Signatures of ideology: the case of the Supreme Court's criminal docket.

• Author: Farnsworth, Ward

TABLE OF CONTENTS I. INTRODUCTION II. RESULTS III. SOME MECHANISMS OF IDEOLOGICAL DECISION IV. IMPLICATIONS V. THE PUZZLE OF SUBSTANTIVE CRIMINAL LAW VI. CONCLUSION I. INTRODUCTION

Everyone suspects that Supreme Court justices' own views of policy play a part in their decisions, but the size and nature of the part is a matter of vague impression and frequent dispute. Do their preferences exert some pressure at the margin or are they better viewed as the mainsprings of decision? The latter claim, identified with legal realism, has been lent some support by political scientists (1) who point out that some justices regularly vote for or against certain kinds of claims (for example, under the Fourth Amendment), (2) or that votes in some areas are broadly predictable according to a single "ideal point" that tries to sum up each justice's preferences, (3) or that justices who dissent from a decision often will not acquiesce to it in future cases. (4) The reason these studies haven't made much of an impression in the legal academy probably is that lawyers and scholars sense many reasons why judges' behavior may follow predictable patterns, not all of them related to their own preferences. Some justices may have ideas about interpretation that happen to produce outcomes friendly to one side or another as byproducts; and a judge's public reputation as a "conservative" or "liberal," to which some of the political science work gives weight in explaining votes, (5) likewise might arise because the judge's interpretive approach happens to yield results that conservatives or liberals like.

A closer look is needed at judicial behavior in cases where the policy stakes are similar but the sources of interpretive dispute are different, the better to reveal which dominates which. The best set of such cases is found on the Supreme Court's criminal docket. Cases involving accused or convicted criminals raise all sorts of legal issues but can be seen to involve a common set of policy stakes: the courts have to referee disputes, often of a zero-sum character, over the advantages to be enjoyed by the government and the accused or convicted defendant. Of course one also can divide up criminal cases into narrower categories that may involve different policies; we will try it later. (6) But the hypothesis that all such cases involve similar rough trade-offs as a matter of policy is a useful starting point. It gives us many decisions to study; cases about criminals usually take up around a third of the Court's docket every term, which is enough to support interesting statistical inquiries and generalizations.

I compiled a database of all the Court's criminal cases over the past fifty years, defining them broadly to include appeals from criminal convictions, questions of search and seizure, disputes over the rules of evidence or of criminal procedure, civil rights claims brought by prisoners against their keepers, and many others: any cases where the government has been on one side with an accused or convicted defendant on the other. (7) The next step was to separate cases involving different types of interpretive issues, which I did as an initial matter by splitting them into two categories: the constitutional and the nonconstitutional. The constitutional cases typically involve defendants' claims that their rights were violated under the Fourth, Fifth, Sixth, Eighth, or Fourteenth Amendments. The nonconstitutional kind involve disputes over the federal rules of evidence and procedure, the meaning of federal criminal statutes, the availability of habeas corpus, and so forth.

2. RESULTS

   The basic results can be seen in two ways. Chart 1 consists of a pair of lines, each based on career data for all of the justices since 1953. One shows how often each justice voted for the government in nonunanimous criminal cases involving constitutional claims. The other shows the same for cases depending on some other source of law--usually a statute or rule. The justices are ordered here according to the data (that is, by the mean of the two lines) to show the alignment between the two trends.

   [ILLUSTRATION OMITTED]

   The lines are notable for both their slopes (the tremendous difference between the justices to the right and to the left) and their similarities (the closeness of the two points with respect to any given justice). A different way to express the result is by removing the justices' names from the graph and instead putting their votes in constitutional cases along the bottom and their votes in nonconstitutional cases along the side. We then have Chart 2, a scatterplot of the twenty-eight justices since 1953 correlating the proportion of their votes for the government in nonunanimous criminal cases of the two types. (8) An increase in the share of votes for the government along one of the dimensions is very likely to mean an increase along the other; a fitted line shows a strong linear relationship between ruling in favor of the government in either situation. The Pearson correlation coefficient (R) is an extremely high .97, accounting for 94% of the variance ([R.sup.2]) (signif. = 0.0000000000000002).

   [ILLUSTRATION OMITTED]

   For lawyers to whom the graphs are gibberish, we can say it this way: knowing how often a justice votes for the government in the constitutional cases gives us a very strong sense of how often he votes that way in the cases not involving the Constitution.

   Interpretive byproducts. The interesting question is why some justices vote for the government so much more often than others--and why this tendency tends to be so similar in different types of criminal cases. The results seem generally consistent with the realist hypothesis: the graphs reveal for each judge a signature of ideology, perhaps invisible on a reading of the judge's opinions but evident enough when the results are viewed as a set. But the point isn't that the decisions are "all politics," or that the justices always vote their policy preferences. We must remember that the cases we are studying are the nonunanimous ones, and that there are others where the left-most and right-most justices agree. (9) The better interpretation is that every case provokes competition between a justice's preferences on the one hand and the legal materials on the other. When the legal materials are very strong, they can produce unanimity despite conflicting preferences. But when the legal materials aren't so strong--when they don't point to a clear answer and leave room for discretionary judgment--the competition is won by the justice's underlying preferences and views of the world. Those views of the world are the same regardless of what provision is at stake in a case; that is why there is so much convergence between the results in cases involving different sources of law. Whether a statute or rule or the Constitution is involved, the important question is simply how clearly the justices think the source of law speaks to the case. That clarity is a function of norms about what sorts of arguments the materials plausibly will accommodate.

   These are natural conclusions from the data, but before settling on them we should consider some alternatives. First, might the results be explained by pointing to interpretive theory? It is well known that some justices tend more often than others to vote against the claims of criminal defendants, but some of those same justices also are the ones who endorse originalism as an approach to constitutional interpretation. Justice Scalia in particular subscribes to the view that the words of the Constitution should be given their original meanings; (10) perhaps some others, such as Justice Thomas, think this as well. One byproduct of originalism might be consistent rulings for the government in criminal cases. The parts of the Bill of Rights concerning criminal procedure limit what the government can do; defendants often argue for updated and expanded readings of those limits, which an originalist might reject on grounds of interpretive principle. Maybe that is too cursory a reading of defendants' claims or of the originalist reaction to them, (11) but this notion of interpretive byproducts at least has surface plausibility; it suggests how judges' tendencies to vote for the government or for defendants could be traced to differences in how they approach the task of interpretation--a legalistic explanation for the large variation in the justices' inclination to vote for the government in constitutional cases.

   But then what about the nonconstitutional results? Of course there are theories of statutory interpretation just as there are theories about how to read the Constitution, and sometimes they are analogous: those who like originalism as a constitutional theory often like textualism as an approach to interpreting statutes, and the two approaches resemble one another; indeed, in Scalia's view they are the same. (12) But I do not know of a story about the consequences of textualism in reading statutes that quite matches the story about the consequences of originalism in reading the Constitution--a story that explains why textualism should be expected to help the government so much more often than defendants. Most of the rules in the criminal cases the Court decides are not written to help one side or the other in the same way that the Bill of Rights was written for the benefit of defendants; and most of the rules and statutes the Court interprets are no more than ten or twenty years old, so there is no reason to expect their original understandings to systematically help one side more than the other. Nor do most other approaches to reading rules and statutes--for example, giving effect to their purposes or studying their legislative histories--seem likely to favor one side consistently. And if we could find theories of statutory interpretation that did tend to favor one side, we still would have the problem of explaining why the...