Wednesday, June 20, 2007

Prosper: A hands-on education in risk management

For some new lenders on Prosper it is a hands-on education in Risk Management. There are several different ways to learn about managing risk. Most colleges have economic courses on risk management, and there have been lots of books written about the topic and how it relates to investing. My favorite book on the topic is a book called Against the Gods: The Remarkable Story of Risk by Peter L. Bernstein. I thoroughly enjoyed reading this book, and highly recommend it.

There are too many important concepts to cover them all in one article, but the two most important items are diversification and risk adjusted return.

Let's say you have $1000 to lend. If you put all of that on one or two loans then you run the risk of that loan defaulting and losing all of your money.

Here are some lenders that did just that:

If these lenders had instead diversified, then like this lender the default would be a small part of the overall portfolio and would not have a significant impact on the overall performance.

This concept extends beyond Prosper and is the reason that financial planners often counsel investors to buy mutual funds or index funds which allow the risk to be spread across a large group of stocks rather than putting "all your eggs in one basket" by purchasing a small number of individual stocks.

The next concept is risk adjusted returns. This is where Prosper's Marketplace Performance Data will help us out. Adjusting for loans that are too recent to go late or default we get the following percentages (by number of loans):

  • AA: less than 1% late or in default
  • A: 3% late or in default
  • B: 7% late or in default
  • C: 10% late or in default
  • D: 13% late or in default
  • E: 28% late or in default
  • HR: 45% late or in default
  • NC: the performance wasn't good on these loans, so Prosper has discontinued allowing people with no credit to borrow

With new lenders, I think there is a great temptation to bid on HR loans. You see a 29% interest rate, and a nice story that explains how the money will be used, and think "Wow! why chase after a 12% A rated loan when there are all these HR loans at 29%?" The answer is that once you subtract the 45% to account for the HR defaults you end up with a return of -16%. At a 29% interest rate on an E loan you end up with a 1% return after accounting for defaults. If you hear someone in the Prosper forums talk about having lost money on Prosper there is always one of two things going on: either they didn't diversify their portfolio or their portfolio is heavily weighted with HR and E loans. I have yet to come across a diversified portfolio that avoided HR and E loans that isn't doing well.

Here are how the stats actually end up by credit grade:

  • AA: Expected Rate 9.5% (Start with an average rate of 11% and subtract 1% for defaults and the 0.5% Service fee)
  • A: Expected Rate 9.5% (Start with an average rate of 13% and subtract 3% for defaults and the 0.5% Service fee)
  • B: Expected Rate 7.5% (Start with an average rate of 15% and subtract 7% for defaults and the 0.5% Service fee)
  • C: Expected Rate 6.5% (Start with an average rate of 17% and subtract 10% for defaults and the 0.5% Service fee)
  • D: Expected Rate 7.5% (Start with an average rate of 21% and subtract 13% for defaults and the 0.5% Service fee)
  • E: Expected Rate -4.5% (Start with an average rate of 24% and subtract 28% for defaults and the 0.5% Service fee)
  • HR: Expected Rate -21.5% (Start with an average rate of 24% and subtract 45% for defaults and the 0.5% Service fee)

Keep in mind that these stats are based on a historical average which may not predict future results. Also, this is based on only about 18 months of data on 3 year loans. Once more longer term data for a larger number of loans is available these types of statistical analysis will become more precise. What these statistics do suggest is that, surprisingly enough, borrowers with A and AA credit end up being the best long term risk adjusted investment, and can be expected to return about 9.5%. Compare this to other investments like CDs or bonds and this is a really great investment return.

As you can
see in my portfolio, 8 of my 15 loans are in the AA category and none are in the E/HR category. The B-D category still has a reasonable 6.5-7.5% expected return so, while not as good as A or AA, I still fund some of those loans if I find one that I think is a better than average loan for that credit group.

If you start lending after having a understanding of these two concepts I think you will do very well. If not, I think you will learn these concepts along the way.

Here is a good example of a lender who learned along the way. He goes by the name of BigGulp and is active in the Prosper Forums. What you will notice by looking at his loan details is that of the first 50 loans he funded, only one was to borrower with a credit rating of A or higher. Several defaults and late loans later he learned his lesson. Now, looking at his most recent 7 loans, 5 of them were to A or higher credit ratings. And, in the past 2 months none of his loans were to HR or E borrowers. So far he has earned about 1% on his portfolio after subtracting for defaults. If you were to take away the E and HR loans from his portfolio he would be at over 10%.

4 comments:

tom said...

Great article - thanks! Based on this report it seems pretty clear to me that it would be wise to avoid ALL E and HR loans. The default rate is just too high to make them worth the risk.

I was surprised at how close the adjusted rate of return is for AA-D. Only a 2% difference.

Anonymous said...

The adjusted returns should be very close, since the rate charged should account for the default risk. I think the adjusted rate even makes sense, when you consider that Prosper most likely competes with CDs and market investments. So somewhere between 5 and 11% would seem about right.`

Alex said...

It's extremely important to acknowledge the negative side of diversification on Prosper. A behavior is risky when a person doesn't know what they're doing - to paraphrase Warren Buffett. In this case, the trick with non-diversified Prosper holdings is to pick someone who you know well, who will pay you an outstanding interest rate, and is very likely to repay.

I am a full funder on HR loans, and occasionally I meet an HR person in real life, build a relationship, and then get them onto Prosper to create a loan.

After that's done, obviously I can't do anything to get them to pay me back - but I've done the leg work up front to pick people I trust and who have prospects for repayment that mismatch their credit score.

suzan said...

Good article about education in risk management