the strategic account and the rule of law jeffrey r. lax · · 2013-05-08the strategic account...

1

The Strategic Account and the Rule of Law Jeffrey R. Lax

What follows are some brief and informal thoughts on the normative implications for the rule of law of the empirical (and theoretical) work on the strategic judicial decision making. They build off of my recent piece for the Annual Review of Political Science, which is also attached here. In effect, this brief introduction is a reader’s guide highlighting what I think of as relevant from that longer piece along with the comments I would insert in the ARPS piece had I paid the attention I should have to normative concerns. In particular, I want to highlight even more what I (in my classes) call the (Barry) Friedman critique—why should anyone outside the field of judicial politics be surprised by X or care about Y? And I think I have been strongly influenced by Vermuele (2005) as well, but am coming to similar conclusions from a different angle.

In the ARPS, I argue that moving forward in judicial politics was hindered by three roadblocks. The first was disagreement over the question “what is law?” (p134-5). My answer, or at least what I suggested as a useful answer, was to think about law as doctrinal instrumentalism---the making of policy through legal rules. In a way, this was also my answer to the question “what is policy?” (p146). In short, we can see the core object of struggle within judicial politics as legal rules. This clearly connects to the rule of law as a law of rules, though I do not mean to invoke the full Scalia-Friedman debate here. Rather, I think our work on judicial politics can best proceed by remembering that when judges interact over policy and law they do so by applying and developing legal rules. (In the paper, I also push for more structure in our analysis of collegial and hierarchical interactions, but that is largely orthogonal to our purposes here.)

What are the implications of judicial interactions over legal rules for the rule of law? I turn here to the hierarchy of claims I attribute to the “strategic account” of judicial politics scholarship (p139-40), which seeks to highlight such interactions as an object of study. These claims characterize a body of work that was for a long time more an empirical body of work than a systematically theoretical one.

• S1: The justices care about more than just the final votes they cast. • S2: The justices have preferences over legal policy and take each other’s positions

and preferences into account when crafting opinions and responding. • S3: The collegial game has actual consequences for legal policy. • S4: Justices might cast insincere merits votes if it means trading the case

disposition for better policy content in the opinion.

At the same time, we should keep in mind some of the normal requirements of the “rule of law.” For my purposes, I steal a list of concepts, but not the nuance of them, from Jamal Greene’s interesting essay for this conference. These requirements are:

§ Law should be part of a process § The process should not be arbitrary § The process should be stable § The process should be replicable § The process should be logic driven.

2

The first strategic account claim (S1) seems completely uncontroversial, even to the purest attitudinalist within judicial politics. But there is a hidden normative element even there. We do want judges to be thinking more broadly than in individual cases. We do want them thinking in terms of more general legal rules, in pursuit of indeed all the five bits of the Rule of Law. So far so good.

What about S2, that the justices interact and accommodate each other’s views? In the ARPS, I called this uncontroversial as well, even invoking the Friedman critique which wondered why we spent so much time establishing such a trivial point. But if we think about S2 normatively, it’s not as uncontroversial perhaps. Even if we are willing to attribute to individual justices good intentions, faithfulness to a philosophy of law and the like, it is not clear that we should be happy that the justices accommodate and compromise with each other. Compromise over principles is a tricky thing. Yes, we can all see why it is politically necessary, but does it not seem strange for justices to regularly compromise and agree to opinion language that does not actually reflect what they believe the law to require? Surely, by S2 we must mean they accommodate more than trivial detail, and yet it is worrisome to think that judges will go along with statements about law, statements that we want to be purely legal, just to get along or reach an end to “deliberations”1 or achieve greater influence over the law and future cases. On the other hand, one can argue that normatively this is a good thing in that the end product is better off for these compromises and accommodations, and that we shouldn’t worry so much normatively at the level of individual judges but rather focus more on the system.

But even there we have problems. As briefly reviewed in the ARPS piece, other work (e.g., Sager & Kornhauser, Lax, Landa & Lax) shows that when we take logical structures like legal rules, “weird things” happen. There isn’t a clean process for aggregating legal rules or even sets of legal findings. Outcomes can therefore be arbitrary, unstable, unreplicable, and illogical. Aggregate legal rules might not resemble the rule believed to be correct by any one judge. They may seem very different in structure from “normal” legal rules.

Turning to S3, it gets even worse. We (many of us) think that the “collegial game” (as Maltzman, Spriggs, and Wahlbeck so wonderfully label the interactions between the justices) matters, that legal policy and the law are affected by the play of the game and how the justices make their choices and maneuver and accommodate. After all, why would we care, if all the strategic account meant was that there was a noisy path to a predetermined end? It must be implicitly believed by strategic account scholars (strategic accountants?) that different rules wind up in place because of the collegial game. So now, even beyond the logical problems of collegial rule aggregation in a passive sense, we have active strategic maneuverings, not even out of some higher normative purpose, but to best influence law as one wants it to be. On the other hand, one could argue that such manipulations themselves do have a higher purpose, that if Justice Brennan in the Maltzman-Spriggs-Wahlbeck account of PA v. Muniz was maneuvering to keep “Sandra[‘s] hands” off of Miranda rights, he was doing so to get better law out of the game. Even here though, if law depends on this game, it seems overly optimistic to hope 1 And we all know they really do not deliberate much in any real sense.

3

this game is non-arbitrary, stable, replicable, and logical. S4, which would mean that in some plays of this game the justices might even sign onto dispositions of cases they do not agree with, seems to make this even worse. And, even if they do not do so in a case directly before them, any time a justice agrees to a legal rule she thinks is the wrong one, she is agreeing to future case dispositions she thinks wrong.

To be clear, all of this affects all the particular questions we study in connection to collegial court politics, such as the effects of opinion assignment, as well as those of separation of powers games, and perhaps even hierarchical games. Neither collegiality nor hierarchy mix well with the Rule of Law.

PL14CH07-Lax ARI 14 April 2011 17:3

The New Judicial Politicsof Legal DoctrineJeffrey R. LaxDepartment of Political Science, Columbia University, New York, New York 10027;email: [email protected]

Annu. Rev. Polit. Sci. 2011. 14:131–57

First published online as a Review in Advance onFebruary 14, 2011

The Annual Review of Political Science is online atpolisci.annualreviews.org

This article’s doi:10.1146/annurev.polisci.042108.134842

Copyright c! 2011 by Annual Reviews.All rights reserved

1094-2939/11/0615-0131$20.00

Keywordscase space, attitudinal and legal models, collegial court, hierarchy

AbstractA new judicial politics of legal doctrine has the potential to resolvefoundational dilemmas and reconcile long-standing and counterpro-ductive scholarly divisions by bringing together legal concerns and po-litical science priorities. This doctrinal-politics approach highlights arelatively new formal apparatus known as the case-space model, and itinvokes close ties between theoretical and empirical work and betweenthe study of judicial behavior and actual legal practices and institutions.The case-space model is an adaption of standard policy-space model-ing, tailored for the distinguishing features of judicial policy making.It allows for ideological differences between judges while expressingthose differences in terms of legal rules that partition fact-filled legalcases into different dispositions. I explore the intellectual origins andprimary contributions of the approach, focusing on how legal policy isaffected by collegiality (the multi-member nature of appellate courts)and hierarchy (the multi-level division of court systems).

131

Ann

u. R

ev. P

olit.

Sci

. 201

1.14

:131

-157

. Dow

nloa

ded

from

ww

w.a

nnua

lrevi

ews.o

rgby

Col

umbi

a U

nive

rsity

on

05/1

7/11

. For

per

sona

l use

onl

y.


A LOOK BACK ON “A LOOKBACK, A LOOK AHEAD”

A decade ago, Epstein & Knight (2000) docu-mented a “sea change” in the field of judicialpolitics: the “strategic account” was replacingthe social-psychological paradigm, which haditself triumphed over both legal formalism andearlier scholarship on judicial strategy. Theiressay, “Toward a Strategic Revolution in Judi-cial Politics: A Look Back, a Look Ahead,” hadthree focal arguments (p. 626): “(1) social ac-tors make choices in order to achieve certaingoals, (2) social actors act strategically in thesense that their choices depend on their expec-tations about the choices of other actors, and(3) these choices are structured by the institu-tional setting in which they are made.”

One goal was the “unification of judicial spe-cialists and their counterparts in other fieldsof political science, as well as of those withinthe law and courts area—especially scholarswho study legal doctrine and those who focuson judicial behavior” (Epstein & Knight 2000,p. 652). But although a sequel a decade laterindeed recounted a “tsunami” of strategic ac-count scholarship (Epstein & Jacobi 2010), thedesired unification of scholars has not occurred.The legal academy still shows little interest inand is given little attention by judicial-politicsscholarship (Friedman 2006, Fischman & Law2009). One culprit is the rigidly divided frame-work popular in judicial-politics research—“legal model” versus “attitudinal model” versus“strategic model”—a framework that obscuresas much as it reveals. “Attitudinal” variablesare pitted against “legal” variables and some-times against “strategic” variables, with the bal-ance sorted out empirically but rarely theoreti-cally. While clarifying key disagreements, thelegal-attitudinal-strategic framework has alsostunted theory development and blocked thesort of fruitful intersections Epstein & Knightlikely had in mind (to be sure, Epstein & Jacobiand others might read this history differently).The opposition encouraged by that frame-work has limited the impact of political science

scholarship on courts and prevented useful bor-rowing across the divisions.

Despite these problems, my own assessmentis optimistic. Recent developments have thepotential to reconcile long-standing divisions,divisions that I argue have hampered progressand have left major questions unresolved oreven unresolvable. This reconciliation—thisnew judicial politics of legal doctrine, or“doctrinal politics” for short—builds on avast body of previous work of a politicalscience bent, particularly the “legal realist”movement as adapted from the legal academy,the “behavioralist revolution” and its empiricalsuccesses, the “strategic revolution,” the rise ofpositive political theory, and the new emphasison “Empirical Implications of TheoreticalModels” (EITM, the NSF-supported peda-gogical program on theory testing). But it alsoputs the primary concerns of the legal academy(and what some political scientists call the legalmodel) at the heart of judicial politics.

Scholars have begun to incorporate legalpractices—not to flash that overused badge “in-terdisciplinary” but to better understand judi-cial politics. It has become more clear that theaccoutrements of law are not just superfluousdetails to be set aside by modelers and othersophisticated political scientists; the idiosyn-cracies of legal policy making are critical forbuilding useful theories, be they formal or in-formal. Meanwhile, theories of adjudication areincorporating the plural and interactive natureof judging to a much greater extent, exploringboth collegiality (the multi-member nature ofappellate courts) and hierarchy (the multi-leveldivision of court systems). At the same time,some progress has been made in burning falsebridges to legislative politics. Some of the in-triguingly different features of explicitly judicialpolicy making are beginning to receive the ded-icated attention they deserve. First-generationmodels borrowed heavily from legislative poli-tics. Now the action is in creating models thatare distinctively judicial.

My goal is to pull together these trendsin the literature to articulate a new “doctrinal

132 Lax

Ann

u. R

ev. P

olit.

Sci

. 201

1.14

:131

-157

. Dow

nloa

ded

from

ww

w.a

nnua

lrevi

ews.o

rgby

Col

umbi

a U

nive

rsity

on

05/1

7/11

. For

per

sona

l use

onl

y.


politics,” an approach that can serve as a uni-fying framework, synthesizing the best of cur-rent and previous work across paradigms. In-deed, I see it as a natural outgrowth of suchwork. Rather than focus on differences acrossthe legal-attitudinal-strategic and other otherdivides, I extract the core messages of each sep-arate approach and show how they can speak toeach other. This synthesized approach has hadstrong successes in recent years both within ex-isting lines of research and in raising new ones,and I argue it shows great promise.

This doctrinal-politics approach has two keyarguments that organize and concretize thestrategic account. One is that judicial prefer-ences and goals are structured by the choicesavailable to judges: dispositions and the rulesthat denote general sets of case dispositions.The second is that judicial choices, rule cre-ation, and rule application are deeply affected bycollegial and hierarchical (horizontal and verti-cal) divisions of power.

I argue in favor of continuing to developthe strategic account into a more concrete ap-proach, rather than a generalist approach high-lighting only that justices interact (see Epstein& Knight 2000, Epstein & Jacobi 2010). Muchof the progress over the past decade has beenmade in line with the following four priorities.

First, our models of courts should recognizejudicial structures. One formal tool for doing sois the “case-space model,” which is designed tocapture them explicitly. This model highlightslegal cases as the vehicles for policy making,but of course judicial policy making is morethan simply case disposition after case dispo-sition. When appellate courts address judicialpolicy more generally, they typically do so inopinions that establish (new or modified) le-gal rules for deciding current and future cases.The possible set of cases is the “case space.” Arule is a partition of the case space into win-ners and losers (see Figure 1). Different judgescan have different preferences over legal rules,which is to say different partitions of the casespace. Battles over law are struggles over parti-tions of the case space. Development of the lawoccurs through articulation of and movements

C E

D

B

A

Dimension 1

Dimension 2

F G

•x1

•x3

•x2

Figure 1This sample case space is two-dimensional, continuous, and bounded. Eachpoint in this space represents a potential fixed case (instead of representing apolicy, as in policy space). On dimension 2, case x1 is the most extreme and x3the least. On dimension 1, x3 is the most extreme and x2 the least. Sevenlettered rules are given. A case is decided as Yes if it falls on or below the ruleand No if it is above (too extreme for) the rule. Thus, all seven rules givenwould say x1 should be decided No. Case x2 is decided as Yes by only rules A, B,and C. Case x3 is decided as Yes by only rules A, B, and F. Rule G at mostaccepts the limiting case at the origin. Rules A, B, C, and F establish a fixedthreshold on each dimension. Rules D and E balance the two dimensions. RuleC and Rule F are conjunctive rules for getting a Yes; the case must lie within thethreshold for dimension 1 and the threshold for dimension 2. Rule A is adichotomous rule for getting a Yes; the case must lie within the threshold fordimension 1 or the threshold for dimension 2. If a Yes outcome is theconservative outcome, then Rule A is the most conservative and Rule G theleast, but not all Rules B through F could be ordered straightforwardly (wecould still say, e.g., Rule D is more conservative than Rule E).

in this partition. A case-space model thus rec-onciles competing modes of analysis of judicialbehavior, integrating the practice of law witha political, policy-seeking perspective. Judicialbehavior can be consistent with legal theory inform and function, while consistent with polit-ical accounts of the incentives faced by judges.

A second priority of doctrinal-politics re-search should be empirical investigation of thestructure of judicial choice. One empirical toolfor doing so that is well suited to the case-spacetheoretical model is the fact-pattern approach,which is already familiar to judicial scholars. Inaddition, we need more data on the content andstructure of judicial preferences and legal opin-ions (that is, more data on how judges want thecase space partitioned).

www.annualreviews.org • The New Judicial Politics of Legal Doctrine 133

Ann

u. R

ev. P

olit.

Sci

. 201

1.14

:131

-157

. Dow

nloa

ded

from

ww

w.a

nnua

lrevi

ews.o

rgby

Col

umbi

a U

nive

rsity

on

05/1

7/11

. For

per

sona

l use

onl

y.


Third, we should continue to study col-legiality and hierarchy, theoretically andempirically, but also focus more on how theyinteract. One new line of scholarship hasdemonstrated striking incentives for doctri-nal choice that arise because of collegiality,hierarchy, and how they intersect.

Finally, scholarship on doctrine should (fol-lowing the lead of Fallon 2001, Shapiro 2006,Heytens 2008, etc.) focus more on how judgesuse doctrine to get what they want (and the workon doctrine flowing from Kornhauser 1992a,bshould receive more attention from politicalscientists). This can feed back into further for-mal and empirical work and vice versa.

Epstein & Knight realized that the strate-gic account would not resolve all scholarly con-flicts. The new politics of legal doctrine will noteither. Rather, to adapt from Epstein & Knight(2000, p. 638), my “broader claim is that thestrategic account [as enacted in this new poli-tics of legal doctrine] not only helps us to eluci-date the ‘right’ questions but it also provides amethodology to help scholars answer them, to

EVIDENCE FOR THE ATTITUDINAL MODEL

The authors claim no decisive test, but rather “clear and con-vincing evidence of the overwhelming importance of the justices’attitudes and values” in predicting the final vote in each case(Segal & Spaeth 2002, p. 97). There is negative evidence thatrules out legal factors (e.g., Segal & Spaeth 2002, 2003; Spaeth& Segal 1999), but see Friedman (2006) and Bailey & Maltzman(2008). The most direct positive evidence is in tables 8.2 and 8.3(Segal & Spaeth 2002, pp. 320, 326): Controlling for the facts ofeach case, ideology scores correlate strongly with vote probabili-ties, and postestimation predictive success is high. If law acts as apressure on all justices (indeed, any force that is constant acrosscases and justices), it can only enter this model through the in-tercept. What the attitudinal model explains then is differencesin voting between justices, differences that match intuitive ideo-logical patterns. Before the model is run, one cannot make pointpredictions (for why Justice Brennan’s attitude score correlatesto a specific search-and-seizure liberal vote rate). Attitudes can-not directly explain the “base” level of liberalness captured by theintercept in the model.

understand the range of choices that contributeto the development of law.”

In the next section, I argue that thedoctrinal-politics approach resolves founda-tional dilemmas in judicial politics. I explain theorigins of the approach and then offer an assess-ment of the literature to date that grounds therecommendations above. Before concluding, Ihighlight the main contributions of this newapproach. The Appendix gives a nontechnicalreview of the case-space model, which mightbe helpful for understanding some of the dis-cussion that follows.

ROADBLOCKS AND A ROADMAPIn moving forward, we have faced three road-blocks, three challenging questions that haveresisted resolution: What is law? What is legaldiscourse? What is policy?

Roadblock 1: What is Law?The founding debate of judicial politics—whether judges find law or make law—yieldedan uneasy relationship with legal doctrine.Whereas legal formalists rejected any role forideology and politics in judicial decision mak-ing, the legal-realism movement and attitudi-nalists rejected law as prime mover. But as a re-sult, any invocation of legal rules, doctrine, orprecedents became suspect. Even legal scholarswho agree that judging can be political com-plain that much political science trivializes lawand the legal enterprise (e.g., Friedman 2006,Tiller & Cross 2006). That is, a significantnumber of legal scholars accept legal realismbut believe ideology plays a smaller role in judi-cial behavior than do attitudinalists (see sidebar,“Evidence for the Attitudinal Model”).

The founding debate between legalists andattitudinalists recognizes only two roles for law:as a constraint on policy making or as a cloak forit. The legal model as defined by Segal & Spaeth(2002) is a set of arguments that ideology has noor little effect on judging and that standard legalconcepts such as precedent and text do almostall the work in deciding cases. Attitudinalists

134 Lax

Ann

u. R

ev. P

olit.

Sci

. 201

1.14

:131

-157

. Dow

nloa

ded

from

ww

w.a

nnua

lrevi

ews.o

rgby

Col

umbi

a U

nive

rsity

on

05/1

7/11

. For

per

sona

l use

onl

y.


say only ideology affects final votes in SupremeCourt cases (and that ideology likely trumps lawat other stages as well). When law is invokedby the justices, it is only to cloak or disguiseideological behavior (perhaps consciously, per-haps as self-delusion). Finally, in the strategicaccount, law is epiphenomenal; it matters asthe output of the judicial process: “[L]aw, asit is generated by the Supreme Court, is the re-sult of short-term strategic interactions amongthe justices and between the Court and otherbranches of government” (Epstein & Knight1998). Or it is a normative constraint to bestrategically accommodated (Knight & Epstein1996). The lack of a mutually acceptable answerto the question “What is law?” is, in my opin-ion, the reason why the various blocs of schol-ars “talked past each other for decades, exhibit-ing fundamental disagreements over how bestto study judicial decisions” (Epstein & Knight2000, p. 653).

Resolution: Doctrinal InstrumentalismOne contribution of the doctrinal-politics ap-proach is that it invokes a different way of think-ing about law (specifically the role of legal rulesand doctrines), which I call “doctrinal instru-mentalism.” In short, judges make policy usinglegal rules. They dispose of cases and more gen-erally craft rules for case dispositions. A case-space model is one formalization of this. Thekey difference in a case-space model is that in-stead of each policy or choice being a pointin space as in the standard policy model, thepoints in the case space are legal cases and a le-gal policy is a partition of the case space into twoor more sets. A case-space model thus placesjudge-created case-sorting rules at the heart ofjudicial policy making. (Note that this also an-swers the question, “What is policy?”)

Law, in this sense, is not merely compatiblewith the view of judges as policy seekers, butmust be understood in order to properly un-derstand such policy-seeking behavior. Think-ing of judges as political creatures does not ob-viate the need to think about cases and rules.

Judicial preferences will be over legal rules thatgovern cases. This resolves the conflict betweenattitudinalists and strategists on one side and le-gal scholars on the other by arguing that pref-erences and law work together, that we canemphasize policy goals but should not neglectthe instruments by which they might be pur-sued. (This does not rule out that law can becloak and constraint as well.) Attitudinalists arecorrect to argue that Supreme Court justiceshave great ideological freedom; strategists areright to prioritize larger policy goals and topoint out complexities in achieving them; andlegalists are right to argue that doctrine mat-ters. The doctrinal-politics approach is a “legalmodel” I hope both attitudinalists and strate-gists can accept. It is also a strategic account thatlargely incorporates attitudinalist priorities andyet should resonate with legal scholars.

Roadblock 2: What isLegal Discourse?In attitudinalism, law is only used to hide pol-itics. As Friedman (2006, p. 266) notes, atti-tudinalism seems to argue that “what lawyers,judges, and legal academics spend years learn-ing, practicing, and theorizing about is mean-ingless, . . . and that the entire discourse of law isan illusion.” As Knight & Epstein (1996) pointout, the justices talk about precedents and otherlegal accoutrements all the time, even in theirprivate memos, which is strange if law is just acloak to deceive outsiders.

Resolution: Legal InstrumentalismAnother possibility, what I call “legal instru-mentalism,” arises from a new line of workin both political science and legal scholarship:To make policy effectively, higher court judgesneed not only to formulate legal rules but alsoto communicate them, which requires engage-ment in the very legal discourse that is dis-missed out of hand by pure attitudinalism. Evena dictatorial judge could not list every possi-ble case along with its desired disposition. Even


Ann

u. R

ev. P

olit.

Sci

. 201

1.14

:131

-157

. Dow

nloa

ded

from

ww

w.a

nnua

lrevi

ews.o

rgby

Col

umbi

a U

nive

rsity

on

05/1

7/11

. For

per

sona

l use

onl

y.


the most ideological appellate judge must makepolicy by telling lower court judges what casefacts to consider and how much, just as a judgebound by legal principle would. They will thuscare about precedents, legal philosophies, andthe like because they wield these as commu-nicative devices to achieve policy goals. Whatseparates legal instrumentalism from attitudi-nalism is that attitudinalism sees legal discourseas only a cloak; legal instrumentalism argues re-liance on legal discourse can affect actual behav-ior and outcomes. What separates legal instru-mentalism from the concept of law as norm isthat reliance on legal discourse constrains pol-icy seeking whether or not a judge respects thenorm of precedent.

The uses and usefulness of legal instrumen-talism are constrained by the limitations of le-gal discourse as tool. (The same could be saidof any tool; consider how the legal-attitudinal-strategic framework both helps and hindersscholarly communication!) Legal discourse isinherently ambiguous, so it can be difficult toconvey exactly how the higher court wants a le-gal rule, generally stated, to be applied specif-ically, in the range of possible cases that canarise. Indeed, if attitudinalists are right thatSupreme Court justices need to hide just howmuch ideology enters into their decisions, theycannot speak to the lower courts freely butmust wink and nod and otherwise communi-cate using legal discourse to convey ideologicalgoals in the language and grammar permittedthem.

To make policy, a policy-seeking judge mustuse the language of judicial policy making. Shemust make use of the standard mechanismsof legal discourse as inculcated in law school,not to figure out the right answer but to enacther preferred (i.e., “right”) answer. To gener-alize from Bueno de Mesquita & Stephenson(2002), it is not that judges care about law in-stead of or in addition to policy; judges careabout law because they care about policy. Law isboth the substance of policy (as per the strategicaccount) and the means of policy making. Us-ing law instrumentally, a judge can cite prece-dent, using actual case examples to “highlight

which factors [she] considers important” andcan use “analogical anchoring” and “analog-ical reasoning” (Shapiro 2006, pp. 315, 325).She must make use of, and is thus constrainedby, the entire legal web in which judicial be-havior is embedded. Although casting a votemay be a relatively trivial act, crafting an ef-fective legal opinion is obviously not. Note thatsome formal models respect legal instrumen-talism in the form of uncertainty and error;they model the justices as actively crafting theiropinions to achieve their desired policy out-comes, inhibited by such uncertainty, ambigu-ity, and complexity (e.g., Bueno de Mesquita &Stephenson 2002, Lax 2007b, Lax & Cameron2007).

Roadblock 3: What is Policy?I now connect the doctrinal-politics approachto the question “What is policy?” and show whyanswering this question is so important for mak-ing further progress building on the strategicaccount. I argue that insufficiently specific an-swers to this question prevented the work in thegeneralized strategic account from reaching itsfull potential. In building on attitudinalism, theimpact of the strategic account was still limitedbecause we lacked a structure for understand-ing the content of the policy battles at its heart.The strategic account is not a specific strate-gic model but rather a “conception of judicialdecision making” consisting of “a range of ex-planations . . . that typically emanate from thefindings of previous empirical investigations—rather than from any overarching theory”(Epstein & Knight 1998, p. xii). The fault maylie on the side of theorists who did not givestrategic empiricists a rich enough apparatus toexplain the obviously complex web of interac-tions they were uncovering. Standard policy-space models only take us so far, as “modelsbased in ‘policy space’ are too far removed fromthe details of adjudicatory practice to explainmuch of judicial activity” (Kornhauser 2008,p. 5). One hears echoes in this complaint of sim-ilar concerns decades earlier in the legislativeliterature:

136 Lax

Ann

u. R

ev. P

olit.

Sci

. 201

1.14

:131

-157

. Dow

nloa

ded

from

ww

w.a

nnua

lrevi

ews.o

rgby

Col

umbi

a U

nive

rsity

on

05/1

7/11

. For

per

sona

l use

onl

y.


In their quest for analytical generality, mostformal theorists had suppressed institutionaldetails . . . , thinking that to include themwould be to specialize and render idiosyn-cratic otherwise general theories . . . . How-ever, . . . by suppressing these details alto-gether, they robbed their own theories of anygenerality by modeling rational behavior inthe most spartan of all institutional settings—that described only by a simple rule of prefer-ence aggregation. (Shepsle 1989, p. 135).

Policy-space models did not offer strategicempiricists or formal theorists a structure forconsidering differences in judicial preferencesnor the full range of choices justices make. Thecase-space model is just such a structure.

Resolution: Case-Space ModelingEpstein & Knight (1998) write of The ChoicesJustices Make; these choices can be modeled asdispositions and rules. When Maltzman et al.(2000) write of the Supreme Court as playinga Collegial Game, I see this game as one overcase-space partitions. This provides the missingstructure for understanding strategic collegialcourt politics as well as hierarchical politics. AsKornhauser (2008, p. 2) puts it,

Models grounded in policy space abstractfrom the specific cases that, in actual adjudica-tion, trigger judicial intervention and are theoccasion for whatever policy-making courtsdo. Policy space models treat policies as fun-damental and cases as nonexistent. Modelsgrounded in case space, by contrast, take casesas fundamental; policies are described in termsof case dispositions. . . . [They] have a certainpriority over policy space models . . . [because]case space provides microfoundations for pol-icy space.

(Other ways to provide structure might alsobe compatible with EITM, of course.) I nowoutline the origins of the doctrinal-politics ap-proach to clarify ideas and relationships acrossliteratures.

Intellectual Origins of theDoctrinal-Politics ApproachLegal formalism, formal legalism, and em-pirical connections. The prioritization of le-gal rules and doctrinal structures flows natu-rally from both the legal literature and politicalscience scholarship. Indeed, the formal incar-nation of this approach in the case-space modelis, when stripped of formalism, standard fare inthe first year of legal education and basic text-books on legal reasoning (see Levi 1949). Thedifference from legal formalism is that rulesare not logically derived “right answers” butrather capture the preferences of judges whocan disagree. The case-space model (see Ap-pendix) originated in the work of Kornhauser(1992a,b, 1995) but is only now gaining realtraction in political science work on courts. Ihave discussed roadblocks to progress from thepolitical science side, but a symmetric problemexisted from the other side. Traditional legalscholarship and interpretive theory ignored thecollegiality of appellate courts (Kornhauser &Sager 1986, Vermuele 2005). The case-spaceframework arose in part to make room for col-legiality in legal theory while still recognizinglegal practices.

The case-space model is an alternative to theparadigmatic policy-space model from politicalscience and economics. The case-space modelis tailored to capture the substance and institu-tional features of judicial policy making, puttingcases and doctrine at the analytic center withoutrejecting a role for judicial preferences. Unlike apolicy-space model, it can easily represent facts,cases, legal rules, and dispositions. Kornhauserspecified the idea of a doctrine as a functionthat establishes equivalence classes of cases tobe decided similarly on the basis of the factualdimensions that make up a case space. He sug-gested various lines of inquiry, some of whichare now receiving greater attention. Cameron(1993) and Grofman (1993) articulated a spa-tial structure within Kornhauser’s general ap-proach and raised further questions. Cameronet al. (2000) formalized a compliance game be-tween a higher court and a lower court in a


Ann

u. R

ev. P

olit.

Sci

. 201

1.14

:131

-157

. Dow

nloa

ded

from

ww

w.a

nnua

lrevi

ews.o

rgby

Col

umbi

a U

nive

rsity

on

05/1

7/11

. For

per

sona

l use

onl

y.


unidimensional case-space model. Lax (2003,2007a) and Landa & Lax (2009) extended dif-ferent types of case-space modeling to study thehigh court as a collegial multi-member court,to explore differences over preferred rules. Lax(2007a,b) and Landa & Lax (2009) also returnedto the multidimensional analysis of case spacestarted by Kornhauser (1992b) and Cameron(1993).

Taking legal rules seriously does not requirejettisoning notions of ideological preferencesand their dominance over Supreme Courtvoting. The case-space approach is not in-compatible with the attitudinal model. Indeed,though not called such, the attitudinal modelis cast in a one-dimensional case space (Segal& Spaeth 2002, p. 90). Each judge has acut-point in this space, so that judges vote“yes” in cases that lie to their left and “no”in cases that lie to their right (or vice versa).The location of each case depends on its facts.The generalization of this one-dimensionalcase space is a multidimensional case-spacemodel, not a standard policy-space model.The doctrinal-politics approach differs fromatttitudinalism in that it focuses more on thelarger policy battles over partitioning the casespace rather than only on case dispositions.

Turning to the legal literature, a newwave of scholarship focuses on how SupremeCourt justices achieve their goals using rulestructures. This work casts justices as pickingthose structures while looking ahead to the ap-plication thereof (Fallon 2001, Friedman 2005,Shapiro 2006, Heytens 2008). As Fallon putsit, the judicial task is that of implementation. Ajudge seeking to implement her preferred setof case outcomes must choose from a variedjudicial toolbox, containing rules, formulas,and tests, which themselves break down intobright-line rules, standards, balancing tests,multiple-prong tests, etc. Shapiro focuses onthe need to communicate these structuresaccurately though actual exemplar case dispo-sitions (see Bueno de Mesquita & Stephenson2002 for a similar argument using a formalmodel of precedent). The case-space model is anatural formalization of these ideas and allows

a fruitful overlap between formal and informalwork on doctrinal choice. See, for example,case-space models of doctrinal choice between“rules” and “standards” ( Jacobi & Tiller 2007,Lax 2007b) and the analysis of other rulestructures (Lax 2007a, Landa & Lax 2009).

One final thread to this intellectual accountflows from the empirical analysis of judicialdecisions in the form of fact-pattern analysis.This empirical approach comports nicely withcase-space theorizing. Building on the founda-tion established by Kort (1957), Segal (1984)and other scholars have demonstrated that thepresence or absence of certain case facts cansuccessfully predict judicial decisions, and therelationship holds both across issue areas andacross courts. This is an empirical invocationof qualitative fact-pattern analysis as practicedin law school. Since Segal’s path-breaking workon Fourth Amendment cases [updated and ex-tended by Segal & Spaeth (1993, 2002) andSonger et al. (1994)], fact-pattern analyses havebeen applied to several other areas of the law.1

In fact-pattern analysis, the court or justice voteis regressed (using logistic regression or pro-bit given the dichotomous dependent variable)on a set of case facts. The resulting coefficientscan be thought of as weights on the variousfacts, measuring how much the fact in ques-tion “pushes” or “pulls” a particular case to-ward receiving a conservative or liberal clas-sification. Cameron et al. (2000) showed howto use the weights from such an approach toform an aggregate case-location score. This ap-proach assumes that the court’s preferred ruleis a balancing test, a weighted function of factsinto a one-dimensional space, creating an ag-gregate score that is compared to a cut-point inthat space. More complicated functions of theseinputs could be assumed in a regression, andjustices could disagree about those functions.One could also connect fact-pattern modelingto multidimensional case-space modeling. An-other approach is that of Kastellec (2010), who

1Kastellec & Lax (2008) provide a list. Friedman (2006) ar-gues that some work in this vein misunderstands the law–factdistinction.

138 Lax

Ann

u. R

ev. P

olit.

Sci

. 201

1.14

:131

-157

. Dow

nloa

ded

from

ww

w.a

nnua

lrevi

ews.o

rgby

Col

umbi

a U

nive

rsity

on

05/1

7/11

. For

per

sona

l use

onl

y.


uses classification trees to estimate the structureof the relationship between facts in their jointproduction of a disposition.

In a highly influential paper, Richards &Kritzer (2002) adapted traditional empiricalfact-pattern analysis into a “jurisprudentialregimes test” to detect the influence of law.This again comports nicely with case-space the-orizing. A jurisprudential regime is a structurefor case decisions in a given area. Such regimesare said to influence the justices themselves, sowe can compare voting patterns before and af-ter key precedents and test for changes in thefact weights. [Lax & Rader (2010a) show, how-ever, that typically employed statistical meth-ods do not reliably test doctrinal change of thisform, and conclude that there is little to noevidence of regime change; see also Segal &Spaeth 2003, Kritzer & Richards 2010, Lax &Rader 2010b.]

The strategic account. I now explain thedoctrinal-politics approach as a natural exten-sion of the strategic account. Pritchett’s (1948)seminal work on the Roosevelt Court, whichbrought legal realism to political science and fo-cused our attention on ideological preferences,and Murphy’s (1964) Elements of Judicial Strat-egy, the first work to fully embrace the strategicaccount, highlighted interactions betweenjustices in the pursuit of legal policy. Theyprovided stylized stories to support the centralthesis but did not systematically analyze strate-gic behavior (Epstein & Knight 1998, p. 629).So far, there was no particular structure inmind for what legal policy was. Jumping aheadto Epstein & Knight (1998), the goal was not toprovide a specific structure but rather to docu-ment “empirical support for the plausibility ofthe assumption of strategic interaction” and forthe frequency and potential impact of such in-teractions (Epstein & Knight 2000, p. 640). Thenext step was to systematically analyze theseinteractions, work spearheaded by Maltzman,Spriggs, and Wahlbeck (Maltzman et al. 2000).These scholars developed state-of-the-art em-pirical models with variables creatively drawnfrom intuitive hypotheses about strategic in-

centives and behavior, producing an impressiveset of data, hypotheses, and tests. The next stepafter that would have been to aggregate thesefindings into a bottom line. But that is wherethe enterprise got stuck, despite the impressivenumber of statistical findings and wealth ofnew data, so that the whole was not necessarilymuch greater than the sum of the parts.

To show what I mean, I break the strategicaccount into a hierarchy of claims, from least tomost controversial.

! S1: The justices care about more than justthe final votes they cast.

Segal & Spaeth (2002, p. 357) concede thatthe “opinion of the Court . . . constitutes thecore of the Court’s policymaking process.” AsEpstein & Knight (1998, p. 50) put it, “Whywould justices bother recrafting their opinionsif they believed that their writings would be ig-nored, if they thought their opinions would lackthe force of law?”

! S2: The justices have preferences over le-gal policy and take each other’s positionsand preferences into account when craft-ing opinions and responding.

Because a judicial opinion in a Supreme Courtcase (not just the disposition) must get majoritysupport to be directly binding on lower courts,justices will join opinions with which they dis-agree in part and make accommodations toother justices when writing opinions. They maycast only sincere final “merits votes” (as in theattitudinal model), but they write or join opin-ions insincerely, making sophisticated/strategicchoices to get to a majority. (Note that attitudi-nalists do not assume the justices are so myopicthat they fail to take strategic constraints intoaccount—rather, they argue these constraintsare effectively negligible when it comes to thefinal vote in each case.)

Most of the empirical evidence for thestrategic account concerns only propositions S1and S2 and aims at research questions that le-gal scholars will not see as important, all relatedto the meta-question, “Are justices strategic?”Indeed, Friedman (2006) argues that too mucheffort in the strategic account seems aimed


Ann

u. R

ev. P

olit.

Sci

. 201

1.14

:131

-157

. Dow

nloa

ded

from

ww

w.a

nnua

lrevi

ews.o

rgby

Col

umbi

a U

nive

rsity

on

05/1

7/11

. For

per

sona

l use

onl

y.


solely at the hegemony of the attitudinalistsand their focus on case dispositions, or speaksagainst a legal model in which the justices areseen as having clean hands, purely devoted towhat they believe to be legally required: “For[strategists], . . . ‘strategic’ often means only thatthe justices are paying attention to one an-other’s arguments . . . . Outside the subfield oflaw and courts, however, does anyone reason-ably doubt that the justices take account of oneanother’s views?” Friedman would likely alsoargue that no one would doubt that justices usetheir opinions to make new law and care aboutthe content thereof (as in Knight 2009, p. 1532).

So far, it is not clear why any of the choicesin this “collegial game” (Maltzman et al. 2000)matter. For example, why does it matter inwhich order justices join the majority opinion,how long it takes them to do so, when sugges-tions get incorporated, etc.? What exactly is itthey are bargaining over and how does it affectlaw? Epstein & Knight (1998) use to great ad-vantage the example of Craig v. Boren (1976),in which Justice Brennan pushed an interme-diate scrutiny standard for gender discrimina-tion claims, despite sincerely preferring a strictscrutiny standard. This is indeed bargainingover a legal rule, as compatible with case-spacemodeling. But this example does not mean thatsuch bargaining did more than achieve the rulepreferred by the median justice, that Brennanhad to compromise in how far he could pull law.A naive legalist might be shocked by this; mostlegal scholars would not.

What is implicit, but generally unexplainedor undocumented—and what I take most strate-gists to believe—is actually a third claim:

! S3: The collegial game has actual conse-quences for legal policy.

Maltzman et al. (2000) make clever use ofan exchange between justices about the caseof Pennsylvania v. Muniz (1990). JusticeBrennan wrote an opinion that restrictedMiranda rights, privately explaining in a writ-ten memo to Justice Marshall that he was beingstrategic: “If Sandra [ Justice O’Connor] hadgotten her hands on the issue, who knows what

would have been left of Miranda?” So the argu-ment is not just that the justices have to com-promise and work with each other, acceptingopinion language they do not like, but that theirinteractions, reactions, and anticipatory ac-tions matter—different preferences and choiceswould play out differently, and not necessarilystraightforwardly. For the Muniz anecdote tospeak against the Friedman (2006) critique, itmust be that if Brennan had not maneuveredas he did, then the exception carved out ofMiranda would have been larger. To speak incase-space language, it must be that he man-aged to keep the cut-point partitioning the casespace from lying at O’Connor’s preferred par-titioning point (assuming she was the medianjustice in a single dimension). If not, then thestrategic account could be accused of explainingonly the path to the median voter result.

Finally, Muniz might also serve as an ex-ample of a still more controversial possibility,since Justice Brennan actually voted to restrictMiranda:

! S4: Justices might cast insincere meritsvotes if it means trading the case dis-position for better policy content in theopinion.

This claim is the only strategic-model claimthat directly attacks the core of the attitudinalmodel.

Why did the strategic account produce solittle evidence at the level of S3? Epstein &Knight (2000, p. 641) recognize the contribu-tions of the empiricists “translating the strate-gic intuition into variables that they include intheir statistical models of judicial behavior” andof the formal theorists seeking to explain ob-served behavior in the form of equilibria analyt-ically derived from explicitly defined incentives.Much was learned even though each strand de-veloped independently. But some argue thereremained too wide a disconnect between theoryand empirics, despite the urges of the EITMmovement (Empirical Implications of Theo-retical Models). Indeed, Cameron (2010) ar-gues that too much of judicial empiricism is notEITM, but “MUSH” (Multiple Un-Structured

140 Lax

Ann

u. R

ev. P

olit.

Sci

. 201

1.14

:131

-157

. Dow

nloa

ded

from

ww

w.a

nnua

lrevi

ews.o

rgby

Col

umbi

a U

nive

rsity

on

05/1

7/11

. For

per

sona

l use

onl

y.


Hypotheses)—that the empirics invoking infor-mal intuitions are not robust enough to estab-lish clear “stylized facts,” nor are the deduc-tions that lead to the hypotheses tight enoughto stand as theory in their own right. MUSHyworks are those that do not advance theory, ei-ther by challenging theorists with interestingunavoidable facts or by making tight deductionsthemselves.

It is difficult to explore judicial interactionswithout thinking more about the structure ofchoice. For example, for Maltzman et al. (2000),what does it mean for a justice closer to the jus-tice authoring the opinion to have “more” in-fluence? Does it move “policy” closer to himand away from the author, in a spatial sense?Why would an opinion author be more willingto move policy toward a proximate justice? Andhow does this affect how other justices respond?Do not these shifts move policy toward or awayfrom the other justices too? Are such accom-modations merely idiosyncratic or cosmetic?In the standard one-dimensional median-votermodel, it is hard to understand these strategicmoves. The most basic challenge to the strate-gic account is, why does it matter who controlsthe majority opinion in the Supreme Court? It isnot obvious that it should: If the Median VoterTheorem applies, then opinion authorship andassignment are irrelevant. If one potential ma-jority author would find it necessary to makeconcessions to build a majority, why would notany other justice have to make the same conces-sions? To put it informally, if the price of form-ing a majority coalition is always the same, whydoes it matter who writes the check? Unless thebargaining process leaves some “surplus” forthe opinion author to capture, it does not mat-ter who writes the majority opinion nor who as-signs it. Much of the strategic literature assumesthe existence of such a surplus but does not ex-plicitly tie it to the strategic-account analysisthat ensues. Subsequent analysis proceeds fromthat point, so it is not tied to any specific ex-planation for this “median justice puzzle.” Lax& Cameron (2001) challenged scholars to solvethis puzzle, and a cottage industry (discussedbelow) has arisen around this very problem.

In the absence of the critical microfounda-tions that might explain the scope of bargainingpower, empirical explanations have invokeda rather mixed bag of motivations (from“egalitarian impulses” to “fostering good will”to “organizational needs”; see Lax & Cameron2007) and have failed to specify sequentiallyrational behavior over the course of bargaining,as if the justices could not think ahead. Forexample, Maltzman et al. (2000) analyze thevarious stages of majority-opinion draftingseparately, chapter by chapter, from studyingopinion assignment to dealing with responsesto the majority-opinion draft. This is chrono-logically correct, but the “wrong” order fromthe perspective of strategic analysis. Opinionassignment is studied before the followingbargaining process is studied, responses tothe initial majority opinion are studied beforeaccommodations to such responses are studied.Empirically, the data at each stage are treatedas though they were not the result of a compli-cated strategic screening process, even thoughthat has been established in the precedingchapter. Theoretically, at each stage, whatwe know about future choices is not invoked.Rather, what are tested are plausible incentivesfor the choices an author or other justice mightmake, instead of the patterns we would observegiven the interaction of these incentives.This tests latent unobservable incentives orstrategies instead of observable equilibriumoutcomes. Cameron (2010) points out that it isnot clear which strategic-account hypothesescould be supported by a rational model andwhether they could be compatibly supportedwithin a single model. This is perhaps why Ep-stein & Knight (2000, p. 642) argued that theinformal-intuitions-as-empirical-hypothesesapproach might best be interpreted as a “tran-sitional bridge between traditional behavioralresearch and strategic analysis.”

The case-space model has moved scholar-ship forward by giving a formal structure to theelements of judicial choice as articulated so wellby the strategists, while also opening up a lineof inquiry about legal doctrine of interest tolegal scholars. At the same time, it can speak to


Ann

u. R

ev. P

olit.

Sci

. 201

1.14

:131

-157

. Dow

nloa

ded

from

ww

w.a

nnua

lrevi

ews.o

rgby

Col

umbi

a U

nive

rsity

on

05/1

7/11

. For

per

sona

l use

onl

y.


existing empirical work (see, e.g., Lax &Cameron 2007). It can attack questions suchas: What exactly is being bargained over?What form do legal preferences and legalpolicies take? How do political preferencesover legal rules interact to shape legal policy?How can legal rules be aggregated? What arethe normative implications of legal doctrineformed by policy seekers?

COLLEGIAL POLITICSIn the remainder of this article, I analyze theliterature on collegial and hierarchical courtpolitics, focusing on case-space models andwork that speaks to similar issues.2 I first con-sider collegial adjudication, then bargainingmodels, and finally empirical evidence.

Collegial Rule Applicationand AggregationWhat consequences does collegiality have foradjudication? The seminal work of Kornhauser& Sager (1986) evaluates how collegiality in-teracts with common issues of legal and inter-pretive theory. The authors compare differentmodes of aggregating judicial choices and sub-choices and articulate a number of dimensionson which collegial adjudication can be assessed,such as consistency and coherence. Too littleof this work has made its way into the politicalscience literature on courts, I suspect becauseit is assumed that coherence is simply a nor-mative goal. However, the ability of a collegialcourt to speak in one articulate voice may affectthe court’s efficacy within the judicial hierar-chy; incoherence might endanger communica-tion with and the management of lower courts.

Easterbrook (1982) argues that collegialcourts could cycle over a list of rule choices(given Condorcet’s Paradox and multidimen-

2I regretfully must “ignore” an important subject of thestrategic account, namely battles between branches of gov-ernment. For a review of formal models of adjudication em-bedded in a constitutional system, see Kornhauser (1999).See Clark (2011) for a recent book-length treatment, withsome of the best evidence to date of such external constraintson judicial choice.

sional Chaos Theorems for policy space). Whatthen does a case-space approach say about con-sistency, that is, about a collegial court mak-ing similar decisions across cases and over time?In any single yes-or-no case heard by a colle-gial court, it would seem that the judges canvote to resolve their conflicts, but even a singlecase can present challenges for collegial adju-dication. Suppose judges are applying a legalrule to a single case, but they disagree as towhether different parts of that rule are satisfiedin that given case (i.e., they disagree as to thelegal findings). Suppose this rule involves mul-tiple findings that in some logical relationship(a conjunctive or disjunctive rule) induce a par-ticular case disposition (e.g., a judge will decide“guilty” if and only if both intent and causa-tion are established). Then a Doctrinal Paradox(Kornhauser & Sager 1986) can arise. The de-cision in the case can vary depending on howthe findings are aggregated: (a) taking majorityvotes on each finding one by one (e.g., first onintent, then on causation) and putting those ma-joritarian findings together or (b) majority votesby judge, with each judge following the legalfindings as she sees them to vote on the aggre-gate finding (guilty or not). Kornhauser (1992b)discusses actual Supreme Court cases in whichthe justices’ findings produce this paradox. Thisparadox has inspired a growing body of litera-ture on collegial rule application and judgmentapplication, spanning legal theory, social choicetheory, and deliberative democratic theory [seeliterature reviews by List (2003) and Landa &Lax (2008, 2009)].

The Doctrinal Paradox means that colle-gial adjudication need not be consistent acrossidentical cases, if the judges aggregate theirviews differently at different points. This aside,Kornhauser & Sager (1986) assert collegialcourts would always reach the same decisionin identical cases, thus yielding consistency.Kornhauser (1992a) extends this to arguethat stare decisis and what he denotes a“result-bound” judicial system will result inpath-dependent but consistent law. Stearns(2000) adds that standing doctrine can induceconsistency.

142 Lax

Ann

u. R

ev. P

olit.

Sci

. 201

1.14

:131

-157

. Dow

nloa

ded

from

ww

w.a

nnua

lrevi

ews.o

rgby

Col

umbi

a U

nive

rsity

on

05/1

7/11

. For

per

sona

l use

onl

y.


Kornhauser & Sager (1986) argue, however,that consistency is not enough. Besides makingconsistent decisions in identical cases, a form oflogical or reasoned consistency across cases isnecessary for a larger body of law to be “coher-ent.” Each judge might have a coherent view,but judges acting collegially might not yield acoherent body of decisions. One concern is thatdifferent judges might prefer different rules.Even if they agree on legal findings, appellatejudges can disagree about which legal findingsmatter and how much. That is, besides the chal-lenge of collegial rule application, they also facea potentially larger challenge, that of collegialrule creation. A newer literature discusses thequestions that arise. If a judge decides cases ac-cording to his or her preferred legal rule, whencan judges sitting on a collegial court come to-gether to create a meaningful legal doctrine?What will the “collegial rule” be?

The starting inquiry is whether there is ameaningful way to construct a collegial rule thatreflects differences among individual judges butstill captures their preferences in a majoritar-ian fashion. Can a higher collegial court give alower court a rule to follow that mimics what thehigher court would do? Kornhauser & Sager(1986) call this requirement “fit” (as part ofwhat they call “representation”; see also discus-sions of “authenticity” and “reliability”). Thekey result of Lax (2007a) is that in case space thejudges can always find a representative rule, alegal rule that captures the effect of their votingcase by case. The de facto rule induced by the setof individually majoritarian case dispositions iscalled the “implicit collegial rule,” or “implicitmedian rule” if the case space is “proper.” TheMedian Rule Theorem states the main point:There will always exist a median rule even with-out a median judge.

However, this de facto rule can differ fromindividual legal rules in that it might not bepossible for the Supreme Court (or other col-legial court) acting as a whole to form the sametype of rule as any individual justice might have.One feature of coherence that can be main-tained under aggregation is properness (a typeof monotonicity), but the implicit collegial rule

might lack properties fulfilled by the justices’individual rules. Lax (2007a) and Landa & Lax(2008, 2009) show when and how this prob-lem can arise. First, the implicit collegial ruleneed not be the rule of any one justice on theCourt. This means that, although philosophicalor other principles may indeed be found to sup-port this amalgamated product, there is no rea-son to believe these will exist; at the very least,the justices may have to go outside their col-lective set of such principles, and the resultingrule loses the presumption of principled justifi-cation that we might associate with the opinionsof justices taken as individuals [as Kornhauser &Sager (1986) suggested]. Moreover, the implicitcollegial rule can be more structurally complexthan any individual rule, indeed quite differentin structure from the individual rules that in-duce it. Another finding is that the preferredrule of the collegial court as a whole can besensitive to the individual rules of many differ-ent justices, each pivotal for different ranges ofcases, even if they are in dissent in the particularcase heard by the Court (Lax 2007a). Replac-ing any of these justices, and not just a singleso-called swing justice, can affect legal policy innontrivial ways. Focusing only on which justiceis the swing voter in a particular case will thenomit much of interest in the collegial formationof legal doctrine.

Kornhauser (1992b) offers the followingstructure for a case space. Doctrine breaks legaldecisions into dichotomous questions. Facts de-termine whether a given issue is satisfied or not;each issue is a bundle of facts that go together.A form of action details and organizes whichfacts are relevant for which issues. Causes ofactions are legal characterizations of issues, andlegal findings must support all these issues for avictory on the cause of action as a whole (a con-junctive test). That is, to win a case, you needwin only one cause of action, but can win anyone (a disjunctive test among causes of action).Causes of action can overlap in terms of whichissues are invoked. A doctrine is a set of one ormore causes of action for a given area of thelaw. Judges can disagree about doctrine, whichis to say they can disagree about the facts of the


Ann

u. R

ev. P

olit.

Sci

. 201

1.14

:131

-157

. Dow

nloa

ded

from

ww

w.a

nnua

lrevi

ews.o

rgby

Col

umbi

a U

nive

rsity

on

05/1

7/11

. For

per

sona

l use

onl

y.


case (though appellate courts do not generallyreview case facts de novo), findings (which canproduce the Doctrinal Paradox), which findingsshould be necessary, and which causes of actionshould be permitted for a given type of legalclaim. (For a typology of disagreements on col-legial courts in policy space and case space, andthe literature thereon, see Landa & Lax 2008.)

In Kornhauser’s terminology, Landa & Lax(2009) study properties of aggregating doc-trines across judges, where each judge’s doc-trine consists of one or more causes of action,but all judges agree as to facts and findings.The implicit collegial rule can require a morecomplicated relationship between the issues in-volved in causes of action and can increase ordecrease the number of causes or action rel-ative to the individually preferred doctrines ofthe judges. In a continuous case space, relativelystraightforward partitions setting a thresholdon each dimension can aggregate into rules thatcannot be represented by simple thresholds oneach dimension (Lax 2007a).

Another form of inconsistency can arise be-cause there are different algorithms by whichjudges might form a collegial rule. The im-plicit collegial rule is equivalent to case-by-casevoting, but a collegial rule can also be con-structed by considering each element (or factoror dimension). Landa & Lax (2009) show thatusing the implicit collegial rule may not satisfymajority will on particular elements of the rule.Even where a majority of judges think an is-sue relevant, the collegial rule might not treatit as relevant. Aggregating individual preferredcase dispositions may lead to different outcomesthan disposing of cases with the rule that is ar-rived at by aggregating individual dimensionsof the preferred rules; and deliberating over acase outcome can yield a different result thandeliberating over the rule first and then apply-ing it to the case. This implies a different type ofdoctrinal paradox, at the level of rule creationinstead of rule application. Landa & Lax (2009)provide necessary and sufficient conditions forthis to occur and show connections to the stan-dard Doctrinal Paradox. Lax (2007a) shows asimilar result where judges have thresholds for

each dimension, disagreeing on the thresholdfor the legal findings on these dimensions.

A final form of inconsistency can arise evenfor a single judge considering connected cases.Most recent work assumes preferences overcases are separable: Everything that is conse-quential for deciding a case is a function of therule and the facts in that case alone, not howother cases are decided. If cases are nonsepara-ble, then a judge may face tension between theinstant case before him and past case disposi-tions. An example of nonseparable preferencesmight be Justice Stewart’s vote in Eisenstadt v.Baird (1972), to extend the right to possess con-traceptives to unmarried people, given his dis-sent in Griswold v. Connecticut (1965), in whichthe Supreme Court held that married coupleshad such a right. Because he could not overturnGriswold, he perhaps preferred to grant the rightto all. Since this means that justices might castfinal merits votes insincerely (with respect tothe single case at hand), this represents a chal-lenge to the attitudinal model. This may also ex-plain why justices sometimes implicitly go alongwith precedents they do not like (see Spaeth &Segal 1999). See Kornhauser (1992a) on pathdependence and separability.

Collegial Rule GamesWhat explains what doctrine the SupremeCourt will choose? Recall that any claim thatcollegial bargaining matters implicitly raises—and preferably explicitly resolves—the “medianjustice puzzle” (how a bargaining window re-mains open given the Median Voter Theorem).Some models below are explicitly in case space,others in policy space. Models also vary in howthey capture the Supreme Court’s bargainingprotocol and how much case-space technologythey invoke.

Monopoly theories. In one family of models,only one justice within the majority has any in-fluence over opinion content, so the opinion isplaced precisely at her ideal point. There arethree contending monopolists: the median jus-tice, the majority-median justice, and the au-thor. The “open-bidding” model of Hammond

144 Lax

Ann

u. R

ev. P

olit.

Sci

. 201

1.14

:131

-157

. Dow

nloa

ded

from

ww

w.a

nnua

lrevi

ews.o

rgby

Col

umbi

a U

nive

rsity

on

05/1

7/11

. For

per

sona

l use

onl

y.


et al. (2005) is the normal Median Voter The-orem model. A variant with the same bottomline is their “median-holdout” model, which as-sumes the median justice refuses to vote forany opinion other than her own ideal point.The majority-median hypothesis is discussed byWesterland (2003) and can arise in the formalmodel of Carrubba et al. (2007), when no jus-tice in the initial majority will accept an offerfrom the opposing side (doing so is either ruledout by assumption or prohibitively unattrac-tive). This yields the Median Voter Theoremapplied within the initial majority coalition. Fi-nally, the author himself might have total con-trol over opinion location. No formal model,but some empirical work, suggests this.

A variation on the median result, wherepolicy reduces to the median rule, but notnecessarily the rule of a median justice, isgiven by Lax (2007a), who considers multi-dimensional case-space bargaining games. Inthe “case-by-case” game and the “opinioncoalition” game (and under some conditionsin the “explicit collegial rule” game), theimplicit collegial rule results. Anderson &Tahk (2007) use a policy-space model to arguethat a dimension-by-dimension median will bestable, assuming that each section of a Courtopinion deals with movement only on one pol-icy dimension and that justices can join sectionsone by one. This establishes, for the judicialsetting, well-known results from social choicetheory about dimension-by-dimension voting.

Influence theories. In another family of the-ories, the opinion author is constrained to takeinto account the preferences of the other jus-tices but still maintains some influence. To startwith non-case-space work (but noting that itmight be possible to reinterpret such work incase space), for Schwartz (1992), the policy al-ternatives available to the opinion author areexogenously fixed, and the author can only con-trol the level of precedent written into the ma-jority opinion. Authorship influences the levelof precedent, with preferences over precedentlevels given by assumption. Maltzman et al.(2000) do not make an explicit prediction as to

where the opinion will fall, but its location ispositively related to author ideology. They ar-gue that opinion assignors will, all else equal,prefer more proximate opinion assignees. ForLax (2001), dissent retards policy change andthis creates a “take-it-or-weaken-it” bargain-ing game, as compared to the normal “take-it-or-leave-it” game. The Court majority is con-strained in cases in which dissent would weakenpolicy change (e.g., when the power of theCourt is threatened, such as in the desegre-gation cases). Finally, in the “agenda-control”variant of Hammond et al. (2005), it is assumedthat only the initial majority opinion author canwrite an opinion. The other justices can onlyaccept that opinion or let policy revert to thestatus quo. (Hammond et al. note it is not clearwhy justices would cede so much power to theopinion writer.) This is then the familiar take-it-or-leave-it bargaining model, in which theauthor’s choice set is the range of points be-tween the status quo and the reflection point ofthe status quo on the other side of the median.

The role of the status quo in this modelis also controversial (Brenner & Whitmeyer2010). What does it mean for the justices todecline to join the majority opinion, but in-stead “choose the status quo”? As Cameron &Kornhauser (2008) point out, this resemblesthe bargaining protocol of a legislature, not theSupreme Court. Legislators vote a single bill upor down against the status quo, perhaps after anamendment process, perhaps with no alterna-tives permitted. The Supreme Court has a verydifferent protocol for bargaining. The justicescan offer alternative opinions, join the major-ity, concur in part, vote with the majority butconcur separately, or dissent. There need notbe a single alternative to the status quo, and thefate of a particular alternative does not rest ona single up or down vote. All this is done in thecontext of a particular case to be decided.

The agenda-control model has the Courtoperate under closed rule, not open rule, in thejargon of legislatures—only one alternative ispermitted on the floor, which is voted up ordown. If there is a status quo legal rule in place,a justice could certainly write an opinion in


Ann

u. R

ev. P

olit.

Sci

. 201

1.14

:131

-157

. Dow

nloa

ded

from

ww

w.a

nnua

lrevi

ews.o

rgby

Col

umbi

a U

nive

rsity

on

05/1

7/11

. For

per

sona

l use

onl

y.


favor of it. But she is not restricted to chooseonly that rule or the one offered by the initialmajority opinion author. And if no one writesto offer this status quo legal rule, there is nomechanism by which the justices may invokeit, other than to get the case dismissed. Evenwhen affirming the disposition from the lowercourts, the justices are free to write their opin-ions as they wish, announcing a new or mod-ified legal rule. Once the Court takes a case,barring dismissal, doctrine never simply revertsto the status quo. (The status quo would seemfar more relevant to case selection; the justiceswould compare the status quo to the expectedvalue of taking the case and having a bargaininggame.)

Lax & Cameron (2001, 2007) present thefirst explicitly case-space model of bargaining.Opinions target a particular partition of theone-dimensional case space, but they have asecond dimension, legal quality, which deter-mines how close the enacted case partitioncomes to the desired case partition (one couldmore simply assume that the justices care di-rectly about quality). Producing higher-qualityopinions is costly, and this creates a wedge thatthe assignee can exploit. The Chief Justice (orother assignor) anticipates the bargaining gameand strategically assigns opinion authorship.The degree of author power, opinion location,and opinion quality vary with bargainingleverage and various model parameters, andtherefore so does assignment. The opinion’scut-point will be located between the medianand the author. In contrast to the proximityprediction of other models, the optimal opinionassignee is one more extreme than the assignor:Given the moderating pull of the median,and given the additional pressure to invest inquality, this produces an opinion closer to theassignee than would assigning to his ideologicaltwin. The model generates a large numberof comparative statics for bargaining andassignment, explaining well-known empiricalregularities previously without microfounda-tion, as well as generating novel predictions.

Carrubba et al. (2007) were the first to ex-plicitly model justices who care both about the

instant case disposition and opinion content.In their model, which I call CFMV (after theauthors’ names), justices can join the majorityopinion or write separately, trading extra valuefrom the majority opinion against expressivepayoffs based on what opinion a justice joins.Other justices can join (adding authority to themajority opinion) or write separately to betterexpress their preferences over policy. Justicesprefer to endorse proximal opinions associatedwith the “correct” case disposition. The modelof Cameron & Kornhauser (2008), or CK, issimilar in many respects. These papers go be-yond opinion location to predict coalition andconcurrence patterns in addition to the case dis-position. Carubba et al. solve their model forthe case in which correct dispositions are val-ued so highly that bargaining is restricted tothe initial majority, setting aside the other casefor the most part. CFMV specifically modelsthe added value from having an official major-ity opinion rather than just a plurality opinionfor the majority disposition. It treats productionof the majority opinion as a public good and in-vokes specific assumptions to make predictionsto avoid free-riding problems and the like. CKalso recognizes that an authoritative majorityopinion is a public good, as is gathering enoughvotes to get the “correct” disposition, but CKresolves this by assuming that only the major-ity author cares about authoritativeness. Legalpractice suggests a big difference between a ma-jority opinion (which counts as precedent for alllower courts) and a plurality opinion, but this isset aside in CK. CK focuses more on strategicdispositional voting than does CFMV.

I next turn to empirical work on opinion “lo-cation,” work that attempts to test the predic-tions produced by the theoretical work above.

Empirical Evidenceof Opinion LocationEpstein & Knight (1998) and Maltzman et al.(2000) present clear if circumstantial evidencethat nontrivial bargaining is occurring (i.e.,something more than median monopoly).Bonneau et al. (2007, pp. 896–97) conclude

146 Lax

Ann

u. R

ev. P

olit.

Sci

. 201

1.14

:131

-157

. Dow

nloa

ded

from

ww

w.a

nnua

lrevi

ews.o

rgby

Col

umbi

a U

nive

rsity

on

05/1

7/11

. For

per

sona

l use

onl

y.


that their agenda-control model better explainsfinal vote data than does a median-monopolymodel. The assumption driving their test isthat the status quo lies in between the groupof justices who voted to grant certiorari andthe group who voted to deny it. If this meansthe status quo tends to fall near the medianjustice already, then the bargaining range issmall to negligible. But more importantly, itis unclear why the justices would want to takecases in which current legal policy is already soclose to the median on the Court. Lax & Rader(2008) present evidence against the monopolytheories (opinion assignment affects changes inmerits votes, which it would not if assignmentwere irrelevant), as well as evidence againstthe proximity-assignment hypothesis, and infavor of the extreme-assignment hypothesisunique to Lax & Cameron (2007). Carrubbaet al. (2007) use concurrence patterns to arguethat opinions do not lie at the median justice.They show that the author, the median, andthe majority-median all affect location, rulingout a pure median-monopoly model (see alsoWesterland 2003). Clark & Lauderdale (2010)develop a scaling model to estimate opinionlocation and ideal points on a common scale,using citations. They present evidence againstthe median- and author-monopoly models, infavor of models that predict location within theinitial majority coalition (the majority-medianmodel outperforms the other monopoly mod-els). Finally, Beim et al. (2010) show, using theClark-Lauderdale scores, that opinion contentresponds to both opinion and dispositioncoalitions—evidence against monopolist mod-els. They also present evidence of “cross-overjoins;” justices do not always prioritize thedisposition over the policy.

HIERARCHICAL POLITICSI first argue that the hierarchical division of ad-judication (combined with a discretionary highcourt docket) raises challenges for empiricalwork on other issues, such as the role of law andideology. I then discuss theoretical and empir-ical work on doctrine making in a hierarchy. I

focus on what is called the agency approach tostudying judicial hierarchy, which highlightsdisagreements between and within levelsthereof. For the team approach, which focuseson hierarchical error rates and error correctionin models of common legal preferences, see Ko-rnhauser (1995, 1999), Cameron & Kornhauser(2006), and Westerland et al. (2010).

The Empirical Implicationsof HierarchyThe Supreme Court and the lower federalcourts play distinct roles. The top tier of the ju-dicial hierarchy, having a discretionary docket,concentrates far more on doctrine—rule cre-ation and articulation—while the bottom tierconcentrates on application of rules to specificcases. This hierarchical division of labor is of-ten downplayed in both empirical and theoret-ical work, which instead operates as though theSupreme Court handles “normal” cases. Workon case selection (e.g., Ulmer 1984, Caldeira& Wright 1988, Perry 1991, Boucher & Segal1995) is set aside.

The implicit picture in much empirical workon the Court, and in some formal models,is that of routine law application, or simplecase sorting, in which the Court can correctmis-sorting by lower courts. But the SupremeCourt is the Court of Last Resort, not Last Re-Sort. Most Supreme Court cases are unusual ornovel in some way, or they are taken becauseenough justices want to change the law. To besure, they take some cases to rectify noncom-pliance or mistakes by lower courts. But theydo not expend most of their docket on routinecase sorting. Indeed, Shapiro (2006) argues theCourt does not do enough of it, so that thereare too few examples of what the high courtwants.

One problem is that we rely on observedSupreme Court cases to study the preferencesof Supreme Court justices, constraints on theirdecision making, the treatment of case facts,the role of law, and ideological change on theCourt. We often try to ascertain the doctrinethe justices want to see applied more generally


Ann

u. R

ev. P

olit.

Sci

. 201

1.14

:131

-157

. Dow

nloa

ded

from

ww

w.a

nnua

lrevi

ews.o

rgby

Col

umbi

a U

nive

rsity

on

05/1

7/11

. For

per

sona

l use

onl

y.


in cases they do not take by inspecting statisticalpatterns in the cases they do take. But the pre-cise selection strategy employed by the justiceswill affect the set of Supreme Court cases weobserve in a given area of the law in a given timeperiod. It is now well known that the cases thatmake it to trial or appeal are a nonrandom setof cases, and so generalizations from these cases(e.g., to infer legal rules) that ignore this selec-tion bias can be highly misleading (see Priest& Klein 1984 and related work). Kastellec &Lax (2008) study selection bias in the contextof various common judicial-politics researchdesigns. They show that ignoring case selectionseverely undercuts inferences about whichcase facts matter to the Court and how much(making it harder to study changes therein, andundercutting jurisprudential regimes tests);about aggregate liberalism in the votes of theCourt and individual justices (and thereforecomparisons thereof ); and about the degree ofcompliance in the lower courts (when extrapo-lating to lower court decisions from high courtfact weights). One way around these problemsis, as per the doctrinal-politics approach, tofocus more on doctrine as described in opinionsinstead of only on dichotomous votes. Wealso might better observe the Supreme Court’sdesired legal doctrine (or jurisprudentialregime changes) by examining lower courtcases, to the extent that such cases are repre-sentative and straightforward applications oflaw.

Selection bias can create particular problemsfor nuanced uses of judicial-ideology scores,which tend to assume a static case distributionand a common case distribution across all issueareas. For this and other reasons, “the inter-pretation of scores as measures of ‘ideology’ inthe strong sense is not warranted by the data.The scores are simply a descriptive summaryof the single dimension that best character-izes differences in merits votes of the justices”(Ho & Quinn 2009, p. 27). [See also Martin& Quinn (2005), Fischman & Law (2009), andFarnsworth (2007).]

Of course, case selection is also a researchopportunity. One important qualification in the

attitudinal model is often forgotten: “Manymeritless cases undoubtedly exist that no self-respecting judge would decide solely on the ba-sis of his or her policy preferences” and “[t]hosethat the Court does decide tender plausible le-gal arguments on both sides” (Segal & Spaeth2002, p. 93). It is not that Supreme Courtjustices will do whatever they want, only thatthey will do what they want in the cases theychoose—which are those in which they can dowhat they want. This suggests another way oftesting the limits of attitudinalism and of as-sessing law as constraint: to focus on the case-selection stage. This boundary assumption maybe where the real action is in the law-ideologydivide, but it is not where enough scholarly ac-tion has been. The Segal-Spaeth position as-sumes only two types of cases, cleanly dividedinto those that allow for attitudinal decisionand those that are legally bound. A continuumof legal-boundedness seems more plausible: Ifthere are some cases that the justices would nottake to exert their attitudes because of clear law,why could they not at least partially respect lawin the cases they do take? Full respect for lawat case selection seems inconsistent with zerorespect for law in case dispositions.

Compliance as De Facto DoctrineThe Supreme Court lacks the most com-mon weapons in principal-agent problems—the power to hire and fire, to reward andpunish—but it does have the power to audit. Inone line of work, incentives for certiorari (audit-ing of the lower courts) and compliance (by thelower court with the higher court’s preferreddoctrine) come together to produce a de factopartition of the case space. Doctrine as appliedwill fall short of the preferred partition due tononcompliance.

According to Cameron et al. (2000), non-compliance arises because the lower courts ex-ploit ambiguity in case facts. The higher courtstrategically and probabilistically audits basedon the lower court’s ideology, observable casefacts, and the lower court decision. A conser-vative (liberal) higher court reviews only liberal

148 Lax

Ann

u. R

ev. P

olit.

Sci

. 201

1.14

:131

-157

. Dow

nloa

ded

from

ww

w.a

nnua

lrevi

ews.o

rgby

Col

umbi

a U

nive

rsity

on

05/1

7/11

. For

per

sona

l use

onl

y.


(conservative) lower courts and only when theymake liberal (conservative) decisions. Empiricalsupport for these propositions is provided.

Lax (2003) shows that collegial court pol-itics and hierarchical politics intersect. The“rule of four,” a nonmajoritarian rule, createsa more credible threat to review cases, whichincreases lower court compliance, and so actu-ally increases majority power. What is counter-majoritarian in appearance is majoritarian in ef-fect. There are also incentives for the justices tohide their true preferences and engage in strate-gic reputation building, to limit the degreeof noncompliance by lower courts. Kastellec(2007) connects hierarchical compliance poli-tics to collegial “panel effects” within the Courtsof Appeals, analyzing formally the effects of awhistle-blower on such a panel (see Cross &Tiller 1998). Clark (2009) extends analysis ofthe type above to en banc review.

Looking DownSome work takes on doctrinal choice explic-itly (in contrast to the de facto doctrine gamesabove). The potential for noncompliance alsohas an effect on what doctrine the higher courtmight choose in the first place. In the legalliterature, Heytens (2008) argues that the rea-sons we observe so much compliance is that theSupreme Court looks ahead when crafting doc-trine. One such inquiry invokes the well-knowndichotomy in the legal literature between rulesand standards. This distinction is defined vari-ously, but usually a “rule” is a determinate ruleelaborated in full a priori, whereas a “standard”involves some degree of indeterminism, requir-ing case-by-case consideration of less than per-fectly concrete factors. The puzzle is why theCourt chooses standards in some areas of thelaw or at some points in time but not others,and why some justices prefer rules while oth-ers prefer standards, often inconsistently acrossdifferent areas of the law.

Jacobi & Tiller (2007) model the choicebetween a determinate rule and an indetermi-nate standard, which depends on the contentof each, the degree of conflict between higher

and lower courts, and bias toward some typesof litigants. The rule and standard chosenbetween are both fixed by assumption, sothe rule is a simplified partition rather thanthe exact partition desired, and the standardis represented as a region of the space leftuncontrolled by the doctrine. The areas ofthe regions in question are given exogenously.Lax (2007b) models the incentives driving thechoice between determinate doctrines in theform of bright-line rules and more flexible orindeterminate doctrines (standards). A standarddiffers by incorporating a factual dimensionthat lacks full transparency or objectivity.The rule and the standard choices are definedendogenously (in terms of content, structure,and legal quality). The trade-off betweenthe optimal rule and the optimal standard isaffected by ideological conflict across the levelsof the judicial hierarchy, judicial expertise, issuecomplexity, issue salience, and the sensitivityof the desired doctrine to varying case facts.Lax’s model then makes predictions for thedegree of transparency, indeterminacy, pre-cision, doctrinal complexity, and lower courtdiscretion, which all emerge endogenously.

In another line of attack, Staton &Vanberg (2008) offer a policy-space model inwhich justices are strategically vague in theiropinions to build institutional strength andprestige, manage limited resources, and deferto those with informational advantages, at theexpense of some policy control. Specificity canincrease compliance by making noncompliancemore visible to external monitors such as thepublic, but can risk laying bare judicial weak-ness if other actors will still not comply. Theirmodel identifies incentives new to the analysisof hierarchy and delegation. In both their modeland Lax’s (2007b), relationships between legalclarity and political factors are nonmonotonic.

Finally, McNollgast (1995, p. 1641) de-scribes a policy-space model of how the Courtmight choose doctrine to maximize compli-ance. Doctrine is a “statement by the SupremeCourt about the range of lower court decisionsthat it finds acceptable.” The Supreme Courtmight induce lower courts to comply by


Ann

u. R

ev. P

olit.

Sci

. 201

1.14

:131

-157

. Dow

nloa

ded

from

ww

w.a

nnua

lrevi

ews.o

rgby

Col

umbi

a U

nive

rsity

on

05/1

7/11

. For

per

sona

l use

onl

y.


granting some permission to be noncompliant,which isolates the remaining lower courtsfor review. The higher court cannot tell thelocation of a lower court decision until it takesa case for review, but it can tell whether thelower court has been compliant. Also, auditingof noncompliant courts is random (subject toa budget constraint).

Looking UpRather than study rules flowing downward fromthe top of the judicial hierarchy, two recent pa-pers investigate rule development percolatingup from below, with the higher court auditingrule selection. This should be the beginning ofan exciting new branch of work on judicial hi-erarchy. Clark & Kastellec (2010) model theSupreme Court’s decision to step in to resolveconflict among the lower courts as an optimalstopping problem in which the Court trades offlearning more about what rule it would preferagainst the costs of allowing conflict to persist.In the case-space model of Carrubba & Clark(2010), the lower court declares a rule and de-cides an instant case. The instant case and hier-archical conflict affect rule creation and audit-ing of the lower court.

CONTRIBUTIONS OFTHE DOCTRINALPOLITICS APPROACHThis section highlights some key works, find-ings, and contributions within the doctrinal-politics approach, with some special attentionto those using case-space models in particular.(The list is not meant to be comprehensive, andI regret any oversights.)

Knight (2009, pp. 1554, 1538) commentsthat “the case-space . . . approach has real ana-lytical promise, but claims that the frameworkhas important methodological implications forempirical studies have yet to be supported” and“there have not been any serious efforts to trans-late the results of the case-space analyses intoan empirically meaningful research agenda.”Although I respect these concerns, I read therecord differently. I think that case-space work

now not only directly raises empirical impli-cations but has engaged such implications. Ialso think that the general approach highlightsgaps in current empirical knowledge, and insome cases calls into question well-establishedfacts. Finally, it seems unlikely that many ques-tions currently being studied with case-spacemodels would have arisen in policy-space ap-proaches, or that a policy-space approach wouldgive more leverage. To be sure, Knight is cor-rect that more needs to be done connectingtheory and data. Empiricists need models thatspeak to the richness of their findings; theoristsneed data on rules and doctrines and changestherein, and not just votes.

A list of key works, findings, andcontributions:

! The incentives for certiorari and lowercourt compliance intersect, leading topredictions for auditing probabilities anddoctrinal enforcement, predictions sup-ported by the data (including the “Nixongoes to China” finding that a conserva-tive Supreme Court does not audit a lib-eral lower court that takes a conservativeaction). These patterns would likely nothave been uncovered without a case-spacemodel (Cameron et al. 2000).

! Clark (2009) presents and tests a theoryof en banc review.

! The case-space model of Kastellec (2007)delineates the conditions under which apotential whistleblower on a lower courtpanel increases compliance with the highcourt.

! The rule of four for selecting cases inthe Supreme Court increases lower courtcompliance, increasing majoritarian con-trol over the lower courts, as shown by acase-space model (Lax 2003).

! Collegiality complicates adjudication andrule creation. Collegial adjudication cre-ates paradoxical complexities even in asingle legal case (Kornhauser & Sager1986). This finding led to an impor-tant literature on this dilemma from bothpositive and normative perspectives (e.g.,Kornhauser & Sager 1993, List & Pettit

150 Lax

Ann

u. R

ev. P

olit.

Sci

. 201

1.14

:131

-157

. Dow

nloa

ded

from

ww

w.a

nnua

lrevi

ews.o

rgby

Col

umbi

a U

nive

rsity

on

05/1

7/11

. For

per

sona

l use

onl

y.


2002, List 2003). Collegial court outputcan depend sensitively on court member-ship, not just on a single swing/medianjustice. Collegial rules can be quite dif-ferent from individually preferred legalrules. Collegiality can force choices be-tween coherence and representativeness(Kornhauser & Sager 1986; Lax 2007a;Landa & Lax 2008, 2009).

! Case-space models of collegial courtbargaining raise and present possibleresolutions to a key puzzle for students ofthe Supreme Court, the median justicepuzzle: Why, given the seeming appli-cability of the Median Voter Theorem,does bargaining not simply devolve to themedian justice, making all the strategicbargaining and opinion assignment westudy irrelevant? There are multiplereasons (e.g., Carrubba et al. 2007, Lax &Cameron 2007, Cameron & Kornhauser2008). These case-space models invokethe actual bargaining protocol of theCourt and give a more specific accountof such bargaining than a generalizedstrategic approach. The connectionbetween the instant case and generalrule simply cannot be studied with astandard policy-space model. Traditionalstrategic-account literature argued thatthe justices of the Court may make “in-sincere” policy choices even if they alwayscast sincere dispositional votes. The case-space literature shows why, how, andwhen this will occur (from hierarchicalinfluences to collegial influences).

! Even in a complex case space and evenwithout a median justice, there will alwaysexist a median rule (Lax 2007a, Landa& Lax 2009). Collegiality alone, then,cannot explain why strategic interactionsaffect policy (as some work seems tosuggest).

! The previous two points together raisea theoretical challenge. The median-ruleresult implies that any other bargaining-induced rule the Supreme Court an-nounces would yield case dispositions

different from those reached if the Courtheard all cases itself. If a lower court thenfaithfully applied the bargained rule, thehigher court would have an incentive toreverse some applications of its own rule.A lower court could even evade the de jurerule and follow the de facto median rule,and a majority of justices might supportthat act of noncompliance. All of the an-swers to the median justice puzzle (howcan bargaining matter if all bargainingcomes down to the median voter?) thencreate a “median rule puzzle”: How canbargained rules that differ from the me-dian rule be enforced, since they will notcapture what the Court majority wants ina set of cases?

! Friedman (2006, p. 267) argues that “byfocusing on votes rather than opinions,real differences in judicial ideology areobscured.” Lax (2007a,b) and Lax &Landa (2008, 2009) unpack these dif-ferences. Such case-space work confirmsthat judicial preferences can interact inways that a standard policy-space ap-proach does not allow us to study andexplores more complicated connectionsbetween ideology and doctrinal prefer-ences than can be understood withoutcase space.

! Data analysis separating opinion coali-tions from disposition coalitions using acase-space perspective finds novel impli-cations of bargaining theory for judicialreplacements, highlighting replacementand peer effects across natural courts(Cameron et al. 2009, Beim et al. 2010).

! Carrubba & Clark (2010) construct andtest a case-space model of rule creation inthe judicial hierarchy from the oppositeperspective, with lower court rules per-colating upward. The instant case beforethe court affects the legal rule adopted, afinding impossible in policy space.

! Case-space models have also been in-voked to study legal development in theform of incremental partitioning overtime. Examples of this include work


Ann

u. R

ev. P

olit.

Sci

. 201

1.14

:131

-157

. Dow

nloa

ded

from

ww

w.a

nnua

lrevi

ews.o

rgby

Col

umbi

a U

nive

rsity

on

05/1

7/11

. For

per

sona

l use

onl

y.


on stare decisis and path dependence byKornhauser (1989, 1992a); analyses of in-cremental partitioning (Cameron 1993,Gennaioli & Shleifer 2007) and theKornhauser (2008) critique of the latter;and Landa & Lax’s (2009) discussion ofincremental partitioning in the context ofproblems of rule aggregation.

! Selection bias is not a unique concern ofthe case-space approach, but the particu-lar bite of it is highlighted by case-spacepriorities (see discussions in Kastellec &Lax 2008, Lax & Rader 2010a).

! Fischman (2011) estimates a structuralmodel (based on a case-space model) ofjudicial panels operating under a norm ofconsensus, showing that there is a highrate of strategic voting, that dissent ratesunderstate disagreement within panels,and that consensus voting obscures theimpact of ideology. This work also gener-ates new estimates of ideology parametersfor individual judges.

CONCLUSIONMuch of the progress made in the past decadeof judicial politics scholarship has been made byreconciling legal concerns and political sciencepriorities in a new judicial politics of legal doc-trine. This doctrinal-politics approach high-lights a relatively new formal apparatus knownas the case-space model, and it invokes close tiesbetween theoretical and empirical work and be-tween the study of judicial behavior and actuallegal practices and institutions. Taking law se-riously and taking adjudication seriously mayboth be required for good modeling of judicialchoice. But taking judicial politics seriously maybe just as important for understanding law andadjudication.

The case-space model allows for ideologi-cal differences between judges while recogniz-ing that these differences will be expressed interms of legal rules that partition fact-filled le-gal cases into different dispositions. Battles overthese partitions are the heart of judicial politics.

APPENDIX: THE CASE-SPACEMODELThe most fundamental unit of judicial policymaking is the disposition of a legal case, whichis presented as a bundle of facts, discovered andrevealed through legal processes such as tri-als. The more general unit of judicial policymaking is a legal rule governing case disposi-tions. Rules, cases, and case facts are core legalconcepts—yet political models of judicial policymaking, formal and otherwise, often pay littleattention to them. When early judicial-politicsscholars separated themselves from the legalacademy to join up with mainstream politicalscience, they ported over a positive political the-ory apparatus (the now standard spatial policymodel) from the study of legislatures. But courtsare not legislatures, and judges are not legisla-tors. Cameron & Kornhauser (2008, p. 1) arguethat further progress requires more attention tothe “institutional features that actually distin-guish courts—especially collegial courts—fromlegislatures.”

Legislatures announce general statutes.Courts take up specific cases even when theyannounce more general rules, yet cases do notexist in the standard spatial modeling apparatus.In using a policy-space model to study courts,does a case raise simply an issue area, creatingan opportunity to make policy for that issue?Is a policy point the outcome within a specificcase? If so, what makes one case different fromany other? What is a legal rule? If it is the policypoint, then what is the case disposition, and howdo rules and dispositions relate? The standardmodeling apparatus cannot readily speak to themost striking feature of appellate court policymaking: the joint production of a case dispo-sition and a general rule. A case-space modelcan.

Case space and policy space share a geo-metric setup with the space containing pointsdefined along various dimensions (a one-dimensional line, a two-dimensional plane,etc.). However, these models differ in whatthese points represent, and therefore in howthey view the structure of preferences and

152 Lax

Ann

u. R

ev. P

olit.

Sci

. 201

1.14

:131

-157

. Dow

nloa

ded

from

ww

w.a

nnua

lrevi

ews.o

rgby

Col

umbi

a U

nive

rsity

on

05/1

7/11

. For

per

sona

l use

onl

y.


choice (see Figure 1). In a policy space, eachpoint is an alternative policy, and a judge typi-cally has a preferred policy point, wanting pol-icy to be as close as possible to this point. Thespace can be unidimensional, a line from liberalto conservative (or across some other quantityor quality), or multidimensional, with eachdimension capturing some aspect of the policyin question. Although on the surface a casespace looks similar to a standard policy space,it differs in the assumptions made about thestructure of choice. Each point represents, nota policy, but a specific legal case. Given a case, acourt chooses a disposition for it. Typically, thisdisposition is a dichotomous judgment for oneside or the other, a yes (Y) or a no (N). Judicialpolicy making is the mapping of these points(cases) to dispositions. The judicial choice isnot “Which point shall I pick?” but rather“Which disposition shall I choose for this givenpoint?”

A case is an exogenously fixed point in thecase space, capturing its location on each fac-tual dimension. There can be one or more fac-tual dimensions, depending on what the judgesthink relevant for the issue in question. Thesedimensions could be, for example, the locationin which a search and seizure took place, thedegree to which a law potentially violating theestablishment clause creates an entanglementwith religion, or something as simple as the ve-locity of a car when pulled over for speeding.Courts then make policy by sorting cases di-chotomously into winners and losers on the ba-sis of the facts in those cases. An evidentiarysearch is admissible or inadmissible. A law vio-lates the establishment clause or it does not. Adriver is speeding or she is not.

In one dimension, in a proper continuouscase space (see Lax 2007a), a rule is itself acut-point dividing Ys and Ns—and may seem

similar to an ideal point in a standard policyspace. In a multidimensional case space, how-ever, a proper rule can no longer be just a pointitself, as it generally takes more than a pointto partition cases into Ys and Ns. To dividepoints in a two-dimensional case space (such asFigure 1), we would need a (one-dimensional)plane curve, such as a line. And so on. In mul-tiple dimensions, differences between the twomodeling apparatuses become more apparent,but even a one-dimensional case space focusesone’s attention differently than does a one-dimensional policy space. [See also Kornhauser(2008) and Landa & Lax (2008), showing thatcase space and policy space are not equivalenteven in one dimension.] Besides dimension-ality, another variation is a dichotomized casespace (each case takes a value of 0 or 1 oneach factual dimension, e.g., Landa & Lax2009) as compared to a continuous case space(with continuous values on each dimensionas in Figure 1). Most policy-space models ofcourts are unidimensional or restrict a seconddimension, to avoid well-known cycling and“chaos” results. In case space, dimensionalityneed not have the same effect; we can explorethe structure of legal rules in more detail.

Although applied case-space modeling is onthe rise, the approach is still relatively new, andthere is still much to be done on the math-ematics of case space and relations to pol-icy space (see Kornhauser 2008). Kornhauser(1999, p. 52) discusses another way to modeldoctrine, a two-dimensional policy space ofpolicy versus precedent/deference (Ferejohn &Shipan 1990, Gely & Spiller 1990, Schwartz1992, Cohen & Spitzer 1994), but argues it“remains inherently political and nonlegal; itmakes no reference to the facts of a case orfeatures of legal discourse that appear in anopinion.”

DISCLOSURE STATEMENTThe author is not aware of any affiliations, memberships, funding, or financial holdings that mightbe perceived as affecting the objectivity of this review.


Ann

u. R

ev. P

olit.

Sci

. 201

1.14

:131

-157

. Dow

nloa

ded

from

ww

w.a

nnua

lrevi

ews.o

rgby

Col

umbi

a U

nive

rsity

on

05/1

7/11

. For

per

sona

l use

onl

y.


ACKNOWLEDGMENTSI thank Deborah Beim, Charles Cameron, Tom Clark, Lucy Goodhart, John Kastellec, LewisKornhauser, Kelly Rader, and Jeff Staton for generous comments and criticisms. I also want toacknowledge the undergraduate and graduate classes in judicial politics that I have taught, as thediscussions in those settings helped me realize what I wanted to say in this article. So did my onethird of a law degree, which, for better or for worse, changed how I think about some of the issuesherein.

LITERATURE CITED

Anderson R, Tahk A. 2007. Institutions and equilibrium in the United States Supreme Court. Am. Polit. Sci.Rev. 101(4):811–25

Bailey MA, Maltzman F. 2008. Does legal doctrine matter? Unpacking law and policy preferences on the U.S.Supreme Court. Am. Polit. Sci. Rev. 102:369–84

Beim D, Cameron CM, Kornhauser L. 2010. Policy and dispositional coalitions on the Supreme Court of the UnitedStates. Presented at Annu. Meet. Am. Polit. Sci. Assoc., Washington, DC

Bonneau CW, Hammond TH, Maltzman F, Wahlbeck PJ. 2007. Agenda control, the median justice, and themajority opinion on the U.S. Supreme Court. Am. J. Polit. Sci. 51(4):890–905

Boucher RLJ, Segal JA. 1995. Supreme Court justices as strategic decision makers: aggressive grants anddefensive denials on the Vinson court. J. Polit. 57:824–37

Brenner S, Whitmeyer J. 2010. Status quo models of Supreme Court decision making: a theoretical evaluation.Presented at Annu. Meet. Am. Polit. Sci. Assoc., Washington, DC

Bueno de Mesquita E, Stephenson M. 2002. Informative precedent and intrajudicial communication. Am.Polit. Sci. Rev. 96(4):1–12

Caldeira GA, Wright JR. 1988. Organized interests and agenda setting in the U.S. Supreme Court. Am. Polit.Sci. Rev. 82:1109–27

Cameron CM. 1993. New avenues for modeling judicial politics. Presented at Conf. Polit. Econ. Public Law,Rochester, NY

Cameron CM. 2010. Comments on “precedent and the implications of stare decisis.” Discussant presentation atAnnu. Meet. Midwest Polit. Sci. Assoc., Chicago, IL

Cameron CM, Kornhauser LA. 2006. Appeals mechanisms, litigant selection, and the structure of judicialhierarchies. In Institutional Games and the U.S. Supreme Court, ed. JR Rogers, RB Flemming, JR Bond,pp. 173–204. Charlottesville: Univ. Virginia Press

Cameron CM, Kornhauser LA. 2008. Adjudication equilibria. Presented at Annu. Meet. Am. Polit. Sci. Assoc.,Boston, MA (current version 2010)

Cameron CM, Park JK, Beim D. 2009. Shaping Supreme Court policy through appointments: the impact ofa new justice. Minn. Law Rev. 93:1820–70

Cameron CM, Segal JA, Songer DR. 2000. Strategic auditing in a political hierarchy: an informational modelof the Supreme Court’s certiorari decisions. Am. Polit. Sci. Rev. 94:101–16

Carrubba CJ, Clark TS. 2010. Rule creation in a political hierarchy. Presented at Annu. Meet. Midwest Polit.Sci. Assoc., Chicago, IL

Carrubba CJ, Friedman B, Martin A, Vanberg G. 2007. Does the median justice control the content of SupremeCourt opinions? Presented at Conf. Empir. Legal Stud., New York, NY

Clark T. 2009. A principal-agent model of en banc review. J. Law Econ. Organ. 25:55–79Clark TS. 2011. The Limits of Judicial Independence. Cambridge, UK: Cambridge Univ. PressClark TS, Kastellec JP. 2010. The Supreme Court and percolation in the lower courts: an optimal stopping model.

Presented at Annu. Meet. Am. Polit. Sci. Assoc., Washington, DCClark TS, Lauderdale B. 2010. Locating Supreme Court opinions in “doctrine space.” Am. J. Polit. 54(4):871–

90Cohen L, Spitzer M. 1994. Solving the Chevron puzzle. Law Contemp. Probl. 57:65Cross FB, Tiller EH. 1998. Judicial partisanship and obedience to legal doctrine: whistleblowing on the federal

courts of appeals. Yale Law J. 107:2155–76

154 Lax

Ann

u. R

ev. P

olit.

Sci

. 201

1.14

:131

-157

. Dow

nloa

ded

from

ww

w.a

nnua

lrevi

ews.o

rgby

Col

umbi

a U

nive

rsity

on

05/1

7/11

. For

per

sona

l use

onl

y.


Easterbrook FH. 1982. Ways of criticizing the court. Harvard Law Rev. 95:802–32Epstein L, Jacobi T. 2010. The strategic analysis of judicial decisions. Annu. Rev. Law Soc. Sci. 6:1–18Epstein L, Knight J. 1998. The Choices Justices Make. Washington, DC: C.Q. PressEpstein L, Knight J. 2000. Toward a strategic revolution in judicial politics: a look back, a look ahead. Polit.

Res. Q. 53:625–61Fallon RH. 2001. Implementing the Constitution. Cambridge, MA: Harvard Univ. PressFarnsworth W. 2007. The use and limits of Martin-Quinn scores to assess Supreme Court justices, with special

attention to the problem of ideological drift. Northwest. Univ. Law Rev. Colloquy 101:143–54Ferejohn J, Shipan C. 1990. Congressional influence on bureaucracy. J. Law Econ. Organ. 6:1–20Fischman JB. 2011. Estimating preferences of appellate judges: a model of “consensus voting.” J. Law Econ.

In pressFischman JB, Law DS. 2009. What is judicial ideology, and how should we measure it? J. Law Policy 29:133Friedman B. 2005. The politics of judicial review. Texas Law Rev. 84:257–337Friedman B. 2006. Taking law seriously. Perspect. Polit. 4:261–76Gely R, Spiller PT. 1990. A rational choice theory of Supreme Court statutory decisions with applications to

the State Farm and Grove City cases. J. Law Econ. Organ. 6:263–300Gennaioli N, Shleifer A. 2007. The evolution of common law. J. Polit. Econ. 115(1):43–68Grofman B. 1993. Public choice, civic republicanism, and American politics: perspectives of a “reasonable

choice” modeler. Texas Law Rev. 71:1541–87Hammond TH, Bonneau CW, Sheehan RS. 2005. Strategic Behavior and Policy Choice on the U.S. Supreme

Court. Palo Alto, CA: Stanford Univ. PressHeytens TJ. 2008. Doctrine formulation and distrust. Notre Dame Law Rev. 83:2045Ho DE, Quinn KM. 2009. How not to lie with judicial votes: misconceptions, measurement, and models. Presented

at Conf. Building Theory Through Empir. Legal Stud., Univ. Calif., BerkeleyJacobi T, Tiller EH. 2007. Legal doctrine and political control. J. Law Econ. Organ. 23(2):326–45Kastellec JP. 2007. Panel composition and judicial compliance on the United States courts of appeals. J. Law

Econ. Organ. 23(2):421–41Kastellec JP. 2010. The statistical analysis of judicial decisions and legal rules with classification trees. J. Empir.

Legal Stud. 7(2):202–30Kastellec JP, Lax JR. 2008. Case selection and the study of judicial politics. J. Empir. Legal Stud. 53:407–46Knight J. 2009. Are empiricists asking the right questions about judicial decisionmaking? Duke Law J. 58:1531–

56Knight J, Epstein L. 1996. The norm of stare decisis. Am. J. Polit. Sci. 40:1018–35Kornhauser L, Sager L. 1993. The one and the many: adjudication in collegial courts. Calif. Law Rev. 91:1–51Kornhauser LA. 1989. An economic perspective on stare decisis. Chic. Kent Law Rev. 65:63–92Kornhauser LA. 1992a. Modeling collegial courts. I: Path dependence. Int. Rev. Law Econ. 12:169–85Kornhauser LA. 1992b. Modeling collegial courts. II: Legal doctrine. J. Law Econ. Organ. 8:441–70Kornhauser LA. 1995. Adjudication by a resource-constrained team: hierarchy and precedent in a judicial

system. South. Calif. Law Rev. 68:1605–29Kornhauser LA. 1999. Appeal and supreme courts. Encyclopedia Law Econ. 7200:45–62. http://encyclo.

findlaw.com/7200book.pdfKornhauser LA. 2008. Modeling courts. In Theoretical Foundations of Law and Economics, ed. MD White,

pp. 1–20. Cambridge, UK: Cambridge Univ. PressKornhauser LA, Sager LG. 1986. Unpacking the court. Yale Law J. 96:82–117Kort F. 1957. Predicting Supreme Court decisions mathematically: a quantitative analysis of the “right to

counsel” cases. Am. Polit. Sci. Rev. 51:1–12Kritzer HM, Richards MJ. 2010. Taking and testing jurisprudential regimes seriously: a response to Lax and

Rader. J. Polit. 72:285–88Landa D, Lax JR. 2008. Disagreements on collegial courts: a case-space approach. J. Constitutional Law

10(2):305–29Landa D, Lax JR. 2009. Legal doctrine in collegial courts. J. Polit. 71(3):946–63Lax JR. 2001. The conflict of consensus: unanimity and dissent in the Supreme Court. Presented at Annu. Meet.

Midwest. Polit. Sci. Assoc., Chicago, IL


Ann

u. R

ev. P

olit.

Sci

. 201

1.14

:131

-157

. Dow

nloa

ded

from

ww

w.a

nnua

lrevi

ews.o

rgby

Col

umbi

a U

nive

rsity

on

05/1

7/11

. For

per

sona

l use

onl

y.


Lax JR. 2003. Certiorari and compliance in the judicial hierarchy: discretion, reputation, and the rule of four.J. Theor. Polit. 15:61–86

Lax JR. 2007a. Constructing legal rules on appellate courts. Am. Polit. Sci. Rev. 101:591–604Lax JR. 2007b. Political constraints on legal doctrine: how hierarchy shapes the law. Presented at Annu. Meet.

Midwest. Polit. Sci. Assoc., Chicago, IL (current version 2010)Lax JR, Cameron CM. 2001. Opinion assignment in the Supreme Court: theory and evidence. Presented at Annu.

Meet. Am. Polit. Sci. Assoc., San Francisco, CALax JR, Cameron CM. 2007. Bargaining and opinion assignment on the U.S. Supreme Court. J. Law Econ.

Organ. 23:276–302Lax JR, Rader KT. 2008. Bargaining power in the Supreme Court. Presented at Conf. Empir. Legal Stud.,

Cornell Univ. School of Law, Ithaca, NYLax JR, Rader KT. 2010a. Legal constraints on Supreme Court decision making: Do jurisprudential regimes

exist? J. Polit. 71(2):73–84Lax JR, Rader KT. 2010b. Rejoinder: The three prongs of a jurisprudential regimes test. J. Polit. 71(2):89–91Levi E. 1949. An Introduction to Legal Reasoning. Chicago: Univ. Chicago PressList C. 2003. A possibility theorem on decisions over multiple propositions. Math. Soc. Sci. 45:1–13List C, Pettit P. 2002. Aggregating sets of judgments: an impossibility result. Econ. Philos. 18:89–110Maltzman F, Spriggs JF, Wahlbeck PJ. 2000. Crafting Law on the Supreme Court: The Collegial Game.

Cambridge, UK: Cambridge Univ. PressMartin AD, Quinn KM. 2005. Can ideal point estimates be used as explanatory variables? Work. pap., Weidenbaum

Cent., Washington Univ.McNollgast. 1995. Politics and the courts: a positive theory of judicial doctrine and the rule of law. South.

Calif. Law Rev. 68:1631–89Murphy WF. 1964. Elements of Judicial Strategy. Chicago: Univ. Chicago PressPerry H Jr. 1991. Deciding to Decide: Agenda Setting in the United States Supreme Court. Cambridge, MA:

Harvard Univ. PressPriest GL, Klein B. 1984. The selection of disputes for litigation. J. Legal Stud. 13:1–55Pritchett CH. 1948. The Roosevelt Court. New York: MacmillanRichards MJ, Kritzer HM. 2002. Jurisprudential regimes in Supreme Court decision making. Am. Polit. Sci.

Rev. 96:305–20Schwartz EP. 1992. Policy, precedent, and power: a positive theory of Supreme Court decision-making.

J. Law Econ. Organ. 8:219–52Segal JA. 1984. Predicting Supreme Court cases probabilistically: the search and seizure cases, 1962–1981.

Am. Polit. Sci. Rev. 78:891–900Segal JA, Spaeth HJ. 1993. The Supreme Court and the Attitudinal Model. New York: Cambridge Univ. PressSegal JA, Spaeth HJ. 2002. The Supreme Court and the Attitudinal Model Revisited. New York: Cambridge Univ.

PressSegal JA, Spaeth HJ. 2003. Reply to the critics of The Supreme Court and the Attitudinal Model Revisited. Law

Courts 13(3):31–38Shapiro C. 2006. The limits of the Olympian court: common law judging versus error correction in the

Supreme Court. Wash. Lee Law Rev. 63:271–337Shepsle KA. 1989. Studying institutions: some lessons from the rational choice approach. J. Theor. Polit.

1:131–47Songer DR, Segal JA, Cameron CM. 1994. The hierarchy of justice: testing a principal-agent model of Supreme

Court–circuit court interactions. Am. J. Polit. Sci. 38:673–96Spaeth HJ, Segal JA. 1999. Majority Rule or Minority Will: Adherence to Precedent on the U.S. Supreme Court.

Cambridge, UK: Cambridge Univ. PressStaton JK, Vanberg G. 2008. The value of vagueness: delegation, defiance, and judicial opinions. Am. J. Polit.

Sci. 52(3):504–19Stearns ML. 2000. Constitutional Process: A Social Choice Analysis of Supreme Court Decision Making. Ann Arbor:

Univ. Mich. PressTiller EH, Cross FB. 2006. What is legal doctrine? Northwest. Univ. Law Rev. 100(1):517–34

156 Lax

Ann

u. R

ev. P

olit.

Sci

. 201

1.14

:131

-157

. Dow

nloa

ded

from

ww

w.a

nnua

lrevi

ews.o

rgby

Col

umbi

a U

nive

rsity

on

05/1

7/11

. For

per

sona

l use

onl

y.


Ulmer SS. 1984. The Supreme Court’s certiorari decisions: conflict as a predictive variable. Am. Polit. Sci. Rev.78:901–911

Vermuele A. 2005. The judiciary is a they, not an it: interpretive theory and the fallacy of division. J. Contemp.Legal Issues 14:549–84

Westerland C. 2003. Who owns the majority opinion? An examination of policymaking on the U.S. Supreme Court.Presented at Annu. Meet. Am. Polit. Sci. Assoc.

Westerland C, Segal JA, Epstein L, Cameron C, Comparato S. 2010. Strategic defiance and compliance inthe U.S. courts of appeals. Am. J. Polit. Sci. 54(4):891–905


Ann

u. R

ev. P

olit.

Sci

. 201

1.14

:131

-157

. Dow

nloa

ded

from

ww

w.a

nnua

lrevi

ews.o

rgby

Col

umbi

a U

nive

rsity

on

05/1

7/11

. For

per

sona

l use

onl

y.

PL14-frontmatter ARI 11 April 2011 17:22

Annual Review ofPolitical Science

Volume 14, 2011Contents

A Life in Political ScienceSidney Verba ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! i

Leadership: What It Means, What It Does, and What We Want toKnow About ItJohn S. Ahlquist and Margaret Levi ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! 1

Examining the Electoral Connection Across TimeJamie L. Carson and Jeffery A. Jenkins ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! !25

Presidential Appointments and PersonnelDavid E. Lewis ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! !47

Understanding the 2007–2008 Global Financial Crisis: Lessons forScholars of International Political EconomyEric Helleiner ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! !67

Presidential Power in WarWilliam G. Howell ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! !89

The Politics of Regulation: From New Institutionalism to NewGovernanceChristopher Carrigan and Cary Coglianese ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! 107

The New Judicial Politics of Legal DoctrineJeffrey R. Lax ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! 131

The Rhetoric Revival in Political TheoryBryan Garsten ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! 159

The Rhetoric of the Economy and the PolityDeirdre Nansen McCloskey ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! 181

The Contribution of Behavioral Economics to Political ScienceRick K. Wilson ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! 201

The Causes of Nuclear Weapons ProliferationScott D. Sagan ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! 225

Network Analysis and Political ScienceMichael D. Ward, Katherine Stovel, and Audrey Sacks ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! 245

v

Ann

u. R

ev. P

olit.

Sci

. 201

1.14

:131

-157

. Dow

nloa

ded

from

ww

w.a

nnua

lrevi

ews.o

rgby

Col

umbi

a U

nive

rsity

on

05/1

7/11

. For

per

sona

l use

onl

y.

PL14-frontmatter ARI 11 April 2011 17:22

The Big Five Personality Traits in the Political ArenaAlan S. Gerber, Gregory A. Huber, David Doherty, and Conor M. Dowling ! ! ! ! ! ! ! ! ! ! 265

ClientelismAllen Hicken ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! 289

Political Economy Models of ElectionsTorun Dewan and Kenneth A. Shepsle ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! 311

Modeling Dynamics in Time-Series–Cross-Section PoliticalEconomy DataNathaniel Beck and Jonathan N. Katz ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! 331

Voting TechnologiesCharles Stewart III ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! 353

Indexes

Cumulative Index of Contributing Authors, Volumes 10–14 ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! 379

Cumulative Index of Chapter Titles, Volumes 10–14 ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! 381

Errata

An online log of corrections to Annual Review of Political Science articles may be foundat http://polisci.annualreviews.org/

vi Contents

Ann

u. R

ev. P

olit.

Sci

. 201

1.14

:131

-157

. Dow

nloa

ded

from

ww

w.a

nnua

lrevi

ews.o

rgby

Col

umbi

a U

nive

rsity

on

05/1

7/11

. For

per

sona

l use

onl

y.

the strategic account and the rule of law jeffrey r. lax · · 2013-05-08the strategic account...

Documents