Default Rates at Lending Club & Prosper: When Borrowers Don’t Pay

Lending-Club-Prosper-Default-Rates

One of the biggest concerns people have with peer to peer lending is default rates – basically, how often these loans are not paid back. It’s a good question. After all, these are unsecured loans we are helping issue, meaning the loans are not tied to any collateral like a house or car. If the borrower cannot repay their loan, there is no real penalty they receive (besides a nasty mark on their credit), so a sizable number of these loans eventually do default.

Today we will answer this important question: what is the default rate in peer to peer lending?

An Annual Default Rate

When people ask, “How many of these peer to peer loans eventually default?”, they are often just asking how much of the investment is lost to non-paying borrowers. “How many?” versus “How much?” is a different question altogether; it is the difference between the number of loans and the total investment (principle).

We care a lot more about the principle. For example, even though 12% of Lending Club’s loans may eventually end in default, these bad loans might drop our overall return by only 3%. Interesting! Why this difference?

Lending Club Charge Off BummersThe answer lies in the way peer to peer lending operates. Peer to peer lending is a rolling continuous investment. After we first open an account, borrower payments will begin to collect as available cash, cash we can then use to invest in additional loans. Eventually there will be a lot of variety within our account. Some loans will be older with more defaults, and some will be younger with fewer defaults. Furthermore, when borrowers default on younger newly issued loans, we lose a lot more of our investment than we would if the loans were older and borrowers had already paid back much of the principle.

Therefore, the best way to think of defaults is how much principle, on average, is lost per year. This is the annual rate of default as highlighted in the next section, and it is a different number altogether from the eventual percentage of principle lost when loans are completed, not to mention the eventual percentage of loans that default. However, all three numbers are interesting to look at, so let’s take a moment to do that.

Lending Club & Prosper Default Rates

peer to peer lending default rates july 2013In the graphic on the right, you can see three different numbers for both platforms. The top number is the annual rate of default. This is the yearly percentage of principle (invested money) lost to defaults since the platforms began. Underneath is the eventual default rate for Lending Club and Prosper’s completed loans. The third number is the historical percentage of completed loans that have ended in default.

The Default Rate in a Nutshell

The most important number in this graphic is the annual default rate, how much of the average investment is lost each year to non-paying borrowers. For Lending Club, it’s 3.0% and for Prosper it is 8.6%. Lending Club continues to highlight this 3.0% default rate throughout their site, for example at their About Net Annualized Return page or in their FAQ. Prosper’s default rate of 8.6% is higher, largely because their loans have historically had higher interest and risk than Lending Club.

Let’s look at how I discovered each of these numbers.

Our Baseline: Annual Rates of Default

This was probably the easiest to discover. I simply went to NickelSteamroller’s Return Forecaster and backtested the historical rate of return for all Lending Club loans issued from 2009 until now (with loss factors found here). Note: I set an Issued Date of January 2009 for Lending Club (and July 2009 for Prosper) because this is when the platforms launched with approval from the SEC.

Lending Club’s annual default rate:
NSR July 13 Default Rate

Prosper Stats July 13 Default RateTrying to find Prosper’s annual default rate was done the same way, but through the Loan Analysis tool at Prosper-Stats.com. Both these tools do an excellent job at highlighting one of the big differences between Lending Club and Prosper. Lending Club has historically had much lower defaults than Prosper, a result of having stricter underwriting (660 minimum credit score to get a loan vs 640 for Prosper), though it seems like this may be changing.

Having a higher default rate has been the price Prosper has paid for offering substantially higher interest rates than Lending Club. HR-grade (high risk) loans at Prosper commonly have a 31% interest rate and a 20% default rate, while the riskiest Lending Club G-grade loans have a 25% interest rate with only a 10% default rate.

As a side note, these higher interest and default rates have allowed lenders like myself to earn a higher return at Prosper (see my 2013Q2 returns).

What is the Annual Default Rate for Completed Loans?

A different number comes up when we look at completed loans only, highlighting the percentage of principle that is eventually lost. For Lending Club, this is 7.7%, and for Prosper it is 6.2%. For the Lending Club number, I went again to the NickelSteamroller tool and checked the box for Process only completed loans (with the same loss factors and Issued Start Date as before).

NSR July 13 Default Rate Completed Loans

Prosper Stats July 13 Default Rate Completed LoansThen I did the same thing for Prosper at Prosper-Stats (setting the date range from 2009-07-01 to 2010-07-24 to isolate their completed loans). It is interesting to point out how low the rate is for completed loans (6.28%) in comparison to the overall annual default rate (8.65%). In my opinion, Prosper’s default rate for its completed loans will probably climb in the coming year. More and more loans default over time, so Prosper’s completed loans should naturally have a higher default rate than uncompleted ones.

What Percent of P2P Loans Eventually Default?

Finally I examined what percentage of completed loans end in default. To get the previously mentioned figure for Lending Club and Prosper, I went to the NickelSteamroller Default Rate chart and averaged the rate of both platform’s completed months. Lending Club has eighteen months of completed history (Jan ’09 – June ’10) and Prosper has ten (Sept ’09 – June ’10). Note: I trimmed Prosper’s July/Aug 2009 since they seemed unrepresentative.

Averaging these completed months had 12.6% of Lending Club’s loans and 14.5% of Prosper’s loans eventually defaulting. See below: for loans solely issued in June 2010, about 14% of Lending Club’s defaulted compared to 16% of Prosper’s.

NSR - Lending Club and Prosper Defaulted Loan Numbers

Interestingly, this data seems to say that 13% of peer to peer loans have historically failed to be paid in full, perhaps typical of unsecured 3-year loans anywhere.

Conclusion: Reinvest Those Returns

Hopefully all this talk of figures and percentages gives you a better picture for how defaults factor into investing at Lending Club and Prosper.

If there is any big lesson to take from all this data, it is the importance of reinvesting your returns. While the default rate for completed loans is high, the overall default rate is much more reasonable (7.7% versus 3.0% for Lending Club). Make sure to log into your account and invest any available cash that has built up, and you can be sure to keep your default rate at a much more tolerable level.

[image credit: Richard Taylor "Sunken boat at sunset" CC-BY 2.0]

Comments

  1. NYD3030 says

    Overall I’m a proponent of p2p lending and I have a small investment in it, but your analysis begs a question for me…

    Given the low annual rate of default, but high percentage of defaults for completed loans, and your subsequent advice to continue investing returns… in some way, isn’t this like a ponzi scheme? A high return is dependent on constantly purchasing new notes, because those immature notes have not had time to default.

    When one exits the accumulation phase of investing and begins withdrawing one’s earnings to live, isn’t the rate of return going to plummet as there are fewer young loans in the portfolio to counteract the aging loans going into default?

    In other words isn’t a high rate of default on completed loans going to get us in the end?

    • says

      Hi NYD. Thanks for the great question. On the one hand, yes you’re right. Many lenders have a higher return when they they invest their initial large chunk, and this return goes down some as they simply reinvest the returns. On the other hand, it is very possible to avoid this eventual default rate I reference in the article.

      Perhaps its better to think of p2p lending less like investing in loans and more like investing in a fund. You can enter and exit, to some degree, whenever you wish, selling loans on the secondary market for a markup whenever you leave, so really there’s no point where you would simply stop investing, and thus there’s no point where you would simply let funds go idle and experience the eventual default rate.

      I can only point at my own returns as evidence for this working. I continue to earn higher than a 10% return, year after year, because I continue to reinvest my returns into quality low-grade loans. Peter over at LendAcademy has been doing this even longer than myself with the same result.

      • Tony says

        Hi Simon, I came across your useful article and appreciate the information and data. Thank you for sharing. However, your conclusion and response to NYD’s comment are rather disconcerting. NYD is exactly right that there is an unrealized loss that is not being shared when you make claims of 10%+ returns.

        Yes, you have made 10% returns over your recent years of p2p lending. However, your returns are only that high because you have reinvested your returns and added additional outside funds. This strategy keeps your average loan age low and therefore skews return numbers in your favor by owning a greater portion of loans that have not yet defaulted (key word: YET). Every loan has to end in three ways: charge-off, full payment, or sold on the secondary market. Your response to NYD is deceiving because one has to exit every loan at some point. To assume you can just “exit” at any point is true, but it comes with a significant loss on the secondary market that should be accounted for when you report your “more than a 10%” return. If the secondary market was such a great way to exit loans before they default, why wouldn’t we do it with all of our loans? P2P lending sites skew their numbers to a lesser extend by issuing more loans than the previous month. This keeps their average loan age lower and their reported returns higher.

        I’m an advocate of p2p lending, but I feel that articles like yours and Peter’s (as well as the p2p websites) are giving false expectations to those interested in p2p lending. I think a more honest way to report returns is based on completed loans ONLY since every loan has to complete. This would provide a more accurate assessment of what a lender could expect to return on any given loan. Alternatively, you could discount the value of all your loans to take into account that 13% percent of your current loans will default (or more if you’re buying low-grade ones).

        I’ve used nickelsteamroller.com to look at historical returns of older, completed low-grade loans, and the returns are more in the 8% magnitude. That is what I expect to make in the p2p lending market for quality low-grade loans. This assumes that even if I use your solution of reinvesting returns (which I do), I will someday have to realize those growing unrealized losses (as my account grows)–either by riding out my loans to the end or selling them on the secondary market at a discount.

        • says

          Hi Tony,
          Thanks for your long comment. Very insightful. Here are some articles which uphold my views on what you say:

          1) It seems possible for some investors to sustain 10% returns, though this will always be the minority of investors: http://www.lendingmemo.com/lending-club-prosper-return-curve/

          2) My return is above 10% and my note age is not that young: http://www.lendingmemo.com/returns-2013-q4/

          3) I think it is fair to say that many people can liquidate their account at par: some with 3% markup, most even, some with a 3% discount. I liquidated one of my Lending Club accounts (this post will be published soon) and overall lost very little principal.

          As to your point, I agree that an 8% return is expected, and quite amazing in this market. I agree that the bulk of investors will probably be around this number (http://www.lendingmemo.com/peer-to-peer-lending-return/). However, if you take advantages of different strategies, even you could amp up your returns to some degree as I have done and others before me have done as well.

  2. Jeff says

    Good article. I work in consumer finance and a lot of people there don’t understand the importance of annualizing loss rates when comparing them to annual interest rates (same goes for fees).

    One note: your eventual default rate methodology aggregates all available “loan-months” equally. This is roughly accurate but is distorted in a young and rapidly growing portfolio, which both of these companies are. The problem is that young loans default less often, which means that if you have a predominantly young portfolio, you are understating losses because the remaining loan-months are, in aggregate, different (worse) than the observed ones.

    The methodology to account for this is called a “static pool”, and it accounts for the fact that the different months of a loan will have different performance (and that different loans have had different numbers of months elapsed). The basic effect is that older performance records (i.e. a month 30 payment) get more weight, since there is less data for these equally important months. This is very similar to your method in mature portfolios, but very different in young ones.

    I’ve done a static pool for Lending Club that I don’t want to share with the world, but based on charge-off date, the “breakeven” point where losses get exactly on track with their linear trend is around month 11 (for 36 month loans). Before that, loans are over-performing their final outcome, and for a long time after that they are underperforming their final outcome (both very young and very old loans have low rates of default, the worst default periods are in the early-middle period).

    Overall, nice analysis.

Leave A Reply