Can There Be Nondiscriminatory “Discrimination”?

Last week, the Supreme Court heard oral arguments in Texas Department of Housing and Community Affairs v. The Inclusive Communities Project, a case in which the Center for Equal Opportunity has been heavily involved.

The question presented in this case is: “Are disparate-impact claims cognizable under the Fair Housing Act?”

Under a disparate-impact claim, discriminatory motive is irrelevant: It need not be alleged nor proved, and it doesn’t even matter if the defendant proves that there was no discriminatory motive. If a policy or procedure results in a disproportion of some sort — not only on the basis of race, color, or national origin, but also religion, sex, or familial status (that is, having children) – then that’s enough, even if the policy is nondiscriminatory by its terms, in its intent, and in its application. The defendant can prevail only by showing – to the satisfaction of a judge or jury who may know or care nothing of the defendant’s needs – some degree of “necessity” for the policy.

The Supreme Court should reject this approach to civil-rights law enforcement.

1. Text. This approach is flatly inconsistent with the Act’s text. The text uses not only the phrase “because of” race but also “on account of” and “based on.” All of these phrases are naturally read to require a showing of discriminatory intent. The phrase “on account of” also appears in a section of the Act that bans coercion and intimidation of those exercising fair-housing rights, and intent is clearly implied there; and the “because of” and “on account of” language also is used to delineate certain fair-housing violations as crimes, and criminal prosecutions cannot be based on a disparate-impact theory. The disparate-impact approach would also render superfluous many of those provisions in the statute regarding the disabled. For instance, the failure to make or allow “reasonable modifications” and “reasonable accommodations” could have been attacked under a disparate-impact theory without those provisions.

* * *

The Obama administration’s brief stresses three brief, scattered in the 1988 amendments to the Fair Housing Act:

Nothing in this title limits the applicability of any reasonable local, State, or Federal restrictions regarding the maximum number of occupants permitted to occupy a dwelling.
Nothing in this title prohibits conduct against a person because such person has been convicted by any court of competent jurisdiction of the illegal manufacture or distribution of a controlled substance as defined in section 102 of the Controlled Substances Act (21 U.S.C. 802).
Nothing in this title prohibits a person engaged in the business of furnishing appraisals of real property to take into consideration factors other than race, color, religion, national origin, sex, handicap, or familial status.

The Obama claim is that there would be no reason for these provisos unless the Act generally allowed disparate-impact causes of action. Of course, it is not particularly surprising or damning that there should be some redundancy in a long statute like this. Legislation is not poetry, and lobbyists are paid to put in suspenders even if a belt might do.

But it is also worth noting that all three provisos involve instances where non-protected characteristics are close enough to protected characteristics that Congress wanted to spell out what was and wasn’t protected a bit more — particularly since Congress was adding “familial status” and “handicap” to the statute at that very time.

To elaborate: The 1988 amendments expanded the Act to cover not just race, etc. but also “familial status” — that is, to make it illegal to discriminate against families with children. But an occupancy limit on, say, an apartment obviously appears on its face to get close to that line.

Another sort of discrimination prohibited by the 1988 amendments was against the handicapped. Those with a record of drug addiction are considered to fall into this category (so long as it is not “current, illegal drug use . . . or addiction”). So here Congress would have been aware that discrimination against someone because of their record of drug crimes gets close to the line of handicap discrimination. The war on drugs was well under way, and no doubt our legislators wanted to make sure no one could claim their bill was soft on drug traffickers.

As for real-estate appraisals, bear in mind again that disability and familial status had just been added to the statute. Designing (or redesigning) property so that it is accessible to the handicapped can affect its value; changing a building so that it is kid-friendly can, too. What’s more, in both instances the result might be a building that is more expensive to construct but less desirable to the large class of individuals who aren’t being accommodated (the able-bodied, and singles or couples). Either way, Congress did not want real-estate appraisers to be held liable for market pricing that was not their doing.

So these provisions were added to clarify in three tricky situations what was and wasn’t disparate treatment — not to ban (only) certain kinds of disparate-impact lawsuits.

2. History. If there is no textual support for a disparate-impact cause of action in the original Act or its 1988 amendments, and since the legislative history points in the other direction as well, the remaining argument to support disparate impact in fair housing law is that some lower courts had recognized a disparate-impact cause of action under the original 1968 version of the Act, and so Congress implicitly endorsed the approach when it reenacted the statute in 1988 with full knowledge of those decisions.

But Congress also knew that the Supreme Court had not resolved this question. During the summer of 1988, while the amendments were before Congress, the Justice Department was arguing to the Supreme Court that it ought to grant review in a New York case and rule against a disparate-impact approach. In other words, Congress could hardly be said to have been endorsing settled case law by passing the 1988 legislation, because no settled case law existed.

3. Deference.There are some newly minted Obama administration regulations that endorse the disparate-impact approach, and it is argued that the Court should defer to that interpretation of the statute. But there are a number of good reasons why these regulations are entitled to little deference, even beyond the fact that they are inconsistent with the plain words of the statute.

The Fair Housing Act has been on the books since 1968, and during that time the executive branch has sometimes endorsed the disparate-impact approach and sometimes not. For example, President Reagan explicitly rejected the approach in signing major amendments to the Act, and his Justice Department argued against it in a brief to the Supreme Court; the Bushes didn’t think much of it, either.

The Obama administration, on the other hand, is attempting to game the system here; it orchestrated a shady deal with the City of St. Paul to get it to withdraw an earlier case, and meanwhile worked on promulgating those new regulations. “We were afraid we might lose disparate impact in the Supreme Court because there wasn’t a regulation,” said Sara Pratt, an official in the Department of Housing and Urban Development.

In any event, the principle of deference ought to be trumped in this case by the “constitutional-doubt canon.” The Justices have repeatedly acknowledged that a statute mandating the disparate-impact approach also can encourage race-conscious decision-making; this of course raises serious constitutional issues. The approach raises further constitutional problems here by altering the state-federal balance in far-reaching ways. For example, it renders race-neutral rules – like rules for preserving order in public-housing projects – suspect; the approach will also result in the federal micromanagement of insurance practices, which is at odds with the McCarran-Ferguson Act, a point emphasized in a recent district court decision striking down the Obama regulations.

4. Coherence.One would also expect that, if a statute contemplates use of the disparate-impact approach, it would answer some fundamental questions like how to measure the kind and degree of disparate impact that is required and what sort of rebuttal is needed. But there’s none of that, and the resulting problems are myriad and severe.

For example, what should decision-makers do if a practice has a disparate impact in one location but not in another? Or if the impact ebbs and flows over time? What should landlords do if a policy (for instance, excluding felons as tenants) has an unfavorable disparate impact on potential tenants of a particular race, but is welcomed by the incumbent tenants who are predominately of that same race?

And what if a practice is favorable for some racial minority groups (say, Asian Americans) but not for others (say, Latinos) – and, what’s more, the opposite is at the same time true for some minority subgroups (thus, unfavorable for Hmong but favorable for Cuban Americans)? And remember, too, that “majority” groups – whites and men and Christians, for example – must be able to bring these lawsuits, too, or you’ve added an even greater equal protection problem.

Thus: (a) a foreclosure policy may have no disparate impact on a particular group in pre-recession 2006, but a severe one in 2009; (b) an income-requirement may have no disparate impact on Latinos in Nashville but a severe one in Denver; this may mean that two companies with identical policies have very different liability risks, or the same company may be liable in one city but not in the other (but should the cities be considered separately if it’s the same company?); and (c) the use of credit scoring may have a disparate impact on Latinos but not Asians, but there may be no disparate impact on Cubans and a severe one on the Hmong.

There’s an even more fundamental problem: It is often hard to say whether the impact a practice has on a group is adverse or not. In the Texas case, for example, the complaint is that low-income housing is being disproportionately located in black areas. But poor black people might prefer to have housing opportunities near where they already live rather than a long way away, and they could complain about the disparate impact of deliberately changing the system so that they had fewer such opportunities. Yes, it might interfere with social engineering that would force blacks to relocate to white areas, but might not even the aim of greater integration be met, at least to some degree (depending on racial breakdown of low-income housing recipients – a breakdown that might vary from city to city and county to county) by encouraging non-blacks (not just whites, but also Latinos and Asians) to live in black areas?

These problems make it difficult to decide not only whether there is a disparate-impact in the first place, but also how to weigh properly the defendant’s rebuttal, which in the public housing context – versus, say, employment – will often involve balancing myriad and hard-to-quantify interests. In sum, it is astonishing to interpret a national civil-rights statute in a way that makes identical conduct in one city illegal while allowing exactly the same conduct in another city, just because of the different racial makeup of the two cities.

* * *

Let me just add three brief points. First, it’s frequently asserted that we must allow “disparate impact” causes of action because actual discrimination (“disparate treatment”) is difficult to prove. But this is simply not true: The overwhelming majority of housing cases brought and won by the federal government are disparate-treatment cases, as anyone who reads the Department’s press releases every day (as the Center for Equal Opportunity does) can attest.

Second, many on the other side are arguing that you need the disparate-impact approach in order to go after segregated housing patterns. But of course that is not true if the segregation stems from actual discrimination. And if it is not, then using the disparate-impact approach raises all the usual problems inherent in the disparate-impact approach. For example, how much racial balancing is to be required? What if the reasons for the racial imbalance reflect voluntary decisions or economic realities? What sort of remedies will be required (like deliberate assignments on the basis of race), and what if those remedies end up hurting people (including minorities) on the basis of race? And so on. Besides, if racial imbalances in housing patterns are a result of voluntary choices by individuals, then it’s unclear why this is a huge evil that the government needs to fix.

Finally, I’d note that the National Association of Mutual Insurance Companies submitted excellent testimony on this issue at November 2013 Congressional hearings, which can be read here (starting at page 110 in the hard-copy pagination).

Here’s hoping the Court does the right thing and puts an end to this nonsense.

Related posts: