Lots of people are talking about race and the mortgage crisis as a result of a New York Times article over the weekend. Now I have no problem thinking of the mortgage crisis as being, in the words of David Harvey, ‘a financial Katrina.’ The devastation in inner-cities has been massive, and the results are still happening:
This is Cleveland; this graph can also be created for Detroit, Baltimore, etc. It is easy to think of it as an urban crisis as opposed to a subprime crisis. People are going to be debating this for a while. I want to point out some high end issues to keep in mind when you think of this, that builds on some of the issues we’ve developed here.
Racism or Entitlement?
Let’s make up two worlds to show how hard this is to investigate with easily obtained data. In the first world, a Commissar from ACORN goes to look at the mortgage book of a subprime lender. Disappointed, he says: “Comrade. Your state mandated obligations of PC loans to poor minorities is woefully low. I want you to write twice as many loans to poor minorities as you do to whites over the next two years, or I shall inform the tin mines of Kolyma that they’ll have another worker starting next week.”
In the second world, corrupt, racist and all powerful subprime lenders force minorities to take loans with terrible prepayment penalties and option arms. Like 80% of subprime loans, minorities refinance their loans every 2 years, paying huge prepayment penalties, in order to avoid the harsh resets. Similarly risky whites have the same initial rates, but their resets happen after five years with a minimum jump, and no prepayment penalties.
So what does this look like in summary statistics? They look exactly the same – minorities get double the loan volume of whites. This is because the representative minority in model 2 takes out his loan twice, because he refinances it. This is key; if we want to look at individual situations, we need to follow individuals, not loans, because the loans recycle so often. We need statistics related to the velocity of the subprime loans themselves, but few, if any, statistics exist like this. Back in the day of the Fordist contract, this wasn’t necessary. We knew when there was prepayments – in 10 years as family upscale to accommodate new family members, and 10 years as families downscale as children leave the house. Now, as consumers are meant to treat their mortgages like interest rate swaps, the old mass measurements have no meaning.
Another issue to watch for in studies is LTV, or loan to value of the houses in question. Studies need to control for LTV if they are going to imply that defaults are the result of irresponsible lending or irresponsible borrowers – if they do not, they are not worth your time.
Your mortgage is $100. To sell the house costs you $1 to pay for a lawyer. If the value of your house when you lose your job is $106, you’ll pay the $1 to sell it, netting $5. If the value is $95, you’ll lose money selling it, so why even bother paying the $1 to the lawyer. Mail in the keys. It’s the economically rational thing to do. So if defaults are high in neighborhoods with high LTV, we shouldn’t think of the people as victims of evil bankers or victims of some sort of Moynihanian cultural boogeyman, but simply people who take Greg Mankiw’s textbooks seriously.
(I know I can hear you saying that’s a terrible thing to do. But I’m just a humble 21st century digital boy, I know no other way to value things. If neoliberalism is about, in the words of Wendy Brown, “extending and disseminating market values to all institutions and social action … the production of all human and institutional action as rational entrepreneurial action,” then that has to cut both ways. Consumers are going to treat their mortgage obligations and bank counterparties like a hedge fund treats a company it is shorting.)
Why would banks discriminate? There’s an argument that markets should clear the floor of racism, since if it is profitable to not discriminate non-racist businesses will drive out racist businesses. The obvious counterargument is racism is profit-maximizing if patrons are racist, not (just) the businessowners – ie, if patrons of the majority in your restaurant won’t eat with a minority, even if it is very profitable to serve minorities you won’t. This doesn’t seem like a good argument for banks, whose patrons are shareholders, who just want returns.
James Kwak (who if you aren’t reading you should be) wonders if minorities have a higher ‘search cost’, and banks know this, and they discriminate and collude accordingly. He bases this off some research which finds this is a good explanation for discrimination in car prices.
I think that part is a stretch. I think that the market for the initial subprime loan was (probably) competitive. The problem comes when the refinancing begins. As we’ve covered before, 80% of subprime loans refinance during the first 30 months, in order to avoid a nasty hit on the mortgage reset. At that point, there is absolutely no reason to believe the market is very competitive. There’s an information asymmetry that exists that consumers can’t likely overcome. It is easy to imagine the banks saying “all is fair on initial clients, but no messing with the fishies someone else can pull into their net” – collusion. And the desperate nature of the clients and the situation (think of Andrews, from the nytimes magazine article, calling his broker to desperately refinance his mortgage) leads the banks to a place where it is easy to gouge on terms.
Looking at summary data, it isn’t easy to determine what kind of loan is what, and teasing out this problem isn’t easy (see document page 17, footnote 24, for more on this here.) I know of no research that covers it. But it is an obvious point for people researching this.
The Obvious Discrimination
Not to hit on a point that I’ve been wailing on for a few weeks now, but I think prepayment penalties are the real place to start with this discrimination search. If a minority and a non-minority are otherwise identical, and the minority has a larger prepayment penalty and a larger reset on their arm, discrimination. Done.
Now here is the fun part – if a minority is riskier, and a non-minority is less risker, and a minority has a larger prepayment penalty, and possibly on the larger reset on the arm, discrimination. Done. Prepayment penalties have nothing to do with the credit worthiness of the borrower – they cover the negative convexity of the bond, which is an issue related to re-investment risks and interest rate risk, not default risk. If anything, it should work the other way around with the riskiness of the borrower – the less risk on the borrower,the higher the prepayment penalty. Since prepayment penalties *should* compensate for the risks on the negative convexity of reinvestments on the market side, not the risks of defaults, then why do all these minorities have prepayment penalties?
I don’t see any reason to think of the hike in the arm reset, where minorities have mortgages that jump up 5% after two years, as a compensation for credit risk- but I might be missing something there. Certainly a higher jump should translate into a lower initial rate. And we have a competing model developed here that explains what is going on there with banks taking bets on minorities. I’m pretty sure this line of argument is a slam-dunk for any lawyer or social researcher looking over the data.
There will be more as time goes on, but these are some items to hopefully guide your own interpretations.