How Judges Think

Richard A. Posner

HARVARD UNIVERSITY PRESS

Cambridge, Massachusetts
London, England
2008
Copyright © 2008 by the President and Fellows of Harvard College
All rights reserved
Printed in the United States of America

A Caravan book. For more information, visit www.caravanbooks.org

Library of Congress Cataloging-in-Publication Data

Posner, Richard A.
How judges think / Richard A. Posner.
p. cm.
Includes index.
ISBN-13: 978-0-674-02820-3 (alk. paper)
1. Judicial process. 2. Judicial process—United States. I. Title.

K2300.P67 2008
347′.012—dc22 2007037826
Contents

Introduction 1

PART ONE THE BASIC MODEL

1 Nine Theories of Judicial Behavior 19
2 The Judge as Labor-Market Participant 57
3 The Judge as Occasional Legislator 78
4 The Mind of the Legislating Judge 93

PART TWO THE MODEL ELABORATED

5 The Judicial Environment:
External Constraints on Judging 125
6 Altering the Environment: Tenure and Salary Issues 158
7 Judicial Method: Internal Constraints on Judging 174
8 Judges Are Not Law Professors 204
9 Is Pragmatic Adjudication Inescapable? 230

PART THREE JUSTICES

10 The Supreme Court Is a Political Court 269
11 Comprehensive Constitutional Theories 324
12 Judicial Cosmopolitanism 347

Conclusion 369

Acknowledgments 379
Index 381
 Introduction

In my youthful, scornful way, I recognized four kinds of judgments;

first the cogitative, of and by reflection and logomancy; second,
aleatory, of and by the dice; third, intuitive, of and by feeling or
“hunching”; and fourth, asinine, of and by an ass; and in that same
youthful, scornful way I regarded the last three as only variants of each
other, the results of processes all alien to good judges.1

Ivan Karamazov said that if God does not exist everything is permitted,
and traditional legal thinkers are likely to say that if legalism (legal for-
malism, orthodox legal reasoning, a “government of laws not men,” the
“rule of law” as celebrated in the loftiest Law Day rhetoric, and so forth)
does not exist everything is permitted to judges—so watch out! Legal-
ism does exist, and so not everything is permitted. But its kingdom has
shrunk and grayed to the point where today it is largely limited to rou-
tine cases, and so a great deal is permitted to judges. Just how much is
permitted and how they use their freedom are the principal concerns of
this book. These concerns have been made especially timely by the star-
tling (to the naïve) right turn by the Supreme Court in its latest term
(ending in June 2007).2 The turn resulted from the replacement of a
moderately conservative Justice (O’Connor) by an extremely conserva-
tive one (Alito), and so underscores the question of the personal and po-
litical elements in judging and thus of the sense in which the nation is
ruled by judges rather than by law. If changing judges changes law, it is
not even clear what law is.
1. Joseph C. Hutcheson, Jr., “The Judgment Intuitive: The Function of the ‘Hunch’ in Judi-
cial Decision,” 14 Cornell Law Quarterly 274, 275–276 (1929).
2. Linda Greenhouse, “In Steps Big and Small, Supreme Court Moved Right: A 5–4 Dy-
namic, with Kennedy as Linchpin,” New York Times, July 1, 2007, § 1, p. 1.

1
2 Introduction

I feel a certain awkwardness in talking about judges, especially appel-

late judges (my main concern), because I am one. Biographies are more
reliable than autobiographies, and cats are not consulted on the princi-
ples of feline psychology. At the same time, I am struck by how unreal-
istic are the conceptions of the judge held by most people, including
practicing lawyers and eminent law professors, who have never been
judges3—and even by some judges. This unrealism is due to a variety of
things, including the different perspectives of the different branches of
the legal profession—including also a certain want of imagination. It is
also due to the fact that most judges are cagey, even coy, in discussing
what they do. They tend to parrot an official line about the judicial pro-
cess (how rule-bound it is), and often to believe it, though it does not
describe their actual practices.4 There is also the sense that judging re-
ally is a different profession from practicing or teaching law, and if you’re
not in it you can’t understand it. I remember when I was appointed re-
ceiving a note from a court of appeals judge in another circuit with
whom I was acquainted, welcoming me to “the club.” This book parts
the curtains a bit.
The difficulty outsiders have in understanding judicial behavior is
due partly to the fact that judges deliberate in secret, though it would be
more accurate to say that the fact that they do not deliberate (by which I
mean deliberate collectively) very much is the real secret.5 Judicial delib-
eration is overrated. English judges traditionally did not deliberate at
all, as that would have violated the ruling principle of “orality,” whereby
everything that judges did had to be done in public so that their behav-

3. A notable example is Harvard law professor Henry Hart’s time-and-motion study of Su-
preme Court Justices: Henry M. Hart, Jr., “The Supreme Court, 1958 Term: Foreword: The
Time Chart of the Justices,” 73 Harvard Law Review 84 (1959), which I discuss in chapter 10.
4. Robert Keeton, a federal district judge and before that a Harvard Law School professor,
acknowledged in his treatise on judging that judges make “value-laden” rulings. Robert E.
Keeton, Keeton on Judging in the American Legal System 15 (1999). But he did not explore the
sources of those values. His treatise has no index entry for either “politics” or “ideology.”
5. Though a pretty open one. “When I first came on the court [the U.S. Court of Appeals for
the District of Columbia Circuit], I imagined that conferences [on cases] would be reflective,
refining, analytical, dynamic. Ordinarily they are none of these. We go around the table and
each judge, from junior to senior, states his or her bottom line and maybe a brief explanation.
Even if the panel is divided, the discussion is exceedingly crisp. The conference changes few
minds. Assignments are made, life goes on.” Patricia M. Wald, “Some Real-Life Observations
about Judging,” 26 Indiana Law Review 173, 177 (1992). Chief Justice Rehnquist described Su-
preme Court conferences similarly. See chapter 10.
 Introduction 3

ior could be monitored;6 hence those seriatim opinions that baffle the
American law student and perhaps the English one as well. In almost all
cases a brief discussion among the judges before deciding enables con-
vergence on a single majority opinion in lieu of a separate opinion by
each judge.
The confidentiality of the judicial process would not matter greatly to
an understanding and evaluation of the legal system if the consequences
of judicial behavior could be readily determined. If you can determine
the ripeness of a cantaloupe by squeezing or smelling it, you don’t have
to worry about the produce clerk’s mental processes. But the conse-
quences of judicial behavior are often more difficult to determine and
evaluate than the consequences even of other professional services, such
as medicine. Many of the decisions that constitute the output of a court
system cannot be shown to be either “good” or “bad,” whether in terms
of consequences or of other criteria, so it is natural to ask whether there
are grounds for confidence in the design of the institution and in the
competence and integrity of the judges who operate it.
The secrecy of judicial deliberations is an example of professional
mystification. Professions such as law and medicine provide essential
services that are difficult for outsiders to understand and evaluate. Pro-
fessionals like it that way because it helps them maintain a privileged
status. But they know they have to overcome the laity’s mistrust, and
they do this in part by developing a mystique that exaggerates not only
the professional’s skills but also his disinterest.7 Judges have been doing
this for thousands of years and have become quite good at it—so good as
to have achieved a certain opacity even to their fellow legal profession-
als, including law professors as well as practicing lawyers. Judges have
convinced many people—including themselves—that they use esoteric
materials and techniques to build selflessly an edifice of doctrines un-
marred by willfulness, politics, or ignorance.
There is nevertheless considerable dissatisfaction with our legal sys-
tem,8 as there is with our system of health care. Like health care, law is

6. Robert J. Martineau, Appellate Justice in England and the United States: A Comparative
Analysis 101–103 (1990).
7. Richard A. Posner, The Problematics of Moral and Legal Theory, ch. 3 (1999).
8. See, for example, Philip K. Howard, The Collapse of the Common Good: How America’s
Lawsuit Culture Undermines Our Freedom (2001); Walter K. Olson, The Litigation Explosion:
What Happened When America Unleashed the Lawsuit (1991).
4 Introduction

said to be too expensive (it certainly costs more per capita than the le-
gal systems of the nations with which we tend to compare the United
States), too intrusive into private and commercial life, too prone to error,
too uncertain, and simply too large (the nation has a million lawyers).
For these reasons it is contended to be a source of immense indirect
costs on top of the expenses to the litigants. The accusations may be
true, though assessing their truth is not the project of this book and is
especially daunting because it is even harder to estimate the benefits of
our legal system than its costs. Legal rights are options that may have
value even if never exercised, but how to value such options? And legal
duties deter harmful conduct—but how effectively is extremely difficult
to determine too.
Supposing the criticisms have merit, the question is whom to blame.
If all that judges do is apply rules made by legislatures or the framers
of the Constitution (or follow precedents, made by current or former
judges, that are promptly changed if they prove maladapted to current
conditions), then the responsibility for the mess (if it is a mess) must lie
with the legislators or the Constitution’s framers, or with the political
process more generally. But suppose that most rules laid down by legis-
lative bodies are all right and the problem is willful judges—judges who
make up their own rules, or perhaps ignore rules altogether, instead dis-
pensing shortsighted justice on the basis of the “equities” of each case,
and as a result create enormous legal uncertainty. The policy implica-
tions and hence the path of reform would depend on which explanation
was correct (both might be). And what if the basic problem is that the
structure of American government, and the American political culture
more broadly, compel judges to make rather than just apply rules of law?
What looks to the critics of the judiciary like willfulness might actually
be the good-faith performance of a vital judicial role, and if judges re-
fused to play it, insisting instead, as some legal thinkers urge (the “le-
galists,” of whom more shortly), on limiting themselves to passively ap-
plying rules made elsewhere, the legal system might be worse than it is.
The answers are bound up with issues of judicial behavior. To il-
lustrate, everyone will agree that contracts are vital to the operation of
markets, and almost everyone will agree that the legal enforcement of
contracts is important to the efficacy of contracts. Contract law is ad-
ministered by judges. (Sometimes they are private judges—arbitrators—
 Introduction 5

but the effectiveness of arbitration depends on the enforceability of arbi-

trators’ awards.) Being a part of the common law, it is also created by
them. The law they create and the way in which they enforce it are delib-
erate acts, just as business decisions and decisions by legislatures are de-
liberate acts. Whether judicially made doctrines and decisions are good
or bad may depend therefore on the judges’ incentives, which may in
turn depend on the judges’ cognition and psychology, on how persons
are selected (including self-selected) to be judges, and on the terms and
conditions of judicial employment. Similarly, American antitrust law is
far more the creation of judicial decisions than of antitrust legislation:
the most important antitrust laws are as skimpy and vague as most pro-
visions of the Bill of Rights. We ought therefore to be interested in how
antitrust law has been shaped by the motivations, constraints, and other
influences that play on judges. The Supreme Court has actually called
the Sherman Act “a common-law statute,”9 and common law is of course
made by judges, not legislators.
The judicial mentality would be of little interest if judges did noth-
ing more than apply clear rules of law created by legislators, adminis-
trative agencies, the framers of constitutions, and other extrajudicial
sources (including commercial custom) to facts that judges and juries
determined without bias or preconceptions. Then judges would be well
on the road to being superseded by digitized artificial intelligence pro-
grams.10 But even legal thinkers who believe passionately that judges
should be rule appliers and unbiased fact finders and nothing more do
not believe that that’s how all or even most American judges behave all
the time. Our judges have and exercise discretion. Especially if they are
appellate judges, even intermediate ones, they are “occasional legisla-
tors.” To understand their legislative activity, one must understand their
motivations, capacities, mode of selection, professional norms, and psy-
chology.
Achieving a sound understanding of judicial behavior is thus of more
than merely academic interest; it is a key to legal reform. Yet its academic
interest is also considerable because of the unusual incentives and con-
straints, so unlike those in most jobs, that shape judicial behavior, espe-

9. Leegin Creative Leather Products, Inc. v. PSKS, Inc., 127 S. Ct. 2705, 2720 (2007).
10. I do not know why originalists and other legalists are not AI enthusiasts.
6 Introduction

cially in the U.S. federal system, and because the analysis of that behav-
ior may offer insights into the general subject of managing uncertainty.11
Uncertainty is as salient a feature of our legal system as expense is of our
medical system, and decision making under uncertainty is a deservedly
important topic in economics, organization theory, and other fields.
Like other writing by judges about judging, this book is heavily influ-
enced by my own judicial experience, consisting of more than a quarter
century as a federal court of appeals judge (seven years as chief judge of
my court), with occasional forays into the district court to preside at tri-
als, mainly civil jury trials. But the mode of the book is scholarly rather
than confessional. In this respect it resembles my book on the regulation
of sexuality, a subject otherwise remote from the study of judicial behav-
ior. That book was motivated by my “belated discovery that judges
know next to nothing about sex beyond their own personal experience,
which is limited,” and one of my aims was to “bring to the attention of
the legal profession the rich multidisciplinary literature” on the sub-
ject.12 Judges, like other “refined” people in our society, are reticent
about talking about sex, but judges are also reticent about talking about
judging, especially talking frankly about it, whether to their colleagues
or to a larger professional audience. This reticence makes the scholarly
study of judicial behavior at once challenging and indispensable.
The book emphasizes positive rather than normative analysis—what
judges do, not what they should do—but I do discuss normative issues
and propose a few modest reforms, as well as making occasional sugges-
tions for further research. Positive and normative analysis cannot easily
be separated when one is dealing with people’s deliberate actions, for
unless they are evil or cynical people, the best explanation for their ac-
tions is unlikely to be that they are deliberately flouting the norms of
their society. If it is deeply wrong for a judge to base a decision on the
flip of a coin, an aleatory theory of judicial behavior is unlikely to be
sound. The grounds of a judge’s decisions may be wrong, but they are

11. About which I have written at length in relation to catastrophic risk and also to the re-
form of the U.S. intelligence system. See my books Catastrophe: Risk and Response (2004); Pre-
venting Surprise Attacks: Intelligence Reform in the Wake of 9/11 (2005); Uncertain Shield: The
U.S. Intelligence System in the Throes of Reform (2006); Countering Terrorism: Blurred Focus,
Halting Steps (2007).
12. Richard A. Posner, Sex and Reason 1, 4 (1992).
 Introduction 7

unlikely to be outside the ballpark of norms and values prevailing in the

society.
The book’s primary focus is on federal appellate judges, including Su-
preme Court Justices (the subject of Part Three, though discussed in the
other parts as well). But there is some discussion of trial judges, state
court judges, judges in foreign nations similar to the United States,13 and
arbitrators (private judges).
I begin with a discussion of the existing theories (attitudinal, strate-
gic, organizational, economic, psychological, sociological, pragmatic, phe-
nomenological, and legalist) of judicial behavior and of the evidence for
and against each. These theories are expounded in a rich literature ig-
nored by most academic lawyers (though this is changing14) and by vir-
tually all judges.15 The theories provide background and support to my
own analysis, which draws heavily on labor economics and on the psy-
chology of cognition and emotion. It is the stress I lay on psychology
that has led me to entitle the book How Judges Think rather than Judicial
Behavior.
My analysis and the studies on which it builds find that judges are not
moral or intellectual giants (alas), prophets, oracles, mouthpieces, or
calculating machines. They are all-too-human workers, responding as
other workers do to the conditions of the labor market in which they
work. American judges, at least, are not formalists, or (the term I prefer,
as it carries less baggage) legalists. Legalists decide cases by applying
preexisting rules or, in some versions of legalism, by employing alleg-
edly distinctive modes of legal reasoning, such as “legal reasoning by
analogy.” They do not legislate, do not exercise discretion other than in
ministerial matters (such as scheduling), have no truck with policy, and
13. That is, nations that have an independent judiciary, as many do not. See, for exam-
ple, Gretchen Helmke, Courts under Constraints: Judges, Generals, and Presidents in Argentina
(2005); Law and Economic Development (Hans-Bernd Schäfer and Angara V. Raja eds. 2006).
14. See, for example, Gregory C. Sisk and Michael Heise, “Judges and Ideology: Public and
Academic Debates about Statistical Measures,” 99 Northwestern University Law Review 743
(2005).
15. The richness is well illustrated by James L. Gibson, “From Simplicity to Complexity:
The Development of Theory in the Study of Judicial Behavior,” 5 Political Behavior 7 (1983). As
the date of his article shows, the literature has been around for many years. That it has not
caught on with the legal professoriat may be in part because of its death of implications for the
understanding or reform of legal doctrine and in part because it challenges the mystique of an
apolitical judiciary, in which lawyers and law professors are heavily invested.
8 Introduction

do not look outside conventional legal texts—mainly statutes, constitu-

tional provisions, and precedents (authoritative judicial decisions)—for
guidance in deciding new cases. For legalists, the law is an autonomous
domain of knowledge and technique.16 Some legalists are even suspi-
cious of precedent as a source of law, because it is infected by judicial
creativity.
But if judges are not legalists, what are they? Might they simply be
politicians in robes? Empirical scholars have found that many judicial
decisions, by no means limited to the Supreme Court, are strongly influ-
enced by a judge’s political preferences or by other extralegal factors,
such as the judge’s personal characteristics and personal and profes-
sional experiences, which may shape his political preferences or operate
directly on his response to a case. No responsible student of the judicial
system supposes that “politics” (in a sense to be explained) or personal
idiosyncrasy drives most decisions, except in the Supreme Court, which
indeed is largely a political court when it is deciding constitutional
cases. Legalism drives most judicial decisions, though generally they are
the less important ones for the development of legal doctrine or the im-
pact on society.
But one must be careful about dividing judicial decisions (or judges)
into legalist and political, or, what is closely related, asserting a Mani-
chaean dualism between law and politics. The dualism works only when
“law” is equated to legalism, and that is too narrow. Justice Scalia was
not stepping out of his proper role as a judge when he said in Richardson
v. Marsh that “the rule that juries are presumed to follow their instruc-
tions is a pragmatic one, rooted less in the absolute certitude that the
presumption is true than in the belief that it represents a reasonable
practical accommodation of the interests of the state and the defendant
in the criminal justice process.”17 This is just as proper a judicial state-

16. “Legal formalists emphasize the specifically legal virtues of the clarity, determinacy, and
coherence of law, and try to sharpen the distinction between legislation and adjudication.
Roughly, they can be divided into rule-formalists and concept-formalists. The former place
more value on determinacy, emphasizing the importance of clear rules and strict interpretation,
while the latter emphasize the importance of system and principled coherence throughout the
law.” Thomas C. Grey, “Judicial Review and Legal Pragmatism,” 38 Wake Forest Law Review
473, 478 (2003). Modern American formalists—comprising what one might call the School of
Scalia—are mainly rule-formalists. Id. at 479. “The most important thing [for Scalia] is that law
should be put in the form of rules wherever possible.” Id. at 499.
17. 481 U.S. 200, 211 (1987).
 Introduction 9

ment as the legalist assertions for which Scalia (he of such pronounce-
ments as that the “rule of law” is the “law of rules”18) is more famous.
This is so even though the statement has political implications. Criminal
defendants are at a disadvantage if a judge’s or prosecutor’s missteps can
be forgiven by the judge’s telling the jury to disregard them, for the bell
cannot be unrung; the jurors cannot exclude what they should not have
heard from their consideration of the defendant’s guilt.
“Law” in a judicial setting is simply the material, in the broadest
sense, out of which judges fashion their decisions. Because the materials
of legalist decision making fail to generate acceptable answers to all
the legal questions that American judges are required to decide, judges
perforce have occasional—indeed rather frequent—recourse to other
sources of judgment, including their own political opinions or policy
judgments, even their idiosyncrasies. As a result, law is shot through
with politics and with much else besides that does not fit a legalist
model of decision making.
The decision-making freedom that judges have is an involuntary free-
dom. It is the consequence of legalism’s inability in many cases to decide
the outcome (or decide it tolerably, a distinction I shall elaborate), and
the related difficulty, often impossibility, of verifying the correctness of
the outcome, whether by its consequences or its logic. That inability,
and that difficulty or impossibility, create an open area in which judges
have decisional discretion—a blank slate on which to inscribe their de-
cisions—rather than being compelled to a particular decision by “the
law.” How they fill in the open area is the fundamental question that this
book addresses, though lurking in the background and occasionally
coming to the fore is the question how they should fill it in.
Although judges often exercise a political judgment in the open area,
“political” is an equivocal term that must be carefully parsed before it
can be usefully applied to judicial behavior. It could refer to a judge
whose decisions reflect his loyalty to a political party. It could refer to a
judge whose decisions faithfully mirror the platform of a political party,
though as a matter of conviction rather than of party loyalty. It could re-
fer to a judge whose decisions reflect a consistent political ideology,
which might be “liberal” or “conservative” and thus correlated (though

18. Antonin Scalia, “The Rule of Law as a Law of Rules,” 56 University of Chicago Law Re-
view 1175 (1989).
10 Introduction

imperfectly) with the Democratic or Republican Party platform, but

which might instead be an ideology embraced by neither major party,
such as libertarianism or socialism. The empirical literature that refutes
legalism as a complete or even approximate description of actual judicial
behavior does not distinguish among these different gradations of “po-
litical.” “Political” could even describe decisions based on purely techni-
cal policy judgments, judgments that involve finding the best means to
agreed-upon ends; any issue of governmental policy is in that sense “po-
litical.” At the opposite extreme, a judge might be “political” in a sense
divorced from policy: he might, like a legislator, use charm, guile, vote
trading, and flattery to induce other judges to go along with him, though
his aim might be to produce legalistic decisions. (He might thus be what
is called in a variety of nonpolitical settings “a good politician.”) The
strategic theory of judicial behavior, discussed in chapter 1, emphasizes
political judging in this “means” rather than “ends” sense. Many legisla-
tors have no policy preferences of their own, but are merely political
brokers for their constituents. Judges, however, unless elected, do not
have constituents.
Ringing changes on the “political” might seem to exhaust the possible
nonlegalist factors in adjudication. It does not begin to. The possible
other factors (call them “personal”) include personality traits, or tem-
perament (and thus emotionality at one end of the temperament spec-
trum and emotional detachment at the other end), which are more or
less innate personal characteristics. They include personal background
characteristics, such as race and sex, and also personal and professional
experience. The political or ideological factors that influence adjudi-
cation may themselves be by-products of personal factors rather than
products of an informed, disinterested, and coolly analytical study of
public issues. Also figuring in judicial decisions are strategic consider-
ations, already alluded to, which need not be related to either the politi-
cal views or the personal characteristics of a judge. A judge might join
the majority opinion in a case not because he agreed with it but because
he thought that dissenting publicly would magnify the effect of the ma-
jority opinion by drawing attention to it. (“Dissent aversion” helps to
explain, as we shall see in chapter 1, the puzzling effect of panel compo-
sition on appellate decisions.) Institutional factors—such as how clear or
unclear the law is, salary and workload, and the structure of judicial pro-
motion—also influence judicial behavior.
1

Nine Theories of Judicial Behavior

There are many positive (that is, descriptive as distinct from normative)
theories of judicial behavior.1 Their primary focus is, as one would ex-
pect, on explaining judges’ decisions. The theories are the attitudinal,
the strategic, the sociological, the psychological, the economic, the orga-
nizational, the pragmatic, the phenomenological, and, of course, what I
am calling the legalist theory. All the theories have merit and feed into
the theory of decision making that I develop in this book. But all are
overstated or incomplete. And missing from the welter of theories—the
gap this book endeavors to fill, though in part simply by restating and
refining the existing theories—is a cogent, unified, realistic, and appro-
priately eclectic account of how judges actually arrive at their decisions
in nonroutine cases: in short, a positive decision theory of judging.
I begin with the attitudinal theory,2 which claims that judges’ deci-

1. For reviews of the literature, see Lawrence Baum, Judges and Their Audiences: A Perspec-
tive on Judicial Behavior, ch. 1 (2006); Barry Friedman, “The Politics of Judicial Review,” 84
Texas Law Review 257 (2005). For an anthology suggestive of the diversity of the literature,
see Supreme Court Decision-Making: New Institutionalist Approaches (Cornell W. Clayton and
Howard Gillman eds. 1999).
2. See, for example, Jeffrey A. Segal and Harold J. Spaeth, The Supreme Court and the Attitu-
dinal Model Revisited (2002); Robert A. Carp and Ronald Stidham, Judicial Process in America
294 (2001) (tab. 10–1); William N. Eskridge, Jr., and Lauren E. Baer, “The Supreme Court’s
Deference Continuum: An Empirical Analysis (from Chevron to Hamdan)” (Yale Law School,
May 11, 2007); Andrew D. Martin, Kevin M. Quinn, and Lee Epstein, “The Median Justice on
the United States Supreme Court,” 83 North Carolina Law Review 1275 (2005); Micheal W.
Giles, Virginia A. Hettinger, and Todd Peppers, “Picking Federal Judges: A Note on Policy and
Partisan Selection Agendas,” 54 Political Research Quarterly 623 (2001); Tracey E. George, “De-
veloping a Positive Theory of Decision Making on U.S. Courts of Appeals,” 58 Ohio State Law

19
20 The Basic Model

sions are best explained by the political preferences that they bring to
their cases. Most of the studies that try to test the theory infer judges’
political preferences from the political party of the President who ap-
pointed them, while recognizing that it is a crude proxy. The emphasis
is on federal judges, in particular Supreme Court Justices. State judges
are of course not appointed by the President, and sometimes the method
of their appointment—for example, by nonpartisan election—makes it
difficult to classify them politically.3
Justices and judges appointed by Democratic Presidents are predicted
to vote disproportionately for “liberal” outcomes, such as outcomes fa-
voring employees, consumers, small businessmen, criminal defendants
(other than white-collar defendants), labor unions, and environmental,
tort, civil rights, and civil liberties plaintiffs. Judges and Justices ap-
pointed by Republican Presidents are predicted to vote disproportion-
ately for the opposite outcomes.
Other evidence of a judge’s political leanings is sometimes used in lieu
of the party of the appointing President, such as preconfirmation editori-
als discussing the politics or ideology of a judicial nominee.4 A neglected
possibility is a fourfold classification in which the intermediate cate-
gories would consist of judges appointed when the President and the
Senate majority were of different parties (“divided government”). How-
ever, Nancy Scherer finds no difference in the decisions of federal dis-
trict judges appointed by “divided” versus “united” government,5 and I
find only a small difference (as shown in Table 16) in the case of federal
court of appeals judges appointed by Republican Presidents. But when

Journal 1635, 1678 (1998). For criticism, see Frank B. Cross, “Political Science and the New
Legal Realism: A Case of Unfortunate Interdisciplinary Ignorance,” 92 Northwestern Law Re-
view 251 (1997); Barry Friedman, “Taking Law Seriously,” 4 Perspectives on Politics 261 (2006).
3. Paul Brace, Laura Langer, and Melinda Gann Hall, “Measuring the Preferences of State
Supreme Court Judges,” 62 Journal of Politics 387 (2000); Carp and Stidham, note 2 above, at
296–297.
4. Jeffrey A. Segal and Albert D. Cover, “Ideological Values and the Votes of U.S. Supreme
Court Justices,” 83 American Political Science Review 557 (1989); Segal et al., “Ideological
Values and the Votes of U.S. Supreme Court Justices Revisited,” 57 Journal of Politics 812
(1995). See also Martin, Quinn, and Epstein, note 2 above, at 1285–1300.
5. Nancy Scherer, “Who Drives the Ideological Makeup of the Lower Federal Courts in a
Divided Government?” 35 Law and Society Review 191 (2001).
6. Some of the classifications used in the data set from which the statistics in Tables 1 and 2
are drawn are erroneous, such as classifying all votes for plaintiffs in intellectual property cases
as “liberal.” I have corrected such errors; for the details of the corrections and a fuller analysis
of the data, see William M. Landes and Richard A. Posner, “Judicial Behavior: A Statistical Anal-
ysis” (University of Chicago Law School, Oct. 2007).
 Nine Theories of Judicial Behavior 21

Table 1 Judicial Votes in Courts of Appeals as Function of United versus Divided

Presidency and Senate, 1925–2002 (in percent)

Republican President Democratic President

Republican Democratic Democratic Republican

Vote Senate Senate Senate Senate

Conservative 55.8 55.9 49.6 55.3

Liberal 37.1 35.9 43.5 37.9
Mixed 7.1 8.2 6.8 6.8
Sources: Appeals Court Attribute Data, www.as.uky.edu/polisci/ulmerproject/auburndata.htm
(visited July 17, 2007); U.S. Court of Appeals Database, www.as.uky.edu/polisci/ulmerproject/
appctdata.htm, www.wmich.edu/⬃nsf-coa/ (visited July 17, 2007). Votes were weighted to
reflect the different caseloads in the different circuits. “Mixed” refers to multi-issue cases in
which the judge voted the liberal side of one or more issues and the conservative side of the
other issue or issues.

Table 2 Judicial Votes in Courts of Appeals as Function of United versus Divided

Presidency and Senate, Judges Serving Currently (in percent)

Republican President Democratic President

Republican Democratic Democratic Republican

Vote Senate Senate Senate Senate

Conservative 66.9 63.2 49.7 57.0

Liberal 25.6 27.0 39.5 35.6
Mixed 7.5 9.8 10.9 7.5
Sources: Appeals Court Attribute Data, www.as.uky.edu/polisci/ulmerproject/auburndata.htm
(visited July 17, 2007); U.S. Court of Appeals Database, www.as.uky.edu/polisci/ulmerproject/
appctdata.htm, www.wmich.edu/⬃nsf-coa/ (visited July 17, 2007). Votes were weighted to
reflect the different caseloads in the different circuits. “Mixed” refers to multi-issue cases in
which the judge voted the liberal side of one or more issues and the conservative side of the
other issue or issues.

the President is a Democrat, it makes a significant difference whether the

Senate is Democratic or Republican, probably because the Republican
Party is more disciplined than the Democratic Party and therefore better
able to organize opposition to a nominee.
Table 2 is similar to Table 1 except limited to currently serving judges.
Notice that the effects of divided government on judicial voting are more
pronounced than in Table 1, consistent with the strong Republican push
beginning with Reagan to tilt the ideological balance of the courts right-
ward. Notice also that federal judicial decisions as a whole tilt toward
22 The Basic Model

Table 3 Ideology of Currently Serving Justices and the Appointing President

President’s Ideology

Conservative Moderate
Justice’s Ideology Republican Republican Democratic

Conservative 4 1 0
Liberal 0 2 2

Table 4 Conservative and Liberal Supreme Court Justices as Function of United

versus Divided Presidency and Senate, Justices Serving Currently

Republican President Democratic President

Republican Democratic Democratic Republican

Justice Senate Senate Senate Senate

Conservative 3 2 0 0
Liberal 0 2 2 0

the conservative end of the spectrum and that the tilt is more pro-
nounced among currently serving judges.
Presidents differ in their ideological intensity, and taking account of
that difference can improve the accuracy of the attitudinal model. Seven
of the nine current Supreme Court Justices were appointed by Republi-
can Presidents, but it is more illuminating to note that four conservative
Justices were appointed by conservative Republicans (Scalia and Ken-
nedy by Reagan, and Roberts and Alito by the second Bush), two liberal
Justices by a Democratic President (Ginsburg and Breyer, appointed by
Clinton), and one liberal and two conservative Justices appointed by
moderate Republicans (Stevens by Ford, Souter and Thomas by the first
Bush). See Table 3.
There is also a divided-government effect in Supreme Court appoint-
ments, as shown in Table 4.
Whatever the method of determining a judge’s political inclinations,
and whatever the level of the judiciary (Supreme Court, federal courts of
appeals—on which there is now an extensive literature7—or federal dis-

7. Christina L. Boyd, Lee Epstein, and Andrew D. Martin, “Untangling the Causal Effects of
Sex on Judging” (Northwestern University School of Law and Washington University School of
Law and Department of Political Science, July 28, 2007); Cass R. Sunstein et al., Are Judges Po-
 Nine Theories of Judicial Behavior 23

trict courts8), the assumed inclinations are invariably found to explain

much of the variance in judges’ votes on politically charged issues. The
hotter the issue (such as abortion, which nowadays is much hotter than,
say, criminal sentencing), the greater the explanatory power of the polit-
ical variable. The attitudinal theory is further supported by the unques-
tionable importance of politics in the appointment and confirmation of
federal judges;9 by the intensity of congressional battles, almost always
politically polarized, over the confirmation of federal judges and partic-
ularly Supreme Court Justices; and by the experiences of lawyers and
judges. Every lawyer knows that the accident of which judges of a court
of appeals are randomly drawn to constitute the panel that will hear his
case may determine the outcome if the case is controversial. Every judge
is aware of having liberal and conservative colleagues whose reactions to
politically charged cases can be predicted with a fair degree of accuracy
even if the judge who affixes these labels to his colleagues would not like
to be labeled politically himself.
Further evidence is the tendency of both Supreme Court Justices and
court of appeals judges to time their retirement in such a way as to maxi-
mize the likelihood that a successor will be appointed by a President of
litical? An Empirical Analysis of the Federal Judiciary (2006); Thomas J. Miles and Cass R.
Sunstein, “Do Judges Make Regulatory Policy? An Empirical Investigation of Chevron,” 73 Uni-
versity of Chicago Law Review 823 (2006); Ward Farnsworth, “The Role of Law in Close Cases:
Some Evidence from the Federal Courts of Appeals,” 86 Boston University Law Review 1083
(2006); Jeffrey A. Segal, Harold J. Spaeth, and Sara C. Benesh, The Supreme Court in the Ameri-
can Legal System 236–242 (2005); Daniel R. Pinello, Gay Rights and American Law (2003);
Frank B. Cross, “Decision Making in the U.S. Circuit Courts of Appeals,” 91 California Law Re-
view 1457, 1504–1509 (2003); David E. Klein, Making Law in the United States Court of Appeals
(2002); Emerson H. Tiller and Frank B. Cross, “A Modest Proposal for Improving American
Justice,” 99 Columbia Law Review 215, 218–226 (1999); George, note 2 above; Richard L.
Revesz, “Environmental Regulation, Ideology, and the D.C. Circuit,” 83 Virginia Law Review
1717 (1997); Sheldon Goldman, “Voting Behavior on the United States Courts of Appeals Re-
visited,” 69 American Political Science Review 491 (1975). For an interesting case study of how
the political preferences of court of appeals judges affect decisions, see Paul J. Wahlbeck, “The
Development of a Legal Rule: The Federal Common Law of Public Nuisance,” 32 Law and Soci-
ety Review 613 (1998).
8. C. K. Rowland and Robert A. Carp, Politics and Judgment in Federal District Courts
(1996); Gregory C. Sisk, Michael Heise, and Andrew P. Morriss, “Charting the Influences on
the Judicial Mind: An Empirical Study of Judicial Reasoning,” 73 New York University Law Re-
view 1377 (1998); Ahmed E. Taha, “Judges’ Political Orientations and the Selection of Disputes
for Litigation” (Wake Forest University School of Law, Jan. 2007), http://ssrn.com/abstract=
963468 (visited Sept. 2, 2007).
9. On which see, for example, Lee Epstein and Jeffrey A. Segal, Advice and Consent: The Poli-
tics of Judicial Appointments (2005); John R. Lott, Jr., “The Judicial Confirmation Process: The
Difficulty with Being Smart,” 2 Journal of Empirical Legal Studies 407 (2005).
24 The Basic Model

the same party as the one who appointed the retiring Justice.10 Still an-
other bit of evidence is what might be called “ideology drift”—the ten-
dency of judges to depart from the political stance (liberal or conserva-
tive) of the party of the President who appointed them the longer they
serve.11 A judge closely aligned with the ideology of the party of the
President who appointed him may fall out of that alignment as new, un-
foreseen issues arise. A judge who was conservative when the burning
issues of the day were economic may turn out to be liberal when the
burning issues become ones of national security or social policy such as
abortion or homosexual rights.
There is more: the outcome of Supreme Court cases can be predicted
more accurately by means of a handful of variables, none of which in-
volves legal doctrine, than by a team of constitutional law experts.12
While there is a high correlation between how a given federal appellate
judge (court of appeals judge as well as Supreme Court Justice) votes for
the government in nonunanimous (hence “close”) constitutional crimi-
nal cases and in nonunanimous statutory criminal cases, there is a low
correlation between the votes of different judges for and against the gov-
ernment in criminal cases.13 Some judges have a progovernment lean-
10. Ross M. Stolzenberg and James Lindgren, “Politicized Departure from the United States
Supreme Court” (University of Chicago and Northwestern University, Mar. 18, 2007); James F.
Spriggs and Paul J. Wahlbeck, “Calling It Quits: Strategic Retirement on the Federal Courts of
Appeals, 1893–1991,” 48 Political Research Quarterly 573 (1995); Deborah J. Barrow and Gary
Zuk, “An Institutional Analysis of Turnover in the Lower Federal Courts, 1900–1987,” 52 Jour-
nal of Politics 457, 467–468 (1990). Another straw in the wind is the surprising finding in a
recent study that Supreme Court law clerks’ self-described political identity (Democratic or Re-
publican) influences the political valence of their Justices’ votes. Todd C. Peppers and Christo-
pher Zorn, “Law Clerk Influence on Supreme Court Decision Making” (Roanoke College, De-
partment of Public Affairs, and University of South Carolina, Department of Political Science,
June 14, 2007).
11. See Andrew D. Martin and Kevin M. Quinn, “Assessing Preference Change on the US
Supreme Court,” 23 Journal of Law, Economics and Organization 365 (2007); Susan Haire, “Be-
yond the Gold Watch: Evaluating the Decision Making of Senior Judges on the U.S. Courts of
Appeals” (University of Georgia, Department of Political Science, 2006).
12. Andrew D. Martin et al., “Competing Approaches to Predicting Supreme Court Deci-
sion Making,” 2 Perspectives on Politics 761 (2004); Theodore W. Ruger et al., “The Supreme
Court Forecasting Project: Legal and Political Science Approaches to Predicting Supreme Court
Decisionmaking,” 104 Columbia Law Review 1150 (2004). The variables are “(1) circuit of ori-
gin; (2) issue area of the case; (3) type of petitioner (e.g., the United States, an employer, etc.);
(4) type of respondent; (5) ideological direction (liberal or conservative) of the lower court rul-
ing; and (6) whether the petitioner argued that a law or practice is unconstitutional.” Id. at
1163.
13. Ward Farnsworth, “Signatures of Ideology: The Case of the Supreme Court’s Criminal
Docket,” 104 Michigan Law Review 67 (2005); Farnsworth, note 7 above.
 Nine Theories of Judicial Behavior 25

ing, others a prodefendant leaning, and these leanings appear to be what

drives their votes in close cases whether the case arises under the Con-
stitution or under a statute—though from a legalist standpoint the text
of the enactment being applied ought to drive the outcome, and there
are huge textual differences between the Constitution and statutes. Apo-
litical judges would not be expected to vote the same way in both types
of case.
All this is not to say that all judicial votes are best explained as politi-
cally motivated,14 let alone that people become judges in order to nudge
policy closer to their political goals. We shall see in subsequent chapters
that to explain the political cast of judicial decisions does not require as-
suming that judges have conscious political goals. No attitudinal study
so finds, and data limitations cannot explain the shortfalls. Even at the
level of the U.S. Supreme Court many cases do not involve significant
political stakes, but that cannot be the entire explanation either. Think
of Oliver Wendell Holmes. The publication of his correspondence after
his death revealed that he was a rock-ribbed Republican, yet he voted re-
peatedly to uphold liberal social legislation (such as the maximum-
hours law at issue in the Lochner case, in which he famously dissented)
that he considered socialist nonsense. He may of course have been an
exception among Supreme Court Justices in this as in so many other re-
spects. He may have few successors in point of political detachment in
today’s more politicized legal culture.
We get a sense of the attitudinal model’s predictive limitations in Ta-
bles 5 and 6, in which judicial votes that lack any political valence are
coded as “other,” and the liberal, conservative, mixed, and other votes
are correlated with the party of the President who appointed the judge
who cast the vote. Notice that apart from the substantial percentage of
votes that were either mixed or other, a large percentage of conservative
votes were cast by putatively liberal judges (judges appointed by Demo-
cratic Presidents) and a large percentage of liberal votes were cast by pu-
tatively conservative judges. Notice, as in the earlier tables, the apparent
trend toward the increased politicization of court of appeals voting re-
14. See, for example, Cross, note 7 above; Cross, note 2 above, at 285–311; Sunstein et al.,
note 7 above; Daniel R. Pinello, “Linking Party to Judicial Ideology in American Courts: A
Meta-Analysis,” 20 Justice System Journal 219 (1999); C. Neal Tate and Roger Handberg, “Time
Binding and Theory Building in Personal Attribute Models of Supreme Court Voting Behavior,
1916–88,” 35 American Journal of Political Science 460 (1991); Sheldon Goldman, “Voting Be-
havior on the United States Courts of Appeals Revisited,” 69 American Political Science Review
491 (1975).
26 The Basic Model

Table 5 Judicial Votes in Courts of Appeals as Function of Party of Appointing

President, 1925–2002 (in percent)

Republican Democratic
Vote President President

Conservative 42.2 37.6

Liberal 28.1 33.3
Mixed 5.9 5.1
Other 23.9 23.9
Sources: Appeals Court Attribute Data, www.as.uky.edu/polisci/ulmerproject/auburndata.htm
(visited July 17, 2007); U.S. Court of Appeals Database, www.as.uky.edu/polisci/ulmerproject/
appctdata.htm, www.wmich.edu/⬃nsf-coa/ (visited July 17, 2007). Votes were weighted to
reflect the different caseloads in the different circuits. “Mixed” refers to multi-issue cases in
which the judge voted the liberal side of one or more issues and the conservative side of the
other issue or issues.

Table 6 Judicial Votes in Courts of Appeals as Function of Party of Appointing

President, Judges Serving Currently (in percent)

Republican Democratic
Vote President President

Conservative 51.2 42.5

Liberal 22.9 33.1
Mixed 7.3 7.6
Other 18.7 16.9
Sources: Appeals Court Attribute Data, www.as.uky.edu/polisci/ulmerproject/auburndata.htm
(visited July 17, 2007); U.S. Court of Appeals Database, www.as.uky.edu/polisci/ulmerproject/
appctdata.htm, www.wmich.edu/⬃nsf-coa/ (visited July 17, 2007). Votes were weighted to
reflect the different caseloads in the different circuits. “Mixed” refers to multi-issue cases in
which the judge voted the liberal side of one or more issues and the conservative side of the
other issue or issues.

sulting from judicial appointments by Republican Presidents. But no-

tice, too, that the differences between the two types of judge, exhibited
in the first two rows of the tables, though significant, are only partial.
And a comparison just of means obscures the fact that the distributions
overlap; some judges appointed by Republican Presidents are less con-
servative than some appointed by Democratic Presidents. This does not
refute the attitudinal model, but it does highlight the fact that the party
of the appointing President is an imperfect proxy for a judge’s judicial
ideology. One reason is that ideological issues important to judges need
 Nine Theories of Judicial Behavior 27

not have salience in political campaigns; capital punishment is a current

example. Another reason is that judges pride themselves on being politi-
cally independent rather than party animals.
An explanation for the attitudinal model’s predictive limitations that
would hold even if all decisions involved significant political stakes is
that a case may pose a conflict between two political values, both of
which are important to a judge, as when, for example, a civil rights suit
(liberal) is brought challenging affirmative action (a conservative bête
noire). One might think that in such a case the political considerations
would cancel and the decision could be attributed to conventional legal
reasoning. But no; the political considerations are unlikely to weigh
equally in the judge’s mind, and if they do not, the heavier may deter-
mine his decision. A notable example is Buchanan v. Warley.15 Decided at
a time when the Supreme Court was strongly disinclined to invalidate
racially discriminatory laws, it nevertheless invalidated a southern ordi-
nance that forbade blacks to live on any block in which whites were in
the majority, and vice versa. The ordinance had blocked the plaintiff, a
white, from selling property to a black. The Court distinguished mere
“social rights”—the right of blacks to associate with whites (and like-
wise of whites not to associate with blacks, a “right” that the whites who
remained in the neighborhood were denied), which the Court had re-
fused to recognize in Plessy v. Ferguson—from “those fundamental rights
in property” that the Fourteenth Amendment was intended to secure to
blacks on equal terms with whites.16 The distinction is not found in the
equal protection clause. Michael Klarman argues persuasively that the
Court simply thought government interference with property rights a
worse affront to personal liberty than segregation of schools and other
public facilities, especially since the person complaining that his prop-
erty rights were being infringed was the white seller.17 The upshot was
that the Court issued a liberal decision, rejecting racial segregation in
private housing.
The attitudinalists’ traditional preoccupation with politically charged
cases decided by the Supreme Court creates an exaggerated impression
of the permeation of American judging by politics.18 Most cases decided

15. 245 U.S. 60 (1917).

16. Id. at 79.
17. Michael J. Klarman, Unfinished Business: Racial Equality in American Law 83–84 (2007).
18. Brian Leiter, Naturalizing Jurisprudence: Essays on American Legal Realism and Natural-
ism in Legal Philosophy 187, 188 n. 22, 192 (2007).
28 The Basic Model

by American courts are neither politically charged nor decided in the

Supreme Court. And to use the political party of the appointing Presi-
dent as a proxy for a Justice’s political inclinations misleadingly implies
that partisan politics pervades Supreme Court decision making. A Presi-
dent invariably appoints most judges (usually more than 90 percent)
from among members of his own political party, but once appointed they
are more likely to want to be good judges than to want to toe anyone’s
political line. You do not find judges saying, even to themselves, “How
would Bill Clinton [or George Bush, etc.], who appointed me, decide
this case?” Nevertheless, in the frequent cases in which a political judg-
ment is required to “close the deal” because legalist analysis of the case
leads nowhere, the judge is likely to lean toward the position that the
political party to which he belongs (or belonged) would support, for it is
usually not an accident that he belongs to that party rather than another.
But “lean toward” is different from “identify with.” Supreme Court Jus-
tices are political, but politically independent. Most of them, indeed, are
outside (either more liberal or more conservative) the range bounded by
the political preferences of the President and the Senate that confirmed
them.19
Any amount of political judging challenges orthodox conceptions of
the judicial process, however, and the attitudinalists have shown that
there is plenty at all levels of the American judiciary (though more, the
higher the level). Yet their findings, while heresy to the legal establish-
ment, have the paradoxical effect of blunting criticisms of the courts as
acting undemocratically when they invalidate legislative and executive
acts. As explained by Mark Graber,

Judicial review is established and maintained by elected officials. Adju-

dication is one of many means politicians and political movements em-
ploy when seeking to make their constitutional visions the law of
the land. Elected officials provide vital political foundations for judi-
cial power by creating constitutional courts, vesting those courts with
jurisdiction over constitutional questions, staffing those courts with
judges prone to exercising judicial power, assisting or initiating litiga-
tion aimed at having those courts declare laws unconstitutional, and
passing legislation that encourages justices to make public policy in
the guise of statutory or constitutional interpretation. Judicial review
19. Michael Bailey and Kelly H. Chang, “Comparing Presidents, Senators, and Justices:
Interinstitutional Preference Estimation,” 17 Journal of Law, Economics and Organization 477,
508 (2001).