Estimated ROI Update & Bug Fixes
June 1st, 2008 | News & UpdatesThis update takes care of a lot of loose ends from the previous release as well as a long awaited ROI update. Once again, thank you for all your feedback.
Lenders in some cases will notice a significant drop in their estimated ROI. This is because of several tweaks made to the ROI calculation including:
- Repurchased loans have been split into 2 categories: Repurchased and Defaulted (Repurchased). If a repurchased loan was repurchased for $0 it will now be reflected as a default. List of recategorized loans can be found here. (jazzpianist, Paul S.)
- Actual default sale prices are now used for calculating defaulted loans. Projected default sale prices will be based on historic sales and the percentages used can be found in the table at the bottom of the FAQ. (This table is updated nightly)
- Properly calculated the loan age for 4+ month lates, defaults and paid loans. (NewHorizon, bamalucky)
Notable bug fixes:
- Fixed a bug where lenders in a watch list sometimes disappeared after a data update. (ccarucci, Chocell)
- Fixed a bug on the lender analysis page where the extended credit data was not working correctly. (SpiritHelper)
- Fixed a bug where keyword searches were not finding every result. (Ken L.)
June 4th, 2008 at 5:56 pm
Thanks for the ROI fix.Now some of these blitzing top lenders may be bitch slapped into reality if they use your site.
June 5th, 2008 at 10:37 am
The ROIs are definitely more realistic now, imho. But I think assuming $0 loss for all loans that are current makes for estimated ROIs that are too high. Not sure what the programmatic/mathematical solution would be, tho’.
On a totally separate note…
I know groups aren’t nearly as popular as they once were, but what do you think about helping to show which are group leaders by changing their little blue person icon to something else?
June 7th, 2008 at 4:11 pm
I noticed one of the people I lended to is now bankrupt according to Prosper (which also consideres the loan as ‘current’).
Do Lendingstats calculations reflect bankrupt status?
June 7th, 2008 at 5:13 pm
Take into account I should say
June 12th, 2008 at 9:34 pm
From my observations, I think Prosper vastly overstates ROI for newer portfolios… Default rates are not linearly correlated with time passed since origination. Is there anyway to use historical data to make a default rate curve and superimpose that on newer loan portfolios, in order to arrive at a more realistic ROI?
June 13th, 2008 at 12:33 pm
On the borrowers-who-lend page, I think the lend-to-loan ratio isn’t being calculated correctly when the borrower already has an active loan.
I think your calculation includes the original loan amount of the active loan? I submit the current principle balance of that loan should be used instead. So as it stands now, this ratio ends up being lower than what I think it should be.
June 24th, 2008 at 2:09 pm
Regarding the chart “Total Loans Funded” at http://www.lendingstats.com/loansFunded …
What with Prosper not selling non-paying loans to JDBs, loans have stopped transitioning to the “defaulted” status for quite some time now. And so comparing loans funded to loans defaulted (as is done in this chart) becomes less and less meaningful, imho, as the number of defaulted loans remains completely flat.
What about, at a minimum, adding a third line to this chart showing all lates? (Maybe change the color of the defaulted line from orange to red - which would be consistent with other charts - and add an orange line for the lates?)
June 26th, 2008 at 4:35 am
newhorizon:
1) Current loans already have future losses ‘baked’ into the ROI. The way we estimate the future losses is based on the historical performance of a particular portfolio.
2) Group leaders should be distinigushable in the next release since the icon is already done.
3) “I think your calculation includes the original loan amount of the active loan? I submit the current principle balance of that loan should be used instead. So as it stands now, this ratio ends up being lower than what I think it should be.”
It should be correct the way it is right now if i’m understanding what your saying correctly. A borrower with 2 active loans should have the total amount borrowed used. (e.g. borrower has 6k lent, has 2 active loans 3k and 3k then the ratio should be 1)
- “What about, at a minimum, adding a third line to this chart showing all lates?”
The chart is actually showing all delinquent loans using the orange line (lates + defaults), I’ll look into adding another line for defaults. (It might just look like a flat line near the bottom so it might not be too useful).
micskill:
- If Prosper shows the loan as current status there is no way for us to write it off as a loss or count it as a bankruptcy unless it was manually done.
mike:
“use historical data to make a default rate curve and superimpose that on newer loan portfolios”
- very interesting idea, i actually thought of something interesting to do with this. (create a chart that shows a breakdown of loan statuses based on the age of the loan. This would answer questions like, how many months into a loan origination does the highest amount of delinquencies occur at?)
June 26th, 2008 at 11:17 am
“It should be correct the way it is right now if i’m understanding what your saying correctly. A borrower with 2 active loans should have the total amount borrowed used. (e.g. borrower has 6k lent, has 2 active loans 3k and 3k then the ratio should be 1)”
If I understand correctly, if a borrower lent 6K, has an active loan for 3K, and has a listing for another 3K, then you calculate the ratio to be 1 on “borrowers who lend” page. 6K / (3K + 3K) Do I have that right?
But if, for example, he’s paid off $2900 of that 3K active loan leaving a balance of $0.1K, then I submit that his lend-to-loan-ratio is actually 6K/ (0.1K + 3K) = 1.93.
(Regarding the “Total Loans Funded” I somehow simply got delinquent confused with defaulted. Sorry about that. I think the flatness of the delinquent line got me thinking the wrong way.)
June 28th, 2008 at 2:46 pm
Yes that’s roughly what I mean. I can calculate it by hand, but a lot of people probably won’t think of doing so.
I actually have a question…
Under what circumstances will Prosper actually transfer a loan into “default” from “delinquent”? This is important, because for tax purposes, I would like faster transfers into “default”. I’ll write them up later, if necessary, showing a gain. But this would defer my payable taxes and enhance my ROI.
On a similar vein, under what circumstances will Prosper sell the loan to a collections agency? Is it at 4 months delinquent, 6 months, etc? This is also important because fresher loans show much higher collection ratios than stale loans, and therefore garner much higher prices.
June 28th, 2008 at 2:59 pm
I guess I also want to expand on my previous point…
From my observations, it takes roughly 10-14 months for a particular loan vintage to become seasoned. Therefore, using prosper’s current ROI calculations, the stated ROI’s for portfolios newer than 8-10 months are VASTLY overstated (because as I said before, the delinquency ratio is not linearly correlated with elapsed time, but Prosper’s calculations assume that it is).
This may be too fine a point, but I believe that this is extremely misleading to potential investors, and may be one cause of some structural inefficiencies in the Prosper system.
(do you ever wonder why, hardly any of the top ROI lenders have portfolios with average ages longer than 360 days? In fact, I believe the top lender with at least 50 loans, with a portfolio age longer than 360 days, has a ROI of 14.45%… )
It wouldn’t be that hard to fix. The portfolio performance should probably start at the aggregated “estimated ROI” shown when making a bid on the main Prosper site. Then you could take actual data and adjust that ROI through time, so that after a few years pass, the actual data dominates the final result.
October 16th, 2008 at 2:19 am
tramadol…
online tramadol. tramadol hydrochloride. tramadol hcl. buy tramadol online. cheap tramadol. buy tramadol online. …