Wednesday, June 8, 2011

Prosper.com - Series E financing analysis

A few days ago, Prosper.com received another venture capital investment.

Investors paid $0.739/share, a slightly higher price than last year's round, but still far far lower than the $9.692/share Prosper raised in 2007.

Here's a quick history of Prosper's venture capital.

04/2005, Series A, 4,023,999 shares for $7,464,450, or $1.875 per share.
02/2006, Series B, 3,310,382 shares for $12,412,301 or $3.776 per share.
06/2007, Series C, 2,063,448 shares for $19,919,009 or $9.692 per share.
04/2010, Series D, 20,340,705 shares for $14,700,000 or $0.723 per share.
06/2011, Series E, 23,222,747 shares for $17,150,000 or $0.739 per share.

Prosper paid an agent $375,000 to raise this money. If we subtract that commission out of the $17,150,000, then Prosper received net $0.722/share for this round.

In recent months, Prosper has been burning (losing) about $1 Million/month. (For example, see the Prosper 10Q filing for the quarter ending 3/31/2011, and leave out the nonrecurring items, such as insurance recoveries. They lost about $3 Million for the quarter.) Prosper's future depends on reducing that burn, eventually breaking even and then becoming profitable. My guess is that with fresh money in hand, they may initially increase the burn rate, as they put money to use for various marketing programs to try to drive the loan origination rate up.

As always, great discussion among P2P and Prosper.com lenders is found on prospers.org. I hope to see you there!

Monday, August 2, 2010

Prosper.com - 07/2010 Late Loan Stats Update

Here's the July 2010 update to to my Prosper.com late loan statistics charts. I was unable to update these charts in May and June, because Prosper had the data offline at that time. The data's back, so here we go...

These charts show statistics for the performance of all prosper.com loans. Each curve represents the set of loans that were created in one calendar month. The vertical axis is the fraction of those loans that have "gone bad", in other words are 1 month late or worse (up to and including default or "charge off" as it is now called). The horizontal axis is the observation date. All data comes from Prosper.com's performance web page.

The worst month so far is still February '07. Of the loans originated by Prosper.com in February '07, 45.3% have now gone bad. Yipes!

More detail can be found in my earlier posts.

Click on the chart to see a larger clearer version.


I terminate each curve when all the loans in that cohort are "done", in other words paid off or charged off. You may notice some of the curves extend to the right farther than you would expect. That's because I'm following the Prosper data.

There are some quirks. For example, I would have expected that loans originated in Dec '06 would all be "done" by now, but the Prosper database shows that they are not. They hold 5 of the Dec '06 loans still in the "30 to 120 days late" category. These are 36 payment loans, with their first payment in Jan '07, and last payment in Dec '09. Prosper holds late loans until they are either paid off or 4 months late. If one of these Dec '06 loans were late on its last payment, one could imagine it might be held thru Jan, Feb, Mar, Apr of 2010, but then it should be charged off. Dec '06 is not the only month containing such anomalies. Not the first quirk in this data, and I'm sure it won't be the last.

Here's a chart of the same data in which each curve has been slid to the left to a common origin. The horizontal axis is now days since loan origination month. Because Prosper's loans changed character when Prosper relaunched in Aug '09 with new credit score limits etc, I have separated this data into two charts, one for the "old" loans Oct'08 and earlier, and a second for the "new" loans Aug'09 and later.



For easy visual comparison, I included two of the "old loan" curves on the new loan chart below. These are approximately the best and worst months of the old loans, so you can see the range of slopes occupied for comparison. The new loans have lower default rates mostly because in Aug'09 Prosper raised the minimum credit score required to apply for a loan.



In July Prosper made several small adjustments. They revamped the "Prosper score" again. This is both good and bad news. No doubt the new score design is an attempted improvement, but the constantly changing baseline makes the score almost useless to lenders. Prosper also raised the origination fees on several credit grades of borrowers. C thru HR credit grade borrowers are now charged a 4.5% up front origination fee for a loan! Many lenders have questioned the wisdom of that price hike. Might increase Prosper's income, but not enough to save the company. To save the company, they have to increase loan volume.



Explanation of methodology can be found in my prior postings in this blog, and in forum discussions on the old prosper forum, now archived at www.prosperreport.com

Many of the very early posts in this blog are still on point, and provide background on prosper, from a lender's perspective. If you're new to this, please read old posts before sending questions.

PS: The very best discussion among P2P and Prosper.com lenders is found on prospers.org. See you there!

Saturday, May 22, 2010

Prosper.com - no more stats

I'd like to update you on Prosper's late loan stats, but I can't.

On May 3rd, Prosper installed new software to process loan payments. There seem to be several problems with the new software.

On the same day, they have removed the Prosper loan performance data from the loan performance statistics web page. The page is still available, but since 5/3/2010 it has displayed the message "Marketplace data temporarily unavailable".

(In the meantime, you can refer to my most recent Prosper loan performance charts.)

Questions from lenders about when the loan performance data will reappear have gone unanswered. (One question from a lender posted on the prosper blog got a response saying the performance data will return, but not when. My own email to prosper on this subject has gone entirely unanswered.)

Meanwhile, the new payment processing software appears to be quite a fumble. In the last three weeks dozens of members have noticed anomalies in payment processing and have notified prosper about same. You can see these discussed in the comments section of the prosper blog, or on the prospers.org forum. It seems likely that most people don't look at the numbers in detail, so would not notice payment processing errors. If dozens are reporting errors, it seems likely that the snafu is widespread. If you have noticed such problems yourself, please contribute to the discussion on the prospers.org forum.

I noticed one myself, and emailed Prosper on 5/10/2010. I wrote:

I'm looking at note 8324-36. Looking at the note detail, I'm having trouble with some of the numbers.

I don't understand the "principal balance" column on the "borrower payments" tab.

The first April-12-2010 payment attempt drops the principal balance from $80.32 to $47.04, which looks about right. Later on April-12-2010, the payment is reversed, which jumps the principal balance to $142.79 . This is impossible.

Later on April-12-2010, another payment attempt is made, dropping the principal balance to $48.59 . At first I wondered why it didn't come back to $47.04 . Then I realized that the difference between these values is two late fees. But why is the guy charged two late fees for one payment?

Finally, I compared the principal balance column on the "lender accounting" tab to the principal balance column on the "borrower payments" tab, and found that they do not agree! One thing that borrower and lender must always agree upon is the principal balance!

Please advise.

I got a response thanking me, but the problem has not been fixed.

I suggest to all lenders that you look over the arithmetic in any loan where there have been late payments, or in which payments have been reversed. Look carefully at late fees. Look to see after a failed payment is reversed whether the principal comes back to what it was before the payment. Look to see whether the borrower tab and the lender tab agree on the loan principal.

Obviously Prosper installed this new software without adequate testing. Their recovery from this fumble is not encouraging. Have all the good software people left the company?

Meanwhile I wonder if the loan performance data will ever return. Seems quite possible that it is gone forever. This is a sad situation.

I remember three years ago when we had a similar situation on Prosper's collections performance data. I repeatedly pointed out that there were obvious errors in the data. Prosper took the collections data offline with the repeated promise that it would return as soon as some software which interfaced to the collections agency was repaired. Later the web page containing the collections data was removed, and the collections performance data was gone forever.

The best discussion among P2P and Prosper.com lenders is found on prospers.org. I hope to see you there!

Tuesday, April 20, 2010

Prosper.com - Series D financing analysis

Prosper has just announced its fourth equity capital round, series D. What follows is a brief analysis of this investment round based on public information.

Prior to the series D round, prosper had 14,752,825 shares of common stock outstanding. (This info from page 82 of Prosper's March 31, 2010 SEC form 10k.)

In the series D round, prosper issued 20,340,705 new shares of common stock for $14.7M. That comes to $0.723 per share. (This info from Prosper's April 20, 2010 SEC form 8K.)

The first thing to notice is that prosper was forced to give away 58% of the company to raise this money. That must have hurt.

How does that price per share compare with prior investment rounds? Here they are for comparison...

04/2005, Series A, 4,023,999 shares for $7,464,450, or $1.875 per share.
02/2006, Series B, 3,310,382 shares for $12,412,301 or $3.776 per share.
06/2007, Series C, 2,063,448 shares for $19,919,009 or $9.692 per share.
04/2010, Series D, 20,340,705 shares for $14,700,000 or $0.723 per share.

In the venture capital business, this is called a "down round". That's one hell of a jump down! From $9.692 to $0.723 is down by a factor of 13.

Although Prosper announced the series D event as a $14.7M round, it brought in only $11.4M of cash. That's because some of the shares in this round were paid for not with cash, but by conversion of the principal and interest Prosper owed on the bridge loans Prosper took out in Nov 2009 and Feb 2010.

If Prosper has continued spending at the rate reported for the quarter ended Dec 31, then it should have about $12.3M in the bank today. I now expect Prosper to begin spending at a greater rate than the December quarter, as it tries to rebuild momentum. If spending increases back to the rate of the quarter ended Oct 31, and revenue stays about the same, then this money would last about a year or so.

Of course I don't know what the future holds.

As always, great discussion among P2P and Prosper.com lenders is found on prospers.org. I hope to see you there!

Wednesday, March 31, 2010

Prosper.com - backloaded months

For a long time now, Prosper lenders (those members who lend to others thru Prosper), have observed that a great many Prosper.com loans originate in the last 3 or 4 days of each month. Each month is "back loaded".

The most often discussed hypothesis is that Prosper rushes at the end of each month to push thru any loans that can be pushed thru to make the best possibly monthly numbers. If so, this would be bad for lenders, because Prosper has a number of duties to perform before a loan originates, and lenders want those duties carried out with care. Prosper is supposed to check the borrower against various fraud databases, receive the post card from the borrower verifying their address, and selectively verify other information the borrower has provided. These checks are an important part of fraud protection on which lenders depend. If Prosper is skipping or skimping on these checks during the mad rush at the end of every month, then we'll have a bunch of lower quality loans in the system, and more defaults.

So do we have an end-of-month rush?

I've charted the daily loan originations for the last 5 months vs day of month. Although the chart is a bit busy, you can plainly see that each month is indeed back-loaded.



Now there are several problems with the chart above. First, each month has a different number of days, so the end-of-months don't line up. Second, there are weekends and holidays scattered around, and no loans originate on those days.

To fix these problems, I squeezed out the weekends and holidays, and then slid all the months over so they aligned on the last day of the month, then averaged the five months shown above. The result is this much cleaner chart, showing business days only, and with month-ends aligned.



Yep. I'd say Prosper's months are very much back-loaded. It really does look like the boss walks in when there are about 3 days left in the month and starts yelling that everything must get out the door by month end.

This is the end of another month, and I've updated my loan growth chart. This chart shows Prosper's loan origination by month, so we can see how the business is growing. February was larger than the preceeding month, but still below November '09, and of course well below anything in 2007 or 2008. The last few months have been pretty much a "no growth" scenario.



Today Prosper announced that they're negotiating with investors for a new investment round, which of course we all expected. In fact they must get this done quickly, because they are nearly out of cash. (See my last few writeups for calculations.) It will be interesting to see the details (ie how much did they have to give away) when they emerge in a few weeks.

Tomorrow is the SEC's deadline for the filing of Prosper's annual report (form 10K). At that time we'll learn a bit about how Prosper has managed their cash during October, Nov, and December. (Unfortunately, this document won't tell us anything about January, February, March. )

The best discussion among P2P and Prosper.com lenders is found on prospers.org. I hope to see you there!

Tuesday, March 2, 2010

Prosper.com - 02/2010 Late Loan Stats Update

Here's the February 2010 update to to my Prosper.com late loan statistics charts.

These charts show statistics for the performance of all prosper.com loans. Each curve represents the set of loans that were created in one calendar month. The vertical axis is the fraction of those loans that have "gone bad", in other words are 1 month late or worse (up to and including default or "charge off" as it is now called). The horizontal axis is the observation date. All data comes from Prosper.com's performance web page.

The worst month so far is now February '07. Of the loans originated by Prosper.com in February '07, 45.3% have now gone bad. Yipes!

More detail can be found in my earlier posts.

Click on the chart to see a larger clearer version.


Here's a chart of the same data in which each curve has been slid to the left to a common origin. The horizontal axis is now days since loan origination month.



Explanation of methodology can be found in my prior postings in this blog, and in forum discussions on the old prosper forum, now archived at www.prosperreport.com

Many of the very early posts in this blog are still on point, and provide background on prosper, from a lender's perspective. If you're new to this, please read old posts before sending questions.

Not much news from prosper.com this month. Lenders are waiting to see Prosper's year-end SEC filing (form 10K) which should issue any day now. This will give us an update on how much cash Prosper has left. Based on the September 30'th quarterly SEC filing, I estimated that Prospser has enough cash to live until April 1, 2010. Some time in the next few days the 10K will issue, and we'll learn how Prosper's management has been managing its cash. If they've been frugal, perhaps they can last a bit longer, or ...

As we discuss the possibility of Prosper's next round of venture capital, it is meaningful to look at the scale of Prosper's business, and how it is growing. A new investor in Prosper's business wants to see growth of course. Prosper's business is not sustainable unless it is able to grow to something like 10x its present size, so growth is absolutely essential. The last few months have been disappointing. Where is the growth? This is sad to see. This company once had enviable buzz. Now their marketing is a horrible failure.



PS: The best discussion among P2P and Prosper.com lenders is found on prospers.org. See you there!

Thursday, February 4, 2010

Prosper.com - bankruptcy postponed

Bankruptcy has been postponed.

At the 11th hour, on February 1, 2010, a group of 6 venture capital firms injected a small amount of money ($2 million) into Prosper.com in the form of a bridge loan. The terms of this loan were (as far as has been disclosed) identical to the terms of the $1 million investment from Nigel Morris on November 10th. These loans will convert to common stock when (and if) someone comes along with the next major chunk of money in the form of an equity investment.

This is a small amount of money on the scale of Prosper's operation. If we assume, as I did in the last post, that Prosper continues to burn cash at the same rate they did between the June and September quarterly reports, this new money is sufficient to keep Prosper operating until April 1st 2010 -- a relatively short time from now. Prosper's management will be working hard to find the next investor before this time is up. I wish them well.

To have a chance of growing to a sustainable revenue, Prosper needs an injection of $20 to $50 Million. Given the difficulties the business has had so far, it seems unlikely that this money will appear in chucks larger than $5 or maybe $10 million at a time. Management has lots of work to do.

While we're on the subject of sustainable business, lets look at how Prosper's loan orginations have grown with time. I estimate that Prosper needs $50 Million/month in loan originations to be sustainable. During the first year, volume grew nicely. After that there were a number of problems, including the SEC shutdown in late 2009.





Unfortunately, Prosper's loan volume after the restart have been comparable to those first months of 2006. Momentum died during the shutdown.

During the first two years Prosper's loan volume was driven largely by PR. In other words, you could see that borrowers came to Prosper after reading stories about Prosper in magazines and newspapers. Every time such a story appeared, you would see a spike in listings. PR is great for a new company, but it isn't sustainable. In the early days Prosper also managed to create an incredible buzz from a highly motivated community. That momentum was killed by Prosper's arrogant attitude toward the community.

For prosper the marketing problem is how to attract growing numbers of borrowers in a sustainable way. (Without PR and without a community?) Prosper has some new folks on board to try to sort this out. Now we get to sit back and see if they solve the puzzle.

And of course if the last 3 years have taught us nothing else, they must attract the right kind of borrowers -- those who will pay you back.

Great discussion among P2P lenders is found at Prospers.org .

Friday, January 29, 2010

Prosper.com is near bankruptcy

Its no secret. Prosper.com has been hemorrhaging cash. Its all in the public SEC filings.

I'm no accountant, but the numbers really jump off the page. Prosper is near bankruptcy.

Prosper's most recent quarterly report, which reports on the company's condition on 9/30/2009 reported $2,079,624 of cash left. That was down from $4,617,954 in the 6/30/2009 report. During this interval, cash was used up at the rate of $27,591 per day.

If you review the quarterly report, you will note that I have excluded a category called "restricted cash". A footnote explains "Restricted cash consists primarily of an irrevocable letter of credit held by a financial institution in connection with the Company’s office lease and cash deposits required to support the Company’s ACH activities and secured corporate credit cards." Gotta keep those ACH transfers rollin', etc. That restricted cash has to stay where it is. Therefore, I only count the cash that isn't in this restricted category.

Assuming a constant cash burn rate, on 11/10/2009 prosper would have had $948,412 of cash left. On that day, Nigel Morris invested $1,000,000., bringing my estimate of prosper's cash to $1,948,412. But then...

Assuming the cash burn rate stayed constant at $27,591 per day, Prosper would have had zero cash as of 1/20/2010, a few days ago!

I'm not saying prosper is bankrupt now. This is merely what would have occurred under this simple scenario. The inside story is surely more complex.

One would expect that as cash dried up, a prudent manager would reduce staff, defer executive salaries, defer accounts payable (ie pay the landlord late, pay the lawyers late, etc). We don't know whether these things have occurred. If they have taken these prudent steps, then another month or perhaps more is possible.

Simple analysis of the public documents indicates that prosper is in desperate need of cash.

The public documents also tells us about prosper.com's liabilities. The most recently quarterly report lists $434,738 accounts payable, $987,128 accrued liabilities, $281,061 long term debt. Those total to $1.7 million. It seems likely that those liabilities are higher now that it was on the 9/30/09 report date. If so, then zero cash in the bank is really net negative $1.7 million.

Oh but then on 11/10/09, Nigel Morris' invested $1,000,000 in the form of a "bridge loan" (per SEC filings), so that increased liabilities to maybe $2.7 million. That money is surely now spent (see above calculation).

Having zero cash and at the same time $2.7 million in liabilities is quite serious.

You may also notice that I haven't counted the loans or notes (two sides of the same coin) that prosper issued after their restart in 2009. That's because these as far as cash flow is concerned, these two items net out. For every loan prosper makes, there is an equal amount of notes issued to lenders.

I haven't tried to calculate what would happen if a bankruptcy were to occur. I've simply calculated what appears to be happening in the operation of the ongoing business.

There are now two possibilities. We could see a new investment of several million in Prosper during the next few weeks, which would breath new life to their balance sheet, or we could see bankruptcy.

Would be a shame to see this great idea go down. With cash in the bank from a new large investment, and new management, I might even become a customer again.

I'm no accountant, but golly gosh, it sure looks like the situation is coming to a head.

The best discussion among prosper lenders is always found at prospers.org

Sunday, January 3, 2010

Prosper.com - 12/2009 late loan stats update

Here's the December 2009 update to to my Prosper.com late loan statistics charts.

These charts show statistics for the performance of all prosper.com loans. Each curve represents the set of loans that were created in one calendar month. The vertical axis is the fraction of those loans that have "gone bad", in other words are 1 month late or worse (up to and including default or "charge off" as it is now called). The horizontal axis is the observation date. All data comes from Prosper.com's performance web page.

The worst month so far is still October '06. Of the loans originated by Prosper.com in October'06, 44.2% have now gone bad.

Feb '07 is comin' up fast with 43.9% gone bad so far.

More detail can be found in my earlier posts.

Click on the chart to see a larger clearer version.


Here's a chart of the same data in which each curve has been slid to the left to a common origin. The horizontal axis is now days since loan origination month.



Explanation of methodology can be found in my prior postings in this blog, and in forum discussions on the old prosper forum, now archived at www.prosperreport.com

Many of the very early posts in this blog are still on point, and provide background on prosper, from a lender's perspective. If you're new to this, please read old posts before sending questions.

Not much news from prosper.com this month. Lenders have not yet heard from new Prosper executive Nick Talwar.

Prosper needs a cash infusion from venture capital source within the next few months. Difficult to predict with precision, but the deadline must be coming up pretty soon now. Maybe end of January? If they become frugal, perhaps the cash could last a little longer. They must be desperately be searching for funds.

PS: Best discussion among P2P and Prosper.com lenders is always found on prospers.org. See you there

Saturday, December 5, 2009

Prosper.com - 11/2009 late loan stats update

Here's the November 2009 update to to my Prosper.com late loan statistics charts. Toward the end of this article, I have included some comments on recent events at Prosper.com.

These charts show statistics for the performance of all prosper.com loans. Each curve represents the set of loans that were created in one calendar month. The vertical axis is the fraction of those loans that have "gone bad", in other words are 1 month late or worse (up to and including default or "charge off" as it is now called). The horizontal axis is the observation date. All data comes from Prosper.com's performance web page.

The worst month so far is October '06. Of the loans originated by Prosper.com in October'06, 44.2% have now gone bad.

October'06 is not alone. Many other months have produced similarly bad performance.

More detail can be found in my earlier posts.

I've started charting some of the loans originated after Prosper's SEC-sanctioned reopening. There weren't enough loans originated in July'09 to make useful statistics, so I've started with Aug'09. In July'09, Prosper raised the borrower credit score cutoff from 520 to 640, a huge change. With these new restrictions on borrowers, the new loans in aggregate should go bad at a much lower rate (ie lower slopes on the new curves). Too early to know much, but the first couple of points do seem to indicate the new loans doing much better.

Click on the chart to see a larger clearer version.


Here's a chart of the same data in which each curve has been slid to the left to a common origin. The horizontal axis is now days since loan origination month.



Explanation of methodology can be found in my prior postings in this blog, and in forum discussions on the old prosper forum, now archived at www.prosperreport.com

Many of the very early posts in this blog are still on point, and provide background on prosper, from a lender's perspective. If you're new to this, please read old posts before sending questions.

There have been several interesting developments at Prosper.com during the last month.

As many of you know, Prosper is a startup, burning thru venture capital, and near the end of its cash. As a result, Prosper needs a substantial new injection of cash within the next 6 months, or else. It was therefore a welcome development when Prosper announced that on November 11 Nigel Morris, a co-founder of Capital One, had made an investment in Prosper. Examination of SEC filings shows us that this investment was only $1 million, a very small amount, and also that this investment was in the form of a bridge loan, at 15% interest, and came with a warrant kicker and a position on the board of directors. This investment is not the substantial investment we've been waiting for. Its just a bridge. Bridge to what we're not quite sure. Does the appearance of Morris signal the beginning of a change of control?

On November 11, 2009, Prosper filed its official quarterly report (10Q) with the SEC. This latest 10Q shows that prosper is burning about $2 million of cash per quarter, and had about $2 million of cash left as of Sept 30th. (Of course the Nigel Morris $1 million was added after Sept 30th.)

In late November, Prosper's chief marketing officer, Catherine Muriel, left the company. Muriel was an embarrassment. Her only visible work was authorship of many of Prosper's official blog entries, on subjects such as thanksgiving leftovers and recommendations to buy Christmas gifts at spencers online, a company that sells pranks such as tear-proof toilet paper, and lets not forget the four part series on online dating. (See the forum thread Crazy Muriel Blogs Again.)

On December 2, 2009, Prosper announced the arrival of Nick Talwar as their new "chief revenue officer". I had never heard the title CRO before. A little research indicates that this title generally includes responsibility for "revenue". I thought that was the CEO's job. Maybe its just a fancy name for the marketing guy. In any case, investors haven't heard from Talwar yet, so we can't yet judge what he's going to accomplish.

On December 2, 2009, Prosper dismissed its accounting firm, Ernst & Young, and hired the much much smaller firm of Odenburg, Ullakko, Muranishi & Co. Although some may find this sinister, I take Prosper's word that this change was made to save money. I don't know anything about OUM, but you can see some of their smiling faces here.


PS: The best discussion among P2P and Prosper.com lenders are found on prospers.org. See you there

Wednesday, November 4, 2009

Prosper.com - Paradise Lost; Deadbeats Won

It was a hot Saturday afternoon, August 5th 2006. I was driving to Long Beach to have dinner with a bunch of folks -- 25 fellow Prosper.com lenders -- who I had never met.

That's not quite true. I had met them over the internet. For several months I had been vigorously corresponding with these folks (day and night) as we worked hard to decipher the risks and necessary disciplines for lending money to anonymous strangers via the internet. We were spending several hours a day reviewing loans, corresponding with borrowers and each other, doing statistical studies, debating our hypothesis about what makes a good borrower. For example.. Historical data said that "homeowners" are better risks, but would that still be true now that housing lenders had gone nuts and strange people were buying with no money down and flipping homes left and right? There seemed to be a million such issues, and a lot of different opinions on each one.

Was I nuts? I had never done anything like this before. A face-to-face meeting with people I knew only by pseudonyms (their Prosper screen names)?

I drove 100 miles, but some came from much larger distances. Prosper CEO Chris Larsen flew down from San Francisco. One lender flew in from New York.

The get-together was arranged by a fellow lender I knew only as DebinVenice. She gave me an address and said "I'll be standing under a big horse." As I drove past the Queen Mary I wondered if there would be a horse or a woman standing under it or what that even meant. (I had never been to a P.F. Changs, so I didn't know they all have giant horse statues out front.)

Sure enough Deb was standing under the horse with a small crowd when I arrived. I felt like I knew all these people. After all we had spent hundreds of hours talking online. Online we had talked mostly about specific borrowers, and various Prosper technical issues and frustrations but by this time we knew a lot about each-others' politics, family life, etc. We did not however know what each other looked like. Screen names don't give you much of a clue about who is male or female, who is black or white, short or tall, etc, and the mental images I had formed needed considerable adjustment.

One couple walked up to me (seeing the screen name on my badge) and thanked me for bidding on their loan. Wow! Borrowers in the flesh! Unexpected. It seemed very personal.

This was the time of the most amazing excitement for the early batch of Prosper lenders.

I remember sitting across the table from Prosper CEO Chris Larsen, as he explained how tough he was gonna' be on fraud and deadbeats. He was gonna make an example of these folks, so borrowers would know that Prosper is tough.

Chris Larsen, CEO of Prosper.com (center) talking with lenders (folks who lend money to strangers over the internat via Prosper.com) over dinner, Long Beach, August 5, 2006. (Photo credit DebinVenice)

This was pretty much the time of peak excitement. What I mean is that lender enthusiasm went downhill after this time. Those folks around that table were the wildest imaginable supporters and advocates for Prosper.com, but none of the folks I sat with at that table (with the exception of Chris Larsen) are still lending via Prosper. Something went wrong.

To those of us who were actively participating, observing how borrowers interacted with lenders, observing how loans were performing, some problems became clear. The population of Prosper borrowers was obviously heavy with fraudsters and deadbeats. A tremendous amount of animosity developed when specific examples of borrower fraud and dishonesty were exposed by lenders time and time again, and Prosper management just didn't want to hear it. Lenders who made the "mistake" of exposing such fraud were routinely censored and censured by Prosper management. Chris Larsen "shot the messenger" instead of learning about the problem and taking action to fix it.

I've written much about the rate at which Prosper loans are going bad. To get some historical perspective, lets look back at the guidance Prosper gave lenders in those early days. Prosper published statistics which developed from credit bureau Experian's experience with credit card customers. The purpose was to help lenders understand the Prosper credit grades, which at that time were just highly quantized Experian credit scores. This data showed that borrowers in Prosper's "AA" credit grade would be expected to default at the rate of 0.20% per year. What actually happened is that loans to AA borrowers went bad at a rate of about 5.00% per year. Wow! That's 25 times worse than Experian's vast experience seemed to predict! Somehow Experian's credit score and their historical credit experience didn't quite predict the behavior of Prosper borrowers. (Note: You can see how "AA" credit grade loans have performed by looking at rateladder's graphs. Find the 1 year point on the X axis and look at the "AA" curve.)

Experian's experience didn't provide the right insight, from which we conclude...

Apparently Prosper attracted a "different" crowd.

Making a successful business in P2P lending requires that you understand this fact. If Chris Larsen had taken the time to realize what was happening, he might have put in place superior fraud checks, serious verification steps, a serious collections department, etc, as appropriate to deal with the substantial fraction of fraudsters and deadbeats who were attracted to Prosper, but that was not to be.

I'd love to give you the details (ie names) and amazing stories of specific deadbeat Prosper borrowers who stiffed me personally. Anecdotes often communicate better than statistics. I won't, because giving you personal information on these borrowers could be interpreted as violating the terms of the lender agreements I signed with Prosper. One of my deadbeats was a former state senator from New Hampshire who is still routinely quoted in the political press. (Rated "A" by Prosper.) One of my deadbeats wrote a book on how to avoid collection agencies! He sells it on the internet! Man that's a deadbeat's deadbeat. (Rated "AA" by Prosper.) Both these guys are highly visible. Its not like all Prosper deadbeats have skipped town. Some sit right in front of your face and don't pay. You'd think that with highly visible people like this stiffing Prosper's lenders, management would want to take some action. Maybe sue some people. Get serious. No, they just had a collection agency call them repeatedly on the phone. Damn weak effort.

In late 2007, Prosper hired a collections expert, Doug Fuller, as VP of Operations. Prosper introduced this new employee as if he were an attack dog who was going to eat deadbeat borrowers for lunch. On October 2, 2007, Prosper published an "interview" of Fuller by a fellow Prosper employee. This was published on Prosper's official forum, on Prosper's web site, but has since been deleted. Fortunately, some lenders kept a copy, and you can read it here. I'll quote a few snips, but you should read the entire interview for yourself. It paints a strong image.

Q: Do you have experience with suing people?
A: Oh yes. Between First Select and Credigy, I have been responsible for making the decision to sue more than 150,000 people. There are a lot of lawyers that can’t claim that number of suits in a lifetime.

...
Q: Okay, so you need to decide who to sue, then what?
A: Put quite simply, my philosophy is this – if you won’t pay, but can (or will in the future) be able to pay, I’m going to sue you. If I sue you I’m going to win.

Q: That sounds kind of arrogant, can you back it up?

A: Courts in seven states have recognized me as an expert at consumer debt litigation. At Credigy, if a case got really nasty, I would go testify live. I refuse to lose.


Q: Really? What’s your win/lose record?
A: In my last 18 months at Credigy, I testified live at 42 trials. My record was 41-1. By the way, I fired the law firm where we lost.

Big hat, but is there cattle?

In January 2008, Fuller selected 66 loans out of the 1800 Prosper.com loans that had defaulted by that time, and he organized a "legal test" in which he would sue these 66 borrowers, and demonstrate that he was able to produce results like the ones he described above. Fuller personally selected these 66 loans. Fuller told us that he picked loans where he believed he had good contact information on the borrower, and where the borrower had the ability to pay. Presumably he also took care to choose loans where he believed he had the best chance to win. (He did not include my state senator or deadbeat's instruction manual author.)

Although Prosper kept the status of this project secret, many courts publish information about ongoing lawsuits, so lenders were able to observe that the lawsuits were not going well. I wrote about this in October of 2008. What I observed is that many of the suits were being dismissed! We couldn't tell exactly what was going wrong, because courts publish limited information.

Some of the lawsuits were dismissed by the court because Prosper was never able to find the deadbeat borrowers to serve them. Oops. Other borrowers were served (ie were found), but the cases were later dismissed by the court. Some suits were being dismissed at Prosper's request! Something more was going wrong. It was mysterious. Sure looked like Prosper was just pulling the plug. Lenders wanted information, but Prosper would not tell us what was happening.

The legal test was a horrible failure. Prosper demonstrated that it doesn't know how to collect on late loans even when it tries, even with expensive lawyers helping. How could this be?

Prosper has been mute on this subject -- for almost two years -- until a few days ago, when Doug Fuller wrote about the results of the legal test project in Prosper's official blog. (Because Prosper has often deleted their blog entries which proved embarassing, I point out that a copy of Fuller's writeup is archived here.) Its a worthwhile read.

Now we know some of the numbers, and more about what went wrong. In summary: Everything went wrong. It is as it appeared. The legal test was a horrible failure. What's worse, to most observers it seems obvious that every step could have been carried out in a much more competent fashion. It was a three stooges operation.

Here's what we learn by reading Fuller's recent admissions. (I'm not the first to dissect Fuller's writeup. I follow the outline given by Prosper lender ira01 here. Read ira's message for more technical legal details.)

They started with 66 hand-picked charged-off loans. Of these, they discovered that 3 borrowers had moved to a different state, so they gave up. (Mind you, it is possible to sue in all 50 states, but Prosper just gave up, preferring the convenience of only suing people resident in Prosper's home state: California.)

Next there were 13 loans where Prosper's legal team was unable to locate the borrower, in order to "serve" him, so they gave up. Its pretty outrageous that 13/66 = 20% of his hand-picked borrowers can't be found.

What did they do to find these people? They don't say. It is possible that they did nothing except try the address they had on file. Fuller offers the excuse that in his experience, the address shown in Lexis is usually accurate, but wasn't in these cases. I guess Fuller believes these people just vanished without a trace. Given Fuller's background, we expected him to be skilled at skip-tracing, or know how to hire someone who was.

Doug offers the observation that Prosper borrowers don't seem to fit the pattern of his prior experience. Here I agree. Prosper borrowers are different. Two years after he joined us, he's beginning to figure it out.

One possibility is that these 13 people never existed. Prosper.com has an obligation to purchase "identity fraud" loans from the original lenders, therefore they don't like to consider this possibility.

There's something else fishy about giving up on these 13 borrowers. In California, if after diligent effort you haven't been able to find the defendant to hand him the paperwork, you can satisfy the requirement for service by printing a notice several times in a newspaper. This allows you to proceed with a suit, and hopefully get a default judgment against the deadbeat, even tho he's hiding from you. Then when you do find him later, you have a legal advantage. Apparently this was too hard for Prosper, so they gave up instead.

Doug then tells us that in 21 more cases, the borrowers filed bankruptcy after Prosper filed suit! Doug also tells us that is a huge unexpected number. Another clue that Prosper borrowers are different.

I do wonder about those 21 bankruptcies. In the past there have been some cases where a borrower told Prosper that he had filed bankruptcy, and Prosper took the borrowers word for it, when in fact there was no bankruptcy. I wonder if they actually checked these cases with the court. I also wonder whether Prosper has filed paperwork with the bankruptcy courts in these 21 cases. Perhaps we'll figure out how to track down these bankruptcies (or lack thereof) another day. (After all, we know the names of most of the defendants, because we found them while searching court records to track down the status of these lawsuits during the 2 years when Prosper was silent.)

66 minus 3 minus 13 minus 21 ... just 29 loans left.

More than half the suits died before they even got started. Remember Fuller's words? "I’m going to sue you. If I sue you I’m going to win. "

At this point Fuller stops giving us precise numbers. He proceeds to explain a number of different things that went wrong, and the fact that he's unhappy about them. He's not the only one.

A typical early step in this kind of lawsuit is to ask the court to grant a "default judgment", meaning that the facts are so clear that there isn't any use wasting time with a trial. (Default judgements are common when the defendant doesn't bother to show up, etc.) Doug explains that they had a little trouble with this: "the vast majority of courts rejected the requested Clerk’s Judgment." Well huh. He doesn't explain why. Dude, we want to know why, because there's something very wrong here. Maybe you should diagnose the problem and fix it. Maybe you guys are putting the carbon paper in backwards.

And a clerk's judgment is just the easiest kind of default judgment. If you fail that, there's another kind you can apply for. So we expect, well you know ... follow-thru. Fuller says that as he was getting ready to do that, Prosper entered its "quiet period", because of Prospers SEC filing, intended to straighten out the regulatory issues Prosper had with their lenders. Fuller decided to not proceed with the necessary court filings during the quiet period. Well the quiet period lasted nearly a year, so this was quite an interruption! Prosper missed legal deadlines and the courts dismissed most of the remaining cases.

There is a bit of a hole in this story. Prosper filed their first S1 with the SEC in October of 2007, before the legal test project had even started! Another hole in this story is the fact that the registration with the SEC has to do with the relationship between Prosper and lenders, and has nothing whatsoever to do with borrowers. I know Fuller agrees with me on this point, because he has used this argument himself. In an earlier blog entry, describing one case which he won, he tells us that the defendant (borrower) raised the issue of the SEC boondoggle, and Fuller was able to counter with an explanation that the SEC matter didn't affect the relationship between Prosper and its borrowers. Given that the courts hearing these cases had no problem with the SEC entanglement, I can only presume that it was overcautious lawyers advising Prosper on the SEC matter who told Doug to shut down. They harmed lenders by doing so.

Doug says he's going to try again on some of the cases that were dismissed. I'm glad to hear it. He didn't say when. We've blown two years on this "test" so far, and we have nothing to show for it.

By the way, we've been out of the quiet period for several months now. If the quiet period was the thing keeping Fuller from proceeding with legal action, he should have fired up the suits again by now, but he has not. I question his intent.

Well, at least they did win at least one case out of 66.

Many times we've been told they were going to treat nonpaying borrowers seriously, but looking back these statements have never been truthful.

While they're firing up the 2nd try on many of the dismissed lawsuits for this initial batch of 66 borrowers, there now are another 7,933 that's right seven thousand nine hundred and thirty three Prosper borrowers who have defaulted on their loans, with nobody even thinkin' about legal action on any of these folks. There is simply no credible story that they're ever gonna get any material legal action going against deadbeat borrowers.

What have we learned?

Prosper borrowers are "different". While most of them are very nice people, like the couple I met in Long Beach, a substantial fraction of Prosper's borrowers are hardened deadbeats, attracted by the sweet odor of naivete. Prosper lenders cannot succeed without better mechanisms to protect against against these zombies.

Deadbeat borrowers are a big problem.

Prosper has a Three Stooges legal strategy.

Prosper management has never taken these issues seriously.

Prosper management's talk about how they would handle deadbeats has been hot air from the very beginning.

It was such a wonderful concept. It seems such a shame.

In order to stay alive, Prosper.com needs a new round of venture capital funding some time in the next few months. If I were one of the VCs considering an investment, I'd be lookin' to replace Chris Larsen. Mr VC, if you're reading this, that is what I suggest you do.

PS: Great discussion among P2P lenders is found at prospers.org .

Sunday, October 11, 2009

Prosper.com - Don't blame the economy

From time to time, someone will try to explain away the poor performance of prosper.com loans by blaming it on the poor economy. Its become a national pastime to use the economy as an excuse these days. But is it true?

To sort this out, I have graphed the default rate of all prosper.com loans side by side with the U.S. unemployment rate. If bad economic conditions cause prosper borrowers to change their behavior, then you should see the prosper.com default rate go up when unemployment goes up, and vice versa.



These two curves don't look correlated at all!

In the last year, unemployment has roughly doubled, which represents a huge change in economic conditions. During that time the default rate on the prosper portfolio has bounced around, but it certainly hasn't gone up in tandem with the unemployment rate.

In a recent blog, Prosper's CFO Kirk Inglis was describing Prosper's new system for rating borrowers, and he said "...the historical performance that underlies the Prosper Rating System is derived from a poor economic environment. As a result the estimates of loss are biased higher than if the economic environment had been more benign."

In other words, he argues that because he did his calculations using loan data from the recent past, when economic conditions were bad, then surely future loans should do better than his model predicts. His unstated underlying assumption is a correlation between prosper loan performance and the economy. I see no evidence for his conjecture. Wishful thinking.

It may seem counter-intuitive, but we can't deny the data. Loan payment performance doesn't seem to track economic conditions.

Boring methodology footnotes: To compute the instantaneous default rate of the entire prosper portfolio, I looked at prosper's performance web page. I asked it to show me all loans (for all time) as observed on the first day of each month. I copied down the number of loans that had defaulted as of that date. Subtracting these numbers for two adjacent months tells me the number of loans that defaulted during that month. To convert this to a default rate, I divide by the total number of loans prosper had originated as of four months earlier.

I chose the four-months-ago total because loans take four months to default. Using a later total would have included loans in the denominator which could not possibly appear in the numerator. Finally, I converted this monthly default rate to an annualized default rate exponentially, ie da = 1-(1-dm)^12 .

This process produces an instantaneous "whole prosper portfolio" default rate. Beware that this number doesn't help us judge the performance of loans prosper originated at any point in time, or at any credit grade, becaue here they are all mixed together. It does allow us to observe trends (if any) in borrower payment behavior over time, so it seems appropriate for this inquiry into how borrowers behave during economic hard times.

You probably note a downward tilt in the default rate curve during the months that Prosper was shut down by the SEC. This is a side effect of the loan portfolio "aging" during that period. No new loans were being added, and the existing loans were all getting older. Therefore the age distribution of loans in the prosper portfolio was changing. This aging produces a lower default rate, because the default rate naturally falls somewhat as loans age. This happens because the portfolio is a heterogeneous mix of loan quality. The bad loans (think HR, E, D, ...) tend to fail early, leaving a more aged portfolio with a higher quality mix.

The unemployment rate shown is the whole-country U-1 series produced by the bureau of labor statistics, and which I obtained from www.economagic.com . This is the unemployment rate that you most often find quoted in the press.

PS: The best discussion among P2P lenders can be found at prospers.org .