[Epistemic Status: I think my argument in the first half of the post is likely true, and my argument in the second half of the post is highly speculative.]
[Disclaimer: I am using the terms ‘male-sexed’ and ‘female-sexed’ in this post due to the stubborn refusal of the trans community to give me better words. Biomedically transitioning trans people are not female-sexed or male-sexed. Neither are intersex people.]

The Blanchard/Bailey theory is essentially that trans women are divisible into two groups: one group tends to transition early, be solely attracted to men, pass better, be unusually feminine as children, and work in typically feminine professions like hairdresser; one group tends to transition late, not be solely attracted to men, pass less well, be masculine as children, and work in typically masculine professions like programmer. The Blanchard/Bailey theory also comes with explanations for these two groups’ motivation in transitioning. Assuming that Blanchard and Bailey’s research is correct even if their theories are not– that these two clusters exist in the surveys they conducted– what could be the explanation for this?

I feel like the “autogynephilia” and “homosexual transsexual” explanations are very unsatisfactory. Autogynephilia does not work like other sexual fetishes. It is relatively rare for a person to upend their entire life to satisfy a sexual fetish; pretty much the only other example I can think of is 24/7 BDSM, which itself often has a strong nonsexual component. Antiandrogens are a commonly used treatment for sexual fetishism, because they lower libido and thus motivation to engage in fetishistic activity; antiandrogens are also a component of HRT, but we don’t see trans women becoming less motivated to transition after they start HRT.

Autogynephilia’s oddness is often explained by saying that it is a peculiar form of fetish known as an erotic target location error, in which a person is aroused by the idea of becoming the thing they’re turned on by. Other examples of alleged erotic target location errors include arousal at the idea of becoming an amputee, autozoophilia, autoplushophilia, and ageplay. (Ageplay is a bit of a weird example, because as far as I can tell there is no evidence that ageplayers are any more likely than anyone else to want to have sex with children, so maybe it’s actually just three examples.)

First of all, the four specific things that are the subject of erotic target location errors are very strange. Why aren’t redhead fetishists aroused by having red hair or foot fetishists aroused by having particularly desirable feet? Erotic infantilism is common, but given men’s preference for twenty-two-year-old women why aren’t there a bunch of men deeply erotically interested in being twenty-two?

Second of all, it is strange that autogynephilia is the only erotic target location error that causes a significant number of people to wish to transition. There are maybe some people with bodily identity integrity disorder (although far fewer than gender dysphorics) and maybe some otherkin. Why aren’t there a bunch of ageplayers insisting that their soul is really ten years old? Why aren’t there a bunch of people who insist that they really are plushies? Why aren’t foot fetishists demanding attractive-foot surgery?

Furthermore, the autogynephilia theory does not even explain the data it purports to explain. Why are trans women disproportionately engineers and soldiers, instead of being randomly sampled from the male population? Why would a fetish make one transition later? Why would it cause one to not pass as well? Surely fetishizing being an attractive woman would cause one to have a lot of motivation to be an attractive woman.

For some reason, everyone gets outraged at the autogynephile theory, but the homosexual transsexual theory seems to me to be equally offensive: the idea that very feminine gay men transition because gay men are attracted to masculine men and straight men are attracted to feminine women, so by becoming a woman they can get a more desirable sexual partner. While this might make sense as a motivation in relatively trans-positive places, I do not think there are a large number of people who become homeless teenagers doing survival sex work so that they can get laid more easily. In addition, it is unclear to me whether transitioning actually improves trans women’s level of sexual and romantic success. Many straight men who are interested in trans women have homophobic and transphobic beliefs, which make them unlikely to commit to a trans female partner, more likely to freak out after sex, and more prone to committing harassment or violence as a way of resolving their cognitive dissonance.

So if it’s not Blanchard and Bailey’s theory, what is it?

It seems to me that relative level of passability is obviously linked to age at transition. Hormones have effects; a person who has had a testosterone-dominant hormone system for forty years will, all things equal, look more like a man in appearance than a person who has had an estrogen-dominant hormone system for much of that time. So those can be folded into a single factor.

Blanchard’s original studies were conducted in the late eighties and early nineties, long before the current wave of trans awareness. In this context, I think it makes perfect sense that trans women who conform very poorly to their assigned gender and who are solely attracted to men would be more likely to transition young. One of the most important aspects in realizing you’re trans, for many people, is meeting other trans people. It lets you realize that transition is possible, that trans people aren’t the pathetic jokes they’re depicted as in mainstream media but people like you. On a more practical level, knowing trans people helps you DIY hormones or get past gatekeepers.

Trans women who are attracted to men would be far more likely to be connected with the gay community, where they could meet other trans women and wind up transitioning fairly young. The same thing’s true of trans women who conform poorly to manhood, because they might think something along the lines of “oh, I’m a flamer, I guess I’m gay.” The ‘feminine’ professions Blanchard and Bailey discuss are mostly, in fact, stereotypically gay male professions (they include “hairdresser” and “sex worker”, but not “nurse” and “secretary”); this could be explained either by male hairdressers being more likely to meet gay men, or by people in the gay male community being more likely to become hairdressers.

Conversely, a trans women who’s attracted to women and who conforms fairly well to her assigned gender would have far more difficulty meeting other trans people, since she had no community that was disproportionately full of other trans people. (Unless, of course, she found a local crossdressers’ group, but even that is quite difficult, given the shame most crossdressers have about their crossdressing.)

With the rise of public awareness about transition and (especially) the Internet, more people are aware of the existence of trans people. So we should expect trans lesbian programmers to be transitioning younger than they did in the past. Anecdotally, this appears to be the case; trans lesbian programmers of my acquaintance regularly transition as young as nineteen or twenty. I await replications of the Blanchard/Bailey study.

So ignoring social factors, we have one group of feminine straight women, and one group of masculine lesbians, bisexuals, and asexuals. (Of course, these are clusters, and not a binary distinction; there are quite a lot of feminine lesbians and masculine straight women.) Now, there are a couple ways one can respond to this. First, one can point out that this is identical to the pattern among cis women: in general, LGBA women are more likely to be gender-non-conforming than straight women, regardless of birth assignment. There might not be all that much to be explained.

However, my inner Blanchard/Bailey theorist is pointing out that there’s still something to be explained. First of all, trans women are much, much more likely to be lesbians than cis women are. Second of all, trans lesbians’ masculinity is different from cis lesbians’ masculinity: cis lesbians are not bizarrely likely to be programmers or mathematicians, and trans lesbians are not that much more likely than average to have buzzcuts.

(Now we’re getting into the highly speculative bit.)

I’ve long been struck by the correlations between gender-non-conformity, being attracted to members of your assigned sex at birth, and gender dysphoria. As the Genderbread Man points out, there is no necessary relationship between these things. And yet tomboys are more likely to grow up to be lesbians. You could say “maybe they’re signalling that they’re lesbians by being gender-non-conforming!”, but that’s a bit hard to square with the tomboys who weren’t attracted to anyone yet and who fully expected to grow up to be straight.

It is almost as if there are three switches, one of which says Figure Out What People Of Your Sex Are Supposed To Do In Your Culture And Do That, one of which says Be Attracted To People Of The Other Sex, and one of which says Feel Strongly That You Are A Member Of Your Sex. And then some factor– perhaps prenatal?– has something like a 50% chance of flipping over each individual switch. So a minority of the population has all three (that is, they are straight feminine trans women), and a lot of people have one or two, but there’s still a strong correlation between the positions of the three switches.

So that’s our first type of trans woman.

Small but interesting studies suggest a correlation between gender dysphoria and autism, as do the anecdotes of myself and others. Note that when I say “autism” here, I am deliberately being very broad and including people with the broader autism phenotype, as well as so-called “optimal outcome” people who are autistic in childhood but do not meet criteria for autism as adults. With the exception of the military, the specific professions in which trans women are overrepresented are also professions in which autism broadly defined is overrepresented. (Possibly including the military? I am not a military person myself, but I can see the attraction among autistics of clear rules, a chain of command, strict routines, and getting to play with machines.) Perhaps the actual division is between trans women whose transness is caused by autism and trans women whose transness is caused by switch-flipping?

And that’s our second type of trans woman.

Now, the million-dollar question is: where are all the trans male autistics? Why aren’t the trans dudes all programmers too? First, it is very possible that there are more male-sexed autistics than female-sexed autistics; while autism tends to present differently in female-sexed people, leading to their underdiagnosis, it is also possible that some of the difference is that there are legitimately fewer female-sexed autistics. It is also possible that– given that autism is likely to be many many different things— the reason that female-sexed autistics present differently is that they’re likely to have a different kind of autism, which doesn’t happen to be linked to gender dysphoria.

Second, the “people connected to the gay community more likely to transition” argument is even stronger for trans men than for trans women. Until recently, most of the people who transitioned were trans women. Presumably, this is because trans men were less likely to be independent than trans women. In a time period where many gatekeepers required that the trans person cut ties with everyone they knew, trans women would be more likely to have useful skills (engineering!) that could be used to rebuild a career. Trans men would be more likely to be housespouses, which does not teach you very many transferable skills and ends as soon as you cut ties with everyone you know. This might also make trans men more reluctant to e.g. found or join crossdresser groups, because if you’re financially dependent on your spouse it’s much less likely they’ll suck it up and deal with it than if your spouse is financially dependent on you.

Therefore, in historical samples, we should expect the vast majority of trans men to be non-autistic. In present-day samples, if #1 is mostly true, we would expect mostly switch-flipping trans men, while if #2 is mostly true, we would expect a population of autistic trans men. Once again, I await replication.