Abstract

Excerpted From: Stuart Minor Benjamin, Kevin M. Quinn and ByungKoo Kim, Twenty-first Century Split: Partisan, Racial, and Gender Differences in Circuit Judges Following Earlier Opinions, 49 Brigham Young University Law Review 367 (2023) (139 Footnotes) (Full Document)

Many researchers have studied the relationship between judges' behavior and their political party, race, or gender. Most of these studies focus on whether judges' votes in a particular set of cases are associated with the judges' party, race, or gender. As we discuss in more detail in Part II, these studies have generally found that Democratic and Republican appointees vote differently in some categories of ideologically salient cases, that judges of different races vote differently in a smaller subset of cases, and that male and female judges vote differently in an even smaller subset of cases.

Judges' votes are very important, but judges of course issue opinions, and the reasoning in those opinions helps to shape later cases. One significant aspect of a given opinion is which earlier cases it follows and which it casts doubt on. Using an original dataset containing all the substantive Shepard's Citations (Shepard's) treatments in all federal appellate majority opinions from 1974-2017 (670,784 opinions in total), we examine how a judge's party, race, and gender are associated with changes in an opinion's likelihood of following an earlier circuit opinion.

The vast amount of data we have allows us to make adjustments for each of the 521 circuit-year combinations in our data. This contrasts with most empirical studies of federal appellate decision making, which do not make such adjustments and include far fewer cases (usually only hundreds of cases). We make these adjustments for all our analyses to reduce the possibility of producing biased results (e.g., by one circuit having a larger number of judges of a particular party, race, or gender in earlier years and another circuit having a larger number in later years, or by turnover on party, race, or gender lines within a circuit over the long period of our study).

What might one expect to find? At one extreme, we might expect no meaningful differences related to partisanship, race, or gender in how later judges substantively treat earlier opinions: judges will follow, say, the canonical case rejecting a claim of ineffective assistance of counsel, and that canonical case will not have elements more likely to appeal to a later judge of the same party, race, or gender as the opinion author. The idea is that when judges choose which opinions to follow, party, race, and gender are irrelevant. On this account, judges are not influenced by the party, race, or gender of the authors whose opinions they follow (and may not even notice the party, race, or gender of the earlier author), and nothing in the earlier opinions of a judge of a particular party, race, or gender will be correlated with anything that a later judge might value. If this account is correct, we would not expect to see any differences in substantive treatments correlated with party, race, or gender.

At the other extreme, we might expect pervasive differences related to party, race, and gender in how later judges substantively treat earlier opinions. The idea is that judges can choose among different opinions on ineffective assistance of counsel (to stick with the example), and they will tend to follow opinions written by judges with whom they share a party, race, or gender, because the language and reasoning in the opinions they follow are closer to the later judges' preferences. As with the possibility that there are no differences based on party, race, or gender, it need not be that judges pay attention to these characteristics of the judges whose opinions they follow. The later judge may simply find the reasoning of a particular earlier opinion attractive for reasons correlated with these characteristics and thus be more likely to choose to follow an opinion written by a judge of the same party, race, or gender even if the judge ignores the identity of the author of the earlier opinion. So a later judge's awareness of the party, race, or gender of an opinion author is not necessary to motivate this account. But for this account to be plausible without such awareness, there must be some element of an opinion correlated with party, race, or gender (such as ideology) such that a later judge is likely to prefer to follow an opinion written by a judge of the same party, race, or gender. If this account is correct, then we would expect to see pervasive differences in substantive treatments correlated with party, race, and gender: judges will be more likely to (consciously or unconsciously) follow opinions written by those of the same party, race, or gender.

Our findings are surprising--and more nuanced than either of these accounts would suggest. Pooling data over the full time span of our data (1974-2017), we find statistically significant, but substantively small, partisan differences and intraparty racial and gender differences in authoring judges' positive treatments of earlier opinions. But looking at the pooled data masks a dramatic development that is the real story: Far from being stable over time, the differences we find are quite small in the early years of our study and rise dramatically over time, becoming large and thus substantively meaningful. Further, the differences are greatest for the most ideologically charged categories of cases. Put differently, we do not see the sort of pervasive partisan, racial, and gender differences over the full time period that would exist if judges were consistently influenced by these factors. Instead, we see a sharp rise in partisan differences, and we see a rise in racial and gender differences within parties (particularly for Republicans). These differences are most dramatic in cases with the most ideological salience. The fact that these racial and gender differences occur within parties highlights that these differences are not the result of statistical associations between partisanship and race or between partisanship and gender. To pick the clearest example, Black and White co-partisan judges treat opinions by Black co-partisans differently in ways that are not only statistically significant but also large and therefore substantively meaningful.

A closer look at our results reveals that the partisan differences in substantive treatments of opinions written by Democratic appointees are larger than the differences in treatments of opinions by Republican appointees. What could explain greater partisan differences in treatments of Democratic opinions than in treatments of Republican opinions? The best explanation involves an accelerating shift among Republican appointees in a conservative direction compared to a steady shift among Democratic appointees in a liberal direction. Such a pattern produces results strikingly similar to what our data show. To be clear, we cannot prove this explanation, but we think it is the most likely one.

Finally, the race and gender findings (which, again, are within party) are particularly interesting, because our study of substantive treatments finds significant Black-White differences and significant Hispanic-White and female-male differences among Republicans, whereas studies of voting have found relatively few significant differences across judges of different races and genders. As with partisanship, these differences have risen in recent years and apply to our broad category of ideologically salient cases (not just the subset of ideologically salient cases that are race-salient and gender-salient).

We think the best explanation of these racial and gender differences is that they are capturing an element of ideology that partisanship does not capture. Our results are not consistent with a desire of judges of a particular race or gender to enhance the status of those of the same race or gender, because that would not explain the rise over time. The temporally increasing racial and gender differences in our data are similar to the increasing partisan differences we find. Both increases align with the widely documented rise in polarization among U.S. elected officials and within U.S. society more broadly. There is evidence that just as a partisan affiliation reveals information about political attitudes and ideology, race and gender are associated with political attitudes and ideology. In light of these correlations, intraparty race and gender provide a finer-grained proxy for a judge's ideological leanings. Consequently, the fact that we see Hispanic-White and female-male differences only within Republican judges provides some support for the proposition that White male Republican appointees are the central contributors to the accelerating rightward move among Republican appointees that we find.

Part of what is striking about the racial and gender differences in substantive treatments is that prior research on merits votes has found differences in voting behavior between female and male judges, or judges of different races, only in limited subsets of cases. By contrast, the intraparty race and gender differences we find suggest the existence of subtle ideological differences operating outside of race- or gender-salient cases--differences that previous studies have failed to identify.

Our results do not support the proposition that judges have always been pervasively influenced by party, race, or gender. Instead, our results provide evidence of increasing partisan, racial, and gender differences among judges in recent years that reflect some combination of greater ideological differences and a greater willingness to let ideological differences influence how opinions are written. Judges have some insulation from the increasing ideological polarization in the country, but that insulation goes only so far.

The rest of the Article proceeds as follows. In Part I, we discuss the importance of majority opinions' substantive treatments of earlier opinions. Substantive treatments are not mere citations. Following an opinion means relying on it as controlling authority. Shepard's is the most studied and accepted source of substantive treatments, and we rely on it here. Part II discusses the empirical literature on differences in judicial behavior, which has focused on party, gender, and race. That literature has focused mainly on judges' votes and found partisan differences in some ideologically salient case categories, racial differences in a few case categories, and gender differences primarily in sex discrimination cases. In Part III we lay out our research questions. We focus on partisan differences as well as intraparty differences with respect to race and gender. In Part IV we present our data and research design, which are unique within the literature. Part V presents our primary results. In Part VI we discuss our findings and some of the interesting questions they present, such as the greater partisan divergence for treatments of opinions written by Democratic appointees than for those written by Republican appointees--a difference that is consistent with an accelerating conservative shift among Republican appointees as opposed to a steady liberal shift among Democratic appointees. A brief conclusion follows.

[. . .]

If we had written this Article twenty years ago, we would have been able to tell a heartening story about ideology--that it was not a significant factor in judges' decisions to follow earlier opinions. But today the story is different: In recent years, polarization has risen along party lines, within both parties for Black versus White judges, and within Republicans for Hispanic versus White and female versus male judges. And these differences are greatest in the most ideologically salient cases. The best explanation is ideological polarization between and within political parties.

And there is an interesting wrinkle: The rising partisan polarization we find in judges' treatment of earlier opinions is greater when a Democrat wrote the earlier opinion. That pattern is consistent with Republican appointees rapidly becoming more conservative and Democrats appointees more slowly becoming more liberal.

More broadly, our data indicate that judges were not particularly ideological in their substantive treatments of earlier opinions in earlier decades, but in the twenty-first century have become increasingly so, with attendant impacts on the shape of legal doctrines. Judges, it seems, are subject to the polarization that affects the rest of us.

Benjamin is the William Van Alstyne Professor of Law at Duke Law School;

Quinn is a Professor at Emory Law School and in the Department of Quantitative Theory and Methods at the Emory College of Arts and Sciences;

Kim is an Assistant Professor at the KDI School of Public Policy and Management.