Entries Tagged 'Statistics' ↓

Prosper 1 Month Late or Worse Curves

The methodolgy or the sql for obtaining the data these curves from the private data export have already been explained in detail, but briefly…

  • The Y axis is the percentage of all loans orginated of a given age that are currently 1 month late or worse…
  • The X axis is the days loan origination. 
  • The curves stop when there is less than 250 loans in the “bucket.”
  • These curves are for all loans of a given Credit Grade.  One would get different curves (flatter slopes=better); for exmaple, if you were to choose loans with 2 or less inquires and 5 years of credit history you would see flatter slopes.

The range is 7-8% for AA all the way to ~55% for HR.  

(Click Graph for Larger Version)

Prosper Vintage Curves

Prosper Vintage Curves

Related Articles

Update of IRR/ROI Values July 2008

Another month another update…  The picture is rather muddled right now between the accrued interest showing up in my account balance even on loans that are technically “charged off”and lending stats changing their algorithm… All I can do is continue to track and hope fore some clarity…  On a positive note I have had no loans in the 1 month late category for most of the last month…  (There are 2 in that category now, but for most of June they were not there…)

I am often asked what is the biggest thing I would do different if I could go back and do it again?  I would follow advice very similar to Rich Credit Debt Loan in the post Making Extra Money With P2P Lending. I have learned a lot about credit risks and the like, but I would have much rather been more successful in my lending endeavours… My single biggest mistake in my lending was to ignore inquires.

Here is my current tracking chart…

Related Articles

Vintage Curve Update — 06/12/2008

Brief Explanation: These curves show the entire set of prosper loan broken down by credit grade and lined up along the x axis on their origination date…  As a loan goes late (1 month or worse) it is counted (either by amount or by count) as late against the population…  The curves stop when the loan population falls below 250 (ie there are 249 or less loans that age or older)…

Recently a study from the University of Maryland claimed a peak default date around month 10 of a Prosper loan.  This would translate into the largest delta  in this graph over a month period.  Does this graph confirm or deny that statement?  Is it conclusive?  Please leave a comment.

Here is the vintage curves by count (click graph for larger version)…

Vintage Curves By Count

Here is the vintage curves by amount (larger loan go late at a higher rate and therefore on a percentage basis you would expect an increase), (click graph for larger version)…

Vintage Curves By Amount

Here is the SQL that I used to pull the underlying data out of the public and private data downloads

DECLARE @DTD int
SET @DTD=30
SELECT
cast(aday-originationdate as int) as 'PIT',
l.creditgrade,
sum(PrincipalBalance+NetDefaults) as 'Amount',
count(l.[key]) as 'Count',
sum(case WHEN (mld.DPD!=0 and
       (mld.DPD+(aday-observationdate))>@DTD) THEN
            PrincipalBalance+NetDefaults ELSE 0 END) as 'AmountLate',
sum(case WHEN (mld.DPD!=0 and
       (mld.DPD+(aday-observationdate))>@DTD) THEN
           PrincipalBalance+NetDefaults ELSE 0 END)/
           sum(PrincipalBalance+NetDefaults) as AmountLatePercentage,
sum(case WHEN (mld.DPD!=0 and
     (mld.DPD+(aday-observationdate))>@DTD) THEN
        1 ELSE 0 END) as 'CountLate',
sum(case WHEN (mld.DPD!=0 and
       (mld.DPD+(aday-observationdate))>@DTD) THEN
       1.0 ELSE 0.0 END)/count(l.[key]) as 'CountLatePercentage'
FROM
loan l
inner join creditprofile cp on cp.listingkey=l.listingkey
inner join LoanPerformance mld on l.[key]=mld.loankey cross join alldays
where
mld.observationdate = ( select top 1 observationdate
from LoanPerformance sub
where sub.observationdate < aday
and sub.loankey=mld.loankey order by sub.observationdate DESC )
and aday < getDate()
and aday >= '02/01/2006'
and l.creditgrade!='NC'
group by
cast(aday-originationdate as int),
l.creditgrade
having
count(l.[key])>250 and
sum(PrincipalBalance+NetDefaults)>0
order by
'PIT'
Related Articles

Prosper in May Set a New Monthly Loan Origination Volume High Water Mark

(sql to produce statistics at the end of post) 

In May, Prosper originated more loan volume than in any previous month roughly 9.6M loans the vast majority were credit grades AA-C.  This is great news for Prosper on 2 fronts: total volume and loan quality. Here is the breakdown…

May 2008 Prosper Origiantions
Credit Grade Loans Amount Amt / Loan % total
AA 223 $2,003,970 $8,986 20.9%
A 225 $1,924,277 $8,552 20.0%
B 270 $2,084,555 $7,721 21.7%
C 388 $1,979,505 $5,102 20.6%
D 295 $1,202,235 $4,075 12.5%
E 82 $185,909 $2,267 1.9%
HR 120 $220,425 $1,837 2.3%
Total 1603 $9,600,876 $5,989

All in all it looks like a very positive step… Can Prosper keep the momentum… stay tuned for next month?

Sql:

Select creditgrade, count([key]), sum(amountborrowed)
from loan
where originationdate>=’5/1/2008′ and originationdate<’6/1/2008′
group by creditgrade

Related Articles

RateLadder IRR/ROI 6/08 Update — 0.57% to 16.64%

My account update for the month shows some general improvement…

One thing to note is that with the loans in the debt sale not sold my IRR calculation is a little out of whack… Hopefully Prosper will create the loan status of “Charged Off” and remove those accounts from my balance which will them cause my IRR to be a clearly representation.

Lending stats changed their ROI algorithm which had a severe effect on my LS ROI… I am not sure the new number is comparable to any previous reading, but I will continue to track it anyway…

Anyway, here is a chart of my updated IRR and ROI values…

Related Articles

RateLadder IRR/ROI 5/08 Update — (1.35%) to 15.3%

Well another month down… Nothing major happened that would effect my IRR/ROI (like a debt sale.)  It was a fairly uneventful month for my Prosper p2p lending account, but it was a good month…  My ROI/IRR is higher on 4/5 of my tracking indices (it will be 5/6 when the statements are available and I update Quicken.) And my EricsCC ROI is within 4 basis points of last month’s value.

When going through my late loans this morning to build these statistics I noticed 2 things.  One it is now very easy to tell which loans are in bankruptcy vs which loans are in collections.  (3 of my 29 1+ late or worse loans are in bankruptcy.)  The second is a more positive development many (more than I have ever seen before) of my late loans have collections payment pending.  In fact, 1 of my 4+ month late loans has a collection payment pending, 3 out of 7 of my 2 month late loans have a payment, and 2 out of 3 of my 1 month late loans have a payment.  Maybe things are looking up in the collections area?  Time will tell.

Anyway, here is a chart of my updated IRR and ROI values…

Related Articles

Quicken ROI Update = 8.99%

With the lender statments taht were released this weekend it is possible to update my Quicken ROI

Last month being a quiet “normal” month one would expect an increase in the quicken value…  My Quicken ROI increased from 8.11% to 8.99%.

Related Articles

The WealthBoy Strict ROI for Prosper Lenders

WealthBoy, whom I met at Prosper Days, has just released The WealthBoy Strict ROI for Prosper Lenders – an open source, sql based ROI calculation for Prosper Lenders.  His calculation makes use of the public and private data files from Prosper.  He was inspired by his Prosper Blog post: The Rule of 72 on Prosper.

TotalROI = (Total Interest Received - Fees - Losses on Defaults) / (Total Loan Originations - Reinvested Loans) 

For RateLadder that equates to

Screen Name: RateLadder_com
Total Bid Count: 324
Total Reinvested Bids: 115
Total Originations (total amount loaned): $16,296.50
Total Investment (total amount loaned excluding reinvested bids): $10,500.92
Total Income (total principal and interest less fees): $6,067.12
Total Profit (total interest less fees and defaults): $1,138.48
Total ROI: 10.84%
Average Loan Age: 11.81 months
Annualized ROI: 11.02%

For the initial release he made no assumptions about the value of late loans. But… 

It is an open source, sql based ROI calculation for Prosper Lenders…  With the calculation being open source it is clear (assuming you can read through about 250 lines of SQL :) exactly how the calculation is performed…  As his sounding board during development I can confirm that WealthBoy put a lot of work into this project.  I think this is a great addition to the Prosper tools.

At the moment it takes about 45 seconds to perform the calculation for the largest lenders at ProProsper…  It will only take longer as more payments are made.  This performance issue is the main reason I have yet to implement it at ProProsper as it will take some back-end engineering on my part to calculate the values in an efficient way in terms of server resources. I expect to add it to ProProsper, but currently I am not able to find the time to do so…  Hopefully soon… Or maybe Eric or LendingStats can add it first…

Anyway, I encourage you to check it out.  Well Done WealthBoy!

Related Articles

Vintage Curves 04 2008 — Amount Late By Credit Grade from Origination

It is time for my monthly update of the amount late vintage curves… The y axis is the percent of $$ 30 more more days late… the x axis is days from origination.  The curves stop when there are less than 250 loans in the bucket.

There are ways to cut the lat percentage shown in these graphs in half and even in half again (6 month inquires and current delinquencies to start with), but I would look at these graphs as a cautionary tale to avoid the lower credit grades and pay attention to the extended credit…

Use the Prosper performance tool to better understand the risk…

Amount LAte Vintage Curves

Related Articles

RateLadder IRR/ROI Update April 2008 — (2.6%) to 14.37%

Compared to last month with the initiation of the lawsuits…  This was a quiet month.  My worst number Model IRR degraded and my best number Default IRR improved.  I will update quicken as soon as the lender statement become available.

Related Articles