It’s well known that some people are paid less than others for the same job, so it should come as no surprise that gay men have traditionally been paid significantly less than straight men with similar jobs.

It’s also unsurprising that the shortfall gay men saw in their pay narrowed as acceptance grew: 92 percent of LGBTQ adults believe that society is more accepting of them than it was even 10 years ago, and it’s natural that this would drive up wages some.

But what is surprising is that gay men seem not only to have narrowed the wage gap but may have surpassed straight men entirely.

According to a new study out of Vanderbilt University, gay men now earn as much as 10 percent more than straight men in the same type of job, on average, even when taking into account factors like education and experience.

Gay women already had a pay advantage over straight women in the past, so as gay men’s pay has gone up, it has brought about the same kind of pay premium lesbians already hadbut lesbians haven’t seen further growth. Whatever the cause, growing LGBTQ acceptance isn’t the full answer.

So what’s going on?

“My initial reaction was: ‘Well, this is obviously wrong,’” says Dr. Christopher Carpenter, the professor of economics who authored the study with his Ph.D. student Samuel Eppink. “[But] once we were convinced there was no data or statistical error, my reaction has been, ‘This is really interestinghow can this be?’”

And no wonderwhat the data shows is a bit baffling. Some obvious explanations are quickly ruled outpeople who identify as queer are more likely to pursue higher education, for instanceand while this is an interesting correlation (are queer people especially inclined to pursue more education, or are people with more education more likely to feel comfortable identifying as queer?), it can’t fully explain the phenomenon Carpenter found.

“If you didn’t account for the fact that sexual minorities have more education,” Carpenter says, “then you’d incorrectly attribute the earnings association uniquely to being a sexual minority.”

But his study doesn’t compare all gay men to all straight men; it compares gay men to straight men with similar levels of education. Same goes for years of experience, skill sets, and the responsibilities of the job.

The associations that emerge from this type of number-crunching are relativethat is, the bump in pay they found won’t look the same for everyone.

“There are important interaction effects going on that we couldn’t explore in this paper,” Carpenter explains, “so we can’t say that this is true for all subgroups within the sexual minority population.”

There’s simply not a large enough sample size to say for sure how the correlation is affected by race and ethnicity, for example. Even so, the data they do have suggests that the effect is most pronounced among (you guessed it) white men. Once you start making comparisons across age, across race, and so on, you’re no longer comparing apples to applesthe differences these other factors introduce will drastically change the differences you can expect to see in pay.

The question remains: Why is this happening? “We don’t have a great way to explain it,” Carpenter says. They can only rule out theories that the data doesn’t supportlike how the fact that men and women are affected differently rules out the possibility that this is simply an LGBTQ thing.

We can also rule out the possibility that discrimination is over. “I want to be very clear: I don’t think these results mean that gay people are not discriminated against,” Carpenter says. “There’s a lot of evidence to show that there’s quite a bit of discrimination.”

Gay and lesbian households are more likely to live in poverty, for instance, and more likely to grapple with unemploymenteffects that hit non-white families especially hard. That’s not to mention that a majority of states still offer LGBTQ workers no legal protections.

So, as nice as it sounds, nobody is getting paid more simply for being gay. Instead, there must be something about being gay that’s leading to a concentration of gay workers in some higher-earning positions, even if it’s not indicative of the community as a whole. Remember, this finding only applies to gay-identified men in the workforce; the survey can’t tell us anything about men who sleep with men but don’t identify as gay, or who don’t claim an income.

“It’s like throwing spaghetti at a wall,” says Carpenter. “What are the plausible explanations that don’t seem first-order crazy, and that might be consistent with the facts?”

Still, he has some theoriesnot so much answers as possible areas for future study, suggested by gaps in what we already know. For one thing, the initial survey didn’t correlate its data to location. “There’s lots of research showing that people in cities earn more than people who aren’t,” Carpenter says, “and there’s lots of research to show that sexual minorities are more likely to congregate to urban areas, relative to heterosexual individuals.” So by choosing city life in higher numbers, gay men may be pushing up their average pay.

There’s also some evidence to suggest that gay men relocate to cities at higher rates than gay women, but because it can’t be correlated directly to the dataset used in this study, this is still in the realm of speculation.

Another sweeping change that seems to have a lot of explanatory power is the shifting composition of gay families. As gay male couples start to form and formally recognize their relationships in greater numbers (getting married, hopping on each other’s health insurance plans), it increasingly gives one partner the option of dropping out of the workforceand it makes sense that it would be the lower-earning one who does.

Over time, lower-earning gay men selectively dropping out of the workforce would drive up the average pay for the gay men who remain. Interestingly, lesbians have traditionally formed and formalized relationships at much higher rates than gay men, so this theory also has the ability to explain why the premium is now showing up among gay men, even while lesbians are holding steady.

We also can’t rule out the possibility of some other indeterminate factor related to gay psychology or socialization, but while we can speculate about intangibles like the existence of a gay temperament or creativity gene, it’s not possible to demonstrate their effect on payat least, not using this type of analysis.

Carpenter’s study raises as many questions as it answers.

“The response that we’ve been getting from academics is a bit head-scratchy, a bit can-this-be-trueand as somebody who wrote many papers finding the opposite, I agree with them,” Carpenter says.

“We absolutely need to see if this result replicates.”