it. What if one could re-sell their Prosper loans…

    Andreas – Gotta avoid those low credit grades! :)

  13. There was a brief article in Parade magazine about Prosper if I remember correctly. Bottom line was that all the risk was born by the lender and that the Prosper lenders were funding at rates well below sub-prime rates that a commercial lender would fund such loans due to competition among the Prosper lenders. Good deal for the borrowers but not for the lenders. Seems your data bears that out.

  14. The credit grades for all my loans are C or higher. The interesting thing is that 3 of the 4 bankrupcies were rating B and the other two were C’s

  15. Jonathan,

    Your analysis that defaults on Prosper increase as credit grade decreases is certainly correct. Testimony to the utility of the Experian Scorex score that Prosper uses.

    However, there is more data available to lenders than just the credit score, and many lenders consider this as valuable if not more valuable than the credit grade.

    Try going to Prosper’s performance page and doing a query on all loans at least one month old (so it will estimate ROI) with 0 current delinquencies, 0 dq’s in the past 7 years, 0 public records in the last 10 yrs, and 0-5 inquiries in the last six months. You’ll find that even the E loans meeting this criteria have performed satisfactorily (7+%), and the higher grades have done even better.

    Edit the criteria again to exclude automatic funding listings (where the borrower is so desperate for cash they decide to forfeit a bid-down of their rate) and ROI on credit grades AA down to D are all at least 8.4%. (Not enough E or HR loans meet the criteria to get an ROI estimate on those grades.)

    I started with Prosper in early 2006 when all the info they gave lenders was the credit grade and DTI. They added the credit details listed above in May 2006, and added even more details including revolving credit balance in Feb 2007. Ten of my 177 loans are late or defaulted, and of those, eight had originated in June 2006 or earlier. Of course the newer loans have had less time to go bad, but my older bad loans went bad early, so I attribute this mostly to (1) improved data from Prosper as time progressed and (2) my own learning curve.

    Don’t get me wrong – Peer to peer lending is risky and I’m not recommending Prosper as an investment to anyone. I’m just saying that if someone *who enjoys the P2P concept* researched the ins and outs first, didn’t send to much money toward Prosper at a time, and bid on strict criteria, they’d probably do OK and possibly also have a great side “play money” investment to diversify away from other investments.

  16. the A and AA loans that I have (about 10 of them) are at interest rates of 10-14%. I have been doing this for about 1 year and have zero defaults or lates. knock on wood. i have mostly A and AA, a few B, and one C.

  17. I’ve started with $300 since February of 2006 – have it on autobid when I have another $50 in there. I haven’t had anyone default and have limited myself to A or AA borrowers. My after-tax return is >10%. I think maybe it’s because I got it really early when the rates weren’t bid down as much? Now I’m having a harder time funding loans based on my ‘pie-in-the-sky’ critera.

  18. Good work on this one. I was on the fence about using prosper as an investment vehicle but you cleared up that ambiguity for me. Thanks for the help. The money is better placed in a good mutual fund.

  19. Starting from April 2007 I invested in $1100 to prosper, having 22 loans. I put 67% of my money on AA and A borrowers. The rest is spread in B, C and HR. So far so good; no late payment, no default. Everage interest rate 12.9%. I am waiting to see what happens in the future…

  20. AA and A ratings are useless on this site since your money will be locked for 3 years with a 7-8% interest.

    On the other hand, you can use many parameters such as no of delinquencies, public records etc. to determine the lowest risk E or HR applicants to get a 20%+ rate. IMO, this is a better way to invest a 3-yr bond.

  21. Fantastic analysis! Thanks

    Here is a question for anyone:

    I hate getting outbid and just want to get my money in the market. If I focus exclusively on borrowers who choose automatic funding, do automatic funding borrowers have any significant difference in default rates or late payment rates, from other borrowers? Have not been able to find that answer in the Prosper database, but it should be simple for them to provide.

  22. The short analysis here is great. Those interested in more stats can find great information on Prosper-dedicated sites such as http://www.lendingstats.com and http://www.ericscc.com. As a current lender, I also find myself in the Prosper forums a lot.

  23. I have been investing with Prosper since Sept 06, purchasing about 3 loans a month. I now have 49 loans at a rate of 16.5%. I have had no defaults so far with 70% of my loans in C grade or worse. I make sure that i dont invest in anymore than 50$ per loan, and that they belong to a 4+ star group, have auto funding and a home owner.

  24. To all,

    I was trying prosper for since August 2006 and now have three AA loans in > 2 months default. That sucks – small money invest

  25. to Opie: Not to burst your bubble, but don’t exclude the possibility that you just have to give your loans more time for lates to pop up. For what it’s worth, I started in February 2006 and have no defaults or lates since November 2006, but overall I have 8 defaults and 9 lates (out of 200+ loans). The $50 per loan strategy is good. However, recent changes on Prosper have very dramatically changed the group dynamic, and I’m not so sure of the impact this has had on the value of the star-rating.

    to Barry: I feel your pain with some delinquent AA’s of my own. Our late AA’s illustrate my point that credit grade should not be used alone to judge Prosper listings – the expanded Scorex data (number of delinquencies, public records, inquiries, etc) should be used to further narrow your search. My late AA’s either (1) had delinquencies or (2) were asking for the full $25K Prosper allows (another red flag IMO).

  26. Excellent post, I think your analysis are highly accurate.

  27. Hi, I would like to have you guys’ opinion on my status. I have a 1 year loan with Prosper, with approximately 19% interest, have never been late and not late on any of my credit cards or loans. Whw ould they lower my rate from a C to and HR in one year evern though I am never late on anything. The only difference since last year is that I closed an auto-lease account after it ended (did not need the car anymore) and I got new loan of a couple of thousand $. But I always pay in time. Additionally, they told me my rating is now lower than 640 (which I cannot seem to find anywhere as it is much higher in Experian) and that it went down for more than 20 points since last year which I have not been able to track either. What would be the reason for someone’s rating to be lowered dramatically although there have been no issues in the credit file at all?

  28. Eva,

    Prosper just recently returned to business after a Quiet Period while they sought SEC approval. Many of their rules have changed since they were open before, including their rating system. What they now call HR borrowers were roughly equivalent to D borrowers under the old system. Most borrowers who ranked as HR and E before can’t even borrow anymore on Prosper.

    Also, Prosper uses a specific product from Experian called Scorex Plus, which is supposed to specifically model risk with unsecured loans. A Scorex Plus score may differ somewhat from the regular Experian score, as it puts more weight on things like credit utilization.

  29. So Prosper just tighten its standards. Peer-to-peer lending is no different than banks and credit unions. There are other ways to determine if people with no established credit history/or first-time borrowers is creditworthy or not rather than soly relying on any credit score. On the following links; watch and listen how Northside credit union rescued this young lady from the payday loan trap before her credit became damaged, and check out the small unsecued payday loan alternative product to help it’s low income members from falling into the debt trap of payday loans.

  30. Great post here Jonathan, thanks. I’d avoided loaning via Prosper after reading before that advertised rates were not matching the actual rates but as you note we need more data like yours. This suggests Prosper may be a very good investment if one sticks to the AA borrowers. There is some intuitive logic to this as well. If all parties act fairly responsibly, prosper eliminates many of the costs of traditional banking, and should be a “win win” deal for both parties.

Trackbacks

  1. [...] was trying to run some more recent numbers to update my most recent concerns about default rates for person-to-person loans at Prosper.com. I was curious because it’s one thing to advertise [...]

  2. [...] rate” at Prosper or simply go with the pre-calculated rate from LendingClub. In the past, I haven’t been too impressed with the rates that resulted on Prosper. Part of the issue, I believe, is due to rates being lower [...]

  3. [...] posts: Review Part 1, Review Part 2, Default rates, Scary [...]

Speak Your Mind

*