EQUALITY

legislative review under Article 14

TARUNABH KHAITAN*

I. INTRODUCTION
THIS chapter provides an analytical overview and limited evaluation of the grounds of legislative
review under the general constitutional guarantee of the right to equality under Article 14. It excludes
legislative review for other reasons (such as for violation of other constitutional rights or because of
lack of legislative competence). Also outside its immediate focus are the (woefully underdeveloped)1
anti-discrimination provisions in Articles 15(1), 16(2), and 29(2)—they are referred to only when
they illuminate the content of Article 14. Other chapters in this Handbook deal with administrative
review, affirmative action, and gender equality.
Within the context of Article 14, my focus will only be on legislative review. What this entails is
not quite as straightforward as might first seem. To begin with, I use ‘legislative review’ to mean
review of legislative acts by judges, rather than review by the legislature. In its adjectival form,
‘legislative’ could characterise at least three distinct things. First, it may be a qualifier that describes
an action based on who the actor is—anything that is done by the legislature is, in this sense,
legislative (actor sensitive). This will include not only primary statutes but also non-statutory actions
of a legislature, such as the adoption of resolutions and procedural rules, or orders punishing
someone for its contempt. It will not include executive ordinances.
In the second sense, an action can be described as legislative depending on how it was done
(process sensitive). Legislation is a particular mode of decision making, which usually includes
consultation, review in standing committees, debate and deliberation, and voting.2 Often, legislative
decisions have to be made democratically—that is, by democratically elected representatives through
an open, fair, and democratic process. There may, however, be occasions when a statute is enacted by
a legislature without any consultation, committee review, debate, or deliberation,3 and therefore may
not be described as adequately legislative in this second sense. Colonial statutes are less ‘legislative’
than post-Independence statutes, and (because of inadequate media and civil society oversight) State
statutes tend to be less legislative in this process-sensitive sense than Central ones.4 Ordinances
issued by the political executive suffer from this lacuna even more deeply. Thus, one could ask not
only whether an act is legislative, but also the extent to which it is so. Whether an action has been
legislative in the procedural sense is likely to be relevant to the standard of review applied, rather
than constitute a ground of review.
In the third sense, the adjective describes actions based on what they entail (product sensitive). In
this sense, what makes an act legislative is its substance. A legislative act comprises general rules
enacted to guide specified conduct.5 In this sense, it may be distinguished from rule application,
which is an essentially administrative exercise looking at particular cases. ‘Legislation’ then is
enacted not just by the legislature, but also by the executive (especially in the form of secondary rules
using authority delegated by statutes, but also as ordinances and policy documents). Even judges
legislate (less controversially) when they determine the procedural rules that ought to govern their
own functioning.6
These three senses have respectively been characterised as actor-sensitive, process-sensitive, and
product-sensitive understandings of the term ‘legislative’. Delineating these three senses of the term
is helpful for the discussion that is to follow. My focus in this chapter remains on the grounds of
legislative review, particularly in the actor-sensitive and product-sensitive sense. However, a full
story of legislative review under Article 14 will need to consider the standards of review debate as
well.7
This chapter explores the two doctrines that have evolved to test the constitutionality of a measure
when faced with an Article 14 challenge: the ‘classification test’ or the ‘old doctrine’ (which I have
labelled ‘unreasonable comparison’) and the ‘arbitrariness test’ or the ‘new doctrine’ (labelled ‘non-
comparative unreasonableness’). I will show that (a) the classification test (or the unreasonable
comparison test) continues to be applied for testing the constitutionality of classificatory rules
(whether or not legislative in character); (b) it is a limited and highly formalistic test applied
deferentially; (c) the arbitrariness test is really a test of unreasonableness of measures which do not
entail comparison (hence labelled non-comparative unreasonableness); (d) its supposed connection
with the right to equality is based on a conceptual misunderstanding of the requirements of the rule of
law; and (e) courts are unlikely to apply it to legislative review (in the actor-sensitive sense). The
way forward is to beef up the classification doctrine to realise its true potential, and abandon the
arbitrariness doctrine with respect to actor-sensitive legislative review.
Writing about Indian constitutional law poses a difficult methodological problem—a staggering
judicial output, contradictory and precedent-blind decisions in constitutionally significant cases by
benches of two or three judges, and under-reasoned judgments make the scholastic task very
difficult.8 In this chapter, I focus primarily on Article 14 cases decided by a bench of five or more
judges. These cases constitute a more manageable dataset, tend to give more attention to doctrine and
precedent, and are constitutionally recognised as being more authoritative than cases decided by
benches of lesser strength.9 This focus is but a guideline—it is impossible to ignore some other
significant cases.

II. THE TRADITIONAL NARRATIVE

Traditionally, many commentators and judges have told a story about the transformation of Article 14
along these lines: in its early years, inspired by the US jurisprudence under the Fourteenth
Amendment, the Supreme Court—led primarily by Das J—devised the classification test to determine
a law’s compliance with Article 14.10 Effectively, this entailed the Court asking two questions: (i)
whether the classification made by the law in question was based on an intelligible differentia; and
(ii) whether the classification had a reasonable nexus with the object the law sought to achieve. The
obvious assumption behind asking these questions was that the right to equality under Article 14 is
engaged only when the law makes a classification (or, arguably, when it fails to make a
classification). The essence of the right to equality was assumed to be comparative. The next stage in
the story involves the characterisation of this inquiry as ‘old’, ‘doctrinaire’, ‘positivist’, ‘narrow’,
etc—an image of a pedantic, unreconstructed, slightly reactionary, and decidedly unfashionable old
man is presented. The traditional narrative then projects the elevation of the administrative law
standard11 of ‘non-arbitrariness’ to a distinct ground of constitutional review under Article 14 in the
1970s as a fresh, progressive, development that unfettered Article 14 from its older doctrinaire
confinement.12 There have, of course, been discordant voices, which have challenged this dominant
narrative forcefully. Even so, the story has come to acquire significant currency, not least because of
its frequent parroting from the bench.13 And yet, the old doctrine refuses to go away.14
The reason for its stubborn persistence is down to the fact that the old doctrine is conceptually
distinct from the new doctrine, and cannot be subsumed simply as an instance of the latter. The
following subsections will explain that the main analytical difference between the two doctrines lies
in the fact that the classification doctrine is essentially comparative, whereas the arbitrariness
doctrine is not essentially so. What this means is that there must be some comparatively differential
treatment between two persons or two classes of persons before the classification doctrine is even
engaged. The arbitrariness doctrine, on the other hand, can be invoked for any sufficiently serious
failure to base an action on good reasons. To put it differently, the classification doctrine interrogates
unreasonable comparisons, whereas the arbitrariness doctrine’s unique contribution is to bring non-
comparative unreasonableness within the ambit of Article 14. Let us now understand each of these
doctrines better, in the context of legislative review.

III. UNREASONABLE COMPARISON

The classification doctrine—meant to attack unreasonable comparison between persons or classes of
persons—is founded upon a highly deferential, remarkably limited, and formalistic equal treatment
principle governing State action. Each of these features is worth emphasising. First, it is based on an
equal treatment principle governing State action. Only a State action that either classifies persons or
fails to classify them can be attacked under this doctrine. The ‘failure-to-classify’ aspect of the
doctrine should not be read to include all omissions. Only when the State fails to classify when
performing some act is the doctrine attracted. In the absence of any relevant State action, no
treatment has been meted out to anyone. It is an action-regarding doctrine, which makes essential
reference to something done. Note that this comment applies only to the classification doctrine. The
alternative, a non-action-regarding equal situation principle, which could have tackled an unequal
state of affairs rather than merely unequal treatment, plays no role in the way the old doctrine
operates. It may well be that the right to equality and non-discrimination guaranteed by the
Constitution (including the provisions on affirmative action) taken as a whole does entail both equal
treatment and equal situation principles. But as far as the classification doctrine is concerned, it only
incorporates an equal treatment principle.
Secondly, the doctrine relates to State action alone, primarily governing vertical relationships
between an individual and the State (although the relationship between two State bodies is technically
within its ambit). What remain outside its scope are horizontal relationships between private persons.
Due to the historically extensive State ownership and funding of industry, the Court has been able to
apply the mandate of Article 14 to State owned or funded airlines, universities, and banks.15 As the
Indian State continues to liberalise and roll back, the number of institutions governed by Article 14
will continue to decrease. At first blush, the text of Article 14—‘The State shall not deny …’—seems
to invite this vertical interpretation. However, the second part of the clause requiring the State to not
deny ‘equal protection of the laws’ could easily have been interpreted to impose some sort of duty on
the State to protect persons from horizontal violations of the right to equality, at least in the limited
way recognised by the US Supreme Court in the landmark case of Shelley v Kraemer. In that case, the
US Court held that the judiciary, as a constituent part of the State, cannot enforce racist covenants in
private housing contracts prohibiting the sale of a property to blacks.16 In a comparable case, the
Indian Supreme Court came to the opposite conclusion, and in the absence of a law prohibiting
discrimination in the housing sector upheld a discriminatory private arrangement as constitutionally
valid.17 What’s more, even the more explicitly horizontally worded provisions such as Article 15(2),
which the founders unambiguously intended to have horizontal effect, have received little judicial
notice.18
Thirdly, the classification doctrine entails an equal treatment principle. We cannot confidently say
the same for the arbitrariness doctrine. Like other fundamental values such as liberty, fraternity, and
dignity, the meaning and scope of equality is contested. Many different conceptions of equality—
formal, substantive, results oriented, opportunity focused, treatment based—have been proposed.
What is common to all these conceptions is that equality entails comparison.19 One can only be equal
to someone or something, and one can meaningfully talk only of equality between at least two entities.
It is conceptually sensible to talk of a person as being free or dignified, without invoking any
relationship that she may have with another. It is nonsensical, however, to speak of a person being
equal without referring to at least one other being. Fraternity is perhaps even more deeply relational,
inasmuch as it requires the existence of some form of a community. In requiring the presence or
absence of some form of classification, the classification doctrine is an egalitarian doctrine. To give
but one example, Mukherjea J found the mere existence of two different criminal procedures to attract
Article 14—the fact that both procedures constituted fair trial and would not have violated any
constitutional right independent of the right to equality was quite irrelevant to this determination.20
The conception of equality that informs the classification doctrine may be weak, formalistic, and
misguided, but it is a principle of equality.
Fourthly, this doctrine is highly deferential in the sense that the Court gives a lot of weight to the
State’s claim about what the facts are, how they ought to be evaluated, and whether they breach
certain norms. To give one extreme, but not entirely atypical, example, in Air India, the Court found a
government declaration to the effect that different wages for air hostesses and assistant flight pursers
was not based on sex as ‘presumptive proof’, which ‘completely conclude[d] the matter’.21 In some
recent cases, the Court seems to have invoked the ‘strict scrutiny’ or the ‘proportionality’ standard,
but there are good reasons to think that in many of these cases these invocations are merely rhetorical
cloaks for the continued use of a highly deferential standard of review.
Finally, and perhaps most significantly, I have claimed that the old doctrine is limited and highly
formalistic. The doctrine is limited because it asks only two of the whole range of questions it could
ask about a classificatory rule. We have seen that these questions are: (i) whether the classification
made by the rule is based on an intelligible differentia; and (ii) whether this differentia has a rational
nexus with the object that the rule seeks to achieve. The doctrine is formalistic because these two
questions are largely content with the prima facie formulation of the rule, and ignore its real-world
impact on persons and groups.
The apparent limitations and formality of the doctrine will become clear once we examine the
whole range of questions that could be asked—and that judges in other jurisdictions often do ask—
about a classifying rule. Before we proceed, however, it is important to note that despite its apparent
limitations and formality, many Indian judges have often misapplied the test to reach the right
outcome. In its current form, the test poses a difficult dilemma to a conscientious judge: if she
prioritises fidelity to law, unjust outcomes come about; on the other hand, if she prioritises fidelity to
justice between parties, she often has to distort and misapply the doctrine, incurring a legality cost.
The dilemma is, of course, not unique to Article 14 cases, and judges do occasionally (and sometimes
justifiably) fudge legal doctrine to arrive at the just conclusion.22 But when this happens all too
frequently, it is time that legal doctrine evolved. Bhagwati J was responding to these infirmities in the
old doctrine when he (mis)prescribed the arbitrariness doctrine as a cure.
But we are getting ahead of ourselves. Let us go back to the possible inquiries that can be
conducted with respect to a classificatory rule. The list that follows is not meant to suggest that all of
these questions ought necessarily to be asked by judges when testing the constitutionality of a
classificatory rule under Article 14. It is provided simply to demonstrate the questions that judges do
not currently ask (except, sometimes, by stealth) and thereby to demonstrate the limitations and
formality of the current doctrine.
It may also be noted that some of these questions go into the ground of review, others concern
setting the appropriate standard of review and the incidence of the burden of proof. Whether the
classification makes an intelligible differentia, for example, concerns the ground of review in the
sense that the answer to this question (along with the answers to other ground-related questions)
directly determines the constitutionality of the rule. A different question, say whether the differentia is
presumptively impermissible (perhaps because it is based on a ground specified in Article 15), may
primarily relate to how anxiously the Court ought to review the rule or how weighty the State’s
justificatory burden ought to be to save the rule.
For any classificatory rule, we can usually identify at least four distinct elements:

1. Right: is the right to equality engaged?

2. Differentia: what classes does the rule create?
3. Objective: what end does it seek to achieve?
4. Impact: what consequences does it subject each of these classes to?

To give an example, let us imagine a rule which says that ‘A contract of employment shall stand
terminated if the employee becomes pregnant.’23 First, the right to equality is clearly engaged because
of the differential comparative treatment of the two classes. The differentia created by this rule is
between those employees who become pregnant and those who don’t (or can’t). Its objective is not
evident from the rule itself, but let us assume that the admitted objective is to avoid the expense and
inconvenience involved in organising replacement cover for the pregnant employee for the period that
she needs to go on maternity leave. Finally, the immediate and direct impact of the rule is that the
employment of pregnant employees is terminated. This, however, only presents a relatively
superficial analysis of the classificatory rule. For each of these elements of the rule, further questions
may be asked.

1. The Right
With respect to the first element, it is clear that if the rule makes a classification, the right to equality
is engaged. But sometimes it is also engaged when, even if the rule does not make any classification
on the face of it, it has a disproportionate impact on different classes of persons (ie, the classification
is made not by the rule’s formal requirements but because of their operation in the real world). Courts
across the democratic world have recognised this as the principle prohibiting disparate impact or
indirect discrimination, although Indian courts have been slow to update the doctrine in this regard. In
our pregnancy example, the rule makes a direct classification on the basis of pregnancy. But it also
makes an indirect classification based on sex, since it is only women who can become pregnant and
will therefore be disproportionately affected by the rule. Another example of an indirectly
discriminatory law would be one that requires bike-riders to wear a close-fitting hard helmet. Such a
law, even though neutral on its face, will have a disproportionate impact on those Sikh men who
believe it is their religious duty to wear a turban. The law may yet be justified given its objectives—
the point simply is that its operation makes a classification between Sikh men and other people. We
will return to the issue of impact in due course.
Furthermore, with regard to the first element, we could ask not only whether the right to equality is
engaged but also whether any other right or value is engaged. This inquiry is possible because
equality is a parasitic right—its engagement often entails the engagement of another interest, which
may sometimes be an important right or value. The rule may not in fact ‘violate’ this other right or
value, the mere fact that it is engaged could be relevant—inequality with respect to a fundamental
right is surely a more serious matter than inequality with respect to a mundane interest. Let me
provide an example to make things clearer. In Anwar Ali Sarkar, the classifying rule dealt with
criminal trials. It was common ground that neither of the two rules constituted a violation of the right
to fair trial, but the fact that the differential treatment was connected with a key fundamental right
ought to be an important consideration.24 Another example may be seen in Re The Kerala Education
Bill reference, where the classification engaged (although it was held not to violate) the freedoms
granted to religious minorities.25 In Subramanian Swamy, the classification was sensitive to the
significant value of corruption-free governance.26 These cases are clearly different from a
classificatory rule of the sort in the Re Natural Resources Allocation reference,27 which engaged the
fairness of the method of allocation of natural resources: an important State interest, but one that does
not constitute a fundamental value or right of the sort engaged in the other three mentioned cases. The
European Court of Human Rights has been particularly sensitive to this parasitic dimension of
equality, since Article 14 of the European Convention of Human Rights only guaranteed that ‘The
enjoyment of the rights and freedoms set forth in this Convention shall be secured without
discrimination.’ It was only in 2000 that Protocol 12 to the Convention expanded the protection
against discrimination to cases where fundamental rights were not engaged. Prior to this Protocol
(and even today for members who haven’t ratified it), the engagement of another fundamental right by
the rule was a jurisdictional prerequisite before Article 14 could be invoked, and therefore went into
the ground of review. The Indian Article 14 (quite sensibly) never had this jurisdictional restriction,
so Indian courts did not feel the need to make this inquiry. However, the engagement of equality with
another fundamental right could be a reason to invoke a higher standard of review.

2. The Differentia
For the second element—differentia—at least two further inquiries may be made: its intelligibility
and its normative permissibility. One could ask whether the differentia is clear or vague; that is,
whether there could be reasonable disagreement over whether a person falls within one group or
another. This is probably what the courts seem to inquire into when they ask if the differentia is
intelligible. Most cases concerning the delegation of excessive and unguided power by the legislature
to the executive are best understood as cases where the intelligibility of the differentia has not been
established—in other words, a failure to classify cases where the discretion should be exercised and
those where it shouldn’t also violates Article 14.28
One could also evaluate the differentia normatively. Most liberal societies have come to accept
that there are certain types of differentia that should not, barring exceptional circumstances, be used
as a basis of legal classification. This is in fact what seems to have been the founders’ intention
behind the non-discrimination clauses in Articles 15(1), 16(2), and 29(2)—to render the specified
differentia (based on sex, race, caste, religion, and other similar personal status) presumptively
impermissible. It is often said that these non-discrimination guarantees merely instantiate the
violations of the right to equality under Article 14.29 This may well be true, but it is hard to believe
that the drafters formulated these provisions simply as illustrations of Article 14, and that they add
nothing to what Article 14 is capable of doing on its own. This scepticism is strengthened by
legislative history. In the earlier drafts of the Constitution, a prototype of Article 15(1) was all there
was under the guarantee of the right to equality. Article 14, in its generality, was added later.30 Surely,
the non-discrimination clauses must have some content independent of Article 14. In particularly
spelling out certain differentia as impermissible, the drafters must have had their presumptive
unacceptability in mind.
Bose J probably also had the normative acceptability of certain differentia in mind when he
exhorted the courts to ask whether the classification would ‘offend the conscience of a sovereign
democratic republic’, viewed in the background of our history.31 Although this normative inquiry did
not become part of the crystallised doctrine, Bose J was surely right. We generally accept that
differentia based on certain personal characteristics—usually constituting a valuable fundamental
choice (such as religion) or which lying outside one’s effective control (such as caste)—are
normatively irrelevant.32 A law that classifies the sellers of tea and those of coffee into two different
groups and imposes differential tax liabilities on them is clearly less problematic than a law which
imposes differential tax liability on Hindus and Muslims (this, incidentally, is exactly the effect of the
income tax law’s recognition of the Hindu Undivided Family as a separate legal person, allowing it to
claim tax deductions not available to non-Hindus). The obvious corollary is that there should be a
higher standard of review for these special cases of classification: an implication that Indian courts
have failed to recognise.33

3. The Objective
Further inquiries can also be made with respect to the third element—the objective of the rule. One
set of questions concerning the objective relate to its genuineness. Is the stated or apparent objective
masking another, more sinister, one? This consideration was clear in Sastri CJ’s judgment in Anwar
Ali Sarkar, where he held that if the law is ‘designed’ to be administered in a discriminatory way, or
is actually ‘administered’ (rather than merely ‘applied’ once in a while) in a discriminatory way, it
will be caught by Article 14, which does not allow ‘colourable legislative expedient’.34 This dictum,
however, sits uncomfortably with the judicial reluctance to inquire into legislative motives in the
context of Article 14.35
Apart from this evidential inquiry into its genuineness, one could also inquire into the normative
legitimacy and the importance of the objective. An objective to establish a theocratic state will not be
constitutionally legitimate (employing constitutional rather than popular morality standards).36 The
objective of administrative efficiency is legitimate, but less weighty than (say) the pursuit of universal
health care. Next, we can also ask whether there is a rational connection between the impugned
measure and the objective in question. As part of this inquiry, judges have often inquired into the
over-inclusiveness and under-inclusiveness of the measure in seeking the stated objective, although
they have generally been tolerant of a considerable degree of misfit.37 We could further ask after the
extent to which the measure is likely to achieve this objective. It may sometimes be the case that the
objective is very weighty, but a classificatory measure achieves it to such a small degree that any
benefits may be outweighed by the harm of classification. Finally, we could ask if the same objective
could be pursued (to a comparable degree) using means that do not restrict fundamental rights. If it
can be, it is not necessary to use the rights-infringing means. Foreign courts typically ask these deeper
nexus questions when conducting a proportionality analysis.

4. The Impact
With respect to impact, one could inquire into the immediate impact of the rule (in our pregnancy
example, the termination of employment), and further material and expressive, if less direct,
implications of the classifying rule on individuals as well as on the groups to which they belong. The
severity of these impacts may also be looked into. Take, for example, the rule in the Air India case,
whose immediate impact was to retire air hostesses at 35 (with the possibility of annual extensions up
to ten years), whereas assistant flight pursers retired at 58.38 The Court’s typically formal inquiry
ignored the impact analysis altogether. The following are the impact-related questions that could be
asked.
First, the immediate impact of this rule is on those air hostesses who lose their job at the age of 35.
Compulsory retirement at an age when one is in the prime of one’s life, especially in light of the fact
that the right to livelihood is a fundamental right flowing from Article 21, makes this a particularly
severe impact.
Secondly, one could inquire into the further impact of the rule on the individuals concerned. Given
that these air hostesses are likely to have spent most of their adult lives in this profession, they are
unlikely to have the skills for a different type of job. At an age when most people are busy climbing
the career ladder, they are faced not just with unemployment, but unemployability. They are likely to
lose all financial independence and, if they are married, become dependent on their spouses. The rule
therefore has very serious implications for their ability to lead an autonomous life.
Thirdly, the impact is not limited to the individuals concerned, but extends to women as a whole.
The rule reduced the already limited employment opportunities available to women, even educated,
middle-class women, who are likely to have been air hostesses. It contributed to the forced
dependence of women on their fathers, husbands, and sons and contributed—in a small but significant
way—to the continued tyranny of patriarchy. It is this implication that makes the rule discriminatory
on the grounds of sex, since it has a disproportionate (and serious) impact on women as a group. The
air hostesses’ argument that the real discrimination is sex based, ‘which is sought to be smoke-
screened by giving a halo of circumstances other than sex’, was considered (and dismissed on flimsy
grounds) only in the context of the Equal Pay Act 1976, and not tested on Article 14.39
Finally, one could ask what social meaning a rule such as this expresses.40 The rule in question sent
a social message that the right place for adult women is in their homes, rather than in the formal
economy, and that it is appropriate to objectify young women for enjoyment of the (mostly male) gaze
of air passengers. This message was strengthened by the fact that pregnancy and marriage also
incurred possible loss of employment for air hostesses in that case. The rule echoed (and, to some
extent, concretised) the idealisation of the role of Indian women as mothers, wives, and daughters; a
phenomenon which has adverse implications for their financial independence and autonomy. Given
that these patriarchal assumptions are already entrenched in our society and that they sustain the
social evil of patriarchy, these expressive implications are very significant.
These questions concerning the impact of the classifying rule concern its operation in the real
world and its effect on real lives. The old doctrine completely ignores the impact question, and is
therefore thoroughly formalistic.

5. A Summary
We can now summarise all the possible questions a court could ask with respect to a classificatory
rule:
1. Right: is the right to equality engaged by the rule?
a. Does the rule have a disproportionate impact on different classes of persons?
b. Was another fundamental right, principle, or value engaged?
2. Differentia: what classes does the rule create?
a. Is the differentia intelligible?
b. Is the differentia presumptively impermissible?
3. Objective: what end does the rule seek to achieve?
a. Is the apparent objective genuine?
b. Is the objective legitimate?
c. Is the objective sufficiently weighty?
d. Is there a rational connection between the impugned measure and the objective?
e. To what extent is the impugned measure likely to achieve the objective?
f. Is the measure necessary to achieve the objective?
4. Impact: what consequences does the rule subject each of these classes to?
a. What immediate consequences does the rule prescribe for the individuals addressed by
it?
b. What are the further material consequences the rule is likely to have for:
– those particular individuals directly addressed by the rule?
– any group to which (most of) the affected persons belong?
c. What are the likely expressive consequences of the rule?
d. How serious are these impacts?
6. Reassessing the Old Doctrine
Some of these questions relate to the ground of review, others concern the appropriate standard of
review and the burden of proof. Of all these possible questions a court could ask when inquiring into
the constitutionality of a classificatory rule, with respect to a classificatory rule, the traditional
classification doctrine only asks the following three:
1. Is the right to equality engaged?
2(a). Is the differentia intelligible? and
3(d). Is there a rational connection between the impugned measure and the objective?

Any classificatory rule satisfies the first question by definition, and at any rate the Court does not ask
it probingly (say, by inquiring into 1(a)). The only real inquiry is therefore into 2(a) and 3(d)—these
together constitute the two limbs of the classification test.41 It may well be defensible to focus only on
these two questions. It is certainly not my case that a court ought to inquire into all of the other
questions in all, or even in any, case. But choosing which questions ought to be asked is an
interpretive choice, one that calls for public justification. Unfortunately, no such justification has been
forthcoming, either from the Court or from scholars, who defend the doctrine in its current form. The
inquiry under the classification test, at least in theory, remains extremely limited. It is also formalistic
because it completely ignores the actual impact of the rule on individuals as well as groups. The
focus remains on the formal elements of the rule—differentia and rational nexus. Since the
classification doctrine does not invite inquiry into any of the other potentially relevant questions
listed above, some judges have been tempted to cheat by somehow incorporating these other
dimensions into the two available to them.42
For example, Gupta J’s dissent in the RK Garg case was really based on the finding that the rule
engaged (and breached) another important principle: that dishonesty ought not to be rewarded. Since
he could not directly access this additional dimension independently, he sought to fit it within both
elements of the available test. He claimed that the classification between dishonest and honest
taxpayers was not intelligible, and that this differentia did not have any rational connection with the
objective (to unearth black money).43 As the majority judgment clearly showed, he was wrong on
both counts. The classification was clear and intelligible; what Gupta J was trying to do was to
update the intelligibility test and turn it into a normative inquiry, by asking if a reasonable fair-minded
person would find it intelligible.44
A similar attempt to fudge the tests to get the desired result was more successful in Subramanian
Swamy. In this case, the objective of the rule requiring prior governmental sanction for an inquiry into
allegations of corruption against senior bureaucrats was to shield them from harassing lawsuits so
that they could exercise their discretionary powers without fear.45 There was much that was wrong
with the rule, but inasmuch as lower-ranking bureaucrats do not wield wide discretionary powers, the
classification between senior and junior officers was clearly intelligible and rationally connected
with the stated objective. The Court struck down the rule, holding that there was no intelligible
differentia.46 The Court could not have found absence of intelligibility in a classification between
higher- and lower- ranking officials unless it was asking it in the normative fashion that Gupta J had
done decades ago. The problem in this case, like the RK Garg case, was that another important value
—the importance of corruption-free administration—was involved. The right way to go about it
would have been to recognise the entanglement of this value with the right to equality and use that as a
basis to conduct a more demanding review. It conducted a strict review anyway, but instead of
justifying it on proper bases and developing the limited doctrine on Article 14, it tried to force a
square peg in a round hole—thereby exacerbating doctrinal confusion.
The Supreme Court has itself recognised that ‘the rule of equality before the law of Article 14’
cannot be equated with ‘broad egalitarianism’.47 Even so, the rule in Article 14 need not remain so
astonishingly limited, formalistic, and confused. It is high time the doctrine evolved. The Court needs
to consider all the questions it could ask under the classification doctrine, and take a considered view
on which of these it should ask in particular categories of cases.

IV. NON-COMPARATIVE UNREASONABLENESS

The ills of the old doctrine have been long recognised. It was in response to these ills that the ‘new’
doctrine was developed. This second doctrinal test to determine the constitutionality of a measure
under Article 14 is called the ‘arbitrariness test’. In a much-quoted paragraph in Royappa (which
incidentally concerned executive action and could have been treated as a matter of administrative
law), Bhagwati J famously railed against the classification doctrine and ‘emancipated’ the principle
of equality from its confinement by holding that mere ‘arbitrariness’ will suffice to constitute a
violation of Article 14.48 The ‘new’ doctrine laid down in Royappa has been affirmed by several
subsequent judgments.49 But controversies continue to dog the doctrine, especially with respect to its
application to legislative review. It seems that Bhagwati J diagnosed the disease correctly, but
prescribed the wrong medicine. Before we delve into it deeper, note that the new doctrine created a
new ground of review under Article 14. It is in the context of a standard of review that the
distinctions between reasonableness, proportionality, strict scrutiny, and similar concepts become
moot.
The content of the arbitrariness doctrine remains unclear. This should hardly be surprising, given
the indeterminacy of the term ‘arbitrary’. Non-arbitrariness was a particularly ill-advised choice of a
test if the purpose of the new doctrine was, as Pal J writing extrajudicially thought it was, ‘to
crystallize a vague generality like Article 14 into a concrete concept’.50 Strictly speaking,
arbitrariness is not the same thing as unreasonableness. Sometimes, it is reasonable to be arbitrary. At
other times, one can be unreasonable without being arbitrary. A decision is reasonable if it is
supported by legitimate and relevant reasons of sufficient weight. A choice between two or more
options is arbitrary only when there is no relevant ground to prefer one option over the other. A
decision based on relevant but illegitimate reasons, or one based on insufficient reasons, would not
be arbitrary.
Consider a rule that fixes the age of consent for sex to be 16 (rather than, say, 15 or 17). Such a law
is clearly arbitrary. In fact, it is impossible to fix any single age as the age of consent for all persons
without it being arbitrary, for the simple reason that different persons reach sexual and emotional
maturity at different ages (the main relevant consideration in determining the age of consent). The only
non-arbitrary way to fix the age of consent would be to individualise it, such that every teenager’s
level of sexual and emotional maturity is constantly evaluated, and only when he or she becomes
sufficiently mature that a competent authority declares him or her to have the capacity to consent to
sex. The clear unreasonableness of the alternative—inasmuch as it entails a serious intrusion into
privacy, will lead to great uncertainty in already unstable teenage lives, and will require a
cumbersome and expensive bureaucracy to administer—should leave us in no doubt that the arbitrary
fixing of the age of consent for everyone at 16 is reasonable.51 In Raju, the Supreme Court rightly
accepted the constitutionality of an arbitrary age of majority in another context.52 The same will be
true of a law fixing the age of consent at 15 or 17. Reason underdetermines where we ought to draw a
line. But some lines are clearly not just arbitrary but also unreasonable—a rule that fixes the age of
sexual consent at 35 is an example.
Take another example. In Javed, a law disqualified anyone with more than two living children from
holding any office in local government.53 One of the arguments mooted before the Court was that the
legislative choice of basing the disqualification on two (rather than one or three, for example)
children made it arbitrary. This is a classic example of a rule that is conceptually arbitrary, but not
unreasonable (on the ground that the disqualification proceeded on having two rather than three
children—the law may well be unreasonable on other grounds, such as excessive intrusion on
privacy). The Court held the legislative choice of two children was not arbitrary.54
We have seen that not everything that we would call arbitrary is also unreasonable. The converse is
also true: an action can be unreasonable without being arbitrary. In Anuj Garg, for example, the
impugned law prohibited women from working as bartenders. The law was not arbitrary in the sense
that it lacked any basis in reason. In the context of a deeply patriarchal society where women are
often seen merely as objects of male desire, and where bars tend to be overwhelmingly male spaces,
the probability that the patrons will harass women bartenders is high. The problem with the decision
to ban women from working as bartenders was not that there was no reason backing it up, but that it
was an illegitimate reason, which further entrenched patriarchy by denying employment to women.
The Court rightly held that the right response would have been to ensure the security of women
bartenders rather than to reduce the already limited options that women have.55 Thus, we have a good
example of an unreasonable rule that is not arbitrary.
Another example of such a provision is to be found in the Air India case. One of the rules
challenged in the case discriminated on the ground of pregnancy in providing for the termination of
the employment contract of an air hostess if she became pregnant. It is clear why the airline had such
a rule—pregnancy of employees is expensive because the airline has to find replacement services and
provide maternity leave. The rule is based on relevant reasons and is not arbitrary in the conceptual
sense. The rule is clearly unreasonable, for it is based on constitutionally unacceptable reasons. As
the Court held, the rule was ‘extremely detestable and abhorrent to the notions of civilised society’
and ‘smacked of utter selfishness’.56 The Court, failing yet again to appreciate the distinction between
arbitrariness and unreasonableness, held the rule to be arbitrary and declared it unconstitutional for
that reason.
What we should surmise from these cases is that what the Court means by ‘arbitrariness’ is really
‘unreasonableness’. The new doctrine has read into the right to equality a right against unreasonable
State action. As a free-standing constitutional right, this is a remarkable development. Of course,
constitutional law traditionally insists on reasonableness when an important constitutional right or
value is interfered with by State action. A reasonableness inquiry usually requires an external trigger
—say, a restriction on the right to free expression under Article 19(1)(a), or a restriction on personal
liberty under Article 21, or even a classification of persons into different groups for unequal
treatment. When these key rights are restricted, the Constitution demands that the restriction must be
reasonable. Thus, in constitutional law, the right against unreasonableness has traditionally been a
parasitic right, piggybacking on other fundamental rights. What the new doctrine has done is to
elevate an administrative law reasonableness standard into a self-standing constitutional ground for
review, without the need for any crutches in the form of other rights or values. This elevation may
have implications for legislative review.

1. Equality, Arbitrariness, and the Rule of Law

Scholars have pointed out several deficiencies in the new doctrine of non-arbitrariness.57 It has been
criticised for its vagueness and for its possible implications for the balance of powers between the
judiciary and the legislature. For our purposes, the most relevant criticism concerns the logical
possibility of locating a free-standing right against unreasonableness within the right to equality. The
problem is that the new doctrine does not require any classification to trigger its inquiry, whereas the
right to equality, whatever else it might be, must essentially be comparative.58 Seervai best
articulated this criticism by suggesting that ‘The new doctrine hangs in the air, because it propounds a
theory of equality without reference to the language of Art. 14.’59 It is this claim that we must now
investigate.
This is how Bhagwati J sought to establish the connection between the new doctrine and the right to
equality:
[E]quality is antithetic to arbitrariness. In fact equality and arbitrariness are sworn enemies; one belongs to the rule of law in a
republic while the other, to the whim and caprice of an absolute monarch. Where an act is arbitrary it is implicit in it that it is
unequal both according to political logic and constitutional law and is therefore violative of Art. 14.60

Possibly, this passage seeks to establish a historical connection between equality and non-
arbitrariness—via the rule of law. The right to ‘equality before the law’, which constitutes the first
part of Article 14, has been variously attributed to the English common law and to the Irish
Constitution.61 Insofar as its roots can be traced back to the English common law, the right to equality
before the law can be seen to manifest itself most securely in the second principle of the rule of law
that the British jurist Dicey articulated thus:
We mean in the second place, when we speak of the ‘rule of law’… not only that with us no man is above the law, but … that
here every man, whatever be his rank or condition, is subject to the ordinary law of the realm and amenable to the jurisdiction of
the ordinary tribunals.62

This rule was, quite literally, a manifestation of equality before the law—everyone was to be subject
to the same law administered by the same courts. Dicey explained further that:
In England the idea of legal equality, or of the universal subjection of all classes to one law administered by the ordinary Courts,
has been pushed to its utmost limit. With us every official, from the Prime Minister down to a constable or a collector of taxes, is
under the same responsibility for every act done without legal justification as any other citizen.63

So far, we can see a historical connection between equality before the law and a particular aspect of
the rule of law that concerns itself with who is governed by the law. This principle of the rule of law
has little to do with arbitrariness. A legal system that exempts public officials from the operation of
ordinary laws is not arbitrary for that reason, even as it violates Dicey’s second principle of the rule
of law. It engages the right to equality because it classifies people into two groups—those who are
public officials and those who are not—and prescribes starkly different consequences for each group
for legal breaches. The second principle deems this unacceptable because this form of unequal
treatment is unreasonable. But there is nothing arbitrary about the distinction.
The principle of arbitrariness is instead to be found in Dicey’s first principle of the rule of law,
concerning not who is governed but how they are governed. This principle, often articulated as ‘the
rule of law, not men’, was expressed by Dicey thus:
We mean, in the first place, that no man is punishable or can be lawfully made to suffer in body or goods except for a distinct
breach of law established in the ordinary legal manner before the ordinary Courts of the land. In this sense the rule of law is
contrasted with every system of government based on the exercise by persons in authority of wide, arbitrary, or discretionary
powers of constraint.64

This first principle prohibits the State’s interference with any person without the authority of law. In
our democratic times, it may seem to be unexceptionable. But at a time when monarchs were used to
mete out favours and penalties on whim, this republican principle65 of legality must have been
revolutionary. In the Case of Prohibitions, the British King James I placed himself as a judge in a
dispute. Standing up to the monarch, a courageous judge insisted that ‘The King in his own person
cannot adjudge any case, either criminal or betwixt party and party; but it ought to be determined and
adjudged in some Court of Justice, according to the law and custom of England.’66 This insistence on
legality reduces arbitrariness in the State’s treatment of its subjects because of certain features that
law tends to possess: law tends to be (and ought normally to be) general, published, prospective, and
stable.67 These features of law—especially when it is administered by independent, professional,
technically trained enforcement bodies—ensures that personal fortunes are not subject to the whims
and fancies of monarchs. The second rule of law principle embodies an element of non-arbitrariness.
But even this first principle of the rule of law is not interchangeable with the principle of non-
arbitrariness. It only obviates one form of arbitrariness—one that proceeds from whimsical executive
determination of rights and liabilities without any basis in general, published, prior law. It does not,
in itself, protect a person from arbitrariness when it is embedded in the law itself.
This brief historical summary shows that the Court arrived from the right to equality to a general
self-standing principle of non-arbitrariness by exploiting an ambiguity between two different
principles of the rule of law (concerning equality and legality, respectively) and by extending
significantly the principle of legality. The connection between equality and non-arbitrariness was
therefore based on a conceptual confusion between different meanings of the rule of law.
If one was less charitable, this judicial manoeuvre could be seen not simply as a misreading of the
historical connections between concepts, but also as a deliberate and unjustified appropriation of
power. After all, Pal J did publicly acknowledge that rights like the right to equality were ‘empty
vessels’ into which ‘each generation pours its content by judicial interpretation’.68 This remarkable
suggestion leaves no role for constitutional text whatsoever. Rights like that to equality are reduced to
mere place holders, which can be filled in, ironically, by whatever the judge arbitrarily thinks is
appropriate. What makes the claim even more startling is that an admittedly vague right to equality
has been sought to be concretised by a similarly vague right to non-arbitrariness.
The connection between the right to equality and the new doctrine becomes even more dubious in
light of our knowledge that the right against arbitrariness is in reality a right against unreasonableness.
Not all forms of unreasonableness entail inequality. Torture is unreasonable, but the wrongness of
torture is not due to it being inegalitarian. The same is true of a whole range of unreasonable acts—
prohibition of criticism of government, ban on religious conversion, denial of access to courts,
prescription of a uniform for citizens, or prohibition on the drinking of tea. All these acts are
unreasonable, but their unreasonableness has little, if anything, to do with inequality. The new
doctrine therefore completely ignores the text of Article 14 and the concept of equality. Equality may
mean many things, but it cannot mean anything. An inquiry into the reasonableness of State action
under the new doctrine no longer requires a prior demonstration that some form of inequality is
involved, as was the case with the old classification doctrine. In the garb of the right to equality, the
new doctrine tries to institute an independent constitutional right against unreasonableness.
The right against unreasonable administrative action is hardly novel. It has been a recognised right
in administrative law at least since the famous British decision in the Wednesbury case.69 What is
remarkable about the new doctrine is that it elevates this free-standing administrative law right
against (a high-threshold) unreasonableness to a constitutional fundamental right, potentially against
all forms of State action. One implication of this move is that a challenge to unreasonable State
action, even if it is only an administrative action, can only be made by following the procedure
prescribed for a fundamental rights claim—that is, by filing a writ in a constitutional court. This
constitutionalisation of administrative law ignores its common law roots, and results in a top-heavy
system where constitutional courts come to arrogate all administrative review powers.

2. Legislative Review and the New Doctrine

The more relevant implication for our purpose is the potential expansion of the unreasonableness
claim to legislative action. In English common law, Wednesbury unreasonableness was available
only against administrative acts. As a self-standing constitutional right in India, it can, at least
theoretically, extend to legislative action. Let us recall the different senses in which we understood
the term ‘legislative’ at the start of this chapter. It is abundantly clear that the Supreme Court has been
applying the right to reasonableness to straightforwardly administrative acts.70 But does it apply to
primary legislation enacted by a legislature and to delegated legislation or ordinances enacted by the
executive?
Let us first deal with whether the new doctrine can be applied to legislative enactments by
legislatures. There are five relevant Supreme Court rulings on the issue: McDowell,71 Malpe
Vishwanath,72 Mardia Chemicals,73 Re Natural Resources Allocation (hereinafter 2G Reference),74
and Subramanian Swamy.75 The evolving doctrinal position appears to be that the new doctrine does
not apply to legislation enacted by a legislature, but one cannot yet assert this with too much certainty.
In McDowell, a three-judge bench refused to test a legislative provision under the new doctrine.
Recognising the distinction between testing the reasonableness of a measure if it restricts a
fundamental right and testing its reasonableness per se, Jeevan Reddy J said, ‘[i]t is one thing to say
that a restriction imposed upon a fundamental right can be struck down if it is disproportionate,
excessive or unreasonable and quite another thing to say that the court can strike down enactment if it
thinks it unreasonable, unnecessary or unwarranted’.76 The corollary was that ‘[n]o enactment can be
struck down by just saying that it is arbitrary or unreasonable. Some other constitutional infirmity has
to be found before invalidating an Act’.77 Statutes violate Article 14 only in the context of a
classification.78 However, in Malpe Vishwanath, another three-judge bench held that a statutory
provision had become arbitrary (unreasonable) due to changed circumstances with the passage of
time since its enactment.79 The third case in our quintet is Mardia Chemicals. Here, yet another three-
judge bench struck down a statutory provision as unreason- able.80 None of these cases involved any
classification. It would have seemed that McDowell was dead.
McDowell was nonetheless resurrected by a five-judge bench in its opinion in the 2G Reference
case. The Court quoted extensively from it with approval, and concluded that ‘A law may not be
struck down for being arbitrary without the pointing out of a constitutional infirmity.’81 The trouble is
that the 2G Reference did not involve legislative review, and therefore its resurrection of McDowell
remains obiter.
Finally, another five-judge bench in Subramanian Swamy was pressed to confirm that arbitrariness
is not available as a ground for legislative review. Because the impugned provision was found to
violate the classification doctrine in any case, the Court’s observations on the new doctrine again
amount to obiter dicta. Even so, the Court ruled out its application indirectly, but setting out the list of
grounds that were available to review legislation:
The fundamental nature and importance of the legislative process needs to be recognized by the Court and due regard and
deference must be accorded to the legislative process. The two dimensions of Article 14 in its application to legislation and
rendering legislation invalid are now well recognized and these are (i) discrimination, based on an impermissible or invalid
classification and (ii) excessive delegation of powers; conferment of uncanalised and unguided powers on the executive, whether
in the form of delegated legislation or by way of conferment of authority to pass administrative orders—if such conferment is
without any guidance, control or checks, it is violative of Article 14 of the Constitution.82

Thus, the Court suggests that there are only two grounds to challenge legislation (or, at least
legislation that has been enacted by the traditional legislative process). These grounds are
unreasonable classification and excessive delegation. I have already argued that the excessive
delegation doctrine is best understood as the flip side of the classification doctrine. It is simply a
failure to classify when classification is warranted: the legislature fails to indicate the distinction
between cases where the delegated power ought to be exercised from those where it ought not to be.
The implication of the 2G Reference and Subramanian Swamy is that the new doctrine is not
available for legislative review. The judicial trend is one of hostility to the application of the
arbitrariness doctrine to legislation (enacted by a legislature). The doctrine can, however, be
clarified only after this hostility forms part of the ratio of a case decided by a constitutional bench.
What about legislation by the executive? It isn’t clear whether the review of ordinances is any
different in this regard from that of primary legislation. But there is some jurisprudence on the review
of secondary legislation enacted by the executive under powers delegated by a statute. The most
important authority on the point is the Air India case, where a rule requiring the termination of the
services of an air hostess on her pregnancy was struck down as arbitrary.83 The authority of the case,
however, is hardly clear. The case involved a clear classification between employees who become
pregnant and those who do not. Although the Court found the rule to be arbitrary per se, it also went
on to hold that pregnancy discrimination was an instance of sex discrimination, and therefore
unconstitutional—thereby admitting the classificatory nature of the rule.84 What complicates matters
further is that McDowell specifically confined the exclusion of the new doctrine to ‘an Act made by
the legislature. [It expressed] no opinion insofar as delegated legislation is concerned.’85 The 2G
Reference opinion, however, misread McDowell: ‘cases where legislation or rules have been struck
down as being arbitrary in the sense of being unreasonable … only on the basis of “arbitrariness” … 
have been doubted in McDowell’s case’.86
But McDowell quite explicitly refused to doubt the application of the unreasonableness test to
delegated legislation. This doctrinal messiness is a result of the lack of clarity over the
distinctiveness of legislation (in the product-sensitive sense) from the legislature (in the actor-
sensitive sense). The question of the applicability of the ‘arbitrariness test’ to the review of
legislation needs settling.

V. CONCLUSION
The following conclusions emerge: (a) the ‘classification test’ (or the unreasonable comparison test)
continues to be applied for testing the constitutionality of classificatory rules; (b) it is a limited and
highly formalistic test applied deferentially; (c) the ‘arbitrariness test’ is really a test of
unreasonableness of measures which do not entail comparison (hence labelled non-comparative
unreasonableness); (d) its supposed connection with the right to equality is based on a conceptual
misunderstanding of the requirements of the rule of law; and (e) courts are unlikely to apply it to
legislative review (at least in the actor-sensitive sense). Article 14 has become a victim of the weak
‘old’ doctrine and the over-the-top ‘new’ doctrine. The former needs expansion and substantiation, the
latter relegation to its rightful place as a standard of administrative review.

* I am grateful to Arthad Kurlekar, Nick Barber, Gautam Bhatia, Aparna Chandra, Abhinav Chandrachud, David Grewal, Sudhir
Krishnaswamy, Manoj Mate, Raju Ramachandran, Gopal Sankaranarayanan, Tamas Szigeti, and the editors for helping in various ways
with this chapter.
1
India has tended to focus on affirmative action rather than anti-discrimination. See Tarunabh- Khaitan, ‘Transcending Reservations:
A Paradigm Shift in the Debate on Equality’ (2008) 43(38) Economic and Political Weekly 8; Tarunabh Khaitan, ‘The Architecture of
Discrimination Law’ in Vidhu Verma (ed) Unequal Worlds: Discrimination and Social Inequality in Modern India (Oxford
University Press 2015).
2
In Malpe Vishwanath Acharya v State of Maharashtra (1998) 2 SCC 1 [16]–[17], the Court noted the extensive consultation that
preceded the legislation. See generally, Jeff King, ‘Deference, Dialogue and Animal Defenders International’ (UK Constitutional Law
Blog 25 April 2013) <http://ukconstitutionallaw.org/2013/04/25/jeff-king-deference-dialogue-and-animal-defenders-international/>,
accessed November 2015.
3
Twenty-seven per cent of the Bills passed in the Lok Sabha in 2009, PRS Legislative Research has calculated, were discussed for
less than five minutes in the House.
4
State of West Bengal v Anwar Ali Sarkar AIR 1952 SC 75 [22], [25]: Fazal Ali J said that a legislation copied verbatim from a pre-
constitutional ordinance could not have borne the Constitution in mind. See also Goutham Shivshankar, ‘Naz and the Need to Rethink the
Presumption of Constitutionality’ (Law and Other Things 13 December 2013) <http://lawandotherthings.blogspot.co.uk/2013/12/naz-
and-need-to-rethink-presumption-of.html>, accessed November 2015.
5
Chiranjit Lal Chowdhuri v Union of India AIR 1951 SC 41.
6
See Constitution of India 1950, art 145(1). Cases involving review of judicial rules are rare, and will not inform the discussion in this
chapter.
7
In general, the standard of review in India has been low and the burden of proving unconstitutionality always falls on the
complainant.
8
Tarunabh Khaitan, ‘Koushal v Naz: A Legislative Court and the Recriminalisation of Homosexuality in India’ (2015) 78(4) Modern
Law Review 672. See also Chintan Chandrachud, ‘Constitutional Interpretation’ (chapter 5, this volume), discussing panchayati
adjudication.
9
See Constitution of India 1950, art 145(3).
10
On Das J’s role in the consolidation of the classification test, see PK Tripathi, Some Insights into Fundamental Rights (Bombay
University Press 1972) 52f.
11
The constitutionalisation of administrative law has done considerable damage to both constitutional law and administrative law in
India. See generally, Farrah Ahmed and Tarunabh Khaitan, ‘Constitutional Avoidance in Social Rights Litigation’ (2015) 35 Oxford
Journal of Legal Studies 607.
12
EP Royappa v State of Tamil Nadu (1974) 4 SCC 3.
 13
Om Kumar v Union of India (2001) 2 SCC 386 [59]: the non-arbitrariness ‘principle is now uniformly followed in all Courts more
rigorously than the one based on classification’.
14
Subramanian Swamy v Central Bureau of Investigation (2014) 8 SCC 682 [58].
15
Ajay Hasia v Khalid Mujib Sehravardi (1981) 1 SCC 722; Air India v Nargesh Meerza (1981) 4 SCC 335; Javed Abidi v
Union of India (1999) 1 SCC 467; Githa Hariharan v Reserve Bank of India (1999) 2 SCC 228.
16
Shelley v Kraemer 334 US 1 (1948). See also Nicholas Bamforth, ‘The True “Horizontal Effect” of the Human Rights Act 1998’
(2001) 117 Law Quarterly Review 34.
17
Zoroastrian Cooperative Housing Society v District Registrar (2005) 5 SCC 632. There was, of course, a material difference
between the Shelley and Zoroastrian Housing cases. In Shelley, the target of the discriminatory covenant was a disadvantaged racial
minority. In Zoroastrian Housing, it was a religious minority that wanted to exclude those belonging to other religions. The role that this
distinction might have played in the reasoning of the Supreme Court is debatable, but it is possible for the Court to distinguish Zoroastrian
Housing if a Shelley-type case comes before it.
18
Constituent Assembly Debates, vol 3 (Lok Sabha Secretariat 1986) 427, 29 April 1947 (RK Sidhwa); Constituent Assembly
Debates, vol 7 (Lok Sabha Secretariat 1986) 651, 657, 661, 29 November 1948 (KT Shah, S Nagappa, BR Ambedkar).
19
This claim does not entail the corollary that discrimination necessarily involves comparison. In fact, the essential competitiveness
of equality is precisely the reason that has led some scholars to see equality and non-discrimination as distinct ideals. See generally, Elisa
Holmes, ‘Anti-Discrimination Rights Without Equality’ (2005) 68 Modern Law Review 175.
20
Anwar Ali Sarkar (n 4) [49].
21
Nargesh Meerza (n 15) [67].
22
Joseph Raz, The Authority of Law (2nd edn, Oxford University Press 2009) ch 10.
23
A similar rule was struck down by a three-judge bench in Nargesh Meerza (n 15).
24
Anwar Ali Sarkar (n 4).
25
Re The Kerala Education Bill AIR 1958 SC 956.
26
Subramanian Swamy (n 14) [63].
27
Re Natural Resources Allocation (2012) 10 SCC 1.
28
Contrast Anwar Ali Sarkar (n 4) and Special Reference No 1 of 1978 (1979) 1 SCC 380.
29
Indra Sawhney v Union of India (1992) Supp (3) SCC 217 [262].
30
Constituent Assembly Debates, vol 3 (Lok Sabha Secretariat 1986) 410, 29 April 1947.
31
Anwar Ali Sarkar (n 4) [95].
32
John Gardner, ‘On the Ground of Her Sex(uality)’ (1998) 18 Oxford Journal of Legal Studies 167.
33
Tarunabh Khaitan, ‘Beyond Reasonableness: A Rigorous Standard of Review for Article 15 Infringement’ (2008) 50 Journal of the
Indian Law Institute 177; Tarunabh Khaitan, ‘Reading Swaraj into Article 15: A New Deal for all Minorities’ (2009) 2 NUJS Law
Review 419.
34
Anwar Ali Sarkar (n 4) [12].
35
In Re The Kerala Education Bill (n 25), for example, the Court refused to inquire into whether the Bill was designed to target
Christian schools.
36
Naz Foundation v Union of India (2009) 160 DLT 277 [79]—although the judgment of the High Court has been overruled, the
constitutional morality point is eminently sensible.
37
See Special Reference No 1 of 1978 (n 28) [134]–[136]; Ram Krishna Dalmiya v Justice SR Tendolkar AIR 1958 SC 538 [16].
38
Nargesh Meerza (n 15).
39
Nargesh Meerza (n 15) [64].
40
Tarunabh Khaitan, ‘Dignity as an Expressive Norm: Neither Vacuous Nor a Panacea’ (2012) 32 Oxford Journal of Legal Studies 1.
41
Tripathi complains that even these two inquiries have sometimes been confused. Tripathi (n 10) 58–59.
42
See Tripathi (n 10) 59f.
43
RK Garg v Union of India (1981) 4 SCC 675 [26]–[31].
44
RK Garg (n 43) [31].
45
Subramanian Swamy (n 14) [63].
46
Subramanian Swamy (n 14) [68].
47
Sanjeev Coke v Bharat Coking Coal (1983) 1 SCC 147 [17].
48
EP Royappa (n 12) [85].
49
Maneka Gandhi v Union of India (1978) 1 SCC 248; Ajay Hasia (n 15).
50
Ruma Pal, ‘Judicial Oversight or Overreach’ (2008) 7 SCC J 9, J16.
51
I am not defending this as an optimal policy choice. A better alternative is a rolling age-of-consent regime, which only prohibits
sexual liaisons involving a young person if the age gap between the parties is more than three or four years. However, even this more
reasonable regime will need to arbitrarily determine who counts as a ‘young person’ and what gap in the ages of the parties is too much
for the law to tolerate.
52
Subramanian Swamy v Raju (2014) 8 SCC 390 [63]: Article 14 will tolerate ‘the inclusion of all under 18 into a class “juveniles” ’.
53
Javed v State of Haryana (2003) 8 SCC 369.
54
Javed (n 53) [8].
55
Anuj Garg v Hotel Association of India (2008) 3 SCC 1 [33], [36].
56
Nargesh Meerza (n 15) [82].
60
EP Royappa (n 12) [85].
57
HM Seervai, Constitutional Law of India, vol 1 (4th edn, Universal Law Publishing 2002) 438; TM Andhyarujina, ‘The Evolution
of Due Process of Law by the Supreme Court’ in BN Kirpal and others (eds) Supreme But Not Infallible (Oxford University Press
2004) 193, 205–11; BN Srikrishna, ‘Skinning a Cat’ (2005) SCC J 3, J9–11.
58
Tripathi (n 10) 101.
59
Seervai (n 57) 438.
61
In Anwar Ali Sarkar (n 4) [43], Mukherjea J identified the English Constitution as the source of the phrase ‘equality before the
law’. In Special Reference No 1 of 1978 (n 28) [72], Chandrachud CJ described it as a basic principle of republicanism, and attributed
its source to the Irish Constitution.
62
Albert Dicey, Introduction to the Study of the Law of the Constitution (9th edn, Macmillan 1939) 193.
63
Dicey (n 62) 193.
64
Dicey (n 62) 188.
65
Subramanian Swamy (n 14) [38].
66
Case of Prohibitions [1607] EWHC J23 (KB).
67
Joseph Raz, ‘The Rule of Law and its Virtue’ (1977) 93 The Law Quarterly Review 195.
68
Ruma Pal (n 50) J15, quoting Learned Hand J.
69
Associated Provincial Picture Houses v Wednesbury Corporation [1948] 1 KB 223 (CA).
82
Subramanian Swamy (n 14) [49].
70
EP Royappa (n 12); Ajay Hasia (n 15). Affirmed in State of Andhra Pradesh v McDowell & Co (1996) 3 SCC 709 [43].
71
McDowell (n 70).
72
Malpe Vishwanath Acharya (n 2).
73
Mardia Chemicals v Union of India (2004) 4 SCC 311.
74
Re Natural Resources Allocation (n 27).
75
Subramanian Swamy (n 14).
76
McDowell (n 70) [43].
77
McDowell (n 70) [43]. Reddy J confuses grounds of review (such as the classification or the arbitrariness doctrine) with possible
remedies. Striking down an unconstitutional provision is but one possible remedy. Often, radical reinterpretation or reading down of a
provision to make it constitutionally compatible will suffice: Githa Hariharan (n 15) [9]; Danial Latifi v Union of India (2001) 7 SCC
740 [33]; Naz Foundation (n 36) [132]. Rarely, a court may even be justified in simply making a declaration of unconstitutionality and
leaving it for the legislature to fix the infirmity: Malpe Vishwanath Acharya (n 2) [31].
78
McDowell (n 70) [43]–[44].
79
Malpe Vishwanath Acharya (n 2) [27], [31].
80
Mardia Chemicals (n 73) [64]. See Abhinav Chandrachud, ‘How Legitimate is Non-Arbitrariness? Constitutional Invalidation in
the Light of Mardia Chemicals v Union of India’ (2008) 2 Indian Journal of Constitutional Law 179.
81
Re Natural Resources Allocation (n 27) [105].
83
Nargesh Meerza (n 15) [82].
84
Nargesh Meerza (n 15) [84]–[87].
85
McDowell (n 70) [46].
86
Re Natural Resources Allocation (n 27) [106] (emphasis added).