A Methodological Critique of The Louisiana Supreme Court in Question: An Empirical and Statistical Study of the Effects of Campaign Money on the Judicial Function

• Author: Robert Newman; Janet Speyrer; Dek Terrell
• Position: Robert Newman is Professor of Economics at Louisiana State University, Baton Rouge. Janet Speyrer is Professor of Economics at the University of New Orleans. Dek Terrell is Professor of Economics at Louisiana State University, Baton Rouge
• Pages: 307-315

Page 307

I Introduction

Vernon Palmer and John Levendis offer a rather confusing and contradictory paper on the impact of campaign contributions on voting behavior by Louisiana Supreme Court justices.¹ The first two sentences provide a very good example of the paper's fundamental flaws. The first sentence asserts that "little" literature exists to guide their study, ignoring over twenty-five years of scholarly work directly related to the question of interest.² The second sentence provides the paper's central thesis-that it supplies statistical evidence that contributions have influenced justices.³ Yet a careful reading of the literature suggests that this paper contains no such evidence. In an even more puzzling twist, footnote 14 of Palmer and Levendis' (hereafter the authors are referred to as P&L) paper states that it will assert no such causal relationship.⁴ Our assessment is the authors have made a serious error in methodology and their conclusions, therefore, should not go unchallenged. Further, this methodological mistake may be compounded by serious errors made by the authors in the construction of their data. Page 308

II Prior Research On Campaign Contributions And Voting Decisions

The first step in any academic study is a careful review of the relevant literature. The first sentence of the P&L paper begins with the observation that, "[t]he effect of campaign contributions on judicial decision making has been the subject of widespread interest and debate, but little empirical research."⁵ Like much of the paper, the first sentence misses the mark. In fact, there is extensive literature in economics investigating the impact of campaign contributions on the decisions of recipients. Understanding the problems with the P&L paper requires first placing it in the context of the literature.

While most of the extant evidence comes from empirical research on the relationship between campaign contributions and decisions of legislators, the methodological issues are identical for examining the same relationship with respect to decisions made by judges. Beginning with Henry Chappell's seminal paper,⁶ the accepted approach for empirical research on this topic must explicitly recognize the probable simultaneity between the effect of campaign contributions on judicial decisions and the effect of judicial decisions on campaign contributions.⁷ His paper is of particular importance because it reveals a fatal flaw in the P&L analysis and points to the appropriate methodology the authors should have employed for estimation of this type of model. The necessity of addressing the simultaneity issue was explicitly stated in an influential study by Thomas Stratman,⁸ which states in the introduction: "All studies addressing the question of whether Page 309 campaign contributions influence congressional voting behavior must address the issue of whether campaign contributions are endogenous in the vote equation. The issue is whether contributions influence the voting behavior or whether the expected voting behavior influences contributions."⁹

This mandate applies equally to all studies addressing the question of whether campaign contributions influence the decisions of judges. Thus, both studies seriously call into question P&L's conclusion of a causal link between contributions and judicial decisions. After Chappell's work, essentially every serious work on the topic must address the fact that there are at least two relationships of interest, not one as their article implies. When there are two sets of decision-makers (judges and contributors), both must be modeled. To assume that one can ignore the decision calculus of contributors is a fundamental error in both the use of economic theory and econometrics. It is difficult to argue that contributors do not take into account judicial temperament and philosophy when deciding how to allocate their scarce dollars. And, it is naïve to assume that a judge's temperament and judicial philosophy cannot be known in advance.

Thus, any serious attempt to address the impact of contributions on the decisions of judges must use econometric techniques that recognize the two-way causality. Assuming away the simultaneity issue, as P&L implicitly have done, represents a fundamentally fatal error in their analysis. The possibility that differences among judges in their decisions or judicial philosophies can influence how campaign contributions are distributed is conceptually identical to the influence legislator predispositions have on the distribution of campaign contributions. This has been well-recognized in economics for decades. For example, Grier and Munger show that specific characteristics of individual legislators attract contributions from some, but not necessarily all interest groups.¹⁰ That is, interest groups that value certain characteristics contribute to the campaigns of legislators who possess those characteristics.

With just a cursory review of the literature, P&L would have found these studies and would have been aware of Stratman's view that all studies must address this issue as central to their analysis. Their failure to even cite these studies, much less address the key issue, reveals a fundamental flaw in their study. However, that is not the only important issue raised by the P&L paper. Many Page 310 economic studies can miss a key item in the literature or err in methodology. But, over time, subsequent research corrects the errors if the study is deemed interesting or research scholars may simply ignore the study, which implicitly deems it as having little value to the discipline.

However, the P&L paper is not the typical academic study. The methodology chosen by P&L, which entirely ignores the simultaneity issue, focuses on voting by specific justices in the Louisiana Supreme Court. The authors conclude that contributions influenced the voting behavior of at least three justices in particular and suggest that it may also be true for the entire court.¹¹ By naming specific justices and incorrectly asserting that they have produced statistically valid evidence that campaign contributions influenced decisions of the Louisiana Supreme Court, the authors risk tarnishing the reputations of longstanding judges with no scientifically valid evidence to support their claims. In this case, the profession's process of simply ignoring poor scholarship or correcting it over time cannot prevent the immediate damage to reputations that the P&L study will produce under the guise of academic research.

This Critique proceeds by first discussing the problems in P&L's methodology. We then focus on conclusions drawn by the authors and the language used to describe the results.

III Problems In The Palmer And Levendis Methodology

A The Effect of Judicial Philosophy on the Decision to Contribute

The key problem in P&L's methodology is that...