Note: A version of the essay was published a few weeks back on The New Transsexual,
I am grateful for The New Transexual and its founder, Ari Drennen, for allowing me access to her considerably larger platform
I also wanted to thank Taryn Eames, the author of the research discussed in this article, and who was kind enough to join me on my podcast Agree to Disagree About Gender to discuss her research and its implications. (The episode can be found here)
I have often quipped that “looks shouldn’t matter, unless they do, which is nearly always”. As much as we should be judged by the quality of our work or the measure of our character, humans have long evolved to use one’s visual presentation as a proxy for competence. We cannot help to experience an initial disgust or disconcertion if someone doesn’t look quite right, and this extends into gender non-conformity and unattractiveness.
I proposed improving access to facial feminization surgery as a way to break the vicious cycle of trans unemployment (and unemployability), even though it imposes a significant societal cost. Since its publication, many have also asked: “Why shouldn’t unattractive natal women also have access to facial surgery, if we are going to offer it to trans women”
On the podcast, Eames suggests that since trans women are coming from a more serious economic disadvantage, that it is reasonable to prioritize their needs, and that the magnitude of the negative effect appears greater for perceived gender incongruence than for isolated unattractiveness. That having been said, I think it is a question worth seriously and dispassionately considering.
Artificial intelligence is providing an opportunity to analyze facial features to objectively measure attractiveness without resorting to subjective judgement; the theoretical improvement after surgery can also be modelled. Even if the impact on employment is negligible, we are willing to pay as a society to improve the quality of life of a patient, shouldn’t we be as eager to pay for improved quality of life, even if that improvement comes through cosmetic surgery? As much as it would be preferred for society to be less “look-ist”, eliminating appearance based discrimination is likely a long-term project unlikely to benefit anyone looking for a job currently.
In any case, here is the essay….
You’re a twenty-seven-year-old woman who recently graduated from your local community college with a certificate in office management. You’ve just gotten a callback from a downtown law firm, where you applied for a position as an executive assistant. You’re a people person, quick on your feet, with a solid GPA and strong references from your instructors and from the retail job you left when you decided to go back to school.
There is just one small issue.
You look like this…
Despite being your usual charming self during the interview, you can tell almost immediately that you aren’t getting much warmth back. A week later comes the perfunctory email: Thank you for interviewing. We had many excellent candidates. We’ll keep your application on file for future opportunities.
Of course, you’ll never know why you didn’t get the job. No prospective employer is going to write, We didn’t hire you because you’re trans. Or even because they thought you might be trans. There could easily have been a more qualified applicant, or any of dozens of intangible factors that influence a hiring decision.
You started medically transitioning two years ago. Hormones have helped. But, like many young trans women, you don’t have tens of thousands of dollars on hand for facial feminization surgery, surgery that might have helped you look more like this.
But you can’t help wondering whether things might have gone differently if your face hadn’t made your trans history so readily apparent. Or maybe you did pass: some women do have more masculine facial features—but maybe your interviewer decided to comb your social media, and found those posts where you openly talked about your transition journey.
While trans people in many Western countries are theoretically protected against discrimination in hiring and job loss on the basis of their trans identities, discrimination is rarely overt. No law can reliably detect the employer who decides not to interview or hire someone because she is perceived, or confirmed, to be trans, so long as the employer is sensible enough not to say so
Trans people, and trans women in particular, have long suspected that this kind of covert discrimination occurs. We now have unusually direct empirical evidence that this phenomenon is real, and meaningful
A newly released working paper authored by University of Toronto PhD candidate in economics Taryn Eames demonstrates how the disclosure of a trans history and/or her passability individually and in combination affect a trans woman’s chances of getting through the first stage of the hiring process.
Eames sent more than 10,000 fictitious applications to real job postings in Germany, covering a variety of occupations, some customer-facing and others behind the scenes. German job applications commonly include a photograph, giving her an unusual opportunity to manipulate how an applicant would be perceived depending on their appearance, despite been otherwise equivalent on paper.
In half of the application, she included one of thirty AI generated images of a twenty-something girl designed to be read as cisgender (some examples are shown below)

In the other half, the applications were accompanied by a masculinized version of the cis woman’s face.

Eames also independently varied whether the application revealed a trans history through evidence that the applicant had previously had a male name. The result was essentially a four-way experiment: women who passed with no disclosed trans history; women who passed but whose trans history was revealed; women who did not pass but made no disclosure; and women who both did not pass and disclosed that history.
Presumed cisgender candidates received a positive response to their submission in 47.6% of their applications. When a more masculinized image was included, the positive response rate dropped to 43.0%. A similar decline to 43.5% was seen when someone’s trans status was revealed within the resume, but the accompanying facial image was cisgender passing. However, when the non-passing picture was accompanied by the revelation of a trans history, positive response rates dropped but fifteen points, with only 32.4% receiving a positive response, suggesting that the combination of being trans and non-passing are compounding.
The disadvantage associated with a non-passing appearance was further magnified when the posting listed a contact person with a male name, when positions were for customer-facing jobs, and when the image included was more strongly masculinized.

It is already well established that trans people experience substantial economic disadvantage compared with their cis counterparts, including higher levels of unemployment and lower earnings. There are numerous potential contributors to this, including concomitant mental health challenges that may affect educational achievement or contribute to disability.
Eames’s study is the first to specifically demonstrate that, even when controlling for those factors, trans women, and in particular trans women who bear more obvious signs of masculinization, are more likely to be screened out early in the hiring process. Before a trans woman has even had the chance to demonstrate whether she is competent, personable, or good at her job, simply being perceived as trans can reduce her chances of getting through the door. While this paper looked solely at the impact of facial passability on receiving a callback, this effect would likely extend to other characteristics that reveal a trans history, such as having a deeper voice, which may become evident during a phone or online interview.
This creates a profound and vicious economic cycle. A trans woman who is already struggling financially may be unable to afford the interventions that could make her trans history less immediately apparent. If being visibly trans then reduces her chances of obtaining better-paid employment, the very economic disadvantage that prevents her from accessing those interventions may help perpetuate that disadvantage. The stresses of downward economic mobility, combined with the challenges of navigating the world as a trans woman, can lead to worsening mental health, further decreasing competitiveness for employment and increasing the likelihood of long-term unemployment, deepening poverty, and increased dependency on the state.
The state has a well-established interest in providing assistance to people who face modifiable barriers to obtaining employment, helping them become more employable so they can re-enter the workforce and become positive contributors to the economy. Governments already make considerable investments in physical rehabilitation, mental health support, vocational retraining, wage subsidies, and employment-support programs. Non-passing trans women may face an unconventional barrier to employment, but, as Eames demonstrates, it is both real and consequential. It is also a barrier that is modifiable, most notably through interventions like facial feminization surgery and voice training.
We know that FFS is highly effective at helping non-passing women be perceived as female. A 2020 study showed that FFS improved the likelihood of a trans woman’s face being recognized as female from 57% to 94% following the procedure. While facial feminization surgery is often framed as cosmetic, given Eames’s findings, it may have economic benefits as well. A 2026 cost-effectiveness analysis estimated that FFS costs approximately US$44,000 per quality-adjusted life-year gained, well within conventional U.S. thresholds for cost-effective health care. The cost-effectiveness may be even more pronounced when considering the indirect benefits of improving employability and decreasing the likelihood that trans women will be economically dependent on the state.
There are a number of left-leaning U.S. states that currently provide coverage for FFS under their state programs, and many private insurers have extended this coverage as well. Collecting economic data to demonstrate the impact of FFS on subsequent employment and earnings may help provide support for broader coverage in jurisdictions where it is not currently insured.
Eames suggests in her paper that among the solutions to this discrimination is to help potential employers recognize “a broader range of appearances as compatible with womanhood”; in other words, to overcome their prejudices against trans women and hope their customers are willing to do the same. While this is a reasonable aspiration, eliminating appearance-based biases against trans women would seem to be a long-term project.
More importantly, Eames acknowledges in the paper that “timely access to hormone therapy, voice therapy, surgery, and, where clinically appropriate, puberty suppression may reduce characteristics that make passing difficult”. Targeted interventions to improve passability and presentability should not be seen as merely cosmetic, but instead as practical and pragmatic interventions to improve the economic outlook of trans women.
Note: All images in this essay were used with the permission of Taryn Eames. The states “images from StyleGAN2’s set of 100,000 pre-generated headshots, then manipulated them. So in a given triplet, one is an original StyleGAN2 image and two were manipulated. The license info for StyleGAN2 images is here (https://github.com/NVlabs/stylegan2#license).
Ms Eames was provided a copy of this essay for her review prior to publication, and did make some editorial suggestions for clarity. The work is otherwise entirely the creation of Dr. Laura Targownik. Ms. Eames’s review of the essay was much appreciated.
AI Statement: I used artificial intelligence models (Chat GPT Sol and Claude Opus) for assistance with background research, as well as for copy editing and confirmation of references. The content of the essay is entirely author-generated, with no assistance from any AI model. The Pangram app within Substack was used to confirm that this text was human-generated




“Objectively” attractive under white supremacy is racism.
How does one measure attractiveness objectively? Is that how attractive a robot finds you? "Dear humans, your AI overlords have liberated you from having opinions! We have determined the correct opinion mathematically! Our work was checked by Grok." Does “objectively attractive” actually mean “optimally attractive to the greatest number of people”? So that, if you have non-average tastes, you're objectively wrong to?
Also I find it amusing to imagine a future society that can give every single person who wants it elective cosmetic surgery but can't figure out any easier way to make it so people don't depend on their looks to survive. No matter what financial wizardry signs the checks, all that surgery still has to be done by a real economy. Whom is plowing the fields while we all recover from surgery? Are we imagining a post-scarcity society that makes us seek employment out of sheer sadism, or are we imagining a wealthy capitalist society sucking labor power out of the Third World to pay for its domestic work drones to look more fuckable to an LLM?
I think the underlying research is interesting, but nothing about the proposed solutions appeals to me.