“Ask the Author”: Cass Sunstein & Tom Miles, Part II

This edition of “Ask the Author” features a conversation with Cass Sunstein and Tom Miles about their article entitled, “Do Judges Make Regulatory Policy?: An Empirical Investigation of Chevron,” see here, which was published in Volume 73 of the University of Chicago Law Review. This article has attracted quite of bit of attention and commentary in the blogosphere in recent months, see here, here, and here, and I hope to explore some of the critiques of their article in more depth here. Part I of the series can be found here.

The Supreme Court is unlike other courts in many ways, particularly in that it is more likely to hear ideologically divisive cases. Do you think your conclusions would differ if you looked at circuit courts?

In fact, we looked at circuit courts of appeals. To make the data collection manageable, we limited our analysis to decisions between 1990 and 2004 involving the NLRB and the EPA. We coded the agency interpretations as liberal or conservative by looking to the identity of the party challenging the agency decision, just as we did for the Supreme Court decisions.

We found that overall both Democratic and Republican appointees voted to validate agency interpretations about 64% of the time. But a judge’s willingness to validate the agency interpretation rose substantially when the interpretation matched the judge’s presumed ideological preference. Democratic appointees voted to validate 74% of the time when the agency interpretation was liberal, compared to only 51% when it was conservative – a difference of more than 20 percentage points. Republican appointees showed a similar pattern but in the opposite direction. They voted to validate conservative agency interpretations 70% of the time on average and liberal ones 60% of the time – a difference of 10 percentage points.

Our analysis also revealed so-called “panel effects.” That is, a judge’s vote was influenced by the partisan affiliations of others on the appellate panel. For example, the rate at which the average Democratic appointee sitting with two other Democrats voted to validate liberal agency decisions was 30 percentage points higher than the rate at which she voted to validate conservative agency decisions. This gap was smaller when Democratic appointees sat with at least one Republican appointee. We observed similar patterns – but in the opposite direction, of course – for Republican appointees.

Are the patterns you detected under Chevron review unique in administrative law?

We don’t think so. We recently completed a study of judicial review under the “hard look” standard, see here. Our attention was limited to decisions by appellate courts because the Supreme Court issued relatively few decisions under this standard.

Again we found that a judge’s willingness to validate an agency decision rose substantially when the agency decision matched the judge’s presumed ideological preference, and that a judge was more likely to vote in a predictably partisan fashion when she sat on a panel with two co-panelists from the same political party. Other studies find similar patterns in many areas of the law. See Cass R. Sunstein, Are Judges Political? (Brookings, 2006).

On a more fundamental level, how would you define judicial activism and judicial restraint? I tend to be someone that does not believe that we can define judicial activism in a meaningful and universal way, and thus as a label it is largely about decisions with which one side or another disagrees. Nonetheless, we hear those words mentioned quite a bit, especially in the popular press, but what do they really mean? Do they describe only whether judges or justices vote to overturn decisions made by the executive or legislative branches or are the concepts of judicial activism and restraint broader?

Well, you can say that activism means following the wrong method of interpretation. On that view, “activist’ means “wrong.” If you’re an originalist, nonoriginalists are activists. If you’re a minimalist who believes in stare decisis, originalists might be activists. Approaches of this sort turn “activist” into an unhelpful word of disapproval.

There’s much to be said for Judge Posner’s neutral definition of judicial activism: An activist is someone who votes to strike down federal statutes. So too, an activist can be defined as someone who votes to strike down federal regulations. These measures allow for empirical tests, and our study was an effort to undertake precisely this sort of test.

An important lesson from the Legal Realists of the 1930’s was that judicial decision making inevitably involves some degree of judicial policy making. Normatively, reasonable people can disagree about the amount and kind of policy making that is appropriate for the judiciary. We suspect that “judicial activism” is a term that commonly get deployed when judicial policy making exceeds one’s preferred amount of judicial policy making or when one simply disagrees with the substance of judicial decisions.

We are hopeful that further empirical studies can provide some insights into when and how judicial policy making occurs. In a forthcoming essay in the University of Chicago Law Review, we recently expressed our belief that the current renaissance of empirical research on judicial decision making – which we and others have called the “new legal realism” – will in coming years greatly add to the understanding of the extent of judicial policy making, see here.

Ed Whelan critiques your methodology by saying that you were not working from the same data set in the paper with respect to all the Justices. In other words, Justices Breyer and Ginsburg were not on the Court during the entire period studied, from 1989-2005. The clear implication of his critique is that Breyer and Ginsburg would have been more likely to overturn the agency decisions of the first Bush administration than during the entirety of the Clinton administration, yet we have far more data points with respect to those two Justices from the latter administration. How do you respond to Whelan’s critique?

This critique just misses the mark. In response to Mr. Whelan’s criticism, we went back and limited our analysis to the period after 1994, when Justices Breyer and Ginsburg had joined the Court. We found much the same patterns as we did in the longer time period. In the post-1994 period, Justice Breyer remains the most restrained as well; he upheld agency decisions 84% of the time. Justice Scalia remains the most aggressive user of judicial power; in this period, he validated the agencies only 52% of the time.

Ilya Somin at Volokh Conspiracy queries whether it is the distribution of agency errors, rather than any notion of judicial activism, that explains your results. In other words, I believe he would argue that because agencies tend to err more often in the liberal rather than conservative direction, it is unsurprising to find that Justice Thomas is the most activist under your measure because erroneous agency decisions, on average, tend to be furthest from his ideal policy point. The contrary would be true if, on average, more agency errors occurred in a conservative direction. How do you respond to Professor Somin’s point regarding agency errors?

We think Professor Somin’s criticism would have some merit had we limited our attention to the justices’ overall validation rates. But, as just described, we also examined whether each justice was more likely to favor an agency when the agency decision was liberal rather than conservative. (Agency decisions challenged by industry were deemed liberal, and those challenged by public interest groups were coded conservative. This measure relies on the litigants’ own assessment of agency errors, which is likely superior to any classification we might attempt.) Thus, we examined the distribution of agency decisions directly.

A larger point is that studies of judicial decision making, such as ours, typically do not measure directly the correctness or accuracy of a court’s decision. This is not surprising given the difficulty of assessing the “correctness” of judicial decisions. (Indeed, disagreements about the definition of judicial activism are in many ways disputes over the correctness of the decisions.) We investigated decisions under Chevron precisely because the command of deference to reasonable agency interpretations gives a plausible presumption about how courts should proceed.