People gave me the "blood from a turnip" argument a year ago when I began complaining about Prosper's incompetent weak nonprofessional collections operation.
That was me. Keep in mind that when you read the following comments, they're geared to address the notion that simply replacing Penncro will significantly increase our returns. I do not believe that will be the case. I will be happy to be proven wrong. It will only benefit us all.
They were curing about 6% of 1 mo late accounts back then. Now they are curing about 20% (of loans recently gone 1 mo late). That's a big improvement. Apparently the turnip has some blood after all.
I disagree with you accounting, at least based on the numbers provided by Prosper here:
http://www.prosper.com/help/topics/lender-collection_agencies.aspxLook down at the Penncro section, and under "Last 3 months" the amounts brought current within one month in the three separate credit buckets are 7.26%/7.43%/5.6%. That's a far cry from the "20% of loans recently gone 1 mo late" that you mention above. (The numbers in the "lifetime" box are a bit better, but nowhere near 20%. At least not for the group you reference.) So, if you want to compare the most recent performance in the "one month late" category, the 3-month box is better. And, those numbers are very close to the 6% that you reference above.
Frankly, Prosper's early collections efforts back then were a complete joke. They are now a little better. There is no evidence at all to support the theory that the turnip now has no blood and they can't do substantially better than they are doing now.
Well... early on, we had no idea what the makeup of the old collections accounts looked like. Since Prosper now breaks the performance down by credit grade, we can only guess what it looked like before. (Compare apples to apples, if you will.) We had little data, and minimal guidance from Prosper. Sub-520 scorex borrowers were allowed to borrow. Back then, I very much postulated that some very very uncreditworthy borrowers were getting funded, and it would bite us in the ass. And, since we were so early in Prosper's existence, it's not unreasonable to assume that prime borrowers took a little longer to become delinquent. So, the early collections statistics were likely very heavily subprime. (Hence my blood from a turnip references.)
Now, months after the fact, Prosper cutoff the sub-520 scores, instituted lender guidance, and made a few other changes. It's entirely possible that the composition of the accounts sent to Penncro has changed. In fact, if you compare the "last 3 months" stats to the "lifetime" stats, by $, you will notice that B-D borrowers in collection make up 5% more of the "3 month" pool than they do in the "life time" pool. (In the lifetime late category, AA/A borrowers make up 10.5% of the pool, B-D 48.2%, and E-HR 41.3%. In the "3-mo" category, AA/A borrowers make up 10.9%, B-D 53.1%, and E-HR 36.0%) Furthermore, if you want to look at the "old" pool vs the "new" pool, one could presumably subtract the dollar amount that is three months late from the dollar amount in the lifetime category, arriving at the dollar amount that was sent to collections, but excluding the loans bucketed in the three month late category. Doing this, E-HR made up 54.6% of the value of the loans sent to collection, and B-D made up 36.1%.
If all of that is confusing, the summary is that is my attempt at showing that the makeup of loans being sent to Penncro now is of much higher quality than the makeup of loans sent to Penncro before. One would expect that better quality borrowers have an increased tendency to pay on their own, regardless of collection activity. Based on the change in composition of the collection pool, and Penncro's very marginal increase in performance in the one-month late category, it is my belief that Penncro's performance is in fact not better, and furthermore, that there is no evidence to support the notion that another collection agency will have statistically better results than Penncro did.
I believe we can do a lot better.
I sure hope we can do better. The best thing we can do is stop lending to lousy borrowers, and in fact, that is already happening. Changing the name of the company running the autodialer and mailing the letters won't change anything.
Do you honestly believe that there is nothing more sophisiticated that can be done beyond sending the late borrower a letter?
Not within the current constructs of the lending agreements, no. By that, I mean there is very little one can do within the 120 day period other than make phone calls, send emails, and write a few letters. The fun doesn't start until the delinquent loan is accelerated, and the borrower has the potential to be taken to court.
Have you ever been on the receiving end of a collection agency's business activities? I have, and let me tell ya, when ya got no job and no money, getting a daily call from the collection agency isn't exactly much of an incentive to pay. Oh, it's *real* easy to not pick up the land line (the number smart borrowers give out on credit applications, and can easily disconnect when necessary) when your buddies are always calling you on your cell.
Furthermore, I believe that Prosper's future depends on it. Not right away. They're still rolling on momentum from "Gee whiz there's this new kinda lending" news stories. That will last for awhile. Eventually if collections doesn't improve, there will be no one to invest.
Oh, I happen to agree with you there. I just happen to believe that the future of Prosper collections is going to be in something similar to the collection agency test that they're running right now. A situation where Prosper lenders benefit from the court action, and not the JDB.
In summary, I don't buy the "can't get blood from a turnip" argument. I just think its stupid and unsupported by the evidence.
What evidence do you want to see? It's an argument far from stupid. I *do* believe that four months of phone calls, letters, and emails are quite ineffective for the subprime crowd, and they won't pay us with money they don't have. The fun comes with garnishments and liens, something that the current arrangement doesn't permit.
So basically, I do agree that collections must be improved, and should be improved. I just don't think they can be done within the current "box" if you will. But I also hypothesized (and showed above) that as the market becomes less subprime, some improvements will occur naturally. A more prime market with more effective techniques will improve returns.
