Journal of Law and the Biosciences, 181–193
Advance Access Publication 3 April 2017
Notes & Developments
e testosterone rule—constructing
fairness in professional sport
Law School, Harvard University, Cambridge, MA 02138, USA
∗Corresponding author. E-mail: Msudai@sjd.law.harvard.edu
is note engages in the public debate over the participation of female athletes with
hyperandrogenism in professional sport. Hyperandrogenism, a condition leading to
higher testosterone levels than would be expected for ‘biological females’, is thought
by many to provide aected athletes with a competitive edge over non-aected ath-
letes.1In 2011, when public aention was drawn to runner Caster Semenya, interna-
tional sporting organizations implemented policies to monitor testosterone levels in
female athletes as part of traditional ‘sex testing’.2e new policy required that female
athletes have a testosterone level under 10 nmol/L in order to be eligible to compete as
female, and recommending that athletes with higher testosterone levels seek medical
advising for further treatment.3
e new policy was criticized for essentially expelling hyperandrogenic athletes from
participating in professional sports unless they undergo dramatic medical modica-
tions, and was recently suspended temporarily by the Court of Arbitration for Sport
(CAS) aer runner Dutee Chand refused to regulate her testosterone levels in order
∗Maayan Sudai is a fellow at the Program on Science, Technology and Society at Harvard Kennedy School
and an S.J.D. Candidate at Harvard Law School.
1CLINICAL FINDINGS IN HYPERANDROGENISM,THE JOHNS HOPKINS MANUAL OF GYNECOLOGY AND OBSTETRICS
538 (Johnson Clark et al. eds., 5th ed., 2015).
2IAAF, REGULAT IONS GOVERNING ELIGIBILITY OF FEMALES WITH HYPERANDROGENISM TO COM-
PETE IN WOMEN’SCOMPETITION (2011) hps://www.iaaf.org/download/download?lename=
20Competition%20-%20In%20force%20as%20from%201st%20May%202011 (accessed Jan. 16, 2017).
[hereinaer IAAF Regulation].
e Author 2017. Published by Oxford University Press on behalf of Duke University School of Law,
Harvard Law School, Oxford University Press, and Stanford Law School. is is an Open Access arti-
cle distributed under the terms of the Creative Commons Aribution-NonCommercial-NoDerivs licence
(hp://creativecommons.org/licenses/by-nc-nd/4.0/), which permits non-commercial reproduction and distri-
bution of the work, in any medium, provided the original work is not altered or transformed in any way, and that
the work is properly cited. For commercial re-use, please contact email@example.com
182 re testosterone rule—constructing fairness in professional sport
to compete as a female.4e CAS decision reenergized the cultural and scientic
testosterone debate, and was marked by the introduction of new arguments regard-
ing whether or not the testosterone rule serves the objective of ‘fairness’ in sports.
Aer briey reviewing the historical origins of the testosterone rule and sex testing
in organized sport, this note describes existing scientic data and its interpretations re-
garding sex. While maintaining that the relationship between a high testosterone level
and athletic performance has not been proven, but also not refuted, the last part oers
preliminary thoughts on three alternative approaches to achieve fairness in sports.
I. THE RISE AND FALL OF THE TESTOSTERONE RULE
A. e formation of the testosterone rule
Professional sports bodies have been troubled by the dicult issue of sex classication
for many years.5Testing for sex was rst enforced by the International Olympics Com-
miee (IOC) in 1968. is practice was implemented in order to mitigate anxieties at
the time that Soviet male athletes would compete as females to increase the Commu-
nist medal count, but also to relax older and more cultural worries of masculinization of
6However, these sex classication tests were never simple to adminis-
ter nor very reliable. Initially, female athletes were required to expose their bodies for
physical scrutiny by examiners; however, mounting complaints led the IOC to employ
a chromosomal test instead.7e rst athlete to challenge these tests was Spanish hur-
dler Maria Patino, whose examination in 1988 by the Olympics’ ‘Femininity Control
Head Oce’ resulted in the discovery of a Y chromosome.8Patino had androgen in-
sensitivity syndrome, a condition in which the cells are not receptive to testosterone.9
Although she was born with a Y chromosome, leading to the development of testes and
the manufacturing testosterone and estrogen, her body reacted solely to the estrogen
she produced and therefore developed in a feminine form (breasts, round hips, etc.).
Patino lived her entire life as a woman and had the strength traditionally expected of
women, but was banned from participating in the 1988 Olympics. Although she was al-
lowed to compete a few years later, the IOC continued to insist that even if the Y chro-
mosome was found to be an unreliable criterion for sex testing, some form of testing
must be done.10
e recent controversy regarding sex verication tests in professional sport was tied
to a dierent dominant biological sex marker: hormone levels, particularly testosterone.
4Chand v. Athletics Federation of India (AFI) & e International Association of Athletics Federations (IAAF),
org%2Fleadmin%2Fuser upload%2Faward internet.pdf&usg=AFQjCNGqgPIuaQX3ATk9iT0hw0UQL
g27A (accessed Jan. 16, 2017) [hereinaer CAS Arbitration].
5For the purpose of this note, I refer to ‘sex’ as the category of biological traits usually aributed to males or
females, dierent from ‘gender’ and ‘gender identity’ the social role aributed by society to individuals, or the
sense of self experienced by individuals.
6ANNE FAUSTO-STERLING,SEXING THE BODY:GENDER POLITICS AND THE CONSTRUCTION OF SEXUALITY,1–4
9Chad Haldeman-Englert, Androgen Insensitivity Syndrome,MEDLINEPLUS (Sept. 11, 2014),
hps://www.nlm.nih.gov/medlineplus/ency/article/001180.htm (accessed Jan. 16, 2017).
10 ANNE FAUSTO-STERLING,supra note 6, at 2.
e testosterone rule—constructing fairness in professional sport r183
Aer the domination of South African runner Caster Semenya in the 2009 Track and
Field World Championship, during which many suspected that she was a man, or
not biologically female,11 the International Association of Athletes Federation (IAAF)
decided to institute new regulations regarding hormone levels for athletes, aimed to
prevent a reoccurrence of the Semenya case.12 e new policy, issued in 2011 aer
18 months of expert consultation, targeted women with hyperandrogenism, dened
therein as a physical condition leading to ‘excessive production of androgens (testos-
terone)’.13 e 2011 IAAF policy maintained that ‘the dierence in athletic perfor-
mance between males and females is known to be predominantly due to higher levels
of androgenic hormones in males resulting in increased strength and muscle develop-
ment’.14 Continuing the desire to preserve ‘fair’ competition that led to the sex separa-
tion in competition in the rst place, the regulations proposed a three-stage diagnosis.
e rst is an initial clinical examination which requires taking a medical history and a
IAAF Regulation, Appendix 2: Medical Guidelines for the
Conduct of Level 1 and Level 2 examinations (2011) hp://
(accessed Jan. 17, 2017).
e second level is a preliminary endocrine assessment, requiring blood and urine
samples to be analysed for specic types of androgenic hormones.16 If the rst
and second-level examinations reveal preliminary signs of aberrant presentation
compared to what would be expected for a biological female, the athlete is re-
ferred to the third level: a full examination, which includes further genetic and
e Longman, Understanding the Controversy Over Caster Semenya, THE NEW YORK TIMES (Aug. 18, 2016),
hp://www.nytimes.com/2016/08/20/sports/caster-semenya-800-meters.html (accessed Jan. 16, 2017).
12 IAAF, IAAF to Introduce Eligibility Rules for Females With Hyperandrogenism, IAAF NEWS (Apr. 12,
2012), hps://www.iaaf.org/news/iaaf-news/iaaf-to-introduce-eligibility-rules-for-femal-1 (accessed Jan.
14 IAAF Regulation, supra note 2, at 1.
15 Id. at Appendix 2.
16 Id at 7.
184 re testosterone rule—constructing fairness in professional sport
Hirsutism scoring sheet according to Ferriman and Gallwey,
IAAF Regulation, Appendix 2: Medical Guidelines for the Conduct
of Level 1 and Level 2 examinations (2011) hp://www.bmj.
20Regulations%20(Final)-AMG-30.04.2011.pdf (accessed Jan. 16,
hormonal laboratory work.17 Aer all data are collected, an expert medical panel
determines whether or not the athlete is eligible to compete based on the ath-
lete’s androgen levels.18 e normal range for males is dened in the regula-
tions as ‘more than 10 nanomoles per litre’ (10 nmol/L). If athletes are past
this threshold, they cannot compete as female, regardless of other aspects of their
B. Conicting decisions regarding the testosterone rule
More recently, two conicting decisions alternatively undermined and restored the
testosterone rule in professional sport. e rst decision was given by the CAS in
favor of Dutee Chand, an Indian sprinter who was banned from competing as a fe-
male due to her high androgen levels. Chand, apparently diagnosed with hyperandro-
genism, was initially required to undergo normalizing medical procedures, hormonal or
17 Id. at 9.
18 Id. at 12.
19 Athletes that are past the 10 nmol/L threshold but have androgen resistance will be allowed to compete as
females on the grounds that they derive ‘no competitive advantage from having androgen levels in the normal
male range’, Id.
e testosterone rule—constructing fairness in professional sport r185
surgical, which would decrease her natural testosterone levels to the permissible
range.20 In her leer to the IAAF Chand wrote:
e high androgen level produced by my body is natural. I have not doped or cheated. If I
follow the IAAF guidelines you have aached, I will have to undergo medical intervention
in order to reduce my naturally produced androgen levels. [...] I feel perfectly healthy
and I have no health complaints, so I do not want to undergo these procedures [...] I
also understand that these interventions will most likely decrease my performance level
because of the serious side eects and because they will interfere with the way my body
has worked my whole life.21
Chand appealed the Athlete Association of India and IAAF’s joint decision to ban
her from participating in competitions, and received a landmark ruling in her favor
in July 2015. e CAS decided that hyperandrogenism regulations were discrimina-
tory toward women, and based on insucient scientic evidence. Although all parties
agreed that lean body mass (LBM) aects athletic performance, they could not agree
on the inuence of testosterone on LBM.22 So although Chand hadn’t rmly estab-
lished that ‘testosterone is not a material factor in determining athletic performance’,23
the evidence brought by the IAAF to support the link between testosterone levels and
athletic performance referred only to exogenous testosterone (externally consumed),
and not endogenous testosterone (naturally produced),24 and did not refute her claim
in full. e panel concluded that the IAAF had not established that hyperandrogenic
females have an unfair advantage or that they perform beer due to their naturally
high testosterone levels.25 erefore, ‘the panel [was] unable to uphold the validity of
the regulations’, and suspended the testosterone rule for a period of 2 years to allow
the IAAF an opportunity to gather supporting evidence concerning the inuence of
endogenous testosterone levels on LBM and athletic performance.26
e second relevant decision came in November 2015 from the IOC Consensus Meet-
ing on Sex Reassignment and Hyperandrogenism. Understanding the growing impor-
tance in recognizing gender identity, the statement was intended to liberalize eligibil-
ity rules for the participation of transgender athletes while still preserving the objec-
tive of ‘fair competition’.27 e updated institutional policy no longer required surgical
20 Juliet Macur, Fighting for the Body She Was Born With, THE NEW YORK TIMES (Oct. 6, 2014), hps://www.
(accessed Jan. 16, 2017).
21 CAS Arbitration, supra note 4, at 9–10.
22 LBMrepresents body weight thatis not fat, so athletes with a highLBM are seen as having an athletic advantage.
See Id. at 39.
23 Id. at 141.
24 Id. at 142.
25 Id. at 155.
26 Id. at 158.
27 INTERNATIONAL OLYMP ICS COMMITTEE,IOCCONSENSUS MEETING ON SEX REASSIGNMENT AND HYPER-
ANDROGENISM 2 (2015), hps://stillmed.olympic.org/media/Document%20Library/OlympicOrg/IOC/
Reassignment-and-Hyperandrogenism.pdf# ga=1.51499305.1851528621.1481057226 (accessed Jan. 16,
2017) [hereinaer IOC Consensus]; See also IOC Rules Transgender Athletes Can Take Part in Olympics
186 re testosterone rule—constructing fairness in professional sport
modication for athletes to complete.28 Instead, the consensus instructed that female
to male transitioning athletes will be able to compete in the male category with no re-
strictions, while male to female (MTF) athletes will have to ‘demonstrate that her total
testosterone level in serum has been below 10 nmol/L for at least 12 months prior to
her rst competition’.29
is consensus statement largely liberalized eligibility requirements regarding
surgery for transgender athletes. Yet it also reinforced the testosterone rule as the de-
marcating criteria between the sexes to constitute fairness between athletes in the same
sex category, contrary to the CAS decision in Dutee Chand’s case. e IOC addressed
the contradiction directly and encouraged the IAAF to ‘revert to CAS with arguments
and evidence to support the reinstatement of its hyperandrogenism rules’.30 e meet-
ing proposed allowing females with hyperandrogenism to compete as males as a solu-
tion to discrimination.31
II. DOES TESTOSTERONE PROVIDE A COMPETITIVE ADVANTAGE?
Policymakers dealing with this topic usually refer to the only empirical evidence avail-
able: two large-scale studies conducted in 2014 designed to measure the hormonal pro-
les of elite athletes. But these studies yielded somewhat conicting results. One study,
oen referred to as GH-2000, was a ‘spin-o’ from a project designed to trace abuse
of growth hormone in sport.32 By the end of the original experiment (conducted in
2012 during the London Olympics), there was sucient serum for the study of hor-
monal proles of 693 elite athletes.33 e blood samples were drawn from 454 males
and 239 female athletes in 15 competition categories within two hours of their com-
petition. Results showed that contrary to what researchers had expected, there was a
substantial overlap in testosterone levels between the sexes, as 16.5 per cent of males
demonstrated low testosterone levels (under 8.4 nmol/L, the lower limit of the nor-
mal reference range for males), whereas 13.7 per cent of females demonstrated high
testosterone levels (above 2.7 nmol/L, the upper limit of the normal reference range
for females).34 However, the most distinctive criterion in dierentiating between male
and female athletes was their LBM,35 as the research established that females have 85
per cent of the LBM of males.36 Researchers believe that these ndings are sucient to
Without Surgery,GUARDIAN (Jan. 24, 2016), hps://www.theguardian.com/sport/2016/jan/25/ioc-
rules-transgender-athletes-can-take-part-in-olympics-without-surgery (accessed Jan. 16, 2017).
28 IOC Consensus, supra note 27, at 2; Simon Hart, IAAF Oers to Pay for Caster Semenya’s Gen-
der Surgery if She Fails Verication Test,T
ELEGRAPH (Dec. 11, 2009), hp://www.telegraph.co.uk/
verication-test.html (accessed Jan. 16, 2017).
29 IOC Consensus, supra note 27, at 2.
30 Id. at 3.
31 Id. at 2.
32 M. L. Healy, et al., Endocrine Proles in 693 Elite Athletes in the Postcompetition Seing,81CLIN.ENDOCRINOL.
294 (2014) [hereinaer GH2000 study].
33 Id. at 298.
34 When comparing between females with testosterone levels above and under the 8.4 nmol/L threshold, re-
searchers found a variety of dierences in many criteria, including fat mass, height, etc. Id. at 295. e study
had also established a strong correlation between weight and body mass index, which correlated less strongly
with LBM. Id. at 297.
35 Id. at 298.
36 Id. at 302.
e testosterone rule—constructing fairness in professional sport r187
account for ‘observed dierences in strength and aerobic performance’ between male
and female athletes, ‘without the need to hypothesize that performance is in any way de-
termined by the dierences in testosterone levels’.37 e researchers additionally sug-
gest that the ndings ‘negate completely the hypothesis concerning testosterone levels
proposed by IAAF/IOC’.38 e authors conclude that hormonal proles of elite ath-
letes dier from the usual reference range, and that ‘the IOC denition of a woman as
one who has a normal testosterone level is untenable’.39
e other study, commissioned by the IAAF and conducted at the 2011 IAAF Track
and Field World Championships in Daegu, South Korea, is referred to as the Daegu
study.40 is study measured testosterone levels among 849 female athletes, with a
goal to estimate the prevalence of hyperandrogenism and other disorders of sex de-
velopment (DSD) among high-level female athletes.41 Results demonstrated that me-
dian testosterone levels among elite female athletes were similar to those of non-athlete
healthy young females (0.69 nmol/L median found in sampled athletes), with the 99th
percentile calculated at 3.08 nmol/L.42 Out of 839 women tested, 9 had testosterone
levels greater than 3 nmol/L, and 3 women had levels above 10 nmol/L.43 Despite the
plausible speculation that high-level athlete women would demonstrate higher testos-
terone levels than their non-athlete counterparts, this hypothesis was not conrmed in
the data.44 On the other hand, the researchers argued that although the link between
testosterone levels and athletic performance had not been directly tested, the much
higher prevalence of hyperandrogenic females among high-level athletes relative to the
general population (approximately 140 times higher in this study), may be ‘indirect ev-
idence for performance-enhancing eects of hyperandrogenic DSD concentration in
female athletes’.45 e next sections will demonstrate how the two studies have been
received and interpreted by both sides to the debate.
A. Natural testosterone levels do not provide a competitive edge, and therefore
hyperandrogenic female athletes should compete as women
Some of the most vocal defenders of the position that higher testosterone levels do not
provide a competitive advantage are Katarina Karkazis, an anthropologist and bioethi-
cist from the Center for Biomedical Ethics at Stanford University, and Rebecca Jordan-
Young, the author of ‘Brain Storm: e Flaws in the Science of Sex Dierences’. In
2012, aer the IAAF and IOC published their new policies on hyperandrogenism,
Karkazis and Jordan-Young argued that despite common belief, the relationship be-
tween endogenous testosterone and athletic advantage is not scientically proven, and
moreover, would be very hard to establish due to the complex set of reactions dierent
37 Relying on a previous study stated to establish the relationship between LBM and athletic performance, a fact
also established in CAS Arbitration, supra note 4, at 39.
38 GH2000 study, supra note 32, at 298.
39 Id. at 294.
ephane Bermon, et al., Serum Androgen Levels in Elite Female Athletes,99J.CLIN.ENDOCRINOL.METAB. 4328
41 Id. at 4329.
42 Id. at 4334.
43 Id. at 4332.
44 Id. at 4333.
45 Id. at 4334.
188 re testosterone rule—constructing fairness in professional sport
individuals have to similar doses of testosterone.46 eir paper oered instead the im-
perfect, yet preferred criterion of ‘legally recognized females’ to determine who could
compete in the female category.47
Aer the two empirical studies were published, Karkazis and Jordan-Young took is-
sue with the Daegu study.48 eir main criticism concerned the fact that the Daegu
study sample excluded 10 women, 5 due to doping and the other 5 due to DSD condi-
tions. Karkazis and Jordan-Young argued that the decision to exclude the ve women
who demonstrated naturally high testosterone levels from the sample resulted from
the misconceived judgement of women with DSD as non-healthy and an abnormal
‘confounding factor’. ey stated that their exclusion created a circular reinforcement
of the object that requires proof, designing the sample to measure normalcy accord-
ing to the pre-assumed values of what normal is.49 ey also replied to critics of the
GH-2000 study who argued that the test overestimated testosterone levels at lower
ranges and that the decision to draw the serum within two hours from the competi-
tion was a mistake in light of data suggesting that male testosterone levels change in
response to competition.50 Karkazis and Jordan-Young argued in response that even
if the test overestimated the testosterone levels of female athletes, this aw would not
account for low levels of testosterone for male athletes, and therefore the overlap in the
GH-2000 study is not refuted. Regarding timing, the authors referenced other studies
representing the broad position in the scientic literature suggesting that it is the ‘type
and duration’ of competition and not the individual’s sex that inuence testosterone
levels aer a competition.51 Karkazis and Jordan-Young proposed that the criterion for
‘female’ athlete competition should probably be LBM and not testosterone levels, as
suggested by the GH-2000 study.52
Following headlines and media coverage of the Dutee Chand case, other inuen-
tial commentators, such as the editors of Scientic American, provided Karkazis and
Jordan-Young support by arming that ‘there is no evidence that they enhance perfor-
mance’ and that ‘this ongoing state of limbo is a mistake. ere is no scientic basis for
barring these women’.53
46 Katrina Karkazis, et al., Out of Bounds? A Critique of the New Policies on Hyperandrogenism in Elite Female Ath-
letes, 12 AM.J.BIOETHICS 3, 8 (2012) [hereinaer Out of Bounds].
47 Id. at 13.
48 Katrina Karkazis & Rebecca Jordan-Young, Debating a ‘Sex Gap’ in Testosterone, 6237 SCIENCE 858 (2015).
49 Id. at 860.
50 Id. at 859.
52 Katrina Karkazis & Rebecca Jordan-Young, e Trouble With Too Much T,THE NEW YORK TIMES
(Apr. 10, 2014), hp://www.nytimes.com/2014/04/11/opinion/the-trouble-with-too-much-t.html (ac-
cessed Jan. 16, 2017).
53 e Editors, Naturally Occurring High Testosterone Shouldn’t Keep Female Athletes out of Competition,
SCIENTIFIC AMERICAN (Aug. 1, 2016), hps://www.scienticamerican.com/article/naturally-occurring-
high-testosterone-shouldn-t-keep-female-athletes-out-of-competition/ (accessed Jan. 16, 2017).
e testosterone rule—constructing fairness in professional sport r189
B. Testosterone levels do provide a competitive edge, and therefore hyperandrogenic
athletes should not be allowed to compete as women
Karkazis and Jordan-Young stand in contrast to David Epstein, author of ‘e Sports
Gene’54, and Alice Dreger, professor of clinical medical humanities and bioethics and
author of ‘Hermaphrodites and the Medical Invention of Sex’,55 who responded to
Karkazis and Jordan-Young arguing that they misrepresented the current state of sci-
ence by relying on the GH-2000 study. Epstein and Dreger instead suggested that the
GH-2000 study was not suciently credible because it was originally designed to re-
search antidoping; the serum samples were taken post-competition; and samples were
taken from variable sports.56
Epstein and Dreger argue that testosterone does provide a competitive advantage.
Epstein aributes many physical features to testosterone which are commonly assumed
to provide athletic advantage: height, size of limbs, fat mass, etc.57 Epstein then cites
a conversation he had with Dreger: ‘...in many sports, the best female athletes can’t
compete with the best male athletes. And everybody knows that, but nobody wants to
say it. Females are structured like a disabled class for... good reasons’.58 eir belief
that a possible overlap in testosterone levels between sexes should not constitute an
obstacle to the argument that a relationship between testosterone levels and athletic
capacity exists is additionally supported by leading geneticist Eric Vilain, who said that
‘it does not maer whether there is some overlap [of testosterone levels between males
and females] or no overlap at all. .. What maers is, what is the best marker to use to set
men apart from women’.59
III. RECONSTRUCTING FAIRNESS
What criteria should maer when we draw lines between groups of athletes, and what
would best serve the principle of fairness? Whose opinion should determine what is
equal and right? IAAF and IOC? Dutee Chand and Caster Semenya? Epstein and
Dreger or Karkazis and Jordan-Young? Fairness can be constructed in various ways,
based on a multitude of criteria regarding athletic capacity, biological, social, nancial
factors, etc. It seems inevitable that every conguration of fairness in sports will seem
right and equal to some and not to others, and will necessarily lead to a dierent dis-
tribution of costs and benets for stakeholders. To conclude, I will suggest preliminary
thoughts on alternative classication systems for competitions in sport that prioritize
biological and social fairness and elaborate on some of the costs and benets that groups
involved in the debate might experience with these systems.
54 DAVID EPSTEIN,THE SPORTS GENE INSIDE TH E SCIENCE INSIDE THE SCIENCE OF EXTR AORDINARY ATHLETIC
55 ALICE DOMURAT DREGER,HERMAPHRODITES AND THE MEDICAL INVENTION OF SEX (1998).
56 David Epstein & Alice Dreger, Testosterone in Sports, THE NEW YORK TIMES (Apr. 16, 2014), hp://
www.nytimes.com/2014/04/21/opinion/testosterone-in-sports.html (accessed Jan. 16, 2017).
57 David Epstein, How Much Do Sex Dierences Maer In Sports? THE WASHINGTON POST (Feb. 7, 2014),
563b86a4-8ed9-11e3-b227-12a45d109e03 story.html?utm term=.19a090644db9 (accessed Jan. 16, 2017).
59 Laura Geggel, Testosterone Rules for Women Athletes Are Unfair, Researchers Argue,LIVESCIENCE (May. 21,
2015, 06:05pm), hp://www.livescience.com/50938-female-athletes-testosterone-olympics.html(accessed
Jan. 16, 2017).
190 re testosterone rule—constructing fairness in professional sport
A. Keep male–female categories, but follow gender identity
Some have suggested that gender identity should be the conclusive criteria instead of
testing ‘biological sex’, or at least should determine the result in questionable cases.60
Using the gender identity criterion would mean that women with hyperandrogenism
would be allowed to compete as women, including MTF transgender athletes, with no
regulation of their testosterone levels. If we think that testosterone levels do not pro-
vide a competitive advantage, then there should be no problem with this rule. But if we
think that high testosterone levels indeed provide a competitive advantage, then this
proposal means women with lower testosterone levels, particularly those at the boom
of the testosterone scale, would be adversely aected by the new policy, whereas MTF
transgender athletes and hyperandrogenic women would benet from it. We do not
know, however, what degree of harm and benets this would confer, as the relationship
between testosterone level and athletic capacity has not yet been fully measured.
But even those who believe that higher testosterone levels lead to beer athletic ca-
pacity might still support a gender rule, even at the price of ‘sacricing fairness’, on the
grounds of respecting individual liberty to identify as one sees t. Alice Dreger sug-
gested that this could be a solution, even if it isn’t ‘fair’: ‘If we are going to say that peo-
ple get to be divided by gender, and we take gender seriously, then we let people play
according to their gender, not according to biological haves or have-nots’.61
B. Neglect male–female categories for a beer biological indicator
is criterion accepts that ‘sex’ is a problematic and unstable biological category, but
more importantly, an ambiguous proxy for achieving fairness. What, then, would be a
beer biological criterion that would maximize fairness? According to the IAAF/IOC,
it should be testosterone level. A less controversial biological marker could also be
LBM, as presented by the GH-2000 study and supported by the vocal critics of the
testosterone rules. In any case, it seems that any agreed biological criteria would still
be divisive, and individuals would be separated according to predened ranges of
For testosterone level to become the new organizing principle, we would need to
understand rst how to create new ‘fair’ groups: what is the range of testosterone levels
that keeps athletes within a similar level playing eld? But even if that goal is achieved,
a new biological criterion might lead to a new organization just in part, as these so-
called nuanced biological indicators are extracted from and aempt to beer distin-
guish between existing male/female categories. us, it makes sense that even a divi-
sion according to a more ‘accurate’ biological marker would probably preserve a strong
male/female dominancy in each group with some ‘outliers’ in the close margins.
If we divide by testosterone level, this means that women with high natural lev-
els (due to hyperandrogenism or otherwise) might nd themselves competing in the
60 RogerPielkeJr,WhoShouldBeAllowedtoCompeteAsaFemaleAthlete?,THENEW YORK TIMES (July 29, 2015),
html? r=1; (accessed Jan. 16, 2017) Arne Ljungqvist & Joe Leigh Simpson, Medical Examination for Health of
All Athletes Replacing the Need for Gender Verication in International Sports: e International Amateur Athletic
Federation Plan, 267 JAMA 850, 851–852 (1992).
61 Jesse Singal, Should Olympic Athletes Get Sex-Tested at All?, SCIENCE OF US(Aug. 18, 2016, 11:04
AM), hp://nymag.com/scienceofus/2016/08/should-olympic-athletes-be-sex-tested-at-all.html (accessed
Jan. 16, 2017).
e testosterone rule—constructing fairness in professional sport r191
group with higher testosterone level athletes, mostly men. If we show that testosterone
levels indeed provide a competitive edge, this group of female athletes, who used to be
at the head of their category, would be adversely aected by such a policy. Additionally,
if we take seriously the best male vs. best female obstacle, such a policy can make the
prospects of a women winning rst place in the upper testosterone level group close
to zero. Similarly, some men with naturally lower testosterone levels might nd them-
selves competing in the lower testosterone level group. ey might benet from having
easier access to the rst places, but would probably be harmed from moving to the lower
ranked group which consists of mostly women and hence likely to be, sadly, less pres-
tigious.62 MTF transgender athletes and athletes with hyperandrogenism benet from
the policy in that they would be spared the adverse consequences of being subjected to
medical monitoring and control of their natural testosterone levels. But it is less clear
whether they would benet or lose in terms of competition status as we cannot know
in which group they would be situated.
C. Start over
One of the strongest arguments against the IAAF and IOC regulation is that even if
we accept that high testosterone levels provide a competitive edge, why of all physical
traits does testosterone get utmost importance? Karkazis and Jordan-Young say that
‘hyperandrogenism should be viewed as no dierent from other biological advantages
derived from exceptional biological variation’,63 and list documented biological condi-
tions that provide an advantage in certain sports, like runners and cyclists who have rare
mitochondrial variations that give them unusual aerobic capacity; basketball players
with acromegaly, a hormonal condition leading to large hands and feet, etc.64 Scientic
American editors also asserted that ‘elite athletes are by denition physiological out-
liers because of their strength, speed and reexes. Natural hormonal variations, similar
to other intrinsic biological qualities—superior oxygen-carrying capacity in the blood,
for example—are part of that mix. e IOC should say so explicitly’.65
e counter argument to this is that women with hyperandrogenism should not be
able to compete in the female category because females are a protected group. Accord-
ing to that logic, ‘Semenya’s dierence puts her outside the protected athletic cate-
gory of “woman”—and that makes it unfair to the other runners if she is allowed to
compete’.66 Another argument, sometimes made sarcastically, is that these were the
lines we have already decided upon as the rules of the game, and that looking to ne
62 On disparate media coverage, see Jennifer Dutcher, Where’s All the Women’s Sports Coverage? NEWHOUSESYRA-
CUSE UNIVERSITY (June. 30, 2015), hps://communications.syr.edu/womens-sports-coverage/ (accessed
Jan. 16, 2017). On sex pay gap, see Collin Flake, Mikaela Dufur & Erin Moore, Advantage Men: e Sex Pay
Gap in Professional Tennis,48I
NT’LREV.SOC.SPORT 366 (2013); See generally Lea Ann Schell & Stephanie
Rodriguez, Our Sporting Sisters: How Male Hegemony Straties Women in Sport,9W
OMEN SPORT &PHYSICAL
ACTIVITY J. 15 (2000).
63 Out of Bounds,supra note 46, at 11.
65 e Editors, supra note 53.
66 Malcolm Gladwell & Nicholas ompson, Caster Semenya And e Logic Of Olympic Competition, THE
NEW YORKER (Aug. 12, 2016), hp://www.newyorker.com/news/sporting-scene/caster-semenya-and-
the-logic-of-olympic-competition (accessed Jan. 16, 2017). Ross Tucker, e Caster Semenya debate, SCIENCE
OF SPORT (July 16, 2016), hp://sportsscientists.com/2016/07/caster-semenya-debate/ (accessed Jan. 16,
192 re testosterone rule—constructing fairness in professional sport
biological variations would be absurd. As a Professor of pediatrics said: ‘to start breaking
down sport classications by specic biological traits, ‘you’d have to run international
competitions like the Westminster dog show, with competitions for every breed”’.67
But is it so farfetched indeed? Looking at the detailed and bifurcated classication
system of the Paralympics by the International Paralympic Commiee (IPC) suggests
that such a classication can indeed be applied successfully. e IPC classication code
takes into consideration the type of impairment, the degree of the impairment, and the
psychical activities required for a particular sport, in an algorithmic combination of fac-
tors.68 In order to be eligible to compete in the Paralympics, an athlete has to have at
least one eligible impairment from the list provided by the IPC, such as impairment in
muscle power, range of movement, leg length dierence, etc. Under this criterion, some
sports are not open to all impairments. For example, while Goalball is open only to vi-
sual impairment, swimming is open to all impairments. Another important criterion of
eligibility is the degree of impairment, called the ‘minimum impairment criterion’. is
requires that the impairment impacts the athlete’s sport performance.69 en every ath-
lete can be assigned to a ‘sport class’, a category that ‘groups athletes depending on how
much their impairment impacts performance in their sport’,70 specically tailored for
each sport. Each class can have athletes with dierent impairments, but all need to have
comparable eects on their performance, a conclusion that is decided upon by trained
and certied classiers who use a point system. Swimming, a category open to all eli-
gible impairments, has 10 dierent ‘sport classes’ that group athletes with various im-
pairments that have a similar activity limitation in swimming, taking into consideration
issues such as arm and leg movement and strength, muscle power, trunk control, etc.71
Somewhat ironically, the idea behind the classication code that distinguishes between
biological dierences so vigorously is to minimize the inuence of biological physics to
the lowest level possible and ‘to ensure the success of an athlete is determined by skill,
tness, power, endurance, tactical ability and mental focus’,72 an explanation that re-
veals the IPCs’ vision of the telos of sport, as vested in the practice of constant exertion
and endeavor, rather than in naturally occurring advantages.
Could this approach similarly be applied to the class of abled athletes? What if we
were to match the 10 categories of impairment to 10 categories of advantages, where we
list all known biological elements that provide a competitive edge, such as LBM, height,
vision, muscle strength, oxygen carrying red blood cells, lung size, etc.? We could then
assign each athlete a numerical grade in relation to the sport they wish to compete in.
Similarly to the IPC classication code, for each sport, the calculation would be dier-
ent, prioritizing specic traits that benet athletes in that particular sport. Using a list of
advantages rather than looking for a single perfect criterion may help mitigate problems
resulting from trying to enforce rules based on non-existing biological dichotomies.
67 EPSTEIN,supra note 54, at 70.
68 OFFICIAL WEBSITE OF THE PARALYMPIC MOVEMENT,hps://www.paralympic.org/classication (accessed Dec.
69 INTERNATIONAL PARALYMPIC COMMITTEE,EXPLANATORY GUIDE TO PARALYMPIC CLASSIFICATION 3
(2015), hps://www.paralympic.org/sites/default/les/document/150915170806821 2015 09 15%
2BExplanatory%2Bguide%2BClassication summer%2BFINAL%2B.pdf (accessed Jan. 16, 2017).
71 Id. at 30.
e testosterone rule—constructing fairness in professional sport r193
is system would constitute a complete reimagining of our current way of
thinking—a plan to ‘start over’. It may be that some sports would remain ‘sexed’ due to
consistent dierences between males and females within the ‘advantages list’, but other
categories would be completely shued, producing new winners and losers. Aer the
2016 Olympic Games in Rio, where three athletes in the Paralympics had beer times
than the gold medalist in the Olympics 1500 m race,73 it is easy to imagine that those
traditionally considered disadvantaged would outdo the ‘advantaged’.
e controversy over whether or not natural testosterone levels provide a competitive
edge is far from over. is note suggests that testosterone should not necessarily be the
focal point of aention in designing fairness in sports: human biology is far too com-
plicated to be represented in a single criterion. e socially contested background that
brought the testosterone rule to life is likely to continue, and the way it preserves and en-
forces past ideologies is by creating a concept of fairness that is grounded in dierences
between males and females rather than other possible categories of biological variation.
is note is agnostic to the question of whether or not testosterone levels provide a
competitive edge, and leaves it open in order to rethink the concept of fairness in a way
that accounts for whatever the answer to this question may be. It ultimately suggests
that the construction of fairness is a public practice of rationalization and rule making,
which can be reimagined over and over, with winners and losers for any agreed system.
I would like to thank Ido Katri, Reut Cohen, and Lihi Yona for very helpful brainstorming on alter-
native paths to fairness. I would also like to thank Holly Lynch and Quinn Walker for their essential
help in revising this note.
73 Maxwell Strachan, Four Paralympians Just Ran e 1500m Faster an Anyone At e Rio Olympics Final,
HUFFINGTON POST (Sep. 12, 2016, 01:05 PM), hp://www.hungtonpost.com/entry/paralympic-1500m-
t13 us 57d6d3a8e4b03d2d459b7e78 (accessed Jan. 16, 2017).