The University of Chicago Law Review The University of Chicago Law Review

Hume's Second-Best Constitutionalism
Author(s): Adrian Vermeule

Source: The University of Chicago Law Review, Vol. 70, No. 1, Centennial Tribute Essays
(Winter, 2003), pp. 421-437
Published by: The University of Chicago Law Review
Stable URL: http://www.jstor.org/stable/1600567
Accessed: 05-06-2016 19:38 UTC
REFERENCES
Linked references are available on JSTOR for this article:
http://www.jstor.org/stable/1600567?seq=1&cid=pdf-reference#references_tab_contents
You may need to log in to JSTOR to access the linked references.
Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at
http://about.jstor.org/terms
JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of content in a trusted
digital archive. We use information technology and tools to increase productivity and facilitate new forms of scholarship. For more information about
JSTOR, please contact support@jstor.org.
The University of Chicago Law Review is collaborating with JSTOR to digitize, preserve and extend
access to The University of Chicago Law Review
This content downloaded from 150.216.68.200 on Sun, 05 Jun 2016 19:38:51 UTC
All use subject to http://about.jstor.org/terms
Hume's Second-Best Constitutionalism
Adrian Vermeulet
In the economic and political science literature on constitutional

choice and constitutional design, David Hume is almost invariably as-
sociated with a master principle of design that is congenial to rational-
actor approaches. I shall call this the knavery principle. As Hume put
it, "in contriving any system of government, and fixing the several
checks and controuls of the constitution, every man ought to be sup-
posed a knave, and to have no other end, in all his actions, than private
interest."' The knavery principle is, however, untenable, as I shall sub-
sequently discuss. So it might seem that Hume has nothing valuable to
tell us about constitutional design.
My concern here is to dispel that impression, and by so doing to
reevaluate and vindicate Hume's contributions. I shall argue that
Hume's principal contribution is not the knavery principle. Indeed
one of his major claims about the British constitutional order was that
it violated the principle, and one of his major projects was to explain
how the mixed British constitution could nonetheless retain its stabil-
ity. Hume's real contribution, rather, is to have pioneered the critical
idea of second-best constitutionalism: the idea that multiple depar-
tures from the optimal or first-best constitutional arrangements might
offset each other, producing compensating adjustments that ensure
constitutional equilibrium. Hume's characteristic mode of analysis, I
shall suggest, is simultaneously to identify both a departure from op-
timal constitutional design and an offsetting institutional adjustment
that compensates for the initial defect.
t Professor of Law, The University of Chicago. Thanks to David Currie, Jack Goldsmith,
Don Herzog, Eric Posner, David Strauss, Cass Sunstein, and participants at a University of Vir-
ginia workshop for helpful comments, to Eric Truett for excellent research assistance, and to the
Russell J. Parsons Research Fund for financial support. Special thanks to Yun Soo Vermeule.
1 David Hume, Of the Independency of Parliament, in David Hume, Essays: Moral, Politi-
cal, and Literary 42 (Liberty Fund 1985) (Eugene F Miller, ed). For works associating Hume with
the knavery principle, see Lewis A. Kornhauser, Virtue and Self-Interest in the Design of Constitu-
tional Institutions, 3 Theor Inq in L 15-16 (2002); Franklin A. Kalinowski, David Hume on the
Philosophic Underpinnings of Interest Group Politics, 25 Polity 355, 369-74 (1993); David M.
Kirkham, European Sources of American Constitutional Thought before 1787,3 USAFA J Legal
Stud 1, 22 (1992). For other versions of the knavery principle, see John Stuart Mill, Considera-
tions on Representative Government, in 19 Collected Works of John Stuart Mill: Essays on Politics
and Society 371, 505 (Toronto 1977) (J.M. Robson, ed); Immanuel Kant, Perpetual Peace 29-30
(Bobbs-Merrill 1957) (Lewis White Beck, ed).
421
422 The University of Chicago Law Review [70:421
The idea of second-best constitutionalism has important implica-

tions for American constitutional law. Many positions and arguments,
in both structural constitutional law and the law of constitutional
rights, appeal to the second-best idea of compensating adjustments; I
shall canvass these arguments and indicate the principal lines of
methodological criticism that can be advanced against them. Most im-
portantly, the ubiquity of compensating adjustments does not entail
that any particular constitutional actor should attempt to supply such
adjustments, least of all judges deciding constitutional cases. I shall
suggest that Hume's conception of second-best constitutionalism im-
plies (or is at least compatible with) a modest conception of constitu-
tional adjudication: Compensating adjustments should be welcomed
where they exist, but judges ought not attempt to identify and pro-
mote them. If judicial attempts to engineer improvements to the con-
stitutional structure often make things worse, not better, then a sec-
ond-best account of constitutional adjudication itself entails that
judges should ignore substantive second-best arguments in constitu-
tional decisionmaking.
Two caveats are appropriate. First, the point of studying Hume's
constitutionalism is presentist, not historical. I mean to plunder Hume
for ideas with fertile applications, and I shall disregard historical fidel-
ity if there is an opportunity to interpret or reconstruct Humean ideas
in ways that enhance their present utility. Second, I shall focus on
Hume's analysis of collective behavior and its institutional determi-
nants, largely ignoring the extensive literature on Humean moral and
political psychology at the level of individuals. Hume's psychology,
with its distinctions among reason, passion, and interest, had great in-
fluence on the Framers, but I shall leave the originalist implications of
Hume's ideas unexplored.
Part I explains the important failings of the knavery principle.
Part II examines Hume's characteristic analytic style, with its appeals
to offsetting violations of ideal institutional design. Part III turns from
British to American constitutionalism, fleshing out the Humean idea
of second-best constitutionalism and examining its implications for
contemporary constitutional interpretation and adjudication. A brief
conclusion follows.
2 See, for example, Morton White, Philosophy, The Federalist, and the Constitution 13-22
(Oxford 1987); Douglass Adair, "That Politics May Be Reduced to a Science": David Hume,
James, Madison, and the Tenth Federalist, 20 Huntington Library Q 343,348-57 (1957) (describ-
ing the influence of Hume on Madison's writings on factions).
2003] Hume's Second-Best Constitutionalism 423
I. KNAVES AND KNIGHTS
The initial puzzle is why anyone might endorse the knavery prin-
ciple as a master principle of constitutional design. As Hume puts it,
"[I]t appears somewhat strange, that a maxim should be true in poli-
tics, which is false in fact."3 It is not true that public officials, or for that
matter people generally, always pursue rational self-interest, except in
the tautological sense that reduces altruistic behavior and morally-
motivated behavior to interdependent utility functions4 and a "taste
for fairness,"'5 respectively. In Hamilton's words, the "assumption of
universal venality in human nature is little less an error in political
reasoning, than the supposition of universal rectitude."6 But the em-
pirical question turns out to be a distraction. The standard defense of
the principle, stemming from work in the economics of constitutional
choice, casts it as a methodological commitment rather than an em-
pirical claim. The knavery principle is factually erroneous, yet (on this
view) represents the best postulate by which to assess constitutional-
design proposals.
The methodological defense of the principle rests on four claims.7
First is a claim from symmetry or parsimony: If people are best mod-
eled as rational maximizers of self-interest in explicit markets, the
same model should be applied to implicit political markets. By itself
the appeal to symmetry does not, of course, tell us whether rational
self-interest is the correct uniform assumption to use, but the implicit
claim is that explicit markets can't sensibly be modeled on any other
assumption. Second is a kind of maximin argument, or more loosely
an argument based on avoiding the worst-case scenario: If the harms
from self-interested official behavior are greater than the benefits of
public-spirited official behavior-if knaves do more harm than
knights do good -then constitutional rules should be designed on pes-
simistic assumptions to minimize the worst possible outcomes rather
than to maximize the best possible outcomes. Third is a selection
claim: Just as non-profit-maximizing behavior is driven out of explicit
markets by the bankruptcy of non-profit-maximizing firms, so too
public-regarding behavior will be driven out of political markets be-
3 Hume, Independency of Parliament at 42-43 (cited in note 1).

4 See, for example, Harold M. Hochman and James D. Rodgers, Pareto Optimal Redistri-
bution, 59 Am Econ Rev 542,542-43 (1969) (using assumptions of interdependent preferences to
model redistribution).
5 Louis Kaplow and Steven Shavell, Fairness versus Welfare 11-12 (Harvard 2002).
6 Federalist 76 (Hamilton), in Benjamin Fletcher Wright, ed, The Federalist 480,483 (Har-
vard 1966).
7 These claims are advanced in Geoffrey Brennan and James M. Buchanan, The Reason of
Rules: Constitutional Political Economy 46-66 (Cambridge 1985), and qualified or repudiated in
Geoffrey Brennan and Alan Hamlin, Democratic Devices and Desires 17-33 (Cambridge 2000).
cause public-regarding officials will lose out to officials who strictly

maximize success in elections and other competitive selection proc-
esses. Fourth is an analogy to Gresham's law: Self-interested behavior
will drive out public-spirited behavior, not by selection processes, but
by forcing public-spirited actors to behave as though they were self-
interested, simply to protect themselves from the depredations of
their self-interested competitors.
All of these claims are questionable. Methodologically, symmetry
or parsimony is always at best a weak and defeasible commitment; it
counsels the theorist to reject unnecessary complexity, adopting like
assumptions to model like settings, but the opposite counsel is that the
theorist should reject excessive simplicity, and should tailor his as-
sumptions to capture relevant differences in different settings.8 To
treat explicit markets and politics under identical assumptions if the
two domains are governed by fundamentally different forces is just as
much a mistake as it is to treat them under different assumptions if
the two domains are fundamentally similar. The appeal to symmetry,
then, is parasitic on a suppressed judgment, or guess, that the same
motivations do in fact dominate both political and economic institu-
tions.
Substantively, the knavery principle is excessively static. It ignores
critical dynamic effects, by virtue of which the best maxim for consti-
tutional design may be to assume that public actors are at least het-
erogeneous in their motivations, rather than uniformly self-interested.
To design institutions on pessimistic assumptions about the motiva-
tions of public actors may simply create a self-fulfilling expectation,
one that itself causes public actors to hold the self-interested motiva-
tions that the knavery principle attributes to them. Two mechanisms
produce this self-fulfilling expectation: the endogeneity of official mo-
tivations and the selection effects of institutional design.9
First, the self-interest assumption may crowd out public-spirited
motivations.'? Providing material rewards for altruistic behavior may
reduce the incidence of the desired behavior if actors are altruistically
motivated."' The reward reduces the cost of the behavior, but the off-
setting effect is that the provision of the reward reduces the utility
that public-spirited actors derive from performing it, if the reward
suggests that the behavior is motivated by venality rather than altru-
8 See Albert 0. Hirschman, Against Parsimony: Three Easy Ways of Complicating Some
Categories of Economic Discourse, 1 Econ & Phil 7,8-9,20 (1985).
9 See Kornhauser, 3 Theor Inq in L at 18-19 (cited in note 1).
10 See Bruno S. Frey, A Constitution for Knaves Crowds Out Civic Virtues, 107 Econ J 1043,
1044-45 (1997).
11 See Richard Titmuss, The Gift Relationship: From Human Blood to Social Policy 223
(Pantheon 1971).
ism. Conversely, on the punishment side, attaching material sanctions

to self-interested behavior may increase the incidence of the unde-
sired behavior. Punishment increases the relative cost of the behavior,
but there are potential offsetting effects here as well: The implicit an-
nouncement that "a fine is a price" 12 may undermine moral or social
norms that otherwise inhibit the behavior, and may signal that viola-
tions of the rule are widespread, thus implying that only chumps or
suckers voluntarily comply. In both cases, material sanctions premised
on the assumption of self-interested behavior can inhibit the internali-
zation of public-spirited motivations. Designing institutions for knaves
may itself beget knaves.
These crowding-out effects suppose that the identity of public of-
ficials is held constant, but that the motivations of officials are en-
dogenous. Public-spirited motivations are driven out by institutional
structures even though the officeholders would have held public-
spirited motivations under alternative structures. A second class of
dynamic models reverses these assumptions, holding motivations con-
stant while treating the identity or selection of officials as endoge-
nous.13 Even if actors' motivations are exogenously fixed, the knavery
principle may effect its self-fulfilling prophecy through the selection of
selfishly motivated actors from a heterogeneous pool of selfish and
unselfish candidates. Altruistic actors experience a differential cost
from operating in an environment that supposes all actors to be self-
ish; imagine a public-spirited official whose opportunities to serve the
public interest are frustrated by the elaborate monitoring devices put
in place to check selfishly motivated officials.14 Altruistic actors who
anticipate this differential cost of officeholding will invest less in seek-
ing public office, all else equal, and will thus tend to leave the field to
selfishly-motivated actors who experience no such costs. This reverses
the selection argument for the knavery principle: Designing institu-
tions for knaves creates a system in which knaves operate comfortably,
while knights decline to take public office.
12 Uri Gneezy and Aldo Rustichini, A Fine is a Price, 29 J Legal Stud 1, 14 (2000) (finding
that the number of late parents at day care increased after a fine for lateness was implemented,
and concluding that "[n]o guilt or shame ... can be attached to the act of buying a commodity at
will").
13 See Brennan and Hamlin, Democratic Devices and Desires at 87-97 (cited in note 7);
Geoffrey Brennan, Selection and the Currency of Reward, in Robert E. Goodin, ed, The Theory
of Institutional Design 256,258 (Cambridge 1996).
14 For examples of this frustration, see Philip K. Howard, The Lost Art of Drawing the
Line: How Fairness Went Too Far 129-31 (Random House 2001).
II. HUME ON BRITISH CONSTITUTIONALISM
If the knavery principle is untenable, at least in any simple form,

what can we learn from Hume about constitutional political econ-
omy? A great deal, I argue, even though Hume provides no overarch-
ing master principle of constitutional design. In this Part, I shall iden-
tify what I take to be the characteristic structure of Humean constitu-
tional argument: Given an irreversible departure from, or violation of,
ideal constitutional design, the best response is not to approximate the
ideal as closely as possible by adopting its remaining components.
Rather, the best response is to violate the ideal along some other mar-
gin, in order to produce an offsetting condition or compensating ad-
justment. In Part II.A, I illustrate the argument in the setting of insti-
tutional competition between Commons and monarchy, while in Part
II.B, I do the same in other settings.
A. Commons and Monarchy
As an exegetical matter, the casual attribution of the knavery

principle to Hume is misleading. Although Hume endorses the princi-
ple as a "just political maxim" established by "[p]olitical writers," his
interest in the principle is complex and ambivalent.'" Hume elaborates
it to mean that constitutional arrangements must, "by the skilful divi-
sion of power," ensure that no one institution of government may ac-
crete so much power as to allow it to exercise unchecked sway over
the whole.16 The critical puzzle for Hume is that the British constitu-
tion of his day violated the knavery principle, yet maintained itself as
an ongoing arrangement. Here I shall examine both the violation that
Hume discerned and the compensating condition that stabilized the
mixed constitutional order.
In Hume's view, a major source of the instability of constitutions
generally is an imbalance between formal or legal power and the de
facto power conferred by property. Social actors lacking legal power
who accrue great wealth can eventually leverage their wealth to de-
stabilize or even overturn nominal constitutional arrangements.
Hume's view is nuanced; he concedes that if the original constitution
of the state allots no power to groups that subsequently become dif-
ferentially wealthy, then "[a] Government may endure for several
ages, though the balance of power, and the balance of property do not
coincide,"' 7 because property holders will lack any pretext for seizing
nominal authority. Yet where the original constitution grants "any
15 Hume, Independency of Parliament at 42-43 (cited in note 1).

16 Id.
17 David Hume, Of the First Principles of Government, in Hume, Essays 32, 35 (cited in
note 1).
share of power, though small, to an order of men, who possess a large

share of the property, it is easy for them gradually to stretch their au-
thority, and bring the balance of power to coincide with that of prop-
erty." 18 Paradoxically, constitutions that exclude wealthy groups or "or-
ders" from nominal power altogether will enjoy greater stability than
constitutions that include them to any degree in the governing
institutions.
The most dramatic example of the accretive process by which a
small share of formal power could be expanded, at the expense of
other institutions, was the swelling power of the House of Commons
during the course of British political history. Far wealthier as a corpo-
rate body than any of the English monarchs, and having "assumed to
themselves the sole right of granting money," the Commons had
gradually reduced the monarchy to such a state of fiscal dependence
that the Commons could easily "wrest from the crown all [executive]
powers, one after another; by making every grant conditional, and
choosing their time so well, that their refusal of supply should only
distress the government."19 By virtue of its financial domination of the
monarchy, the Commons had attained so much power that "it abso-
lutely commands all the other parts of the government."20
This state of affairs violated the knavery principle. Because no
"skilful division of power",21 checked the might of the Commons, the
mixed British constitution looked as though its continuance depended
solely on the "good-will of our rulers"22-which according to the knav-
ery principle is no guarantee at all. The division of power needed to
check institutional ambitions had failed entirely; what then kept the
Commons from overmastering the British constitutional order?
Hume's solution to this puzzle appeals to a sophisticated set of com-
pensating adjustments that kept the Commons within the bounds of
the mixed constitution-not by legally constricting the legislators' op-
portunities, as in the formal separation and division of powers, but in-
stead by preventing them from fully exploiting their opportunities.
The key stabilizing condition was the ability of the crown to buy
support from a sufficiently large faction within the Commons, through
the grant of lucrative offices, rotten boroughs, and perhaps outright
bribery, to block harmful legislation:
[T]he interest of the body is here restrained by that of the indi-

viduals.... [T]he house of commons stretches not its power, be-
18 Id.
19 Hume, Independency of Parliament at 44 (cited in note 1).

20 Id.
21 Id at 43.
22 Id at 42.
cause such an usurpation would be contrary to the interest of the

majority of its members. The crown has so many offices at its dis-
posal, that, when assisted by the honest and disinterested part of
the house, it will always command the resolutions of the whole so
far, at least, as to preserve the antient constitution from danger.23
On this picture, we are to imagine a rabidly republican faction of

the House that is animated by a desire to overpower the monarchy,
but is outvoted by a coalition between honest and disinterested legis-
lators who desire to uphold the mixed constitution, on one hand, and
legislators who have sold their votes to the crown, on the other. The
motivational assumption here is that legislators comprise two types,
both public-spirited and rationally self-interested; legislators of the
second type can be turned against the collective interest of the Com-
mons by exploiting their private interests as individuals.
As this last formulation suggests, however, we must understand
Hume as positing a collective action problem among the Commons.
With sufficient coordination, self-interested legislators could combine
with republican ideologues to leverage the Commons' fiscal power,
forcing the monarchy to grant offices, boroughs, and perquisites to the
self-interested legislators who would otherwise sell their votes to the
crown. The collective action problem arises from the asymmetrical
structure of property-holding across institutions. Although the Com-
mons is, taken as a whole, far wealthier than the crown, the crown is
far wealthier than any individual legislator. "[M]uch less property in a
single hand will be able to counterbalance a greater property in sev-
eral; not only because it is difficult to make many persons combine in
the same views and measures; but because property, when united,
causes much greater dependence, than the same property, when dis-
persed."24
We may read this brisk passage as positing two causes of the
Commons' collective action problem. First, and straightforwardly,
Hume is suggesting that the transaction costs of assembling a majority
coalition to coerce rents from the monarchy are prohibitive. Given the
status quo bias of legislative institutions, the monarchy need only buy
enough legislators to block hostile action by republican ideologues.
Second, and more speculatively, we may read Hume as suggesting that
the monarchy's asymmetrical wealth allows it to engage in exploita-
tive vote-buying.25 Each legislator (at least of the venal variety) who is
23 Id at 45.
24 David Hume, Whether the British Government Inclines More to Absolute Monarchy, or to
a Republic, in Hume, Essays 47,47-48 (cited in note 1).
25 For modem accounts of the collective action problems involved in vote-buying, see Saul
Levmore, Voting with Intensity, 53 Stan L Rev 111, 122-41 (2000); Russell Hardin, Morality
within the Limits of Reason 93 (Chicago 1988).
offered a modest office or perquisite in return for his vote supposes

that if he refuses, the monarch will simply buy the vote from some
other legislator. Calculating that it is best to sell his vote, no matter
what others do, each legislator thus sells for a relative pittance, even
though the votes taken as a block are of great value. In this way the
monarch may assemble the necessary defensive coalitions on the
cheap.
Parenthetically, it's not clear that Hume's analysis correctly iden-
tifies a violation of the knavery principle. The key institutional
mechanism that compensates for the Commons' overmastering
power-the monarchy's ability to exploit collective action problems
within the legislature-itself seems to require the presence of self-
interested actors. On this view, Hume errs by equating the knavery
principle with a requirement that a "skilful division of power" obtain
across institutions. The latter requirement is indeed violated by the
unbounded power of the Commons, but it is simply a checks-and-
balances principle that operates at the level of institutions, whereas
the knavery principle operates at the level of individual motivations.
Hume's own argument shows that a mixed constitution may maintain
its stability without formal checks-and-balances, given the right distri-
bution of knavish individuals within institutions.
This criticism may itself be wrong on exegetical grounds. Hume's
account, while assuming the presence of some self-interested actors,
denies that all actors are self-interested; the account assumes, critically,
that some legislators are public-spirited, precisely the motivational
heterogeneity that the knavery principle denies. Even if the criticism
were correct, however, it would merely show that Hume's argument
rests on a false substantive premise. As far as constitutional theory is
concerned, Hume's contribution isn't the substantive analysis, but the
methodological insight that offsetting departures from ideal constitu-
tional design may wash out. On Hume's view, the optimal constitu-
tional design would embody a "skilful division of powers," and would
also remain free of the crown's corrupt influence. If the first goal is
impossible to attain, however, the second goal, even if feasible, may no
longer be desirable. Hume's idea is that attaining the ideal state of
corruption-free government would prove disastrous in conditions
where corruption alone keeps the formally unbounded power of Par-
liament in check. A violation of the second ideal is a necessary coun-
terweight to the irreversible violation of the first.
B. Other Examples
I have claimed, perhaps too boldly, that the conjunction of offset-

ting violations represents the characteristic structure of Humean insti-
tutional argument. So far I have provided only one example, albeit a
very general and important one. I shall make no attempt to substanti-

ate the more general exegetical claim in watertight fashion. Instead I
shall merely gesture at a range of institutional structures that Hume
treats through the lens of offsetting violations or compensating ad-
justments, leaving for the next Part the task of making the analytic
structure of this form of argument more precise. Consider the follow-
ing Humean claims:
1. An executive elected at reasonably frequent intervals is the
best structure (at least in a mixed government containing both repub-
lican and monarchical elements). Suppose, however, that it is a given
that the executive will serve for life. It does not follow that optimal
design will approximate the ideal as closely as possible by making the
life-tenured executive an elected office. The rents that can be obtained
in such an office will be so high that factions and individuals will in-
vest heavily, and unproductively, in obtaining it. (As Hume puts it, "a
crown is too high a reward ever to be given to merit alone, and will
always induce the candidates to employ force, or money, or intrigue, to
procure the votes of the electors."26) Better for the life-tenured execu-
tive to be a hereditary monarch, so that the office is permanently as-
signed to one family, and the timing of the monarch's accession is de-
termined by a random variable (the death of the previous monarch).
2. The republican ideal is that the people should govern them-
selves through direct, collective deliberation rather than through indi-
rect representation. Yet "the people, collected in a body,... [are] quite
unfit for government,"27 because they are subject to sudden and vio-
lent gusts of passionate, nondeliberative opinion. Given the inevitable
violation of the deliberative ideal, intelligent institutional design re-
sponds by an offsetting violation of the ideal of collective self-
government of all by all: Laws should be enacted by a representative
body, with each representative subject to binding instructions from his
local constituents. The latter proviso in effect "disperse[s the people]
in small bodies," where they are "more susceptible both of reason and
order; the force of popular currents and tides is, in a great measure,
broken.""'
3. The mixed British constitution must inevitably decay towards
either a pure republic or an absolute monarchy; which is preferable?
A perfect republic is unattainable, and an imperfect republic will, "af-
ter many convulsions, and civil wars, find repose in absolute monarchy,
which it would have been happier for us to have established peaceably
26 David Hume, That Politics May Be Reduced to a Science, in Hume, Essays 14,18 (cited in
note 1).
27 Hume, First Principles at 36 (cited in note 17).
28 Id.
from the beginning."29 Rather than attempting to approach the ideal

republic as closely as possible, better to minimize the disruptive con-
sequences of inevitable failure by choosing failure at the outset.
We might easily quibble with these analyses, on either substantive
or methodological grounds. The first argument overlooks that would-
be hereditary monarchs may unproductively invest resources in at-
tempted rebellions or usurpations, just as would-be elected monarchs
might invest in corrupt electioneering. The third argument, although
notable for its sophisticated pessimism, confuses ex ante and ex post
perspectives: The exhausted "repose" obtained in absolute monarchy
may only be possible to nations that have already passed through the
interim tumults of attempted republicanism and, thus, may not be ob-
tainable by deliberate choice from the beginning. The structure of the
argument, however, is in each case striking and memorable, in part be-
cause of the implicit appeal to offsetting violations of constitutional
ideals.
III. SECOND-BEST CONSTITUTIONALISM IN AMERICAN LAW
On the account I have suggested, Hume's major methodological

contribution is not the knavery principle, but rather to have antici-
pated30 the modern idea of second-best institutional design. The anal-
ogy here is to a technical idea in economics: If perfect efficiency can-
not be obtained, efficiency is not necessarily maximized by approxi-
mating the first-best conditions as closely as possible; the second-best
outcome might, in principle, be obtained by departing from the first-
best conditions in other respects as well.31 Whatever the technical mer-
its of the analogy, it has by now found wide application in other fields,
including consequentialist political theory. Put negatively, institu-
tional analysis must be alert to the fallacious approximation assump-
tion: the erroneous belief that the best policy is the one that approxi-
mates an unobtainable ideal as closely as possible. In positive terms,
violating our ideals in one respect may require that we violate them in
others as well in order to best approximate the state of affairs that we
would obtain absent any violations. In this Part, I shall indicate some
29 Hume, Whether the British Government Inclines toward Absolute Monarchy at 53 (cited
in note 24).
30 And perhaps inspired or influenced; this is a historical question that I lack competence
to answer.
31 See R.G. Lipsey and R.K. Lancaster, The General Theory of Second Best, 24 Rev Econ
Stud 11 (1956).
32 See, for example, Bruce Talbot Coram, Second Best Theories and the Implications for In-
stitutional Design, in Goodin, Theory of Institutional Design 91 (cited in note 13); Robert E.
Goodin, Political Ideals and Political Practice, 25 Brit J Polit Sci 37, 52-56 (1995).
33 See Avishai Margalit, Ideals and Second Bests, in Seymour Fox, ed, Philosophy for Edu-
cation 77,77 (Daf-Chen 1983).
applications of these ideas to American constitutional law and canvass

some of the resulting methodological problems.
Consider the following areas in which claims of second-best off-
sets or compensating adjustments routinely surface:
1. If constitutional doctrine has permitted excessive delegations
from Congress to the President-excessive on either an originalist34 or
a functional35 account-then a laudable compensating adjustment
would be to allow the legislative veto, congressional restrictions on the
appointments power, and other structural innovations intended to
check presidential power, even if those innovations would otherwise
be unconstitutional.36
2. If the business of the executive has become increasingly politi-
cal, as opposed to technical or administrative, a functionally valuable
adjustment would be to treat the Constitution as establishing a "uni-
tary executive," ensuring that a politically accountable official (the
President) may control discretionary bureaucratic power- even
though the unitary executive lacks an originalist pedigree.37
3. If the President's veto power has been effectively undermined
by the increasing incidence of omnibus legislation, which puts the
President to an illegitimate or all-or-nothing choice, then a laudable
compensating adjustment would be to treat the Constitution as afford-
ing the President a line-item veto, even if that is not the best textual or
original interpretation of the veto power.38
4. If the growth of the national government's jurisdiction has
exacerbated collective action problems within Congress, making pork-
barrel spending uncontrollable, the line-item veto might be a useful
adjustment for this reason as well.39
34 See Gary Lawson, Delegation and Original Meaning, 88 Va L Rev 327, 333-34 (2002)
(asserting that the Constitution "does contain a discernible, textually grounded nondelegation
principle that is far removed from modern doctrine").
35 See David Schoenbrod, Power without Responsibility: How Congress Abuses the People
through Delegation 99-154 (Yale 1993) (arguing that delegation is both undesirable and unnec-
essary). For rejoinders to both the originalist and functionalist critiques of delegation, see Eric A.
Posner and Adrian Vermeule, Interring the Nondelegation Doctrine, 69 U Chi L Rev 1721,1733-
36,1743-54 (2002).
36 See, for example, Martin S. Flaherty, The Most Dangerous Branch, 105 Yale L Rev 1725,
1828-39 (1996); Abner S. Greene, Checks and Balances in an Era of Presidential Lawmaking, 61
U Chi L Rev 123,184-96 (1994).
37 See Lawrence Lessig and Cass R. Sunstein, The President and the Administration, 94
Colum L Rev 1,2-3 (1994).
38 For functional arguments in favor of the line-item veto, see Robert Justin Lipkin, The
New Majoritarianism, 69 U Cin L Rev 107,121-22 (2000); Steven G. Calabresi, Some Normative
Arguments for the Unitary Executive, 48 Ark L Rev 23,78-81 (1995).
39 See Calabresi, 48 Ark L Rev at 78-79 (cited in note 38) ("[T]he line-item veto would, if
everything else stayed the same, clearly lead to lower levels of district-targeted spending and
redistribution.").
5. Another, larger adjustment that might offset the growth of na-

tional power is to expand individual rights through expansive inter-
pretations of the Bill of Rights, including the Ninth Amendment.40
All of these examples have the characteristic structure of the sec-
ond-best argument. In the first example, "[a] first-best world would
have neither delegations nor legislative vetoes, but a world with both
delegations and legislative vetoes is closer to the correct constitutional
'baseline' than is a world with only delegations." 41 The same pattern
applies, with appropriate modifications, for the other examples. All
this reprises the structure of Hume's argument, in which one depar-
ture from the first-best arrangement, the unchecked growth of parlia-
mentary power, is offset by another departure, the use of corrupt in-
fluence by the monarchy.
Second-best constitutionalism can be appended to any approach
to constitutional interpretation; the idea is, in itself, neutral across
first-best theories. Sometimes second-best constitutionalism is associ-
ated with or even equated with interpretation as "translation,",42 in
which the interpreter adapts founding-era commitments to changed
circumstances by recasting or abandoning constitutional texts and the
Framers' specific expectations.43 There is, however, no necessary link
between the two ideas. Translation is merely a version of purposivism.
It assumes an originalist criterion for identifying first-best constitu-
tional arrangements; it differs from standard versions of originalism
only insofar as it boosts the level of interpretive generality, from spe-
cific texts and original intentions to large-scale structures, commit-
ments, and purposes. Second-best constitutionalism is equally com-
patible with first-best interpretive criteria that are not originalist or
purposivist, as when the first-best constitutional arrangement is identi-
fied purely on consequentialist grounds. Consider the argument that
the legislative veto is a valuable adjustment of the constitutional
framework because the best constitutional design, now unattainable,
would for purely functionalist reasons bar sweeping delegations to the
President.
Despite its ubiquity, the appeal to second-best constitutionalism
is problematic on both conceptual and pragmatic grounds. A standard
conceptual objection is that the policy of adjustment is indeterminate,
40 See Randy E. Barnett, Reconceiving the Ninth Amendment, 74 Cornell L Rev 1, 26

(1988).
41 Gary Lawson, The Rise and Rise of theAdministrative State, 107 Harv L Rev 1231,1252-
53 (1994).
42 See Lawrence Lessig, Fidelity in Translation, 71 Tex L Rev 1165, 1189-92 (1993).
43 See Michael J. Klarman, Antifidelity, 70 S Cal L Rev 381, 394-95 (1997) (describing
translation as the idea "that one can avoid the vices of both dead-hand control and uncabined
judicial subjectivity by 'translating' the Framers' concepts into modern circumstances").
as the interpreter may choose the margin on which the adjustment is

made."4 If sweeping delegations produce excessive presidential power,
why adjust by upholding the legislative veto, as opposed to, say, grant-
ing Congress the commander-in-chief power? Here as elsewhere in
constitutional interpretation, however, the indeterminacy point is only
partly persuasive; there are easy cases for second-best constitutional-
ism as well as for first-best interpretive theories. If the growth of om-
nibus legislation has undermined the veto power, we need no elabo-
rate theoretical apparatus to appreciate that permitting the (otherwise
suspect) line-item veto is a more fitting compensating adjustment
than, say, making the veto immune from congressional override.
A related conceptual objection is that second-best constitutional-
ism arbitrarily assumes a wide "transactional frame" that includes
both the rule or condition that departs from optimality and the com-
pensating adjustment-both sweeping delegations and the legislative
veto.45 If, however, the interpreter's responsibility is simply to decide
the narrower question whether the legislative veto is constitutional in
itself, without regard to its relationship to other doctrines, then up-
holding the veto as a compensating adjustment will be beyond the
scope of the task at hand. Note that this objection, unlike the first,
does not hold that the size of the transactional frame is inevitably in-
determinate. The frame may itself be chosen by reference to whatever
first-best interpretive approach the interpreter holds. An originalist
might ask whether the framers or ratifiers thought the relevant sub-
jects were related, so that constitutional design changes along one
margin should be calibrated with choices along the other; a conse-
quentialist might ask whether addressing constitutional questions
through larger or smaller transactional frames produces better prag-
matic results.
There are also important pragmatic objections to second-best
constitutionalism. An ambiguous example is a passage from Washing-
ton's Farewell Address:
If, in the opinion of the people, the distribution or modification of

the constitutional powers be in any particular wrong, let it be cor-
rected by an amendment in the way which the constitution desig-
nates. But let there be no change by usurpation; for though this,
in one instance, may be the instrument of good, it is the custom-
ary weapon by which free governments are destroyed. The
precedent must always greatly overbalance, in permanent evil,
44 For a powerful critique of translation theory on similar grounds, see id at 395-400.

45 See Daryl J. Levinson, Framing Transactions in Constitutional Law, 111 Yale L J 1311,
1313-14 (2002) (discussing generally the concept and significance of transactional frames in con-
stitutional analysis).
any partial or transient benefit to which the use can at any time
yield.4
We might interpret this passage to counsel any of three things, of

which only the third is a pragmatic injunction. First, and most crudely,
the reference to "transient benefit" might amount to a condemnation
of usurpation for nonconstitutional ends. But most usurpations are
carried out in the name of the Constitution itself, rather than in the
name of some other value; compensating adjustments, whether or not
usurpative, are typically justified on the ground that they better ap-
proximate the "true" constitutional order than does seriatim enforce-
ment of particular provisions. Second, the passage might suggest that
even constitutional ends (approximating the first-best constitutional
structure) can only be accomplished through constitutional means
(a textual amendment rather than a compensating adjustment). Third,
we might understand it as a pragmatic, empirical prediction about the
long-term effect of a regime that permits compensating adjustments.
Even if adjustments better approximate the constitutional first-best in
any particular instance, or in the short term, their long-term effect will
be self-defeating, because they undermine public respect for the rule
of law. Unfortunately, this argument rests on an incomplete compari-
son between the regime that permits compensating adjustments and
the regime that bars them. If constitutional amendments rarely occur,
then the latter regime might inflict greater damage on the rule of law,
because it forces citizens to live under an obsolete constitution whose
structure has been rendered incoherent by partial, uncompensated
changes.
The most important pragmatic issue arises when, as is usually the
case, second-best constitutionalism is taken to suggest not only a mode
of constitutional interpretation but also a mode of constitutional ad-
judication. On one view, grounded in comparative institutional compe-
tence, common-law judges ought not take second-best considerations
into account.47 The legislative process is better suited than the litiga-
tion process to forestall unintended consequences and to identify
compensating adjustments on margins remote from the questions at
hand in particular cases, so the common-law judges should focus solely
on the local effects of the rules they adopt. Judges should stick to local
tasks, based on a global assessment that any more ambitious judicial
posture will have damaging global effects. To be sure, there is a super-
ficial air of paradox about this admonition. If the judges can assess the
global effects to which this reasoning appeals, why can't they assess
46 John Marshall, 5 The Life of George Washington 596 (AMS 1969).

47 See Thomas S. Ulen, Courts, Legislatures, and the General Theory of Second Best in Law
and Economics, 73 Chi Kent L Rev 189,217-19 (1998).
the global effects of judicial decisions as well? But there isn't really
any paradox. This view simply holds that judges can and should under-
stand what it is they do not understand, that they should know the
limits of their own knowledge. That second-best adjustments are de-
sirable does not entail that any particular institution or actor should
attempt to identify and implement them.
But the implications of this pragmatic critique for constitutional
adjudication are ambiguous. In the common-law setting the counsel
that judges should ignore second-best considerations is at least clear,
whatever its merit: The judges decide cases on strictly local grounds,
and the legislature may, if it chooses, overrule the judges after consid-
ering second-best effects. But this picture has no unique mapping onto
the constitutional setting, because constitutional decisions cannot be
overruled through ordinary lawmaking. If courts invalidate local struc-
tural innovations that, like the legislative veto, are impermissible
taken in isolation, they thereby ignore any appeals to the offsetting ef-
fects of the innovation along other margins -an approach that ob-
serves the injunction to ignore second-best considerations. Yet invali-
dation prevents the compensating adjustment from being made by any
institution, short of obtaining a constitutional amendment. If the law-
making process really is better suited than the judiciary to identify
necessary second-best adjustments, perhaps judges should simply de-
fer to politically-developed structural innovations in constitutional
cases, rather than invalidating them on the basis of a strictly local
analysis.
In my view, this ambiguity is irresolvable. Both systematic defer-
ence and systematic seriatim enforcement of local constitutional pro-
visions are compatible with the pragmatic critique. But this does not
mean that the critique is so spongy as to be worthless. What it does bar
is the ambitious idea that judges should evaluate global consequences
on a case-by-case basis,48 attempting to identify and uphold all and
only those structural innovations that will indeed compensate for
prior departures from the optimum. Despite its apparent modesty, that
procedure is a counsel of perfection, one that assumes a heroic judicial
capacity to identify the global effects of particular institutional inno-
vations (as opposed to the global effects of a general mode of judicial
decisionmaking). The case-by-case procedure is itself suspect on sec-
ond-best grounds: If the judges' information and computational ca-
pacities are far from the ideal, they ought not simply attempt to imi-
tate ideal judges to the best of their abilities; that is the approximation
'8 See INS v Chadha, 462 US 919,978 (1983) (White dissenting) (urging that the legislative
veto should be upheld to compensate for the demise of the nondelegation doctrine); Peter B.
McCutchen, Mistakes, Precedent, and the Rise of the Administrative State: Toward a Constitutional
Theory of the Second Best, 80 Cornell L Rev 1, 37-38 (1994) (elaborating White's claim).
assumption again. Again, there is no paradox in proposing this sort of

second-order decision, in which judges act in awareness of the limita-
tions of their own information and abilities.49
CONCLUSION
My hope here is to have convincingly recharacterized and vindi-

cated Hume's contributions to constitutional political economy.
Rather than being associated with the rather simplistic knavery prin-
ciple, Hume should be associated with a sophisticated complex of
ideas centering on second-best constitutional design (to describe his
contributions in deliberately anachronistic terms). With the basic idea
established, the overview of conceptual and practical issues in Part III
simply mines a lode of insights that Hume discovered, one that consti-
tutional theory has barely begun to exploit and will never exhaust.
49 See Cass R. Sunstein and Edna Ullmann-Margalit, Second-Order Decisions, 110 Ethics
5,7 (1999).

The University of Chicago Law Review The University of Chicago Law Review

Uploaded by

Copyright:

Available Formats

The University of Chicago Law Review The University of Chicago Law Review

Uploaded by

Document Information

Original Description:

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

The University of Chicago Law Review The University of Chicago Law Review

Uploaded by

Copyright:

Available Formats

Hume's Second-Best Constitutionalism

Author(s): Adrian Vermeule

In the economic and political science literature on constitutional

The idea of second-best constitutionalism has important implica-

I. KNAVES AND KNIGHTS

3 Hume, Independency of Parliament at 42-43 (cited in note 1).

cause public-regarding officials will lose out to officials who strictly

ism. Conversely, on the punishment side, attaching material sanctions

II. HUME ON BRITISH CONSTITUTIONALISM

If the knavery principle is untenable, at least in any simple form,

A. Commons and Monarchy

As an exegetical matter, the casual attribution of the knavery

15 Hume, Independency of Parliament at 42-43 (cited in note 1).

share of power, though small, to an order of men, who possess a large

[T]he interest of the body is here restrained by that of the indi-

19 Hume, Independency of Parliament at 44 (cited in note 1).

cause such an usurpation would be contrary to the interest of the

On this picture, we are to imagine a rabidly republican faction of

offered a modest office or perquisite in return for his vote supposes

I have claimed, perhaps too boldly, that the conjunction of offset-

very general and important one. I shall make no attempt to substanti-

from the beginning."29 Rather than attempting to approach the ideal

III. SECOND-BEST CONSTITUTIONALISM IN AMERICAN LAW

On the account I have suggested, Hume's major methodological

applications of these ideas to American constitutional law and canvass

5. Another, larger adjustment that might offset the growth of na-

40 See Randy E. Barnett, Reconceiving the Ninth Amendment, 74 Cornell L Rev 1, 26

as the interpreter may choose the margin on which the adjustment is

If, in the opinion of the people, the distribution or modification of

44 For a powerful critique of translation theory on similar grounds, see id at 395-400.

We might interpret this passage to counsel any of three things, of

46 John Marshall, 5 The Life of George Washington 596 (AMS 1969).

assumption again. Again, there is no paradox in proposing this sort of

My hope here is to have convincingly recharacterized and vindi-

You might also like