New Cars, Mortgages, and Race

Like most forms of hardship in our society, the foreclosure crisis is disproportionately affecting minorities. The New York Times conducted a study of foreclosures in the New York area and found, among other things:

Defaults occur three times as often in mostly minority census tracts as in mostly white ones. Eighty-five percent of the worst-hit neighborhoods — where the default rate is at least double the regional average — have a majority of black and Latino homeowners.

Well, that might simply be a function of poverty: statistically speaking, minorities are more likely to be poor, and therefore more likely to become delinquent on their mortgages. But I don’t think it’s that simple.

First, whether you are likely to go into default is a function not just of your income, but of both your income and the size of your mortgage. As Tanta wrote,

The belief that subprime borrowers are “poor people” has taken root so deeply that you need a jackhammer to rip it out. The capacity C of traditional underwriting was, of course, always relative to the proposed transaction. A lower-income person buying a lower-priced property was, you see, not a case of subprime lending; assuming a reasonable credit history, it was a prime loan.

Let’s say you start out with a middle-income person who, like most, might be vulnerable to an economic downturn. There are a couple things you can do that will make him more likely to go into default when the downturn (or an illness, or a divorce) hits. You can push him into too big a mortgage. In Tanta’s words,

The trouble with the low-income prime loan was that it was a small prime loan. And that there were, in many market areas, more lower-income people than lower-priced properties. Both industry greed—wanting to make the biggest loans possible to make the biggest profits possible—and industry overcapacity, combined with ever less-affordable housing in the employment-rich population centers, brought us to a situation in which we might not have started with poor people, but they were certainly poor by the time we got done putting them into too much loan to buy too much house. There are subprime borrowers you find. There are those you create.

You can also give him a higher-rate mortgage than he could otherwise qualify for. According to the Times article:

Roughly 33 percent of the subprime mortgages given out in New York City in 2007, [Secretary of HUD Shaun] Donovan said, went to borrowers with credit scores that should have qualified them for conventional prevailing-rate loans.

In general, high-rate mortgages account for a larger proportion of mortgages in minority communities, even after taking median income into account. According to sociologist Gregory Squires,

We see these loans heavily concentrated in poor neighborhoods and targeted to minority neighborhoods. There is some evidence that these neighborhoods were actually targeted — that lenders have gone after people whom they think are less sophisticated borrowers, including single women and the elderly. . . .

Credit rating and income would and does explain some of the patterns. But when you control for those, segregation is also a factor. . . . In those metro areas where segregation is highest, the share of loans that are subprime goes up.

If you are disproportionately steering minority borrowers into higher-rate mortgages, then of course they will suffer a higher rate of defaults and foreclosures than you would predict solely from their other characteristics (income, credit rating, etc.).

But why would you do this? I mean, it’s obvious why an individual broker would: higher commissions. But why would a competitive industry as a whole do this?

In principle (by which I mean first-year microeconomics textbook principle), the free market should eliminate most forms of commercial racism. If a minority homebuyer is a good risk at 6%, then Good Bank willing to lend at 6% will get his business, and Bad Bank who insists on 9% will lose his business; over time, Good Bank should drive Bad Bank out of business. There are probably some doctrinaire free-market people who would insist that this in itself proves that racism does not occur: if banks are charging higher rates to minorities, then the existence of the free market proves that those minorities are higher risks.

But there is ample evidence of racism in areas where the correlation of race and credit risk is not an issue. Almost fifteen years ago, in “Race and Gender Discrimination in Bargaining for a New Car,” Ian Ayres and Peter Siegelman sent pairs of testers into auto dealerships, armed with identical back stories and identical bargaining scripts, and even wearing similar clothing. (All testers said they would provide their own financing for the car.) Black men were initially offered twice the initial markup (price minus dealer’s cost) as white men; after negotiation, black men ended up paying three times the markup of white men.

Ayres and Siegelman discuss two sets of theories why racism might occur. The first set is based on animus: car dealers don’t like minorities and therefore treat them worse, even though it is bad for their bottom line. The second set is based on statistical inference: car dealers might think that being a minority is likely to be correlated with higher search costs or lower access to information, either of which might make the buyer more likely to accept a high asking price.

In the case of mortgages, search costs might have been a factor. Perhaps an equilibrium developed where minority neighborhoods were dominated by high-interest lenders, so if you were going to enter that neighborhood, it made more sense to charge the going rate and split the high profits with your competitors than to undercut them and trigger a price war. Or perhaps there are information disparities that correlate with race, even after controlling for income. Taking advantage of people because they are poorly informed, as far as I know, is completely legal.

In retrospect, this seems like an area where better consumer education could have played a role. Back to the Times:

Upper-income black borrowers in the region are more likely to hold subprime mortgages than even blacks with lower incomes, who often benefit from homeownership classes and lending assistance offered by government and nonprofits.

Maybe the economic debacle we are all living through will lead to better personal finance education, both in school and for adults. That’s one silver lining to hope for, since an end to racism is almost certainly far off.

By James Kwak

24 thoughts on “New Cars, Mortgages, and Race

  1. An alternative explanation is that, once the ability to repay became of no concern to the brokers, they embarked on a search for anyone and everyone likely to sign whatever papers were put in front of them.

    If you really want to make a charge of racism that passes the laugh test then you should be able to show that mortgage brokers and lenders shied away from pushing white people into ill-advised subprime loans when they had the chance. This idea, given what has happened, is nonsense on its face.

  2. Ah yes, the traditional five Cs of lending: Capacity, Collateral, Credit, Class, and Color.

    I hate to be “doctrinaire”, but this sounds like a massive market inefficiency which one would think somebody would exploit eventually. Why haven’t they?

    “Economic illiteracy” and “information disparities” do not explain it. Even “racist brokers” does not explain it. Eventually, honest merchants and brokers should appear to steal the business, even if the customers are stupid. The only explanation you propose:

    Perhaps an equilibrium developed where minority neighborhoods were dominated by high-interest lenders, so if you were going to enter that neighborhood, it made more sense to charge the going rate and split the high profits with your competitors than to undercut them and trigger a price war.

    …is possible, I guess. But collusion between all of the lenders in every neighborhood?

    Do these statistics hold up when you normalize for income?

  3. Ohhh… You mean “economic illiteracy” to the point of not even recognizing a better deal, so even if one is offered, the consumer will reject it?

    Also possible, I guess.

  4. Nemo: ““Economic illiteracy” and “information disparities” do not explain it. Even “racist brokers” does not explain it. Eventually, honest merchants and brokers should appear to steal the business, even if the customers are stupid.”

    The trouble is, “eventually” can take hundreds of years. :(

  5. so what does this tell you?

    “Roughly 33 percent of the subprime mortgages given out in New York City in 2007, [Secretary of HUD Shaun] Donovan said, went to borrowers with credit scores that should have qualified them for conventional prevailing-rate loans.”

  6. I believe there’s some problems with financial education’s effectiveness as well, perhaps because there isn’t a non-partial source for it. Often times, the education materials are provided by the banks themselves, so there may be some self-interest in NOT giving enough information.

    Perhaps we should try to create more community credit unions?

  7. That’s exactly what I was thinking, Min.

    Nemo: This is precisely the response I envisioned when I read James’s “doctrinaire” (I prefer to refer to it as market Panglossianism, eg: Since the invisible hand is good and omnipotent, and since it chose this equilibrium out of all possibilities, this equilibrium must be good—in fact, this equilibrium is the best of all possible equilibriums).

    Could it possibly be that the springs that pull this system back into equilibrium are a bit rusty when it comes to traditionally underserved minority neighborhoods? In fact, so rusty that they can’t possibly respond within the timeframe of the housing boom-bust cycle? So instead you get cycles of opportunistic lenders and no lenders, never quite making it to the equilibrium where these people are well served.

  8. Alternative causes for higher minority mortgage defaults exist and data needs adjustment to make NY Times’ claim statistically valid.

    Minorities’ percentages of home ownership were increasing so have greater percentage of minority homeowners with mortgages and with higher remaining principal balances as percentage of home value and in dollars than in non-minority group.

    Unemployment, loss of income, and uninsured medical expenses are also primary causes of mortgage defaults. Minorities have higher unemployment rates during this recession and are more likely not to have health insurance. Also, maybe more likely to work in jobs where hours are cut backed in recession and their weekly wages have declined.

    Need to compare minorities against similar group. Need to adjust non-minority default rate for lower unemployment, lower mortgage balances and lower unexpected medical expense costs.

    When two groups are balanced, default rate will most likely be the same, which is why just about every time Fed researchers look at problem they do not find a minority, mortgage discrimination problem.

    Higher mortgage foreclosures are not as much due to a subprime problem as it is due to other social and economic factors.

  9. It would be “doctrinaire” would be to assume that no racism exists in lending (thanks to the invisible hand). It would also be “doctrinaire” to assume a priori that no behaviors could possibly be correlated with race.

    I myself am doing neither. I am simply pointing out that whenever your claims imply that there is money sitting there for the taking but somehow nobody is choosing to take it, then you have some explaining to do. And I have yet to see much in the way of explanations.

    Is it possible that some people paid higher interest rates because their loans were worse credit risks, and those loans are now defaulting also because they were worse credit risks? (The quote about credit scores is interesting, but credit scores are not the only thing that determine credit risk. Loan-to-value and debt-to-income ratios also matter.)

    Again, how do these numbers look once you normalize for income?

  10. I’m reminded of the old economist/young economist $20-bill-on-the-street joke… the money isn’t sitting there for the taking. It’s hard to get and real businesses have lower-hanging fruit than to get into a price war with the incumbent loan shark over a couple of points on a low-income mortgage. There are a million obvious reasons why this market might be inefficient in the real world. I’m not claiming racism is necessarily the main one, just that the statement that people who present solid evidence for it have to also explain why the invisible hand failed in this case is silly. Equilibrium theory comes with some very strong assumptions and when you relax one of them just a bit you can get weird stuff. Pretty much all of them are taking a nap in this market.

    Also, the quote says that the Gregory Squires research does control for both credit rating and income, so I’m not sure what you want there.

  11. Always loved that joke. Out of curiosity, when exactly was the last time you found a $20 bill on the street? (In my case, the answer is “never”, but perhaps I am just unlucky.)

    To stretch the analogy a bit: We are not talking about a single $20 bill on the street. We are talking about something systemic. If you came across a street covered in $20 bills and everybody was walking by without touching them, would you bend over and start collecting? After all, the market CAN be inefficient… But personally, that would not be my initial assumption.

    I seriously doubt we are talking about “a couple of points on a low-income mortgage”, because if we were, there would not be so much discussion about this issue.

    OK, so I should have read the background material more carefully first. To be precise, here is what I want: I want to know whether high-interest loans, when made to minorities, are less likely to default than similar high-interest loans made to non-minorities. Or, I want evidence that the higher rates of default are entirely an effect, not a cause, of higher interest rates.

    Convince me that the higher rates of interest are due to some factor other than efficient pricing of the risk of default. Again, “credit rating and income” are not sufficient to capture credit risk.

    I realize these loans were made disproportionately to minorities. So demonstrate that these same minorities are not disproportionately poor credit risks. Show that actual rates of default are lower among minority sub-prime borrowers. That would be compelling evidence. (And it may well be possible. But I do not see it among the references cited here.)

  12. I truly think it comes down to level of financial sophistication…and unfortunately, there isn’t enough of that in lower income neighborhoods. I would guess the number of first time minority home owners jumped with the housing boom. The first time you do anything, you’ll be more willing to listen to and believe what you’re told. Does race play a part? I can’t see how it doesn’t in some cases but my guess is the main driver here is exploitation of the under-informed. What would be great is if every credit application/conversation from here on out was recorded…and instances of fraud would negate the contract.

  13. Once, when I was ~12 or so. I picked it up. I’m pretty sure collecting the surplus from competitors’ overpricing of mortgage loans for low-income customers would have been slightly harder, but perhaps I’m just overestimating the difficulties inherent in running a business.

    The difference between our perspectives seems to be that when trying to apply a highly stylized and largely empirically disproved theory, my initial assumption IS that it doesn’t apply in a complicated real-world situation, unless I can convince myself that at least some of its many stringent assumptions are met. Equilibrium theory is a cute rule of thumb to talk about broad market dynamics, but it’s not scientific law by any stretch.

    So, I guess we’re at Occam’s razor time:

    I find it more likely that the market for financial services in low-income communities is distorted away from equilibrium by a number of factors including things like systematic under-provision of (all types of) services and probably a bit of good, old-fashioned racism.

    Your position is that in fact the people giving out these mortgages were properly calculating the inherently-higher default risk above and beyond their credit scores and income levels (perhaps through good old fashioned community banking practices?) and it just looks like discrimination when really it’s good lending practices representing a market in equilibrium… no?

    I guess there is a point where the “price” in this market is so fuzzy as that we can convince ourselves that whatever rates the market comes up with “reveal the true risk” of these borrowers. Certainly lenders who charged customers as if they were more risky than the rest of the market thought they were from plugging in FICO and income look pretty smart today. My gut is that these folks were shady and taking advantage of the population… compared with the argument that they were actually the smartest mortgage brokers around, it just seems more likely to me, but YMMV.

  14. For the record, my position is that I do not know, based on the evidence presented here.

    The answers do not need to come from ideology. Certainly a lot of these loans were made in the past few years, and they are defaulting in significant numbers right now. So it should be possible to answer: For loans made at interest rate X%, what default rates are being observed for each racial group?

    If lenders were over-charging minorities — i.e., over-pricing their credit risk — then high-interest loans made to those minorities ought to be defaulting at a lower rate than similar high-interest loans made to non-minorities. And there should be plenty of data available at this point to make or to refute this case.

  15. Additionally, credit scores are not a very good way to judge what a loan’s interest rate and default rate should be across demographic groups. Credit scores have more validity for assessing risk of default within demographic groups.

    An August 2007, a Federal Reserve study, “Report to the Congress on Credit Scoring and Its Effects on the Availability and Affordability of Credit” received a lot of press.

    The report found, “Consistently, across all three credit scores and all five performance measures, blacks, single individuals, individuals residing in lower-income or predominantly minority census tracts show consistently higher incidences of bad performance than would be predicted by the credit scores. Similarly, Asians, married individuals, foreign-born (particularly, recent immigrants), and those residing in higher income census tracts consistently perform better than predicted by their credit scores.” (p. 89).

    PDF version of report

    HTML version of report

  16. James says in his post, “In principle (by which I mean first-year microeconomics textbook principle), the free market should eliminate most forms of commercial racism.”

    Is there principle in business? Seems not, at least not in the housing industry.

    Is there racism in the mortgage industry? I don’t think the stats cited by the NY Times tell a complete story.

    Are the minority neighborhoods harder hit by layoffs? If so, that’s not an indication of racism at the origin of the loan.

    How much down payment did these buyers put down? Were they in reality renting via the sub prime mortgage?

    Is there inequity in availability of affordable credit in lower income neighborhoods? There seems to be much that is not available in low income neighborhoods, libraries, stores, video rental stores, etc. Is that racism or the impact of poverty?

    Coming on the heels of the NY Times piece by the (white) NY Times economics reporter who wrote about his own personal economics crisis, the subprime crisis does not seem racial in its origins. (Unlike redlining, the racist process of limiting access to homes due to race.)

    In NYC and throughout the country, people wanted more house than they could afford.

    The mortgage industry profited hugely on these desires.

    “Principles” were not at all a factor in the subprime crisis.

    I’d also like to see a more current version of the 15 yr old study on racism in car sales. MUCH has changed in those years. Not saying we’ve reached the promised land, but we’ve inched a lot closer since that time….

  17. Nemo,

    A big problem when you start to tear apart the data is the overwhelming presence of prepayment penalties and differentials in option arm amounts.

    If a minority has 10 FICO points worse than a non-minority, and has, say, 25 basis points more in a interest rate on their 2/28 arms, but also has a 6-month interest prepayment tab in the first two years, plus an additional 3% over the non-minority on the arm, is that racism?

    I’d argue that the real story is in those last two things, things we normally associate with non-credit risk on the borrower end (unless they have really good credit for prepayment penalties!)

  18. More precisely, as I’m sure you know Nemo but other readers might not, prepayment penalties *should* compensate for the risks on the negative convexity of reinvestments on the market side, not the risks of defaults. Yet why do all these minorities have prepayment penalties?

    80% of subprime has prepayment penalties, while only 2% of prime do. As far as I understand, this is where the minority problem comes in, but it isn’t clear how to quantify it, as there is debate on how it pays off risks and also contributes to defaults.

    James, my guess is that the market is probably competitive on the initial mortgage but a clusterf*** on the refinancing, where suddenly one bank has an informational edge on other banks, and where collusion to not mess with the fishies that have been caught in someone else’s net is more realistic. Though this is difficult to dig out of the data. (see page 17 here for a discussion.)

    Remember that 80% of these subprime loans refinance within 30 months – if that refinancing isn’t efficient, suddenly there can be major problems quickly.

  19. I am not an economist, but when Tanta’s imprimatur is on something, it gets cred with me.

    One observation not in the article for which I can vouch on a local observation level is that a fair number of the ‘point of contact’ salespeople who ‘made’ the sales were also minorities, just as greedy and unscrupulous as their whiter higher-ups, who are in no way exonerated by this fact.

  20. I don’t have a lot of background with this so parden my ignorance, but hasn’t the government (both parties) been specifically pushing home ownership for minorities and/or disadvantaged (poor), i.e. Freddie/Fannie, for the past 20+ years or so? Could this be considered predatory if they end up defaulting? How much money have these orgs lost over the last couple of years?

    Seems to me that some are playing both sides here. In good times, racism is claimed for lack of available credit and in bad times, racism is claimed because greedy lenders gave money out (when pressured by gov’t to do so) to individuals that shouldn’t have had it in the first place.

  21. Nemo: “The answers do not need to come from ideology. Certainly a lot of these loans were made in the past few years, and they are defaulting in significant numbers right now. So it should be possible to answer: For loans made at interest rate X%, what default rates are being observed for each racial group?”

    Please consider this, from Milton Recht: “Unemployment, loss of income, and uninsured medical expenses are also primary causes of mortgage defaults. Minorities have higher unemployment rates during this recession and are more likely not to have health insurance.”

    So minorities should default at higher rates. Now, for the sake of argument, suppose that high unemployment rates and lack of health insurance are caused by racism. Also, suppose that the lender harbors no personal racial bias, but is aware that minorities default at higher rates, all else being equal. It would be rational then, to charge higher interest to minority borrowers than to non-minority borrowers who are the same except for not being in a minority. The racist bias lies elsewhere.

    But racism is also a property of social systems. It offends our sense of fairness that a person should have to pay (by a higher interest rate on a loan) for being a victim of racism (by having a greater probability of being laid off). The lender who takes minority status into account is a part of the racism of the system.

    The sociological experiments demonstrate that business people and lenders do take minority status into account, whether consciously or unconsciously. It may be right, in economic terms, to do so, because doing so gives a better assessment of risk. But that does not mean that it is not part of systemic racism.

  22. All these smart people, and nobody knows this one?

    1) The process by which you obtain and/or maintain a banking license in this country is extremely racist, and selects for equivalent racism. Go back to the way minority-owned banks were treated during the S&L crisis, and tell me where, exactly, minorities were supposed to find a non-racist banker? Does such a thing exist in ANY country in this world?

    2) If you look at the pattern of property crimes against the poor by the rich going all the way back to the 1870s, you see that rich people in this country figured out something a very long time ago: if you steal from or cheat white people, judges and legislatures are likely to object. But if you steal from black people, the judges and the legislators are more likely to blame the black victims. Racist brokers pushed toxic loans into black communities because they knew that if they peddled those same loans to white people, courts and legislatures might stop and/or punish them. By pushing those loans into the black community, they evaded prosecution, and will almost certainly get to keep that money.

    I’d like to think that a black president who used to be a community organizer would do something about this. Yep. I surely would like to think that. Sadly, he’s far too busy being a “uniter.”

  23. More nonsense.

    How could they adjust for income when many of the sub-prime,alt-A,option arm, etc. loans were essentially liar loans where the stated income was FALSE.

    That in itself tells me doing any analysis on the data will lead to false conclusions.

    Perhaps minorities did pay higher rates, but we certainly don’t need ANOTHER government intervention to “correct” this market “inefficiency”.

    I do remember that Congress along with Freddie and Fannie consistently lowered lending standards (no down payment! no credit no problem!) for the sake of having more “homeowners”.

    Obviously there is NO free lunch. No one is a homeowner until they’ve paid off the bank.

Comments are closed.