In many ways, Merrick Garland was a safe choice for President Obama to nominate to the Supreme Court. This is in part because Garland’s credentials — degrees from Harvard University, work as a federal prosecutor and service as chief judge of the U.S. Court of Appeals for the District of Columbia — are impeccable. But it is also because he has a reputation as being ideologically moderate.

Is Garland’s reputation as a moderate judge accurate? Some have argued that Garland has deferred to administrative agencies when it benefits unions but not when it benefits corporations, suggesting that Garland may be quite liberal.

Our research that measures the ideology of judges based on the ideology of the clerks they hire, though, suggests that Garland is not very liberal, but actually more moderate — technically, center-left.

How do we measure Garland’s ideology?

Because Garland has been a federal judge for 18 years, one might imagine we could measure his ideology by drawing on his many court decisions. This would be similar to how scholars measure the ideology of Supreme Court justices. For example, the best measures we have to estimate judicial ideology — the Martin-Quinn scores for the Supreme Court Justices — compare how the justices vote relative to each other.

But Garland is a federal appellate judge, and they generally sit in panels of just three and do not all vote together — making it significantly harder to use their votes to estimate their ideology. In fact, little quantitative information exists about the ideological leanings of lower-court judges.

Given the difficulties of using judicial decisions to estimate the ideology of lower-court judges, scholars have attempted to develop other methods. One example is the Judicial Common Space (JCS) score, which measures judges’ ideology based on the ideology of their appointing presidents and home-state senators. But this measure does not use any information from, for instance, the 18 years that Garland has actually served as a judge.

In another attempt to measure the ideology of lower-court judges, a terrific recent post for The Monkey Cage by Tom Clark and others analyzed Garland’s voting in en banc hearings. In these special cases, an entire court (not just a three-judge subset) hears the case being argued. One concern with using voting behavior in en banc cases to estimate ideology is that these cases are not representative of the full spectrum that judges hear and may not provide a comprehensive measure.

Our recent research exploits another aspect of judicial behavior to estimate ideology: the hiring of clerks. Most federal appellate judges — who form the likely candidate pool for Supreme Court nominees—hire four clerks per year. An appellate judge serving for 18 years — as Garland  has — would have hired approximately 72 clerks over that time period.

The premise of our ideology measure is that “birds of a feather flock together.” In other words, one would expect that judges are likely to hire clerks that have ideological leanings similar to their own. Likewise, clerks tend to seek out and accept clerkships with judges whose ideologies they share. Based on this starting point, we combined the most comprehensive data sources on the identity of clerks with a measure of the clerks’ political leanings estimated from their campaign donations. (More technically, we matched clerks to their “CFscores” from the Database on Ideology, Money, and Elections).

We then average the political leanings of the clerks hired by a judge to produce a measure of ideology that places the judge on a spectrum from very liberal to very conservative. Our research shows a strong correlation between this ideology measure and other existing measures for Supreme Court justices, which adds credibility to our assumption that clerk ideology sheds light on that of their hiring judges. And although other credible measures of Supreme Court ideology already exist, an important feature of our approach is that it can be applied to lower-court judges as well, including Garland.

Where does Garland fall on the ideological spectrum?

The following figure presents Garland’s clerk-based ideology score relative to the clerk-based ideology scores of the current Supreme Court justices. In the figure, a value of 0 represents the ideology of the average judge on the federal judiciary, and each one unit move away represents one standard deviation from the average. Negative values mean that the judge is more liberal than average and positive values mean that the judge is more conservative than average.

As the figure shows, based on the clerks Garland has hired, his ideology score places him in a center-left position. In particular, Garland appears to the right of Stephen Breyer, but substantially to the left of Anthony Kennedy and the other conservatives on the court (including the justice he has been nominated to replace, Antonin Scalia).

Although these data suggest that Garland would be the median justice if he were confirmed to the Supreme Court, one possible concern is that lower-court judges may hire very different kinds of clerks once they are “promoted,” for example, clerks who are more ideologically extreme. If true, this pattern would call into question our estimate of Garland’s ideology as center-left.

But our data suggest that this is not a major concern. For the two judges we observe being promoted to the Supreme Court — Sotomayor and Alito — the average ideology of the clerks they hired on the appellate court is quite similar to the average ideology of the clerks they hired after being confirmed to the Supreme Court, as shown in the figure below.

Of course, there are limitations with estimating judicial ideology based on the campaign contributions of their clerks. One limitation is that our measure is based entirely on the clerks for whom we were able to find corresponding campaign contributions. Although 41 percent of clerks appear in our donations data, compared with only about 5 percent of the general public, it is still possible that a different picture would emerge if we were able to observe the ideological leanings of every clerk.

Another limitation is that our measure is based on the donations clerks make over the course of their lifetimes. It is possible that the ideology of clerks as estimated in our data is different than it was at the time of the clerkship. But our research gives us reason to think that these concerns are fairly minor.

So where does this leave us? Clearly, Garland’s nomination would shift the court — and the court’s median ideology — significantly to the left. But based on the clerks he has hired, there is good reason to think that Garland is not as liberal as some believe and would in fact be the most moderate member of the court’s liberal bloc.

Adam Bonica is an assistant professor of political science at Stanford University.

Adam Chilton is an assistant professor at the University of Chicago Law School.

Jacob Goldin will be an assistant professor at Stanford Law School in the fall.

Kyle Rozema is a post-doctoral fellow at Northwestern University Pritzker School of Law.

Maya Sen is an assistant professor at Harvard University’s John F. Kennedy School of Government.