Equality, the Family, and Unpaid Domestic Work: The Judgment of the Kenyan High Court in MW v AN

In an interesting judgment delivered earlier this month, the High Court of Kenya at Nakuru held that the housework and care-work performed by a female spouse (the plaintiff) entitled her to an equal share of the matrimonial property at the time of the dissolution of marriage. The facts of MW v AN were that the parties were married in 1990, separated in 2003, and divorced in 2011. The dispute centred upon the fate of a house constructed at Nakuru. While the house was registered in the name of the male spouse (the defendant), the plaintiff argued that she had taken out extensive loans to finance the purchase of the land and the construction of the house. Moreover, despite having a job herself, she had been the sole care-giver in the family. The defendant, for his part, argued that not only had he bought the plot on his own, but had also been providing financial contributions towards the upkeep of his wife.

Matheka J observed that Section 6(7) of the Matrimonial Property Act of 2013, matrimonial property “vests in the spouses according to the contribution of either spouse towards its question, and shall be divided between the spouses if they divorce or their marriage is otherwise dissolved.” In Echaria v Echaria, it had been held by the Court of Appeal that where there was a “substantial but unascertainable contribution” by both parties, a default rule of equal division would apply. The question, of course, turned upon the meaning of the word “contribution.”

In this context, Matheka J observed that “contribution” would have to include not only tangible financial contribution, but also the “unseen” contribution of housework and care-work. In paragraph 38, she observed:

This other part of mothering, housekeeping and taking care of the family is more often than not not given any value when it comes to sharing matrimonial property. It is easy for the spouse working away from home and sending money to lay claim to the whole property purchased and developed with that money by the spouse staying at home and taking care of the children and the family. That spouse will be heard to say that the other one was not employed so they contributed nothing. That can no longer be a tenable argument as it is a fact that stay at home parents and in particular women because of our cultural connotations do much more work (house wives) due to the nature of the job … hence for a woman in employment who has to balance child bearing and rearing this contribution must be considered. How do we put monetary value to that process where a woman bears the pregnancy, gives birth, and takes care of the babies and where after divorce or separation she takes care of the children single handedly without any help from the father of the children … Should this court take this into consideration when distributing matrimonial property where the husband as in this case is left in the matrimonial home where the wife rents a house to provide shelter for herself and the children? I think it should count, especially where the husband has not supported the raising of the children, has not borne his share of parental responsibility.

Furthermore, this would have to be determined by evidence:

It is time that parties took time to give evidence, sufficient enough to support the value to be placed on the less obvious contribution. It is unfair and unjust for one party to be busy just making their money (the ‘seen’ income) while the other is doing two or three other jobs in the family whose income is ‘unseen’ and then claim this other one did nothing. This attitude is so entrenched we still hear women especially who are housewives say: sifanyi kazi (literally I do not do any work) simply because they do not leave the home to go earn money elsewhere. (paragraph 39).

Consequently, Matheka J held that notwithstanding the fact that the matrimonial property was registered in the name of the husband, the maximum “equality is equity” would apply, and that consequently “the property be valued, sold and each party have 1⁄2 share of the proceeds of the sale.”

The judgment of Matheka J is important because of the explicit recognition it gives to “unseen” and unpaid housework, within the context of domestic relationships; as has been well established by now, across the world and across societies, within the institution of the family, the burden of such work is gendered in nature (see, e.g., The Second Shift) – and often, unseen and unpaid domestic work by the female spouse is what “frees up” the male spouse to enter the labour market and engage in the kind of financially remunerative work that, ultimately, results in (for example) matrimonial property being bought with “his” money, and therefore registered in his name. Thus, departures from traditional notions of property are essential in order to do justice in and within the institution of the family.

It is important to contextualise this judgment, both within the framework of Kenyan and comparative law. In Kenya, the default position used to be (as in many other countries) that only financial contributions were to be taken into account in calculating respective shares in the matrimonial property upon dissolution of marriage. Explicitly seeking to change this, the Kenyan Constitution of 2010 contained Article 45(3), which – borrowed from CEDAW – states that “Parties to a marriage are entitled to equal rights at the time of marriage, during the marriage and at the dissolution of the marriage.” In her book, Equality in Kenya’s 2010 Constitution (2021), Dr. Victoria Miyandazi notes that the intention behind Article 45 was, inter alia, to address “harmful practices such as … unequal claims to matrimonial property upon divorce.” (see pg 42) In Agnes Nanjala Williams vs Jacob Petrus Nicholas Vandergoes, the Court of Appeal directly applied Article 45 between two private parties to mandate an equal division of assets between the spouses, even in the absence of a statutory framework (“horizontal application of rights”).

This position, however, was arguably overruled by the Matrimonial Property Act of 2013, which required judges to take into account the relative contributions of the spouses (as indicated above), but also explicitly specified that the word “contributions” included “domestic work, childcare, and companionship.” The Matrimonial Property Act was challenged in Federation of Woman Lawyers on the basis that the displacement of the 50% rule in favour of “non-monetary contributions” would restore the gendered inequality within marriage, based on the difficulty of calculating non-monetary contributions. This challenge, however, was rejected by the Court.

In that context, the judgment in MW v AN is important, as it essentially restores the position of the default equality rule where there is evidence of “non-monetary contribution”, and allays fears that judiciaries that might not have entirely broken out of patriarchal norms will use the vagueness of the statutory clause to devalue housework or care-work.

Furthermore, this is a position that has been advanced by progressive courts across the world. Perhaps the most outstanding example is New Zealand, where the Property Relations Act of 1976 established a presumption of equal sharing at the time of dissolution, and specifically provided that financial contribution was not to be treated as weightier than non-financial contribution. In numerous judgments interpreting the Property Relations Act, the New Zealand courts have interpreted it with a few towards fulfilling the statutory purpose of achieving the “equal status of women in society”, holding, for example, that wherever the provisions of the Act were ambiguous, the default presumption would be in favour of the property being matrimonial/joint (and therefore, subject to equal division).

Indeed, Matheka J’s language is also remarkably similar to a 1992 judgment of the Colombian Constitutional Court. In Sentencia No. T-494/1992, the Constitutional Court was considering the eviction of a widow from the matrimonial home; the widow’s non-monetary contributions had not been taken into account in determining whether or not she had a legal interest in the home. The Constitutional Court noted that such a position would have the effect of “invisiblising” domestic work, and deepen inequalities within social relations. The Court went on to question the “artificial” distinction between “productive” and “non-productive work”, and noted that refusal to factor in unpaid domestic work would violate the Colombian Constitution’s guarantee of equality and non-discrimination.

The judgment of the Kenyan High Court, thus, joins a global constitutional conversation of how institutional inequalities within the family may be judicially redressed; and it also, I submit, advances the goals of Article 45(3) – itself a fascinating constitutional provision. For these reasons, it deserves careful study by students of comparative constitutional law.

Guest Post: Analysing the Constitutionality of the Reverse Onus Clause in the UP “Love Jihad” Law

[This is a guest post by Shamik Datta and Ishika Garg, and first appeared on the Proof of Guilt blog. It is cross-posted here with permission.]

In February, 2021, the Uttar Pradesh Prohibition of Unlawful Conversion of Religion Act, 2021(Hereinafter “the Law” or “the Uttar Pradesh Law”) was passed by the Uttar Pradesh State Assembly. Section 3 of the Uttar Pradesh Law lays down the offence punishable under the legislation, and reads as: 

“No person shall convert or attempt to convert, either directly or otherwise, any person from one religion to another by use or practice of misrepresentation, force, undue influence, coercion, allurement or by any fraudulent means or by marriage nor shall any person abet, convince or conspire such a conversion”

In relation to this offence, the contentious Section 12 of the Law crafts a reverse onus clause. The clause reads as follows: 

“The burden of proof as to whether a religious conversion was not effected through misrepresentation, force, undue influence, coercion, allurement, or by any fraudulent means or by marriage, lies on the person who has caused the conversion and, where such conversion has been facilitated by any person, on such person.”

A reverse onus clause can be understood as one that places the burden of proof upon the accused once the prosecution establishes certain foundational facts. In the absence of such a clause, the onus would be on the prosecution to show that the conversion was coercive. However, in the present case, this clause would require the accused to prove that the conversion was consensual, as long as the prosecution establishes the basic foundational fact of the occurrence of the act of conversion. In this post, the authors shall analyse the constitutionality of this clause by subjecting it to the four-fold test laid down by Justice Gupte in his separate opinion, in the case of Shaikh Zahid Mukhtar v. State of Maharashtra  [(2017) 2 AIR Bom R 140]. In doing so, the authors shall rely on various judgements that have upheld the constitutionality of reverse onus clauses in other statutes. 

An analysis of the clause, through Justice Gupte’s four-fold test.

Justice Gupte laid down the four-fold test to scrutinise the constitutionality of reverse onus clauses, on the anvil of the fundamental rights of equality and liberty enshrined under Article 14 and 21 of our Constitution (Paragraph 213, Pages 231-232). The four conditions stressed upon were as follows –

  1. Is the State required to prove enough basic or essential facts constituting a crime so as to raise a presumption of balance of facts (considering the probative connection between these basic facts and the presumed facts) to bring home the guilt of the accused, and to disprove which the burden is cast on the accused?
  2. Does the proof of these balance of facts involve a burden to prove a negative fact?
  3. Are these facts within the special knowledge of the accused?
  4. Does this burden, considering the aspect of relative ease for the accused to discharge it or the State to prove otherwise, subject the accused to any hardship or oppression?

We shall now examine the clause through each prong of this test. 

The first condition of the test requires the prosecution to establish certain basic foundational facts which must have a rational connection with the presumed facts, in a way that makes the latter highly probable. This presumption of ‘balance of facts’ (considering the probative connection between the basic facts and the presumed facts) is a prerequisite for presuming the guilt of the accused. In the landmark case of Noor Aga Khan v. State of Punjab [(2008) 16 SCC 417], the Supreme Court upheld the constitutionality of Sections 34 and 54 of the Narcotics Drugs and Psychotropic Substances Act, 1985. The presumption of guilt only arises, when the prosecution proves certain foundational facts, which would shift the burden on the accused to rebut this presumption. 

This crucial requirement of proving foundational facts was reiterated in Dharampal Singh v. State of Punjab [(2010) 9 SCC 608] and Bhola Singh v. State of Punjab [(2011) 11 SCC 653]. The evidence presented by the prosecution must have probative value, which essentially means that it must sufficiently point towards the conclusion that the accused committed the crime. Only when the prosecution satisfies this standard can it be considered just for the accused to displace such presumption of guilt. In the context of NDPS, this standard of proving initial facts would be satisfied once the prosecution established that firstly, there was contraband and secondly, that the same was in the conscious possession of the accused. Similarly, in the recent case of Justin @ Renjith v. Union of India, the Kerala High Court stressed on the importance of necessitating the prosecution to prove certain foundational facts such as the age of the victim, and the occurrence of the alleged incident, in order to uphold the constitutionality of the reverse onus clause in the Protection of Children from Sexual Offences Act, 2012

Contrary to this requirement of establishing basic facts, the Uttar Pradesh Law presumes the guilt of the accused without the establishment of a probative connection between the basic facts and the criminalised act of religious conversion. This is because the only foundational fact to be established in the present case is that of there having been a conversion. In our view, this fact alone is not sufficient to establish a probative connection in this regard. To better understand the problem associated with shifting the burden of proof under the Law without the establishment of a probative connection, we must turn our attention towards the numerous instances where the Law has been misused. From these instances, it becomes clear that even in cases where the conversion took place with the consent of the ‘victim’, the accused has had to face an FIR lodged by the family members of the converted person, or a third party. When such is the ground reality, one must question the absence of proving the non-consensual aspect of conversion from the ambit of the basic foundational facts. The mere fact of there having been a conversion, without establishing the non-consensual nature, cannot suffice to establish the probative connection between the facts and the criminalised act under the Law. Therefore, the Law does not rationalise the connection between the basic and presumed facts and hence, does not fulfil the first condition laid out in the four-fold test.

In P.N. Krishna Lal v. Govt. Of Kerala [1995 Supp (2) SCC 187], the Supreme Court recognised that the casting of burden of proof on the accused cannot be allowed when such presumption is raised in relation to the proof of negative facts. Building on this, the second condition of Justice Gupte’s test enquires whether the aforementioned ‘balance of facts’ involves a burden to prove negative facts. The Uttar Pradesh Law criminalises conversion via ‘misrepresentation, force, undue influence, coercion, allurement, or by any fraudulent means or by marriage’. Section 2(a) of the Law defines ‘allurement’ as:

Allurement means and includes offer of any temptation – (i) any gift, gratification, easy money or material benefit either in cash or kind. (ii) employment, free education in reputed school run by any religious body; or (iii) better lifestyle, divine displeasure or otherwise.

In the present case, the negative fact to be proved by the accused is that no allurement occurred due to their actions. Broad terms like ‘allurement’ and ‘any temptation’, once alleged by the prosecution, require an analysis of the mind of the victim at the time of the commission of the offence, on the basis of whether the party actually felt ‘tempted’ by the actions of the accused. Even if the test of preponderance of probabilities is applied, it is unthinkable how the accused can reasonably or fairly be expected to disprove this subjective feeling of temptation, once alleged by the prosecution. Such a burden imposes an unreasonable standard of proving negative facts on the accused. 

On this account, the Uttar Pradesh Law also fails the third condition of the test which requires the balance of facts to fall within the special knowledge of the accused. As argued by the authors, the proving of negative facts does not lie within the special knowledge of the accused in the kind of situations dealt with by the Law. As emphasised in Justin @ Renjith v. Union of India, Parliament is justified in reversing the burden of proof in context of aspects which lie exclusively within the domain of the accused’s knowledge. This case is distinguishable in the present discussion, as it deals with sexual offences committed under the Protection of Children from Sexual Offences Act, 2012. The rationale for such justification was based on the fact that there may not be any eye witness to the incident committed. However, this rationale cannot be deemed to apply in a fool-proof manner in the context of religious conversions, and the authors argue that no other comparable justification is present either. 

It is undisputed that the reversal of burden of proof can only be allowed in those cases where it is evident that there exist certain special facts within the knowledge of the accused, which he can prove or disprove easily. Knowledge, or mens rea, is imputed to the accused in cases such as those of possession of firearms, where mere possession leads to presumption of a harmful purpose [Shaikh Zahid Mukhtar v. State of Maharashtra, (2017) 2 AIR Bom R 140]. However, merely proving the basic foundational fact that a religious conversion took place, cannot be deemed to be in furtherance of a harmful purpose. This is because in multiple cases of religious conversion, it has been noted that the accused had the consent of the converted person, which is not an offence punishable under the Uttar Pradesh Law. The mere proof of conversion does not lead to a presumption that the act was in furtherance of an unlawful, let alone a harmful purpose. Thus, the standard for imputing knowledge to the accused is not satisfied in the present case. With this in mind, the Uttar Pradesh Law fails to fulfill the third condition of the test. 

This leads us to the fourth condition of the test, which enquires whether the burden of proving innocence will subject the accused to any hardship or oppression. In addition to the distress mentioned above, there is also a psychosocial aspect to be considered. In India, religious conversions are widely associated with social stigma and seen as ‘immoral’. This leads to the viewing of the accused as a ‘wrong-doer’, even when the conversion is performed with the consent of the converted person. This situation is further exacerbated when the Law is misused extensively by the prosecution. The justification of reverse onus clauses in light of ‘easier convictions’ cannot be accepted as a higher conviction rate could only increase the possibility of wrongful convictions. A conviction under a reverse onus clause does not necessarily signify guilt, and may just be the undesired result of the accused not being able to satisfy the high standard of proof he is burdened with. An example of the possibility of such wrongful convictions is found in a recent case, where the government explicitly admitted that it did not possess the requisite evidence to prosecute two Muslim men accused of converting a Hindu woman. The woman supported the men, stating that her husband had wrongfully tried to frame them because they supported her when she faced domestic violence from her husband. In all such cases, the hardship faced by the accused is evident and it is clear that the Law does not meet the fourth condition of the test. 


When assessing the validity of Section 12 of the Law, ignoring the inextricable link between constitutional and criminal law would be a grave mistake. The above analysis of the reverse onus clause under the Law, leads to the conclusion that the Law fails to meet the criteria of the four-fold test. Moreover, the rationale behind the law of curbing ‘unlawful conversions’, cannot be justified when analysed in comparison to reverse onus clauses in other statutes, whose constitutionality has been previously upheld by the court. In light of these considerations, the authors believe that the reverse burden of proof clause needs to be reconsidered. 

Affirmative Action as a Remedy for Dispossession: The Judgment of the Ugandan Constitutional Court

On 19 August 2021, the Constitutional Court of Uganda handed down an interesting judgment in United Organisation for Batwa Development in Uganda v The Attorney-General. The case concerned the rights of the Batwa, an indigenous community in South-West Uganda. The Petitioners argued that over a long period of time, spanning many decades, the Batwa had been systematically dispossessed of their lands, first by the colonial government, and then by the actions of the government of Uganda. This constituted a breach of their constitutional rights under the 1995 Constitution of Uganda (such as the rights to life, equality, and a set of social, economic and cultural rights), and triggered a range of remedies, ranging from financial compensation to recognition of the right to access traditional forest land.

There are two features of this judgment that repay careful study. The first is the extent of evidence that the Petitioners brought on record to demonstrate the historical ties of the Batwa community with the land in question. This evidence took the form of individual testimonies, combined with expert analysis submitted by way of affidavit (see pgs 20 – 32 of the judgment); this expert analysis included material from the colonial archives, which demonstrated how the colonial government had displaced the Batwa in order to create game reserves. The Respondents were unable to rebut this evidence, arguing only that as of date, the Batwa were not located on forest land, and their access to it was limited to procuring some forest produce. As the Court recognised, that, of course, was a non-sequitur, because the displacement was historical.

From a legal point of view, it is the second feature that is particularly interesting. The Court located the remedy within Article 32 of the Ugandan Constitution, which provides for affirmative action:

Notwithstanding anything in this Constitution, the State shall take affirmative action in favour of groups marginalised on the basis of gender, age, disability or any other reason created by history, tradition or custom, for the purpose of redressing imbalances which exist against them.

The Constitutional Court held that the text of Article 32 indicated an affirmative obligation upon the government to undertake affirmative action measures, tailored to two requirements: first, it must be established that the claimant group is marginalised (call this the “entitlement” requirement); and secondly, that the nature of the measure must be such that redresses these structural imbalance (call this the “fit” requirement). In the present case, on the basis of the evidence before it, the Court found that it was beyond cavil that the Batwa had been – and continued to be – a marginalised group. This fulfilled the entitlement requirement, and thus triggered an obligation upon the State to introduce affirmative action measures.

In other words, therefore, the Court held that a generic affirmative action provision in the Constitution created a right in favour of marginalised groups, and vested in them the standing to approach the Court for enforcement. This is as it should be; however, it is also important to note that a generic affirmative action guarantee (as in Article 32(1) of the Ugandan Constitution) is nothing more than the principle of substantive equality, expressed in more concrete terms. Logically, therefore, what the Ugandan Constitutional Court accomplished through Article 32(1) ought, in theory, to be accomplished in Constitutions that do not have a specific affirmative action guarantee, but nonetheless subscribe to a theory of substantive equality. More to the point, in the Indian context, this shows once again how the Indian Supreme Court’s repeated statement that “there is no right to reservation” is entirely at odds with its simultaneous jurisprudence that the Indian Constitution guarantees a right to substantive equality. It is illogical to hold, on the one hand, that there is a right to substantive equality, while also holding, on the other, that people who have been deprived of this right have no remedy for it.

Back to the issue: the Ugandan Constitutional Court then went on to hold that the tailoring of affirmative action measures could not be undertaken purely through evidence by way of affidavit. It was a fact-intensive enquiry. Under Article 137(4) of the Ugandan Constitution, it was open to the Constitutional Court to refer a matter to the High Court “to investigate and determine the appropriate redress.” Availing of this provision, the Constitutional Court directed the High Court to expeditiously hear evidence and determine the matter, keeping in mind that affirmative action measures put in place “do not expose the Batwa people to further exploitation, are practically effective, and are enjoyed by all the Batwa people.” (pg 46)

It is of particular significance to note that upon finding that a State obligation of affirmative action was triggered, the Constitutional Court did not remand the matter to the government, to devise an appropriate affirmative action programme; instead, it remanded the matter to the court below, to hold (what appears to be) a trial on the issue.

Now if you look at Article 32(2) of the Ugandan Constitution, which immediately follows the affirmative action provision, it is provided there that “Parliament shall make relevant laws, including laws for the establishment of an equal opportunities commission, for the purpose of giving full effect to clause (1) of this article.” This formulation is a familiar one in the context of comparative constitutional law: rights that require structural or institutional solutions to implement are often followed by a sub-clause that delegates their implementation to the legislature (see e.g., Article 17 of the Indian Constitution, prohibiting “untouchability”). It is therefore noteworthy that the Constitutional Court nonetheless held that the crafting of affirmative action measures was a judicial task, to be undertaken after judicial consideration of evidence. In this, the judgment is somewhat reminiscent of the Indian Supreme Court ruling in NALSA v Union of India, which specifically held in favour of affirmative action for the transgender community, and the concrete form that it would take (enforcement, of course, ran into hurdles).

Consequently, the judgment sets down the crucial principle that: (a) under the Constitution’s generic affirmative action principle, it is open for groups to directly approach the Court as rights-bearers, and trigger an obligation upon the State to fashion an affirmative action programme; and (b) if the claimants are successful in meeting the entitlement requirement, then the remedy will be judicial in nature, i.e., the Court will hear evidence and determine affirmative action remedies that meet the “fit” requirement. Taken together, these principles represent a highly progressive approach to affirmative action, which firmly grounds it as a right, that can be enforced by claiming specific remedies, which can be granted even where the government chooses to drag its feet and not fashion an appropriate affirmative action scheme. Of course, this approach comes with its own risk: as has been seen around the world, the exclusion of the government from the task of fashioning social welfare schemes, and the transposal of that power to the judiciary, can bring with it its own set of pathologies. Given that, students of comparative constitutional law may be particularly interested in following these proceedings through the High Court, where the question of the appropriate affirmative action scheme will be debated in the coming weeks.

Guest Post: The New Pension Rules 2021 and the Freedom of Speech of Government Employees – II

[This is Part Two of a two-part series by Harsh Jain and Eeshan Sonak. Part One may be found here.]

On 31st May 2021, the Central Government notified an amendment to Rule 8(3A) of the Central Civil Services (Pension) Rules, 1972, that restricts retired officials in intelligence/security related organizations from critically commenting on issues relating to current policy that fall within the ‘domain’ of the organizations they served. Failing to seek prior clearance could result in the withdrawal or withholding of the official’s pension. Rule 8 of the CCS Pension Rules relates to officials’ “pension subject to future good conduct.” Previously, ‘good conduct’ meant that the official must not be convicted of a serious crime by a court of law, or be found guilty of grave misconduct. The explanation to the rule clarifies that grave misconduct would mean the disclosure of a secret obtained while holding office under the government, as mentioned in Section 5 of the Official Secrets Act, 1923. The amendment is an attempt to transgress these accepted restrictions, and punish retired officials who use their expertise to criticize the government and point out the flaws in its governance.

In the previous part of this article, we have argued depending on the nature of their work, the government can, in the interests of discipline, efficiency, and confidentiality, impose certain reasonable restrictions on its employees above and beyond those applicable to other citizens. However, what constitutions ‘reasonable restrictions’ is left open for the courts to adjudicate depending on the facts and circumstances of each case. We now argue that the amendment imposes an unconstitutional restriction to free speech for it is unreasonable, disproportionate, vague,  and overbroad. But before moving into these arguments, we wish to make two preliminary submissions that shall have a bearing on the rest of our arguments.

Firstly, pension is not a bounty, charity or a gratuitous payment, but an indefeasible right of every employee. It not an ex-gratia payment that depends on the sweet will of the employer, but a staggered form of payment for the past service rendered. Such was the view taken in D.S. Nakara v. Union of India (1982).

Secondly, restrictions on free speech must pass a higher threshold of reasonability vis-à-vis other rights. Justice Nariman in Shreya Singhal v. Union of India (2015), noted that “restrictions on freedom of speech must be couched in the narrowest possible terms.” Justice V.S. Deshpande, writing extra-judicially, discussed how the liberal wing of US Supreme Court judges advocated a doctrine that the freedom of speech and expression is preferred to the less important rights such as the right to property. According to them, a restraint on the freedom of speech and expression is prima facie unconstitutional unless the state can justify it. The Supreme Court of Colombia has also held that any limitation on freedom of expression “is presumed to be suspect, and therefore must be subject to a strict constitutional analysis” (see also a report by the Special Rapporteur for Freedom of Expression by the Inter-American Commission on Human Rights). Justice Deshpande was of the firm view that these principles are applicable in India as well.

On the subject of the constitutionality of the New Pension Rules, 2021: the Indian jurisprudence when it comes to the Article 19 rights of government employees is clear. Any restrictions of the Article 19(1) rights of government officials must pass the test of reasonableness laid down in Articles 19(2) to 19(6). As we have argued previously, the threshold of reasonableness for restrictions on the rights of government employees might be lesser than that of ordinary citizens and must be decided by the courts depending on the facts and circumstances of each case. It is our submission that the new amendment to Section 8(3A) of the Central Civil Services (Pension) Rules, 1972 does not pass the test of reasonableness laid down in Article 19(2).

In Sakal Papers v. Union of India (1961), the Supreme Court held that the State cannot place restrictions on the freedom of speech and expression in the general interest of the public. Free speech restrictions are permissible only on any of the eight subject matters contained in Article 19(2). Further, in Superintendent, Central Prison, Fatehgarh v. Ram Manohar Lohia (1960), the Supreme Court held that the restriction must have a ‘proximate’ or ‘imminent’ connection with one of the grounds under Article 19(2). This ‘proximity’ test was strengthened in Shreya Singhal v. Union of India (2015), where the Supreme Court struck down Section 66A of the Information Technology Act by distinguishing between ‘advocacy’ and ‘incitement’ and holding that only the latter could be constitutionally prohibited (more on the proximity test here).

The unamended Rule 8(3A) which was introduced by way of an amendment in 2008, prohibited the publication of materials without prior clearance only when it related to subjects that might affect the sovereignty and integrity of India, the interests of the State, relations with a foreign State, or might lead to incitement of an offence, i.e., most of the recognised grounds in Article 19(2). The amended Rule 8(3A), however, restricts publication of any material which might fall under the domain of the official’s organization, including any expertise or knowledge gained by virtue of working in that organization. Such a restriction goes beyond the subject matters specified in Article 19(2).

Furthermore, it goes against the doctrine of proportionality. The Supreme Court, in cases like Modern Dental College v. State of Madhya Pradesh (2016) and K.S. Puttaswamy v. Union of India (2018), endorsed a four-limb test of proportionality conceptualised by Aharon Barak, the former President of the Supreme Court of Israel. According to the test, a measure restricting a constitutional right must, first, have a legitimate goal; second, it must be a suitable means of furthering the goal; third, there must not be any less restrictive but equally effective alternative; and fourth, the measure must not have a disproportionate impact on the rights holder. It is the third limb or the “least restrictive measure” test that is of relevance in the present context. It requires that the State assess all possible alternatives and use the least restrictive method in pursuance of its goals. The State must factually demonstrate with evidence that the alternatives do not achieve the goal in a more effective manner than the restriction it proposes or else the restriction would be ruled unconstitutional (previously discussed here). A similar standard was used by the Supreme Court in Internet Mobile Association of India v. Reserve Bank of India (2020).

The purported aim of the amendment to Rule 8(3A) was to prevent sensitive information from being put in the public domain by retired intelligence/security officials. As discussed earlier, the unamended rule already prevented the publication of any material pertaining to the security of the State, relations with a foreign State, interests of the nation, etc. Such a limited restriction balanced the right to freedom of speech and expression of the officials and the social importance of effectively restricting such a right. The new rule puts a blanket restriction on the publication of any material within the domain of an official’s organization. The State must factually demonstrate with evidence that the new rule protects sensitive information more effectively than the alternative previously in place. In case it fails to do so, the rule must be held to be unconstitutional for being disproportionate.

The amendment also suffers from the vices of vagueness, overbreadth, and of producing a chilling effect. As was held in Chintaman Rao v State of MP (1950), “so long as the possibility of [a statute] being applied for purposes not sanctioned by the Constitution cannot be ruled out, it must be held to be wholly void.” In State of Bombay v. F.N. Balsara (1951), the Court struck down sections of the Bombay prohibition law that criminalized “frustrating or defeating the provisions of the Act” by, inter alia, noting that the impugned words “are so wide and vague that it is difficult to define or limit their scope.” This is believed to be the origin of the vagueness doctrine and also the first time the Supreme Court gave a nod to the concept of overbreadth (see more on the roots of vagueness here).

Similarly, in Kartar Singh v. State of Punjab (1994), the Court held that “laws should give the person of ordinary intelligence a reasonable opportunity to know what is prohibited, so that he may act accordingly” and said that if the provisions are not clearly defined, the enactment is void for vagueness. Furthermore, the Supreme Court has incorporated the concept of ‘chilling effect’ into Indian jurisprudence through cases such as Kameshwar Prasad, Khushboo v. Kanniammal (2010), and most recently in Shreya Singhal, where it applied all the three doctrines of vagueness, overbreadth, and chilling effect to strike down Section 66A of the IT Act. It said: “Section 66A is cast so widely that virtually any opinion on any subject would be covered by it, as any serious opinion dissenting with the mores of the day would be caught within its net. Such is the reach of the Section and if it is to withstand the test of constitutionality, the chilling effect on free speech would be total… therefore, [it would] have to be struck down on the ground of overbreadth.”

The newly amended rule fails to precisely define the expression “domain of the organization” and gives it an open-ended definition: “including any reference or information about any personnel and his designation, and expertise or knowledge gained by virtue of working in that organisation.” As was the case with Section 66A, the amendment is likely to be misused in selectively censoring any opinions that portray the government in bad light. Alternatively, it would result in self-censorship of retired officials as, in the words of Lt. Gen. H.S. Panag (retd.), “which self-respecting retired government official would like to seek permission from her/his former junior or fight a prolonged legal battle to get his pension restored?

An article written by Siddharth Varadarajan that interviews retired officials captures the sheer overbreadth of the new rules. Varadarajan writes that the amendment would prohibit former R&AW officials from writing on foreign policy, or former Intelligence Bureau officials from writing on communal violence or mishandling of internal security issues or even domestic politics since the IB’s domain includes those. Retired officials, by their years of experience, are best suited to comment on policy matters and convey this information to the general public. They are equipped with the requisite knowledge to call out the government’s mishandling of issues or to suggest ways of improvement through open letters. The amendment prohibiting them from expressing themselves on areas in which they have special expertise is clearly hit by the test of overbreadth.

For these reasons, the courts must recognize that the amendment is unreasonable, disproportionate, vague, and overbroad, and must therefore declare it to be unconstitutional.

Guest Post: The New Pension Rules 2021 and the Freedom of Speech of Government Employees – I

[This is a guest post by Harsh Jain and Eeshan Sonak.]

In an earlier post on this blog, Gautam Bhatia discussed in depth the right to free speech of government employees. The post was written in 2016 in the backdrop of a proposed change to the All India Service (Conduct) Rules, 1968, which would prohibit government officials from ‘criticising’ the government on social media. Five years later, we now see an amendment to the Central Civil Services (Pension) Rules, 1972, which goes even further in its effect: it prohibits retired officials in intelligence/security related organizations from publishing anything that falls within the ‘domain’ of the organization they served. A failure to seek prior clearance can lead to an official’s pension being withheld or withdrawn.

In this article, we present a challenge to the constitutionality of the amendment on the anvil of Article 19(1)(a), by examining whether, and to what extent, it is permissible to place greater restrictions on the freedom of speech and expression of retired officials vis-à-vis other citizens. To do so, we trace the decisions of the Supreme Court on the constitutional rights of government employees, and somewhat differ from Bhatia in our conclusion.

Bhatia writes that “in its relationship with its employees, the government assumes two faces: as an employer, with the power to enforce discipline and unity, and as the State, which must respect fundamental rights. The scope of government employees’ rights to freedom of expression and association, therefore, depends upon which of those faces the courts have considered to be the dominant one, and to what extent.” While adhering to this framework, we argue that unlike the narrative of a vacillating Supreme Court presented by Bhatia, its decisions can be read as being consistent in substance and reasoning, though questionable in their conclusions.

Constitutional Rights of Government Employees

The Supreme Court decided on the validity of terms of service that restrict government employees’ rights for the first time in P Balakotaiah v. Union of India (1957). A few railway employees, who had been terminated for engaging in “subversive activities” by organizing a general strike in association with communists, challenged their termination before the Supreme Court as a violation of their fundamental right to form associations. A Constitution Bench of the Supreme Court upheld their termination and stated that the orders of termination:

… do not prevent them from continuing to be communists or trade unionists. Their rights in that behalf remain after the impugned orders precisely what they were before. The real complaint of the appellants is that their services have been terminated; but that involves, apart from Article 311, no infringement of any of their constitutional rights.” As Bhatia puts it, the underlying premise behind the Supreme Court’s refusal to invoke Article 19(1)(c) was that  “since there is no antecedent fundamental right to government employment, the government is free to make employment conditional on the requirement that employees do not associate with communists.

Five years later, in 1962, two Constitution Benches of the Supreme Court took a radically different approach from Balakotaiah. In Kameshwar Prasad v. State of Bihar the Court struck down a rule that prohibited government servants from participating in demonstrations or strikes concerning their conditions of service. The Court noted that entering government service does not disentitle a person from claiming the freedoms guaranteed to every citizen, and therefore, any restriction on Article 19(1)(a) or (c) would have to satisfy the reasonableness test under Articles 19(2) or (4) respectively. It held that in prohibiting all forms of demonstrations, without showing any proximate link with public disorder, the rule was overbroad and hence, void. Similarly, a few months later in O.K. Ghosh v. E.X. Joseph, the Supreme Court struck down a rule prohibiting government servants from joining associations not recognized by the government.

However, in 1998, a two-judge bench of the Supreme Court yet again changed tack. In M.H. Devendrappa v. Karnataka Small State Industries, the Court upheld the dismissal of an employee on the basis of a rule which prohibited employees from doing “anything detrimental to the interests or prestige of the Corporation.” The Court reasoned that “a rule which is not primarily designed to restrict any of the fundamental rights cannot be called in question as violating Article 19(1)(a).” To this extent, as Bhatia notes, it is an incorrect position of law as the Supreme Court, in Bennett Coleman v. Union of India, abandoned the “object and form” test for the violation of fundamental rights in favour of the “effects” test.

Bhatia then goes on to discuss how the Court in Devendrappa wriggled out of the difficulty when faced with the “conflicting precedents” in Balakotaiah and Kameshwar Prasad/O.K. Ghosh. Describing the complexity involved, Bhatia says that “Kameshwar Prasad and O.K. Ghosh incorrectly interpreted Balakotaiah, and were then themselves incorrectly interpreted by Devendrappa, giving us, at the end of the day, an initial Constitution Bench decision and a later two-judge bench decision on one side, and two Constitution Bench decisions in the middle on the other.”

It is this part of Bhatia’s criticism that we disagree with, for reasons that will shortly become clear. To clarify, we do not seek to defend Balakotaiah or Devendrappa, and in fact, believe – as Bhatia points out – that they are strewn with glaring errors. Instead, we seek to focus on those aspects of the judgments that lie at the very core of the constitutional rights of government employees and argue that Kameshwar Prasad and Balakotaiah are not entirely “conflicting precedents” and that there is scope to reconcile them when looking at their core aspects. Amidst the numerous differences in reasoning and conclusions, there runs a commonality that is consistent through all four judgments. The following three principles consistently emerge from them:

  1. Government employees cannot be said to form a class apart to whom the rights guaranteed by Part III do not, in general, apply.
  2. Depending on the nature of their work, duties they are required to discharge, and information they are privy to, reasonable restrictions above and beyond those applicable to other citizens may be placed on Government employees in the interest of discipline, efficiency, and confidentiality.
  3. The question as to what constitutes ‘reasonable restrictions’ is subjective, and is left open for the courts to determine on a case-by-case basis.

So far as the first principle is concerned, Kameshwar Prasad (para 18) and O.K. Ghosh (para 11) unequivocally say in no uncertain terms that government servants can claim fundamental rights. We believe that this principle is also implicit in Balakotaiah and Devendrappa. Balakotaiah took the view that the appellant’s rights under Article 19(1)(c) have not been infringed since they can continue to be communists or trade unionists, but they have no fundamental right to be continued in employment by the State. Whatever be the correctness of this view, there is, arguably, a tacit recognition that the State cannot infringe upon the fundamental rights of government employees.  

Now coming to Devendrappa, we believe that its emphasis and repeated re-iteration of the second principle impliedly affirms the first principle. Devendrappa focuses on the reasonableness of the restrictions and says:

Proper exercise of rights may have, implicit in them, certain restrictions. The rights must be harmoniously construed so that they are properly promoted with the minimum of such implied and necessary restrictions. In the present case, joining government service has, implicit in it, if not explicitly so laid down, the observance of a certain code of conduce necessary for the proper discharge of functions as a government servant. That code cannot be flouted in the name of other freedoms. Of course, the courts will be vigilant to see that the code is not so widely framed as to unreasonably restrict fundamental freedom. But a reasonable code designed to promote discipline and efficiency can be enforced by the government organisation in the sense that those who flout it can be subjected to disciplinary action.

Contrary to Bhatia’s claim that the Court in Devendrappa tried to wriggle out of the difficulty of “conflicting precedents,” it highlighted Kameshwar Prasad and O.K. Ghosh as shining examples of how the courts have made sure that the impingement on the fundamental rights of Government employees is minimal. It concluded that “a proper balancing of interests of an individual as a citizen and the right of the State to frame a code of conduct for its employees in the interest of proper functioning of the State, is required.In our opinion, this iteration alone captures all three principles that we culled out.

Kameshwar Prasad and O.K. Ghosh also advance a similar proposition. Kameshwar Prasad says that “if the rule had confined itself to demonstrations of a type which would lead to disorder, or which would fall under the other limiting criteria specified in Article 19(2) the validity of the rule could have been sustained.” This means that government employees may have special rules or restrictions as long those restrictions fall within Article 19(2). If certain actions by government employees may lead to disorder, which if done by ordinary citizens would not lead to the same, then those rules can be sustained under a 19(2) enquiry. Kameshwar Prasad then goes on to give specific two instances where such rules would be sustained: that of an income tax officer mandated to maintain secrecy of documents under the Income Tax Act and an election officer mandated to do the same under the Representation of the People Act. The Court held that rules regulating the use of information obtained by such employees in the course of their duties by virtue of their official position do not infringe the right to free speech.

Bhatia argues that the very narrowness and specificity of these examples directly contradict the broad interpretation that Devendrappa placed upon Kameshwar Prasad, namely that “fundamental rights challenge could be avoided on the ground of requiring proper discharge of duties by government servants.” We believe that Devendrappa did not hold that a fundamental rights challenge could be avoided on this ground, but that it would have to be tested on a different threshold of ‘reasonableness’ by taking into consideration the interests of discipline, efficiency, and confidentiality in the discharge of their duties.

Similarly, O.K. Ghosh also held that “there can be no doubt that government servants can be subjected to rules which are intended to maintain discipline amongst their ranks and to lead to an efficient discharge of their duties. Discipline amongst government employees and their efficiency may in a sense, be said to be related to public order.” It added, however, that “a restriction can be said to be in the interests of public order only if the connection between the restriction and the public order is proximate and direct.” In the particular facts of that case, the Court struck down the law as unconstitutional holding that “it is difficult to see any direct or proximate or reasonable connection between the recognition by the government of the association and the discipline amongst, and the efficiency of, the members of the said association.”

Thus, the first two principles establish that owing to the nature of their work, which may entail consuming sensitive information, government employees can be subjected to stricter regulations vis-à-vis other citizens. It is also established that these regulations must be ‘reasonable’ in line with Article 19(2). What is left unanswered is the question of what constitutes ‘reasonable restrictions.’ We believe that it is this question that the courts have answered inconsistently, if at all. Balakotaiah did not undertake any reasonability enquiry and went too far in holding that the government can terminate its employee even for the proper exercise of her fundamental right. Devendrappa too, in our opinion, erred in upholding the dismissal on the ground of proper functioning of the government organisation. In contrast, we believe that Kameshwar Prasad and O.K. Ghosh reflect the correct exposition of law. Regardless, an undeniable observation that arises from this analysis is that the courts have adopted different standards of reasonability when testing the constitutionality of the law or when applying it to different cases.

Hence, we conclude that what constitutes a reasonable restriction in the special circumstances of government employees is left undecided. It is up to the discretion of the courts to determine whether a restriction is reasonable after considering the particular factors of a given case like the nature of employment and the extent to which the law restricts one’s fundamental right. It is this subjectivity that we grapple with in the next part of this article, where we present a challenge to the constitutionality of the amendment to the Central Civil Services (Pension) Rules, 1972. We call this amendment the ‘New Pension Rules, 2021’ because of how drastically it expands the requirement of good conduct after retirement; so much so that it is not a stretch to say that it introduces a new code in itself for availing pension.

However, before we proceed to discuss the New Pension Rules, there remains unaddressed one last argument by Bhatia: the doctrine of unconstitutional conditions. Bhatia places the incipience of this doctrine into Indian jurisprudence in the case of Re Kerala Education Bill (1958). The doctrine prohibits the State from denying citizens a benefit by making access to that benefit conditional upon citizens’ abstaining from exercising any or all of their fundamental rights (discussed in greater depth in another post by Bhatia; available here). Bhatia argues that this pronouncement impliedly overruled Balakotaiah because though there is no antecedent right to government employment, making one’s employment conditional upon her abstaining from exercising her constitutional rights to free speech and association would no more be permissible.

While we agree that the reasoning advanced in Balakotaiah would not stand the test of the unconstitutional conditions doctrine, we believe that it is important to clarify that this doctrine cannot be held as absolute, and is itself subject to implied limits of reasonable restrictions. Since the unconstitutional conditions doctrine originated in the United States, let us see some of the US Supreme Court decisions relating to this doctrine.

Professor Emeritus Thomas McCoy of Vanderbilt University discusses the limitations and proper application of the unconstitutional conditions doctrine. He says that if a restriction is reasonably necessary for the effective performance of the contract, the unconstitutional conditions doctrine will not apply, and the court will uphold the contractual restriction on speech. For example, a position with the Central Intelligence Agency is routinely conditioned on an individual’s acceptance of significant restrictions on her freedom to speak about classified information to which she will have access. In Snepp v. United States (1980), the Supreme Court upheld the speech-restrictive condition in the CIA employment contract saying that the unconstitutional conditions doctrine will not apply since such restrictions are essential to the effective performance of the CIA agent’s duties. Similarly, in Rust v. Sullivan (1991), the Court refused to apply the doctrine holding that a government employee engaged in family counselling using government-specified speech, contractually agrees to forego the advocacy of her own viewpoints during the counselling time. However, the Court has also ruled that the Government cannot discourage unfavourable speech unrelated to contract performance. In Elrod v. Burns (1976), the Court held that a city government cannot offer employment as a police officer on the condition that the employee refrains from making speeches in her spare time that are critical of the mayor’s political views.

In India as well, secrecy and non-disclosure are essential requirements of certain positions. Kameshwar Prasad itself gives two such positions: income tax officers and election officers. Thus, we see that it is permissible to restrict free speech to the extent it involves classified information. Therefore, the doctrine of unconstitutional conditions does not affect the application of the three principles we culled out as commonalities in all four judgments.

(Part Two to follow.)