I’ve been pretty harsh on Prosper’s Portfolio Plans (PP) since they were introduced in October 2007.
This last January I wrote,
“While the introduction of Portfolio Plans was a great idea for the rookie lender, I just don't think that they will perform as well as estimated.”The good thing is that I was right. The bad thing is that I was, err… right. I must admit that I only looked at one PP slice and that the results weren’t as bad as I expected them to be, but this is what I found:
In the slice I looked at (Balanced PP, Slice 2, B grade, no autofund), loans originating during the first version of Prosper’s PP criteria (Oct 29, 07 – Jan 9, 08) had a higher default rate than given in the estimate, and lower performance page ROI than estimated. What a surprise. Of course were comparing loans of different aging, and all sorts of other stuff that muddy the results (roll rates, etc…), but the performance page is the best (and only) tool to look at this stuff.
(Since I don’t want to detract from my main point, the details of this can be found at the end of this post.)
But in the 9 odd months since Prosper first introduced its Portfolio Plans, the PP criteria have changed dramatically.
I looked at a similar B slice in the latest version (July 2, 2008), and the criteria set for this slice can be seen below:
The original B non-autofund slice is here:
What a difference. Prosper Portfolio Plans are really growing up.
Looking at the historical data this slice is based on, this is really good stuff. The performance page shows a default rate for this slice at -2.21%. Not too shabby.
Some more fun facts about this set of criteria:
- The data only holds 75 loans for this slice (grade B). Pretty thin. Looking at the results for the AA and A grades as well, it’s still not too bad. AA grade holds 352 loans, and A grade holds 155 loans
- Even though the data shows a default rate for this criteria in the AA, A, and B grades, there are NO LATES (must be from Prosper’s funky roll rate calculation)
- Re-running the criteria to set an observation date for today showed only minimal changes to default rates and resulting ROI for AA, A and B grades. Still NO LATES.
The icing on the cake for me has been the change introduced last month where the minimum bid rates for each lender are adjustable. Unfortunately they still default to the minimum bid rate shown in the criteria pages, but each lender still has the chance to adjust these as needed. So, my other big beef recently about the conservative PP’s nauseatingly low minimum bid rates is no longer a concern.
I said I would be watching you, Portfolio Plans. I’m very happy to see this change and to see how much you’ve grown up…
~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~
Nitty-gritty data from first section of post:
Here is the criteria from the very first Portfolio Plan B slice, non-autofund (dated Oct 29, 2007). This criteria remained in effect until Nov. 26, 2007. The Nov. change only adjusted bid rates, and was in effect until Jan 10, 2008.
The performance page was run using the criteria above, with origination dates between Oct. 29, 2007 and Jan. 9. 2008, with an observation date of July 25, 2008.
You can see the performance page yourself here for loans falling within the B slice of the original PP criteria.
http://www.prosper.com/lend/performance.aspx?af=2&esba=23&gm=0&gr=0%2c1%2c2%2c3%2c4%2c5&hw=0&iba=255&ibid=0&iwatch=0&lc=0%2c1%2c2%2c3%2c4%2c5%2c6%2c7&lq=&maxAmt=25000&maxDTI=1000000&maxFund=1&maxGrpTLC=1000000&maxND=0&maxQ6=2&minA=0&minAA=0&minAmt=0&minB=0&minC=0&minD=0&minDTI=0&minE=0&minFund=0&minGrpTLC=0&minHR=0&minNC=0&minND=0&minQ6=0&occ=&od=07%2f15%2f2008&oer=01%2f09%2f2008&osr=10%2f29%2f2007&plcgd=&plp=0&sf=10&sh=0&sn=&tg=0While the published default rate at the time was 2.80% (Oct 29, 07 – Nov 25, 07) and 2.98% (Nov 26, 07 – Jan 9, 08), today the default rate for loans originating in this time period is showing 4.27%.
The difference in ROI was not as significant. ROI was estimated 9.10% (Oct 29, 07 – Nov 25, 07) and 9.01% (Nov 26, 07 – Jan 9, 08), while the link above shows an ROI of 8.78%.
But, you must remember that many of the more decent loans (based on other criteria besides the PP criteria) were bid down below the minimum PP bid rate, so PP lenders did not join in these. (I cannot break out the performance page by interest rate.) Also, these loans are very young. With the link provided above, we can track the performance of these over time. I expect the default rate to increase and ROI to decrease. Who knows.
LAW 1: The absence of parent or pet owner in the borrower image is directly related to the underlying risk.
LAW 2: The attractiveness of the borrower image is directly proportional to the underlying risk.
LAW 3: The amount of clothing in the borrower image is inversely proportional to the underlying risk.
LAW 4: The value of the Borrower’s car compared to your own is directly proportional to the underlying risk.
