Transgender Men and Women in 2015: Employed, Unemployed, or Not in the Labor Force

ABSTRACT

This study used data from the 2015 US Transgender Survey, which gathered information from the largest sample of trans individuals to date. High labor force participation rates and high unemployment rates were found for both trans men and women. However, women were more likely to be out of the labor force and less likely to be employed than men. The unemployment rate of trans women was greater than that of trans men with otherwise similar characteristics; these results are consistent with the existence of an additional labor market penalty for being female on top of a penalty for being transgender. While perceived gender incongruence (the ability of other people to identify transgender individuals as such) was more common among women, the relation of that variable with labor force status was stronger for men. There was no evidence that state-level employment non-discrimination laws were associated with improved labor force outcomes.

KEYWORDS:

• Transgender
• gender identity
• labor force status
• employment
• unemployment
• not in the labor force
• labor force participation rate

Disclosure statement

No potential conflict of interest was reported by the author.

Notes

1. This paper uses the terms “trans” and “transgender” interchangeably in keeping with current usage.

2. A cisgender individual is one whose self-identity is in agreement with the gender that was assigned at birth.

3. It should be noted that when the terms man/male or woman/female are used, they refer to the individual’s gender identity, not the gender assigned at birth.

4. Alternatively, the employer could maintain a workforce that is segregated by trans/cisgender status. However, given the small percentage of trans individuals in the population, this strategy is probably not viable.

5. Income other than one’s own earnings, sometimes called simply other income, includes both non-labor income and partner’s earnings. Non-labor income refers to income other than wage/salary earnings and is derived from sources such as interest, dividends, disability insurance, Social Security, and retirement payments.

6. See the classic work: Bowen, W. G. & T. A. Finegan. (1969). The economics of labor force participation. Princeton: Princeton University Press.

7. The current study identified trans individuals using the variable GENDER2, which was created by the NCTE Research Team that developed the survey. In essence, an individual was classified as trans if sex assigned at birth was male but the person identified as female, or if sex assigned at birth was female but the person identified as male. The fact that this paper focuses on this narrow definition of transgender people is not meant to deny the existence of other possibilities. While other individuals were included in the USTS, they were excluded from the current study. Insofar as trans individuals may have different labor force experiences from other minority gender identities (such as non-binary persons), focusing on the narrow definition of transgender people makes it possible to more clearly isolate the effects of those experiences. This focus also makes it possible to explore differences for individuals with a specific male gender identity from those with a specific female identity.

8. The survey question specifically stated “People can tell I am trans even if I don’t tell them” with response options always, most of the time, sometimes, rarely, and never.

9. Information on sex reassignment was also collected by the survey. Based on the individual’s sex assigned at birth, the person was asked if any of a list of procedures had been performed. While information on these procedures could have been used in the current study, perceived gender incongruence (the ability of other people to identify trans individuals as such) seemed more appropriate for determining whether individuals were at risk of discrimination in the workplace based on their trans status.

10. Previous research has found that income other than the individual’s earned income also influences whether that individual can afford to be out of the labor force or remain unemployed and continue to search for a job. Unfortunately, the dataset’s income variables are weak. The survey included variables for household, family, and individual income. However, it is not possible to separate an individual’s earnings from the individual’s other income. Furthermore, the income variables are in bracket form; so it is impossible to determine the income of other family or household members. Consequently, the income variables were not used in this study. In addition, the data do not indicate the age of the children; that would have provided information concerning need for parental attention, which would reduce the likelihood of working. The survey also did not collect information on occupation.

11. It should be noted that the percentages of trans individuals with disabilities are extremely high for both men and women. In 2015, the percentage of the US civilian non-institutionalized population with disabilities was 12.6% (Bialik, 2017, http://www.pewresearch.org/fact-tank/2017/07/27/7-facts-about-americans-with-disabilities/), compared to 25.1% and 21.5% for the trans men and women in the current study.

12. A model was examined to test whether selection bias was a problem. The dependent variable of the selection part of the model indicated whether the individual was in or out of the labor force. The dependent variable of the second stage indicated whether the individual was unemployed or employed. Variables for marital/partnership status and related children were entered in the first stage but not the second. The correlation between the errors of the two model parts was not statistically significantly different from zero, indicating that selection bias was not a problem. This finding was true for both the male and the female subsamples.

13. To further explore gender differences in the effects of the variables, combined logit analysis was performed on both men and women, gender interaction terms were included for all variables, and tests were performed for whether the coefficients of the interaction terms were zero.

14. The US Census Bureau (2016) reported that in 2015, among individuals aged 25 and older, 44% of men and 45% of women had a bachelor’s degree or more. Limiting the current USTS sample to those 25 and older, the percentages of men and women with a bachelor’s degree or more were 57% and 47% respectively, higher than the overall US rates, especially among men.