Free Updates

Exclusive content to your inbox for FREE!

Analysis of Prosper’s Top 50 Lenders

by Peter Renton on May 3, 2011

I have spent several hours in the past week looking at the top investors on Prosper. One of the truly great things about Prosper from an investor perspective is that you can see the returns of your fellow investors. Now, most likely you won’t know who is behind these investments but you can learn plenty of details about how they invest.

This is all thanks to third party sites like Lendstats and Ericscc that take advantage of the way Prosper records the information on their database. For my analysis I looked at the top 50 lenders on Prosper who have invested in at least 200 loans and who have an average loan age of at least 9 months. I am also only looking at loans on Prosper 2.0 (those loans that originated after the SEC imposed quiet period that ended in July 2009). Here a link to this query on Lendstats. Ericscc also has a tool to analyze top Prosper lenders but I used the Lendstats tool for my analysis here.

High Yielding Returns Here

The ROI for the top 50 lenders according to Lendstats range from 22.4% down to 13.7% on the low end. Every investor here is earning above average returns and some are doing extremely well. One useful feature in Lendstats is it shows the criteria (on average) these investors are using when selecting which loans to invest in. This is what I have found most interesting.

Here are the typical criteria for the average top investor on Prosper:

Average ROI16.98
Loan Age10.9
Average number of loans497
Average loan size$6,283
Open credit lines8.8
% Credit Utilization59.4
Delinquency past 7 yrs29.1%
Current delinquency17.8%
Public record past 10 yrs22.2%
Annual borrower income$61,911
Number of inquiries1.6
Debt-to-income 29.0%
Revolving debt$17,311
Average credit history14.9 years

So what can we learn from these numbers? Well for starters, delinquencies appear to be less important than I first thought. I have been selecting loans based on a zero tolerance for delinquencies but by doing so, I could well be missing out on some good borrowers. Without exception, every one of these top 50 lenders allowed some borrowers with delinquencies (both current and 7 year) within their portfolio. Other factors I also found interesting were the high public records percentage, the relatively high borrower income and high number for credit history length.

If you don’t understand some of the terms in the above table you should do some research on interpreting credit reports. Here is a good place to start.

The Focus is on High Risk Loans

It is stating the obvious, but these investors are focused on the high risk loans, Prosper Ratings of C and below. These are the loans that offer the highest returns, but also the highest risks of defaults and every investor here has experienced several defaults. While these defaults have certainly impacted their ROI, because they have kept defaults to a minimum, (7% on the high end down to 1.3%), their returns have remained high despite these defaults.

Now, these loans, at an average age of 10.9 months, are still young. But they are old enough that the unscrupulous borrowers who deliberately default on a loan have been weeded out for the most part. While the ROI of these investors will likely continue to decline on these loans, I think it is fair to say as a group these investor returns will stay well above the average.

After looking at all these loans I would like to point out one lender who I am keeping an eye on. This is sweety075 (hat tip to Ken at Lendstats). She (I am assuming this lender is a she) doesn’t have the best return but she does have the best default rate among the top 50 investors. With 306 loans invested in Prosper 2.0 loans, she has a default rate of 1.3% with just four defaults. She is investing in loans with an average interest rate of 23.1% but with just four defaults she is enjoying an ROI of 20.61%.

I am tweaking my investment criteria going forward based on my research here. So, what do other investors think? Happy to hear reactions, positive or negative.

{ 23 comments… read them below or add one }

Dan B May 3, 2011 at 12:14 pm

I don’t know………..I’m looking at my Prosper acct. & Ericscc says my return is 20.4%, while lendstats says 6.2%, so I’m going to go with Ericscc since it sounds better to me. :)

Reply

Peter Renton May 3, 2011 at 12:48 pm

@Dan, I think both sites struggle to estimate ROI accurately for new accounts. I used to have a wide difference between the two, but it is getting better. For my account (with an average loan age of 5.4 months), it shows on Lendstats at 26.1%, on Ericscc at 25.4% and my real world return based on my XIRR() calculation has me at 21.9%. But it is still very early days (every note is current with three fully paid) so I am not reading too much into it. Although I am putting new money to work at Prosper (as well as Lending Club).

Reply

Wiseclerk May 3, 2011 at 3:58 pm

Well, since you restricted the search to post-relaunch loans, I doubt that this tells much about the expected performance.
This includes too many loans that are too young to be able to default.
Try the same search with setting ‘use all loans’
and you get a very different picture:
http://www.lendstats.com/lenderlists/lenderlists.php?crit=roi&order=ASC&type=lender&age1=9&age2=&loans1=200&loans2=&lent1=0&lent2=&rem1=0&rem2=&pr2=&pr1=&cg2=&cg1=&roi1=&roi2=&pronly=0

Reply

Peter Renton May 3, 2011 at 4:14 pm

@Wiseclerk, Thanks for your comment. I fully realize that the numbers are very different if you include all loans. The reason I chose post re-launch loans only is that Prosper 2.0 is a very different animal from Prosper 1.0 as I am sure you would agree. If we include all loans, I feel like we are comparing apples and oranges.

I disagree that too many loans are too young to be able to default. With an average loan age of 10.9 months most of the loans here will have had several months where they could have defaulted. From analysis others have done we know that more loans default in the first six months than in any other six month period. Will there be more defaults? Absolutely. But I stand by these numbers and my prediction that most of these investors will outperform the averages.

Reply

Dan B May 3, 2011 at 6:40 pm

I’ve only been on Prosper for about 20 minutes so I really don’t know…………..Is there a consensus that the first 1 1/2 yrs of Prosper 2.0 has produced substantially better results than the first 1 1/2 yrs of Prosper 1.0??

Reply

Peter Renton May 3, 2011 at 9:22 pm

Without question. Just look at this chart from Fred93 that I have shared before. While it is not up to date you can see the steep curve of the loans from the first year and a half. Contrast that with loans from July 2009 through June 2010 and you have a total of 6.1% defaults and 3.9% 30+ days late. I would guess these numbers are a third what they were in Prosper 1.0.

By reading back through the many blog posts written about Prosper circa 2007 and 2008 most investors were very unhappy. People were losing money left and right, you just don’t see that with Prosper today. I am still amazed they made it through intact from that time period.

Reply

Dan B May 4, 2011 at 1:01 am

Ok, I see that it is a massive improvement. But still a combined 10% defaulted & late loans total for notes whose average age is likely 5 months+/- is hardly ideal.

Then again, given the high interest rates involved………..it might just be palatable. I don’t know, I’m going to have to think about this more. Any thoughts ?

Reply

Peter Renton May 4, 2011 at 9:25 am

@Dan, The percentages I am referring to above were for loans originated from July 2009 through June 2009 so they would have an average age of around 15-16 months now.

I think you hit the nail on the head with your comment about the high interest rates. I would say a 6.1% default rate with a 15 month loan average is pretty darn good. Lendstats puts the ROI on all these loans at 8.9%, so with a little filtering it would not be hard to achieve 12% or more. BTW, this ROI goes to 11.9% on all loans in the C – HR ratings (with an 8.7% default rate). These are the loans I am focused on with my Prosper investments.

Reply

Dan B May 5, 2011 at 4:27 am

I’m really confused. When I go to Lendtstats I see that since 7/09 there were 10260 loans issued at Prosper. But when I add up all the loans from AA to HR I come up with a total of 3852. Add loans that were paid off early & loans issued but not yet performing & we’re up to 6110. Add defaults of 330 & all the “lates” & were only at 6790. Where are the remaining 3400+ loans hiding??

Reply

Peter Renton May 5, 2011 at 9:17 am

@Dan, How are you adding up all the loans from AA to HR? I just did the same thing where I went through each rating individually and the numbers all added up for me.

Reply

Dan B May 5, 2011 at 11:33 am

@Peter………..I just used the numbers that are supplied on the performance summary page for a total number of loans after 7/09……..which is listed as 10260. Then clicking on AA thru HR individually & adding them as I described above.

Reply

Peter Renton May 5, 2011 at 12:06 pm

OK. I think I know what the problem is now. When you click on AA through HR on Lendstats you are not selecting all loans, you are selecting a filter of loans based on what Ken has considered to be criteria that will produce the best return.

For example, when you click on AA, you get 550 loans generating an ROI of 7.18% (as of right now). But these are a subset of AA loans, not all loans. If you look at all the input fields near the top of the screen you will see the various values that Ken has worked out will give the best ROI. The A through HR are just examples to get you going. I have used them as starting points but ended up using my own criteria. I hope that is clear. One day I am going to produce a video about Lendstats because it is not intuitively obvious how to use it.

Reply

Dan B May 5, 2011 at 4:18 pm

Thanks Peter, that must be it.

Reply

KenL May 5, 2011 at 8:46 pm

Hi Peter hi Dan,

Yeah, you have explained everything pretty well. Yeah, on the home page those are filters. I have made a separate filter for each Prosper Rating and I named them as such. I really should do something to clear that confusion.

Wiseclerk, I believe you can make quality estimates of ROI based on young loans. In order to do that though you, need to estimate losses on all delinquent loans, and LendStats does that. Losses are estimated on all loans from 1 to 120 days late and these estimations are summed with the defaults.

Ericscc has at least a few problems with its ROI calculation. First of all it counts no losses on loans which are 1 to 30 days late. Second, ericscc does not calculate interest earned, it estimates it based solely on age. This works fine, until irregular or early payments occur. Thirdly, Ericscc is using the wrong value of principal balance on many of the defaulted loans which results in sometimes wildly inflated ROI. Just take a look at the lender CountOnMyBids and you’ll see what I mean.
http://www.ericscc.com/lenders/countonmybids
and at lendstats CountOnMyBids’ ROI is completely different
http://www.lendstats.com/loansearch/loanfilter.php?type=lender&lender=countonmybids

Erics has me at 15% and Lendstats has me at 5.2%. 15% sure would be nice, but if it were true I’d have over $20k more in earnings, and even though I look to try and find that extra $20k I just don’t see it anywhere, lol. I know it’s a downer, but my overall ROI really is only 5.2%

Reply

KenL May 5, 2011 at 9:46 pm

I just want to add that there are a couple other criteria which do not show up in the lender profiles at lendstats (yet) that top lenders may use. These are whether borrowers are also lenders (blenders), and repeat borrowers.

At Prosper there are a lot of borrowers who are also lenders, sometimes these are the most sure thing you can get and they sometimes have high rates. You can find a list of blender listings here http://www.lendstats.com/blenders.php

There are also a ton repeat borrowers, they also perform very well if they have an extended borrowing history. At lendstats on the loan filter page there are 3 repeat borrower filters, called PB1, PB2, and PB3 and they work great.

Reply

Dan B May 6, 2011 at 5:42 am

@KenL…………Are you saying that there are Prosper borrowers who are borrowing at high interest rates & that they are also lending at Prosper…………… at presumably lower interest rates than they are borrowing at?

Reply

Peter Renton May 6, 2011 at 8:35 am

@Ken, Thanks for chiming in. I need to point out, though, that you are being modest. While your overall return at Prosper is 5.2%, in the post SEC registration time period you are killing it with a 20.7% return. This compares somewhat favorably with the 0.5% you received in Prosper 1.0 :-).

I noticed the PB filters, as in previous borrowers, and I am planning on doing a blog post soon based just on that. Previous borrowers who have never had a late payment are giving some of the best returns you can get. I hope Lending Club adds that as a selection criteria soon.

Like @Dan, I would like some clarification on what you meant about the blenders. If a blenders is borrowing at a high rate, wouldn’t that really hamper their ability to lend profitably?

Reply

KenL May 6, 2011 at 8:47 am

Yes, there are some. Usually the lenders have pretty good credit and get decent rates, but sometimes there are lenders who have bad credit and can’t get such a good rate. Blenders also do tend to pay the loans back faster than most other lenders.

Reply

KenL May 6, 2011 at 9:08 am

Thanks Peter, yes by all standards I’m really kicking butt in 2.0 Prosper. I still actually can’t believe how well I’m doing. And I’ve got enough lending experience to realize the returns I am getting are real. They aren’t going to come crashing down in the next weeks or months. I do however expect my returns to come down some, but not much. After funding 241 defaults with Prosper 1.0 I know first hand that defaults happen and now I’m managing my risks/rewards as accurately as I can.

Reply

KenL May 6, 2011 at 9:16 am

About the blenders. Most who are borrowing are not borrowing to re-lend. The loan is usually for some other need. Since they usually have a good understanding of how interest works, they tend to pay loans back faster than other non-lender borrowers.

Reply

Peter Renton May 6, 2011 at 12:04 pm

@Ken, Thanks for the clarification. I am still not convinced blenders are a good risk for the money. Do you have any numbers to back up your hunch? I don’t see any performance related data about blenders on Lendstats. Of course, we would easily be investing in a blender and not know it.

Reply

KenL May 6, 2011 at 3:27 pm

@Peter, no numbers, yet. But I’m working on it. My biggest default (from 1.0) was actually a blender. $1770 lent and I didn’t get a dime back. This was someone who must have been disillusioned and wanted to recoup his money and then some. I remember bidding on it and realized I made a big mistake after the fact. This lender was getting horrible returns and stopped lending a full year before I bid on him. However due to lack of quality information, I didn’t realize his returns were so bad. These days, I generally bid on blenders who are getting good returns.
his listing:
https://www.prosper.com/invest/listing.aspx?listingID=394170

his lender profile at lendstats
http://www.lendstats.com/loansearch/loanfilter.php?type=lender&lender=youngmogul

I made that bid long before lendstats existed and largely because of this mistake I designed the blender page at lendstats so that I would never make such a mistake again.

Reply

Peter Renton May 6, 2011 at 3:58 pm

@Ken, Curious. I can see how an investor might be angry about his returns on Prosper 1.0 (like many of the folks on Prospers.org), but that is an interesting way to try and recoup your money. Certainly someone without a high standard for ethics.

I read his loan listing and he certainly talked up a believable story, I guess that came from reading all those loan descriptions as an investor.

Reply

Leave a Comment

Notify me of followup comments via e-mail. You can also subscribe without commenting.

{ 1 trackback }

Previous post:

Next post:

Real Time Analytics