Guest Post: Rekindling the Proportionality Test for Protective Discrimination under Article 15(3) [Part I]

[This is a guest post by Kartik Kalra. It is the first of a two-part series on the evaluation and application of a rekindled framework to test protective discrimination under Article 15(3).]


Article 15 consists of a set of prohibited grounds, limiting Parliament’s ability to treat classes differently based on religion, race, caste, sex and place of birth. It creates an exception for “women and children” within the Article, letting the state make sex-based classifications in the form of “special provisions.” The history of Article 15(3) is widely considered one of perpetuating sex-based stereotypes in the name of protection, with some authors locating its role in “justifying provisions which [are] downright derogatory”.

In this two-part series, I examine the history of case law on Article 15(3), propose a rekindled proportionality test to save discriminatory laws under that Article, and apply this framework to test the saving of pensionary benefits and their cessation on a widow’s remarriage. At times, striking down stereotype-perpetuating laws solely because of their stereotypical nature may produce socially sub-par outcomes, for some stereotypes are often true and laws are made in cognizance of the same. Section 125 of the CrPC is one such example, which creates a regime of maintenance only for women, children and parents, excluding men from its domain. The pensionary regime, which will be discussed in the following post, is another such example: the state recognizes that following the death of their husbands, women may be dependent, and makes provisions for pensions accordingly. Once a widow remarries, the state considers that they may no longer depend on it, for they would look to their new husband for their sustenance. This logic, while based on a quantifiably true understanding of the economic dependency of women over men, feels uncomfortable. A law of this kind, despite taking cognizance of true stereotypes, also risks actively perpetuating them. The effects of validating such a law would reaffirm the dependency stereotype, causing a disadvantage to women as a class. Simultaneously, women being the beneficiaries of maintenance or pensionary regimes, are also advantaged by such laws.

Navigating this line between recognizing and entrenching stereotypes is akin to walking on eggshells, for there exists no determining principle to save some stereotype-perpetuating laws within Article 15(3) and strike down others. In this piece, I attempt to develop such a determining principle in the form of a four-pronged test that is based on quantifiability, historical disadvantage, lesser-discriminatory means, and the balancing of advantages and disadvantages. In order to do so, I first assess Article 15(3) in light of quantifiably true stereotypes, followed by positing this new test and demonstrating its functionality.

Stereotype Perpetuation and Article 15(3)

The usual account of the evolution of Article 15(3) jurisprudence is as follows: there is one line of cases where the victimization of women occurs in the name of ostensible protection, to which cases such as Yusuf Abdul Aziz v. State of Bombay and Leela v. State of Kerala belong; and that there is a more recent line of cases that includes Anuj Garg v. Hotel Association and Joseph Shine v. Union of India that deny the saving of stereotype-perpetuating laws within Article 15(3). The former line is considered to lie in a paradigm of romantic paternalism that portrays women as the weaker sex that needs the continued protection of men. One author considers romantic paternalism an unwritten bargain where “economic support and protection given by the male [is exchanged] for subordination in all matters, sexual service and unpaid domestic work”. In exchange for the ostensible protection that the state gives women under Article 15(3), it presumes their subordination in the private realm. The latter line of cases does away with this paradigm, taking into consideration the historical disadvantages faced by women that presently necessitate special provisions, and limiting the protective ambit of Article 15(3) only to such provisions.

This separation does not, however, hold true for case law on quantitatively-verifiable stereotypes that also confer a discernible advantage on women. Many laws take into account a state of economic dependency that women have on men: compassionate appointments to government jobs, maintenance u/s 125 of the CrPC, sex-based reservations, and the pensionary regime are some examples. This can also be factually assessed: only a fifth of India’s female population is working, as opposed to nearly four-fifths of men. Laws recognizing this stereotype and making disparate treatment have been subject to multiple constitutional challenges. In State of Andhra Pradesh v. P.B. Vijayakumar, a rule preferring women over men in state employment was under challenge. Recognizing the historical disadvantages faced by women that necessitated a rule that officially prefers women over men, the Court acknowledged that special provisions are primarily intended to address the historic economic exclusion of women:

7. The insertion of clause (3) of Article 15 in relation to women is a recognition of the fact that for centuries, women of this country have been socially and economically handicapped. As a result, they are unable to participate in the socio-economic activities of the nation on a footing of equality. It is in order to eliminate this socio-economic backwardness of women and to empower them in a manner that would bring about effective equality between men and women that Article 15(3) is placed in Article 15. Its object is to strengthen and improve the status of women.

In Ramesh Chander v. Veena Kaushal, the constitutionality of Section 125 of the CrPC was challenged on the ground that it provides for maintenance only for women, excluding men. This was saved on the ground that it recognizes an existing reality and functions as a measure of social justice (¶9). Chaturbhuj v. Sitabai reaffirmed this purpose of the Section 125, pointing to the general state of economic dependency of all classes eligible for maintenance (¶5); Ramesh Rege v. Gauri Rege held that a law providing for maintenance for unmarried daughters is based on their economic dependence and confers an advantage on them (¶8); Cyril Britto v. Union of India held that a law prohibiting the arrest of women judgement-debtors would be saved due to their economic dependance, for the law serves as a “recognition of a reality” (¶8);  Shrikrishna Eknath Godbole v. Union of India reaffirmed the role of a law preventing the arrest of women judgement-debtors to concern their economic dependence (¶9,10); Dattatreya Motiram More v. State of Bombay upheld a law reserving seats for women by referring to the advantage being conferred on a historically excluded group:

7. The proper way to construe Article 15(3), in our opinion, is that whereas under Article 15(1) discrimination in favour of men only on the ground of sex is not permissible, by reason of Article 15(3) discrimination in favour of women is permissible, and when the Stale docs discriminate in favour of women, it does not offend against Article 15(1). Therefore, as a result of the [j]oint operation of Article 15(1) and Article 15(3) the State may discriminate in favour of women against men, but it may not – discriminate in favour of men against women.

In the above set of cases, we can observe the invocation of distinct determining principles to save laws under Article 15(3): Ramesh Chander relies on quantitative verifiability, P.B. Vijayakumar focusses on remedying historical disadvantage, and Dattatraya More relies on the conferral of an advantage on women in the form of reservations of seats. Another principle has been invoked in Anuj Garg:

50. “The test to review such a protective discrimination [under Article 15(3)] statute would entail a two-pronged scrutiny:

(a) the legislative interference (induced by sex discriminatory legalisation in the instant case) should be justified in principle,

(b) the same should be proportionate in measure.

Further, consider the following remark in Vasantha v. Union of India, a case assessing the constitutionality of a rule prohibiting the employment of women in factories at night:

72. In the present case, the provision is not a protecting provision so that it could be tested with reference to Article 15(3) or Article 15(4) of the Constitution but it is a restriction and therefore, the validity of such restricting provision has to be tested in the anvil of Article 15(1) of the Constitution.

The Court, therefore, drew a line between protective and restrictive provisions, indicating that a law curtailing the rights of women cannot be saved under Article 15(3). This was also invoked by Chandrachud J. in his concurrence in Joseph Shine:

189. Article 15(3) encapsulates the notion of “protective discrimination”. The constitutional guarantee in Article 15(3) cannot be employed in a manner that entrenches paternalistic notions of “protection”. This latter view of protection only serves to place women in a cage. Article 15(3) does not exist in isolation…Neither Article 15(1), nor Article 15(3) allow discrimination against women. Discrimination which is grounded in paternalistic and patriarchal notions cannot claim the protection of Article 15(3).

It can, therefore, be observed that Courts have offered fragmented, and at times, intuition-based reasonings to save some laws under Article 15(3) but reject others. In the following section, I coalesce these distinct reasonings into a four-pronged test to determine whether a law can be saved by Article 15(3).

A Test to Evaluate Protective Discrimination

At this stage, it must be noted that all cases except Anuj Garg and Joseph Shine addressed laws based on stereotypes whose quantitative veracity was verifiable, which ultimately concerned the socioeconomic disadvantage and economic exclusion of women. Preventing their arrest due to judgement-debt, calling for special maintenance provisions, and the reservation of seats are all ultimately exercises in the recognition of stereotypes. Laws made in pursuance of the dependency stereotype, therefore, have been unwaveringly upheld.

Stage 1: Quantitative Verifiability

On this basis, I propose the first prong of the test, which would require the law to be made in pursuance of a stereotype that is quantitatively verifiable. If the law is based on a stereotype that isn’t true at all, no question of its saving under Article 15(3) arises. If the stereotype of lack of safety in establishments selling alcohol is false, then all matters end there, and the law cannot be saved within Article 15(3). Consider the use of this prong in Joseph Shine: since the law was based on the stereotype of a woman’s lack of sexual autonomy, which is undoubtedly false, the law cannot be saved as a “special provision” under Article 15(3).

Stage 2: Remedying Historical Disadvantage

The use of the second prong is essential to paint a true image of the reasons necessitating the use of Article 15(3). Its conceptual underpinnings are reflected adequately in the abovementioned extract from P.B. Vijayakumar, which pertains to the structural economic exclusion of women that causes the present status of economic inequality, necessitating remedial measures. Article 15(3) has the same conceptual underpinnings as that of reservations under Articles 15(4) and 16(4), well-summarized in Subba Rao J.’s dissent in T. Devadasan v. Union of India, approved in all cases thereafter:

26. Centuries of calculated oppression and habitual submission reduced a considerable section of our community to a life of serfdom. It would be well nigh impossible to raise their standards if the doctrine of equal opportunity was strictly enforced in their case. They would not have any chance if they were made to enter the open field of competition without adventitious aids till such time when they could stand on their own legs.

The use of Article 15(3), therefore, must be cognizant of the reasons that necessitate it. Like reservations under Articles 15/16(4), special provisions are created in a context where the historic political and economic exclusion of women has continuing implications. Arguments of petitioners in Cyril Britto or Shrikrishna Godbole, which point to the unfairness in creating criminal immunity for women in offences that men are imprisoned for, miss the larger picture of historical disadvantage that necessitates such laws. A stereotype-perpetuating law, therefore, must have its roots in remedying historical disadvantages.

Stage 3: Availability of Lesser Discriminatory Alternatives

The third prong is an exercise in finding lesser-discriminatory means that can reach the same end that the law seeks to achieve. Essentially, this prong is a test of over-inclusiveness: is there a possibility of the creation of a narrower class to whom the impugned stereotype-perpetuating law can apply, which can avoid the vice of essentialism? The test of over-inclusiveness has been used in equality jurisprudence in, inter alia, Indian Hotel and Restaurant Association v. Union of India, which held the prohibition of any dance performance in some establishments to be bad due to the simultaneous prohibition of both sexually inappropriate and ordinary dances, making it over-inclusive (¶20); Navtej Singh Johar v. Union of India, which held the inclusion of both consensual and non-consensual sexual intercourse u/s 377 to be over-inclusive (¶221); and Inspector Ravina v. Union of India, which held the inclusion of all forms of unwillingness, that which was the result of the exercise of bodily autonomy and that which concerned one’s true unwillingness as a ground to disqualify one from promotion to be bad for over-inclusiveness (¶12).

The state, therefore, would have the obligation to prove that there is no narrower class to whom the stereotype-perpetuating law can apply, which would prevent over-inclusiveness and the portrayal of an entire sex-based class a certain way.

Stage 4: Balancing Between Benefits and Burdens

It is evident from case law that the conferral of a benefit on women is necessary for the law to be saved by Article 15(3), and a restriction cannot masquerade as a protection. Malhotra J.’s concurrence in Joseph Shine also relied on the following reasoning:

274. The true purpose of affirmative action is to uplift women and empower them in socio-economic spheres. A legislation which takes away the rights of women to prosecute cannot be termed as “beneficial legislation”.

There is, at the same time, harm that could result from the saving of laws under Article 15(3), which pertains to the perpetuation of stereotypes. Given the balancing between the conferral of benefits on women and the harm resulting from the perpetuation of stereotypes, this prong may be considered akin to the last prong of the traditional proportionality test. The state, therefore, would have the burden to prove that the benefits being derived from via the protective measures are greater than the harm caused due to the perpetuation of stereotypes.

This stage, like its traditional proportionality counterpart, is also value-based and requires a subjective assessment of the value accorded to particular principles. Thus, even though the exclusion of men from the domain of Section 125 of the CrPC has the effect of perpetuating a stereotype of women as dependent, a judge may come to a conclusion that the benefits being conferred on women via this law outweigh the disadvantages caused by the perpetuation of stereotypes. The treatment of a law of this kind under the third prong, however, is a different issue.

Navigating the Thin Line Between the Recognition and Entrenchment of Stereotypes

The simultaneous evaluation of these four prongs is necessary in order to navigate the thin line between the recognition of a stereotype and its active entrenchment. While Anuj Garg posits a two-pronged test, its context pertains to stereotypes that are neither quantitatively true (lack of safety for women in establishments selling alcohol), nor have the effect of conferring an advantage on women, for the law caused their economic exclusion. For laws that concern a quantitatively verifiable stereotype, an immediate declaration of unconstitutionality may cause a devastating levelling down: the state can argue that there is no scope for maintenance provisions for women if economic independence is the constitutionally mandated starting point. This was the disagreement between O’Regan and Kriegler JJ. In the South African Constitutional Court’s judgment in President of South Africa v. Hugo, a case concerning a remission of prison sentences of mothers with children younger than twelve years. In a challenge to the exclusion of men constituting the perpetuation of sex-based discrimination due to the assignment of child-care duties to women, O’Regan J. made the following remark:

112. To determine whether the discrimination is unfair it is necessary to recognise that although the long-term goal of our constitutional order is equal treatment, insisting upon equal treatment in circumstances of established inequality may well result in the entrenchment of that inequality.

O’Regan J. considered that the denial of the present stereotypical reality of women engaging in child-rearing, while desirable in the long run, would produce an undesirable outcome of their continued incarceration at present. While Kriegler J. ultimately concurred with the majority on the validity of the remission, he considered its underlying reasoning in women’s status as caregivers of children to be constitutionally unsound:

80. In my view the notion relied upon by the President, namely that women are to be regarded as the primary care givers of young children, is a root cause of women’s inequality in our society. It is both a result and a cause of prejudice; a societal attitude which relegates women to a subservient, occupationally inferior yet unceasingly onerous role. It is a relic and a feature of the patriarchy which the Constitution so vehemently condemns.

The navigation of the issue of quantifiably verifiable stereotypes must, therefore, occur with great caution. While the production of socially sub-par outcomes due to an unrealistic commitment to constitutional ideals is undesirable, so is the indefinite perpetuation of sex-based stereotypes. In order to navigate this, the four-pronged test would perform the following tasks: the first prong of quantitative verifiability will ensure that only those rules concerning true stereotypes can be saved by Article 15(3), with the perpetuation of false stereotypes being entirely excluded from that Article’s domain; the second prong will ensure that there is a nexus between the allegedly protective provision and its historical underpinnings, ensuring that a true image of the need of such a law is painted; the third prong ensures that wherever a lesser-discriminatory alternative in the form of a narrower class is available, it is employed instead of the impugned law; and the last prong of the weighing of advantages and disadvantages creates an additional possibility of saving stereotype-perpetuating laws where the Court assesses the benefits being conferred on women to outweigh the disadvantages via the perpetuation of stereotypes.

On this basis, the rekindled test to save stereotype-perpetuating laws within Article 15(3) comprises of the following prongs:

  1. the law must be based on stereotypes that are quantitatively true and verifiable;
  2. the stereotype must be perpetuated in pursuance of remedying historical disadvantages faced by women;
  3. there must be no availability of a lesser-discriminatory alternative than the impugned law; and
  4. the advantage conferred on the historically disadvantaged group (i.e., women) must be greater than the disadvantage generated via the perpetuation of stereotypes.

As I have demonstrated, the present regime on saving laws within Article 15(3) functions in an ad-hoc manner, offering very little in definite determining principles to save some laws but reject others. Even Joseph Shine, which is widely lauded for affirming the true role of Article 15(3), doesn’t rely on a strict principle to save some laws but reject others, relying only on the notion that laws “entrench[ing] paternalistic notions of protection” cannot be saved by the Article. For all practical purposes, this doesn’t leave us in a position better than Cyril Britto and Ramesh Chander, both of which posit that laws made in the “recognition of a reality” can be saved under Article 15(3). The navigation of the line between recognizing and entrenching stereotypes, I submit, can best be done via this four-pronged test.

2 thoughts on “Guest Post: Rekindling the Proportionality Test for Protective Discrimination under Article 15(3) [Part I]

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s