The Core Of An Uneasy
Case For Judicial Review
The Harvard community has made this
article openly available. Please share how
this access benefits you. Your story matters
Citation Richard H. Fallon, The Core Of An Uneasy Case For Judicial Review,
121 Harv. L. Rev. 1693 (2008).
Published Version http://www.harvardlawreview.org/media/pdf/
fallon_judicialreview.pdf
Citable link http://nrs.harvard.edu/urn-3:HUL.InstRepos:11222677
Terms of Use This article was downloaded from Harvard University’s DASH
repository, and is made available under the terms and conditions
applicable to Open Access Policy Articles, as set forth at http://
nrs.harvard.edu/urn-3:HUL.InstRepos:dash.current.terms-of-
use#OAP
1693
VOLUME 121 MAY 2008 NUMBER 7
© 2008 by The Harvard Law Review Association
ARTICLE
THE CORE OF AN UNEASY CASE FOR JUDICIAL REVIEW
Richard H. Fallon, Jr.
TABLE OF CONTENTS
I. WALDRON’S ARGUMENT THAT OUTCOME-RELATED REASONS ARE
INADEQUATE TO SUPPORT JUDICIAL REVIEW..........................................................1701
A. Waldron’s Assumptions...................................................................................................1701
B. Which Assumptions Do What Work?...........................................................................1701
II. PREFERRED RIGHTS AND OUTCOME-RELATED REASONS TO SUPPORT
JUDICIAL REVIEW..............................................................................................................1704
A. Distinguishing Errors of Underenforcement and Overenforcement of
Individual Rights............................................................................................................1704
B. The Limits of the Outcome-Based Case for Judicial Review: Contestable
Premises and the Burdens of Judgment.......................................................................1710
III. PROCESS-BASED REASONS AND POLITICAL LEGITIMACY .....................................1715
A. Waldron’s Process-Based Argument — And Its Limits .............................................1716
B. Political Legitimacy and Its Sources ...........................................................................1717
C. Comparative Democratic and Political Legitimacy....................................................1718
1. Anchoring Assumptions............................................................................................1718
2. Transitional Questions..............................................................................................1720
3. Judicial Review Without Entrenchment................................................................1720
4. Entrenched Rights and Judicial Review................................................................1723
(a) Entrenched Rights Without Judicial Review...................................................1723
(b) Entrenched Rights Coupled with Entrenched Judicial Review....................1726
IV. NOTES ON THE DESIGN OF A SYSTEM OF JUDICIAL REVIEW ..............................1728
A. Judicially Reviewable Issues.........................................................................................1728
B. Scope of Review ..............................................................................................................1732
C. Choosing Strong or Weak Judicial Review .................................................................1733
D. Judicial Review in Societies that Are Not Well Ordered..........................................1734
V. CONCLUSION.......................................................................................................................1735
1694
THE CORE OF AN UNEASY CASE FOR JUDICIAL REVIEW
Richard H. Fallon, Jr.∗
The best case for judicial review in politically and morally healthy societies does not
depend (as is commonly believed) on the idea that courts are more likely than
legislatures to define vague rights correctly. It rests instead on the subtly different
ground that legislatures and courts should both be enlisted to protect fundamental rights
and, accordingly, that both should have veto powers over legislation that might
reasonably be thought to violate such rights.
In developing this case for judicial review, Professor Fallon proceeds by confronting
recent, influential, philosophically probing arguments against judicial review by
Professor Jeremy Waldron. Professor Fallon concedes arguendo that, as Professor
Waldron argues, courts are no better than legislatures at defining rights correctly, but
maintains that the crucial question is not whether courts or legislatures are less likely to
err, but which kinds of errors are most important to avoid — those that result in rights
being overprotected or those that result in rights being infringed. Insofar as judicial
review can be designed to prevent errors in just one direction, involving failures to
protect rights adequately, then judicial review may be supportable even if courts are no
better than legislatures at identifying rights correctly. Professor Fallon also argues,
contra Professor Waldron, that judicial review can actually contribute to the political
legitimacy of an otherwise democratic scheme of government when the demands of
political legitimacy are understood correctly.
Professor Fallon’s revised justification for judicial review, which does not presume
courts to be better than legislatures at identifying fundamental rights, has important
implications for how judicial review should be practiced. It implies a diminished role
for courts in cases in which fundamental rights are pitted against one another, such that
the overenforcement of one entails the underenforcement of the other. It also implies
that courts should withhold review when legislatures conscientiously seek to protect one
fundamental right without plausibly threatening another.
or a long season, the desirability of judicial review of legislation
was a complacent assumption of American constitutional, political,
and moral thought. A vigorous debate percolated about how courts
should interpret the Constitution, but not much serious discussion
addressed whether judicial review should exist at all. Now matters
have changed. Although debate continues concerning how courts
should make constitutional decisions, distinguished critics have begun
to argue for a fundamental rethinking of the role of courts in a democ-
–––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
∗ Ralph S. Tyler, Jr. Professor of Constitutional Law, Harvard Law School. I am grateful to
Heather Gerken, Don Herzog, Allan Hutchinson, Daryl Levinson, Frank Michelman, Martha
Minow, Warren Postman, John Pottow, Richard Primus, Eric Rakowski, Sam Scheffler, Mike
Seidman, and Adrian Vermeule, and to participants in workshops at the University of Michigan
Law School and UC Berkeley School of Law, for extremely helpful comments on earlier drafts.
Elizabeth Barchas provided outstanding research assistance.
F
2008] AN UNEASY CASE FOR JUDICIAL REVIEW 1695
ratic culture such as ours.1 Some advocate the total abolition of judicial
review.2
Having heard the critics, I now believe that the affirmative case for
judicial review needs to be partially revised if judicial review is to be
defended successfully on the moral high ground of liberal political theory.
In a nutshell, the best case for judicial review in politically and
morally healthy societies does not rest (as has often been asserted) on
the idea that courts are more likely than legislatures to make correct
decisions about how to define vague rights of the kind commonly included
in bills of rights — on notions, for example, that courts are peculiarly
well designed to function as “forum[s] of principle.”3 The best
case, as Frank Cross also has argued,4 rests instead on the subtly different
ground that legislatures and courts should both be enlisted in
protecting fundamental rights, and that both should have veto powers
over legislation that might reasonably be thought to violate such
rights.
A suggestive, albeit not perfect, analogy comes from the federal
jury system in criminal cases, under which a defendant cannot be convicted
without the unanimous agreement of the jury,5 and each of the
twelve jurors must vote to acquit unless persuaded that the defendant
has been proven guilty “beyond a reasonable doubt.”6 If the concern
–––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
1 See, e.g., LARRY D. KRAMER, THE PEOPLE THEMSELVES: POPULAR CONSTITUTIONALISM
AND JUDICIAL REVIEW (2004).
2 See, e.g., MARK TUSHNET, TAKING THE CONSTITUTION AWAY FROM THE COURTS
154–76 (1999).
3 See RONALD DWORKIN, A MATTER OF PRINCIPLE 69–71 (1985) (characterizing the Supreme
Court as “the forum of principle”).
4 Frank B. Cross, Institutions and Enforcement of the Bill of Rights, 85 CORNELL L. REV.
1529, 1576 (2000) (arguing that judicial review may be justified even if a judiciary lacks “any intrinsic
advantage in constitutional interpretation and enforcement” because “adding an additional
check on government action will enhance the liberty the Bill of Rights offers”).
5 The analogy of multiple jurors having vetoes to the vetoing effect of judicial review is admittedly
not perfect. Whereas the Supreme Court can conclusively nullify an enactment that it
believes to be unconstitutional (at least in the absence of a constitutional amendment), a single
dissenting juror cannot absolutely block a conviction, due to the possibility of a retrial in which a
conviction might subsequently be obtained. But see Jeffrey Rosen, After ‘One Angry Woman,’
1998 U. CHI. LEGAL F. 179, 191 (reciting conclusions of a study in which “[o]nly 34 percent of
[hung jury cases were] actually retried” and asserting that among “11-to-1 hung-jury cases that
are retried, juries convict only 58 percent of the time, hanging again or acquitting in 42 percent of
the cases” (emphases omitted)). In addition, where judicial review operates as a safeguard only of
the interests that would be harmed by a legislative enactment, the requirement of jury unanimity
applies to what otherwise would be majority decisions to acquit, as well as to convict, a defendant.
See Jason D. Reichelt, Standing Alone: Conformity, Coercion, and the Protection of the
Holdout Juror, 40 U. MICH. J.L. REFORM 569, 570 (2007) (“The federal criminal trial system and
every state, with the exception of two, require unanimous verdicts for conviction or acquittal in
felony trials.”).
6 See In re Winship, 397 U.S. 358, 363 (1970) (quoting Davis v. United States, 160 U.S. 469,
493 (1895)).
1696 HARVARD LAW REVIEW [Vol. 121:1693
were simply to get correct judgments about whether the accused had
committed a crime, decisions by majority vote, pursuant to a preponderance
of the evidence standard, would produce more accurate outcomes.7
Instead, we skew the system in a pro-defendant direction
based on the premise that errors resulting in mistaken convictions of
the innocent are morally worse, and thus more important to avoid,
than erroneous acquittals of the guilty.8 In other words, we care less
about minimizing the overall number of errors than about minimizing
the errors in a particular direction — a situation that might also obtain
with respect to judgments involving individual rights.9
In developing this case for judicial review, my argument proceeds
by confronting and then refuting the most profound challenge to judicial
review that has achieved prominence in the law reviews — Jeremy
Waldron’s argument in The Core of the Case Against Judicial Review.10
In that article, Waldron first debunks what he calls the “outcome-related”
argument on which he takes the affirmative case for judicial
review to depend — the argument that courts are more likely
than legislatures to identify individual rights correctly. In any reasonably
democratic society that satisfies four “quite demanding”11 but
“not unrealistic”12 conditions,13 he argues, there is no reason to think
courts more likely than legislatures to determine unerringly what
rights people actually have14 (on the assumption, which I too shall
adopt for purposes of this Article, that debates about constitutional
rights at bottom are or ought to be debates about moral rights).15 The
most important of these conditions are that legislators take questions
involving individual rights seriously and that “reasonable disagreement”
exists about their correct resolution.16 (Waldron does not chal-
–––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
7 See LARRY LAUDAN, TRUTH, ERROR, AND CRIMINAL LAW: AN ESSAY IN LEGAL
EPISTEMOLOGY 29–30 (2006) (examining traditional principles of criminal law to show that
standards such as the requirement of proof beyond a reasonable doubt are designed to ensure that
errors “will be predominantly false acquittals rather than false convictions”).
8 See id. at 63 (collecting authorities that so assert).
9 See Cross, supra note 4, at 1592 (arguing that “false negatives (underenforcement of constitutional
freedoms) are both more serious and more likely to occur than false positives (overenforcement
of constitutional freedoms)” and calling “for a decision rule that does not necessarily
minimize all mistakes but rather minimizes false negatives”).
10 Jeremy Waldron, The Core of the Case Against Judicial Review, 115 YALE L.J. 1346 (2006)
[hereinafter Waldron, The Core Case]. Waldron has advanced similar arguments before, see, e.g.,
JEREMY WALDRON, LAW AND DISAGREEMENT 211–312 (1999), but he writes in the most recent
article, “I am not satisfied that I have [previously] stated in a clear and uncluttered way what
the basic objection [to judicial review] is.” Waldron, The Core Case, supra, at 1352–53.
11 Waldron so characterizes them. Waldron, The Core Case, supra note 10, at 1401.
12 Id. at 1402.
13 For discussion of those conditions, see infra Part I.
14 Waldron, The Core Case, supra note 10, at 1360.
15 See id. at 1367, 1385–86 & n.110.
16 See id. at 1364–69.
2008] AN UNEASY CASE FOR JUDICIAL REVIEW 1697
lenge judicial review of action by executive and bureaucratic officials,17
who may often have less opportunity for debate and deliberation
before they must act in ways that might put rights at risk.) When
these conditions are met, Waldron maintains, the fact that courts and
legislatures differ in their assessments provides no ground for thinking
courts more likely to be right.
This much of Waldron’s argument I am inclined to accept, at least
for the sake of argument. Many if not most arguments that courts
should be presumed better than legislatures at determining whether
legislation violates individual rights have a troublingly elitist cast —
especially if one follows Waldron in assuming that the kinds of rights
commonly incorporated into bills of rights are moral rights. Virtually
without exception, judges and Justices are well-educated members of
the upper or upper-middle classes who have been socialized to accept
professional norms.18 The preference for having a small number of
lawyers in robes resolve contested questions about individual rights
almost inevitably reflects one or another species of antipopulism,19 frequently
coupled with highly idealized portraits of the few who wield
judicial power.
In expressing unease about the elitism implicit in traditional defenses
of judicial review, I do not mean to suggest that courts have no
institutional advantages in resolving disputed rights questions.20 Perhaps
most important, processes of judicial review can be and often are
structured to permit courts to address challenges to the constitutionality
of legislation only after potentially unforeseen implications have
manifested themselves. Even in cases in which the passage of time
and the accrual of further information help, however, judicial decisions
could be made subject to legislative override, with the effect that this
particular judicial advantage would largely disappear.21
Moreover, the judicial branch may labor under some relative disadvantages
too — at least if we assume that judicial review will be
practiced more or less as it traditionally has been. As Waldron points
out, courts have an understandable tendency to confront issues that
–––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
17 See id. at 1353–54.
18 See John Ferejohn, Independent Judges, Dependent Judiciary: Explaining Judicial Independence,
72 S. CAL. L. REV. 353, 369 (1999) (noting that judges “are likely to bring to their work
the perceptions of an upper middle class, educated, largely male, and largely white elite”); see also
Paul Brest, Who Decides?, 58 S. CAL. L. REV. 661, 664 (1985).
19 See KRAMER, supra note 1, at 241–46; RICHARD D. PARKER, “HERE, THE PEOPLE
RULE” 56–58 (1994).
20 Waldron does not maintain the contrary. See Waldron, The Core Case, supra note 10, at
1376.
21 Waldron acknowledges that courts may have an initial advantage in assessing as-applied
challenges to legislation, but he apparently believes that the balance of argument supports only a
“weak” form of judicial review in which a court may declare that important issues of rights are at
stake, while leaving responsibility for resolving those issues with the legislature. Id. at 1370.
1698 HARVARD LAW REVIEW [Vol. 121:1693
are fundamentally moral in legalistic, logic-chopping terms in which
head-on engagement with the contending moral considerations often is
almost wholly washed out.22 Roe v. Wade23 stands as Exhibit One for
this proposition,24 but other examples abound.
In response to claims that courts often grapple clumsily with appeals
to moral principle, some would undoubtedly reject Waldron’s
premise that the kinds of rights included in bills of rights typically are
moral rights, not historically bounded references to be interpreted in
accord with the original understanding of constitutional language. But
he advances a general philosophical argument, not one about the merits
of originalism as an American constitutional theory, and his challenge
deserves to be answered on its own terms. In any event, constitutional
adjudication in the United States has enough nonoriginalist
elements that Waldron’s arguments bear on U.S. practice and on
whether that practice ought to be reformed even if — as I agree —
U.S. courts do not regard constitutional rights entirely as moral rights.
Having dismissed the outcome-based case in favor of judicial review,
Waldron advances “process-based” arguments purporting to establish
that judicial review is unfair and “illegitimate” in any society in
which his four conditions hold.25 In essence, he claims that judicial
review deprives political majorities of the right to democratic selfgovernance
on important issues. This process-based objection might
possibly be overcome, he acknowledges, if courts were likely to be better
than legislatures at resolving disputed questions about moral
rights.26 But having adjudged the outcome-related case for judicial
review to be “at best inconclusive”27 in societies defined by his four assumptions,
Waldron concludes that the process-based or legitimacy argument
in favor of majority decision by legislatures ought to prevail
and that judicial review should be abolished.28 If judicial review could
ever be defended successfully, he argues, it would need to be as a corrective
for “pathologies” in a society’s culture or political system.29
Waldron is a distinguished political theorist. He has no partisan ax
to grind. His claims have achieved influence already, and they will
undoubtedly prove influential in the future.
–––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
22 See id. at 1383–85.
23 410 U.S. 113 (1973).
24 See Waldron, The Core Case, supra note 10, at 1383 (noting that the fifty-page opinion in
Roe devoted only a few paragraphs to the moral issues at stake and that “the ‘reasoning’ is
thread-bare”).
25 See id. at 1386–95.
26 See id. at 1375 (acknowledging the “Gordian knot” that would exist if outcome-related reasons
supported, while process-related reasons opposed, judicial review).
27 Id.
28 See id. at 1375–76.
29 Id. at 1406; see id. at 1352, 1401–05.
2008] AN UNEASY CASE FOR JUDICIAL REVIEW 1699
As I shall demonstrate, however, a fallacy pervades The Core of the
Case Against Judicial Review. Even if courts would be no more likely
than the legislature to decide questions of individual rights correctly in
a society in which legislators took rights seriously, Waldron’s conclusion
that outcome-based reasons fail to support judicial review would
not follow. As I have suggested, his argument would not go through if
some rights deserve to be protected by multiple safeguards or veto
powers. If errors of underprotection — that is, infringements of rights
— are more morally serious than errors of overprotection, and if a few
other plausible conditions obtain, then there could be outcome-related
reasons to prefer a system with judicial review to one without it.
Waldron’s process-based argument also collapses insofar as there
are outcome-related reasons to support judicial review. The fairness of
procedures depends crucially on the ends that they seek to accomplish.
If judicial review promotes morally better outcomes than would exclusive
legislative definitions of disputed rights, then reliance on judicial
review is not unfair, nor does it, as Waldron maintains, necessarily lack
“legitimacy.”30
As Waldron uses the term, a political regime is legitimate insofar as
its design and composition provide good reasons for those who disagree
with particular laws and decisions nevertheless to respect and
obey them.31 In arguing against judicial review, Waldron frequently
equates “political legitimacy” with “democratic legitimacy.” When decisions
emerge from processes in which everyone has a voice and a
vote, he argues, even the losers have reason to accede to the outcome
on the ground that their views were fairly taken into account.32 Although
Waldron is correct that democratic legitimacy is important, he
fails to acknowledge that political legitimacy — which is a broader
concept — can have multiple sources. Even if judicial review is relatively
lacking in specifically democratic legitimacy, one good reason for
citizens to respect political decisions with which they disagree is that
those decisions issue from institutions that are well designed to safeguard
individual rights. If judicial review reduces the likelihood that
important rights will be infringed, then it may actually enhance, rather
than undermine, a governmental regime’s overall political legitimacy.
Although my arguments against Waldron add up to an affirmative
case in favor of judicial review, my defense is sufficiently qualified and
contingent to qualify as “uneasy.” I argue that if certain conditions exist
— as I am inclined to believe that they do and as Waldron does not
demonstrate to be impossible or even unlikely — then judicial review
–––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
30 See id. at 1386–93.
31 See id. at 1386–87.
32 See id. at 1386–89.
1700 HARVARD LAW REVIEW [Vol. 121:1693
would be desirable. Among the most centrally important premises
supporting the case for judicial review are these: (1) Even if courts are
not better overall at identifying rights violations than are legislatures,
courts have a distinctive perspective that makes them more likely than
legislatures to apprehend serious risks of rights violations in some
kinds of cases. (2) Legislative action is more likely to violate fundamental
rights than legislative inaction. (3) Some rights are more important
than others and, accordingly, are more deserving of protections
against infringement. (4) A system of judicial review can be so designed
that the moral costs of such overenforcement of rights as judicial
review would produce will likely be lower than the moral costs
that would result from such underenforcement of rights as would occur
in the absence of judicial review.
Although I would guess that all of these premises are satisfied in
most western liberal democracies, I do not attempt to prove that they
actually hold, either in the United States or elsewhere — a job that
would be Herculean at the very least. In this Article it will be enough
if I can, first, develop a case for judicial review that does not depend
on the premise that courts are more likely than legislatures to define
rights correctly and, second, call attention to the kinds of assumptions
on which I believe that any argument either for or against judicial review
needs to depend. I shall also, near the end, venture some
thoughts on how the case for judicial review that I make in this Article
would influence the form that judicial review ought to take.
The Article unfolds as follows. Part I lays out Waldron’s argument
that the balance of outcome-related reasons provides inadequate support
to justify judicial review in nonpathological societies. Part II then
criticizes Waldron’s reasoning about the balance of outcome-related
considerations and, in the course of doing so, develops an affirmative
case for judicial review. Assuming for the sake of argument that judicial
review would not increase the number of cases in which questions
of rights were decided correctly, Part II supports judicial review as a
means of minimizing the number of cases in which underenforcement
occurs. Part III argues, contra Waldron, that judicial review can actually
contribute to the political legitimacy of a scheme of otherwise democratic
government when the demands of political legitimacy are
understood correctly. Part IV offers a brief, tentative exploration of
how judicial review ought to be practiced if courts were not presumed
better than legislatures at identifying rights, but viewed instead as
providing a potentially costly but also potentially valuable veto power
over rights-threatening legislation. This Part argues that courts should
exercise only the most deferential review or should even withhold review
altogether in cases in which the legislature conscientiously seeks
to protect one set of fundamental rights without plausibly threatening
another (as may frequently happen in the United States when Con-
2008] AN UNEASY CASE FOR JUDICIAL REVIEW 1701
gress legislates pursuant to Section 5 of the Fourteenth Amendment).
Part V supplies a brief conclusion.
I. WALDRON’S ARGUMENT THAT OUTCOME-RELATED REASONS
ARE INADEQUATE TO SUPPORT JUDICIAL REVIEW
In mounting his case against judicial review, Waldron makes four
central assumptions. If these assumptions hold, he argues, then the
case for judicial review will fail.33 Although I shall assume for purposes
of argument that all of Waldron’s assumptions obtain, it is important
to state them clearly in order to understand their role in his
“core case.” It is also important to understand how Waldron’s carefully
stated assumptions relate to his other arguments that outcomebased
reasons provide no clear support for judicial review.
A. Waldron’s Assumptions
First, Waldron assumes that the society’s democratic institutions,
including its legislature, are in “reasonably good working order.”34 In
other words, the society has tolerably good even if not perfect institutions
of political democracy.
Second, the society has a relatively well-functioning judicial system
capable of applying the law in an impartial even if not perfect way.35
Third, the society believes in and is committed to protecting individual
rights.36 There are few if any skeptics who regard the idea of
moral rights as nonsensical. Rather, the society, or most of its members,
believe that in principle questions involving claims of individual
rights have correct answers.
Fourth, despite the society’s commitment to the idea of individual
rights, widespread, reasonable disagreement exists about what rights
people have.37 For example, people disagree about whether the right
to religious freedom encompasses a right to be excused from obedience
to otherwise generally applicable laws, about whether freedom of
speech subsumes a right to donate vast sums of speech-generating
money to political campaigns, and about whether there is a right to
abortion.
B. Which Assumptions Do What Work?
Given his four assumptions, Waldron believes courts no more likely
than legislatures to arrive at the correct answer to questions of what
–––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
33 Id. at 1360.
34 Id.; see id. at 1361–62.
35 See id. at 1363–64.
36 See id. at 1364–66.
37 See id. at 1366–69.
1702 HARVARD LAW REVIEW [Vol. 121:1693
rights people have. As I have stated already, I am prepared to accept
for purposes of argument Waldron’s narrow conclusion that courts
would be no more likely than legislatures to identify rights correctly if
all of his assumptions held. (This, obviously, is a very big “if.” I mean
to leave open the question whether particular legislatures are in “reasonably
good working order” with respect to their solicitude for rights.)
Nevertheless, it is worth parsing Waldron’s argument with care, for he
fails to make explicit which of the premises carry which parts of the
argumentative burden that The Core of the Case Against Judicial Review
assumes.
On one possible interpretation, Waldron’s fourth assumption, involving
reasonable disagreement about the content of rights, does most
or all of the work of establishing that courts could not be better than
legislatures at deciding correctly what rights people have. If people
reasonably disagree about rights, we might interpret Waldron as arguing,
then we must explain not only how they disagree, but also how it
is reasonable for them to disagree; and if disagreement is reasonable
because no epistemically reliable method of identifying rights exists, as
he may seem to suggest,38 then it is impossible to say that one decisionmaker
is more likely than another to decide correctly. In short,
once the reasonableness of disagreement is acknowledged, it follows
inexorably that there can be no outcome-related reason to support judicial
review.
Although it is possible to imagine an opponent of judicial review
relying on the assumption of reasonable disagreement in this way, I do
not understand Waldron to do so in The Core of the Case Against Judicial
Review.39 Waldron states his four assumptions early on, but
when he later considers and rejects three arguments that courts might
be better than legislatures at identifying rights — that courts are likely
to do better (1) because courts focus their deliberations on concrete
cases, not abstractions, (2) because courts have special competence in
interpreting legal texts such as bills of rights, and (3) because courts
carry an obligation to provide reasoned explanations of their decisions40
— he nowhere suggests that reasonable disagreement suffices to
defeat these arguments. Rather, assuming that questions involving
rights have correct answers, he maintains that claims for courts’ comparatively
greater competence fail either because courts do not in fact
–––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
38 See id. at 1368 (assuming “that our condition is not one in which the truth of the matter
[about moral questions] discloses itself in ways that are not reasonably deniable”).
39 This may be Waldron’s position in other writing. See, e.g., Jeremy Waldron, The Irrelevance
of Moral Objectivity, in NATURAL LAW THEORY 158 (Robert P. George ed., 1992).
40 See Waldron, The Core Case, supra note 10, at 1379–86. See also Cross, supra note 4, at
1536–50 (criticizing these traditional defenses of judicial review).
2008] AN UNEASY CASE FOR JUDICIAL REVIEW 1703
possess specifically asserted advantages or, alternatively, because any
judicial advantage is offset by a judicial disadvantage.41
Nor do I believe that the mere fact of reasonable disagreement
could persuade anyone who was not a rights-skeptic of the impossibility
of one institution’s being better than another at getting the right
answer to questions involving individual rights.42 For me to believe
that reasonable disagreement about rights is possible, while continuing
to believe in a nonskeptical way that people nevertheless possess particular
rights, I have to believe that I have identified the pertinent
moral truth in a relatively, even if not perfectly, epistemically reliable
way. If, for example, I am a utilitarian, then I must believe in the relative
reliability of the processes of reasoning that led me to adopt utilitarianism
in preference to Kantianism and other rival theories — even
if many reasonable people disagree in light of what John Rawls called
the “burdens of judgment” that inescapably attend this conclusion.43
Similarly, if I think that the facts are pertinent to some or all moral
judgments, yet believe nonskeptically in the validity of some claims of
moral right, then I must think that there is a relatively epistemically
reliable way of determining the relevant facts.
It lies far beyond my ambition in this Article to identify which
moral theory, if any, is correct, or to determine how any kind of fact
would most reliably be determined. My only point here is that acknowledgment
of reasonable disagreement does not preclude reasoned
judgments about what is right and wrong or about how rights are
most reliably identified. And from this point, another follows: If people
can reasonably believe not only that there are truths about rights,
but also that there are relatively epistemically reliable means of discovering
those truths, then the fact of reasonable disagreement cannot,
by itself, prove that one institution could not be better than another at
reaching correct outcomes. If, for example, courts routinely reasoned
from better supported moral premises or used more reliable delibera-
–––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
41 Among the advantages of legislatures, he includes their tendency to focus directly on underlying
moral considerations, see Waldron, The Core Case, supra note 10, at 1384–85, whereas courts
become preoccupied with issues of textual exegesis and interpretive methodology, see id. at 1381–
82.
42 Waldron’s core case against judicial review involves societies in which the rights-skeptical
position is merely “an outlier.” Id. at 1365.
43 JOHN RAWLS, POLITICAL LIBERALISM 54–58 (1993). Rawls used the term to refer to “the
many hazards involved in the correct (and conscientious) exercise of our powers of reason and
judgment,” id. at 56, which give rise to “reasonable disagreement,” id. at 55, and explain why “it
is not to be expected that conscientious persons with full powers of reason, even after free discussion,
will all arrive at the same conclusion,” id. at 58. To say that a decision is subject to the burdens
of judgment is not to say that there can be no right answer, see id. (noting that some reasonably
disputable judgments “may be true”), but it is to acknowledge that knock-down
arguments cannot always be expected.
1704 HARVARD LAW REVIEW [Vol. 121:1693
tive procedures than did legislatures, then courts could be expected to
reach more accurate conclusions.44
On a better interpretation of Waldron’s argument, his first and
third assumptions — that legislative institutions are in reasonably
good order and that the society and its members, including its legislators,
are seriously committed to protecting individual rights — must do
the principal work of supporting his conclusion that there is no sufficient
outcome-based reason to want judicial review. On this interpretation,
the key point is that judges have no greater capacity for moral
insight, nor do they follow processes of deliberation more likely to lead
them to insight, than morally conscientious legislators.
II. PREFERRED RIGHTS AND OUTCOME-RELATED
REASONS TO SUPPORT JUDICIAL REVIEW
As I have said, I shall stipulate for purposes of argument that if
Waldron’s first and third assumptions were to hold, and that if the sole
question were whether courts or legislatures are more likely to identify
rights correctly, there would not be sufficient outcome-based reasons to
adjudge courts conclusively better than legislatures.45 Crucially, however,
whether courts or legislatures will do better at identifying moral
rights correctly is not the sole question to be resolved in arguing for or
against judicial review. Once an alternative question is put on the table,
involving whether one or both institutions should be charged with
rights protection, Waldron’s argument loses its potency, and a powerful
variant on the traditional case for judicial review emerges.
A. Distinguishing Errors of Underenforcement
and Overenforcement of Individual Rights
Although judicial review permits courts to invalidate legislation, it
ordinarily imposes no impediment to the legislature’s respecting or
protecting individual rights by refraining from actions that it thinks
would violate them. In so saying, I put to one side the well-known argument
of James Bradley Thayer that the availability of judicial review
predictably diminishes legislatures’ willingness to deliberate seriously
and independently about questions involving rights.46 Within
the bounds of argument that Waldron’s premises define, I take
–––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
44 I do not mean to suggest that only moral realists, who believe in objectively right answers to
moral questions, could have good reasons to support judicial review. As I shall explain below,
relativists or pragmatists could also be warranted in favoring judicial review.
45 My actual view is that courts would be more likely to reach correct decisions at least about
whether general legislation violates individual rights as applied to particular cases, some of which
the legislature may not have anticipated.
46 See James B. Thayer, The Origin and Scope of the American Doctrine of Constitutional
Law, 7 HARV. L. REV. 129, 155–56 (1893).
2008] AN UNEASY CASE FOR JUDICIAL REVIEW 1705
Thayer’s view to be ruled out by the assumption that a society’s democratic
institutions, including its legislature, take rights seriously.
Quite apart from argumentative stipulations, it would seem to me to
be dramatically imprudent for a society that thought its legislature did
not currently take rights seriously to abolish judicial review in hopeful
anticipation that the legislature would thereafter change its ways.47
In any event, if we assume that the prospect of judicial review creates
no ex ante impediment to a legislature’s respecting individual
rights, the question to be addressed in assessing judicial review is not
necessarily whether courts or legislatures would perform better at
identifying rights if we had to charge only one institution with this
task. The question, instead, is whether a society might have good outcome-related
reasons to enlist both the legislature and the courts in
protecting individual rights.48 Or, to put the question slightly differently,
might a society reasonably want to create multiple veto points so
that governmental action could not occur if either a court or the legislature
thought that the action would violate individual rights?49
–––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
47 But cf. TUSHNET, supra note 2, at 66 (acknowledging that concern for the Constitution currently
ranks low among congressional concerns and asserting that “this situation may have arisen
in part because of judicial review”); id. at 194 (suggesting that “[p]erhaps it is time” to abolish judicial
review).
48 Cf. William N. Eskridge, Jr. & John Ferejohn, Virtual Logrolling: How the Court, Congress,
and the States Multiply Rights, 68 S. CAL. L. REV. 1545, 1547 (1995) (arguing that “separation of
powers, federalism, and judicial review foster a lawmaking regime that produces multiple opportunities
for rights creation and fewer opportunities for rights negation,” which results in “the multiplication
of rights over time”); Douglas Laycock, Federalism as a Structural Threat to Liberty,
22 HARV. J.L. & PUB. POL’Y 67, 75 (1998) (describing the Fourteenth Amendment’s “multiple
enforcement mechanisms,” including the judiciary, Congress, the Executive, and the states, and
the possibility that if any one branch of government took action, “the Fourteenth Amendment
would be safe”).
49 See Cross, supra note 4, at 1576 (explaining that “the multiple vetoes concept relies on the
benefit of adding judicial review on top of congressional and executive action”). Other scholars
have argued that regardless of whether courts are more likely to identify rights correctly than are
legislatures, judicial review might be defended on the ground that it creates an additional barrier
to the enactment or enforcement of legislation. See, e.g., Richard A. Epstein, The Independence of
Judges: The Uses and Limitations of Public Choice Theory, 1990 BYU L. REV. 827, 846 (arguing
that views on judicial independence may depend on what people think about “the number of hurdles
that must be crossed before the state can impose its commands on individual citizens”); Julia
D. Mahoney, Kelo’s Legacy: Eminent Domain and the Future of Property Rights, 2005 SUP. CT.
REV. 103, 129–31 (arguing that “even if judges have inferior ability to calculate the social costs
and benefits of eminent domain,” judicial review can “add value” by providing “multiple veto
points in rearranging property rights”); Kermit Roosevelt III, Constitutional Calcification: How
the Law Becomes What the Court Does, 91 VA. L. REV. 1649, 1704 & n.182 (2005) (arguing that
the “chief value of judicial review” in cases in which the court may not be better than the legislature
at balancing costs and benefits is “that the judiciary may serve as a second negative, and one
less susceptible to the temporary excesses of popular sentiment that can infect legislatures”).
1706 HARVARD LAW REVIEW [Vol. 121:1693
Systems of “strong” judicial review, which Waldron inveighs
against,50 institutionalize an affirmative answer to this question. In
doing so, however, they need not depend on the premise that a court is
more likely than the legislature to rule correctly on whether any particular
claimant truly possesses a right. Rather, a stronger case for judicial
review in morally and politically nonpathological societies rests
on the assumption that if either a court or the legislature believes that
an action would infringe individual rights, the government should be
barred from taking it.
Two analogies suggest the force of this argument. One, to which I
have alluded already, comes from the jury system and the demand
for unanimous verdicts in federal criminal cases.51 The reason to require
unanimous verdicts — as, additionally, to apply a beyond-areasonable-doubt
standard of proof52 — is not that this requirement
will maximize the number of correct determinations of whether accused
persons committed alleged offenses. For that purpose, a decision
by majority vote coupled with a lower standard of proof would serve
better. Nevertheless, we may reasonably believe that we have a good
outcome-related reason to give every juror a veto over what otherwise
would be a guilty verdict and, what is more, to ask each to exercise
her veto power unless satisfied that the government has proven the defendant
guilty beyond a reasonable doubt. The reason, obviously, is
that errors that result in the conviction of the innocent are more morally
disturbing than errors that result in acquittals of the guilty.53 In
light of that assessment, we have adopted a system that minimizes the
most morally grievous errors, even if that system leads to more of the
less grievous errors, and indeed to more total errors, than would an alternative.
Just as there may be outcome-related reasons to prefer a
system that minimizes erroneous convictions of the innocent, so there
may be outcome-related reasons to prefer a system that minimizes violations
of individual rights — even at the cost of increasing the number
of cases in which individual rights claims are mistakenly upheld.54
An even closer analogy involves the multipart system of lawmaking
created by the United States Constitution, which has parallels in the
constitutions of some other nations. As a matter of structural design,
–––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
50 See Waldron, The Core Case, supra note 10, at 1354–59 (distinguishing “strong” judicial review,
which authorizes courts to nullify or modify statutes, from “weak” judicial review, under
which courts can signal their concerns that statutes violate rights but cannot decline to enforce
statutes on that ground).
51 See supra pp. 1695–96.
52 See In re Winship, 397 U.S. 358, 368 (1970).
53 See LAUDAN, supra note 7, at 63 (collecting authorities that so assert).
54 See Roosevelt, supra note 49, at 1661 (explaining that courts may adopt decision rules in an
attempt to reduce certain types of errors, rather than to “minimize the total number of errors
without reference to kind”).
2008] AN UNEASY CASE FOR JUDICIAL REVIEW 1707
the Constitution makes it difficult for majorities to legislate. Separate
majorities of both houses of a bicameral legislature are required, and
the President also possesses a veto power.55 The underlying premise,
plainly, is that it is presumptively worse for legislation to be enacted
than not enacted, largely because of the threat that legislation might
violate individual rights, and that multiple veto points should therefore
exist.56
–––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
55 See U.S. CONST. art. I, § 7.
56 See, e.g., THE FEDERALIST NO. 51, at 322 (James Madison) (Clinton Rossiter ed., 1961)
(“Ambition must be made to counteract ambition.”); Samuel Freeman, Constitutional Democracy
and the Legitimacy of Judicial Review, 9 LAW & PHIL. 327, 352–55 (1991).
A third analogy — on which I am inclined to place less weight, because it may tend to presuppose
(rather than help to establish) the desirability of judicial review — involves a number of judicially
crafted rules or doctrines embodying the premise that it is better for constitutional rights
to be overenforced than for them to be underenforced. Among the best known of these is the rule
of Miranda v. Arizona, 384 U.S. 436 (1966), under which a confession may not be introduced in a
criminal trial unless the defendant was apprised
that he has the right to remain silent, that anything he says can be used against him in a
court of law, that he has a right to the presence of an attorney, and that if he cannot afford
an attorney one will be appointed for him prior to any questioning if he so desires.
Id. at 479. Miranda overenforces the Fifth Amendment’s prohibition against compelled selfincrimination,
for not every confession elicited in the absence of a Miranda warning would necessarily
be coerced. But the Supreme Court evidently determined that it was too difficult for courts
to identify impermissible coercion on a case-by-case basis, and it opted for overenforcement in
preference to the underenforcement that was likely to result if the burden rested on criminal defendants
to prove that their wills were overborne.
A preference for rights to be overenforced rather than underenforced also finds expression in
First Amendment cases affirming that “[f]reedoms of expression require breathing space.” Hustler
Magazine, Inc. v. Falwell, 485 U.S. 46, 52 (1988) (quoting Phila. Newspapers, Inc. v. Hepps, 475
U.S. 767, 772 (1986) (internal quotation marks omitted)). Under the First Amendment overbreadth
doctrine, for example, courts will facially invalidate statutes that apply to some constitutionally
unprotected speech, such as obscenity or fighting words, but that also extend to some protected
speech and are “substantially overbroad.” See generally Richard H. Fallon, Jr., Making
Sense of Overbreadth, 100 YALE L.J. 853 (1991). When statutes are overbroad, courts could continue
to enforce them against unprotected speech and hold them unconstitutional only as applied
to speech that is constitutionally protected. But the Supreme Court has said that First Amendment
rights are so precious that facial invalidation is appropriate to avert the “chilling” of protected
speech, see, e.g., Bd. of Airport Comm’rs v. Jews for Jesus, Inc., 482 U.S. 569, 576 (1987),
even though an overenforcement of speech rights occurs as a result. For discussion of other doctrines
that deliberately “overprotect” rights based on a calculation that it is better for them to be
overenforced than underenforced, see, for example, Richard H. Fallon, Jr., Judicially Manageable
Standards and Constitutional Meaning, 119 HARV. L. REV. 1275, 1303–06 (2006), and see also
Mitchell N. Berman, Constitutional Decision Rules, 90 VA. L. REV. 1 (2004).
If courts sometimes overenforce constitutional rights, it is at least equally true and important
that they sometimes underenforce constitutional rights as well. See, e.g., Lawrence Gene Sager,
Fair Measure: The Legal Status of Underenforced Constitutional Norms, 91 HARV. L. REV. 1212
(1978); see also Fallon, Judicially Manageable Standards and Constitutional Meaning, supra, at
1299–303. Even when constitutional norms are judicially underenforced, however, the availability
of even limited judicial enforcement will presumably result in fewer cases of rights being infringed
than if sole responsibility for defining and enforcing rights were confided in the
legislature.
1708 HARVARD LAW REVIEW [Vol. 121:1693
It bears emphasis that one need not be a moral realist in order to
see the force of these analogies or to accept the argument that they
support. Relativists and skeptics can, and frequently do, retain firstorder
moral views — personal convictions about right and wrong on
the basis of which they are prepared to act — with their relativism or
skepticism pertaining only to second-order questions involving the
foundations for those views.57 It is therefore wholly possible for a second-order
relativist or skeptic to believe, as a matter of first-order
conviction, that it is more morally troublesome for rights to be underenforced
than overenforced and to support judicial review based on
this judgment — just as it is possible, and indeed presumably familiar,
for second-order relativists to endorse the requirements of unanimous
jury verdicts and proof beyond a reasonable doubt.
Although it is reasonable in principle to believe it better to err on
the side of too much rather than too little protection of rights under
conditions of reasonable disagreement, an argument that supports judicial
review on this basis obviously risks proving too much. If multiple
vetoes are good, why stop with the legislature, the President, and
the courts? Why not establish other institutions with veto powers or
insist on unanimous consent before any legislation can be enacted?
Any good answer to questions such as these must have two related
parts. First, a sensible balance needs to be struck. Any scheme that
relies on fallible human beings to identify and enforce moral rights will
bring hazards with it, including the costs of errors of over- and underenforcement.58
Under these circumstances, no sound approach could
wholly discount the costs of errors in one or the other direction.59 Although
most of us think it worse to convict one innocent defendant
than to let three or five or perhaps nine guilty persons go free, we do
not structure the criminal process on the assumption that it would be
better to let thousands escape accountability than to risk ever punishing
a single innocent.
Second, in striving for a reasonable balance, the strategy of assigning
vetoes to multiple institutions makes most sense insofar as the recipient
institutions possess distinctive perspectives tending to make
them more sensitive than others to some morally pertinent considerations.60
A reason to give the President a veto power over legislative
–––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
57 See generally THOMAS NAGEL, THE LAST WORD (1997); Ronald Dworkin, Objectivity
and Truth: You’d Better Believe It, 25 PHIL. & PUB. AFF. 87 (1996).
58 See ADRIAN VERMEULE, JUDGING UNDER UNCERTAINTY: AN INSTITUTIONAL THEORY
OF LEGAL INTERPRETATION 255–76 (2006) (categorizing pertinent costs).
59 But cf. Cross, supra note 4, at 1577–78 (arguing that “even a wildly incompetent court
would have a constitutional benefit as a backstop to screen out unconstitutional legislation”).
60 My argument in this paragraph parallels the argument for federal habeas corpus review of
state criminal convictions advanced in Robert M. Cover & T. Alexander Aleinikoff, Dialectical
Federalism: Habeas Corpus and the Court, 86 YALE L.J. 1035 (1977). The authors defend federal
2008] AN UNEASY CASE FOR JUDICIAL REVIEW 1709
action is that the executive and legislative branches will likely have
different, potentially valuable vantage points from which to assess
competing interests and values. Similarly, a reason to give courts a
veto power is that courts are likely to have a perspective that may
make them more sensitive than legislatures to some possible rights violations
even if the evidence on whether that perspective is better overall
may indeed be “inconclusive.”61 Perhaps most obviously, courts
typically decide cases upon concrete facts, some of which even highly
competent legislators may not have foreseen. Furthermore, a distinguishing
feature of judges’ professional training and mission involves a
solicitude for rights as they have historically been understood. Historic
understandings may of course have been wrong in some, even many,
instances.62 Nevertheless, judges’ professionally ingrained instincts
and processes of judgment are likely to differ from those of legislators
and to be better adapted to reflecting such imperfect wisdom about the
content of rights as our legal tradition embodies.63 In the context of
historic understandings of and anxieties about judicial power, courts
can also be asked and expected to discharge their reviewing functions
with reasonable restraint.
This, then, is the core of the strongest case for judicial review in the
kind of nonpathological society with which both Waldron and I are
concerned: errors that result in the underenforcement of rights are
more troubling than errors that result in their overenforcement, and
judicial review may provide a distinctively valuable hedge against errors
of underenforcement.
–––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
review not on the ground that federal courts are more likely than state courts to define constitutional
rights correctly, but on the basis that jurisdictional redundancy is especially desirable because
state and federal courts are likely to have different perspectives on how rights ought to be
defined. Id. at 1046–54.
61 Although “there are a large number of obviously nontotalitarian societies — Great Britain,
New Zealand, the Netherlands, Sweden, and France, for example — that survive quite nicely
without” robust, American-style judicial review, those countries tend to be less protective than the
United States of claimed rights involving “criminal procedure, freedom of the press when it is irresponsible,
freedom of speech for the truly evil (Nazis, Klansmen, and child pornographers, for
example), and a strong separation between church and state.” Frederick Schauer, Judicial Supremacy
and the Modest Constitution, 92 CAL. L. REV. 1045, 1066 n.100 (2004). According to
Schauer, “[i]t may be wrong for the United States to be so divergent, but setting these countries
out as models suggests a willingness to accept fewer defendant’s rights, free speech rights, free
press rights, and separation of church and state rights than exist in the United States.” Id.; see
also Frank B. Cross, The Relevance of Law in Human Rights Protection, 19 INT’L REV. L. &
ECON. 87, 92–93 (1999) (concluding on the basis of empirical data that “[j]udicial independence”
is “significantly associated with greater political freedom, suggesting a prominent role for the law
and courts in the protection of freedom”).
62 See generally Adrian Vermeule, Common Law Constitutionalism and the Limits of Reason,
107 COLUM. L. REV. 1482 (2007) (skeptically probing arguments that the historic survival of
common law decision rules attests to their likely wisdom).
63 See generally David A. Strauss, Common Law Constitutional Interpretation, 63 U. CHI. L.
REV. 877 (1996) (defending judicial review partly on Burkean grounds).
1710 HARVARD LAW REVIEW [Vol. 121:1693
B. The Limits of the Outcome-Based Case for Judicial Review:
Contestable Premises and the Burdens of Judgment
Having now laid out a basis for supporting judicial review, I should
clarify the limits of my argument. Taking seriously Waldron’s claim to
have mounted a persuasive affirmative case against judicial review, I
have maintained that his argument is insufficient to dislodge reasonable
beliefs that there are good outcome-related reasons to support judicial
review. But I have meant to hedge my argument with appeals
to what it would be reasonable to believe, rather than claiming
straightforwardly to have demonstrated the desirability of judicial review,
because my argument depends on a number of unproven assumptions,
some of which may be only implicit in what I have said so
far.
Although it would mark a great advance if I could now make a
clinching demonstration of the validity of the assumptions on which
my argument rests, I fear that I must leave matters of proof more or
less where they stand. It is important, however, that I should identify
the central moral, conceptual, and empirical premises on which an argument
such as mine must ultimately depend.
First, I have assumed that even if courts cannot be shown to be
better than legislatures at resolving disputed rights questions, courts
are likely to have a distinctive perspective, involving both a focus on
particular facts and a sensitivity to historical understandings of the
scope of certain rights, that would heighten their sensitivity to some
actual or reasonably arguable violations that legislatures would fail to
apprehend. Although this premise is obviously contestable, even critics
of judicial review do not generally deny that courts have some advantages
in resolving rights claims. Waldron, for example, argues only
that the case for thinking courts more likely than legislatures to resolve
rights disputes correctly is ultimately “inconclusive,” not that it is
clearly insupportable.64
Second, I have assumed that legislative action is more likely to violate
fundamental rights than is legislative inaction. This is a commonly
held assumption that underlies a variety of familiar governmental
structures — such as bicameralism requirements and provisions for
presidential vetoes — that are designed to stop political majorities
from enacting their preferences as readily as possible into law. Indeed,
Waldron himself speaks favorably of bicameralism as a well-advised
check against “legislative pathologies”65 (notwithstanding the obvious
possibility of reasonable disagreement over which popular dispositions
count as pathological).
–––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
64 Waldron, The Core Case, supra note 10, at 1375.
65 Id. at 1361 & n.47.
2008] AN UNEASY CASE FOR JUDICIAL REVIEW 1711
Nevertheless, the assumption that governmental action is more
likely to violate rights than governmental inaction is certainly open to
challenge. One might contest it on the ground that a well-designed
constitution would include a substantial number of “positive” rights to
governmental assistance.66 Or a challenger might maintain that the
realization of many important moral rights requires affirmative governmental
action even if the relevant rights are not formally denominated
in a bill of rights. Moreover, even if positive rights are put to
one side, a risk certainly exists that judicial review could sometimes
frustrate legislative efforts to protect fundamental rights, depending on
how constitutions are first written and then construed.
Dred Scott v. Sandford,67 in which the Supreme Court ruled that
Congress could not forbid slavery in the federal territories, offers a
case in point. Waldron cites a different example, Lochner v. New
York,68 in which the Supreme Court invalidated a statute setting
maximum working hours for bakers. Although I am uncertain
whether Lochner should be regarded as a case in which the courts
stymied legislative efforts to protect fundamental individual rights,
rather than simply thwarted the legislature from implementing a humane
policy, the example deserves attention. In a study of the recent
introduction of bills of rights and judicial review into Israel, Canada,
South Africa, and New Zealand, Ran Hirschl concludes that whereas
those countries’ constitutional courts have “fortif[ied] and expand[ed]
the boundaries of the private sphere in the context of freedom of expression
and religion, freedom of movement, the right to privacy (including
reproductive freedom), and formal equality,”69 they have “utterly
failed to promote progressive or egalitarian notions of distributive
justice in a meaningful way.”70 In some instances, Hirschl writes,
courts enforcing bills of rights have actually “shield[ed] the economic
sphere from attempts to reduce socioeconomic disparity through regulatory
and redistributive means.”71
–––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
66 See, e.g., Mark Tushnet, Social Welfare Rights and the Forms of Judicial Review, 82 TEX.
L. REV. 1895, 1919 (2004) (describing positive rights recognized in foreign constitutions and arguing
that “[t]he experience of constitutional courts around the world suggests that scholars of U.S.
constitutional law would do well to rethink the conventional wisdom about the inclusion of social
welfare rights in judicially enforceable constitutions”). But cf. Cass R. Sunstein, Against Positive
Rights, in WESTERN RIGHTS?: POST-COMMUNIST APPLICATION 225, 225 (András Sajó ed.,
1996) (asserting that it is “a large mistake, possibly a disaster” to include positive rights in constitutions
in countries transitioning from communism to a market economy).
67 60 U.S. 393 (1857).
68 198 U.S. 45 (1905); see Waldron, The Core Case, supra note 10, at 1348 (making this
reference).
69 RAN HIRSCHL, TOWARDS JURISTOCRACY: THE ORIGINS AND CONSEQUENCES OF
THE NEW CONSTITUTIONALISM 118 (2004).
70 Id. at 14.
71 Id. at 218.
1712 HARVARD LAW REVIEW [Vol. 121:1693
The lesson that emerges from Dred Scott, Lochner, and Hirschl’s
study — and, some would say, from recent decisions by the U.S. Supreme
Court invalidating legislation enacted by Congress under Section
5 of the Fourteenth Amendment72 — is that the question whether
legislative action is more likely to threaten or protect important individual
rights has crucial empirical and predictive as well as moral dimensions.
Much depends on the status quo, the kinds of injustices
that it includes, and the likelihood of successful legislative or judicial
action to correct current moral shortfalls.73
As presently advised, I believe that those designing and assessing
constitutions in most western liberal democracies could reasonably assume
that legislative action is more likely than inaction to violate fundamental
rights.74 This judgment obviously underlies bills of rights
that create negative but not positive rights — and even Hirschl, in expressing
frustration that judicial review has not done much to solve
problems of distributive inequality, hedges in his assessment of
whether judicial review has, on balance, done more harm than good.75
Moreover, because economic regulatory and redistributive legislation
of the kind involved in Lochner-era cases does not typically threaten
any fundamental rights at all, it should not be assumed that judicial
review will result in the invalidation of such legislation (even if the
possibility cannot be rejected out of hand).
I take very seriously, however, the possibility that some categories
of cases should be exempted from judicial review, or at least subjected
to judicial review of only a very narrow scope. Among other things, a
multiple-veto-opportunities argument in favor of judicial review has
no traction in cases in which legislatures seek to promote the rights of
one class of citizens without threatening the fundamental rights of another.
Nor does my argument apply in zero-sum controversies in
which fundamental rights are pitted against one another and the over-
–––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
72 See, e.g., Kimel v. Fla. Bd. of Regents, 528 U.S. 62 (2000) (invalidating congressional legislation
expanding rights to freedom from age-based employment discrimination).
73 Cf. Freeman, supra note 56, at 361 (noting that “whether judicial review is appropriate for a
particular democratic constitution is a strategic question”).
74 Cf. THE FEDERALIST NO. 78 (Alexander Hamilton), supra note 56, at 469 (arguing that
“the courts of justice are to be considered as the bulwarks of a limited Constitution against legislative
encroachments”).
75 Hirschl writes, for example, that “[w]hereas the constitutionalization of rights does have
crucial importance in affirming marginalized identities and enhancing the status of individual
freedoms, its independent impact on ameliorating the socioeconomic status of historically disenfranchised
groups is often exaggerated.” HIRSCHL, supra note 69, at 168 (emphasis added). As
the italicized language indicates, Hirschl is not always clear about the relationship between his
claims that judicial review has failed to advance and that it has actually thwarted interests in
progressive redistribution, and his judgment of the overall balance of the costs and benefits for
judicial review for disadvantaged groups frequently seems equivocal.
2008] AN UNEASY CASE FOR JUDICIAL REVIEW 1713
enforcement of one entails the underenforcement of the other. I shall
more fully discuss the challenge posed by such cases in Part IV.
The third, related assumption on which I have relied is that some
rights are more fundamental than others and, accordingly, deserve
more protections against infringement through the provision of multiple
veto opportunities. This assumption is also contestable. It is at
least arguable — indeed, Waldron himself has maintained in other
writing — that every claim of individual right asserted against democratically
enacted legislation also implicates the majority’s “right” to
collective self-government.76 If collective self-governance is a right,
and if it stands on the same plane of moral importance as such rights
as freedom of speech and religion, then there could be no sound reason
to err on the side of overprotecting the rights that bills of rights customarily
recite.
As I have suggested, Waldron may in fact believe that the right to
self-government possesses the same moral status as many of the rights
commonly listed in bills of rights. He develops an argument that emphasizes
the importance of rights of democratic self-government —
with the term “right” used in a strongly moralized sense — in his important
book Law and Disagreement.77 Interestingly, however, he
places little if any weight on this argument in The Core of the Case
Against Judicial Review. The reason may involve his third assumption:
that the societies that come within his core case hold a serious
commitment to the protection of rights that they have memorialized in
a bill of rights.78 In most bills of rights, a general right of political majorities
to govern themselves by processes of majority decisionmaking
does not occupy the same status as fundamental individual rights protected
against legislative infringement.
In any event, the view that certain core individual rights are more
morally important than a general right to democratic self-government
seems to me both widely held and, at the very least, reasonable. Nevertheless,
to defend this view would take a very long argument that I
shall not attempt here.
The fourth assumption undergirding my thesis that outcome-based
arguments support judicial review is that a system of judicial review
can be so designed that the total moral costs of the overenforcement of
rights that judicial review would likely produce will be lower than the
moral costs that would result from the underenforcement of rights that
–––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
76 See WALDRON, LAW AND DISAGREEMENT, supra note 10, at 213; see also Roosevelt, supra
note 49, at 1662 (explaining that some citizens may support judicial restraint because when a
court “erroneously strikes down a valid governmental act, the American people . . . have been denied
the ability to govern themselves”).
77 See WALDRON, LAW AND DISAGREEMENT, supra note 10, at 233–54.
78 See Waldron, The Core Case, supra note 10, at 1364–66.
1714 HARVARD LAW REVIEW [Vol. 121:1693
would likely occur in the absence of judicial review. This is a very
large assumption, and I shall revisit it in Part IV, which considers
more specifically how judicial review ought to be designed, after Part
III has introduced some further pertinent considerations. For now,
suffice it to say that any assumption that judicial review can be so designed
that its moral benefits will exceed its costs rests on tricky assessments
along at least two dimensions. It requires predictions about
comparative numbers of errors and about the comparative moral costs
of over- and underenforcement under circumstances in which reasonable
disagreement exists both about what would count as over- and
underenforcement and about what their respective moral costs would
be.79 In developing my argument, I have advanced no formula for
making these calculations.80
My failure to work out responses to challenges to my crucial fourth
assumption might appear to concede one of Waldron’s main contentions:
given the difficulty of producing an analytically rigorous formula
for determining whether, when, and to what extent judicial review
might lead to morally better outcomes than would otherwise occur,
perhaps we should just dismiss outcome-based arguments as ultimately
“inconclusive.” But I think not. For anyone considering
whether outcome-related reasons come out for or against judicial review,
three possible answers stand out: first, the balance of outcomerelated
reasons supports judicial review; second, the balance of outcome-related
reasons opposes judicial review; third, outcome-related
concerns are sufficiently uncertain that the decision whether to have
judicial review should be made on other grounds entirely. What merits
emphasis is that the third conclusion is as much subject to the risks of
moral and empirical error as either of the first two. If outcome-related
concerns matter in principle in deliberations about whether to support
or oppose judicial review, the question of comparative moral costs
needs to be answered, whatever the difficulties in making the requisite
calculations.
–––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
79 Other considerations also matter, including the “uncertainty costs” that exist when it cannot
be known in advance whether legislation will ultimately be upheld as valid. See VERMEULE,
supra note 58, at 275. In believing that a multiple-veto-points case for judicial review must depend
on a sophisticated, multifactored calculation of costs and benefits, I differ with Frank Cross,
who maintains that the overenforcement of rights is so categorically preferable to underenforcement
that judicial review would be desirable even if courts were “typically wrong and much less
capable than Congress” in making decisions about when to uphold rights claims. Cross, supra
note 4, at 1577–78.
80 Cf. WALDRON, LAW AND DISAGREEMENT, supra note 10, at 252–53 (noting that instrumental
arguments for choosing one or another mechanism to adjudicate claims of rights frequently
commit the fallacy of “presuppos[ing] our possession of the truth about rights in designing
an authoritative procedure whose point it is to settle that very issue”).
2008] AN UNEASY CASE FOR JUDICIAL REVIEW 1715
The question of comparative moral costs is not, of course, a binary
one, for a crucial variable involves the nature or scope of the judicial
review that courts might practice. As I have noted, no one would
think that criminal defendants should go free in every case that occasions
even the smallest scintilla of doubt about the defendant’s innocence.
Similarly, no one would think that the rules governing judicial
invalidation of legislation should mandate the striking down of every
statute that could even colorably be argued to violate someone’s rights.
In designing a morally defensible system of judicial review, two interrelated
questions thus need to be resolved. One involves which
rights and other constitutional norms the judiciary should be enlisted
to enforce. A related question is how stringent judicial review ought to
be in order to get the balance of moral costs and benefits into an optimal
or at least a defensible alignment. For example, should courts exercise
independent judgment in identifying rights violations, should
they invalidate legislation only in cases of “clear mistake” by the legislature,81
or should they employ standards of review of some intermediate
scope?82 Once again, I have not attempted to work out answers to
questions such as these, though I shall say a bit more about them in
Part IV. Instead, I have assumed that it is reasonable to believe that
such questions can be resolved satisfactorily, even if not perfectly or
with unanimous accord.
Although the assumptions on which my case rests are, I repeat,
very large — and the case itself thus is somewhat uneasy — I believe
that any persuasive case for judicial review would need to rely on assumptions
that are similarly strong. Surely this is true of the familiar
claim that courts are better than legislatures at resolving rights-based
disputes correctly under conditions of reasonable disagreement.
To conclude, despite Waldron’s argument to the contrary, it is reasonable
to believe that a constitutional democracy with a well-designed
system of judicial review would produce a morally better pattern of
outcomes than a political democracy without judicial review, even if
Waldron’s four conditions hold, under circumstances that plausibly
exist.
III. PROCESS-BASED REASONS AND POLITICAL LEGITIMACY
Waldron’s core case against judicial review also has a processbased
component,83 and it, too, merits consideration and a response.
–––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
81 See Thayer, supra note 46, at 144; see also VERMEULE, supra note 58, at 270–76 (also defending
a version of the clear-mistake rule).
82 See Roosevelt, supra note 49, at 1661–63 (arguing that courts may adopt decision rules of
varying levels of scrutiny based on assessments of the costs of error and the desire to minimize
certain kinds of error).
83 See Waldron, The Core Case, supra note 10, at 1375, 1386–95.
1716 HARVARD LAW REVIEW [Vol. 121:1693
In formal terms, the analysis that I offer in this Part parallels my response
to Waldron’s outcome-based arguments in Part II: my criticisms
of Waldron’s case against judicial review add up to, or at least
suggest, a qualified argument that actually supports judicial review on
grounds of fairness and political legitimacy.
Although this conclusion can be stated simply, my argument in this
Part is intricate, reflecting the complex interrelationships among concepts
or ideals of (1) the morally best pattern of substantive outcomes,
(2) procedural fairness, and (3) political legitimacy. A preliminary
mapping of this Part’s organizational structure may therefore prove
helpful.
After briefly recounting Waldron’s process-based argument against
judicial review, I begin by arguing that the fairness of decisionmaking
procedures depends heavily on whether they are well designed to
achieve substantively good outcomes. Waldron treats issues of fairness
and legitimacy as independent of any outcome-based considerations
that might support judicial review — for he, of course, believes that
none exists. But if judicial review reduces rights violations, this benefit
is clearly relevant to the issue of fairness, and it ought to matter to
legitimacy as well.
Legitimacy, however, is a less intuitive concept than is fairness. In
order to show that the legitimacy of decision procedures may depend
partly on their capacity to generate substantively just outcomes, it is
necessary to unpack the concept of political legitimacy. In doing so, I
distinguish between overall political legitimacy, which is the ultimate
concern of democratic political theory, and specifically democratic legitimacy,
which is a vital source of political legitimacy, but not the
only one. In light of this distinction, I maintain that judicial review
could contribute to the overall political legitimacy of a constitutional
regime insofar as it helps to minimize fundamental rights violations,
even if it lacks democratic legitimacy.
It is not enough, however, to say that judicial review might make
some contribution to overall political legitimacy even if it lacks specifically
democratic legitimacy, for the concern remains that at least some
degree of democratic legitimacy might be necessary for overall political
legitimacy. In other words, the possibility must be confronted that judicial
review might so far deprive a political system of its necessary
modicum of specifically democratic legitimacy as to render that system
illegitimate overall. In the final section of this Part, I consider and
dismiss arguments to this effect, subject to the proviso that certain assumptions
must hold.
A. Waldron’s Process-Based Argument — And Its Limits
Waldron’s process-based argument against judicial review rests on
what he characterizes as “the well-known fairness arguments” in favor
2008] AN UNEASY CASE FOR JUDICIAL REVIEW 1717
of “the principle of majority decision.”84 According to him, this is the
preferred mechanism for resolving disputed issues of governmental
policy “[w]hen we disagree about the desired outcome, when we do not
want to bias the matter up-front one way or another, and when each of
the relevant participants has a moral claim to be treated as an
equal.”85
If Waldron were correct that there are no good outcome-related
reasons to want judicial review, his fairness arguments would carry the
day. But if he is wrong about the balance of outcome-based considerations,
then his fairness arguments also lose their potency. It is almost
too plain for argument that the fairness of procedures depends on the
nature of the substantive ends that the procedures are designed to promote.
For example, it is not unfair to use nonmajoritarian voting procedures
in criminal trials; to accord each juror the power to veto a
guilty verdict is wholly justifiable in view of the goals of the criminal
justice system. Similarly, most of us do not think it unfair to put control
of interest rates and the money supply in the hands of an independent
Federal Reserve Board,86 even though these matters can engender
reasonable contestation. Outcome-based considerations make
this allocation of power fair within the context of an otherwise largely
democratic government.
B. Political Legitimacy and Its Sources
As I have written elsewhere, the term “legitimacy” admits of diverse
usages.87 Appraisals of legitimacy can occur in sociological
terms, referring to prevailing public attitudes toward political authorities
and, in particular, to popular dispositions to respect those authorities’
claims to obedience.88 Alternatively, legitimacy can function as a
moral concept, measuring whether people ought to regard a political
regime, institution, or decision as having an entitlement to respect that
transcends the substantive correctness of an immediate object of dispute.89
In The Core of the Case Against Judicial Review, Waldron’s
concern involves moral legitimacy.90
–––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
84 Id. at 1387–88.
85 Id. at 1388.
86 But see Jon Elster, Don’t Burn Your Bridge Before You Come to It: Some Ambiguities and
Complexities of Precommitment, 81 TEX. L. REV. 1751, 1779 (2003) (noting that central banks
such as the Federal Reserve “have been widely criticized for lack of democratic legitimacy”).
87 See Richard H. Fallon, Jr., Legitimacy and the Constitution, 118 HARV. L. REV. 1787,
1789–90 (2005).
88 See id. at 1795–96.
89 See id. at 1796–1801.
90 See Waldron, The Core Case, supra note 10, at 1387–89 (introducing the concept of political
legitimacy as necessary to explain to those who disagree with political decisions why they none-
1718 HARVARD LAW REVIEW [Vol. 121:1693
No less than fairness, political legitimacy is relative to the ends that
an institution seeks to achieve. One possible ground for the political
legitimacy of a decision is that it emerged from open processes in
which all citizens had equal voice and votes.91 But there are grounds
for political legitimacy besides the majority-decision principle. Another
good reason to respect a decision and to regard it as deserving
obedience is that it issued from a process designed to reach generally
sensible substantive decisions while making infringements of individual
rights as unlikely as reasonably possible.92
When the diversity of potential sources of political legitimacy
comes into view, what Waldron can claim to have established convincingly
is, at most, that judicial review lacks the specifically democratic
form of legitimacy that adheres to decisions reached through majoritarian
processes. When other sources of political legitimacy enter the
calculus, the possibility emerges that judicial review might actually
promote, not detract from, the overall legitimacy of a governmental
regime if it provides important assurances against rights violations.
C. Comparative Democratic and Political Legitimacy
In claiming that judicial review might enhance the legitimacy of a
system of government, I need to introduce some qualifications. Although
there are multiple possible sources of political legitimacy, democratic
legitimacy matters deeply. A scheme of government that
wholly lacked democratic legitimacy, such as a hereditary dictatorship,
might strike us not only as seriously deficient in democratic legitimacy,
but also as lacking political legitimacy more generally. This judgment
might well persist even if the dictator put a number of safeguards in
place to protect against violations of the individual rights that typically
appear in bills of rights — except, of course, for voting and political
rights. It is therefore necessary to take seriously whether judicial review
sufficiently robs an otherwise generally democratic scheme of
government of its democratic legitimacy that it threatens the system’s
political legitimacy overall.
1. Anchoring Assumptions. — Although the question whether judicial
review might diminish or even demolish overall political legitimacy
bears asking, it is also nebulous. No set scale for calibrating political
legitimacy exists, and the standards of assessment can be at best
loosely defined. I shall proceed, however, on the basis of a number of
–––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
theless ought to regard those decisions as worthy of respect and as having a moral claim to
obedience).
91 See id. at 1387–88.
92 Cf. Randy E. Barnett, Constitutional Legitimacy, 103 COLUM. L. REV. 111, 145 (2003) (asserting
that “a legitimate lawmaking process is one that provides adequate assurances that the
laws it validates are just”).
2008] AN UNEASY CASE FOR JUDICIAL REVIEW 1719
assumptions, beginning with three conceptual assumptions that I mean
to be neutral with respect to the legitimacy of judicial review. First, as
I have signaled already, I assume that the purpose of legitimacy inquiries
is to determine whether citizens have a moral reason to accede to
political decisions with which they substantively disagree93 or whether
government officials are morally justified in coercively enforcing the
law.94 Second, I assume that in every case this question has either a
yes or a no answer — that is, that it is either minimally legitimate or is
not (even if there is reasonable disagreement about which answer is
correct). Third, however, I assume that among minimally legitimate
regimes some are more legitimate, in the sense of more closely approximating
an ideal of perfect legitimacy, than are others.95 Accordingly,
judicial review might affect political legitimacy in either of two
ways: it might make a political regime either legitimate or illegitimate
in the minimal sense, or it might affect whether a minimally legitimate
scheme of government is more or less legitimate.
Even with these anchors in place, inquiry into the legitimacy of judicial
review risks cascading into self-defeating abstraction. To meet
this difficulty, I shall make four additional, substantive assumptions
about the political regimes to which judicial review might be added or
from which it might be subtracted. Although these assumptions are
deliberately chosen to make the case for the legitimacy of judicial review
seem plausible, none of them, to borrow a term from Waldron, is
“unrealistic.” Moreover, all reflect the modest aspirations of this Article,
which aims not to prove decisively that every society should adopt
judicial review, but only to show that there are plausibly imaginable
yet nonpathological circumstances under which judicial review would
contribute to political legitimacy overall.
First, I shall assume that a system with judicial review produces
fewer violations of fundamental rights than would a system without
judicial review.
Second, I shall assume that a system that includes judicial review
also comprises reasonably well-functioning democratic institutions,
such as a legislature, with large responsibilities for the enactment of
law and the design of governmental policy.
Third, I shall assume that the decision to establish judicial review
emerged from a majoritarian or supermajoritarian process at Time
One, even though it subsequently constrains majority decisionmaking
at Time Two and thereafter.
–––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
93 See Waldron, The Core Case, supra note 10, at 1386–87.
94 See Fallon, supra note 87, at 1800.
95 Id. at 1796–99 (distinguishing “ideal” from “minimal” legitimacy).
1720 HARVARD LAW REVIEW [Vol. 121:1693
Fourth, I shall assume that the citizens living at Time Two, or at
least many of them, feel a psychological sense of connection to their
Time One predecessors as co-participants in an enterprise of collective
self-government extended over time.
2. Transitional Questions. — Although all of my anchoring assumptions
are important, the third substantive assumption most directly
frames the questions raised by Waldron’s arguments: is it politically
legitimate for democratic majorities or supermajorities at Time
One to establish judicial review as one element of a multi-part system
of government, notwithstanding actual or anticipated protests that judicial
review is unfair because it will make it more difficult for political
majorities to enact their preferences into law at Time Two? This
question has many possible variants, of which I shall consider just
two. First, I shall discuss the political and especially the democratic
legitimacy of judicial review when judicial review is not entrenched in
a society’s constitution. Absent entrenchment, judicial review exists,
but it could be abolished by majority vote in the legislature, which
could also override any particular judicial decision. Second, I shall
consider the distinct issues that arise when judicial review is entrenched
in a double sense — when the rights protected by judicial review
cannot be altered or abolished except by supermajoritarian procedures,
and when the institution of judicial review is itself protected
against abolition by mere legislative majorities.
3. Judicial Review Without Entrenchment. — If a majority of citizens
in an otherwise reasonably democratic society wants to establish
or maintain nonentrenched judicial review as a means of protecting
nonentrenched rights, I can see no plausible objection of political legitimacy.
As I have argued, judicial review might reasonably be
thought to enhance substantive justice by minimizing infringements of
important rights. Moreover, when judicial review is not entrenched
and when it protects rights that are also not entrenched, judicial review
does not so much derogate from the principle of majority decision
as implement it: democratic majorities that want judicial review for
outcome-based reasons are entitled to legislate accordingly.
I am not sure whether Waldron agrees with this assessment. In
The Core of the Case Against Judicial Review, he states at the outset
that the target of his argument is strong or entrenched, not weak or
nonentrenched, judicial review.96 Later, however, he advances an argument,
almost in passing, that appears to assert the unfairness of judicial
review that is subject to legislative override. Even if the legislature
can amend the provisions on which courts have based their
decisions to invalidate legislation, it is unfair, he writes, to have a sys-
–––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
96 See Waldron, The Core Case, supra note 10, at 1354.
2008] AN UNEASY CASE FOR JUDICIAL REVIEW 1721
tem of judicial review that “will have credentials in the political culture
that raise the stakes and increase the burden associated with the
amendment effort” and thereby result in “the deck . . . be[ing] stacked”
against those who dislike what a court has done.97
Although Waldron’s passing remark should raise a red flag for
anyone otherwise prepared to assume without analysis that nonentrenched
judicial review is wholly immune to fairness objections, it is
not easy to identify the premises on which an argument to the contrary
might rest. On the strongest possible interpretation, the majoritydecision
principle would condemn any rule that impedes current majorities
from enacting their preferences as frictionlessly as possible into
law. Although this extreme interpretation would explain a belief that
even nonentrenched judicial review is politically illegitimate, the position
that it reflects seems chimerical.98 Because democratic government
cannot exist in the absence of rules to constitute it, decisionmakers
at Time One need to be able to structure or constitute the forms of
political democracy — the voting rules, the offices for which elections
will be held, the requirements for the enactment of valid legislation,
and so forth — that will obtain thereafter unless they are authoritatively
changed.99 Moreover, virtually every electoral mechanism and
rule of parliamentary procedure may have some effect in frustrating
the wishes of current majorities. Why are elections so far apart? Why
must so many rules and procedures — requirements of multiple readings,
opportunities to prolong debate, and so forth — stand in the way
of enacting a bill into law? Whether they are wise or unwise, many
rules established at Time One do not so much constrain political de-
–––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
97 Id. at 1394. The pertinent passage reads as follows:
[D]efenders of judicial review claim that if legislators disagree with a judicial decision
about rights, they can campaign to amend the Bill of Rights to explicitly override it.
Their failure to do this amounts to a tacit democratic endorsement. This argument is
flawed because it does not defend the baseline that judicial decisionmaking establishes.
Amending a Bill of Rights characteristically involves a supermajority; or if it is a British-
or New Zealand-style statute [that can be overruled by a simple legislative majority],
it will have credentials in the political culture that raise the stakes and increase the
burden associated with the amendment effort. If our disgruntled citizen Cn asks why
the deck should be stacked in this way, the only answer we can give her refers back to
judicial decision. And that has already been found wanting.
Id.
98 See Eric A. Posner & Adrian Vermeule, Legislative Entrenchment: A Reappraisal, 111 YALE
L.J. 1665, 1686 (2002) (noting that “[t]he Condorcet paradox . . . casts doubt on the premise that a
simple majority can in normal circumstances even be identified,” due to the phenomenon of cycling,
in the absence of institutional structures and mechanisms of agenda control).
99 In theory, at least, there is a major puzzle about how an initiating institution such as a constitutional
convention could organize itself without a prior settlement of voting rules. See JON
ELSTER, ULYSSES AND THE SIRENS: STUDIES IN RATIONALITY AND IRRATIONALITY 95
(rev. ed. 1984) (hypothesizing that in the case of an “original assembly” the “circumstances [may]
make for a unity and unanimity in the face of which procedural questions evaporate”).
1722 HARVARD LAW REVIEW [Vol. 121:1693
mocracy as establish and empower it100 — even though the effect of
some is to “raise the stakes and increase the burden” that must be
borne by those wanting change.
Beyond concerns about simple coherence, ideals of democratic legitimacy
need to be tested against and brought into a sensible alignment
with broader ideals of political legitimacy, which presumably attach
significance to the likelihood that structures of (reasonably)
democratic government will produce good substantive outcomes — so
that bicameralism can be preferred over unicameralism, for example,
not because it is more democratic, but because it is likely to yield better
legislation. To put the point slightly differently, there may be competing
conceptions of democratic legitimacy.101 Given a choice, one
might reasonably prefer a conception that coheres as well as possible
with the best ideal of overall political legitimacy, including the concerns
about substantively good government that the broader ideal
imports.
On this point, Waldron would not appear to disagree with my
analysis, at least in the abstract. For example, he not only assumes
that bicameralism requirements accord with democratic ideals, but actually
supports bicameralism on the ground that it furnishes a prudent
check against possible majoritarian pathologies.102 He gives no hint,
moreover, that he would withdraw his praise in cases — such as that
of the Constitution of the United States — in which exacting requirements
for the enactment of legislation were put into place for the purpose
of stopping transient majorities from violating individual rights,
notwithstanding reasonable disagreement about those rights’ proper
contours.
Measured against premises that Waldron clearly accepts, the argument
that nonentrenched judicial review would be politically or even
democratically illegitimate thus seems unpersuasive. The same conclusion
emerges from reflection on the untenable implications of the
premise that it is democratically illegitimate for political majorities to
establish institutions “that will have credentials in the political culture
that raise the stakes and increase the burden” of reversing their decisions
by majority vote in the legislature.103 This premise would sug-
–––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
100 See STEPHEN HOLMES, PASSIONS AND CONSTRAINT 134–77 (1995); Larry Alexander,
What Is the Problem of Judicial Review?, 31 AUSTL. J. LEGAL PHIL. 1, 3–4 (2006).
101 See, e.g., John Ferejohn & Lawrence Sager, Commitment and Constitutionalism, 81 TEX. L.
REV. 1929, 1933 (2003) (noting that “[d]emocracy is an essentially contested concept”); Freeman,
supra note 56, at 332–39 (criticizing a purely procedural or majoritarian conception of democracy
and defending an alternative social contractarian conception that also incorporates a demand for
respect for basic individual rights).
102 See Waldron, The Core Case, supra note 10, at 1361 & n.47.
103 As Posner and Vermeule argue, much ordinary legislation has the same effect:
2008] AN UNEASY CASE FOR JUDICIAL REVIEW 1723
gest, for example, that it is politically illegitimate for states to vest responsibility
for drawing electoral districts in unelected, nonpartisan
commissions, the decisions of which could be overturned by a partisan
majority in the legislature.104 As I have suggested already, reliance on
a mechanism of this kind could more aptly be described as implementing
decisions made by a fairly democratic process than as unfairly
“stack[ing]” the “deck” against those who appeal to the principle of
majority rule.
4. Entrenched Rights and Judicial Review. — If some constitutive
rules of democratic government that were enacted at Time One can
make it difficult for Time Two majorities to enact their preferences
into law without affronting the ideal of democratic legitimacy, the
question arises why, or if, entrenchment is different. If bicameralism
requirements and presidential vetoes are democratically acceptable devices
for protecting individual rights, then what, if anything, is so objectionable
from the perspective of democratic legitimacy about Time
One decisions to entrench certain rights105 — especially if the entrenchments
are subject to supermajoritarian override and thus make
it difficult but not impossible for Time Two majorities to enact their
preferences into law?106
(a) Entrenched Rights Without Judicial Review. — Because entrenched
rights could exist even in the absence of judicial review, with
the legislature trusted to define and enforce them, it is an important
preliminary question in considering Waldron’s argument against judicial
review whether he opposes all entrenchments of rights or just entrenched
rights coupled with entrenched judicial review. In The Core
of the Case Against Judicial Review, Waldron elides this question.107
In Law and Disagreement, however, he suggested that he finds all entrenchment
to be incompatible with democratic ideals:
–––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
The Endangered Species Act imposes a political cost on later congressional majorities
with different views about species protection — majorities who might be willing to allow
or even promote animal takings and habitat destruction absent the Act but who are
not willing to pay the price attendant upon repealing the Act to accomplish those
purposes.
Posner & Vermeule, supra note 98, at 1696.
104 Such commissions have been proposed or established in a number of states to preserve the
basic fairness of the electoral system. See generally Scott M. Lesowitz, Recent Development, Independent
Redistricting Commissions, 43 HARV. J. ON LEGIS. 535 (2006).
105 For a lucid introduction to debates about this issue since the eighteenth century, see
HOLMES, supra note 100.
106 It bears noting that entrenchment comes in degrees, reflecting differences in the obstacles
that would need to be overcome at Time Two in order to displace a norm entrenched at Time
One. For discussion, see infra section IV.C.
107 Waldron assumes that a society committed to rights will have a bill of rights, though he
“mak[es] no assumption that the Bill of Rights is entrenched . . . . I want to leave that open.”
Waldron, The Core Case, supra note 10, at 1365.
1724 HARVARD LAW REVIEW [Vol. 121:1693
To embody a right in an entrenched constitutional document is to
adopt a certain attitude towards one’s fellow citizens. That attitude is
best summed up as a combination of self-assurance and mistrust: selfassurance
in the proponent’s conviction that what he is putting forward
really is a matter of fundamental right and that he has captured it adequately
in the particular formulation he is propounding; and mistrust, implicit
in his view that any alternative conception that might be concocted
by elected legislators next year or in ten years’ time is so likely to be
wrong-headed or ill-motivated that his own formulation is to be elevated
immediately beyond the reach of ordinary legislative revision.108
Waldron’s argument that entrenchment reflects a kind of insult by
Time One decisionmakers directed toward their partners and successors
in democratic self-governance moves too quickly. Not all entrenchments
of decisions inherently involve insults. For example, individual
human beings can have good, noninsulting reasons to
entrench current decisions against future reversal even with respect to
themselves, as illustrated by the much cited case of Ulysses and the
Sirens.109
The argument that entrenchments are inherently insulting also ignores
the constitutive function that constitutions sometimes perform.
At Time One, it may be an open question for those drafting or being
asked to assent to a constitution whether or with whom to join in a
proposed project of collective self-government.110 Under these circumstances,
it may not be unreasonable to seek entrenchment of assurances
without which a proposed collective venture would appear unattractive
or even unacceptable. When the question at issue is who
“one’s fellow citizens” will be, the suggestion that bargaining for entrenched
guarantees reflects an unworthy “combination of selfassurance
and mistrust” seems misplaced.
Any general claim that all entrenchments of rights necessarily violate
the minimal requirements of democratic legitimacy seems similarly
unsustainable. To begin with, some rights, such as voting and free
speech rights, might actually be necessary for democratic legitimacy to
exist. One source of democratic legitimacy — not the only source, but
–––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
108 WALDRON, LAW AND DISAGREEMENT, supra note 10, at 221–22.
109 See, e.g., ELSTER, supra note 99, at 36–37. In citing a case of individual precommitment, I
do not mean to suggest that the entrenchment of norms in constitutions — in which one generation
attempts to bind successor generations as well as itself — are similar in all or even most respects.
I mean only to argue that attempts at precommitment do not necessarily convey an insult.
110 Cf. Ferejohn & Sager, supra note 101, at 1930 (noting the likelihood that “those who expected
to be threatened by the new national government” created by the Constitution of the
United States “insisted that certain of their rights were guaranteed, or committed to, as a price for
accepting the new government”).
2008] AN UNEASY CASE FOR JUDICIAL REVIEW 1725
an important one — is substantive consonance with democratic
ideals.111
In addition, rules adopted at Time One that make it difficult for
political majorities to enact their preferences into law at Time Two
may possess at least some democratic legitimacy insofar as they were
themselves established through democratic processes at Time One and
insofar as the citizens at Time Two regard their Time One predecessors
as participants in a continuing regime of collective self-government.112
On the scale of democratic political legitimacy, it counts for something
(though surely not for everything) that a rule that frustrates majoritarian
decisionmaking at Time Two was itself enacted through a fair
democratic process at Time One.
Finally, it matters crucially that democratic legitimacy is a property
of political regimes overall, not just of their components.113 Even if
one element, such as judicial review, is relatively lacking in democratic
legitimacy, democratic legitimacy may remain a property of the regime
as a whole — which, as I have insisted repeatedly now, may gain in
political legitimacy overall by foregoing aspirations to maximize specifically
democratic legitimacy in some of its elements. In a country in
which robustly democratic institutions make the bulk of important political
decisions, the right question about entrenchments of rights is not
whether they are maximally democratic, but whether they are otherwise
justified within a politically legitimate regime that derives some
of its legitimacy from specifically democratic sources.
–––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
111 See RONALD DWORKIN, FREEDOM’S LAW 15–34 (1996); Michael J. Klarman, Majoritarian
Judicial Review: The Entrenchment Problem, 85 GEO. L.J. 491, 499–501 (1997).
112 My way of putting this point may appear to conflate an empirical psychological issue —
involving how people living at Time Two view their Time One predecessors — with a fundamentally
moral issue of political legitimacy. But I do not think so. The very idea of democratic legitimacy
necessarily presupposes the existence of a community whose practices should at least
presumptively be democratic. In considering the specifically democratic legitimacy of any claim
that a rule is binding in the present because it was adopted by a majoritarian or similar procedure
in the past, it always matters how those living in the present reasonably understand their connection
with past actors. Most of us do not doubt the democratic legitimacy of legislation enacted by
Congress in 1789 or 1889, even though none of us was alive then, any more than we doubt the
democratic legitimacy of legislation enacted by Congress in 1989 or 2006. But cf. Thurgood Marshall,
Reflections on the Bicentennial of the United States Constitution, 101 HARV. L. REV. 1, 2–5
(1987) (questioning “the wisdom, foresight, and sense of justice” of the Framers due to their discrimination
on the basis of race and gender). By contrast, any suggestion that we were bound
today by a decision reached by majority vote of the British Parliament, whether before or after
the Revolution, would be wholly lacking in democratic legitimacy in the United States because
the requisite sense of participation in a shared, ongoing project of democratic self-government
would be absent.
113 See Frank I. Michelman, Is the Constitution a Contract for Legitimacy?, 8 REV. CONST.
STUD. 101, 105 (2003) (maintaining that legitimacy judgments appraise “an entire system, or practice,
or ‘regime’ of government”).
1726 HARVARD LAW REVIEW [Vol. 121:1693
(b) Entrenched Rights Coupled with Entrenched Judicial Review.
— The strongest objection to judicial review undoubtedly involves entrenched
constitutional provisions, especially those guaranteeing individual
rights, that are backed by entrenched judicial review. Waldron
argues forcefully that because the appropriate specification of rights
may be subject to reasonable disagreement, fairness requires the resolution
of disagreement by democratically accountable institutions.114
The first thing to be said about this claim — even insofar as it appeals
to the ideal of specifically democratic legitimacy — is that it
seems at the very least to sweep too broadly. As John Hart Ely argued,
one vital function of entrenched judicial review can be to promote
democratic legitimacy by ensuring that elected representatives
adhere to democratic norms.115 Certainly officeholders can have incentives
to enact measures tending to disable political opposition.
With this hazard in view, the entrenchment of judicial review to protect
the constitutive norms of political democracy could serve as a
safeguard of, rather than a threat to, the democratic legitimacy of constitutional
government.
In conjuring the prospect that a legislature might trample on rights
of political participation or otherwise subvert democratic selfgovernment,
I have of course imagined that two of the four assumptions
on which Waldron rests his case against judicial review might
cease to hold: a legislature that disregarded citizens’ rights and concerns
of democratic legitimacy would not be “in reasonably good working
order,” nor would it manifest a commitment to preserving rights.
Accordingly, this might look like the kind of “pathological” situation in
which Waldron acknowledges that judicial review might indeed be
warranted.116 Moreover, because the case would come under Waldron’s
express exception for pathological societies, it might appear not
to weaken his argument that entrenched judicial review should be rejected
as democratically illegitimate in societies whose legislatures are
in good order.
But this line of reasoning ignores the distinction between Time
One, when a constitution is established, and Time Two, when the conditions
existing at Time One may have ceased to exist. Once this distinction
is recognized, it seems plain that decisionmakers at Time One
might reasonably entrench judicial review of legislation alleged to
threaten the constitutive norms of political democracy, not because
they believe their society to be pathological at Time One, but to establish
a bulwark against pathologies that might develop in the future.
–––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
114 See Waldron, The Core Case, supra note 10, at 1389–93.
115 See JOHN HART ELY, DEMOCRACY AND DISTRUST: A THEORY OF JUDICIAL REVIEW
135–79 (1980); see generally Klarman, supra note 111.
116 Waldron, The Core Case, supra note 10, at 1352, 1401–06.
2008] AN UNEASY CASE FOR JUDICIAL REVIEW 1727
Surely a society can take precautions against the possibility of future
legislative pathologies without thereby betraying itself as pathological
already.117
Of course, judicial review under most constitutions aims to protect
individual rights on a broad basis, not just to safeguard the constitutive
norms of political democracy. Insofar as entrenched judicial review
extends to asserted violations of rights not directly linked to political
democracy, its claim to specifically democratic legitimacy grows
weaker. But a slippage into overall political illegitimacy need not
result.
Even in cases that trigger reasonable disagreement, the force of
Waldron’s argument erodes under the weight of a number of mutually
reinforcing considerations, most if not all of which should now be evident.
First, as I have emphasized, political legitimacy has sources besides
democratic legitimacy. If good reasons exist to think that entrenched
judicial review may produce a morally better pattern of
results than would otherwise occur, then entrenched judicial review
derives (some) political legitimacy from its functionality.
Second, the adoption of judicial review through relatively democratic
processes at Time One may endow it with a continuing residue of
democratic legitimacy at Time Two, even in cases involving reasonable
disagreement about the appropriate specification of disputed rights.
Recognizing once more that legitimacy is a matter of degree, I do not
mean to claim that the entrenchment of judicial review is necessarily
perfectly democratically legitimate at Time Two, just that democratic
adoption at Time One may count for something.118
Finally, insofar as the issue is the democratic legitimacy of the political
system as a whole, rather than just the corner involving judicial
review, a scheme of government that includes judicial review can still
register decently high even on the scale of democratic legitimacy. In
assessing overall democratic legitimacy, it will matter greatly if a political
regime’s legislature is “in reasonably good working order” and if,
as I have suggested, many exercises of entrenched judicial review may
serve more to protect than to threaten political legitimacy.
When all of these considerations enter the analysis, Waldron fails to
demonstrate the political illegitimacy of judicial review even in cases
involving controversial applications of vague moral rights. If judicial
review makes a contribution to the overall moral quality of a society’s
–––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
117 Waldron himself so acknowledges when he defends a preference for bicameral over unicameral
legislatures on the ground that “unicameral arrangements can easily exacerbate other legislative
pathologies.” Id. at 1361 n.47.
118 The degree of entrenchment may also matter. Democratically established entrenchments
that are subject to displacement by supermajority voting requirements may be less troublesome
than are absolute or more nearly absolute entrenchments.
1728 HARVARD LAW REVIEW [Vol. 121:1693
political decisions, then judicial review might actually enhance the
overall political legitimacy of an otherwise reasonably democratic constitutional
regime.
IV. NOTES ON THE DESIGN OF A SYSTEM OF JUDICIAL REVIEW
In advancing a case for judicial review that rests on the desirability
of multiple veto points to protect fundamental rights, even on the assumption
that a well-functioning legislature takes rights seriously, I
have so far simply assumed that a system of judicial review could be
so designed that the moral costs of the resulting likely overenforcement
of rights would be less than the moral costs of the underenforcement of
rights that could be anticipated otherwise. For reasons stated above,
proof of the validity of this assumption exceeds my ambitions in this
Article. Nevertheless, the questions whether there should be judicial
review and how judicial review ought to be designed are too intertwined
for any complete separation between them to be wholly satisfactory.
Before concluding, I therefore want to say a few words about
how my argument in favor of judicial review — which makes no assumption
that courts are better than legislatures at resolving questions
of rights correctly — bears on how judicial review should be designed
and administered. For purposes of exposition, I shall separately address
three issues that are in fact pervasively interconnected. First,
which claims of rights should courts review? Are there some issues
with respect to which judicial review is unnecessary or unwise? Second,
what should be the scope of judicial review, or how searching
should it be? Third, should judicial review be entrenched against legislative
override (“strong”) or subject to legislative displacement
(“weak”)?
A. Judicially Reviewable Issues
Because I have rested my argument for judicial review of legislation
on the premise that some rights are more fundamental than others
and therefore more deserving of protection, my core case extends only
to the kinds of fundamental rights characteristically protected in bills
of rights and does not necessarily apply directly to “ordinary” liberty
rights to freedom from governmental regulation.119 Claims of rights to
–––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
119 Although I have accepted for purposes of argument that courts are no better than legislatures
at identifying rights correctly, I have not similarly assumed that courts have no comparative
advantages over executive officials. Executive officials must frequently make snap, nondeliberative
decisions that even they themselves might not endorse after thinking through the implications
for individual rights. Judicial review of executive action thus seems to me a sound practice that
my argument in no way calls into question, regardless of the nature of the rights at stake. Cf.
TUSHNET, supra note 2, at 163–64 (arguing that a system without judicial review of legislative
2008] AN UNEASY CASE FOR JUDICIAL REVIEW 1729
liberty in many Lochner-era cases thus might be excluded entirely
from judicial review under my argument. Alternatively, to anticipate a
“scope of review” issue, judicial scrutiny might be minimal in such
cases — a possibility that of course tracks current constitutional practice
in the United States.120 Or courts’ decisions in cases of this kind
might be subject only to weak judicial review even if strong judicial
review occurred in other cases.
Similarly, my argument for judicial review does not encompass
cases in which the legislature enacts its interpretation of fundamental
rights into law and the resulting legislation does not threaten the fundamental
rights of others.121 To take an example from U.S. constitutional
law, my argument for multiple veto opportunities to protect
fundamental rights would ordinarily provide no foundation for judicial
review of congressional action to protect congressionally defined rights
under Section 5 of the Fourteenth Amendment.122 If no fundamental
rights are threatened, then my argument would not apply. Nor would
my argument establish the desirability of judicial review to protect any
structural constitutional norms not directly safeguarding fundamental
rights (including norms of constitutional federalism).123 There is, I
–––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
action should have “a reasonably robust law of ultra vires . . . that denies any legal effect to acts
by a government official outside the bounds of authority granted the official by the law”).
120 See, e.g., FCC v. Beach Commc’ns, Inc., 508 U.S. 307, 313 (1993) (explaining that “[i]n areas
of social and economic policy, a statutory classification that neither proceeds along suspect lines
nor infringes fundamental constitutional rights must be upheld against equal protection challenge
if there is any reasonably conceivable state of facts that could provide a rational basis for the classification”);
see also Richard H. Fallon, Jr., The Supreme Court 1996 Term—Foreword: Implementing
the Constitution, 111 HARV. L. REV. 54, 69 (1997) (describing how the “rational basis” test and
other “[n]onsuspect-content tests call for judicial scrutiny pursuant to standards reflecting strong
presumptions of constitutional validity”).
121 Cf. Cross, supra note 4, at 1589 (arguing for a “one-way ratchet” whereby “whichever national
institution provides the greatest protection of rights prevails”).
122 The Supreme Court has recently engaged in close oversight of congressional efforts to “enforce”
the Fourteenth Amendment based on its judgment that Congress may not define the Constitution’s
substantive guarantees more broadly than the Court has defined them. See City of
Boerne v. Flores, 521 U.S. 507, 519 (1997) (holding that by attempting to provide more protection
to free exercise rights than had the Supreme Court, Congress exceeded its Section 5 powers, since
“[l]egislation which alters the meaning of the Free Exercise Clause cannot be said to be enforcing
the Clause”). The Court once took a more relaxed approach, positing that Congress’s powers under
Section 5 mirrored the broad reach of its powers under the Necessary and Proper Clause as
articulated in McCulloch v. Maryland, 17 U.S. (4 Wheat.) 316 (1819). See Katzenbach v. Morgan,
384 U.S. 641, 651 (1966) (holding that “the McCulloch v. Maryland standard is the measure of
what constitutes ‘appropriate legislation’ under § 5,” which is “a positive grant of legislative
power authorizing Congress to exercise its discretion in determining whether and what legislation
is needed to secure the guarantees of the Fourteenth Amendment”).
123 Cf. JESSE H. CHOPER, JUDICIAL REVIEW AND THE NATIONAL POLITICAL PROCESS
(1980) (arguing that judicial review is appropriate in cases involving individual rights, but not in
those involving federalism issues or the respective powers of Congress and the President). It is
frequently argued that judicial review of separation-of-powers and federalism issues is appropriate
because the ultimate purpose of the separation of powers and federalism is to protect rights.
1730 HARVARD LAW REVIEW [Vol. 121:1693
should hasten to add, nothing in my argument that precludes judicial
review to enforce structural constitutional norms, but the case for judicial
review in such cases would need to depend on other arguments
than those that I have advanced.
A potentially even larger implication of supporting judicial review
based on my argument, rather than on the alternative premise that
courts are better than legislatures at identifying fundamental rights
correctly, would come in cases in which claims of fundamental rights
are plausibly (even if not correctly) asserted on both sides. For example,
Congress might think certain “accommodations” of religious institutions
or believers necessary to protect their fundamental right to religious
freedom, whereas those burdened by an “accommodation”
might contend that it violates their fundamental right not to be compelled
to support religious practice or belief.124 Or a legislature might
conclude that prohibitions against hate speech are necessary to afford
vulnerable minority groups the equal protection of the laws, while opponents
claim that a prohibition against hate speech interferes with the
fundamental right to freedom of expression.125 In cases in which the
legislature has striven conscientiously to determine which of two competing
fundamental rights claims deserves to prevail, my argument —
which presupposes that legislatures are as likely as courts to decide
correctly — implies that there is no good reason to want judicial review.
The argument for preferring overenforcement to underenforce-
–––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
See, e.g., Clinton v. City of New York, 524 U.S. 417, 449–53 (1998) (Kennedy, J., concurring) (disputing
the notion that separation-of-powers infringements do “not threaten the liberties of individual
citizens”); Akhil Reed Amar, Of Sovereignty and Federalism, 96 YALE L.J. 1425, 1426
(1987) (arguing that federalism “is designed to protect, not defeat . . . individual rights”). But
while judicial review of federalism and separation of powers questions plausibly helps to protect
liberty in a general sense, such review is generally unnecessary to protect fundamental rights, and
my multiple-veto-opportunities argument rests on the premise that fundamental rights have a
claim to special judicial protection that other liberty interests do not under conditions of reasonable
disagreement.
124 Cf. Corp. of the Presiding Bishop of the Church of Jesus Christ of Latter-Day Saints v.
Amos, 483 U.S. 327 (1987) (upholding a federal law exempting religious organizations from an
otherwise applicable prohibition against employment discrimination on the basis of religion even
though the effect of this exemption was to allow employees to be dismissed based on their religious
practices and beliefs).
125 Cf. R.A.V. v. City of St. Paul, 505 U.S. 377, 391 (1992) (striking down a city ordinance prohibiting
bias-motivated crime because “[t]he First Amendment does not permit St. Paul to impose
special prohibitions on those speakers who express views on disfavored subjects”); Am. Booksellers
Ass’n, Inc. v. Hudnut, 771 F.2d 323, 329 (7th Cir. 1985) (declaring unconstitutional on free expression
grounds a city’s antipornography ordinance even while accepting the ordinance’s premise
that pornography that subordinates women “leads to affront and lower pay at work, insult and
injury at home, [and] battery and rape on the streets”); Frank I. Michelman, Conceptions of Democracy
in American Constitutional Argument: The Case of Pornography Regulation, 56 TENN.
L. REV. 291, 307–09 (1989) (noting Hudnut’s refusal to balance the interest in free expression
against the “women’s claims to liberty and equal protection” infringed by unregulated
pornography).
2008] AN UNEASY CASE FOR JUDICIAL REVIEW 1731
ment of rights has no bite in genuinely zero-sum contests between
competing claims of fundamental rights.126
Obviously, however, it may often be disputable whether a particular
case does in fact involve a collision of rights or whether it would be
reasonable to believe that a collision exists. It is far from obvious
what the institutional mechanism for resolving such questions ought to
be. One possible approach would have courts decide such issues at the
threshold of every case and then decline further review if a collision
between fundamental rights exists or could reasonably be thought to
exist. Another approach would assume that actual collisions of fundamental
rights are likely to occur so rarely, and that the task of distinguishing
actual from purported collisions of fundamental rights is
likely to be so conceptually complex and politically divisive, that
courts should simply proceed to decide all cases involving claims of
fundamental rights, even if it would be desirable in principle to withhold
review in zero-sum cases. Division and resentment would seem
inevitable if a court had to hinge its decision of a disputed question on
the ground that the contentions of one side were not only mistaken,
but also unreasonable. And an approach under which courts exercised
judicial review in all disputed cases, even though the rationale for their
doing so would not in principle apply universally, would accord with
one of the assumptions on which I have said that my multiple-vetoopportunities
defense of judicial review rests — that legislative action
is more likely to violate fundamental rights than legislative inaction.
Other possibilities are also imaginable.
Although my argument for judicial review does not directly mandate
any single approach to cases involving claimed tensions between
fundamental rights, my tentative judgment is that it would be best to
have courts determine at the threshold whether a case could reasonably
be thought to involve an actual collision of fundamental rights
and, if so, either go no further or certify any decision on the merits as
subject to legislative override. As I have said, when rights disputes
could reasonably be thought to be zero-sum games, the best argument
for judicial review runs out if legislatures are presumed to be as likely
as courts to identify rights correctly.
–––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
126 See Cross, supra note 4, at 1604–05 (arguing that a “multiple vetoes” rationale for judicial
review identifies no comparative judicial advantage in cases involving conflicts of rights and that
“difficult tradeoffs are best made by more majoritarian institutions”). A multiple vetoes argument
also fails to support judicial review in any case in which there is reasonable disagreement concerning
whether a single governmental action either violates an individual’s fundamental rights
or, to the contrary, is necessary to respect the fundamental rights of that very same person. This
situation would exist if, for example, it was reasonable to believe that a partly race-based scheme
of public school assignments was necessary to enforce a particular schoolchild’s fundamental right
to an education untainted by past public racism and that race-based assignments would violate
the same child’s fundamental right not to be stigmatized through treatment as a racial token.
1732 HARVARD LAW REVIEW [Vol. 121:1693
B. Scope of Review
In cases that could not reasonably be thought to involve zero-sum
collisions of fundamental rights, every increment in the stringency of
judicial scrutiny diminishes the risk that a constitutional right will be
erroneously underenforced, but increases the likelihood of erroneous
overenforcement. With the anticipated moral costs of these effects
needing to be weighed against one another, there is no reason to assume
that the same balance should prevail with respect to every fundamental
right — any more under my rationale for judicial review
than under a scheme of judicial review predicated on the assumption
that courts possess a special competence for identifying rights correctly.
In American constitutional law, for example, different fundamental
rights are currently protected by a diverse array of tests.127 My proposed
multiple-veto-opportunities rationale would easily support a
similar diversity. It would also comport with the current judicial practice
of propounding doctrinal tests that implement vague constitutional
norms without necessarily reflecting their semantic meaning.128 To
take perhaps the easiest, best known example, the Miranda129 rule
calling for the exclusion of confessions from criminal trials unless suspects
in police custody have received specified warnings does not precisely
reflect the meaning of a constitutional norm forbidding coerced
confessions. Rather, it implements that norm by establishing rules that
are readily administrable by police and judges alike.130
My framing of the stakes in weighing the costs of anticipated overand
underenforcement of fundamental rights does, however, put scopeof-review
questions in a different light than do familiar discussions of
how to trade off rights against competing governmental interests. In
designing a system of judicial review, the question — as I have emphasized
— is not solely one of how to weigh rights against other interests
based on the supposition that we actually know what rights people
have. Rather, the challenge is to determine the ultimate standard that
a court should apply in determining whether to invalidate legislation
when no initial baseline agreement concerning the content of rights exists
and there is no reason to think courts better than legislatures at resolving
disputed questions correctly. One question is how a court,
–––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
127 See Fallon, supra note 120, at 67–73 (providing a typology of “relatively common kinds of
tests” adopted by the Court “to help define constitutional limits on governmental powers”).
128 See generally RICHARD H. FALLON, JR., IMPLEMENTING THE CONSTITUTION 37–44
(2001) (developing the thesis that many constitutional tests are judicially constructed devices for
implementing constitutional norms, not reflecting their original or semantic meanings); Berman,
supra note 56.
129 Miranda v. Arizona, 384 U.S. 436 (1966).
130 See Fallon, Judicially Manageable Standards and Constitutional Meaning, supra note 56, at
1305–06.
2008] AN UNEASY CASE FOR JUDICIAL REVIEW 1733
thinking independently, should determine how to define a fundamental
right. Another closely related but separate question is whether a court
should invalidate legislation whenever, in its independent judgment, a
fundamental right has been violated or whether — under circumstances
of reasonable disagreement — it should accord more or less
deference to the legislature’s determination. For a system-designer, the
ultimate question has to be how to weigh the risks and costs of underenforcement
against those of overenforcement when a court and the
legislature, acting in presumptive good faith, reasonably disagree
about whether a fundamental right has been violated.
Relatively deferential review is undoubtedly an option. Indeed,
such review is probably a more attractive option when judicial review
is predicated on the assumption that fundamental rights should be protected
by multiple veto opportunities than if courts are thought more
likely than legislatures to resolve disputed questions correctly.
C. Choosing Strong or Weak Judicial Review
A final, crucial design question is whether judicial review should be
strong or weak. As should already be evident, however, strong and
weak judicial review are not binary alternatives, but labels that mark
areas along a spectrum; rights may be more or less entrenched, as may
the guarantee of judicial review as a mechanism to enforce fundamental
rights.131 The Constitution of the United States, for example, includes
several varieties of entrenchment rules. Under Article V, the
amendment of most constitutional rules requires approval by twothirds
majorities in both houses of Congress and by legislatures or
conventions in three-fourths of the states.132 But Article V also provides
even more stringently “that no State, without its Consent, shall
be deprived of its equal Suffrage in the Senate,” and it once established
absolutely that no amendment could be enacted barring the slave trade
before 1808.133
My multiple-veto-opportunities argument in favor of judicial review
implies that judicial decisions defining and enforcing fundamental
rights ought to have some degree of invulnerability to override by
ordinary legislation. But it is impossible to give an abstract specification
of exactly where along the spectrum of possible degrees of entrenchment
the optimal solution lies. There are too many interrelated
–––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
131 The most extreme form of entrenchment would purport to be absolute and to bind in perpetuity.
By contrast, a very weak, if not the weakest, form would forbid displacement of a current
norm by the ordinary majoritarian requirement of fifty percent of the legislature plus one, but
permit displacement by a vote of fifty percent plus two.
132 U.S. CONST. art. V.
133 Id.
1734 HARVARD LAW REVIEW [Vol. 121:1693
variables.134 In a genuinely nonpathological society of the kind that
Waldron imagines, only a mild form of entrenchment would seem to
me to be desirable. Under conditions of reasonable disagreement, democratic
institutions should not be too disabled from making periodic
reassessments of where and how to strike the balance in weighing the
comparative risks and moral costs of the over- and underenforcement
of fundamental rights.
Once again, however, it cannot simply be taken for granted that a
currently nonpathological society will remain that way, and entrenched
judicial review may provide a stronger hedge against pathologies that
might develop in the future. The desirability of relatively more or less
entrenched judicial review may thus depend on the pathological proclivities,
if any, of particular societies — even when no assumption is
made that those proclivities are socially or politically dominant at a
particular time.
D. Judicial Review in Societies that Are Not Well Ordered
Perhaps out of an excess of caution, I want to conclude this Part by
reminding readers that my arguments with respect to constitutional
design have all proceeded on the assumption that Waldron’s four conditions
are satisfied and, in particular, that the legislature is conscientiously
committed to respecting individual rights. I have not purported
to judge how fully the United States or any other society
realizes the predicates on which Waldron’s arguments and mine depend.
Nor have I purported to assess how judicial review should be
structured in societies in which the assumptions that I have followed
Waldron in making are, to one or another extent, less than wholly
satisfied.
I hope, however, that the importance of the subjects that I have not
addressed will not create the impression that I, following Waldron,
have not plowed rich ground. Too often American constitutional theorists
have focused their analyses exclusively on how the Supreme
Court of the United States should interpret the Constitution of the
United States under current circumstances. The proper domain of
constitutional theory is broader. Certainly it should encompass inquiries
into what provisions for judicial review (if any) ought to exist in
constitutions for all societies whose people and legislatures are seriously
committed to respecting rights.
–––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
134 Among other things, the appropriate degree of entrenchment might depend on the scope of
judicial review of legislative action. If, for example, courts operated under a scope-of-review formula
that permitted them to invalidate legislation only in cases in which they thought that the
legislature had made a “clear mistake” in finding no infringement of fundamental rights, then a
stronger degree of entrenchment might be appropriate than if it were assumed that courts would
exercise wholly independent judgment in identifying rights violations.
2008] AN UNEASY CASE FOR JUDICIAL REVIEW 1735
V. CONCLUSION
In this Article I have advanced parallel, substantially overlapping
negative and positive theses. The negative thesis is that Jeremy Waldron’s
provocative arguments in The Core of the Case Against Judicial
Review do not succeed. The positive thesis is that judicial review is
reasonably defensible within the terms of liberal political theory, even
for nonpathological societies, if, but only if, a number of assumptions
are valid.
Waldron’s core case against judicial review rests heavily on premises
and arguments suggesting that courts may be no better than legislatures
at resolving disputes about individual rights correctly. But the
best outcome-based reason to support judicial review is one that Waldron
overlooks. The affirmative case for judicial review need not depend
on the assumption that courts can identify rights more accurately
than legislatures. Rather, the most persuasive case maintains that both
institutions should be enlisted in the cause of rights protection because
it is morally more troublesome for fundamental rights to be underenforced
than overenforced.
Waldron’s affirmative case that judicial review is unfair and politically
illegitimate also fails, and the arguments so demonstrating point
once again to grounds of fairness and legitimacy on which judicial review
can be affirmatively defended. The fairness and political legitimacy
of procedural mechanisms depend on the ends that they serve.
If judicial review is reasonably designed to improve the substantive
justice of a society’s political decisions by safeguarding against violations
of fundamental rights, then it is not unfair, nor is it necessarily
politically illegitimate. Political legitimacy can flow from multiple
sources. Even insofar as judicial review lacks specifically democratic
legitimacy, the democratic character of other elements of a political regime
can partly compensate for this deficiency. And a shortfall in democratic
legitimacy may ultimately be outweighed, as a matter of
overall legitimacy, by the contribution that judicial review can make to
the protection of individual rights.
In concluding, however, I must emphasize once more that the arguments
for judicial review that I have advanced in this Article are
qualified and contingent in several ways. The case for judicial review
depends on a number of premises that seem to me likely to be true and
that Waldron has not refuted, but that would be difficult and in some
cases perhaps impossible to establish with knock-down, rationally irresistible
arguments. Some of these premises are moral or conceptual,
others empirical. Among the important but contestable premises undergirding
my case for judicial review are that even if courts are not
better overall at identifying rights violations than are legislatures,
courts have a perspective that makes them more likely to apprehend
serious risks to some kinds of fundamental rights; that errors that re-
1736 HARVARD LAW REVIEW [Vol. 121:1693
sult in the violation of fundamental rights are typically more morally
disturbing than errors that result in the erroneous overenforcement of
fundamental rights; and, perhaps most crucially, that a system of judicial
review can be so designed that the total moral costs of the overenforcement
of rights that judicial review would likely produce will be
lower than the moral costs that would result from the underenforcement
that would occur in the absence of judicial review. Any philosophical
case for judicial review, and probably against it as well, needs to
depend on reasonable, but also perhaps reasonably contestable, premises
such as these. This is an important point in its own right, with
which debate about judicial review needs to come to grips.