Home > Uncategorized > Republicans would let car dealers continue racist practices undeterred

Republicans would let car dealers continue racist practices undeterred

November 11, 2015

There’s an upcoming House Bill, HR1737, that would make it easier for auto dealers to get away with being racist. It’s being supported by Republicans* and is being voted on in the next couple of weeks. We should fight against it.

The issue centers on the problematic practice of “dealer markups,” discretionary fees that brokers slap on after the credit risk of a given borrower has been established. It turns out that these fees vary in size and are consistently bigger for blacks and Hispanics. Which means that if a number of people of different races but a similar credit history walk into a car dealership and buy a car, the minorities will typically end up paying more. This is illegal discrimination under the legal tool called the theory of “disparate impact.”

The Consumer Financial Protection Bureau (CFPB) has been making a valiant effort to hold accountable the financiers behind these loans. For political (read: lobbying) reasons the CFPB doesn’t have regulating power over auto dealers directly, but they do regulate the bankers that supply the money, and they’ve been nailing those bankers for unfair practices. For example, there’s an ongoing case against Ally Financial for upwards of $80 million, which would go to compensating the victims.

Here’s the amazing thing: Ally Financial doesn’t claim their loans weren’t racist. They simply claim that they were less racist than the CFPB thinks, and that the methodology that the CFPB used to measure the racism is flawed. The Republicans agree, and they’re trying to remove the CFPB’s ability to enforce disparate impact violations altogether**.

So, just to recap, the Republican argument is: if you can’t specify exactly how racist this practice is, then you can’t stop people from doing it at all. It’s a dumb and dangerous argument. It is, in fact, exactly what disparate impact was meant to avoid.

A little background on why the measurement is so involved. In mortgage lending, the race of the borrower is recorded. In fact the race of every applicant is recorded, so that later on people can go back and see if there are racist practices going on. This is not so for auto lending, however. That means that when the CFPB suspects racism in auto lending, they have to impute the race of the applicant based on the information they know. Then, once they have partitioned the borrower population into “probably minority” and “probably white” subgroups, they measure the extent to which the “probably minority” gets overcharged. If it’s substantial, they charge the lender with discrimination and allot damages.

The opponents of the CFPB methodology claim that the race estimates are inaccurate; in particular, they charge that white people are sometimes being mistakenly labeled black or Hispanic. But consider this. That error, of overestimating minorities, would be alleviated by the second step, because the measurement of the extent of racism would be diluted if the “probably minority” group contained a bunch of white people, for the very reason that white people don’t suffer from racism. So it’s not even clear that the amount of damages being awarded is inflated; it’s just that the damages aren’t provably being sent to exactly the right people.

In statistical terms, we are worried about false positives – white people who might receive a compensation check for racist practices – and the proposed solution is to ignore all the actual victims of racist practices – so all the true positives. It’s throwing the baby out with the bathwater.

Even so, the CFPB’s race model is an imperfect methodology, as all models which depend on proxies are. And hey, you can try it yourself here.

Let’s take a step back. What’s the long-term goal here? It’s most definitely not to stand by, watching car dealerships screw minorities, and then after the fact to step in and assess damages. The actual goal is to put an end to the racist policies themselves.

What would that look like? Well, BB&T, another big financing bank for car loans, is changing from the dealer markup system to a system of flat fees which solves this particular problem (there are others!). From the article:

The bank didn’t attribute the move to pressure from regulators, but in a written statement the bank cited “fair and equal treatment of all consumers,” and said it was launching a “nondiscretionary dealer compensation program.”

To state the situation frankly: the anti-discrimination policies that the CFPB has been developing puts pressure on car dealers’ banks, and thus on car dealers, to deal fairly and transparently with their customers. And pushback against those policies is a vote to keep shady, discriminatory negotiations with car dealers that, generally speaking, screw minorities.

From my perspective, as a data scientist who studies both algorithmic unfairness and institutional racism, this is just the beginning of a much larger debate around what kind of processes we can audit with algorithms, what constitutes evidence of discrimination, and how quickly – or slowly – the laws are going to catch up with technology. It’s a litmus test, and it’s on the verge of failing badly.

* and some Democrats as well. See the full list of supporters here.

** Technically, the CFPB would still be able to enforce Fair Lending laws, but given that their methodology would be scrapped, it’s not clear how they’d actually go about doing that.

Categories: Uncategorized
  1. November 11, 2015 at 10:43 am

    I was never a fan of Richard Cordray who runs the CFPB either as he’s just another lawyer, They got it backwards in not having a tech run the agency, and then backed up with legal help so we have Cordray “learning” as they said when he was appointed. Today he’s still learning and he does go after low hanging fruit but the agency could be a lot more powerful but I do keep in mind it operates under the Office of the Comptroller of the Currency which doesn’t have the best reputation and skills as top bank regulators outside of collecting the “fines”.

    The agency spend a ton of money collecting consumer credit card transactions from Argus for one who does the research and “scores” us. CFPB is a big customer there as well as banks, credit card companies and anyone who wants to try and figure out some behavioral aspects of what we do as consumers. Have had conversations with World Privacy folks over this organization and their so called “scoring”, which of course ends up like everything else, scores you right out of access with a lot of flawed data out there.



  2. captain obvious
    November 11, 2015 at 4:32 pm

    re:”Here’s the amazing thing: Ally Financial doesn’t claim their loans weren’t racist. ”

    The government did not claim that Ally’s loans were “racist”. The reports of what Ally said come from legal pleadings, in which a defendant is generally not allowed to respond to claims that were not made. They had no reason to disclaim racism per se because the loan applications didn’t ask for borrowers’ race.

    The government basic claim isn’t that Ally itself did anything unfair, but that Ally’s internal statistical procedures for monitoring the compliance of its loans with disparate impact law were insufficient, so that Ally should be held responsible for an estimate of unfair behavior by car dealers. Thus, there isn’t much for Ally to talk in their pleadings except the validity of the statistical methods.

    The government’s core argument is (in items 36 of the complaint,
    http://www.justice.gov/sites/default/files/crt/legacy/2014/01/13/allycomp.pdf ) :

    “the interest rate markups [set by car dealers] charged by Ally to consumers are separate from, and not controlled by, the adjustments for creditworthiness and other objective criteria
    related to borrower risk already reflected in the [algorithmically determined] buy rate”

    However, in the government’s own models, 87 percent of the “disparate impact” goes away after controlling for creditworthiness of the borrowers. Those regression models include (imputed) race, and thus compare the relative importance of race and creditworthiness.


    If 87 percent of the effect goes away when adding credit score to the models, and race is also a variable in the regression, this suggests that what the car dealers are doing is mostly the opposite of what the government says they were doing. It is basically an “adjustment for creditworthiness and other objective criteria … already reflected in the buy rate”. A profit-maximizing dealer would want to guess which borrowers are more willing to afford a high rate, and if their judgements load on the Credit coefficient more than the Race coefficient (ie., low-credit borrowers have fewer options for cheap loans) this is being done much more fairly than the government implies.


    • November 11, 2015 at 4:42 pm

      Here’s the thibg, though. The credit risk of these borrowers was already addressed by their APR. This is added markup that is above and beyond credit risk.

      In other words, although it might be true that dealers find it easy to charge people more just because they can, that doesn’t make it fair. And it also doesn’t come under the reasonable business practice exception for disparate impact.


      • captain obvious
        November 11, 2015 at 4:58 pm

        I think the issue (the one I raised, at least) is that the government’s arguments are refuted by their own models, and
        they had these models in hand when going to court, so it is either incompetence or an abuse of process.

        If CFPB had set the penalty an order of magnitude smaller based on the model that uses creditworthiness, that would at least be a logically consistent application of disparate-impact theory. They did detected a possible race effect, such as dealers doing the race-neutral profit maximization more often on black customers than whites. But this is a second-order correction to the race-neutral thing that the government discovered but demands be ignored.


        • November 11, 2015 at 5:07 pm

          I disagree that the government is being inconsistent. The argument I made above explains that the CFPB does not need to “account for creditworthiness” when computing the disparate impact, exactly because the markups are placed on the loan deal after the APR has already been algorithmically decided.

          Your point would be valid if the CFPB were complaining about charging blacks more overall; given that black people tend to have worse credit scores, it makes “business sense” that they are charged more overall. The point here is that, after taking into account credit scores, on TOP of that credit risk, they are still charged more.

          You are suggesting that we should let markups also be determined, more or less, by credit risk, but that’s not a business need. So disparate impact applies.


  3. JRH
    November 16, 2015 at 6:24 pm

    Curious that the majority of Financial Services Democrats voted yes on this, even despite the relatively conservative make-up of the Dems on that committee. Was particularly surprised to see Terri Sewell, who represents Alabama’s Black Belt, vote yes. Less surprised after seeing this: http://www.opensecrets.org/politicians/contrib.php?cycle=Career&cid=N00030622&type=I ($32,500 from the National Auto Dealers Association, which doesn’t include the $5,000 they gave her “leadership PAC”). Pretty good return on that investment, it looks like!


  4. captain obvious
    November 24, 2015 at 7:18 pm

    ((“I disagree that the government is being inconsistent. … CFPB does not need to “account for creditworthiness” when computing the disparate impact, exactly because the markups are placed on the loan deal after the APR has already been algorithmically decided.”))

    That was the gov’t argument, and it is inconsistent in several ways.

    1. The lender and car dealers are being treated as inseparable for purposes of legal responsibility. In that case, the markup is irrelevant, the analysis should be of the whole loan. The government certainly did that analysis, since it is easier statistically, is a more straightforward argument to make legally, and is closer to the philosophy of disparate impact.

    The markup is only an accounting fiction from the customer’s point of view. The loan as a whole is what the consumer actually pays; it is more directly tied to the lender, since it includes the lender APR in addition to the markup; and the amounts of money (hence the possible level of disparate impact) are larger by an order of magnitude. The analysis of the markup, if the disparate impact were concentrated within that, would be nothing more than an additional piece of evidence in the case against the loans as a whole. The 87 percent figure I quoted from the CFPB study indicates that it’s not particularly concentrated there. So there is a strong possibility that the government analyzed the actual loans, and didn’t have a statistical case.

    2. The government “solved” this likely problem, the lack of a real case, by inventing a distinction between APR and markup. This also works well politically by scapegoating the dealers, who are unpopular, while holding responsible the lenders, who have the money.

    3. The distinction between markup and APR is a shell game. Economically and in business terms, there is no fundamental difference, because the consumer does not interact with the APR or the markup on their own, or know what they are. The lender is using the dealers to fine-tune its rate algorithm based on human observation and giving them a piece of the profits as payment for that.

    4. The markups do control the APR (opposite to what gov’t argue) because the formula determining the APRs is calibrated based on the level of markups the lenders can expect. The lender is trying to optimize their entire loan portfolio, based on APR + markup + volume of loans + market share + (everything else). The markups control their ability to sell the loans to dealers, and the APR influences their ability to sell to consumers through the dealers. Changing one number means other numbers change to compensate when running the lender’s optimization process. If the lender thinks that 4.4 percent is the right average interest rate on its loans, and dealers mark up by 10 percent, then its APRs will be set around 4 percent, which is lower than without the markup.

    5. As the last example suggests, if every dealer in the country employed the race-independent process of taking the APR and adding 10 percent to it (multiplying by 1.1),
    that isn’t discriminatory, but it would probably have some amount of disparate impact given blacks’ lower average creditworthiness indicators. Which comes to your remark:

    ((“Your point would be valid if the CFPB were complaining about charging blacks more overall; given that black people tend to have worse credit scores, it makes “business sense” that they are charged more overall. The point here is that, after taking into account credit scores, on TOP of that credit risk, they are still charged more.”))

    As I just explained, there is a model in which they are not charged more (same algorithm for everyone), and what they are charged is determined by credit scores, but the shell game of separating the markup allows the government to claim that the charges are not explained by the credit score. But of course they are.

    ((“You are suggesting that we should let markups also be determined, more or less, by credit risk, but that’s not a business need. So disparate impact applies.”))

    If lenders optimizing their profits is not a business need, there is no such thing as business need. It’s what they are in business to do. A “business need” requirement beyond that amounts to saying that any non-algorithmic component to price setting such as haggling, bargaining, and markups and such has to be specially justified. But if profit is not a sufficient justification, the government is just saying that it wants all such practices eliminated. I do think the economy and the regulatory environment have been moving in that direction, but I also don’t see how shifting the burden to the defendant can be legitimately done by that sort of argument.


  1. November 17, 2015 at 1:52 pm
  2. November 17, 2015 at 3:15 pm
Comments are closed.
%d bloggers like this: