My position is supported by evidence. Your position is supported by flawed evidence – namely, evidence that does not control for crucial variables. You continue to hold to the position that a “significant number of lenders let prejudice get in the way of their decisions” without addressing the flawed methodology of the evidence upon which you rely.
You have made a claim. You have the burden of proof on that claim. You have presented evidence. That shifts the burden to me. I have presented contrary evidence that shows a serious flaw in your evidence. That shifts the burden back on to you. You have not met your burden on that point.**
My arguments and evidence have mostly been relegated to lending practices. It would be error to construe them beyond that area.
At any rate, even in the lending area, you have not made a case that racial discrimination “continues to happen.” You have failed to control for crucial variables. See above.
Dewey, I think you are generally having the best of this aspect of the discussion, but you do have an assumption in here that I have not seen made explicit.
If you go back to your original list of criteria, can you establish that every one of those points is genuinely necessary to establish good loan risks? For example, “that study shows that whites fare better on debt-to-income measures, while blacks do better on loan-to-value measures, with the two more or less canceling each other out.” They may appear to cancel each other out for the purposes of comparison, but are both of them true measures credit worthiness? Or is it possible that the aspects where blacks or whites are penalized more frequently are simply some “this should work” test that the banks imposed (not necessarily to impose color tests, but simply because it felt right to them some years ago), that are now being used without reconsideration despite having an unfair impact on loan practices among blacks?
As an outside analogy: We all (or, at least, all of us old people) remember the brouhaha about lowering standards to allow women to join fire departments. In some cases the standards really were lowered (there are a lot of fire companies out there), however, in at least one instance, the “lowered standard” was a chimaera. Much was made of the fact in one set of tests that women were not being held to the standard of carrying X amount of weight over their shoulders and lots of people leaped up to decry being stuck in a building where some woman firefighter was going to be unable to do the necessary work to save them.
On re-examination, however, it turned out that the women were still required to be able to move the minimum weight standard, they were simply not required to hoist it over their shoulders where a man’s center of balance provides him an advantage for that technique. (And, in fact, for a number of the carries, the training manual recommended that the loads be dragged or otherwise carried below the smoke, giving the women an advantage for real-life situations.) The test had simply been put in place, years earlier, for an all-male fire company when it provided a general test of strength, but it was not based on the actual requirements of a person fighting a fire.
Your list of credit qualifications certainly appears neutral, but I wonder how much of it may be irrelevant–and may be setting up barriers based on unconsidered legacy practices.
So, I ask once again (and will keep asking until you or someone actually answers it, instead of saying it doesn’t matter). Why do you think that blacks are less creditworthy to a bank than whites?
Occam’s razor tells me it’s discrimination, but I’ll gladly listen to other ideas.
Not to answer for Dewey but my understanding of Occam’s Razor is:
In the case of the study Dewey cited a large number of factors are being accounted for. Occam’s Razor tells me that the credit worthiness algorithms used by banks have a disparate impact on blacks based on neutral color-blind data.
tomndebb’s supposition about credit worthiness being based on outdated or inaccurate criteria may be one reason. We’ll all have to see if that hypothesis bears fruit. There may be other reasons. And while I’m sure that some individuals are racist and that some of those individuals work in banks and that there are certainly examples of blacks being discriminated against by said individuals I’m frankly skeptical of the claim that racism is endemic to banking institutions.
That would still require that there are large differences in one or more of the factors between blacks and whites as groups. That’s the question I keep asking. What are those differences, and for that matter why do they exist? Is there some bizarre scientific connection between melanin levels and financial acumen? Why won’t anyone answer this question?
No. But there could conceivably be different lifestyles (and, hence, financial priorities) among different cultural groups.
Two points to remember:
The comparisons are always between the extremely heterogeneous white populations and the somewhat more homogeneous black population. Would we see the same disparities if we broke down the numbers among different ethnic groups or rural vs urban vs suburban whites? I don’t know.
The black community was subjected to financial pressures to which the white communities were not for over 100 years. Those experiences may have shaped spending and saving patterns in ways that the banks have traditionally chosen to disparage. This could be a sign of previous institutional racism that exerts a current pressure that no one has recognized.
[Anecdote alert!] My grandfather was subjected to a very bitter lawsuit on the death of his parents regarding the inheritance. It tore up that family. At that point, he chose to never own property so that his children would never find themselves fighting over his estate. (Pretty dumb move for an otherwise quite smart guy, but emotional influences are not rational.)
Now, extend that concept into the sphere of black relations with banking institutions since 1865. Is it possible that the black community has chosen financial patterns that will keep their money out of the hands of greedy,racist bankers–and those practices are now coming back to haunt them?
I do not know that the criteria to get a loan are or are not discriminatory, per se. At this point, I would not accept either the charges of racism or the claim of financial neutrality at face value.
This is the same questions that I’m asking, additionally I don’t trust the study that Dewey posted because there’s no data that accommpanies it that backs up the claims that are made with respect to the financial status of the two groups. Additionally, the study makes claims that the statistics from the Acorn study clearly dispute, and the study admits that lenders are using a different rule of thumb when evaluating minority applicants.
How do we account for this data?
How do we account for this happening in other cities besides Chicago?
It could be any number of reasons, including the effects of past and ongoing discrimination. I’ll cede that just for the sake of argument (though that could be a debate in and of itself). But the salient point for the current discussion is thus: it is not the bank’s job to remedy the wrongs of the world.
Even if past discrimination means that blacks as a general rule have poorer credit histories, riskier collateral, higher debt-to-income ratios, and other troublesome factors on their credit applications, it is not discrimination on the part of the bank if it turns down a black applicant due to those race-neutral reasons. The bank is only blameworthy if it turns down an applicant because they are black.
It isn’t 1965, either. Even Selma isn’t Selma anymore.
The study I cite to clearly cites to its data sources.
This:
Indicates a misunderstanding of the study in question. First of all, the ACORN study does not “clearly dispute” the findings of the Fed study, because the ACORN study only controls for income. IOW, the Fed study is designed to correct for factors that studies like the ACORN study fail to take into account.
Second, the Fed study admits that different lenders will use different “rules of thumb” as to the weight they give various economic indicators – one lender may consider credit history more important than debt load, another lender vice-versa. The study then attempts to control for those differing weights in the formulas it uses to reach its conclusions. Far from being a weakness of the study I cite, its attempt to control for such variability between lenders is actually a strength.
Well, that’s clearly a question on which various lenders will disagree. Some lenders will give more weight to credit history, others to existing debt load, others to other factors. Indeed, the precise formula a bank uses is generally considered an important proprietary asset – everyone thinks their formula is best. And every bank’s goal is to tune their formula to maximize loans written while simultaneously minimizing defaults. At any rate, that’s a thing for the market to shake out – as long as the banks are using race-neutral criteria, then can’t in any meaningful way be engaged in racial discrimination.
Cheap shot, I know, but it would appear that you would oppose rules that provide de facto support for minorities while supporting rules that provide de facto disparagement.
(I would also point out that the examples you provided of de facto harm were not, in fact, apparently neutral rules that had a de facto result. The old literacy tests were clearly not neutral, beginning with the Grandfather Clauses and extending through the practice of giving blacks with elementary school educations college tracts to read, black M.D.s Chinese texts in the original Chinese characters, and allowing the test to be “graded” at the whim of the local voting official. The Hopwood case was also not an example of a neutral test with a de facto result as, according to your presentation, the Admissions Office was going outside the rules to manually sort minority applicants into categories for which they had not achieved a sufficient score.)
Not quite. What I wrote above about de facto rules was meant to apply to rules that are facially neutral, but are in fact designed to act as a quota system – a plan to bring in quotas through the back door, if you will. That isn’t the case with the lending practices we’re discussing – unless you’re willing to say that using criteria like credit history and collateral value are actually intentional attempts to engage in back-door discrimination.
I haved no desire to assert that they are deliberately discriminatory. However, given the proprietary (and, hence, hidden) nature of the weighting, I suspect it might be worthwhile to see whether there are unintended–and irrelevant ot the loan–consequences of the criteria in use.
If, like the “carry X pounds above the shoulders” rule, they exhibit de facto discrimination without actually protecting the banks’ investments, then I think they ought to be reconsidered. (And these would significantly differ from the Howard case in which neutral rules were not followed.)
Now we’re getting somewhere. Just for the sake of argument, wouldn’t attempting to remedy an obvious disparity, even if based on past wrongs, as opposed to just current ones, be a worthwhile thing?
If banks are solely looking at a precise set of unbiased criteria in determining loan acceptance, I would agree with you completely in this regard. Unfortunately, many financial institutions have “override” methods, wherein borderline cases (say credit scores falling between 600-650, for example) are decided on a case by case basis, at the discretion of certain people (in cases I’m familiar with, it was the manager of operations over the loan processing department). Considering the large number of applications that sometimes fall in this grey area, discrimination could conceivably have a major effect. If, and only if, a financial institution stuck strictly to its criteria, and if its criteria was not intentionally weighted to give an edge to white customers, then you can eliminate discrimination from the picture.
Do you have any ideas for eliminating the human factor from the process, while leaving the banks some autonomy in their criteria?
The short answer is, they are reviewed with regularity. Banks hire people with expertise in both financial and statistical analysis to develop their formulas on an ongoing basis. Irrelevant data will be stripped from the formula because it makes the loan process inefficient – it takes manpower and computing time to process those applications, and extraneous data adds to that time unnecessarily.
At any rate, if you have a problem with any of the criteria suggested thus far, or any of the criteria used in the Fed paper, I think it’s incumbent on you to demonstrate a prima facie argument for why that data is irrelevant to a loan decision. Otherwise, I think the presumption should lie that the banks are not interested in wasting time and that thus the criteria are valid.
No, it would not. The bank is not a charity, nor does it exist to remedy discrimination in other areas. The bank exists to make money by making profitable loans to debtors who will not default.
I mean, really, what are you suggesting? That banks loosen their lending criteria for blacks solely because black debtors may be higher credit risks due to the vesitges of Jim Crow and its ilk?**
Discretionary cases such as those you describe are, virtually by definition, going to represent a small proportion of the bank’s total loans written. I have a hard time believing they would sway the statistics in a material fashion.
Here’s some more fuel for the fire. While this study doesn’t directly address the study cited by Dewey it does shed a little more light on the situation and perhaps raises an interesting question.
One of the early studies made in response to the original Boston study hypothesized that if minorities were being held to higher credit worthiness standards (i.e. they were being discriminated against) then that higher standard should result in a lower loan default rate for minorities as a group. The thinking being that if they were required to be better credit risks that should translate into lower default rates. In fact the study found that minorities defaulted on loans at a greater rate. The study I cited above notes however that they did not account for credit history (incidentally, Dewey’s study does). When taken into consideration the poorer FICO credit rating of minorities on average could easily explain the higher loan default rate of minorities in general (credit rating having a good predictive value in regards to determine loan default rates).
The conclusion the study makes is less than earth shattering. Basically that the claims of the original loan-default model (which attempted to show there was no/very little discrimination) are not nearly as strong as they originally seemed. In addition the study states it would be “hazardous” to try and draw any other substantial conclusions based on the data available. In other words, it’s inconclusive. The jury is still out.
The study does bring one question to mind. Why do blacks and hispanics have lower FICO scores on average across the board in comparison to whites and asians? Presumably the algorithm used to calculate FICO scores has no variable for race. So why the disparity? I suspect cultural factors are the bugbear.
I have no specific argument against the banking procedures and I agree that if someone does, it is their responsibility to establish the actual errors, if any. I have already noted in this thread that it may be culturally based actions within the affected groups that may be triggering the statisitics we see.
(On the other hand, having been the person who actually gathered the data to make financial decisions for various companies in several industries, I find the notion that all such decisions are based solely on rigorous applications of facts and logic to be a bit surreal. I have yet to encounter an industry that has no sacred cow assumptions that are off limits to actual inspection.)