Privacy, Data Protection, and the National Population Register

The constitutionality of the National Population Register [“NPR”] – authorised by the Citizenship Rules of 2003 – is presently under challenge. Grounds of challenge include questions about excessive delegation, the doctrine of ultra vires, disproportionate burdens upon rights, and so on. At its heart – as the Citizenship Rules show – the NPR is a data collection exercise, involving the collection, storage, and processing of personal data (for the purposes of preparing a National Register of Indian Citizens [“NRIC”]). Consequently, issues around data protection and the right to privacy (as set out in Puttaswamy) come to the fore.

In this context, the Internet Freedom Foundation has prepared an important briefing paper that examines the interface between the NPR and the ongoing legislative discussion around the proposed Personal Data Protection Bill. IFF’s intervention is particularly important because it engages in a reflective analysis of the NPR and the PDP, in a manner that brings out crucial shortcomings in both.

Constitutional principles stipulate that in cases of non-consensual collection of personal data (as in the NPR), at the minimum, (a) data protection best practices, including data minimisation and purpose limitation must be followed, (b) data collection must follow the principles of necessity and proportionality (especially if rights are affected). Within this framework, some issues come to the fore immediately. The first is the disjunction between the statutorily stipulated “purpose” of the NPR, and the public discourse around it. The IFF Brief points out that – as per the Citizenship Rules of 2003 – the only statutorily mandated use to which NPR data can be put is the preparation of an NRIC. In public discourse, however, the government has claimed that the purpose of the NPR is to ensure targeted delivery of welfare schemes. This – naturally – finds no mention in the Citizenship Rules. This is a classic case of “function creep”, and was witnessed for many years in the case of Aadhaar (until under pressure because of litigation, the Aadhaar Act was passed in 2016).

For this reason, the IFF Brief suggests that Section 5 of the PDP make clear that purpose limitation requires the specific and concrete purposes for which the data will be collected to be “anchored” in primary legislation. This is important for three reasons: first, once the purposes are set out, the proportionality of data collection can be examined with respect to each purpose, separately (for example, in the Aadhaar Case, the Supreme Court found that the proportionality standard was not met for bank accounts and mobile phones, but was met for welfare subsidies and income tax); secondly, it ensures that citizen and resident databases are not open-ended enterprises that – once data is collected – can be used for any purpose at any time; and thirdly – and following from the first two reasons – it rules out interlinking of databases and the accompanying profiling and surveillance (which was explicitly deemed unconstitutional by the Supreme Court in the Aadhaar judgment).

Indeed, the IFF Brief points out how a rigorous implementation of purpose limitation – along the lines suggested above – would bring out clear and evident shortcomings in the NPR process itself. If the statutorily stipulated purpose of the NPR is to prepare an NRIC (caveat: contrary to IFF’s suggestions, this is part of the Rules, and has not been set out in primary legislation), then only that much data as is required for the purposes of establishing citizenship may be collected (caveat #2: this is without prejudice to the overall arguments against the NPR). Thus, for instance, the NPR process cannot mandate the collection of Aadhaar details (as Aadhaar is not a proof or even an indication of citizenship), and the NPR database cannot be seeded with Aadhaar (as has already been partially done).

The IFF Brief points out, in addition, that the PDP Bill’s broad exemption clauses enable the possibility that the entire NPR exercise could be taken out of the ambit of data protection principles altogether. This is obviously a serious problem, and it is worthwhile to note that in the original Srikrishna Bill, the Government was not authorised to simply dispense with the requirements of necessity and proportionality when engaged in non-consensual data collection. The potential non-application of the PDP Bill to an exercise as important and sensitive as the NPR, therefore, demonstrates an urgent need to tighten the Bill’s provisions in order to ensure constitutional compliance.

In sum, therefore, the NPR process is precisely the kind of large-scale, nationwide data-gathering and database-creating exercise that requires principles of data protection to be applied rigorously. When, however, we measure the NPR and the PDP against those constitutional principles (as set out in Puttaswamy and elsewhere), we find important shortcomings in both. The IFF Brief provides an important starting point from which to identify these shortcomings, and also indicates what a strong data privacy regime would look like in practice.

Proving Citizenship: Lessons from the African Court on Human and Peoples’ Rights

Ochieng v Tanzania and Penessis v Tanzania are two recent decisions of the African Court of Human and Peoples’ Rights, that demonstrate a humane and sensitive approach to the vexed question of citizenship challenges. Both cases involved challenges by individuals who had been declared non-citizens by Tanzanian authorities.

In Ochieng v Tanzania (22 March 2018), Tanzania deported a declared “non-citizen” to Kenya, who then “deported” him back, leaving him stuck in what the Court called a “no-man’s land” on the border. The basis of the Tanzanian authorities’ decision was that the Mr. Ochieng had forged his documents to obtain his passport – and, in particular, on the basis of “discrepancies” between the documents and the result of an investigation conducted in his village. The Court noted that deprivation of nationality could only be done in compliance with international standards, including the four-pronged test of proportionality. In order to determine whether Tanzania had complied with these standards, the first question that arose was: on which party did the burden lie to demonstrate Mr. Ochieng’s nationality, or lack thereof? In paragraph 80, the Court answered that:

… the instant case, the Applicant maintains that he is of Tanzanian nationality, which is being contested by the Respondent state. In the circumstance, it is necessary to establish on whom lies the burden of proof. It is the opinion of the Court that, since the Respondent State is contesting the Applicant’s nationality held since his birth on the basis of legal documents established by the Respondent State itself, the burden is on the Respondent state to prove the contrary.

The State of Tanzania made a number of arguments that would be familiar to Indian readers, such as spelling discrepancies between documents and testimonies, conflicting oral testimonies, and so on. The Court rejected these arguments, noting that as Tanzania had neither (a) prosecuted Mr. Ochieng for forgery, or (b) conducted a DNA test to determine his link to his parents (who were admittedly Tanzanian citizens) despite him having offered to do so in order to resolve the contradiction between the testimonies, it had failed to discharge its evidentiary burden.

Penessis v Tanzania involved somewhat similar facts. Mr. Penessis was convicted of being an illegal immigrant in 2010, and then placed in detention (also a familiar story for Indian readers). He lost his appeals in the higher courts, which eventually passed orders for his deportation. As in Ochieng’s Case, the dispute revolved around documents: Mr. Penessis claimed that he had a valid birth certificate and Tanzanian passport and was a citizen by birth; the State of Tanzania, on the other hand, argued that he also had British and South African passports, and that Tanzania did not allow for dual citizenship.

Going beyond its judgment in Ochieng’s Case, the African Court referred to international legal instruments to note that “the right to nationality is a fundamental aspect of the dignity of the human person.” (paragraph 87) Consequently, the Court went on to note that “a person’s arbitrary denial of his/her right to nationality is incompatible with the right to human dignity.” (paragraph 88) These observations were crucial, because they went to the heart of the same question that had occupied the Court in Ochieng’s Case: in a dispute over citizenship status, where did the burden of proof lie?

The Court noted, at paragraph 92:

… the burden of proof lies with the alleging party and shifts to the other party only when discharged. Having said that, the Court is of the view that this principle is not static and may be subject to exceptions especially in circumstances where the alleging party is not in a position to access or produce the required proof; or where the evidence is manifestly in the custody of the other party or the latter is entrusted with the means and prerogatives to discharge the burden of proof or counter the alleging party.

That said, citing various other judgments, the Court went on to note:

… when it comes to human rights, this rule cannot be rigidly applied and there must be an exception among other circumstances, where … the means to verify the allegation are likely to be controlled by the state … In such cases, the … the burden of proof is shared and the Court will assess the circumstances with a view to establishing the facts … The International Court of Justice (lCJ) in the Nottebohm Case (Liechtenstein v. Guatemala)2s also held that to determine a nationality link, it is necessary to take into account the very important social factors which bind the Applicant to the Respondent State. (paragraphs 93, 95)

What followed from this, therefore, was that:

… the Applicant who alleges that he holds a certain nationality bears the onus to prove so. Once he has discharged the duly prima facie, the burden shifts to the Respondent State to prove otherwise. (paragraph 96)

On the facts, the Court noted that (a) the State of Tanzania had alleged that the Mr. Penessis’ birth certificate was forged, but had adduced no evidence to prove that; (b) that the facts of Mr. Penessis’ birth were testified to by a person who (according to his birth certificate) was his mother, and that therefore, the initial presumption of citizenship had been established. The Court then observed that:

At this juncture, the Court further notes that all the documents tendered by both parties are copies or certified copies and that neither of the parties adduced originals of the documents used as evidence. In the circumstance, the Court is of the opinion that the Respondent State, as a depository and guarantor of public authority and custodian of the civil status registry, has the necessary means to correctly establish whether the Applicant was a Tanzanian, South African or a British citizen. The Respondent State could also have obtained and produced concrete evidence to support its assertion that the Applicant has other nationalities … In view of the aforesaid, the Court considers that there is a body of documents especially the certified copy of the birth certificate and the certified temporary travel document issued by the competent authorities pending finalisation of the passport, establishing that the Applicant is Tanzanian by birth and that the Respondent State has not been able to prove the contrary. It therefore finds in conclusion that the Applicant’s right to Tanzanian nationality has been violated, contrary to Article S of the Charter and Article 15 of UDHR. (paragraphs 102 – 103)


In both Ochieng and Penessis, therefore, the African Court provided important insights into the vexed issue of burden in citizenship cases. According to the Court, because the bases of proving citizenship (documentary evidence) was in the possession of the individual, the initial burden lay upon him. However, once prima facie citizenship had been demonstrated, the burden would shift to the State, and the State would be held to a strict standard of proof, given the human rights consequences of deprivation of nationality and statelessness.

Compare, however, these judgments with the recent judgment of the Gauhati High Court in Noor Begum v Union of India. The Gauhati High Court upheld a decision of the Foreigners’ Tribunal rejecting the citizenship status of Noor Begum – a decision that involved dismissing no fewer than eight official documents. Curiously, there was no evidence that the documents were forged. What the Court held was that some of the documents were not admissible because they had not been proved by the Issuing Authorities, and even the document that was admissible (a school certificate) did not make it “the end of the matter”, as the Headmaster who had issued it had not been examined. Additionally, Noor Begum produced documentary evidence of the citizenship status of her projected grandfather and father, and oral evidence of her projected mother to draw that link. However, the High Court disaggregated these two pieces of evidence, holding that (a) the documentary evidence of Noor Begum’s projected grandfather and father was worthless because it did not establish her link to them, and (b) the oral evidence of her projected mother (that would have established that link) was worthless because it was only oral, and with no supporting documents! The Court justified this by noting that “the relevant facts being within the knowledge of the procedee, the therefore, the burden of proving citizenship absolutely rests upon the procedee.”

The approach of the Gauhati High Court is in stark contrast to the approach of the African Court of Human and Peoples’ Rights. As both Courts observed, the initial burden of proof rests upon the person claiming citizenship (the logic being that that person would be in the possession of relevant information required to substantiate that claim, and it would be difficult for the State to “prove a negative”, so to say). But while the African Court went on to hold that this was only an initial burden, which would flip once a prima facie case was made out, the Gauhati High Court turned it into an “absolute burden” (there is no warrant for that in the Foreigners Act). The High Court then placed a series of essentially onerous conditions upon the applicant (such as having to examine the headmaster of a school she graduated from twenty years ago) which – it should be obvious – would be very difficult to fulfil. Noticeably, in none of this was there any allegation that any document was forged, or that the oral testimony was false; this is specifically important because – as the African Court correctly noted – when it comes to the question of forgery, it is actually lies more within the government’s power to demonstrate that (its own official) documents have been forged, rather than for the applicant to (now) “prove the negative”.

It is submitted that the international presumption against statelessness and the link between stateless and human dignity requires a nuanced and sensitive approach towards the question of burden of proof in citizenship cases. This is especially true in countries like Tanzania and India, where there exist large numbers of marginalised and vulnerable people, whose access to documents is limited to start with. The approach of the African Court represents the correct balancing of interests, and – it is submitted – would be a better fit with Indian constitutional commitments, as opposed to the zero-sum approach of the Gauhati High Court. Perhaps that task will lie with the Supreme Court.

The Citizenship (Amendment) Act Challenge: Three Ideas

Thus far, the constitutional debate around the Citizenship (Amendment) Act has been framed around the following arguments: (a) does the grant of immunity and citizenship to a select group of migrants violate the principle of “reasonable classification” under Article 14, by virtue of the individuals and groups it excludes?; (b) does the selection of groups lack any “determining principle”, and is therefore unconstitutionally arbitrary?; and (c) by privileging religious persecution over other forms of persecution in claims to citizenship, does the CAA violate the basic feature of “secularism”?

In this post, I attempt to move beyond these basic arguments which have – by now – run their course in the public sphere. Beyond reasonable classification, arbitrariness, and secularism, I will suggest that there are deeper reasons to hold the CAA unconstitutional. As it will become clear, a closer engagement with these reasons will require us to rethink some of our long-held assumptions about Indian constitutional law. As I shall argue, however, these are not radical or off the wall arguments, but rather, implicit within constitutional practice. I shall argue, first, that the principle of equality under the Indian Constitution has moved beyond the classification and arbitrariness tests (as I have argued before); secondly, that – contrary to a widespread assumption in our legal culture – citizenship laws deserve greater judicial scrutiny instead of judicial deference; and thirdly, that notwithstanding the language of Article 11 of the Constitution, there exist implied limitations upon Parliament’s power to confer or withdraw citizenship – limitations that flow from the existence of equally important and fundamental constitutional principles. 

The Evolving Idea of Equality

In the 1950s, heavily influenced by American jurisprudence, the Indian Supreme Court adopted the “classification test” for determining violations of the guarantee of equal treatment under Article 14. The “classification test”, as everyone knows, required that in order for a law to pass Article 14 scrutiny, there must exist (a) an intelligible differentia between the individuals or groups that are subjected to differential treatment, and (b) a rational nexus between that differentia and the State’s purpose in framing the law. Right from the beginning, however, there was a dissenting tradition at the Supreme Court that recognised this approach to be excessively formalistic and constrained. In Anwar Ali Sarkar, for example, Vivian Bose J. asked what “substantially equal treatment” might mean in “the democracy of the kind we have proclaimed ourselves to be.” As Bose J. understood at the time, equality could not be divorced from more fundamental ideas about democracy and republicanism.

In the coming years, the Supreme Court made various attempts to break out of the shackles of the classification framework. For example, it evolved the “arbitrariness” standard – which is, only now, being given flesh and bones, in some of the recent judgments of Nariman J. It also held that the State “purpose” would have to be “legitimate” – i.e., it added a third, more substantive, prong to the classification test. The real breakthrough came, however, with the 2018 judgments in Navtej Johar and Joseph Shine. In reading down Section 377 and 499 of the Indian Penal Code, Constitution Benches of the Supreme Court advanced a richer and more substantive vision of equality, that was also in line with global best practices. In short, the Supreme Court shifted the focus from “reasonable classification” to the idea of disadvantage. True equality – as we can intuitively sense – is about identifying disadvantage, about identifying the axes of diadvantage, and then working to remedy them. 

To recognise and identify disadvantage, however, the law requires proxies. It is here that the observations of Indu Malhotra J., for example – as highlighted in a previous post – become important. As a shorthand for identifying disadvantage, constitutional courts all over the world have asked whether legislation picks out people on the bases of “personal characteristics” that they (a) have no control over, (b) are powerless to change, or can only change at great personal cost. Take, for example, the idea of “race”: a person does not choose the race into which they are born, and cannot – obviously – change their race in any meaningful way. Laws that pick out people on the bases of race for differential treatment, therefore, presumptively violate the principle of equality (unless, of course, they are designed to remedy racial disadvantage, through affirmative action programmes, for example). 

It is this richer and more substantive vision of equality and equal treatment that demonstrates the unconstitutionality of the CAA in starkest terms. Each of the three “conditions” under the CAA – country of origin, religion, and date of entry into India – are effectively beyond the control of the individuals the law is targeted at. A person cannot choose which country they were born in, which religious community they were born into, and when persecution forced them to flee into India. But the CAA takes the category of migrants living in India and divides them precisely on these three bases. This is why it goes against the basic tenets of equality. 

Citizenship Laws and Standards of Review 

Another common argument that is invoked by the defenders of the CAA is that issues of citizenship and migration are firmly within the domain of sovereign State powers, and the scope of judicial intervention is highly limited. Courts must – or so the argument goes – defer to the State’s decision regarding who will be granted citizenship, and how. This argument has been repeated so often over the years, and so frequently, that it has by now acquired the immovable weight of a mountain. But the most cursory examination will reveal, however, that this mountain is made of straw. 

Let’s go back to the basics. What was the original justification of judicial review in a democratic society? What justified an unelected Court striking down laws passed by democratically-elected legislatures? The answer, of course, was that the primary role of the Court was that of a counter-majoritarian institution. It existed to check the excesses of majoritarianism, on the understanding that true democracy meant something more than brute majority rule. For this reason, in its famous Carolene Products footnote, the US Supreme Court noted that the role of the Court was particularly important in cases involving “discrete and insular minorities.” Why? Because it were these minorities that faced the greatest difficulties in articulating their interests through the normal channels of (majoritarian) democratic governance. The task of the Court, essentially, was to come to the rescue of those whom the political process – formally or effectively – excluded from equal participation. Thus, for instance, if there is a country where same-sex relations are viewed with opprobrium by a large segment of the population – to the extent that the LGBTQ+ community is permanently excluded from access to political power, as nobody else will ally with them – the Court is justified in subjecting laws targeting that community to stringent scrutiny. 

It should be obvious that migrants – or refugees, as the case may be – fall squarely within this category. As they cannot vote, they are formally excluded from participation in the political process. More than any other vulnerable or marginalised group in the country, they have no say in the laws and policies that will impact them. For this reason, laws that affect citizenship status in the manner that the CAA does, must be subjected to the highest threshold of judicial scrutiny, rather than the lowest. 

Harmonising Constitutional Principles: Sovereign Powers and Conditions of Entry

In a constitutional democracy, no power is absolute. Constitutional authorities are established by – and owe their existence to – the Constitution, and the powers they exercise flow from that same Constitution. In some cases, these powers are limited in express terms. For example, Article 13 of the Constitution expressly limits Parliament’s power of law-making by making it subject to the fundamental rights chapter. 

Article 11 of the Constitution – that deals with citizenship – contains no such express limitation. It gives to Parliament the right to “regulate citizenship by law”, and allows Parliament to make “any” provision with respect to acquisition and termination of citizenship, and “all other matters” relating to citizenship. Commentators have pointed to the width of these words to argue that in matters of citizenship, Parliament has virtually unlimited power (apart from the usual touchstone of the fundamental rights chapter).

What this argument ignores, however, is that express limitations are not the only manner in which constitutional authorities are constrained. As noted in Kesavananda Bharati, there also exist implied limitations that flow from the structure of the Constitution. When – and how – do we discern implied limitations? For the purposes of this post, a short answer will suffice: power under the Constitution to do “x” is limited at the point at which doing “x” will frustrate or destroy another, equally important constitutional principle. This principle was most recently reiterated by the UK Supreme Court in Miller v The Prime Minister, where the British Prime Minister’s power to “prorogue” Parliament was held to be limited by the constitutional principle of representative democracy, according to which it was Parliament’s function to scrutinise and debate important legislation. It was found that the Prime Minister’s prorogation – just before the deadline for Brexit – had the effect of denying Parliament an adequate opportunity to debate the proposed EU Withdrawal Bill, and was therefore unconstitutional. 

What is the implied limitation in the present case? The answer is the constitutional principle of secularism. Secularism – as Kesavananda Bharati held – is a basic feature of the Indian Constitution (independent of its subsequent insertion into the Preamble during Indira Gandhi’s Emergency). The Indian Constitution commits us to being a secular polity. The key issue, then, is that can the conditions of entry into the polity (determined by citizenship law) be such that they frustrate the character of the polity itself. The answer, obviously, is no. In other words, therefore, there is an implied limitation upon the power under Article 11 to grant or withdraw citizenship, that does not permit Parliament to pass any such law that would negate the secular character of the polity – in this case, through the backdoor, by creating conditions of entry where religious claims become determinants of citizenship. To put it in a single sentence: the principle of secularism acts as an implied limitation upon Parliament’s power to legislate on citizenship. Parliament, therefore, has all powers to prescribe conditions of citizenship except and insofar as such conditions frustrate the Constitutional commitment towards preserving a secular polity. 


Sterile debates over “reasonable classification”, “rational nexus”, and “sovereign powers” can only take us so far. More than that, they serve as conceptual prisons that stop us from thinking more deeply about the idea of equality, the link between equality and democracy, and what the Constitution really asks of us. In recent years, Indian constitutional jurisprudence has begun to liberate itself from that conceptual prison, and has articulated a richer vision of equality and democracy. The CAA challenge now gives the Supreme Court an opportunity to further develop – and evolve – that jurisprudence.

[Disclaimer: the author is involved in two of the petitions challenging the constitutionality of the CAA.]

Guest Post: The Wrath of CAAn – on Citizenship, Secularism, and Equality

[This is a guest post by Abhinav Chandrachud.]

In a paper posted on SSRN, I outline what I consider to be the key arguments for and against the Citizenship (Amendment) Act, 2019 (“CAA”).

Hidden Constitutional Premises:

To begin with, the paper interrogates the premise that the CAA violates the secular fabric of India’s Constitution and argues that there were certain hidden premises in the citizenship provisions of the Constitution that were not very secular at all. Two waves of immigration from West Pakistan to India occurred after 1 March 1947. In the first wave, Hindus and Sikhs came here in large numbers. In the second wave, Muslims who had left India for Pakistan tried to come back home. This created problems in India – the homes of Muslim “evacuees” were being used to rehabilitate Hindu and Sikh “displaced persons”. In July 1948, therefore, a permit system was introduced which made it very difficult for Muslim “evacuees” to return to India if they had property back home. Articles 6-7 of the Constitution indirectly entrenched this preference for Hindu/Sikh “displaced persons” and discriminated against Muslim “evacuees” who wanted to come back home. They did so by making it far easier for those who had come to India prior to 19 July 1948 (i.e., the date of the introduction of the permit system), presumed to be mostly Hindus and Sikhs, to become Indian citizens, while denying Muslim “evacuees” citizenship unless they had obtained an elusive permit for resettlement.

The CAA Is Discriminatory:

That does not mean, however, that the CAA is constitutionally valid. The conditions which existed between 1947-50 (viz., housing shortages, a mass exodus of people, a charged communal environment caused by an influx of refugees) no longer exist today. A persuasive argument can be made that the CAA is discriminatory for five reasons: (i) it leaves out other religious communities in the subject countries (e.g., Jews, atheists, agnostics, Shias, Ahmadiyas); (ii) it ignores other countries in India’s neighborhood (e.g., Sri Lanka, Bhutan, Myanmar); (iii) the cut-off date of 31 December 2014 is arbitrary (though bright line rules tend to be arbitrary, the CAA’s bright line rule undermines its own ostensibly humanitarian objective of protecting those who suffer religious persecution); (iv) it ignores non-religious persecution (e.g., persecution on grounds of race or sexual orientation); (v) the relaxation of the residence requirement from 11 years to 5 years is palpably arbitrary (why should a Parsi fleeing religious persecution from Iran have to reside in India for 11 years to seek citizenship by naturalization, while a Parsi fleeing religious persecution from Afghanistan has to wait only 5 years?)

Underinclusion, or the Chief Justice Patanjali Sastri Defence

However, proponents of the CAA might argue that Article 14 of the Constitution has its limits. In an early case, Chief Justice Patanjali Sastri held that classification brought about by the legislature need not be “scientifically perfect or logically complete”. Building on this doctrine, the Supreme Court has, over the years, held that a pragmatic (not doctrinaire) approach will be adopted in classification cases. The court will frown upon an overinclusive statute but not an underinclusive one (unless there is no fair reason to exclude those left out). The CAA is underinclusive. The state may argue that the categories it excludes are based on degrees of harm and that classification need not be scientifically perfect.

However, this argument is not very persuasive. The classification under the CAA must be tested in the light of the ostensible purpose of the statute. The aim of the statute, according to the Statement of Objects and Reasons, is to protect communities which “faced persecution on grounds of religion” in the subject countries. If that is so, the exclusion of the four categories mentioned above (i.e., other religious minorities in the subject countries; religious minorities in other countries; those who suffer religious persecution after the cut-off date; those who suffer non-religious persecution) lacks an “adequate determining principle” altogether, and therefore is not a simple case of “under-inclusion”; rather, it might fall foul of the “manifest arbitrariness” test.

The CAA is only part of the problem:

However, the CAA is only a part of the problem with Indian citizenship law. Amending the CAA to undo its shortcomings will still not save genuine Indian citizens from being disenfranchised under a National Register of Citizens (NRC). Some 19 lakh people in Assam have been left out of the NRC. Non-Muslims who are excluded from the NRC will only be able to seek refuge under the CAA if they can prove that they came to India before 31 December 2014 from one of the subject countries fleeing (or fearing) religious persecution, which may not be as easy as one might think (despite the “sympathy” provision in Section 8 of the Foreigners Act, 1946).

The paper identifies four problems with Indian citizenship laws: (i) the reverse burden of proof; (ii) the abandonment of citizenship by birth; (iii) the failure to grant a safe harbor to “dreamers”; and (iv) the procedural defects in the Foreigners Tribunals.

The burden of proof in citizenship cases is on the alleged foreigner. Relying on similar laws in countries like the U.S., the Supreme Court of India has upheld this evidentiary rule in Sonowal I, without realizing that the socio-economic condition of the citizens of countries like the U.S. is very different from that in India. An Indian citizen may not, on account of illiteracy, poverty or inadequate infrastructure, be able to prove his citizenship like an American citizen would.

After the 2004 amendment to the Citizenship Act, 1955, the principle of citizenship by birth in India has been abandoned. Now, anyone born in India after 1987 has to prove either than one or both parents are Indian citizens (in some cases, that one parent is not an illegal immigrant). Proving the place of birth of one’s parents is hard for anybody. However, this is especially harsh on some like orphans and transgender persons. How will orphans, who never knew their parents, be able to prove that their parents were Indian citizens? How will transgender persons (some of whom are abandoned at birth) be able to do so?

Indian citizenship law operates harshly on “dreamers” (the term “dreamer” is used in the U.S. to describe the child of an illegal immigrant who arrived illegally in the U.S. at a very young age). A child born in India to illegal immigrants has only known India to be his/her homeland since childhood. That child may now be an adult. Even so, is it not unfair to now deport that person to a country he/she has never known? Similarly, a child who accompanied his/her illegal immigrant parents to India has also known only India to be his/her homeland since childhood and the illegal immigration to India was not his/her fault. Why should such persons, even if they are now adults, be deported overseas?

The judges of Foreigners Tribunals lack security of tenure, without which there cannot be an independent judiciary. They hold office not during “good behaviour” (e.g., for a fixed, non-renewable term in office, or until a retirement age), but for short, renewable terms. Further, though “members” of the tribunal are supposed to have judicial experience, even retired civil servants have recently been accepted as members of these tribunals, contrary to the principles laid down by the Supreme Court. The fact that there are no statutory appeals against these tribunals accentuates these problems. Though a writ remedy is available against their “opinions”, High Courts will not enter into complicated questions of fact while entertaining writs.

Guest Post: The Constitutionality of the Citizenship (Amendment) Act – A Rejoinder

(This is a guest post by Varun Kannan.)

I have had the benefit of reading two illuminating pieces written on this blog by Nivedhitha K and Shivam Singhania, on the constitutionality of the Citizenship (Amendment) Act, 2019 (‘CAA’). While Nivedhitha and Shivam arrived at different conclusions with respect to the constitutionality of the CAA, they also expressed different opinions on whether Indu Malhotra J’s concurring opinion in Navtej Johar made an interpretative advance to the traditional classification-based test under Article 14, and whether a religion-based classification is permissible.

In this post, I aim to argue that Section 2 of the CAA fails to satisfy even the traditional classification test of Article 14, and that it can be struck down without any need to resort to the Navtej Johar decision, or the test of ‘manifest arbitrariness’. Another aspect of Section 2 of the CAA which has not received sufficient attention is the cut-off date of 31st December 2014. I shall also argue that prescription of this cut-off date in the statute is violative of Article 14.

The change in eligibility criteria that has been brought about by the CAA

Before proceeding with the analysis, let us reproduce the exact wording of Section 2 of the CAA, which seeks to amend Section 2(1)(b) of the Citizenship Act, 1955 (‘the parent Act’) by inserting the following proviso:

“Provided that any person belonging to Hindu, Sikh, Buddhist, Jain, Parsi or Christian community from Afghanistan, Bangladesh or Pakistan, who entered into India on or before the 31st day of December, 2014 and who has been exempted by the Central Government by or under clause (c) of sub-section (2) of section 3 of the Passport (Entry into India) Act, 1920 or from the application of the provisions of the Foreigners Act, 1946 or any rule or order made thereunder, shall not be treated as illegal migrant for the purposes of this Act;”


The proviso inserted makes three distinct classifications: (i) a religion-based classification i.e. the person must belong to either of the six communities referred to above, and must also be a victim of religious persecution; (ii) a country-based classification i.e the person should be from Afghanistan, Pakistan or Bangladesh, and should have suffered from religious persecution in either of these three counties ; and (iii) a classification based on the cut-off date i.e. whether the illegal immigrant has entered India on or before 31st December 2014. It is important to note here that the requirement of ‘religious persecution’ has not been mentioned anywhere in the text of the CAA. This requirement has been mentioned only in the Statement of Objects and Reasons of the CAA, and in certain notifications issued by the Ministry of Home Affairs in 2015 and 2016 – that the CAA refers back to – which have exempted the same category of illegal immigrants from the penal consequences of the Passport (Entry into India) Act, 1920 and the Foreigners Act, 1946. (The relevant 2015 notifications can be found here and here).

It is also instructive to refer to Section 3 and Section 5 of the CAA. Section 3 inserts Section 6B to the parent Act, and inter alia states that upon fulfilment of additional conditions that may be prescribed, individuals referred to in the newly inserted proviso shall be eligible to receive a certificate of naturalisation or registration. Section 5 amends the Third Schedule of the Parent Act to reduce the naturalization requirement from 11 years to 5 years, for individuals who fulfill the criteria prescribed by Section 2. This highlights that the CAA does not automatically grant citizenship to certain communities, and that it only changes the eligibility criteria for obtaining Indian citizenship for individuals from certain religious communities who have come to India from three neighboring countries. This change in the eligibility criteria which makes three distinct classifications referred to above forms the basis on which we shall analyze whether Section 2 of the CAA satisfies the traditional classification-based test of Article 14.

Does Section 2 of the CAA satisfy the traditional classification-based test under Article 14?

Under the traditional classification-based test of Article 14, as laid down in decisions such as Anwar Ali Sarkar a classification created through legislation must satisfy a two-pronged test – (i) it should be founded on an intelligible differentia, which distinguishes those that are grouped together from those that have been excluded; and (ii) the differentia must have a rational nexus with the object that the statute seeks to achieve. In Anwar Ali Sarkar, it was also held that the application of Article 14 cannot be precluded on the basis of whether there was clear evidence of the Legislature’s intention to discriminate, and held that the application of Article 14 is premised on whether discrimination is a necessary consequence of the statute.

The validity of the religion and country-based classification

Keeping this in mind, let us jointly analyze the religion-based and country-based classification created by Section 2 of the CAA. Section 2 changes the eligibility criteria for obtaining citizenship only for six communities (Hindu, Sikh, Buddhist, Jain, Parsi or Christian) who have come to India as illegal immigrants from Afghanistan, Pakistan or Bangladesh, and have been victims of religious persecution. The Statement of Objects and Reasons of the CAA, as mentioned in the Citizenship (Amendment) Bill, 2019 (‘CAB’) inter alia states that – “The constitutions of Pakistan, Afghanistan and Bangladesh provide for a specific state religion. As a result, many persons belonging to Hindu, Sikh, Buddhist, Jain, Parsi and Christian communities have faced persecution on grounds of religion in those countries. Some of them also have fears about such persecution in their day-to-day life where right to practice, profess and propagate their religion has been obstructed and restricted”.

However, Pakistan, Afghanistan and Bangladesh are not the only neighboring countries that have a specific state religion. There are two other neighboring countries i.e. Bhutan and Sri Lanka, which also have a specific state religion, which is Buddhism. There exists clear evidence that Tamils (who also subscribe to the Hindu faith) from Sri Lanka and Christians from Bhutan have been subjected to religious persecution. In fact, as Arun Janardhanan points out in the Indian Express, around 1 lakh Tamils from Sri Lanka who had fled from the State due to anti-Tamil riots continue to live in India as of today. Section 2 of the CAA fails to include such other persecuted communities.

In fact, while Hindus and Christians from Pakistan, Bangladesh and Afghanistan have been included, the Hindu Tamils from Sri Lanka and Christians from Bhutan have been excluded from the ambit of Section 2. Additionally, Section 2 of the CAA also does not include the Ahmadi community of Pakistan, who constitute a sect of Muslims who have been severely persecuted in Pakistan. The exclusion of three communities mentioned above shows that discrimination in terms of eligibility criteria for citizenship is a necessary consequence of Section 2 of the CAA.

Taking into account these exclusions, the Union Government may refer to the Statement of Objects and Reasons to contend that the classification is applicable only to those countries that have Islam as the specific state religion, and that the communities covered are only those communities in these three countries who do not believe in Islam. This may also be used to justify the exclusion of Ahmadi Muslims, to state that they are persecuted on sectarian grounds, as they do not subscribe to a certain version of Islam in Pakistan. However, such a classification is not based on an ‘intelligible’ differentia for three reasons, which shall be outlined below.

First, the classification is based on the premise that religious persecution takes place only in those neighboring countries that have Islam as the specific state religion. But, as discussed above, religious persecution also takes place in Sri Lanka and Bhutan, which have Buddhism as the State religion, and not Islam. The case of Tamils in Sri Lanka and Christians in Bhutan highlights that persecution takes place even in two other neighboring countries that do not subscribe to Islam.

Second, the classification is based on the presumption that only communities who are non-believers in the state religion are persecuted, and that only minority communities are persecuted in an Islamic country.  The persecution of Ahmadi Muslims in Pakistan for not subscribing to a certain version of Islam rebuts this presumption. As Ahmadis are persecuted for not subscribing to a certain version of Islam, their persecution does have a ‘religious’ angle to it. This renders the distinction between ‘religious’ and ‘sectarian’ persecution a matter of mere semantics, which cannot be used to argue that the classification is based on an intelligible differentia. In any case, the distinction between religious and sectarian persecution is applicable only to Ahmadis, and is inapplicable to the exclusion of Sri Lankan Tamils and Bhutanese Christians.

Third, the religion and country-based classification takes into account only religion-based persecution, and does not take into account certain other forms of persecutions, such as persecution that is political in nature. For instance, Tibetans are persecuted in China for political reasons, and even they stand excluded from the protective umbrella of the CAA. For these reasons, the religion and country-based classification is not just and sound, and cannot be said to be based on an ‘intelligible’ differentia.

Furthermore, instead of viewing this classification as a case of ‘under-inclusion’ by the Legislature, it is better understood as one that leads to discrimination as a necessary consequence. This is a case of unequal treatment of individual human beings within the territory of India, who are placed in a similar situation, and is not akin to other classic cases of ‘under-inclusion’ such as Chiranjit Lal Choudhary, where in a government take-over case, it was argued that other similarly placed companies were ‘under-included’ by the Legislature. The CAA accordingly leads to discrimination as a ‘necessary consequence’, as certain communities already in India as “illegal immigrants” from three neighboring countries are granted a different eligibility criteria for citizenship, to the exclusion of others (such as Sri Lankan Tamils) who are similarly placed. This also goes against the basic tenet of Article 14, which is that ‘equals should be treated equally’.

As the classification is not based on an intelligible differentia, it fails to satisfy the classification test of Article 14. However, for the sake of argument, let us assume that the Supreme Court finds an intelligible differentia, and analyze whether the second prong of the test is satisfied i.e. whether the differentia has a rational nexus with the purpose that the statute seeks to achieve. In decisions such as Subramanian Swamy v. CBI, the Supreme Court has added another prong to the rational nexus requirement, by holding that the purpose that the statute seeks to achieve should itself be legitimate. It was also held (in the context of Section 6-A of the Delhi Police Special Powers Act) that if the object of the statute is itself discriminatory, then that discrimination cannot be justified on the ground that there is a rational nexus with the object that the statute seeks to achieve.

As discussed above, the Union Government may present the object of Section 2 of the CAA narrowly, by contending that it seeks to provide benefit to persons belonging to non-Islamic communities, in the three neighboring countries that have Islam as the state religion. The Union Government may contend that the CAA changes the eligibility criteria for only those communities who have been victims of religious persecution, for belonging to a community that does not believe in the state religion i.e. Islam. While a rational nexus may be established with this object, I would submit that for the same reasons discussed above, the object of Section 2 of the CAA is in itself discriminatory. The reasons mentioned above can be summarized as follows:

i.         The religious persecution of Sri Lankan Tamils and Bhutanese Christians shows that religious persecution exists even in those countries that have a state religion apart from Islam, and that religious persecution is not limited solely to Islamic nations;

ii.         The persecution of Ahmadi Muslims in Pakistan highlights that even communities that subscribe to the state religion can be persecuted. As Ahmadi Muslims are persecuted for not subscribing to a certain version of Islam, their persecution also has a ‘religious’ angle, and cannot be categorized as persecution that is merely ‘sectarian’ in nature;

iii.         Persecution of communities is not restricted solely to religious grounds, as evidenced through the persecution of Tibetans in China, which is done on political grounds. The object of the statute hence includes one form of persecution to the exclusion of other forms of persecution which are equally significant.

All communities mentioned above have been excluded from the ambit of the CAA. Hence, the object of the religion and country-based classification created by Section 2 of the CAA is in itself discriminatory. Applying the test laid down in Subramanian Swamy, such discrimination cannot be justified on the ground that it has a rational nexus with the object that the statute seeks to achieve.

We may also make reference here to the Supreme Court decision in Deepak Sibal v. Punjab University. Here, it was held (in context of admission to educational institutions) that while the Legislature is entitled to prescribe the source from which admission to an educational institution is to be made, such a source (which selects the beneficiaries of the legislation) is also a ‘classification’. As it is a ‘classification’, it must also satisfy the traditional Article 14 test. This decision can also be applied to the classifications created by the CAA. The Union Government cannot successfully argue that the requirements of Article 14 are satisfied by defining the purpose of the CAA in terms of the selection of the beneficiaries i.e. in terms of illegal immigrants of non-Islamic communities, who have come to India from three neighboring Islamic nations. To illustrate further, a law which confers red-haired businessmen with a tax exemption cannot be defended on the ground that the object of the law is in itself to provide a benefit to solely red-haired individuals, as such an object would by itself be discriminatory.

For the reasons mentioned above, the religion and country-based classifications fail to satisfy either of the requirements of the traditional classification test.

The validity of the cut-off date-based classification

Section 2 of the CAA also prescribes that in order to take benefit of the amendment, the illegal immigrant should have entered into on or before 31st December 2014, which is the cut-off date. This creates a third tier of classification, between illegal immigrants (who belong to either of the six communities and have come to India from Pakistan, Afghanistan and Bangladesh) who have entered India before 31st December 2014, and those who have entered India after the cut-off date. This cut-off date was not present in the 2016 version of the CAB, and was inserted only in the 2019 version of the Bill, which is now the CAA.

As Arunabh Saikia argues, this cut-off date may have been put in place to assuage concerns of groups from the North-east regions, who expressed an apprehension that the absence of a cut-off date could lead to an endless influx of immigrants. The Union Government may also justify the cut-off date by contending that the CAA is a logical continuation of the 2015 and 2016 notifications referred to above, which exempted the same category of illegal immigrants referred to in the CAA from the penal consequences of the Passport (Entry into India) Act, 1920 and the Foreigners Act, 1946. (For ready reference, the relevant 2015 notifications are available here and here). These notifications also prescribed the same cut-off date of 31st December 2014.

This however cannot be a ground to show either an intelligible differentia or a rational nexus, as the Notifications can easily be amended through an executive order and brought in conformity with the date of coming into force of the CAA. Furthermore, there can be no intelligible distinction based on whether the same category of illegal immigrants entered India before or after 31st December 2014. Such a distinction is not just and sound, as there exists no parameter to justify the exclusion of the same category of illegal immigrants, solely on the ground that they entered India after the end of a particular date.

If a person has entered India after being a victim of religious persecution, he should technically be eligible for protection irrespective of the date on which he entered India. If the CAA would have prescribed that illegal immigrants who entered India on or before the date of commencement of the CAA would fall within its protective umbrella, then such a cut-off date would have some nexus with the goal of preventing a large scale influx of illegal immigrants. But, creating a classification through a cut-off date that is almost five years prior to the commencement of the CAA is not founded on any intelligible differentia. Such a differentia also does not have any rational nexus with the purported objective of providing protection to communities who have suffered from persecution in Islamic nations.


For the reasons mentioned above, all three classifications created by Section 2 of the CAA fail to satisfy the traditional test of Article 14. Hence, even without resorting to the subjective and more controversial test of ‘manifest arbitrariness’ or the Navtej Johar decision, Section 2 of the CAA, can be declared as unconstitutional.

The Constitutionality of the Citizenship (Amendment) Act: A Response

[This is a guest post by Shivam Singhania. The piece was written before the passage of the Bill into an Act, and the constitutional challenge to the Act. References to the “Bill”, therefore, may be understood as references to the Act.]

This piece is in response to the piece on this blog titled “The Citizenship (Amendment) Bill is Unconstitutional”. Respectfully disagreeing with the author, I shall endeavour to address the arguments against its constitutionality, and also chart out a path within the bounds of the settled judicial precedent on Article 14.

The bill changes amends the Citizenship Act on two counts – by inserting a proviso in Section 2(1)(b), and by amending clause (d) of the Third Schedule. The first count, i.e., the proviso, provides that ‘any person belonging to Hindu, Sikh, Buddhist, Jain, Parsi or Christian community from Afghanistan, Bangladesh or Pakistan’ will not be treated as an illegal immigrant under the Act, subject to three conditions: (a) he/she entered India on or before December 31, 2014, (b) is exempted by the Central Government under Section 3(2)(b) of the Passport (Entry into India) Act, 1920, (c) is exempted from the application of Foreigners Act, 1946.

(Note: The Passport Act empowers the Central Government to make rules requiring persons entering India to possess passports. It also gives the Central Government power to exempt, conditionally or unconditionally, any person or class of persons from complying with such requirements. The Foreigners Act empowers the Central Government to make orders prohibiting, regulating, restricting entry of non-citizens, i.e., foreigners into India. The Central Government also has the power to exempt any individual or class or description of foreigner from the application of the act. The exemptions for Bangladeshi and Pakistani nationals were issued in 2015, and included Afghanistan nationals in 2016. These exemptions have not been challenged. Further, the same classification exists for the purpose of issuing Long Term Visas (LTV) to nationals of these three countries. The same has also not been challenged.)

The second count, i.e., amendment to the Third Schedule, relaxes the time-criteria for naturalization from at least 11 out of 14 years to at least 5 out of 14 years of residing in India, or being in service of the Government, for ‘any person belonging to Hindu, Sikh, Buddhist, Jain, Parsi or Christian community from Afghanistan, Bangladesh or Pakistan’.

Admittedly, the constitutional questions rest on the issue of discrimination and non-adherence to ‘equal protection of law’.


The Constitutional position, right from Anwar Ali Sarkar to a catena of subsequent cases, is clear that the equality principle envisages like to be treated alike, and consequently allows unlike to be treated differently. Such differentiation, however, has to stand the test of intelligible differentia, i.e. that the basis or principle guiding the creation of different groups meant to be treated differently by the law should be ascertainable and sound. Further, such differentia should have a rational nexus to the object of the law, meaning that t should directly further the purpose that the law seeks to fulfil. Subsequently, the test of arbitrariness was also added in E.P. Royappa.

The classification in this amendment, as rightly described by the author, is on two counts: first – religion of the target group, i.e., Hindus, Sikhs, Buddhists, Jains, Parsis and Christians; and second – country of origin of these groups, i.e., Pakistan, Bangladesh and Afghanistan.

It will be shown in the course of this essay that the principles that ground the above classifications are intelligible and not under-inclusive and serve the object of the amendment and overall, are not arbitrary.


Over and above the generality of equality principle laid out in Article 14, Article 15 provides for prohibition of discrimination on specific grounds, such as race, sex, and also religion. This means that any such basis of classification, even if intelligible, cannot pass muster of Article 14 for reason of the specific prohibition in Article 15.

However, Article 15 applies only to citizens. On the other hand Article 14 applies to ‘any person’. Thus, as settled in Chandrima Das and Indo-China Steam Navigation cases, a non-citizen cannot take the benefit of Article 15. Thus, with respect to non-citizens, a classification which has religious inklings cannot be dismissed at the threshold; it will pass muster if it passes the tests of Article 14 discussed above.

We now address the author’s argument based on the concurring judgement of Justice Malhotra in Navtej Johar v Union of India. The author’s inference from certain portions of Justice Malhotra’s opinion Navtej Johar is misplaced. The quoted paragraph – “Race, caste, sex, and place of birth are aspects over which a person has no control, ergo they are immutable. On the other hand, religion is a fundamental choice of a person. Discrimination based on any of these grounds would undermine an individual’s personal autonomy” – lays out the underlying constitutional intent of Article 15. It does not, and is not meant to address the point of whether such grounds transcend Article 15, and apply as a bar upon classification ipso facto, including for non-citizens under Article 14

The following exposition from the judgment, admittedly, is unequivocal: “Where a legislation discriminates on the basis of an intrinsic and core trait of an individual, it cannot form a reasonable classification based on an intelligible differentia”. However, what begs an answer, and none is provided in clear terms by any judicial precedent, is whether certain grounds, beyond those in Article 15 and thus applicable to non-citizens, are barred at the threshold as the bases of any classification; if yes, then what are these grounds and do ‘intrinsic and core trait(s) of an individual’ qualify as such grounds?

Essentially, the submission of the author, relying on Justice Malhotra’s opinion, is that any classification with semblance to religion is impermissible, since like Article 15, which applies only for citizens, Article 14 contains an inherent bar upon classification on certain grounds, one of which is the notion of core trait/choice/personal autonomy (and religion is a part of that). This, however, has little basis in existing equality jurisprudence. The tests are clear and any classification can exist so long as the tests of Article 14 are met. Beyond that, no inherent barriers have been created for classification under Article 14.

In any case, the proposition that classification rests solely on religion should be taken with scepticism. In the CAA, religion is qualified by persecution. Therefore the two should be considered together. Religion persecution has also been recognized in the Refugee Convention as a form of persecution. Therefore, it cannot be considered as a capricious standard of marking the lines, given that the exact target groups are sought to be protected as recognized by International Convention. Further, the classification is not to create new benefits, but is only to accelerate an existing process as a measure of protection for recognizable and internationally accepted target group that, due to their precarious situation, need such protection. It is akin to any law which may be framed for instance to protect LGBT individuals from countries which provide for capital punishment for them. In that case, it will be a constitutional disservice to accept an argument that they cannot be allowed a dignified life in India because the law classifies on the basis of sex/sexual orientation.


A classic fallacy with the arguments on under-inclusiveness, also reflected in the author’s piece, is that its proponents enlarge either the principle determining classification, or the object of the law beyond their actual bounds, to argue that the classification is narrower than it ought to be.

It will be shown that the author’s purported principles determining the classification is deliberately broad, and the actual principle is more qualified. Therefore, when the purpose is defined properly, the classification is not under-inclusive.

The author, in her piece, considers four principles and finds them inadequate. For convenience, they are reproduced below –

  1. “Principle 1: Pakistan and Bangladesh were part of British India. Illegal immigrants from there could still generically be considered of Indian origin. However, with the inclusion of Afghanistan, it is evident that the classification is not based on the principle of divided India and undivided India.
  2. Principle 2: Afghanistan, Pakistan and Bangladesh have a State religion. However, the classification cannot be on the basis of a State religion, as Sri Lanka prescribes Buddhism as the State religion.
  3. Principle 3: Degrees of harm. In Chiranjit Lal Chowdhury it was held that the legislature is free to recognize the degrees of harm and confine the classification to where harm is the clearest. However, if the CAB is based on the degrees of harm then the Rohingyas of Myanmar ought to be included as the 2013 UN report states that the Rohingyas are the most persecuted in the world.
  4. Principle 4: The classification might be limited to singling out persecuted religious minoritiesHowever, on this logic, Sri Lankan Eelam Tamils must also be included, as the Tamil Eelams are persecuted based on religion (Hinduism) and ethnicity.”

The classic problem of “inadequate principle” is at issue here. The above four principles, if applied, result in the law leaving out some group who ought to qualify. However, it is quite possible that the alleged principles are not actually the principles on which the classification is based.

In this context, let us begin by noting that laying down adequate determining principle for classification in a law lies at the doorstep of the Parliament.

It is submitted that here, the determining principle is not solely based on Partition, or harm. It is also not only about state religion or secular states, or about religious minorities. Determining principles cannot be laid in water-tight compartments, and it is not necessary that the entirety of one ground has to make a determining principle (for instance “all minorities”, or “all persecuted”). More than one ground can qualify the other grounds if the determining principle is specific (for instance, not all minorities but religious minorities, or not all persecuted but persecuted for being a religious minority).

While one set of determining principles takes in the Partition (Pakistan and Bangladesh), and the unique situation of Afghanistan vis-a-vis India, another possible set of principles is not Partition-related. India sees its responsibility as a regional leader in its neighbourhood with a thriving democracy and secular credentials to protect certain classes of foreigners under threat in their own countries. Among many forms of persecution, for now, India considers itself sufficiently prepared to be able to imbibe as citizens through a relaxed process of naturalization, those persecuted on the basis of religion. The next question is, who all and from where all?

The principle is members of communities (religion) not being the state religion of the respective nation. Essentially, it is a loose combination of Principle 2 and 4 described by the author. In other words, the broad sphere of ‘religious minority’ is qualified by its application only with respect to nations with state religions and on minorities being non-believers of the state religion. This is clearer with the altered text of bill, as being circulated in the media, by which the phrase “…persons belonging to minority communities, namely, Hindu, Sikh…” has been replaced with “persons belonging to Hindu, Sikh…”. [Editor’s Note: this is now the text of the Act.]

Therefore, it can no more be argued that the genus was ‘minority community’, and that only six of them cannot be protected (i.e., under-inclusion). The principle is shortened from ‘minority community’ to a principle which is qualified by two other components – applicable to states with a state religion, and applicable for religions not being the state religion. In other words, the principle is narrower than presented. Hence non-inclusion of some is not because of under-inclusion, but because they are not in conformity with the principle for classification. (such as Ismailis or Shias, being another sect of Islam and not a separate religion than Islam), or Myanmar, not included for not having a state religion.

Pakistan, Afghanistan and Bangladesh have a state religion, i.e., Islam. The Sri Lankan Constitution gives Buddhism the foremost place and the Myanmarese Constitution only recognizes the special position of Buddhism as the faith practiced by majority. In any case, if it is still considered that Sri Lanka has a state religion, courts may also take cognizance of data, if provided by the Union of India, of the alleged dwindling rate of the minority population in Pakistan, Afghanistan (less than 1% presently) and Bangladesh (from 23% at independence to 9% presently) of the communities sought to be protected vis-à-vis Sri Lanka (28 % in 1953 to 20% presently) to gather the mischief that the amendment seeks to arrest.


The classification under Article 14 does not have inherent barriers of impermissibility at the threshold level, and nor does it have to conform to Article 15 when invoked with respect to non-citizens. In this case, the only test needs to be satisfied is the intelligible differentia-rational nexus-non-arbitrariness test of Article 14. The principle underlying the classification is not merely minority or religious minority, but of belonging to group other than the state religion of Pakistan, Bangladesh and Afghanistan (i.e. states with a constitutionally prescribed state religion). The classification with respect to states with state religions, for the benefit of persons not following the state religion, bears a rational nexus with the object of the amendment in protecting such persons facing threat of persecution for reason of non-conformity with their state religion and being eligible for relaxed criteria of immigration and naturalisation in India, and in furtherance of India’s responsibility as the leader in its nation with strong democratic values and emphasis on dignity of life to support victims of religious persecution in theocratic states. The classification, having semblance to a category of vulnerability recognized in international convention and supported empirically (possibly) is therefore not arbitrary.

Guest Post: The Citizenship (Amendment) Bill is Unconstitutional

[This is a guest post by Nivedhitha K. The piece was written before the latest draft of the Bill – which exempts certain North-Eastern states from the operation of the Amendment – was made available.]

The BJP-led Union government is determined to table the Citizenship (Amendment) Bill, 2019 in this parliamentary session, after it lapsed earlier in the year. The bill proposes to amend Section 2 (b) of the Citizenship Act – which defines ‘illegal immigrants’ – by excluding “persons belonging to minority communities, namely, Hindus, Sikhs, Buddhists, Jains, Parsis and Christians from Afghanistan, Bangladesh and Pakistan” from the ambit of “illegal immigrants”. The bill also reduces the period of residence in India for the acquisition of Indian citizenship through naturalization to six years from the earlier period of twelve years. Thus, under the amendment, these ‘minority communities’ from Afghanistan, Bangladesh, and Pakistan are no longer ‘illegal’ immigrants, and they can obtain Indian citizenship through naturalization if they have resided in India for six years.

The amendment makes two classifications: first, a classification based on religion by excluding Hindus, Sikhs, Buddhists, Jains, Parsis, and Christians from the ambit of illegal immigrants; and secondly, a classification based on country, by restricting the benefit of acquiring citizenship through naturalization to minority immigrants only from Afghanistan, Pakistan and Bangladesh.

Religion-based classification

The first argument for the unconstitutionality of the CAB is that it enacts an impermissible, religious-based classification under Article 14 of the Constitution. To understand why, we first need to look at Article 15. Article 15 provides for specific grounds – religion, race, caste, sex, place of birth – that cannot form the basis for discrimination. Thus, the grounds in Article 15 indicate impermissible discrimination, i.e the law should effect discrimination based on these grounds to be violative of Article 15. In Navtej, Chandrachud J and Indu Malhotra J did not declare section 377 of the IPC as violative of Article 15 only because it effected classification based on “sex”. Rather the test used was whether the classification was based on Article 15 grounds, and if so whether it effected discrimination. Therefore, for the violation of Article 15, both classification based on the grounds and discrimination in effect will have to be proved. However, the  Citizenship Amendment Bill cannot be subjected to Article 15, because that Article is only applicable to citizens (immigrants, by definitions, are not citizens).

This takes us to Article 14. The traditional test for an Article 14 violation requires the classification to have an intelligible differentia and a reasonable nexus with the legislative object. In Anwar Ali Sarkar, it was explained that intelligible differentia means that there must be a yardstick to differentiate between those included in, and excluded from the group. However, in Navtej, an important interpretive advance was made upon this. Indu Malhotra J, in her judgment, infused Article 15 grounds into Article 14. She interpreted intelligible differentia to mean reasonable differentia. She required the intelligible differentia test to fulfil two sub-tests: one there must be a yardstick to differentiate between those included in and excluded from the group, and two, that yardstick must itself be reasonable. She observed: “Where a legislation discriminates on the basis of an intrinsic and core trait of an individual, it cannot form a reasonable classification based on an intelligible differentia”. Therefore, the yardstick is reasonable only if it is not based on an intrinsic and core trait of an individual. She then referred to Article 15 grounds to explain that they form an intrinsic and core trait of an individual. She noted: “Race, caste, sex, and place of birth are aspects over which a person has no control, ergo they are immutable. On the other hand, religion is a fundamental choice of a person. Discrimination based on any of these grounds would undermine an individual’s personal autonomy.” Therefore, contrary to Article 15, the test for Article 14 violation is based on impermissible classification. Impermissible classification means that a particular principle cannot be used to classify because it is constitutionally irrelevant. Religion is a facet of personal autonomy, and a classification based on it is an impermissible classification.

The infusion of impermissible classification into the intelligible differentia test is justifiable for two reasons. First, while the objective of Article 15 is anti-discrimination, the objective of Article 14 is to provide equal protection of laws. Thus, Article 14 will be violated if the classification is unreasonable, while Article 15 will be violated only if there is discrimination based on the classification. Second, the interpretation of Article 14 on the lines of Article 15 aligns with the argument that fundamental rights are not watertight compartments. This understanding began with Fazl Ali J’s dissent in A.K Gopalan v. The State of Madras on the interpretation of the ‘procedure established by law’ clause. The interpretation was cemented through the evolution of the golden triangle in Maneka Gandhi v. Union of India, and strengthened by Chandrachud J’s interpretation of Article 26 in the Sabarimala case.  Therefore, classification based on religion is prima facie impermissible and violative of Article 14.

The second argument against religion-based classification in this case turns on its inability to fulfil Article 14’s nexus prong. The object of the amendment is to ‘protect those who have faced religious persecutions in Afghanistan, Pakistan, and Bangladesh’. However, by excluding Muslims from the category of ‘persecuted’, the amendment is based on the false premise that only minorities face religious persecution in a Muslim-majority country. The amendment makes an easy – but untrue – classification between minority and majority religion. This assumption is similar to assuming that all Hindus in India are treated alike irrespective of caste. However, within the majority Muslim religion, there is persecution based on sect. For example, in Pakistan, the Shias face religious persecution. Ahmadiyyas who align themselves with the Sunni school also face persecution. Unless persecution of a sect within the majority religion is recognised, the classification – of majority and minority – will have no nexus with the object of protecting those who face religious persecution.

Country based classification

The country based classification violates Article 14 as it fails the “manifest arbitrariness” test. Nariman J in Shayara Bano noted, “manifest arbitrariness, therefore, must be something done by the legislature capriciously, irrationally and/or without adequate determining principle. Also, when something is done which is excessive and disproportionate, such legislation would be manifestly arbitrary.” The common threads that run through the three selected countries in the CAB are that they are Muslim-majority countries, and that they are India’s neighbours. The first rationale has been addressed above. If, however, the relevant countries have been selected because they are India’s neighbors, the exclusion of the other neighbors – such as Sri Lanka and Myanmar where people face religious persecution – must be justified. However, there is no justification because there is no adequate determining principle that guides the classification. The following points elucidate the absence of a determining principle for the classification.

  1. Principle 1: Pakistan and Bangladesh were part of British India. Illegal immigrants from there could still generically be considered of Indian origin. However, with the inclusion of Afghanistan, it is evident that the classification is not based on the principle of divided India and undivided India.
  2. Principle 2: Afghanistan, Pakistan and Bangladesh have a State religion. However, the classification cannot be on the basis of a State religion, as Sri Lanka prescribes Buddhism as the State religion.
  3. Principle 3: Degrees of harm. In Chiranjit Lal Chowdhury it was held that the legislature is free to recognize the degrees of harm and confine the classification to where harm is the clearest. However, if the CAB is based on the degrees of harm then the Rohingyas of Myanmar ought to be included as the 2013 UN report states that the Rohingyas are the most persecuted in the world.
  4. Principle 4: The classification might be limited to singling out persecuted religious minoritiesHowever, on this logic, Sri Lankan Eelam Tamils must also be included, as the Tamil Eelams are persecuted based on religion (Hinduism) and ethnicity.

Therefore, it is evident that the exclusion of the other neighboring countries where people face religious persecution is not justified, because the inclusion of these three countries is not based on any determining principle be it Indian origin, state religion, the degree of harm, or of persecuted minorities. Thus, the country based classification is violative of Article 14 as it suffers from the vice of manifest arbitrariness.

Under-inclusiveness and Deference

Having established the violation of Article 14, it is also necessary to address two counter arguments that arise. The first counter argument is that under-inclusiveness cannot render the law unconstitutional. The second counter argument is that the court would have to exercise deferential review – by showing deference to legislative wisdom – while deciding cases on citizenship, refugees etc.

Indian courts have permitted under-inclusive laws on grounds of administrative necessity and legislative experimentation (see State of Gujarat v. Ambika Mills). However, in N.P Basheer v. the State of Kerala, it was held that under-inclusiveness would withstand scrutiny under Article 14 only when it is ‘nominal’. However, the decision neither explains nor defines the word ‘nominal’. Nonetheless, there is ample evidence now for what cannot be considered “nominal”: for example, in the case of Sowmithri Vishnu, the constitutionality of the adultery provision was challenged. The court observed that Section 497 was under-inclusive but allowed deference to legislative wisdom. However, in Joseph Shine, the court performed a volte face and struck down section 497 because it violated the ‘personal autonomy’ of women. The importance of personal autonomy in Article 14 challenges is also evident from the Supreme Court’s decisions in Indian Young Lawyers (Sabarimala) and Navtej. Hence, if the differentiation violates personal autonomy – in this case, religion – under-inclusiveness is certainly not ‘nominal’.

What of the argument that deference to legislative wisdom is required on issues of citizenship, refugees and the like, as they fall under the realm of legislative policy? On the contrary, I suggest that this argument should be turned on its head: precisely because the issue concerns citizenship, deference is undesirable. It has already been argued above that the court should not allow deference to issues that concern personal autonomy. In Navtej, Indu Malhotra J and Dipak Mishra (writing for himself and A.M. Khanwilkar) referred to Article 15 grounds as illustrative of personal autonomy. They also accepted the analogous grounds argument, by holding Section 377 violative of Article 15 because it discriminates on sexual orientation which is a ground analogous to grounds specified textually in Article 15. Therefore, very little deference must be allowed to grounds under Article 15, and grounds that are analogous to Article 15 grounds, since they are based on personal autonomy. The argument now is that citizenship is a ground analogous to Article 15 grounds, since it is an important facet of personal autonomy. The Canadian Supreme Court in Andrew v. Law Society of British Columbia held that citizenship is a ground analogous to Section 15 of the Canadian Charter of Rights and Freedom – a provision which is similar to Article 15 of the COI – because it is a ‘personal characteristic’. It was observed that analogous grounds must be determined based on the “place of a group in the socio-political-legal fabric of the society.” Where a number of important rights are accrued on the basis of citizenship, non-citizens as a minority whose interests are brushed off would fall within the analogous category. Hence, the heightened level of scrutiny to a classification based on personal characteristics (grounds analogous to Article 15, and grounds in Article 15) should prevent the court from applying a deferential review.


Consolidating the arguments above, religion-based classification is violative of Article 14 of the Constitution because religion is a constitutionally impermissible ground for classification, and the classification does not have any nexus with the object of the bill. The country-based classification is manifestly arbitrary and violative of Article 14, as it is not based on any determining principle. The argument that under-inclusiveness cannot be a ground for unconstitutionality is countered by explaining that a law that is grounded on personal autonomy (i.e religion) is not merely a case of nominal under-inclusiveness. And the argument on application of deferential review to issues concerning citizenship and refugees is inapplicable, as citizenship falls within the ambit of analogous grounds, for which deference should not be allowed.

The Constitutional Challenge to S. 6A of the Citizenship Act (Assam Accord): A Primer

Tomorrow, a Constitution Bench of the Supreme Court will commence hearings in the constitutional challenge to Section 6A of the Citizenship Act. The case comes up for hearing as the result of a referral order under Article 145(3) of the Constitution, passed by a bench of two judges in Assam Sanmilita Mahasangha vs Union of Indiawho framed thirteen questions of law to be decided by a Constitution Bench.

Tomorrow’s hearing may be a brief one. The Bench has indicated that it is unlikely to hear the matter unless all counsel commit to finishing within seven working days. Given the scale and complexity of some of the questions (as we shall see), as well as the number of intervention applications that were allowed after the referral, this is unlikely. In light of the fact, however, that even if it is not heard at the present, it is likely to be taken up soon after the vacations (in July or August), I shall provide a brief primer to the case.


Migration has been a source of social and political conflict in the border-state of Assam at least the middle of the 19th century. During the framing of the citizenship provisions of the Constitution during the Constituent Assembly Debates, the representative from Assam highlighted issues pertaining to large-scale migration from Bengal, its impact upon the indigenous population and culture, and asked for specific constitutional provisions to deal with the issue. Ultimately, however, the Constitution contained only skeletal provisions on citizenship – in particular, to deal with the Partition – and left the issue to be addressed by Parliament. Article 5 of the Constitution incorporated the broad jus soli principle of citizenship, stipulating that all those who had their domicile in India at the time of the commencement of the Constitution, would be citizens if they were born here, if either of their parents were born here, or who had been ordinarily resident for not less than five years. Articles 6 and 7 were the Partition provisions, dealing with migrations to and from Pakistan, and fixing 19th July 1948 as the “cut-off date” for citizenship. And to clarify that these provisions were only dealing with the special situation created by the Partition, Article 11 contained an overriding clause authorising Parliament to legislate for citizenship. Parliament did so in 1955, with the Citizenship Act, and a special law for Assam titled the Immigrants (Expulsion from Assam) Act of 1950.

To regulate the entry of migrants into India, the colonial government had passed the Foreigners Act of 1946, which continued even after Independence. This Act conferred powers upon the government to prohibit entry of foreigners, among other things. In 1964, acting under the authority of the Act, the Government promulgated the Foreigners Tribunal Order. This Order authorised the Government to establish Tribunals to determine questions of nationality, in accordance with the provisions of the Foreigners Act. Consequently, the Constitution of India, the Citizenship Act of 1955, the Foreigners Act of 1946, and the Foreigners Tribunal Order of 1964 comprised a comprehensive statutory regime dealing with both substantive and procedural questions of citizenship and migration.

Meanwhile, issues of migration continued to cause conflict in Assam. Matters came to a head during the run-up to the Bangladesh War of 1971, where in fact a massive influx of refugees into India from (what was then) East Pakistan was cited as one of the reasons for India’s involvement in the war. The issues did not cease even after 1971, however, because it was perceived that many of “illegal immigrants” were being put on voting rolls by political parties attempting to create faithful constituencies. Ultimately, this led to a state-wide student movement called the Assam Agitation, which lasted six years, from 1979 to 1985. The movement was sometimes punctuated by violence, including the Nellie massacre of 1983. It was finally brought to a close in 1985, with the signing of the Assam Accord between the Government of India, and the leaders of the movement.

The Assam Accord, S. 6A of the Citizenship Act, and the IMDT Act 

The Assam Accord was effectively a political compromise between the government and the leaders of the Assam Agitation. While providing for two separate cut-off dates for regularisation of migrants (an issue we shall discuss in a moment), the Accord also contained provisions for the development of Assam, as well as obligating the Government to see that “the international border shall be made secure against future infiltration by erection of physical barriers like walls, barbed wire fencing and other obstacles at appropriate places.”

Section 6A of the Citizenship Act – introduced through an amendment in 1985 – was the legislative enactment of the legal part of the Assam Accord. Section 6A divided “illegal” immigrants of Indian origin (i.e., those whose parents or grandparents were born in undivided India) who came into Assam from Bangladesh into three groups: those who came into the state before 1966; those who came into the state between 1966 and 25th March, 1971 (the official date of the commencement of the Bangladesh War); and those who came into the state after 1971. The first group (pre-’66) was to be regularised. The second group (’66 – ’71) was to be taken off the electoral rolls, and regularised after ten years. The third group (’71-onwards) was to be detected and expelled in accordance with law.

Section 6A, therefore, was a special citizenship law for Assam, hammered out as a result of a political settlement. Meanwhile, two years before the Accord and S. 6A, the Parliament had also passed the Illegal Migrants (Determination by Tribunals Act) of 1983. This Act authorised the Government to set up Tribunals for the purposes of determining whether migrants were illegal. Under the Act, the Government framed the Illegal Migrant Rules of 1984. The Act and the Rules, taken together, made some departures from the procedure under the Foreigners Act and the Foreigners Tribunal Order: for example, the procedure for making a reference to the Tribunal was made more onerous, the burden of proof was shifted from the State to the individual, and so on.

Consequently, the statutory regime governing migration to Assam now became Section 6A of the Citizenship Act, read with the Illegal Migrants Act of 1983, and the Illegal Migrant Rules of 1984. While the Government defended this regime on the basis of protecting minorities, who were genuine citizens of India, from persecution they were also attacked as being too lax on illegal migration, and making it almost impossible to deport illegal migrants.

The Judgment in Sarbananda Sonowal vs Union of India

The Illegal Migrants Act and Rules were challenged before the Supreme Court in Sarbananda Sonowal vs Union of India. A three-judge bench of the Supreme Court held that the statutory regime, with its reversal of the burden of proof clause (placing the burden of proof upon the State rather than the alleged illegal migrant), and its procedural requirements of filing applications (“… accompanied by affidavits sworn by not less than two persons residing within the jurisdiction of the same police station in which the person referred to in the application is found, or residing, corroborating the averments made in the application.“), was insufficient to check the problem of illegal migration. Relying upon a 1998 report by the Governor of Assam, the Supreme Court held that there was a flood of Bangladeshi migrants into Assam, which the statutory regime had failed to check. This, the Court held, amounted to “external aggression” against the State of Assam, and under Article 355 of the Constitution, it was the duty of the Union to protect every state against external aggression. Holding the statutory regime of the Illegal Migrants Act and Illegal Migrants Rules to be directly responsible for this failure, the Court held the Act and Rules to be unconstitutional.

After Sarbananda Sonowal, therefore, the Tribunals under the IMDT ceased to function, and the statutory regime reverted to Section 6A of the Citizenship Act, and the Foreigners Act and the Foreigners Tribunal Order. The State’s attempt to get around this through passing the Foreigners Tribunal (for Assam) Order of 2006 was also struck down by the Court in Sarbananda Sonowal (II)

The Referral Order

It is in this context that the challenge to Section 6A of the Citizenship Act came before the Supreme Court. In his referral order, Justice Nariman framed thirteen questions of law:

“(i) Whether Articles 10 and 11 of the Constitution of India permit the enactment of Section 6A of the Citizenship Act in as much as Section 6A, in prescribing a cut-off date different from the cut-off date prescribed in Article 6, can do so without a “variation” of Article 6 itself; regard, in particular, being had to the phraseology of Article 4 (2) read with Article 368 (1)?

(ii) Whether Section 6A violates Articles 325 and 326 of the Constitution of India in that it has diluted the political rights of the citizens of the State of Assam;

(iii) What is the scope of the fundamental right contained in Article 29(1)? Is the fundamental right absolute in its terms? In particular, what is the meaning of the expression “culture” and the expression “conserve”? Whether Section 6A violates Article 29(1)?

(iv) Whether Section 6A violates Article 355? What is the true interpretation of Article 355 of the Constitution? Would an influx of illegal migrants into a State of India constitute “external aggression” and/or “internal disturbance”? Does the expression “State” occurring in this Article refer only to a territorial region or does it also include the people living in the State, which would include their culture and identity?

(v) Whether Section 6A violates Article 14 in that, it singles out Assam from other border States (which comprise a distinct class) and discriminates against it. Also whether there is no rational basis for having a separate cut-off date for regularizing illegal migrants who enter Assam as opposed to the rest of the country; and

(vi) Whether Section 6A violates Article 21 in that the lives and personal liberty of the citizens of Assam have been affected adversely by the massive influx of illegal migrants from Bangladesh.

(vii) Whether delay is a factor that can be taken into account in moulding relief under a petition filed under Article 32 of the Constitution?

(viii) Whether, after a large number of migrants from East Pakistan have enjoyed rights as Citizens of India for over 40 years, any relief can be given in the petitions filed in the present cases?

(ix) Whether section 6A violates the basic premise of the Constitution and the Citizenship Act in that it permits Citizens who have allegedly not lost their Citizenship of East Pakistan to become deemed Citizens of India, thereby conferring dual Citizenship to such persons?

(x) Whether section 6A violates the fundamental basis of section 5 (1) proviso and section 5 (2) of the Citizenship Act (as it stood in 1985) in that it permits a class of migrants to become deemed Citizens of India without any reciprocity from Bangladesh and without taking the oath of allegiance to the Indian Constitution? 

(xi) Whether the Immigrants (Expulsion from Assam) Act, 1950 being a special enactment qua immigrants into Assam, alone can apply to migrants from East Pakistan/Bangladesh to the exclusion of the general Foreigners Act and the Foreigners (Tribunals) Order, 1964 made thereunder?

(xii) Whether Section 6A violates the Rule of Law in that it gives way to political expediency and not to Government according to law?

(xiii) Whether Section 6A violates fundamental rights in that no mechanism is provided to determine which persons are ordinarily resident in Assam since the dates of their entry into Assam, thus granting deemed citizenship to such persons arbitrarily?”

As we can see, these referral questions raise a host of complex issues about the interaction between the State’s sovereign power of conferring citizenship, the right to equal treatment, and the right to preservation of culture and identity; the interaction between rule of law and citizenship provisions arising as a result of political settlements; and the impact of a possible judgment of unconstitutionality upon vested rights that have stood for decades.

The Aftermath

After the referral order, some further petitions were filed, that were tagged with the main case. These included a petition asking that Section 3 of the Citizenship Act be read in a manner that children of illegal immigrants, when it came to Assam, ought not to be granted citizenship, on the basis that Section 6A was a comprehensive provision dealing with the issue of migration and citizenship in Assam. Petitions were also filed challenging the Foreigners (Amendment) Order of 2015 and the Passport (Entry into India) Amendment Rules, 2015, which stated that “persons belonging to minority communities in Bangladesh and Pakistan, namely, Hindus, Sikhs, Buddhists, Jains, Parsis and Christians who were compelled to seek shelter in India due to religious persecution or fear of religious persecution and entered into India on or before the 31 st December, 2014” would be granted exemption from application of the Foreigners Act and the Passport Rules. It is unclear whether the Court will take up these additional issues for hearing as well.

Either way, we shall know more about the progress of this case tomorrow.

(Disclosure: The writer is assisting the Respondents (AASU) in defending the constitutionality of S. 6A)