Tushnet & Bugaric, What Is Constitutionalism

1
What Is Constitutionalism?
Downloaded from https://academic.oup.com/book/38889/chapter/338031927 by Harvard University Library user on 19 February 2023

I. Defining Constitutionalism
What is constitutionalism? The only sensible follow-up is this: why do you want
to know? Scores of books have been written offering scores of descriptions of con-
stitutionalism: descriptions of the real world of constitutions and prescriptions
for the authors’ “best” account of constitutionalism as a system of values. Each
author has in mind some purpose for which she is providing the description. As
should be clear already, our purpose is to provide some leverage on the analysis
of contemporary populism.
That purpose sets some important guidelines for our answer to the question
posed in the chapter’s title. Because we are interested in contemporary populism,
we are interested in contemporary constitutionalism. We aren’t going to spend
time on describing “Athenian” constitutionalism, or “medieval” constitution-
alism, although for other purposes examining those and many other forms of
constitutionalism would be quite valuable.
More important, we aren’t going to discuss constitutionalism as what
philosophers call a “regulative ideal”—a reasonably comprehensive set of values
that we could use as a benchmark against which we could measure actual con-
stitutional performance. Consider a relatively simple question: Are laws against
hate speech consistent with constitutionalism as a regulative ideal? On some ac-
counts, clearly no—such laws violate ideals of freedom of expression. On other
accounts, clearly yes—such laws are required to ensure that everyone can partici-
pate in civic life on an equal basis. Proponents of each view will of course provide
elaborate arguments rooted in stories about individual autonomy, democratic
self-governance, and the like to justify their particular view of constitutionalism
as a regulative ideal. Precisely because each view is reasonable—as are many
others—we think it unhelpful to rely on “regulative ideal” versions of constitu-
tionalism in our inquiry into the relation between contemporary constitution-
alism and contemporary populism.
Of course, we do need some benchmark—some rough idea of what we mean
by “constitutionalism”—if we are to make any progress toward our goal. The
“regulative ideal” approach to the question relies upon what we call a “thick” ac-
count of constitutionalism. Thick accounts build a fair amount of detail into the
story. A thick account might for example provide answers to questions about the
Power to the People. Mark Tushnet and Bojan Bugarič, Oxford University Press. © Mark Tushnet and Bojan Bugaricˇ
2021. DOI: 10.1093/oso/9780197606711.003.0002
10 The Framework
kinds of political demonstrations that must be permitted—where the “must be”

means “constitutionalism requires that they be permitted.” We could provide
other examples, but the point should already be clear: thick accounts are detailed,

though perhaps not to the level of resolving questions about hate speech regula-
tion or other quite specific matters. Thick accounts can operate at a quite gran-
ular level, as “regulative ideal” accounts as we describe them do, or at a middle
level of abstraction.1
The thicker the account, the more readily will we be able to identify tensions
between constitutionalism and populism. Yet, the thicker the account, the more
readily will we be able also to identify tensions between constitutionalism and
any real-world political practices. Consider a relatively simple mid-level abstrac-
tion: contemporary constitutionalism, it might be said, requires that eligibility to
vote must be close to universal. Every jurisdiction that we know of excludes chil-
dren from the franchise. What about exclusion of long-term residents, lawfully pre-
sent, who aren’t citizens of the jurisdiction? Exclusion of those convicted of serious
crimes while they are imprisoned? For some period afterward? Should we regard
the rather large number of nations with one or more of those practices as falling
short of the requirements of contemporary constitutionalism? Perhaps we should.
Yet doing so may well dilute the sting of assertions that other shortfalls are more
serious.
We pursue a different strategy here. We offer a relatively thin account of con-
temporary constitutionalism. Its principles are stated at a reasonably high level
of abstraction: the judiciary must be reasonably independent of political control,
methods of determining popular support for policy choices must be reasonably
reliable, and a few more.2 Many different systems, some of them populist, satisfy
the requirements of a thin constitutionalism.3
1 We think it worth noting that the linkage between constitutionalism and some version of lib-
eralism, and in particular with democracy, developed in the early modern period, well after the
concepts of constitutionalism and democracy came on the scene. In a sense, then, our “thin” consti-
tutionalism can be treated as a branch of a tree of which thicker versions are other branches.
2 For a similarly thin definition of constitutionalism, see Wojciech Sadurski, Constitutional
Democracy in the Time of Elected Authoritarians, 18 International Journal of Constitutional
Law 324 (2020), at p. 328 (listing “free, fair, and regular elections,” “civil and political rights, in par-
ticular those which are instrumental to unconstrained political communication necessary for a dem-
ocratic electoral choice,” “separation or dispersion of powers,” and “the rule of law . . . [including]
constitutional rules which [the government] cannot change at will whenever political expediency so
demands”). Sadurski’s exposition of these principles, though brief, suggests that he would give them
somewhat more content that we would. See also Jeremy Waldron, Political Political Theory
(2016), ch. 2 (“Constitutionalism: A Skeptical View”); Tom Ginsburg & Aziz Z. Huq, How to Save
a Constitutional Democracy (2018), at p.10: “Our definition aims to be as minimalist as possible
without simply equating democracy with elections alone.”
3 Paul Blokker, Populist Understandings of the Law: A Conservative Backlash?, 13 Participazione
é Conflitto 1433 (2020), offers a sociolegal account of what he calls “embedded constitutional de-
mocracy,” in which the idea of constitutional democracy takes different forms depending upon the
socio-historical context. Perhaps our “thin constitutionalism” can be understood as our attempt to
dis-embed the idea of constitutionalism as much as possible, to take account of all the variations that
What Is Constitutionalism? 11
We don’t contend that thin constitutionalism is the most attractive version of

constitutionalism. For us it is mostly a heuristic device that allows us to lay out
the complex connections between populism and constitutionalism more clearly

than relying on a thicker idea would. We do think that it captures something
like a least common denominator of constitutionalism: Whatever else you think
constitutionalism requires, you’ll agree that it requires at least thin constitu-
tionalism. What’s more, thicken the account even a bit and we’re going to find
reasonable disagreement: you say that constitutionalism requires this specific
thing, but I can point you to constitutional systems that seem to me to do just fine
without it.
We worry that too much of the discussion of the relation between constitu-
tionalism and contemporary populisms is unproductive between participants in
the discussion have thick versions of constitutionalism in mind—and too often
those thick accounts either are undefended or conflict with other thick accounts
being put forth, in which case the participants are talking past each other. Our
hope is that the idea of thin constitutionalism can provide a common ground
for discussion. It can do so in the negative sense that we should all agree that
a populist government that fails to satisfy thin constitutionalism’s requirements
is indeed anti-constitutional. And, for us more important, the idea of thin
constitutionalism can help clarify some of the criticisms offered of some con-
temporary populisms, in which the language of constitutionalism is used as a
vehicle for expressing substantive disagreement with the regime’s policies. The
idea of thin constitutionalism brings out the essentially political—rather than
constitutional—nature of objections to specific populisms.
Before identifying what thin constitutional requires, though, we must note
one problem associated with thin accounts. As we have suggested, thick accounts
make it too easy to find fault with actual practices. Thin accounts, though, may
make it too difficult to do so. The reason is straightforward. Each high-level
abstraction can be brought to ground through a large number of what we call
“specifications,” each of which is reasonable on its own but many of which are
inconsistent with others.4
Again, the hate speech example is useful. Thin accounts of constitutionalism
may require free expression and civic equality. Hate speech laws might be in-
consistent with free expression (on some views) but required by civic equality
(on some other views). We think that this disagreement is an entirely reasonable
actually occur. (Complete dis-embedding is impossible without abandoning “constitutionalism” as

an analytic category.)
4 Specifications can be understood as the way in which reasonable disagreements about constitu-
tional meaning manifest themselves. For a discussion of such reasonable disagreements, see Jeremy
Waldron, Law and Disagreement (1999).
12 The Framework
one—which implies that having laws against hate speech and prohibiting them
as unconstitutional are both consistent with constitutionalism on a thin
understanding.

As we will show in our discussion of populism and in our case studies, rea-
sonable alternative specifications of abstract principles are everywhere. Much
of the asserted tension between constitutionalism and populism occurs because
populist regimes adopt a specification that critics don’t like even though it is,
in principle, a reasonable specification. As we argue in what follows, we believe
that sometimes the specifications as they operate in particular circumstances are
troublesome, but the circumstances, not the specifications, should be the focus
of concern.
II. The Elements of Thin Constitutionalism
Thin constitutionalism today has four elements.
• Majority rule: Policy over a wide range of issues is determined by the

preferences of a popular majority today.
• Entrenchment: Some policies can’t be changed by a simple majority of
today’s voters. These policies involve both rights and structures for making
decisions (including structures for determining what unentrenched poli-
cies a contemporary majority prefers).
• Judicial independence: Judges who resolve disputes over the application of
policies in individual cases should be independent of direct political con-
trol (that is, shouldn’t be responsive to politicians themselves chosen by the
people), but should be directly accountable to “the law” and indirectly ac-
countable to the people.
• Politicians and political parties. Office- holders and office-seekers—
politicians—organized in parties seek popular support for the programs the
politicians propose.
These elements seem simple enough, but when we examine each in detail, real
complexities emerge. And the elements interact in complicated ways.5
5 A cautionary note: we always have to keep in mind that we are discussing thin constitution-
alism. Real-world constitutions narrow the range of policies subject to choice by popular majori-
ties by entrenching some policies and having an extensive list of entrenched rights and structures.
For example, the current constitution of Brazil entrenches a balanced-budget requirement, that of
Germany commits the nation to social welfare policies, some nations give constitutional status to
their central banks, and others create ombuds offices with protections akin to those given judges.
Some or all of these might be good ideas, but none, we think, are required by a thin understanding of
contemporary constitutionalism.
A. Majority Rule
Contemporary governments can address an enormous number of issues: tax

policy, environmental degradation, the scope of privacy rights against the gov-
ernment and against private companies, and much more. They needn’t take
up any particular policy at any moment, of course, but when they do address a
policy issue, thin constitutionalism requires that the policies in place be chosen
by contemporary majorities. More precisely, it requires that a majority vote taken
today determines what the policy shall be going forward, no matter what some
past majority has chosen—and only until voters in the future choose a different
policy.6 Today’s electorate might not put existing environmental policy up to a
vote, accepting for the moment the policy choices made by their predecessors,
but that too is a choice made by today’s voters.7
As we will see, the range of policy choice open to majority rule is wide but not
unlimited because some policies are entrenched against contemporary majority
choice. More important for the moment, the principle of majority rule has to be
implemented by real-world institutions. How do we know what policies contem-
porary majorities want?
Thin constitutionalism requires only that majority preferences be reliably
determined—unless the procedures for determining those preferences are them-
selves entrenched.8 The first question, though, is, “majority of whom?” The quick
answer seems simple enough: a majority of adult citizens.9 Thin constitution-
alism allows for some narrowly defined groups of adult citizens to be excluded
from the count: citizens living abroad for an extended period and, more con-
troversially, for some time persons convicted of some crimes. Contemporary
populisms tend not to challenge the “adult citizen” requirement, so we don’t
focus more attention on it.
More important, every actual mechanism for finding out what policies today’s
majority prefers is imperfect. We’ll examine several mechanisms in a moment,
but take elections as the simplest example. Not everyone will vote in every
election (put another way, compulsory voting, while perhaps a good idea, isn’t
required by thin constitutionalism—or, indeed, by contemporary constitution-
alism itself). And we can be pretty confident that in any nation voting rates will
6 In British constitutional theory A.V. Dicey made this the centerpiece of his definition of par-
liamentary supremacy. Recent controversies over the existence of “constitutional” statutes in the
United Kingdom have revived interest in this definition. We discuss some of these controversies in
Chapter 6, on Brexit.
7 Thin constitutionalism is thus not intrinsically unstable—does not put all policy choices up for
grabs at every moment—because voters (and political leaders) decide which policies to reconsider.
8 This is the first place where we see how the elements of thin constitutionalism interact.
9 Again, specific national constitutions might include more people in the group whose preferences
count—in some places, for example, young people, in others long-term resident noncitizens.
Excluding such groups is consistent with thin constitutionalism, though.
14 The Framework
fall into patterns. The most widespread one is probably that the well-educated
and well-to-do vote at higher rates than the less well-educated and the less well-
to-do. Thin constitutionalism—in contrast to constitutionalism as a regulative

ideal—accepts these patterns of effective limitations on voting (or other ways of
determining preferences). It requires only that policy choices result from some
reasonably reliable mechanism for determining those preferences. And contem-
porary populists don’t defend their policy choices on the ground that they reflect
“true” majority preferences in the face of patterned effective limitations on deter-
mining preferences, so again we put this issue aside.
To summarize: thin constitutionalism requires that when the nation’s people
can make policy choices—which is to say, when they choose policies on subjects
not constitutionally entrenched against majority decision—the choices should
match the preferences of a majority of the nation’s adult citizens as revealed
by mechanisms that are reasonably reliable tools for determining what those
preferences are. The tools don’t have to be perfect—they only have to be good
enough.
There are a large number of (imperfect) mechanisms for determining the
majority’s preferences. Thin constitutionalism allows the use of pretty much any
mechanism that produces reasonably reliable measures of those preferences.
The simplest, of course, is reasonably free and fair elections. We know, of course,
that elections can measure majority preferences imperfectly. The simplest case
is probably the most common: political parties compete in elections by offering
platforms that address many subjects. A party or coalition can win a majority
because voters prefer its platform as a whole to the platform the losing party
offers. That doesn’t mean, though, that a majority prefers every item in the win-
ning party’s platform over the corresponding item in the losing party’s platform.
A stylized example drawn from the U.S. experience: a conservative party wins
on a platform promising tax cuts and substantially less environmental regula-
tion. The party could win an election focused on the tax issue because a majority
might really like the idea of tax cuts and only mildly oppose reducing environ-
mental regulation. Or another example: an unexpected policy issue—for ex-
ample, how to respond to a novel pandemic—arises during a government’s term
in office; the majority that supported its election couldn’t have had preferences
about the policy response to the novel issue, and so the government’s actions
might or might not reflect majority preferences.
That’s why thin constitutionalism allows for the use of other mechanisms for
determining a contemporary majority’s preferences. Referendums can supple-
ment elections, for example. And again, we know that referendums are imperfect
mechanisms, mostly because they typically offer voters a simple “yes/no” choice
on policies about which the voters might well have more nuanced views. (We
present a case study of the Brexit referendum later to illustrate this problem.)
Innovations in the technology of democracy now offer additional mechanisms

for determining (reasonably reliably) the majority’s views. We discuss some of
these innovations in the final part of this book. They include deliberative polling,

delegating policymaking power to a randomly selected group of people rather
than to elected representatives, and participatory budgeting built from the
ground up. Thin constitutionalism allows for the use of any of these mechanisms
so long as they are reasonably well-designed to provide a reasonably reliable
measure of majority preferences. And, once again, perfection isn’t required: the
measure of reasonableness should probably be, roughly, whether the chosen
mechanism is at least as good as reasonably free and fair elections are as a method
of determining majority preferences.
This brief discussion opens up some of the complexities associated with the
idea of majority rule as a basic principle of thin constitutionalism. If reason-
ably free and fair elections provide the benchmark for determining whether
some other mechanism for determining majority preferences is good enough,
what makes an election reasonably free and fair? And, what does thin consti-
tutionalism have to say about cases where two ways of determining majority
preferences seem to conflict (or where one mechanism is entrenched and others
prohibited)? This is perhaps the most difficult issue, because its answer depends
upon an answer to the deeper question: Is majority rule a substantive or a proce-
dural requirement of thin constitutionalism?
Free and fair elections. For elections to be reasonably free, candidates and
•
parties have to be able to present a full range of policy options to the public:
of course, that the current government should be replaced, but also that
taxes should go up or down, that public services should be expanded or
privatized, and so on through the list of all matters on which policy can be
made. Reasonably free and fair elections, that is, require freedom of political
expression.
We don’t want to offer a full theory of free expression here.10 Any such ac-
count, though, has to begin with the observation that freedom of expression
always competes with other important values—social stability, personal
privacy, civic equality, and much more. Thin constitutionalism requires
only that the balance struck between political expression and those com-
peting values be reasonable. And that, we think, is going to vary with so-
cial and political circumstances. Some restrictions on political expression in
highly polarized nations with a history of civic violence might be reasonable
even when the same restrictions elsewhere would be unreasonable. New
10 For the views of one of us, see Mark Tushnet, Advanced Introduction to Freedom of
Expression (2018).
16 The Framework
information technologies might alter the balance among the concerns for
autonomy, public understanding, and other considerations underlying the
general account of free expression.

For elections to be reasonably fair, voters have to be able to cast their votes
as they wish, and the votes that they cast have to be counted. Thin consti-
tutionalism prohibits physical intimidation of voters and fraudulent vote
counting. Subtler forms of coercion, such as an implicit threat that someone
who votes the “wrong” way will lose her job, are also inconsistent with thin
constitutionalism, although detecting such coercion might be so difficult as
to make this criterion unusable in practice.
Substance and procedure in majority rule. Consider a nation whose constitu-
•
tion says that all lawmaking power resides in the legislature, and doesn’t au-
thorize referendums. A social movement agitates for a referendum on some
issue, saying that the legislature’s position on the issue doesn’t reflect what
today’s majority wants. Of course, the movement can put pressure on the
legislature in all sorts of ways, including organizing a “referendum” on the
issue.11 But is the procedure set out in the constitution the only way to deter-
mine majority preferences that have legally binding effects?
Or go back to our example of a conservative political party elected on a
platform of tax cuts and deregulation. Suppose high-quality opinion polls
regularly show high levels of support for maintaining or even tightening en-
vironmental regulations. The presidentially controlled environmental regu-
latory agency authorizes an increase in the amount of air pollution a factory
can emit. Can a citizens’ group get a court order barring the deregulation on
the ground that the opinion polls provide a better measure of a majority’s
preferences than does the president’s campaign platform?
These are fancy ways of asking whether majority rule is a substantive prin-
ciple or a procedural one. On the substantive view, the majority’s preferences
govern once we know what they are, no matter how we gain that knowledge—
by elections, referendums, opinion polls, or any other source. On the procedural
view, the only ways that count for determining the majority’s preferences are the
mechanisms set out in the constitution—which could but need not include the
full list of mechanisms for determining a majority’s preferences.12 So, to go back
to our examples, if the nation’s constitution bars referendums, no purported ref-
erendum creates binding law no matter how overwhelming the majority is, and
the factory can increase the amount of pollution it spews into the air.
11 Proponents of Catalan independence organized a public “consultation”—really, a referendum

without binding legal effect—on the issue. The Spanish courts held the consultation unlawful and the
central government has prosecuted those who organized the consultation for (in effect) sedition.
12 Our point here has to be qualified by our discussion in the next section of this chapter of how
entrenched provisions—including those for determining majority preferences—can be modified.
We can imagine a political theory in which the substantive view prevails.

But, we think, constitutionalism no matter how thin has some irreducible
procedural core.13 One specific payoff is important in terms of the history of

populisms. An elected president can’t use decrees beyond those authorized
by the Constitution to impose what the president contends are the views
of today’s majority in the face of legislative resistance. That payoff is less
important today, where populist presidents usually have support from the
legislature.14
There’s a broader payoff, though. Thin constitutionalism means that polit-

ical leaders speaking for a majority have to use the institutions in place to ad-
vance their policies, or change those institutions according to rules for changing
them. Where the constitution in place entrenches some policy or procedure for
adopting policies, thin constitutionalism means that the constitution has to be
amended to replace that policy or procedure. As we will show, critics of populism
sometimes contend that populist enthusiasm for constitutional amendments
shows why there’s a tension between populism and constitutionalism. But, as
long as populists use the constitution’s amendment rules to change the constitu-
tion, they are acting in a manner consistent with thin constitutionalism.15
A final point: the word “reasonable” is doing a lot of work in our explana-
tion of thin constitutionalism. It’s quite difficult to offer anything more specific,
and people will often disagree about whether some policy choice is reasonable.
Note, though, one important feature: all the choices we’ve discussed—that is, all
the choices subject to the test of reasonableness—are themselves produced by
the institutions of majority rule. The question then is whether those imperfect
institutions have operated well enough to produce something we can fairly call a
reasonable though of course contestable choice.
We suggest two perspectives on that question. The first is for observers or
critics of the system. We think they should imagine that the choice had been
made by a government whose motives they didn’t suspect, and ask themselves
13 From different traditions, the work of Lon Fuller in jurisprudence and Joseph Schumpeter and
Robert Dahl in political science converge on this conclusion. See Lon Fuller, The Morality of
Law (1964); Joseph Schumpeter, Capitalism, Socialism, and Democracy (1942); Robert
Dahl, Dilemmas of Pluralist Democracy (1982).
14 Venezuela since around 2015 might be thought a counterexample, but as we argue in our case
study of Venezuela, though Hugo Chávez, who died in 2013, was a populist leader, his successor
Nicolas Máduro is a straightforward authoritarian.
15 We note here, and discuss in more detail in Chapter 4’s case study of Hungary, that the substance
of constitutional amendments adopted even in regular form can be inconsistent with thin consti-
tutionalism when they undermine the ability of current policy to reflect current majority views by
making it exceedingly difficult to displace a government whose policies no longer reflect those pre-
ferred by the majority. This is a different problem from the one discussed in the text, that populist
enthusiasm for amending constitutions is itself an indication of an anti-constitutional mindset.
18 The Framework
whether they would regard the choice as a reasonable one. The second is an insti-
tutional perspective: should courts assessing the choice find it outside the range
of reasonable options a democratic majority should have available? We take that

question up later in this chapter.
B. Entrenchment
Thin constitutionalism, we have argued, requires that some things—specifically,

some principles of free expression and civic equality (in connection with free and
fair elections) and some procedures for determining majority preferences—be
“entrenched.” The idea of entrenchment is intuitive: entrenched provisions can’t
be changed by the procedures used to enact and repeal ordinary laws—the laws
on which majority rule prevails.
Thin constitutionalism— indeed, we think, any sensible view of
constitutionalism—does not assume that whatever happens to be entrenched
in a constitution is good. A constitution might protect individual privacy “too
much,” in the sense that it shields from public view facts that voters would find
relevant to their choice of representatives. Or, it might protect privacy too little,
allowing the press to degrade targets of sensationalistic stories with little or no
public benefit. A constitution might make it too difficult to expand social welfare
provisions by requiring a balanced budget, or too easy to do so by guaranteeing a
long list of social welfare rights.
As we’ve noted, proposing to “de-entrench” some existing constitutional pro-
vision shifts the subject from a list of policies that can be changed only by some
supermajority to the list of policies that can be changed by majority rule. Doing
so isn’t in itself anti-constitutional. What matters is whether the substance of the
entrenched provision matters so much that it has to be protected from change by
current majorities (and subsequent change if the majority’s preferences change).
The rules dealing with financing election campaigns in the United States pro-
vide a useful example. As interpreted by the U.S. Supreme Court, principles
of freedom of expression severely restrict legislatures’ power to develop rules
restricting contributions to and expenditures by political candidates and parties.
Critics of those decisions have proposed a constitutional amendment that would
authorize legislatures to adopt “reasonable” regulations of campaign finance.
Would “amending the First Amendment” in this way be anti-constitutional?16
16 The example isn’t quite what we need, because the proposal can be characterized as an amend-
ment aimed at correcting the Supreme Court’s misinterpretation of the First Amendment and
restoring the amendment’s proper interpretation. To clean up the example, imagine that the consti-
tution entrenches a specific set of rules about campaign finance, which a current majority comes to
believe, reasonably, no longer fits the nation’s conditions.
Other well-functioning constitutional democracies have quite different systems

of financing political campaigns, including systems of complete public financing
with private contributions and expenditures prohibited and systems that limit

private contributions and expenditures. Perhaps there are reasons specific to
the United States to be concerned that legislatures wouldn’t adopt similar sys-
tems, or that such systems would operate badly in the United States even if they
operate reasonably well elsewhere. But, we think, arguments about such a pro-
posal should be conducted at that level of detail, not at the level of concern about
“constitutionalism.”
An important form of entrenchment occurs when legislation must pass
through what political scientists call “veto gates.” A veto gate exists at each point
in a multistage policymaking process where approval by some person or body is
required before a proposal moves to the next stage. A legislative committee can
be a veto gate; so can a lower house, or an upper house, or the president—or a
constitutional court. The more veto gates there are, the greater the chance that
actual majority preferences won’t be reflected in binding law. Veto gates block
the adoption of new policies, which means that they induce a bias in favor of
the status quo, even in the face of majority preferences as revealed outside the
legislative process, for example in opinion polls. Another metaphor is that some
constitutional provisions are speed bumps, slowing down movement toward a
policy’s implementation without definitively blocking the policy.17 From one
point of view, speed bumps offer a chance for additional deliberation about the
policy’s wisdom; from another, slowing progress toward implementation gives
opponents more time to build support for repealing the policy.
Ordinary majorities can’t change entrenched provisions they believe to be bad
ones, or eliminate an entrenched veto gate when some piece of legislation fails
to get through it or an entrenched speed bump that’s slowed down implemen-
tation dramatically. But thin constitutionalism does allow amendments to en-
trenched provisions by supermajorities. Many amendment rules exist: requiring
two-thirds or three-quarters approval rather than a simple majority, requiring
approval by a referendum, requiring approval by two successive parliaments
with an intervening election.18
17 For the term “speed bumps,” see Tom Ginsburg, Aziz Z. Huq, & Mila Versteeg, The Coming
Demise of Liberal Constitutionalism, 85 University of Chicago Law Review 239 (2018), at p. 253
(“At best, constitutional design features serve as speed bumps to slow the agglomeration and abuse of
political power; they cannot save us from our worst selves completely.”).
18 We’ve used examples of written constitutions with entrenched provisions and written amend-
ment rules. Entrenchment occurs even where the constitution is unwritten and sometimes in the
interstices of written constitutions. Entrenchment where the constitution is unwritten comes in the
form of settled practices and strong norms protecting them against change. These norms mean that
a politician who wants to change the entrenched rules ordinarily has to build up a great deal of sup-
port for change, more than she needs to secure changes in ordinary policy. Or, where the politician
manages to implement a change in norms without supermajority support, the changed norm won’t
“stick” when the politician leaves the scene.
20 The Framework
Amendments become sensible when experience reveals that the constitution’s

authors made a mistake.19 When nationally organized political parties came
onto the scene, the initial U.S. system for choosing a president and vice president

led to a constitutional crisis. After the crisis was resolved, the Constitution was
amended to prevent its recurrence. Amendments make sense when accumulated
experience leads to a better understanding of what mechanisms promote good
governance. As Thomas Jefferson memorably put it in 1816:
Some men look at constitutions with sanctimonious reverence, and deem them
like the ark of the covenant, too sacred to be touched. They ascribe to the men
of the preceding age a wisdom more than human, and suppose what they did to
be beyond amendment. I knew that age well; I belonged to it, and labored with
it. It deserved well of its country. It was very like the present, but without the ex-
perience of the present; and forty years of experience in government is worth a
century of book-reading; and this they would say themselves, were they to rise
from the dead.20
Finally, technological developments can make new techniques of governance

available.
A further point about entrenchments and their alteration by amendments ac-
cording to prescribed rules: some constitutions contain, or have been interpreted
to contain, provisions that simply can’t be amended at all within the rules the
constitution sets out (so-called “eternity” provisions). The German Basic Law,
for example, says that it can’t be amended to eliminate its protection of “human
dignity” or the nature of the state as “social and federal” under the rule of law.
The Indian Supreme Court has held that amendments to the constitution’s basic
structure—which includes commitments to federalism, secularism, and judicial
independence—are unconstitutional.
Amendment rules and eternity provisions themselves might be badly
designed—or so today’s people might think. An easy case: suppose the amend-
ment rule requires approval by three-quarters of the people, but 70 percent be-
lieve that that rule is too stringent. A harder case for many: suppose secularism
is protected by an eternity clause but three-quarters of the people think that their
country would be better off with a modest religious establishment. An even more
difficult case: the constitution’s structure as a federal republic is protected by an
eternity clause but a substantial majority comes to believe that economic change,
19 At the U.S. constitutional convention George Mason said that an amendment procedure was
needed because the new constitution “will certainly be defective.” Similar observations were made
during the French constitutional convention a few years later.
20 Thomas Jefferson to Samuel Kercheval, July 12, 1816, available at https://founders.archives.gov/
?q=Ancestor%3ATSJN-03-10-02-0128&s=1511311111&r=2, archived at https://perma.cc/97S5-NXM3.
changes in social conditions, and advances in the technologies of governance

now make a fully centralized government better than a federal one.21
Does thin constitutionalism address eternity clauses or amendment rules that

from the point of view of today’s supermajority are overly stringent? The standard
answer, we think, is that if you don’t like the constitution you have, if only be-
cause of its amendment rules or eternity clauses, you can scrap it—technically,
replace rather than amend it.22 Some constitutions, though, say that they can be
amended by some supermajority rule but replaced only through some even more
stringent process.23 And then the obvious question: What if people think that the
replacement rule is too stringent?
Here we want to truncate our presentation. Today’s populists all seem rea-
sonably happy to use existing amendment rules to implement the changes they
favor, as to both substantive rights and procedures for enacting laws (including
changes in judicial power, which some populists have come to see as the most
significant veto gate in their systems).24 For that reason we would say today’s
populists seem mostly happy to comply with thin constitutionalism’s require-
ment, if it be one, that the constitution can be amended only by adhering to the
amendment rules as they happen to be.25
What about amendments adopted pursuant to existing amendment rules
that change the amendment rules themselves? As with other amendments, such
amendments might be problematic if, for example, they increase the difficulty
of amending provisions that make it difficult to displace a government in power.
Other amendments to the amendment rules might not be troublesome in thin
constitutional terms. Once again, we think that the focus should be on the sub-
stance of the amendments and their actual effect in the circumstances at hand
rather on the fact of amendment itself, even the amendment of amendment rules.
We feel some obligation, though, to say something about the largely the-
oretical question about getting rid of bad eternity clauses or overly stringent
21 Some but not all of these difficulties can be handled by creative constitutional interpretation—
finding that some substantial move toward centralization actually is consistent with federalism prop-
erly understood, or finding that some public recognition of a religion is consistent with secularism
properly understood.
22 Sometimes the terminology is “substitution” rather than “replacement,” but the underlying idea
is the same.
23 See Rosalind Dixon & David Landau, Tiered Constitutional Design, 86 George Washington
Law Review 438 (2018).
24 They have sometimes pressed the definition of “amendment” and “replacement,” with critics
contending that what populists have described as an “amendment” is actually a “replacement” that
didn’t go through the more stringent processes for replacements. We provide a few examples in our
case studies, but our general view is that here too a standard of reasonableness applies: Is it reasonable
to treat a change as a (mere) amendment rather than as a replacement?
25 A possible counterexample is provided by the 1999 constitutional revision in Venezuela, which
expressly departed from the requirements in the constitution-in-place for replacing the constitution.
The case is complicated by the fact that the (non-populist dominated) high court held that departing
from the constitution’s requirements was lawful. For a more complete discussion, see Chapter 3.
22 The Framework
amendment rules.26 We think that they can be done away with through what has
to be understood as a revolutionary transformation of the constitutional order,
one that expressly breaks the bounds of the existing constitution.27 Relative to

the existing constitution, revolutionary transformations are extralegal.28
C. A Short Note on Why Thin Constitutionalism

Isn’t about Limiting Government Power
By now readers may be impatient at our failure to say anything about what they
might think at constitutionalism’s heart: constitutionalism, on one common
view, is about limiting government power, to protect the natural liberties people
have and want. The well-regarded Stanford Encyclopedia of Philosophy’s first
sentence on constitutionalism is this: “Constitutionalism is the idea, . . . that gov-
ernment can and should be legally limited in its powers.”29
We don’t include an expansive idea of limited power in our conception of thin
constitutionalism. Of course, whatever happens to be entrenched operates as a
temporary limitation on government power. And equally of course “rule of law”
requirements that are within our conception of thin constitutionalism are them-
selves limitations on government power. Thin constitutionalism places some
limits on pure majoritarianism, but “limiting government power” isn’t the ulti-
mate goal of constitutionalism, as some formulations occasionally suggest. The
reason: we limit government power to protect liberty. Sometimes, though, exer-
cising power protects liberty. The most obvious cases involve the use of national
power to control abuses of power by petty local tyrants who corruptly abuse their
public positions to put people under their thumb. These cases are easy even if we
define “liberty” as something that can be threatened only by public power. That’s
26 The question is not entirely theoretical. The U.S. Constitution provides that each state, no matter
what its population, shall have equal representation in the upper house (currently, two senators per
state), and that no amendment can change the principle of equal representation. Many in the United
States, though almost certainly not a supermajority, believe that that “eternity” provision is a bad one.
And courts are sometimes protected by an eternity clause or, more often, by a judicially created “basic
structure” doctrine.
27 For clarity’s sake we note that a prominent recent work, Bruce Ackerman, Revolutionary
Constitutionalism (2020), employs a notion of revolutionary constitutionalism that is, we think,
quite a bit different from the one so thinly outlined here. Exploring our disagreements would blur our
focus on our primary effort to discuss populism and constitutionalism.
28 The scholarly literature on this point divides between those who argue that revolutionary trans-
formations are not constrained by law in any sense, and those who contend that such transformations
are bound by some (rather thin) specifications of fundamental human rights. We take no position on
this dispute.
29 Wil Waluchow, “Constitutionalism,” The Stanford Encyclopedia of Philosophy (Spring
2018 ed., Edward N. Zalta ed.), available at https://plato.stanford.edu/archives/spr2018/entries/con-
stitutionalism/.
one conceivable definition, but not the only one—and so not a definition that
should be built into the idea of constitutionalism.
Suppose we care about human liberty because it’s a good thing for people to

have. Specifically, liberty is good because it lets people choose the kind of life
they want to live, free from domination. Private power can limit people’s choices,
sometimes more dramatically than public power does: think of a state that barely
is able to keep civic peace, and large employers who offer jobs on the condition
that their workers do pretty much whatever the employer wants pretty much all
of the time. Or compare the everyday intrusions on privacy emanating from gov-
ernment (traffic cameras, perhaps) with the intrusions on privacy built into the
algorithms of our social media giants. Exercising public power against private
power can protect individual liberty understood in this way.30
Our conclusion is that the idea that constitutionalism is about limiting public
power is one possible version of constitutionalism, but it’s not one that everyone
committed to constitutionalism does (or, in our view, should) accept—and so
isn’t part of thin constitutionalism.31
D. Judicial Independence
Thin constitutionalism as described to this point requires only that policies

across a broad range be made according to the preferences of a contemporary
majority as determined by reasonably reliable mechanisms, some but not all
of which are entrenched in the Constitution, and that entrenched policies be
changed only according to the rules regulating constitutional amendments. Thin
constitutionalism also deals with how we enforce and apply the law under the
heading of “separation of powers.”
The “veto gate” concept gives us some insight into the way thin constitu-
tionalism deals with the separation of powers. The concentration of power in
a single institution, we’ve all learned, is the very definition of tyranny. The body
that makes the law shouldn’t implement it, and the body that enforces the law
shouldn’t interpret it. Separation of powers protects our liberty. And, the folk
wisdom that two heads are better than one implies that separation of powers
30 Here are two relevant quotations from classic liberal writers: Montesquieu—liberty means that
“the government must be such that one citizen cannot fear another citizen”; Locke—“The Injury and
the Crime is equal whether committed by the wearer of a Crown, or some petty Villain.” Both are
quoted in Steven Kautz, “On Liberal Constitutionalism,” in The Supreme Court and the Idea of
Constitutionalism (Steven Kautz et al. eds. 2011), at p. 32.
31 Martin Loughlin, The Foundations of Public Law (2010), provides a good account of the
long history of republican and “political” constitutionalism, ideas that don’t rest on principles of neg-
ative liberty.
24 The Framework
often improves the quality of legislation by getting several institutions to agree

before a law becomes effective.
Everyone knows that two heads are better than one, but also that too many

cooks spoil the broth. A system with too many veto gates and speed bumps
can produce an overall body of law that is worse than one with fewer—at least
if you’re not a committed libertarian who thinks that more law is always worse
than less law.32 Policymaking at the national level in the United States has been
stymied by “gridlock” resulting from having too many veto gates. We don’t think
anyone has produced a decent theory telling us how many veto gates and speed
bumps are optimal. We’re reasonably confident, though, that there’s no general
argument against reducing the number of veto gates and speed bumps from
whatever it happens to be in any constitutional system. Eliminating them all, of
course, means consolidating power in a single person, typically an authoritarian
dictator—and that’s clearly inconsistent with thin constitutionalism. Beyond
that there’s little general to say. Sometimes going from seven veto gates to five
won’t be a problem; sometimes it might be.
Thin constitutionalism requires only that there be some separation of powers,
but not much beyond that—with one important qualification: thin constitution-
alism clearly requires that laws be interpreted and applied by courts that are in-
dependent of the legislature and the executive government.
After policies are adopted, they have to be applied. And application will some-
times be controversial. The legislature adopts a law saying that factories that emit
pollution use the “best available technology” to minimize their emissions. A fac-
tory faces a fine for using something less than the best available technology; its
owners say that there’s actually nothing better to deal with the precise pollutant
the factory emits. Thin constitutionalism requires that courts ultimately be avail-
able to resolve this dispute.
We write “ultimately” because thin constitutionalism doesn’t require that the
dispute be immediately sent to a court for final resolution, or even that a prelim-
inary resolution be made by a court, or even that the court refuse to give defer-
ence to a preliminary resolution made outside the courts. Thin constitutionalism
requires only that a court be available to take a look at the dispute and ensure that
the outcome conforms to the anti-pollution law the legislature enacted.
How deep does that look have to be? Clearly, a law can be applied to a person
only when doing so is consistent with a reasonable interpretation of the law. So,
the courts have to interpret the law. Does thin constitutionalism require more
than that? Most contemporary constitutions—but not all—allow the courts to
32 More precisely, by preserving the status quo a larger set of veto gates and speed bumps leaves a
wider range of policies to be determined by the choices made by private parties regulated by private
rather than public law.
hold statutes unconstitutional or “disapply” them in specific cases.33 Almost all

allow the courts to hold executive actions unlawful when the actions aren’t au-
thorized by either the constitution (when the executive uses “decrees” to exercise

its prerogative powers), or by statutes (when the executive uses a different category
of “decrees,” the term used in civil law systems, “secondary legislation,” the term
used in the United Kingdom, or “regulations,” the term used in the United States).
And almost all tell the courts to interpret statutes to be consistent with the con-
stitution where doing so is within the broad bounds of reasonable interpretation.
We note, though, that even where constitutional review is authorized it is not
comprehensive—courts won’t hear some constitutional claims because, for ex-
ample, the challenger wouldn’t benefit from winning on the constitutional issue
or, more important, because the constitution commits the substantive consti-
tutional question to resolution by the political branches (this is known as the
“political questions” doctrine). And some constitutions allow legislatures
using ordinary majority rules to “override” judicial decisions finding a statute
unconstitutional.
We’re agnostic on whether thin constitutionalism requires some form of con-
stitutional review by the courts. That some constitutional systems that almost
everyone would treat as “constitutionalist” get by without robust constitutional
review suggests that it doesn’t.34 But, even if it does, we think the existence in
decent constitutional systems of an “override” power and the political questions
doctrine make it clear that quite thin constitutional review is enough. Context
matters, of course: There’s some reason to believe that long-established democ-
racies can get by with rather weak forms of constitutional review, and some ev-
idence, albeit quite spotty, that on occasion robust review can protect what law
professor Samuel Issacharoff calls fragile democracies from shattering.35 That
the evidence isn’t tremendously strong cautions against including robust review
as a requirement of thin constitutionalism.
For thin constitutionalism the courts’ domain, then, is statutory interpreta-
tion and to some modest degree review of statutes and executive actions for con-
sistency with the constitution. Within that domain, how are courts to behave?
33 Jan-Werner Müller shows how after 1945 “a highly constrained form of democracy” emerged
as a response to the totalitarian excesses of the 1930s and 1940s. Key parts of this “post-war con-
stitutional settlement” were powerful constitutional courts. Jan-Werner Müller, Contesting
Democracy: Political Ideas in Twentieth Century Europe (2011), at pp. 128–40.
34 In designing his proposed constitutional court, Hans Kelsen didn’t include a provision for re-
view of claims that legislation violated individual rights. The post-war paradigm of constitutional law
does require such review, and we think it reasonable to conclude that contemporary constitution-
alism does so as well. But, as we argue in the text, that requirement can be satisfied in ways compatible
with populism’s commitment to majority rule.
35 Samuel Issacharoff, Fragile Democracies: Contested Power in the Era of
Constitutional Courts (2015). See also Carles Boix, Constitutions and Democratic
Breakdowns (2005);
26 The Framework
The standard answer is that constitutionalism, no matter how thin, requires

that courts be independent, but that’s pretty clearly an incomplete answer. The
standard example of lack of judicial independence is “telephone justice”: be-

fore deciding a case, a judge calls a politician and ask what result the politician
wants.36 Beyond that, though, the idea of judicial independence gets compli-
cated, because we want judges to be accountable—to something—as well as
independent. Accountability matters because a completely independent judge
could enforce the law based on his biases against one of the parties or on her idio-
syncratic views about what good public policy requires (by relying on a distorted
interpretation of a clear statute, for example).
Spelling out what accountability means turns out to be surprisingly difficult.
One dimension seems simple enough: judges should be held accountable to “the
law.” Their decisions should rest entirely on reasons drawn from the legal system
itself. That’s one reason telephone justice is bad: it leads judges to make decisions
based on “politics” in a narrow sense, and no legal system worthy of the name
(or at least no legal system that satisfies thin constitutionalism’s requirements)
includes such reasons within the set of permissible ones.
The complications come in when we try to move beyond that. Some legal
systems allow judges to refer openly to policy considerations when they inter-
pret statutes, for example, while others insist that judges rely entirely upon the
statute’s text (itself a complicated proposition). Some legal systems allow judges
to refer to unwritten principles of fundamental human rights when they assess a
statute’s constitutionality; others insist that judges rely only on the written text.
Put generally: what counts as a decision according to law varies from one legal
culture to another. In our view, thin constitutionalism requires that judges justify
their decisions by adverting to the kinds of reasons their legal culture treats as
legal. We write “adverting to” deliberately. We think it a mistake to say that thin
constitutionalism requires that decisions “be” consistent with the law because, as
always, there’s going to be reasonable disagreement about what the law actually
requires. Critics of a decision might say that the judge is only pretending to rely
on reasons made available in the law, and sometimes that’s going to be correct—
but distinguishing between pretense and reasonable disagreement is almost al-
ways going to be an exercise in political rather than legal analysis.
That accountability to the law directs us to examine what kinds of reasons the
legal culture recognizes leads us to a second dimension of accountability. The
legal culture is necessarily an elite culture, which creates a tension between ju-
dicial independence (understood to require accountability to law) and majority
rule.37 Thin constitutionalism resolves the tension by requiring some sort of
36 Or the politician calls the judge to tell her how to rule.

37 In the field of constitutional law, Alexander Bickel’s term “the countermajoritarian difficulty”
captures this concern.
accountability to the public, typically through the mechanisms for appointing

and removing judges. Here too contemporary constitutionalism recognizes
many possibilities. In Australia and Canada judges on high courts are appointed

by the prime minister acting alone, though subject to rather strong cultural
norms; in the United States they are nominated by the chief executive and con-
firmed by the upper legislative house; in states within the United States, judges
are elected;38 in most nations judges are chosen by a judicial selection committee
composed of a mix of politicians and representatives of civil society.39
In light of this variety, we think it impossible to say more than that thin con-
stitutionalism permits (and may require) some reasonable form of judicial ac-
countability to the public. Typically questions about judicial independence and
accountability don’t arise until politicians propose to change the form of judi-
cial appointment or removal. Sometimes critics describe such changes as threats
to judicial independence even if they are on their face a shift from one reason-
able version of accountability through appointment to another, which itself
would have been fine had it been in place all along.40 Such criticisms introduce
a more general issue for thin constitutionalism—the problem of what we call
retrogression—to which we now turn.
E. Two Related Problems—

Specification and Retrogression
We suggested earlier that criticism of constitutional developments that invoke

notions of constitutionalism that are thicker than the thin constitutionalism
we’ve described are quite likely to be problematic. Here we elaborate that sug-
gestion, with the goal of showing that the problems lie in what are ultimately
disagreements with the policies aided by the constitutional developments: criti-
cism that takes the form of concerns about constitutionalism conceals the under-
lying concern, which is with the substantive policies the criticized government is
pursuing.
We’ve already introduced some of the examples we’ll use to develop this ar-
gument. Suppose the critic has a moderately thick notion of constitutionalism
that includes ideas about freedom of expression, equality, individual autonomy,
judicial independence, personal privacy, limits on executive power, and so on
through the list of everything that’s contained in most modern constitutions.
38 For a case study of judicial elections in Bolivia, see Chapter 8.

39 Removal mechanisms include impeachment and, more important, term limits and age limits on
sitting judges.
40 Our case studies of the judicial appointment process in India and Israel in Chapter 8 examine
this question.
28 The Framework
Almost always these ideas will be expressed in general terms. The U.S.
Constitution says that “Congress shall make no law . . . abridging the freedom
of speech”; the German Basic Law says, “Every person shall have the right freely

to express and disseminate his opinions in speech, writing and pictures and to
inform himself without hindrance from generally accessible sources.” The U.S.
Constitution says, “No state shall . . . deny to any person . . . the equal protec-
tion of the laws.” The German Basic Law says, “All persons shall be equal before
the law.”
These provisions are typical in the level of generality with which they state
the governing principles. In specific cases or in connection with specific pieces
of legislation, though, the general principles can be brought to ground in many
different ways. To take the most obvious example, essentially every statute treats
some people differently from others and the principle of “equality before the
law” has to take that obvious fact into account. Consider a problem the Japanese
constitutional court faced. We know that the principle of equality before the
law allows the government to punish murderers more severely than robbers.
A statute imposed greater punishment on people who killed one of their parents
than on those who killed strangers. Is that statute consistent with equality before
the law? (The court held that it wasn’t.) What about a ban on video game and pool
parlors near schools but not farther away? The South Korean constitutional court
upheld a ban on billiard parlors near elementary and high schools but didn’t let
the government ban them near kindergartens and universities.41
Or, to return to earlier examples, the principle of free speech can be specified
in ways that make hate speech regulation permissible or unconstitutional: For
the first, specify the principle as “no discrimination on the basis of content”;
for the second, specify it as “restrictions on expression must be proportional
to the harm caused by the regulated expression.” Judicial independence can be
protected by vetting judges through a judicial appointment commission or by
insisting that the legislature and executive agree on judicial appointments.
Of course, you can define constitutionalism as requiring one or another of
these specifications down through the entire list. If you do, though, we’re going
to end up with hundreds of competing definitions. And almost all of them
are going to be reasonable ways of specifying the general principles of consti-
tutionalism. That means that thickly specified constitutionalisms—and the
plural is important—can’t be the basis for critical evaluation of constitutional
developments. Or at least it can’t be such a basis when the developments shift a
nation’s constitutional order from one reasonable specification to another. Only
41 See Shingori Matsui, Why Is the Japanese Supreme Court So Conservative?, 88 Washington
University Law Review 1375 (2011), at pp. 1388–89; Yoon Jin Shin, “Proportionality in South
Korea: Contextualizing the Cosmopolitan Rights Grammar,” in Proportionality in Asia (Po Jen
Yap ed. 2020), at pp. 90–91.
our thin constitutionalism supports criticisms that reach beyond parochial views
of what constitutionalism requires.
That’s the general argument—a system consisting of reasonable specifications

of foundational constitutional principles can’t be criticized as anti-constitutional
except under quite specific circumstances (circumstances that we explore in our
case study of Hungary as a “Frankenstate”). But our statement has itself been
quite abstract. We make it more specific with some examples.
Free expression: libeling public officials. Public officials have to expect that
•
someone sometime will criticize them quite severely for what they’ve done
either in their public roles or in their private lives. That criticism can harm
their reputations. When the criticism is unfair—for example, when it rests
on a factually false premise—a legal system might reasonably give the public
official a remedy for the harm to reputation. That’s what the law of libel does.
We’ll assume, we think accurately, that contemporary constitutionalism
bars nations from adopting the classical common law rules of libel law: strict
liability for publishing reputation-damaging falsehoods, a presumption
that false statements do harm reputation, the burden of proving truthful-
ness on the publisher. Even with those elements ruled out, many possibil-
ities remain. The United States gives false statements about public officials
a high degree of protection: the official loses unless she can show that the
falsehoods were published by someone who either knew that they were false
or acted with “reckless disregard” of their truth or falsity. Elsewhere public
officials can win if they show that the publisher failed to comply with pro-
fessional standards for verifying factual statements before publishing them.
In yet other nations the officials can win if they prove that the statements
were false and provide evidence that the statements did indeed harm their
reputations.
Now consider a government that proposes to change the libel rules cur-
rently operating—say, from the stringent U.S. rule to the somewhat less
speech protective “journalistic standards” one. Is this intrinsically a “retro-
gression”—an erosion of constitutionalism?
Here are two arguments why it isn’t. First, the government might provide
reasons for the change, pointing for example to changes in the media envi-
ronment that have generated a much larger number of scurrilous and false
allegations about public officials that had been the case when the highly
speech-protective rule was adopted (with the effect, perhaps, of making it
more difficult to attract people to public service). If that’s a reasonable argu-
ment, we think it difficult to describe the change as a retrogression; it’s just
a change. Second, and more straightforward, if the “journalistic standards”
rule had been in place from the beginning, we wouldn’t worry that the
30 The Framework
nation was somehow falling short of constitutionalism. Changing from one

reasonable specification of freedom of speech to another is, again, just a
change.

Suppose, though, that we’re confident that the government’s proposal is
motivated by a desire to punish its critics. Within broad limits, shouldn’t the
answer be that it depends upon whether the critics are right or wrong—not
in their factual assertions, but in their core criticisms? If the government’s
pursuing good policies, scurrilous and false allegations against it are bad in
part because they impede its ability to carry out those policies. And if the
government’s pursuing bad policies, we want to minimize the risks critics
face, including risks of lability under libel law.
• Judicial independence: packing the courts. Suppose a progressive government
takes office with an ambitious program—a “Green New Deal” or an innova-
tive tax program to soak the rich. The principle of majority rule tells us that
the government should be given a fair shot to implement its program. The
government’s legal advisers look at the veto gates the government’s program
has to get through. They identify three: a crucial committee in the lower
house that’s likely to have jurisdiction over key aspects of the government’s
program; a practice—the filibuster—in the upper house that in effect allows
a minority to block the program; and the nine-person constitutional court
whose current majority seems likely to obstruct the government’s ability to
implement its program. The advisers make three proposals: send the legisla-
tion to some other committee by changing the lower house’s rules, eliminate
the filibuster, and add two judges to the constitutional court. The advisers
note that the constitution expressly gives each legislative house the power to
determine its own rules and that nothing in the Constitution entrenches the
number “nine” as the court’s size. Before enacting the rest of its program, the
government implements all three recommendations and packs the court by
adding two members. Is this “court-packing” anti-constitutional?42
The same arguments we’ve deployed earlier work here as well. Changing
the legislature’s rules as the constitution permits can’t plausibly be
described as anti-constitutional. The same goes for changing the court’s
size. An eleven-member constitutional court is well within contemporary
parameters, and no one would blink if the court had had eleven members
all along. Each of the three actions removes a different veto gate. The only
reason to allow the removal of the first two but worry about removing
the third would be some sort of quasi-religious veneration for the courts.
42 The problem gets more complicated if the court’s size is constitutionally entrenched. All is well if
the government’s majority is large enough to secure a constitutional amendment. If it isn’t, the ques-
tion we discussed earlier—what to when the amendment rule in place is too stringent—arises.
Maybe that sort of veneration is a good thing (though we doubt it), but we’re
reasonably confident that constitutionalism doesn’t require it.

F. Politicians and Political Parties
We’ve already relied on the final component of thin constitutionalism—the im-

portant role that politicians and political parties play, especially but not exclu-
sively when majority preferences are determined by elections.43 Politicians and
their parties present political agendas to the public, and voters choose among the
parties on the basis of the agendas they offer.
The other components of thin constitutionalism are almost entirely norma-
tive, identifying the “oughts” that define thin constitutionalism. This component
has some normative content but is primarily descriptive: under contemporary
circumstances reliably determining what a majority prefers is as a matter of fact
greatly facilitated by politicians and political parties.
The focus here is on why voters (as a shorthand for “the public”) support a
politician or a political party, not on the politicians. A politician might quite
cynically offer a political agenda to the public not because she believes that
implementing it would be good for the country but simply because she thinks
it’s an agenda that a majority will support, and she wants to be elected (and re-
elected) simply because she like to exercise power (or thinks that she’ll benefit
financially more from being elected than from anything else she might do). The
politician’s motivation doesn’t matter, from the point of view of thin constitu-
tionalism; the agenda the voters favor does.
One important qualification: consider a cynical politician who simply wants
to exercise power. He wins by offering an agenda a majority favors (with the
imperfections we’ve noted earlier about how party platforms only imperfectly
reflect majority preferences). Then he proposes constitutional changes, using
the existing amendment rules, that actually facilitate the implementation of that
agenda. By removing some veto gates or speed bumps, the changes make it easier
for the politician to accomplish anything and, at some point, the politician’s un-
derlying motivation to retain power comes to the fore. He implements a pro-
gram that increases his power but actually doesn’t have majority support, but
with the veto gates and speed bumps destroyed, there’s almost no way other than
domestic turbulence to stop the now autocratic leader from hanging on to power.
43 Politicians and political parties can propose referendums and other mechanisms for deter-
mining majority preferences outside of elections, and can take positions on referendums and the like.
Our exposition is eased by focusing solely on elections, but our general points apply to politicians’
role in every preference-determining mechanism.
32 The Framework
This is indeed a problem for thin constitutionalism. But, as we’ll argue

in detail later, it’s a problem associated with majority rule, not populism. The
problem arises from the possibility that a proto-authoritarian can win an elec-

tion by presenting a platform that a perhaps temporary majority prefers to the
alternatives on offer. And the most obvious remedy, we think, is to hold elections
rather frequently as a way of making it more likely that the government’s agenda
continues to have majority support.44
We might want to thicken constitutionalism to prevent this outcome. Some of
our case studies suggest, though, that sometimes elections are enough to prevent
proto-authoritarians from entrenching themselves. And, as important, thick-
ening constitutionalism to prevent this outcome necessarily limits the ability
of contemporary majorities to get what they want—it limits majority rule. At
the level of constitutional theory and design, the relevant questions are the fol-
lowing: How great is the risk that a proto-authoritarian will win enough elections
to become a true authoritarian ruler? How important are the policies that will be
blocked by the thicker constitutional provisions used to prevent that outcome?
And how likely is it that those policies would be adopted in a thin constitutional
system? Our most general argument is that those questions can’t be answered
simply by saying that the politicians we worry about (today) happen sometimes
to present themselves as populists, but have to be addressed by looking in detail
at specific national conditions. Or, in an oversimplified tag line: we should worry
about proto-authoritarian politicians, not populist politicians.
III. Constitutionalism, Norms, and Guardrails
The constitutionalism we’ve been discussing so far, thin or thick, is all about
institutions—elections, courts, and more. For some time now, though, it’s been
clear that there’s more to constitutionalism than institutions. Constitutionalism,
again whether thin or thick, can be sustained only if there’s something like a
pro-constitutional culture behind it, a commitment by most of a nation’s people
to the idea of constitutionalism itself. And, we’ve come to understand, a pro-
constitutional culture is sustained by norms and guardrails, to use terms popu-
larized by political scientists Steven Levitsky and Daniel Ziblatt.45
Norms and guardrails are political practices that guide the behavior of
politicians and ordinary citizens. For Levitsky and Ziblatt, the most important
thing is that you treat the people you disagree with about politics as wrong but
44 Of course, if the opposition doesn’t offer an attractive alternative, as our case study of Venezuela
suggests, elections aren’t going to help.
45 Steven Levitsky & Daniel Ziblatt, How Democracies Die (2018).
not your enemies. They have misguided views about what’s good for the country,
but they’re working with you in the common political endeavor of making the
country better. That’s the most general one, but it gets worked out in more partic-

ular practices: You can protest when the legislature adopts policies you disagree
with but you can’t threaten the people you oppose with violence; you can go to
court to find out whether they’ve done something unconstitutional, but you have
to accept defeat when the courts rule against you—continuing the fight through
ordinary political means.
Of course, we agree that a pro-constitutional culture, norms, and guardrails
are important, yet as lawyers with some feeling for political science we have to be
modest about what we can usefully say.
Where does a pro-constitutional culture come from? The classic answer is,
“The family, education, religion, and some forms of economic organization.”
Responding to social developments, we (and others) would add “civil society”
broadly understood and certain types of political parties (those in which mem-
bership requires active engagement within the party) to the list.46 Lawyers
can’t contribute much to a discussion of how those institutions promote a pro-
constitutional culture. We can of course say things about the ways in which
constitutions can insulate the institutions from outside intrusion, thereby giving
them space to be pro-constitutional. Yet, when they have that space, they can be
anti-constitutional as well, and lawyers have nothing to say about how, for ex-
ample, a family should be organized so as to make it likely that its members will
be pro-rather than anti-constitutional. (We can’t resist moving outside our dis-
ciplinary lane a bit, though, to observe that in our view the world of the twenty-
first century places enormous pressures on the institutions traditionally thought
to foster a pro-constitutional culture. Our discussion at the book’s conclusion
of some pro-democratic methods of populist governance might be construed
as describing new institutions that can perform the pro-constitutional function
that these traditional ones no longer can.)
When we move from culture to norms and guardrails, we think a legal per-
spective can be helpful, though the legal story we tell is a bit depressing. The
problem starts with the observation, obvious as soon as it’s stated, that norms
aren’t written down (if they were, they’d be what we’ve called institutions). That
generates quite a few difficulties. We describe them here in an abstract form; our
46 A word about two items on the list: some classical political theorists supported what’s known
as the “doux commerce” theory, according to which the ordinary face-to-face interactions between
traders induced mutual tolerance of disagreements (after all, traders had to bargain with each
other). Agrarian theorists argued to the same effect about how the daily activity of farming and sel-
ling what one raised created communities of mutual tolerance. And not all political parties have
pro-constitutional effects. For an argument that elite-dominated conservative parties have some-
times facilitated the displacement of democracy by autocracy, see Daniel Ziblatt, Conservative
Parties and the Birth of Democracy in Europe (2017).
34 The Framework
discussion in Chapter 8 of judicial reform proposals in the United States gives the
arguments concrete content, and Chapter 10 examines the arguments in detail.
Here we preview the arguments we make there.

When we look at some specific political practice or some series of political
actions, we can’t be sure that there’s a norm or guardrail in place at all—or, per-
haps more precisely, when some people do something that other people disagree
with, the claim that the former group are trashing a norm or guardrail can be-
come just another version of the underlying political disagreement: “You say that
there’s a norm that we’re violating, but you’re wrong: this particular thing we’re
doing operates in a norm-free domain where you can act pro-constitutionally
simply by exercising whatever political power you happen to have.”
Even when people can agree that there’s a norm in the premises, they can—and
will—disagree about what the norm is and so they will disagree about whether
one side is breaching a norm. The reason is obvious: by definition norms aren’t
written down; they’re content is inferred from practices people have engaged in
in the past (recent or remote).
As lawyers familiar with the common law tradition, we do know something
about inferring content from practices. Legal rules in a common law system
aren’t written down in canonical texts. They are inferred from decisions made by
judges in the recent or remote past. We figure out what the rules are by standard
methods of common law reasoning: We describe a prior case as resting upon a
principle broader than the narrowest one necessary to justify the result, and then
apply that principle to the problem we’re facing; we say that a court’s statement of
a broad principle to explain its result was dictum and that the case should be read
in light of the specific facts at hand; we draw analogies for the present case from
decisions dealing with similar but slightly different facts; and much more.
Arguing about what norms are in place, and about whether something contro-
versial violates a norm, is just like that. The effect is to reproduce the basic disa-
greement about policy in the discussion of norms and their purported violation.
A pro-constitutional culture, norms, and guardrails might well be centrally
important to sustaining constitutionalism, and such a culture can be nurtured
in many institutions, though whether it actually is nurtured depends a lot on
how those institutions are structured. As lawyers our contribution must be lim-
ited: some skepticism about whether we get any purchase on the relation be-
tween constitutionalism and populism by focusing on norms and guardrails.
Thin constitutionalism might be enough to sustain democracy, and more de-
mocracy might lead ordinary people to commit themselves even more deeply to
democracy, a possibility we hint at in Chapter 12. We aren’t completely agnostic
about the ability of the institutions that legal scholars tend to pay most atten-
tion to, to sustain a pro-constitutional culture, but we don’t want to exaggerate
the contributions we can make with the distinctive tools that lawyers can deploy
and use.

IV. Thin Constitutionalism’s Attractions
Consider a person who supports a political agenda. She is committed to de-

mocracy, which means that she wants to see that program implemented, but
only if she can persuade a majority of her co-citizens to support it as well. She
also is committed to principles of constitutionalism. We believe that no matter
what her political agenda, she would be satisfied with the principles of thin
constitutionalism—majority rule, amendable entrenched provisions (with the
possibility of replacing the existing constitution if she garners sufficient sup-
port from others), an independent judiciary. And—this is the central point—she
might not be satisfied with anything thicker because her political agenda might
be inconsistent with anything thicker. Eternity clauses are the most obvious
example, but so would be thick conceptions of free expression, individual au-
tonomy and privacy, and civic equality.
Here’s another way of making this point. Spell out what you think constitu-
tionalism requires, and ask yourself whether someone with a substantially dif-
ferent political agenda from yours would be unreasonable were she to reject the
constraints on policy choices that version imposes. We believe that only thin
constitutionalism fills the bill—that only anti-democrats can reasonably reject
its principles.47
The payoff for the purposes of this book is simple and clear. we should examine
each version of contemporary populism to see whether it is consistent with thin
constitutionalism. Asking for more, in constitutional terms, from any specific
version of populism is to conceal disagreement with that populism’s program. It
puts a constitutional mask over political disagreement. When that occurs, as we
argue in the following chapters, we believe that scholarly honesty requires that
we insist that the mask be taken off. And, though we are less certain about this,
we hope that unmasking the politics underneath the nominally “constitutional”
criticism will enhance the political effectiveness of the critics’ positions.
47 This defense of thin constitutionalism resembles well-known arguments in political theory—

the idea associated with John Rawls that we try to figure out the principles of justice we would choose
behind a veil of ignorance, and the related but somewhat different idea that principles of justice have
the characteristic that no one could reasonably reject them. We should be clear that we are not relying
upon those arguments to justify thin constitutionalism, only pointing out the similarity between
them and our argument.

Tushnet & Bugaric, What Is Constitutionalism

Uploaded by

Copyright:

Available Formats

Tushnet & Bugaric, What Is Constitutionalism

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Tushnet & Bugaric, What Is Constitutionalism

Uploaded by

Copyright:

Available Formats

1

Downloaded from https://academic.oup.com/book/38889/chapter/338031927 by Harvard University Library user on 19 February 2023

kinds of political demonstrations that must be permitted—​where the “must be”

Downloaded from https://academic.oup.com/book/38889/chapter/338031927 by Harvard University Library user on 19 February 2023

We don’t contend that thin constitutionalism is the most attractive version of

Downloaded from https://academic.oup.com/book/38889/chapter/338031927 by Harvard University Library user on 19 February 2023

actually occur. (Complete dis-​embedding is impossible without abandoning “constitutionalism” as

Downloaded from https://academic.oup.com/book/38889/chapter/338031927 by Harvard University Library user on 19 February 2023

II. The Elements of Thin Constitutionalism

Thin constitutionalism today has four elements.

• Majority rule: Policy over a wide range of issues is determined by the

A. Majority Rule

Contemporary governments can address an enormous number of issues: tax

Downloaded from https://academic.oup.com/book/38889/chapter/338031927 by Harvard University Library user on 19 February 2023

Downloaded from https://academic.oup.com/book/38889/chapter/338031927 by Harvard University Library user on 19 February 2023

Innovations in the technology of democracy now offer additional mechanisms

Downloaded from https://academic.oup.com/book/38889/chapter/338031927 by Harvard University Library user on 19 February 2023

Downloaded from https://academic.oup.com/book/38889/chapter/338031927 by Harvard University Library user on 19 February 2023

11 Proponents of Catalan independence organized a public “consultation”—​really, a referendum

We can imagine a political theory in which the substantive view prevails.

Downloaded from https://academic.oup.com/book/38889/chapter/338031927 by Harvard University Library user on 19 February 2023

There’s a broader payoff, though. Thin constitutionalism means that polit-

Downloaded from https://academic.oup.com/book/38889/chapter/338031927 by Harvard University Library user on 19 February 2023

Thin constitutionalism, we have argued, requires that some things—​specifically,

Other well-​functioning constitutional democracies have quite different systems

Downloaded from https://academic.oup.com/book/38889/chapter/338031927 by Harvard University Library user on 19 February 2023

Amendments become sensible when experience reveals that the constitution’s

Downloaded from https://academic.oup.com/book/38889/chapter/338031927 by Harvard University Library user on 19 February 2023

Finally, technological developments can make new techniques of governance

changes in social conditions, and advances in the technologies of governance

Downloaded from https://academic.oup.com/book/38889/chapter/338031927 by Harvard University Library user on 19 February 2023

Downloaded from https://academic.oup.com/book/38889/chapter/338031927 by Harvard University Library user on 19 February 2023

C. A Short Note on Why Thin Constitutionalism

Downloaded from https://academic.oup.com/book/38889/chapter/338031927 by Harvard University Library user on 19 February 2023

D. Judicial Independence

Thin constitutionalism as described to this point requires only that policies

often improves the quality of legislation by getting several institutions to agree

Downloaded from https://academic.oup.com/book/38889/chapter/338031927 by Harvard University Library user on 19 February 2023

hold statutes unconstitutional or “disapply” them in specific cases.33 Almost all

Downloaded from https://academic.oup.com/book/38889/chapter/338031927 by Harvard University Library user on 19 February 2023

The standard answer is that constitutionalism, no matter how thin, requires

Downloaded from https://academic.oup.com/book/38889/chapter/338031927 by Harvard University Library user on 19 February 2023

36 Or the politician calls the judge to tell her how to rule.

accountability to the public, typically through the mechanisms for appointing

Downloaded from https://academic.oup.com/book/38889/chapter/338031927 by Harvard University Library user on 19 February 2023

E. Two Related Problems—​

We suggested earlier that criticism of constitutional developments that invoke

38 For a case study of judicial elections in Bolivia, see Chapter 8.

Downloaded from https://academic.oup.com/book/38889/chapter/338031927 by Harvard University Library user on 19 February 2023

Downloaded from https://academic.oup.com/book/38889/chapter/338031927 by Harvard University Library user on 19 February 2023

nation was somehow falling short of constitutionalism. Changing from one

Downloaded from https://academic.oup.com/book/38889/chapter/338031927 by Harvard University Library user on 19 February 2023

Downloaded from https://academic.oup.com/book/38889/chapter/338031927 by Harvard University Library user on 19 February 2023

We’ve already relied on the final component of thin constitutionalism—​the im-

This is indeed a problem for thin constitutionalism. But, as we’ll argue

Downloaded from https://academic.oup.com/book/38889/chapter/338031927 by Harvard University Library user on 19 February 2023

III. Constitutionalism, Norms, and Guardrails

Downloaded from https://academic.oup.com/book/38889/chapter/338031927 by Harvard University Library user on 19 February 2023

Downloaded from https://academic.oup.com/book/38889/chapter/338031927 by Harvard University Library user on 19 February 2023

kinds of political demonstrations that must be permitted—where the “must be”

actually occur. (Complete dis-embedding is impossible without abandoning “constitutionalism” as

• Majority rule: Policy over a wide range of issues is determined by the

11 Proponents of Catalan independence organized a public “consultation”—really, a referendum

Thin constitutionalism, we have argued, requires that some things—specifically,

Other well-functioning constitutional democracies have quite different systems

E. Two Related Problems—

We’ve already relied on the final component of thin constitutionalism—the im-

47 This defense of thin constitutionalism resembles well-known arguments in political theory—