Strategic Behavior and Policy Choice on the U.S. Supreme Court

Vol. 16 No.1 (January 2006), pp.87-91

STRATEGIC BEHAVIOR AND POLICY CHOICE ON THE U.S. SUPREME COURT, by Thomas H. Hammond, Chris W. Bonneau and Reginald S. Sheehan. Stanford: Stanford University Press, 2005. 328pp. Cloth $65.00. ISBN: 0-8047-5145-5. Paper $24.95. ISBN: 0-8047-5146-3.

Reviewed by Chad Westerland, Department of Political Science, University of Arizona. Email: cwesterl [at] email.arizona.edu

Strategic alternatives to the attitudinal model have both implicitly and explicitly invoked rational choice models to explain the choices justices make. Until now, this literature has been missing a fully integrated model of the sequential decisions made by the Court. With STRATEGIC BEHAVIOR AND POLICY CHOICE ON THE U.S. SUPREME COURT, Thomas Hammond, Chris Bonneau, and Reginald Sheehan have attempted to fill this major void. The authors apply a coherent and clearly specified spatial model to each stage of the Court’s decision making process. The goal is to use the model to generate a set of behavioral predictions that distinguish between strategic and sincere decisions at each stage in the process.

The book is organized into three sections. Part I gives an initial outline of the assumptions that underlie the rational choice model that is to be applied and provides a review of the problems the authors have with the attitudinal model and, to a lesser extent, the existing invocations of strategic alternatives. Chapter 2 consists of a list of seven pairs of distinctions between different theories of decision making, which incidentally is the first of many similarly presented lists of ideas (for example, distinction 2 is itself a list of six problems with theories of attitude activation). The aim is to justify the rational choice model advanced later, since the seven distinctions serve as the fundamental assumptions for the authors’ model. The distinctions vary in importance, but as I will discuss below, a central distinction is the question of whether or not a status quo policy position must be explicitly modeled. The existence of a status quo is essential to the model advanced in the book, as it provides a reference for all decisions. In fact, the model is not possible without a well defined status quo.

Chapter 3 defends the theoretical perspective advanced in the previous chapter by critiquing the problems the authors have with the attitudinal model. The authors have four problems (and three issues) with the attitudinal model as developed by Segal and Spaeth (1993, 2002). Hammond, Bonneau, and Sheehan find Segal’s and Spaeth’s invocation of attitude-activation theories in the first version of SCAM to be inconsistent with the version advanced in SCAMR. They argue that more than the final vote needs to be explained, given that justices make multiple decisions before the vote on the merits and that the Court sets policy with written opinions. Six reasons are also given for why justices’ merit votes might reflect strategic considerations. After the determination of a need for a [*88] strategic model of decision making, the authors note that advocates of strategic supplements to the attitudinal model have failed to provide formal models of strategic behavior, which results in an inability to distinguish adequately between sincere and strategic behaviors.

Part II is the application of the model. After a superfluous chapter on the general utility of formal models (Chapter 4), the assumptions and definitions of the model are given. Some of Chapter 5 merely introduces rudimentary concepts central to any spatial model, such as ideal points and the shapes of utility functions. More importantly, several essential assumptions behind the model are presented. In particular, there exists a status quo, justices have perfect information about the preferences of other justices, cases are unidimensional and independent, justices cannot write concurring opinions, and opinion writing is costless.

I imagine there will be a fair amount of variance in how readers respond to these initial assumptions. I will briefly discuss two of my concerns. The existence of a status quo policy is to some extent an obvious assumption to make. The authors state that “[i]f a justice does not know the location of the status quo at every stage . . . he or she could scarcely make any rational decision at all” (p.85). The Court, however, specifically devotes its resources to cases in which the legal state of the world may be less than clear. To be sure, the status quo might simply be the policy location of the lower court decision under review or existing Supreme Court precedent, but then what is the status quo when there is circuit or state and federal conflict or in a case with novel issues? Further, assuming perfect information about the location of the status quo ignores a very strong rationale for why the Court’s process exists in the first place. The authors ignore, for example, Tim Johnson’s work (2001, 2004) on oral arguments. If orals are used by justices specifically to gather information about the policy implications of the case, then it is reasonable to infer that this includes determining the location of the status quo. The decision not to allow for concurring opinions is also troubling. The rationale for doing so is that the authors cannot explain why justices would do so within the context of their model (p.93). Clearly, the option to agree with the outcome but not with the location of the policy set in the majority opinion should have implications for how opinion coalitions are formed.

In presenting the model, the authors start with the Court’s final decision and work backwards. Chapter 6 applies the spatial model to the final vote and coalition formation. The model is clearly presented, and the results and implications are thoroughly examined. The model itself is straightforward. Nine justices arrayed in a unidimensional policy space align in various coalitions depending on the location of the status quo and the location of the proposed new policy. Also, three variants of the model on how justices respond to draft opinions are offered, which are really three different sets of assumptions. In the open-bidding model, justices freely and aggressively distribute responses to draft opinions, and in the median holdout version, [*89] justices will not vote until policy moves as close as possible to their respective ideal points. The result of both versions is simply the median voter theorem. Setting the policy at the median’s ideal point is always the equilibrium outcome. In the agenda control model, justices passively respond to draft opinions, meaning justices will either endorse or not endorse the opinion. If the opinion is better than the status quo for a justice, it will be endorsed. The result from this version of the model is that an opinion writer may be able to move policy away from the median and closer to her ideal point. The authors also show that in any of the three versions all coalition sizes are possible since the final coalition size will depend on the location of the status quo. At the end of Chapter 6 (and the following three chapters), the authors provide a useful and easily digestible summary of the theoretical results.

Chapter 6 is critical since forward-thinking justices will be making decisions at earlier stages based on where policy will eventually be set. If policy will be set at the median’s ideal point, then opinion assignment, conference voting, and cert decisions are not very interesting. Who cares who writes the opinion or who assigns the opinion if all of the justices know that the opinion will eventually converge on the median’s ideal point? If O’Connor is the median, then policy set forth from a coalition of Scalia, Thomas, Rehnquist, Kennedy and O’Connor is no different from the policy set forth by a coalition of Stevens, Souter, Breyer, Ginsburg, and O’Connor. This means that if the open-bidding and median holdout models are correct, then there is no need to model any of the other stages, with the possible exception of the cert decision. The authors recognize this obvious fact, and note that if either the open-bidding or the median holdout model is true, then there is no need to continue (pp.133-137). Thus, the agenda-control model must be persuasive for the rest of the book to have any appeal. The absence of any serious discussion about which of the three variants is most likely suggests the authors are agnostic, but given their own assumption of costless opinion writing, the agenda control variant is surely the least likely model. Why would policy motivated justices passively accept any opinion that is better than the status quo but worse than another majority preferred outcome, if a costless behavior can achieve that outcome? Presumably, if writing a majority opinion is costless (or even relatively cheap), then drafting alternative opinions would be easy enough. And since all justices are assumed to have perfect information about other justices’ preferences, convergence to the median would not be difficult. The agenda-control variant would have to be an endogenously created and maintained norm of deference to opinion writers, but no justification in Chapter 6 is given for why such a norm would ever be considered.

Chapter 7 presents the opinion assignment model under the agenda control variation. Although the model shows self-assignment is always best (not surprising if the opinion writer matters and writing is costless), the authors state that “the empirical literature suggests that self-assignment is often not possible” (p.142). This [*90] assertion is not explored further. However, the model shows that, if the agenda variant holds and if self-assignment is not an option, then the opinion assignment can change the Court’s policy output. This is a function primarily of the distance between the median and the status quo; the closer the median is to the status quo, the less assignment will matter because that gives the opinion writer less room to maneuver. The authors thoroughly explore the implications of the agenda control model for opinion assignment.

Chapter 8 models the conference vote. Again, in the agenda control variation, the ability of the opinion writer to pull policy away from the median induces incentives for strategic conference voting. Under this condition, the model becomes considerably more complicated. Justices have incentives to misrepresent their positions in the conference, and the Chief Justice, who should always vote with the majority in order to assign the opinion, must determine whether or not conference positions are sincere. Since the goal of the associates is to get the assignment, at least one justice would be likely to express support for the Chief Justice’s ideal point. Once again, this stage of the model is carefully developed and clearly presented. The authors conclude the chapter by noting that if the agenda model is not accurate, then “the complex logic that we have developed in this chapter will be largely irrelevant” (p.212).

The final chapter in Part II concerns certiorari. The results here are obvious: sincere justices will always vote to grant cert if their ideal point is not the status quo, but if the justices are strategic, then cert will be granted if the median’s ideal point is not exactly on the status quo. The cert model is consistent across the three opinion response variants. In light of the essential (yet uncited by the authors) work of Cameron, Segal, and Songer (2000) and Lax (2003), the cert model presented in Chapter 9 can hardly be seen as an advance.

Part III consists of a chapter on the empirical implications of the model and a chapter on directions of future research. Chapter 10 details briefly what the model ultimately means for each stage in the process. Hammond, Bonneau, and Sheehan also discuss how the models may generally inform research on strategic decision making on the Court. Chapter 11 conducts a reexamination of the essential assumptions of the model and presents a cursory discussion of how legal considerations might work within the context of the model. The authors mention in Chapter 11 that the agenda control model may be justified as a time saving device, as justices may not want to trouble themselves with converging to an equilibrium (p.266-7), but given the importance of this particular variation, this discussion is not sufficient. While many of the theoretical points are given further consideration, an opportunity is missed since neither of the concluding chapters do much to connect the theoretical results of the model to the existing literature.

In the Preface, the authors state that their book should be seen as an extended “if . . . then” argument. This point certainly [*91] must be kept in mind when working through the model. This does not mean, however, that all “ifs” are equally interesting. The absence of a strong argument for the agenda control model makes me wonder about the ultimate payoff of Hammond’s, Bonneau’s, and Sheehan’s work, since otherwise they are simply stating that the median voter theorem applies to the Supreme Court. Regardless of the state of the literature, that is not a major revelation.

REFERENCES:

Cameron, Charles M., Jeffrey A. Segal, and Donald R. Songer. 2000. “Strategic Auditing in a Political Hierarchy: An Informational Model of the Supreme Court’s Certiorari Decisions.” 94 AMERICAN POLITICAL SCIENCE REVIEW 101–116.

Johnson, Timothy R. 2001. “Information, Oral Arguments, and Supreme Court Decision Making.” 29 AMERICAN POLITICS RESEARCH 331-351.

Johnson, Timothy R. 2004. ORAL ARGUMENTS AND DECISION MAKING ON THE UNITED STATES SUPREME COURT, by Timothy R. Johnson. Albany, NY: State University of New York Press.

Lax, Jeffrey R. 2003. “Certiorari and Compliance in the Judicial Hierarchy: Discretion, Reputation and the Rule of Four.” 15 JOURNAL OF THEORETICAL POLITICS 61–86.

Segal, Jeffrey A. and Harold J Spaeth. 1993. THE SUPREME COURT AND THE ATTITUDINAL MODEL. New York: Cambridge University Press.

Segal, Jeffrey A. and Harold J Spaeth. 2002. THE SUPREME COURT AND THE ATTITUDINAL MODEL REVISITED. New York: Cambridge University Press.

*************************************************