Friday, March 31, 2006

Living in interest-ing times, standing up for principle (or is that my principal), and other punny clichés best left to the dogs...

Alternatively, The netizens are restless.

Several interesting things happened today on Prosper.com (in order of significance to me as an investor, not in order of how riled up they made me or others):
1)Default rates have now been affirmed to be annualized. 'nuff said.
2)The rate cap on loans has been changed to 24% from 36%.
3)It was explicitly reiterated that group leaders are not to use any credit reportage stuff at all to determine membership in groups (see below for details)
Other fairly good site-design, data provision, group ranking algorithm stuff got lost in the furor. HTML explosions have finally been banished from messaging, apparently…

Side notes:
I also got to speak over the phone w/Chris Larsen (CEO of Prosper, founder of E-Loan, and, in the latter capacity, the first person to push FICO (credit) scores out to the public at large) for a fairly decent chunk of time today, which was a bit odd. I mean, my day job allows me that kind of access to CEO’s and CFO’s all the time, but it is fairly weird when my track record of posting on an internet forum brings that about…

I don't think that I, as a fairly discerning investor/lender, will be much hurt by the changes at Prosper in the short run (or even that much in the long run, as long as Prosper survives). Turmoil tends to benefit the paranoid, after all. If anything, despite my oft-stated pursuit of high risk loans, my weighted average interest rate is 26% or so (I haven't been loaning simply by the numbers, except in ~.25 of my original, smaller allocation to Prosper -- which converts to ~6% of my current allocation), and I've been investing a fair amount in people constrained by rate caps under 24%, but otherwise well suited for loans.

1)From the very start, default rate data has been handled rather poorly -- and sadly enough (for others, since I’ve never assumed otherwise), my worst case analysis was right. That is, default rates given are annualized. This actually has a more concrete effect on lenders and borrowers than anything else -- we've been told that they're for 3 year terms, then 2 year, then 2 year but annualized, etc -- when one is investing in poorer credit, this is a HUGE difference in expected returns.

2/3) I've already said a lot on the subject on the Prosper forums
(and relearned, for the umpteenth time, the valuable lesson that one shouldn't mix math and being 'riled up).

I primarily think that the rate cap change was:
a)handled poorly/disclosed poorly/discussed poorly (though this was perhaps intended to prevent listing runs)
b)sucks horribly for HR's who don't happen to perfectly fit the "join the affinity group to lower your rates" model, especially in light of reiterations that group membership is not to be conditioned upon credit reports.
c) interesting re: "Group leaders may not obtain or use a credit report, require prospective members to provide a credit report, or to provide authorization to obtain a credit report, as a condition to group membership" This seems to leave open the possibility of asking for reports, or more properly, anonymized portions of reports, under other circumstances. For instance, so long as group membership is not conditioned upon credit report, after you accept someone, can't you then ask for a partial/anonymized report when deciding how heartily to endorse someone who is already a group member?

Or, if asked by a lender who is not a group leader? Or, verging close to the spirit-vs-word-of-the-law-line here, if spontaneously sent to group leaders by borrowers (since action upon such things is impossible to ascertain)? For the record, I'm out of the group leader game, except for people I know from the off-line world, so please don't send my unsolicited reports or assume that I'm trying to skirt the rules for my own benefit here. I'm merely hoping to probe (as I did before with default rates and a million other things), an area that remains foggy, and hopefully, will be better clarified by Prosper.

I was overly (though again, perhaps less than most) impressed that Chris Larsen ended up giving various lenders and/or posters (including myself) calls over the course of the day, and that he was surprisingly responsive (and unsurprisingly, a little hoarse by the end of the day). [Actually, to clarify, I was impressed by his responsiveness, especially in light of various news reports that paint him as a fairly public interested but outspoken/stubborn guy w/ “fire in his eye."]

(Please comment below if you also got a call – I’ve only been able to speak to one other person about it, and I’m very curious about what else he said – or else click on my profile and email me.) Of course, I rapidly reminded myself that doing spin control is clearly in his best interest (as a large equity stakeholder), and that he only did so in the wake of the massive (for the internet) furor amongst lenders. And, somewhat damningly, he still hasn't chosen to make any public announcements about the motivations behind the rate cut, but has instead has made various oral comments to various lenders, perhaps with the ulterior motive of allowing calming but uncommitted comments to emanate outwards and become increasingly disseminated, while being fairly able to distance himself from the results. (Again, I’d be pleasantly, and only mildly surprised if he did put out some sort of message onsite, but I’m still waiting for it – and I suspect it would have helped matters greatly had the change been given the prominence it deserved, rather than buried in a bunch of other updates within a forum posting.)

A rough recap of his comments, again in descending order of significance:
1.Prosper is very concerned about the influx of HR 36% instafunding money that seems to be adversely selecting (I'd say particularly disturbing to me was an E rated guy I previously noticed, whose post literally consisted, in its entirety, of "I need $10k -- I'm willing to pay the maximum rate").Chris wants to protect lenders, and move the dominant model back toward that of HR's seeking credit enhancement through groups. He admits that the math makes it fairly hard on HR’s, but credit enhancement via affinity group could make it happen.

Fine, but really, this isn't the ideal way to address it. The rationale is plausible, but from my "what if I were living on Planet (Traditional) Economicus (where everyone's an amoral Einstein w/perfect self control, and no one does anything, ever, except out of self interest)," I'd give it a 55% honesty rating. (I’d maybe give it 75% ‘cuz I’m not that cynical, and Chris Larsen the guy is probably a resident of Planet Earth. Plus, on Earth, I’d be underweighting the genuine possibility that having made a fair amount of money, his utility starts getting more and more enhancement from public interest-y success – not a one to one trade, but still...)

a)After all, guess what – the sketchy HY loans are not funding fully anyway, so intelligent lenders seem to be aware of the problems w/blindly going into such auctions. And, the less sophisticated/well funded ones aren’t filling those auctions regardless.
b)Prosper's going to end up with a whole lot more HR 24% instafunding money that seems to be adversely selecting, so if they're trying to clear the pool of such people, it ain't gonna work that quickly by only reducing rates (which doesn’t rule out other silent screening measures they may be implementing in conjunction with this one, of course). If anything, it tends to make it harder to separate the sketchy from the nonsketchy by compressing them even closer together. Also, what happened to the old measure of nonsketchiness – loan history? Look for the guy who’s been relisting repeatedly, increasing interest rate gradually, and you get a pretty good sense – though again, you might not catch the nonsketchy but desperate.
c)Given the default rate imbroglio, who was bidding on any auction w/a real sense of expected returns, with or without adverse selection? C'mon -- we've gone from being told that HR default rates over 3 years are 19.1%, to 47.05% (making various basic assumptions, yadda yadda).
d)While he didn't broach this topic, I already was aware of an upcoming ABC special on Prosper. From a business angle, the rate cap remains perhaps understandable, in light of the expected effects of the program, which is sure to bring in a deluge of people (not just Georgians from the Clark Howard show). I’d still argue that more proactive management could have prevented much of the ickiness that's expected to ensue, and much of the ickiness that this change could have been predicted to bring about.
e)Prosper still seems fairly fixed on affinity group model, rather than verification-oriented groups – just an observation. I still think there’d be plenty of economic value to the Prosper investment club system – every month, 25 lenders each come up with one vetted loan, and everyone invests $50 in each one. No one busts their butt to do due diligence, but diligence and diversification still occur.

2.He believes that HR credit enhancement can occur through groups, and is more than willing to consider the possibility (my nonexact take on his phrasing/presentation, though he is undoubtedly motivated to give the impression of seriously considering this) of raising rates in the future -- after the model settles the way he, and other microlending-inspired individuals, believe it should settle. Comments already on record.

3.A group of features that he emphasized could be realized in the near to medium term, that I already think are great ideas/have asked repeatedly on forums for, and that he wanted kept confidential for now. I’ll respect that. By the way, I spoke over the phone w/one of the other lenders he spoke to – w/o breaking confidentiality, we were able to rapidly ascertain that we’d been told about the same things, and to keep them confidential.

Some speculation, within confidentiality:
a)He genuinely wants some of the loudest posters to know (or believe) that these changes are in the works, so that they’ll calm down (or, in that kraazy internet lingo that all the kids are talking these days, STFU).
b)He knows that most people told to keep mum won’t, and as a result, he’ll have all the benefits of positive publicity with none of the drawbacks, given the hedging potential of plausible deniability -- in case the desirable changes fail to come to fruition/take a lot longer to materialize, or are spun from thin air. The latter seems overly cynical – if nothing else, I’d assume that he’s a good enough businessman to recognize good ideas when he sees them.
c)I need to start reapplying more of my analytical tendencies to my real work.

4.I now know which user Chris Larsen is – and no, I’m not telling. And no, it isn’t skibum. And yes, it wasn’t a big surprise (to me, at least, but then, I over-research everything).


One other thing – the update has changed group ranking metrics. I’m a fairly big fan of this, as I’ve already pointed out how easy it is/was to game the system, if one were willing to forgo the interest on a decent chunk of money during the desired high-ranking time – which I was. Though, I’d also point out that my group was pretty damn good, and I felt that it lived up to its gamed ranking.

And, another comment. I'm well aware that I'm developing the rep of being Homo Economicus. However, I'd add that I'm, if anything, a behavioral economist. Do I believe that various behavioral tenedencies (affinity effects, reciprocity, etc) can be leveraged so that much of the dead weight loss from an artificial rate constrains can be sidestepped? In microlending contexts, sreu, but in the substantively different environment of Prosper? Well, not without more evidence. But would I bet against it? No. Would I rather 24% be the cap if that would work -- well, yes, but I am a bleedign heart liberal, right?

Finally, I'm skipping all the business analysis related stuff -- I clearly believe that Prosper can survive w/24% rates and no HR borrowers (neither of which I'm assuming wil necessarily be the case, nor can I make irrefutable arguments for), and I understand why as a business looking to dominate the marketplace, forgoing some of the extremes makes sense -- so would anyone else, I'd hope. What Prosper really needs, again, is better disclosure -- c'mon, do you have to make the Prosper forum boards your drug of choice in order to discover this stuff? There's gotta be a better balance between insane lack of transparency (again, see default rates) and then having the CEO call up individual lenders (again, definitely appreciative here) to deal w/the backlash.

Edit 4/1/01: Eventually, a Prosper representative made an official forum post basically affirming the main points of what Chris Larsen mentioned, including upcoming improvements -- specifically, various opt-in measures for borrowers to provide more summary information for lenders to view (along the lines of my "derogs and negs" summary). Hopefully, this isn't Prosper's "April Fool's!"


Read more!
/body>