Monday, November 30, 2015

The Consumer Financial Protection Board And Anyone Involved In Creating It Should Be Ashamed. The Rest Of Us Should Be Advised.

Here's a hypothetical for you to consider using the variables below:

- I live in a suburban neighborhood
- I just bought a new car.
- I got the lowest possible APR due to my excellent credit score
- My last name is Mann

- My neighbor also just bought a new car
- My neighbor's last name is Morris
- My neighbor's APR was 30 basis points higher than mine because his credit score is lower

Now the Consumer Financial Protection Board, that great defender of hapless consumers of all races, colors and creeds, gets wind that there may be some shady goings on in the finance departments of auto dealers.  They suspect that these dealers, who are so blinded by their visceral hate of anyone who is not of the caucasian race, refuse to sell them cars unless they agree to pay more for them than their white counterparts.  It has to be true because, well, just because.  All that remains is to prove it and bring the criminals to justice.

So the CFPB has to decide how to go about proving it.  They demand the names and addresses of scores of customers from the nation's auto dealers and their financing arms. There's a problem though. Auto loan applications don't call out race.   The CFPB is at an impasse, but they're determined to get to the bottom of this so they have two choices:

1)  They can begin the tedious process of calling loan applicants to determine if they are in fact minority or white, as a preliminary step in deciding whether or not an ultimate determination of discrimination on the basis of race is warranted; or

2)  They can guess about the races of those applicants and use their guesses to bring charges and levy fines against the auto dealer's finance companies.

If they pick number one, the process will be very slow, but the information obtained will be useful and accurate and, accordingly, the judgments would be seen as fair, as long as they were consistent with the data.

If they pick option two, well that's just silly.  And on top of that it would drive up the costs of auto purchases for all consumers as auto dealers would more than likely adjust their rates consistently upward rather than downward to comply with the government's illogical (based on common procedures for underwriting risk) ideas about fairness and consistency with respect to auto loans.  But since, as one writer put it, "the CFPB doesn't take any risk and it bears no accountability."  Who cares if prices go up as long as they go up for everybody, regardless of their risk profile?  That would seem to be the CFPB's point of view.  Fairness prevails.

So there's the hypothetical scenario.

Which option do you think the ultra responsible and highly accountable CFPB selected? If you said option 1, you would be making a perfectly rational selection.  But you would be flat wrong.

Instead of treating businesses with the same fairness it claims to be seeking on behalf of consumers, the CFPB uses an algorithm that takes the last name and home address of the auto consumer and uses it to predict whether or not the consumer is a minority or white. There is no subsequent follow up to determine the level of accuracy of the guess.

We're not talking hypotheticals anymore.  This is really happening.  See the actual tool being used below:

How the Government Predicts Race and Ethnicity

The Consumer Financial Protection Bureau and the Justice Department predicted borrowers’ race and ethnicity by using an algorithm that analyzed 

last names and addresses.  They are now awarding $80 million in a car-lending discrimination settlement based on those findings. 

Auto lenders and car dealers say the technique is prone to errors.

Enter a surname and address to see how accurate the prediction is for you.
Note: Hispanic may be of any race. Names and addresses submitted are not retained by the Journal. Full addresses are sent to the Census Bureau to determine geographic region. This may produce slightly different results than the geocoding service used by CFPB. Entering an address will provide more accurate results than a ZIP Code.
Source: Consumer Financial Protection BureauU.S. Census Bureau, Department of Justice

So, solely on the basis of a bunch of guesses using the tool above, the CFPB has extorted over $200 million to date from several large finance companies, including Ally Financial, who coughed up $80 million without admitting wrongdoing so that its government application to become a holding company, which is critical to its continued survival, wouldn't be denied by bureaucrats on the basis of trumped up fairness in lending violations.

Oh, by the way, the CFPB's guesstimator determined that there is an 81.68% chance that I am white and 12.52% chance that I am black.  Bubba would beg to differ.

The guesstimator further determined that my neighbor, who is all white (and alright), had only a 71.53% chance of actually being white and 23.92% chance of being black.  I'm not making this up, I swear.  

Sadly, if the hypotheticals above were in play and the race element were determined using the guesstimator, the CFPB might actually us that data to bring racism charges against some unsuspecting dealer.

Though the CFPB was created by the Democratic Party, purportedly to protect the little guy, the issue just discussed is not about politics at all.  It is, however, about the little guy.  And it's worth noting, that every business, no matter its size, is a little guy when compared to the U.S. government.  Ally Financial has enormous resources at its disposal, yet it was brought to heel by a government whose influence over its citizens has continued to grow.

Thomas Jefferson warned that, "A government big enough to give you everything you want, is a government big enough to take away everything that you have."  

We should all be aware of that and should act accordingly.


No comments:

Post a Comment