Don’t Assume That Gay Couples Face Mortgage Discrimination

A study by Iowa State researchers that has gotten some media attention at places like NBC News finds that mortgage lending to same-sex applicant pairs is associated with higher rates of loan rejection and slightly higher interest rates. The study is already being cited in support of the Equality Act, a bill that, among many other provisions expanding the scope of federal law, would extend the federal Fair Housing Act to cover sexual orientation. 
There are reasons, however, to approach the findings with caution. To begin with, lenders currently have an economic incentive to underwrite loans correctly and compete for all profitable business. Beyond that, studies that find positive (even if thin) evidence of discrimination tend to get reported and amplified heavily, while those with null results get ignored.
The first point to get on the table here is that same-sex couples are decidedly *not* distributed randomly across all sorts of neighborhoods. In particular, in common observation, male couples have long tended to be overrepresented in neighborhoods that are undergoing various stages of renovation, often associated with upward movement of real estate values. Some of these neighborhoods are run-down or even crime-ridden when the same-sex couples start moving in. Some have finished the process of “arriving” and have become pricey and desirable (although the gay couples might choose to move on at or before that point). 
There are at least two possible mechanisms by which these neighborhoods could have higher mortgage denial rates for reasons unrelated to any animus on the part of lenders: they might include more marginal/troubled neighborhoods in the early stages of comeback (which will have higher default rates for multiple reasons). Separately, renovators have different needs in the mortgage market than those who buy new (or buy fully renovated).  Loans in comeback neighborhoods might be exposed to the complications of renovation by being, for example, hybrid construction loans, multi-family, residential-plus-commercial, or lacking in conventional amenities such as off-street parking. 
One transaction I know about from firsthand experience, for example, was nonconforming in several of these ways: it was a construction loan premised on the idea that a large chunk of cash was going to fix up the old house, it was a two-unit, and one of the units was commercial. Such a loan is often hard to place in the conventional mortgage market, which is geared toward standard servicing and packaging for resale into predictable secondary markets, as favored by mass lenders and mortgage servicers. The alternative, as in the case I know about firsthand, can mean turning to an individualized deal on slightly less favorable (but still fair and competitive) terms.  
Sure enough, if you turn to the write-up at Iowa State, you find that it reaches quite a significant conclusion that goes unmentioned in the more popular NBC report:  
“What they found was somewhat surprising. In neighborhoods with more same-sex couples, both same-sex and different-sex borrowers seem to experience more unfavorable lending outcomes overall. The researchers say the findings should raise enough concern to warrant further investigation.”
In other words, the study itself leaves gay couples’ higher turn-down rate looking like something of an artifact: the observable result was linked instead to the neighborhoods where they take out mortgages. Incidentally, the researchers were able to control for some variables such as types of mortgage, which did allow them to take into account some of the differences that could be observed and quantified. But just as surely, especially working with data sets gathered for other purposes, they could not control for all the ways one neighborhood (or for that matter one individual loan) differs from another. 
These are not findings on which one would want to premise any new legislation restricting liberty of contract.