Income verification and self-reporting

So, one of the cornerstones of Prosper.com lending is the Debt-to-Income Ratio, or DTI. It roughly takes all of a borrower's minimum monthly payments, adds them together, and then divides that number by the borrower's monthly income. Because income is self-reported on Prosper, many lenders complain that its usefulness is nonexistent. While I certainly agree that there are various tweaks available to the existing system, I think that DTI is good enough for most loans, and that we're hardly bound for a lemon market. There are plenty of signalling possibilities -- if nothing else, income verifications can be performed with the help of the borrower and his or her employer.

Further detailed thoughts on the matter:

1.Sometimes people are honest -- even online.
2.Many people assume prosper has ways of checking your info out that give it, and big info-sucking corporations a little too much credit. This is not entirely without basis -- I, for one, was surpised to find that the year of one's car for a car loan is embedded in your credit profile data (and not like, account name=TOYOTA 2005 LOAN) -- and this datalet becomes a credit profile confirmation at times.
3.Many people fear being caught lying, maybe more than they should -- probabilistically speaking, not ethically.
4.Others would lie, but rationally believe that the upside isn't worth the risk of being caught.
5.Spot checks help -- knowledge of spot checks (even exaggeration of frequency) help even more.
6.Consider it a form of resume inflation -- just another cost of doing business in America. People tend to exaggerate, but most people do it in proportion to their starting point, and are rationally cautious about doing so, so just scale down the resume-equivalents a notch and lender estimates will be roughly right.
7.The most egregious examples tend to stick out like a sore thumb, even if prosper fails to catch them.
8.Standard practice in the credit card industry, etc, is to not verify for credit lines of 25k or less; insofar as income exaggeration contributes to default rates, the statistics for default that we see already take this into account, to some small degree.

My (sometimes) immodest proposals:
1.Prosper states that it is a felony when you lie on their application (but there's obviously minimal risk of prosecution). I suppose they could make it much more prominent, like in SCREAMING RED LETTERS w/vintage 1998 website cool WAV music and dancing animated GIF flames, along with a clickthrough path that requires you to click 20 dialog boxes.

Or, make an example of some guy who rounds up his income from 97,550 to 98,000 -- throw the book at him, Johnny!

2.Prosper spot checks should be/are probably focused on outliers -- particularly (numerically) unlikely combinations. Not sure what these criteria are, but I assume a HR individual w/ 300k annual income in the back woods of Michigan (don't get me wrong, I like back woods, I like Michigan, some of my best friends are from Michigan) will get the hairy eyeball more than the 80k income guy w/a C rating in midtown Manhattan. Prosper -- the kindlier, friendlier tax auditor you've always dreamed of.

3.More important than the existence of random spot checking, I think, would be the clear and prominent announcement to all borrowers that random income checking occurs and will result in loans (and teeth) being pulled. As I recall, when I checked out the borrowing process side of the equation, the announcement was either not present (beta site), or else it was not particularly prominent -- someone who has more recently submitted a loan request might want to weigh in here (I was, and will soon be, purely a lender). Insert Orwellian site announcements, "Lender cellardoor has been caught falsifying his aerobic exercise time to monthly mortgage debt ratio -- he will be forced to pay off his house by doing sponsored walk-a-thons. We know your IP address -- you could be next, "and you're done.

Strictly speaking, at that level, who needs actual income checking? A sufficiently strongly worded disclaimer that income checking does occur will have the same deterrent effect (though perhaps this would be illegal insofar as lenders are misled as to the degree of assurance they can have in the markets).

After all, take personal tax audits for people who aren't filthy rich -- presumably, the deterrence effect is far more valuable than the actual value of successful (positive) revisions to non-rich people's tax payments due to audits. After all, the main problem with pretending to audit w/taxes is that taxes are frequently discussed and done every year -- a policy of prominent announcements of audits w/o concommitant actual auditing, would rapidly be discovered by taxpayers (not to mention that IRS workers wouldn't all be able to keep a secret, etc). However, the borrower community is much more dispersed, and less likely to be as persistent as the lender community.

4.Every credit card app I've seen typically has some sort of "you attest that everything you say is true, and you acknowledge that there can be criminal penalties if you lie" clause. Anyone know if this might be the equivalent of a "ski lift clause?" The story goes, lots of ski lifts have people sign papers waiving their right to sue the ski lift owner, no matter what happens to them; typically, signing such a document is legally meaningless/unenforceable, but still dissuades many people from suing in the wake of minor (or even not so minor) injury. I mean, you get the deterrent effect even if there's no real legal penalty to lying. On the other hand, I'd assume that large corporations like AMEX, don't typically use ski lift clauses. Not because they wouldn't even if they could get away with them, but because the marginal benefit isn't worth the headline risk.

This is Emmanuel Goldstein, signing out.