Friday, August 31, 2007

Linkfest

Some more interesting reading stumbled upon by yours truely...

Lies, damn lies, and statistics in PhD candidate publishes empirical analysis of Prosper:

The results are certainly interesting. One thing that catches my eye is the homeowner determination. Based on Matt's analysis, homeowners are actually a greater risk of default than non-homeowners. Yet, according to the dataset Kumar used, being a homeowner is not a significant variable but Prosper lenders bid down homeowner loans. A careful look at Kumar's dataset could help increase returns for lenders.

ZCommodore has a good overview on the HR implosion:

Over 50% of all loans on Prosper during November and December last year were to E grade borrowers and below (including NC). These borrowers have been shown to default at alarmingly high rates, much higher than the supposed 19.1% Experian estimate promoted by Prosper.

How did this happen? Here's the answer:

Prosper Lending Review spots an entertaining NY Times article on Prosper:

Jonathan Last wrote a humorous article about Prosper, Need a Loan? Usury for Beginners, for the Wall Street Journal which details his own experience becoming a Prosper lender. Not only does he become a lender, he becomes a loan shark.

RateLadder reports on the results of 3 loans sold during the latest defaulted debt sale:

CreditGrade: AA
...
$36.63 / $143.87 = 25.46% of principal

CreditGrade: C
...
$4.06 / $43.26 = 15.95% of principal

CreditGrade: D
...
$3.83 / $45.60 = 8.4% of principal

RateLadder later posted Pensioner's results.

Thursday, August 30, 2007

Concensus Ratings

LendingStats' has started rating listings based on who's bid on them.

A few people have inquired about how the “Consensus Rating” is being generated on the Listing Picker page under Lending Strategies. The idea behind the Listing Picker is to identify listings bid on by lenders who have good estimated ROIs. The more lenders with good ROIs the higher the consensus rating becomes.

The formula that generates consensus ratings is as follows: Look for all lenders with an average portfolio age of greater then 4 months on a listing. If there are more then 3 such lenders then get the average estimated ROI of these lenders. Only listings that have an average estimated ROI of > 5% will be shown on the Listing Picker page.


This is an interesting analytical addition to Lending Stats' technology. I'll be watching this closely.

Accessing Prosper's Performance Data Part 3

This is the third post in a series on accessing Prosper's Marketplace Performance engine. Disclaimer: If you're not at least a bit geeky, this may bore you to tears. That's the danger of reading a blog run by an engineer.

In the last two posts, I covered some of the various parameters that are accessible to crafty users and how to extract useful statistics from the webpages Prosper creates. In this mad dash, I skipped over a whole bunch of parameters that can wildly adjust the loan performance. Time to fill those in.

The parameters below are part of the extended options (see the "Show More Options" section) in the Prosper Marketplace Performance page, and they're quite fun to play with.

Field
Type
Default
Min
Max
Description
minND
Integer
0
0
?
Minimum Now Delinquent
maxND
Integer
? (max)0
?
Maximum Now Delinquent
mnAD
Integer
0
0
?
Minimum Amount Delinquent
Minimum amount ($) currently delinquent
mxAD
Integer
? (max)0
?
Maximum Amount Delinquent
Maximum amount ($) currently delinquent
minD7
Integer
0
0
?
Minimum Delinquent In The Last 7 Years
maxD7
Integer
? (max)0
?
Maximum Delinquent In The Last 7 Years
minPR10
Integer
0
0
?
Minimum Public Records In The Last 10 Years
maxPR10
Integer
? (max)0
?
Maximum Public Records In The Last 10 Years
mnPRM
Integer
0
0
?
Minimum Public Records In The Last 12 Months
mxPRM
Integer
? (max)0
?
Maximum Public Records In The Last 12 Months
minQ6
Integer
0
0
?
Minimum Number Of Inquiries In The Last 6 Months
maxQ6
Integer
? (max)0
?
Maximum Number of Inquiries In The Last 6 Months
minFR
Integer
0
0
?
Minimum Number Of Years To First Credit
maxFR
Integer? (max)0
?
Maximum Number Of Years To First Credit
mnCCL
Integer0
0
?
Minimum Number Of Current Credit Lines
mxCCL
Integer? (max)0
?
Maximum Number Of Current Credit Lines
mnOCL
Integer
0
0
?
Minimum Number Of Open Credit Lines
mxOCL
Integer? (max)0
?
Maximum Number Of Open Credit Lines
minCL7
Integer0
0
?
Minimum Number Of Total Credit Lines
maxCL7
Integer? (max)0
?
Maximum Number Of Total Credit Lines
mnRC
Integer
0
0
?
Minimum Revolving Credit Balance
mxRC
Integer
? (max)
0
?
Maximum Revolving Credit Balance
mnBU
Float
0.0
0.0
1.0
Minimum Bankcard Utilization
mxBU
Float
1.0
0.0
1.0
Maximum Bankcard Utilization
esba
Integer
63
0
63
Employment Status
Binary encoded number. To specify status, add up the values for all requested criteria. 1 = "Not Availble", 2 = "Full-Time", 4= "Part-Time", 8 = "Self-Employed", 16 = "Retired", 32 = "Not Employed". For example, for Full-Time, Part-Time, and Retired, add 2 (Full) + 4 (Part) + 16 (Retired) = 22.
mnES
Integer
0
?
?
Minimum Length Of Status
Specified in months
mxES
Integer
? (max)
0
?
Maximum Length Of Status
Specified in months
iba
Integer
255
0
255
Annual Income
Binary encoded number. To specify income, add up the values for all requested criteria. 1 = "Not Displayed", 2 = "$0 Or Unable To Verify", 4 = "$1-$24,999", 8 = "$25,000-$49,999", 16 = "$50,000-$74,999", 32 = "$75,000-$99,999", 64 = "$100,000+", 128 = "Not Employed". For example, for $50k-$74.9k and $25k - $49.9k, add 8 (25-49.9) + 16 (50-74.9) = 24.
Thus ends this 3-post series. I'll be following up to see if any fields were improperly documented and to fill in any missing details. Happy performance searching.

Wednesday, August 29, 2007

Tanta On Troubled Borrowers

Tanta over at Calculated Risk (I'm a huge fan) had a very appropriate commentary on borrowers who get in over their heads.

I remain convinced that there's something wrong with blaming the financially inept for not realizing that they are financially inept, when those who are supposed to be financially ept--loan officers, brokers, financial counselors, advice columnists in business publications--spent the last several years refusing to tell them that they were financially inept.

Of course people who are in over their heads are surprised. They lacked the skills necessary to understand what "over their heads" might mean.
The replies to Tanta's remarks are equally interesting for thinking through it all.

Prosper lenders, you need to be responsible for your money (as mortgage brokers should've been responsible). Lenders need to be comfortable that the borrower understands their finances, or they need to walk away from the bid.

Accessing Prosper's Performance Data Part 2

This is the second post in a series on accessing Prosper's Marketplace Performance engine. Disclaimer: If you're not at least a bit geeky, this will bore you to tears. That's the danger of reading a blog run by an engineer. This post, specifically, covers non-financial topics like programming. You've been warned.

In the last post, I outlined how to get specific performance information from Prosper and documented some parameters that can be accessed. Now to suggest a use for this. Imagine that you want to trace the default history for loans from January, 2006, and see how they've aged over time. This can be done by modifying the origination start date, origination end date, and the observation date. We set the origination start (osr) to 01%2f01%2f06, or 01/01/2006, and the origination end (oer) to 01%2f31%2f06, or 01/31/06. To collect all the data, sweep the observation date from 02/01/06 (02%2f01%2f06) to 08/01/07 (08%2f01%2f07). Remember, the observation date says "pretend it's this date when looking at the data". From here, read the data into Excel, OpenOffice, or your favorite spreadsheet, and process away. Sounds difficult? It's not, and here's how to do it.

To collect all the various sets of data from Prosper, use a tool that'll automate downloading web pages, like cURL. It lets you specify the URL and the saved file. If you're running Linux (like we here at Prosperous Land do), it can be accomplished with a bash script looking something like this:

curl -o od_feb06.html http://www.prosper.com/lend/performance.aspx?osr=01%2f01%2f06&oer=01%2f31%2f06&od=02%2f01%2f06
curl -o od_mar06.html http://www.prosper.com/lend/performance.aspx?osr=01%2f01%2f06&oer=01%2f31%2f06&od=03%2f01%2f06
curl -o od_apr06.html http://www.prosper.com/lend/performance.aspx?osr=01%2f01%2f06&oer=01%2f31%2f06&od=04%2f01%2f06

And so on until you've collected all the necessary dates. If you're creative, a script could automate most of this. And, to convert the html files into comma separated values (.csv) files that can be imported into any spreadsheet, the a Python program will do the trick. It will extract the performance information from the html files. If we call the program table_extract.py, use it as 'table_extract.py mar06.html > mar06.csv'

From here, you can munge and process the data like a statistical fiend. Just remember to be kind to Prosper's server and not request searches too quickly.

The next installment will wander away from the programming realm and cover even more parameters that can be searched. Stay tuned!

Tuesday, August 28, 2007

Accessing Prosper's Performance Data Part 1

This is the first post in a series on accessing Prosper's Marketplace Performance engine. Disclaimer: If you're not at least a little geeky, this may bore you to tears. That's the danger of reading a blog run by an engineer.

Prosper has been developing an API that developers can use to create 3rd party applications. It's progressing (slowly), but seriously lacking on capturing loan performance information. To the probable chagrin of the Prosper folks, I've gone through the trouble of documenting the HTTP GET fields for the Prosper Marketplace Performance engine, the first step in cracking open the performance data.

Loan Criteria

Here are the basic criteria that Prosper lets you manipulate via their webpage.

Field
Type
Default
Min
Max
Description
osr
text
13 Months Ago
n/a
n/a
Origination Start
Only consider loans originated after this date
Encoded in MM/DD/YYYY format as:
MM%2fDD%2fYYYY (%2f is URL-safe version of '/')
oer
text
25 Months Ago
n/a
n/a
Origination End
Only consider loans originated before this date
Encoded in MM/DD/YYYY format as:
MM%2fDD%2fYYYY (%2f is URL-safe version of '/')
od
text
Today
n/a
n/a
Observation Date
Observe the statistics using information known by this date. Encoded in MM/DD/YYYY format as:
MM%2fDD%2fYYYY (%2f is URL-safe version of '/')

Listing Criteria

Field
Type
Default
Min
Max
Description
minDTI
float
0
0
1000000Minimum DTI
Use real number instead of percentage (use 0.1 for 10%)
maxDTI
float
1000000
0
1000000Maximum DTI
Use real number instead of percentage (use 0.1 for 10%)
minAmt
float
0
0
25000
Minimum Loan Amount
maxAmt
float
25000
0
25000
Maximum Loan Amount
af
integer
0
0
2
Auto Funding
0 = "No Preference", 1="Include only", 2="Exclude"
hm
integer
0
0
2
Homeowner
0 = "No Preference", 1="Include only", 2="Exclude"

Group Criteria

Field
Type
Default
Min
Max
Description
gm
integer
0
0
1
Group Member
0 = "Include anyone", 1 = "Only include group members"
sh
integer
0
0
1
Group Rewards
0 = "Include anyone", 1 = "Only include groups with no group rewards"
gr
text
0%2c1%2c2%2c3%2c4%2c5
n/a
n/a
Group Rating
Include groups with the following ratings
0 = "No Rating", 1 = "1 Star", 2 = "2 Stars", 3 = "3 Stars", 4 = "4 Stars", 5 = "5 Stars". Assemble the text string by separating every desired star rating by '%2c'. For 1, 3, and 5 start groups, it would be 1%2c3%2c5

Putting It Together

Now that the fields are defined, how are they used? They're passed in on the HTTP URL. So, to start with, the performance webpage is 'http://www.prosper.com/lend/performance.aspx'. To set specific fields, we add '?field1=value1&field2=value2&field3=value3' to the default URL for all the fields we want. Here's some specific examples.
In the next few posts, we'll cover what to do with the web pages once they're generated and some of the extended options.

Monday, August 27, 2007

Prosper Pitching to E-Bay Sellers

I was ready to call it a week and then I saw a post over at Money Smart Life that had a juicy tid-bid. Prosper hosted a workshops through E-Bay targeted toward small business financing. All the people selling random stuff on E-Bay, and Prosper's lenders are underwriting it

Need a small business loan and not sure where to get it? Or, are you certain a loan is what you need? In this informative workshop, Prosper Marketing Director, and former eBay PowerSeller Program Manager, Shira Levine will demonstrate how financing can help your business. We’ll cover the five basic questions every business should ask themselves before you finance, how to know where you stand financially, and the importance of cash reserves in your company. Then we’ll cover financing basics: debt vs. equity, credit scores and reports, and common pitfalls to avoid. The workshop will end with how to create an excellent loan listing on Prosper that will get your business funded. We’ll also look at how eBay PowerSellers are using Prosper both to borrow for business and to make money using Prosper’s referral program.

I'll hand it to Prosper for a few different things. They're targeting a entrepreneurial group that's perfect for their loan sizes (typically $5k - $25k), and they've got an ex-E-Bay staffer, Shira Levine, driving the effort. I'm amused by the comment in the forum dancing around the lowest allowable FICA score, but bring on the borrowers!

Saturday, August 25, 2007

Latest Debt Sale

The official announcement forum had the results from the latest debt sale:

We completed a debt sale this week which will be evident in participating lender accounts over the next few business days.

When a loan is sold, it is considered to be a default on the Prosper site and in the performance metrics. If one of your loans was sold, you will receive an email notifying you of the default and indicating the sale amount. Sale proceeds will be transferred directly into your Prosper account a few days after you receive your email notification.

Here are the details of the sale:
• Eligible loans were 122 days past due as of July 26, 2007, provided the loan was not part of any bankruptcy filing
• 309 loans were sold
• Price range: 1.8% - 26% as a percent of principal balance

Pricing is determined solely by the debt buyer and can vary from sale to sale. Several factors were used to determine pricing in this sale, with credit grade being a primary reason. Homeownership, which was a major factor in prior debt sales, was not used by the debt buyer as a pricing factor in this sale.
• Weighted average prices by credit grade:
--AA-A = 23%
--B-D = 13.3%
--E, HR, NC = 8.1%

We anticipate the next debt sale will occur in December.

Of interest was that the debt buyer ignored home ownership. This was not the past default sale experience. If this continues, it will simplify default modeling for lenders slightly. It also looks like they fetched more for defaulted AA-A grades compared to the May, 2007 sale (19% high in May, no specific grades mentioned)

Friday, August 24, 2007

Prosper vs S&P 500: Apples To Apples Time

Prosper sent out a marketing e-mail last night pitching the improved Prosper returns compared to the S&P500. Their blurb is below if you were somehow unlucky enough to miss out on it.
ZCommodore was all over their disclaimers:

First, they are comparing a year of returns on Prosper with 2 years on the S&P 500. The dates are different as well. I don't know what the returns looked like for the most recent year of the S&P 500 but I suspect they'd be different if they chose the same year.

Ok, fair's fair. Lets get the S&P 500 vs Prosper for the same dates. To be extremely fair, we'll use Prosper's dates.

Rate of return shown is the average net annual return on Prosper loans originated between 7/22/06 and 7/22/07 to borrowers with AA credit grades who have 0 delinquencies and 0 to 2 credit inquiries on the their credit record, as of 8/23/07. For more information, go to http://www.prosper.com/lend/performance.aspx.

The S&P 500 index was 1240.49 on 7/21/06 (as close as Google will get me), and close at 1462.50 on 8/23/07. That's a 17.9% return over 13 months, pretty respectable and corresponding to an approximately 16.5% annualized return. Oops.

And then there's those darn taxes. If I had invested in the S&P 500 via an ETF or Spider and sold on the 23rd, I'd be paying capital gains tax rates of 15% this year versus Prosper which requires me to pay my regular going tax rate of 25% every year that I receive interest income. So, lets run the numbers, shall we.

S&P 500
Prosper
Initial Investment
$10,000
$10,000
Value After 13 Months
$11,789
$10,997
Gains
$1,789
$997
Tax Rate
15%
25%
Taxes
$268
$249
Post-Tax Gains
$1521
$748
After Tax Returns
14.04%
6.9%
As they say on MythBusters, this one is BUSTED!

Thursday, August 23, 2007

Another Snafu At Prosper

Prosper seems to be having one thing after another nipping at them. First there was the blast of planned and unplanned outages. Then there was the lender uprising that was contained before the forums were entirely shutdown. The latest is that their ACH clearing house (Wells Fargo) had some problems with transactions:

Our banking partner, Wells Fargo, is experiencing technical difficulties with their ACH money transfer system today. This may mean that ACH transfers will be delayed. Here is what you may experience.

As a Lender:
If you transferred money in or out of the system, there may be a delay coming into or out of your bank

In the next week, you may see a “dispute” on a Borrower payment you recently received

As a Borrower:
You may see a delay in either a payment processing from your bank account or, if you are getting a loan, a delay in obtaining the loan money

As a Member:
If you are setting up a new account with Prosper, the bank account verification may be delayed in setting up the new challenge deposits

We anticipate this problem being resolved in a timely manner and will post an update here as soon as we know more. You may also receive an email with this notification.

Wednesday, August 22, 2007

Ben Stein Editorial

The NY Times had an editorial by Ben "Bueller" Stein on the irrational behavior of lenders. Even though it was originally published in March, I think it's relevant for Prosper lenders.

But wait a minute. What is a junk bond if not a loan to a subprime borrower? Why do lenders think that those bonds are called “below investment grade?” It’s not just a formality or a quaint term. Over long periods, junk bonds really do default at rates far greater than investment-grade bonds. That’s why smart lenders used to demand very substantial premiums for making those loans.

Now, in a mania of adolescent “we know better than our stodgy parents” behavior, lenders lend to junk borrowers for not wildly more than they charge Exxon Mobil.

Or, to translate into Prosper terms: what do you mean I can't expect HR and NC loans to have the same default rates as AA? This is the new economy and Web 2.0 (TM). It's different this time.

Right...

Tuesday, August 21, 2007

Sub-prime Mess And Prosper

This blog post got me up and thinking on the sub-prime mess and how it interacts with Prosper. Since I'm sure there will be lots of articles to follow, I wanted to stake out my positions before the posts are flying fast and furious. I've been following this slow-motion train wreck for about a year and 1/2 on Calculated Risk's blog (and, as an aside, you can see that CR did great diligence going back into 2005).

By all accounts, the basis of the sub-prime implosion was two related problems: people who shouldn't have been given loans were, and "creative" loans were given in place of easy to understand ones. Sometimes this was done fraudulently, other time criminally, but times were good, homes were appreciating, and everyone was an easy refinance away from calamity. And then the houses stopped appreciating, refinancing wasn't easy, and people face the worst possible outcome: default and home loss. All of the financial engineering, securitization, collateralized debt obligations, stated incomes, slicing, and dicing exacerbated the situation because someone made money at every step and there was little incentive to do proper diligence with everyone else lending like crazy people.

It should be noted that about 6 months ago, the Prosper lenders went through a similar meltdown. There was a collective realization that borrowers with HR credit ratings shouldn't be getting $25,000 loans at 15% because they were horrible risks. In the Prosper situation, there wasn't much of a wind up - Prosper loans are not complicated, came due quickly, loud lenders provided feedback, and the market corrected itself as statistics became available. It was dangerous for lenders who jumped in with both feet at the beginning, but this underscores the importance of having a good model before lending. This resulted in a flight to quality, with AA - C grade loans getting much more attention despite their lower interest rates.

There are lessons and warnings for Prosper lenders from the sub-prime mess, however. First, avoid stated income loans. Prosper does verify the claimed incomes of borrowers before a fully funded listing proceeds to a loan. Look for borrowers with verifiable income and check a borrower's budget against their income. Be careful with listing descriptions that claim a spouse's income - it's not verified and you're trusting them to state it correctly. Second, guard against collateral damage from the sub-prime implosion. Realtors, mortgage brokers, construction workers, and house flippers will be under financial stress as the housing industry slows down. Tighten up your lending criteria if a listing mentions the real-estate industry in their job or listing description or if they live in danger areas. Ask borrowers about their mortgage terms, and avoid California, Arizona, and Florida unless you're ready to put in the additional vetting work, for these states are over-exposed to the sub-prime contagion.

Monday, August 20, 2007

Linkfest

A bit of linkage to start into the week.

On the Prosper Upgrades

  • Major Prosper Upgrade: "There is a major Prosper Upgrade tonight… I apologize for not trying any other these features, but I am on vacation an thus here is a reprint of the announcement from Prosper Andrew…"
  • Prosper improves lending and borrowing experience: "Prosper just announced some significant site and policy updates. Loan listings now include borrower's past activity, borrowers can make payments with their Prosper cash balance, borrower activity including all past activity will be reported to TansUnion in addition to Equifax, and the referral program has been extended through the end of the year. These changes have been well received by the Prosper community after the recent controversy."
  • I was right about one thing: "I was right about the fact that Prosper was preparing a new release. Last night Prosper released a few new changes. The biggest one in my mind was the addition of a second credit bureau that they are reporting to. They still are only checking Experian when pulling credit data but if putting dings on two credit bureaus for late borrowers helps encourage them to pay up, more power to them".
  • Even more data transparency - Prosper adds missing link: Prosper.com now makes the link between loans and listings available. With this new feature now everyone can see the status of the loan payments of every borrower. On the borrower profile page there is now a "Loans" tab.
Online Riot (Sigh)
  • Open rebellion of lenders in Prosper forums: "Looking into the lender section at the Prosper.com forums part of the posting lenders seem to be in open rebellion. Titles of threads from the last two days include "FLASH:Prosper bans $100.000 lender", "Prosper Mng. Living Under a Rock", "Shooting the Messenger", "What happens if Prosper goes under?", "I am done lending on Prosper", "Hello Prosper Moderator", "Lender's WHO are DONE with Prosper", "My letter to Prosper", "Open letter to John Witchel""
  • Prosper Blenders: "On Prosper an individual that is both a borrower and a lender has been given the nickname blender. ... It seems that after a week or so of the Prosper Forums ranting about this issue and a suspension of a major (>100K) lender for impersonating a Prosper employee, that Prosper has finally said what it will do about the issue"
  • After online riot Prosper restricts some lenders, bans others from forums: "... there has been an uprising in the official Prosper forums over the last couple of days over this issue. Lenders have asked repeatedly for Prosper to force these blenders to repay their own loans with the payments they receive from their borrowers. The issue became heated in the forums over the last two weeks ..."
RateLadder relates a conversation with rgf on how default timing impacts ROI that's worthy of a read

I had a very interesting email exchange with RGF. (I emailed him after using his forum posts in articles last week).

Here is a portion of that email exchange on default rates over time and their affect on yield.

Friday, August 17, 2007

Prosper Lender-side Dynamics

There have been some shifting dynamics going on that I've noticed as, at best, some anecdotal experiences. They've altered the lending game a little bit, but are mostly signs of a more complex financial ecosystem.

First, cloud9, the group leader for Have Money - Will Bid, has taken to putting up large some of cash for listings about 2 days before they're due to expire. For example, this one, where he bids $3100 on a $4900 loan. He's bid at the original interest rate, so his large contribution rolls off as the last minute bidders and standing orders show up. The strategy is working to get his group's loans funded (the group ranges from 3 - 9 loans funded a day over the last few weeks), but his actions could be pushing possibly marginal loans into the funded category. Either way, it has the effect of pushing down interest rates - good for borrowers, mixed bag for lenders.

Second, I'm seeing larger bids showing up on the lending side, which is driving down interest rates. I am finding that it's not unusual to have multiple >$1000 bids on a listing. This specific listing has 1 $1250, 3 $1000, and 3 $500 bids. My suspicions are that there is more money flowing into Prosper for lending as it's become more credible and that there's been a flight to quality (i.e. non-HR loans) once the pioneers realized that a poor credit rating means something.

And finally, I've been having a lot more cancellations happen for listings that got 100% funded. For my current bidding bender, I've had 7 / 36 loans that completed get canceled. That's almost 20% and a black eye on the borrowers. Since I didn't see such a high rate previously, I'm caught trying to guess at the cause. Admittedly, I've moved into the D space, but either the quality of borrower has gone down, or Prosper has tightened up their review standards.

For the first two, if you have confidence in how you model the value of the loan and your underwriting standards, these new dynamics won't change much. It does lower the interest rate that loans will fetch, however, so this will push more loans below your bid price.

As to the cancellation rate, I'm disappointed that it's so high, but I'm finding some comfort in it. Presumably Prosper is making some effort to verify information and, in failing to do so, are preventing either fraudulent (why, of course I made $75k last year) or careless (did I forget to send in last years tax return?) borrowers from having their loans issued. This improves the quality of those loans that do fund.

Wednesday, August 15, 2007

Lending Efficiency

Since I'm still having procrastination problems with writing about default models, I started thinking about what my time spent working on Prosper loans is worth. That got me looking at the revenue generated from Prosper loans and how my time was invested in getting that revenue. Lets start by looking at the interest income over 3 years for a bid of $50 (the Prosper minimum).

Interest Rate
Interest Income
10%
$8.08
14%
$11.52
18%
$15.07
22%
$18.74
26%
$22.52
Doesn't seem like much does it. After all, $50 at 10% per year should be at least $15, right? That's the magic of the amortizing loan schedule. The interest earned decreases every month as the borrower repays a portion of the $50 borrowed.

As to time spent, my off-the-cuff guess is that I spend around 5 minutes per bidded listing (most of that time is passing on the rest of the listings), and only a third of those listings will successfully become loans (between not being fully funded, outbid, and having the loan canceled). This is sufficient to determine my hourly wage for being a bidding monkey.
Interest Rate
Interest Income
Hourly Income
10%
$8.08
$32.32 / hr
14%
$11.52
$46.08 / hr
18%
$15.07
$60.28 / hr
22%
$18.74
$74.96 / hr
26%
$22.52
$90.08 / hr
This is, of course, excluding early repayments, defaults, and time spent researching loan strategies. If only life were like the 26% loans. The right strategy to improve your efficiency is to invest more per bid (say $100) instead of chasing those D, E, and HR listing, but be careful to stay diversified with a very low percentage (2-3%) of your Prosper money in any given loan.

Tuesday, August 14, 2007

Zipcodes With High Defaults

I got really frustrated with Prosper when they removed the cities from the borrower's listings. This is why:

A study for CNNMoney.com by RealtyTrac, an online marketer of foreclosure properties, showed that 139 of California's ZIP codes fell within the top 500 for total foreclosure filings in the United States. The next highest count for any state is less than half that at 72 and is in another sun-belt state - Florida.

...

RealtyBid spokeswoman, Daphne Shannon, said, "The Midwest has always been very solid for us, but the properties we're seeing are moving across the country - they're from California, Arizona and Nevada."

Call me cynical, but I don't want to lend to people in areas with high mortgage defaults and foreclosures. I'm thinking that if someone has defaulted on their mortgage and is considering bankruptcy to hold things together, an unsecured loan is a trifle. I keep hoping that Prosper will add state information to their Market Performance engine, but until then, continue to highly scrutinize listings from California, Nevada, Florida, and Arizona.

Monday, August 13, 2007

LendingStat Is Back

LendingStat, a Prosper statistics site, is back after a long period without database updates:

We’re back after a longer than expected hiatus. We apologize for the delay between updates. There’s a saying that things happen in threes, well, in a perfect storm of sorts, our web programmer needed to take a leave of absence last month. Between that, the Prosper data changes, and our backend changes/server migration, we’ve been trying to pick up the slack.

I can now go back to excitedly clicking every few days to see my anticipated ROI. Twitch. Ok, more seriously now, it's insightful to be able to see what lenders are up to.

Saturday, August 11, 2007

It's All About The Income

I've been playing around with Prosper's Marketplace Performance tool to overcome the limitations present in the exported Prosper data (no detailed credit information). While playing around with some initial queries (if you're trying to figure out how to change things, look for the not-so-easy-to-find "Edit Criteria" link), I stumbled onto some statistical confirmation of something that I've believed for a while: income matters.

This seems obvious. Now matter how well intentioned someone is, they can't repay their lender if they have no cash. Therefore, borrowers disclose their incomes. Prosper is supposed verifying income levels if a loan funds. It's the borrowers who low-ball their official incomes for the listing and then claim in the description that they've got more money available (it's just hard to prove it, trust me) that scare me. As the mortgage industry is learning, stated income loans are risky.

Prosper only has about 6 months of data with income, so this is only a first peek, but the results are interesting. Below are the percentage of distressed loans (late or worse) in the recovered datasets.

Income
AA
A
B
C
D
E
HR
Income >$25k
0.54%
0.99%
0.53%
2.13%
2.55%
4.86%
9.35%
Unstated or <$25k 0.00%
0.75%
2.53%
3.20%
4.64%
11.39%
20.39%

While this isn't enough information to predict default rates, it's a clear indicator that a verified income is equivalent to elevating the borrower by one credit grade for all but the best borrowers (or, conversely, not having verified income is equivalent to a drop of one credit grade).

Tuesday, August 7, 2007

Finding Default Information

RateLadder has been collecting default sale information from various lenders (pensioner and L5 come to mind) to get a picture of the default recovery for lenders. This is a bit labor intensive and requires the cooperation of other lenders. I've recently spotted a quicker way to collect default statistics.

The Prosper Marketplace Performance page provides all the necessary information. After doing a search, look under Performance Data at Defaults (D) and Net Defaults (ND).

The amount recovered for the lender after fees is (D - ND). To calculate the recovered percentage, use (D-ND)/D. I've pulled a few of the favorites just for fun (I'm odd like that), but I didn't have the patience to do a cross-correlation with auto funding.

Criteria
AA
A
B
C
D
E
HR
All
14.2%
19.9%
15.8%
19.2%
11.5%
9.9%
8.5%
Homeowners
n/a
19.9%
22.8%
21.2%
19.7%
18.7%
21.7%
Non Homeowners
14.2%
19.9%
12.1%
14.7%
8.5%
7.2%
6.7%
The AA grade is slanted because there's only 1 loan that's gone through a default sale. The A grade is just a freaky coincidence (and an implication that home ownership doesn't matter here). I know this much - if I dared mess with E and HR loans, I'd rate home ownership pretty highly.

Friday, August 3, 2007

Rapid Transfer Details

While trolling through the Prosper forums a while ago, there was some commentary on express transfers for Prosper lenders (also from RateLadder). It caused money transfers from banks to post immediately to the lender's Prosper account instead of the usual 4 business days.

I got contacted by a Prosper Lending Services Manager. I suspect the call was because I've been doing lots of transfers to fund more loans (which, incidentally, is what's been distracting me from posting these last few weeks) and they like happy lenders. He was friendly enough, and I did have the opportunity to ask about the rapid transfer service.

To qualify, a lender needs at least $2500 in active loans. I have interpreted this to mean principle in active loans, but once you hit $2500, that's splitting hairs. Once this happens, the service is automatically enabled and the lender can get the rapid transfer on any amount greater than $500 and less than 20% of the active loan amount.

I'm guessing this'll show up in the next non-maintenance post in the announcement forums, but only time will tell.

Thursday, August 2, 2007

Tip To Borrowers - Lenders Can Do Math!

I don't make a habit of picking on specific borrowers, but after writing a few tips for borrowers, I wanted to highlight what not to do. In this listing, the potential borrower wants a loan of $8000 and claimed a gross income of ~$16600 and listed expenses of ~$5000.

MY INCOME and EXPENSES:I make about $200,000 per year and my husband stays home with the kids. My expenses Mortgage ($2,500/mo), Insurances (300/mo), Utilities (500/mo), Two car payments ($757 and $712/mo respectively), three credit cards that never really have much of a balance (maybe $100/mo payment), Student Loans ($170/mo). I also have to cover some of my own travel in my income I stated as well as pay my taxes.

Even ignoring food and leisure, quick math implies that there should be some money left over (pessimistically assume taxes average 35%). $16600 * 0.65 - $5000 = $5790. So why do you need the loan again? Ah, yes, that little detail about covering travel expenses.

thank you for your question. I have to pay for my travel expenses out of my income and since I travel 100% of the time, typically my expenses for a month are around $6,000 or $70K per year. I also have to pay my taxes on my income as I am self-employed. I end up with some extra money per month which I try to put into the kids' education fund and save for vacations as well as a regular savings account. I could go for several months without work since I am a contractor & rely on savings.

Oh, yes, that one expense that's greater than all the other expenses combined. Yes, that's why they need the loan. Would the inclusion of this expense stop me from bidding? Nope. Would the act of ignoring a significant expense make me wonder what else the borrower is failing to disclose? Yup. Will this stop me from bidding? You betcha!