Initially, the text above the chart read, “Judge Garland’s score is based on his time as a federal appeals court judge.” The Times itself noted later in the day that this was incorrect: Depicted on the chart were Lee Epstein and co-authors’ “judicial common space scores,” and Garland’s score was based on the ideology of his appointing president, Bill Clinton. Those data were originally collected by one of us (Giles, along with co-authors) and then linked to data on how Supreme Court justices vote. However, as others have pointed out, it is not necessarily sensible to claim that Garland is as liberal as Justice Ginsburg or Justice Stephen G. Breyer simply because they were nominated by the same president.
What would scores based on Garland’s time on the D.C. Circuit look like? Fortunately, one of us (again, Giles) has collected data that might permit us to answer this question. As with all circuit courts of appeal, the D.C. Circuit sometimes decides cases in three-judge panels and sometimes “en banc” – that is, with all of the circuit’s judges participating. With funding from the National Science Foundation, Giles has been in the process of collecting data on the votes cast in all en banc cases heard by the D.C. Circuit from 1938 to 2008.
Using these data, we estimated ideology scores for all judges sitting on the D.C. Circuit who participated in these cases. We used a standard model from psychometrics and ideal point estimation in political science, known as an item-response theory model. Simply put, the model assumes there is a single dimension along which judges can be arrayed according to their propensity to vote together in cases. Overlap among judges who vote together over time allows us to compare them to one another. That dimension is typically assumed to be, or interpreted as, ideology. In our model, lower scores indicate “liberal” and higher scores “conservative.”
The figure below displays the results of this exercise. The curve is a “smoothed” histogram reflecting the distribution of all circuit court judges over the time period. Unsurprisingly, most judges fall somewhere in the middle of the ideological distribution, with extreme liberal and conservative judges less common.
We then plot ideology estimates for a sample of circuit court judges who eventually served on the Supreme Court or are well-known enough to provide reference points. The figure suggests that Garland is indeed a moderate. He is close to the center of all of the judges who have served on the D.C. Circuit since 1938. He is considerably to the right of well-known liberal judges such as Judge Abner Mikva or Judge Harry Edwards. However, he is to the left of well-known conservatives such as Justices Antonin Scalia and Clarence Thomas. Indeed, our estimates place him right at the middle of the set of current justices on the U.S. Supreme Court who also served on the D.C. Circuit prior to 2008 (when our data end).
These scores should be taken with a sizable grain of salt, for three reasons. First, judges in the D.C. Circuit do not vote on the same set of issues as Supreme Court justices. Therefore, measures of ideology based on voting may translate imperfectly from one venue to another. (Note Scalia’s surprising status as a moderate in the figure.) Second, ambitious circuit court judges may be motivated not only by judicial philosophy but also by the hope of an eventual nomination to the high court. This could be particularly problematic, given the common reference, repeated by Obama on Wednesday, to the D.C. Circuit as the “second-highest court in the land.” On the other hand, while this incentive might muddy the waters somewhat, it’s not entirely clear how it cuts: Sometimes, the president’s incentive is to nominate a moderate (e.g., in the face of a hostile Senate), and at other times an extremist (in the face of a restive base). Finally, due to limitations of the data, these measures are estimated with a substantial amount of uncertainty. This uncertainty is particularly acute for judges who participated in few en banc votes.
That being said, we note that the ordering of these scores for eventual Supreme Court nominees mirrors, to a large extent, that derived from scores based on their eventual voting patterns on the Supreme Court, as well as popular perceptions of the court. And in a number of cases, the differences between the eventual justices are statistically significant. More broadly, we can state confidently, for example, that Harry Edwards is more liberal than Ruth Bader Ginsburg, who is more liberal than Merrick Garland, who is more liberal than than Antonin Scalia was, who was more liberal than Clarence Thomas. (We cannot say much about Chief Justice Roberts, because he participated in very few en banc cases during his two-year tenure on the D.C. Circuit.) In particular, our method estimates that there is a greater than 98 percent probability that Merrick Garland is “in between” Ginsburg and Scalia. In other words, that he is comparatively moderate.
Of course, the million-dollar question is whether Garland is to the left or right of Anthony Kennedy. This is a not a question we can answer with these data. Prior to his stint on the Supreme Court, Kennedy served on the 9th Circuit, deciding an entirely different set of cases than those considered by the D.C. Circuit. Thus, we cannot evaluate his propensity to vote with the various other judges on the D.C. Circuit.
Does this exercise answer every question about Garland’s legal philosophy? Not even close — these estimates are very coarse, and should be treated accordingly. Nonetheless, we believe that providing them is an instructive alternative to the view that a judge on the D.C. Circuit is the ideological clone of the president who nominated him. This is particularly the case when we consider presidents’ incentives to select would-be justices strategically.
Tom Clark is Asa Griggs Candler Professor of Political Science at Emory University.
Sanford Gordon is Professor of Political Science at New York University.
Micheal Giles is Fuller E. Callaway Professor of Political Science, Emory University.