Topics

Observers have suggested that adding sources of interpretation tends to increase interpreter discretion. The idea is embedded in a quip, attributed to Judge Harold Leventhal, that citing legislative history is like “looking over a crowd and picking out your friends.” Participants in debates over interpretive method have applied the idea to the proliferation of other sources as well, including canons of construction and originalist history. But the logic of “more sources, more discretion” has escaped serious testing. And predicting the effect of source proliferation is not a matter of logic alone. The empirical study of how information loads affect behavior has grown dramatically in recent decades, though almost without notice in legal scholarship on interpretive method.

This Article tests the logic and evidence for “more sources, more discretion.” The idea turns out to be incorrect, without more, as a matter of logic. Adding sources tends to reduce the chance of discretion using a simple model of interpretation. This starter model depicts judges as aggregators of source implications, and it draws on basic probability theory and computer simulations to illustrate. The analysis does change if we allow judges to “spin” or “cherry pick” sources, but without much hope for limiting discretion by limiting sources. Of course, judges will not always behave like machines executing instructions or otherwise follow the logic of these models. Thus the Article goes on to spotlight provocative empirical studies of information-load effects, develop working theories of interpreter behavior, and present new evidence.

After emphasizing that interpreters might ignore additional information at some point, the Article tests three other theories. First, an extended dataset casts doubt on an earlier study that linked a growing stock of precedents to increased judicial discretion. Adding to the pile of precedents seems to have no simple pattern of effect on discretion. Second, existing studies indicate that increasing information loads might prompt judges to promote the status quo, and new data suggest that this effect depends on the type of information added. The number of sources cited in appellant briefs appears to have no effect on judges’ willingness to affirm—in contrast with the number of words and issues presented, which may have opposing effects. Third, an expanded dataset supports an earlier finding that judges who face a large number of doctrinal factors might weight those factors in a quasi-legal fashion. This time-saving prioritization does not seem to follow conventional ideological lines.

With simple intuitions in doubt, thoughtful work remains to be done on the effects of source proliferation. Observers interested in judicial discretion have good reason to look beyond source proliferation to find it. And observers interested in institutional design have good reason to rethink the range of consequences when information is added to our judicial systems.

In administrative law, it is generally assumed that once an agency promulgates a final rule, its work on that project—provided the rule is not litigated—has come to an end. In order to ensure that these static rules adjust to the times, therefore, both Congress and the White House have imposed a growing number of formal requirements on agencies to “look back” at their rules and revise or repeal ones that are ineffective.

Our empirical study of the rulemaking process in three agencies (N = 462 revised rules to 183 parent rules) reveals that—contrary to conventional wisdom—agencies face a variety of incentives to revise and update their rules outside of such formal requirements. Not the least of these is pressure from those groups that are affected by their regulations. There is in fact a vibrant world of informal rule revision that occurs voluntarily and through a variety of techniques. We label this phenomenon “dynamic rulemaking.” In this Article, we share our empirical findings, provide a conceptual map of this unexplored world of rule revisions, and offer some preliminary thoughts about the normative implications of dynamic rulemaking for regulatory reform.

Testing the Constitution

We live in the age of empiricism, and in that age, constitutional law is a relative backwater. Although quantitative methods have transformed entire fields of scholarly inquiry, reshaping what we ask and what we know, those who write about the Constitution rarely resort to quantitative methodology to test their theories. That seems unfortunate, because empirical analysis can illuminate important questions of constitutional law. Or, at least, that is the question to be tested in this Symposium.

We brought together a terrific group of scholars with a unique assignment. We paired distinguished constitutional thinkers with equally accomplished empiricists. We asked the law scholars to identify a core question, assumption, or doctrine from constitutional law, and we asked the empiricist to take a cut at answering it, or at least at figuring out how one might try to answer it. We understood that their answers might be preliminary at best, that the questions might be resistant to easy answers. This is so, in part, because empiricism is as much a means of refining questions as it is a way of answering them.

The balance of this Foreword is, in a sense, an introduction to the idea that more serious empirical analysis can further both constitutional law scholarship and constitutional law decisionmaking. Hence our title: Testing the Constitution.

Attorneys General as Amici

An important strain of federalism scholarship locates the primary value of federalism in how it carves up the political landscape, allowing groups that are out of power at the national level to flourish—and, significantly, to govern—in the states. On that account, partisanship, rather than a commitment to state authority as such, motivates state actors to act as checks on federal power. Our study examines partisan motivation in one area where state actors can, and do, advocate on behalf of state power: the Supreme Court. We compiled data on state amicus filings in Supreme Court cases from the 1979–2013 Terms and linked it up with data on the partisanship of state attorneys general (AGs). Focusing only on merits-stage briefs, we looked at each AG’s partisan affiliation and the partisanship of the AGs who either joined, or explicitly opposed, her briefs. If partisanship drives amicus activity, then we should see a strong negative relationship between the partisanship of AGs opposing each other and a strong positive relationship between those who cosign briefs.

What we found was somewhat surprising. States agreed far more often than they disagreed, and—until recently—most multistate briefs represented bipartisan, not partisan, coalitions of AGs. Indeed, for the first twenty years of our study, the cosigners of these briefs were generally indistinguishable from a random sampling of AGs then in office. The picture changes after 2000, when the coalitions of cosigners become decidedly more partisan, particularly among Republican AGs. The partisanship picture is also different for the 6% of cases in which different states square off in opposing briefs. In those cases, AGs do tend to join together in partisan clusters. Here, too, the appearance of partisanship becomes stronger after the mid-1990s.

Oliver Wendell Holmes’s notion of the marketplace of ideas—that the best test of truth is the power of an idea to get itself accepted in the competition of the market— is a central idea in free speech thought. Yet extant social science evidence provides at best mixed support for the metaphor’s veracity, and thus for the view that the truth of a proposition has substantial explanatory force in determining which propositions will be accepted and which not. But even if establishing an open marketplace for ideas is unlikely to produce a net gain in human knowledge, it may have other consequences. We illustrate how to empirically study the consequences of establishing or restricting a communicative domain. Our focus is on time, place, and manner restrictions, and we examine two potential natural experiments involving speech buffer zones around polling places and health care facilities providing abortions. Using a regression discontinuity design with geocoded polling information for over 1.3 million voters in two high-density jurisdictions (Hudson County and Manhattan), we provide suggestive evidence that speech restrictions in Hudson County reduced turnout amongst voters in the buffer zone. By failing to cue voters of the election, speech restrictions may have unanticipated costs. And using difference-in-differences and synthetic control matching with state-level data from 1973 to 2011, we illustrate how one might study the impact of speech restrictions around health care facilities. Although the evidence is limited, Massachusetts’s restrictions were accompanied, if anything, by a decrease in the abortion rate. Buffer zones might channel speech toward more persuasive forms, belying the notion that the cure for bad speech is plainly more speech.

An Empirical Study of the Roberts Court

Constitutional law casebooks, generations of constitutional lawyers, and the Justices themselves say that the Court is more likely to depart from precedent in constitutional cases than in other types. We test this assumption in cases decided by the Roberts Court and find, at odds with earlier studies, that the data provide inconclusive support for it. Other factors, especially criticism of precedent by lower courts and lawyers, are more consistent and stronger predictors of the Court’s decisions to depart from precedent. These findings have interesting implications for lawyering, teaching, and judging in the constitutional law context.

Supreme Court doctrine grants special protection against laws that “chill” protected speech, most prominently via the overbreadth doctrine. The overbreadth doctrine permits persons whose own speech is unprotected to challenge laws that infringe the protected speech of third parties. The Court has not generally applied overbreadth and the other speech-protective doctrines to other constitutional rights even though other rights could also be subject to a chilling effect. The case law simply assumes that the chilling effect only acts on the exercise of speech, and that this justifies treating speech differently from other rights.

We tested these assumptions with respect to abortion rights. By comparing abortion rates with state laws over a two-decade-plus period, we found a statistically significant correlation between laws forbidding late-term abortions and the reduction of not only late-term but also “near-late-term” abortions, i.e., abortions in the roughly one month before the period in which abortions are forbidden. That effect persists even after controlling for potentially confounding variables, such as the number of abortion providers and pro-life public opinion. Moreover, the effect is not limited to the year of enactment or associated with failed policy initiatives, suggesting that the impact is due to the law itself rather than associated publicity. These findings are consistent with, and strongly suggestive of, a chilling effect on abortion providers and/or women seeking abortions. This result undermines the implicit assumption that the chilling effect is unique to laws regulating speech and vindicates the general proposition that laws can chill the exercise of constitutional rights beyond their literal coverage.

Testing the Harm of Campaign Spending

In its landmark campaign finance decision Citizens United v. FEC, the Supreme Court found that favoritism or influence over elected officials gained by wealthy campaign contributors does not—in the absence of outright corruption—give rise to the sort of constitutional harm that would justify restrictions on campaign spending. The Court was also insistent that any perceptions of ingratiation would not undermine the electorate’s faith in democracy. This paper challenges the doc- trinal and empirical underpinnings of those assertions. We argue that a loss of faith by the electorate implicates a central constitutional value and is a sufficiently compelling interest to justify campaign finance regulation. We also demonstrate empirically that the Court should not have been so confident that the elecorate’s faith in democracy is unaffected either by the appearance of influence or access due to campaign spending or by independent expenditures.

The Supreme Court’s decision in Shaw v. Reno established an “analytically distinct” constitutional claim of racial gerrymandering for majority-minority districts drawn predominantly on the basis of race. The case was and continues to be controversial, because the precise nature of the injury caused by such districts has been a persistent source of debate. Shaw districts did not minimize a group’s representation, but rather they communicated an “expressive harm” due to signals they sent to the electorate and representatives that the jurisdiction relied too much on race in the construction of a district. Such districts, the Court argued, communicated racial stereotypes that individuals belonging to the same racial group were politically interchangeable, despite their many social and economic differences. This paper tests the “Shaw hypothesis” with recent survey data. We find no patterns in racial attitudes based on the shape and racial composition of a congressional district. We do, however, find substantial and expected gaps among racial groups concerning attitudes toward the practice of majority-minority districting, Section 5 of the Voting Rights Act, and job approval of the respondent’s representative.

Standard-form contracting is the engine of the mass-market economy, yet we know little about what drives it and what factors are associated with its evolution. Understanding change and innovation of the substance, length, and complexity of fine print in the consumer context can help regulators identify sources of potential intervention as well as help them evaluate the effectiveness of mandatory disclosure regimes, which are commonly used as consumer protection tools. This Article studies the rate, direction, and determinants of change in consumer standard-form contracting. We examine what changed between 2003 and 2010 in the terms of 264 mass-market consumer software license agreements. Thirty-nine percent of contracts materially changed at least one term, and some changed as many as fourteen terms. The average contract became more pro-seller as well as several hundred words longer. The increase in length is not due to the use of simpler language. Contract readability has been constant: The average contract is as readable as an article in a scientific journal. The variance of contract length has grown, as has the variance in overall pro-seller bias, resulting in reduced contract standardization over time. Firms that were younger, larger, or growing, as well as firms with inhouse counsel, were more likely to change existing terms and to introduce new terms to take advantage of technological and market developments. Contracts appear to respond to litigation outcomes: Terms that were increasingly enforced by courts were more frequently used in contracts, and vice-versa. The results indicate that software license agreements are relatively dynamic and shaped by multiple factors over time. We discuss potential consumer protection implications as a result of the increased length and complexity of contracts over time.