Prompted by several recent high-visibility killings by police officers, the U.S. civil rights enforcement regime is the subject of focused attention at the national, state, and local levels. Much of the discussion has centered on the barriers that prevent victims of civil rights violations and their families from obtaining relief through civil litigation. Of all of these barriers, qualified immunity—a powerful judicial doctrine that shields government officials, including those in law enforcement, from being held personally responsible for constitutional violations—has received the lion’s share of attention. Broadly, qualified immunity can bar a damages remedy in civil rights cases even when plaintiffs can establish that their constitutional rights were violated; if defendants can show that the law governing their conduct was not “clearly established,” they are entitled to immunity from suit even if their conduct violated the Constitution. For this reason, the immunity doctrine has been the subject of withering criticism among policymakers, advocates, and academics who argue that the protection comes at too great a cost to justice and accountability.
But while the academic literature criticizing immunity doctrine is both too broad and too deep to summarize in a sentence or even a paragraph, it is essential to note that, for many years, qualified immunity has not been subject to sufficient empirical scrutiny. Anecdotally, no empirical study is required to show that the U.S. Supreme Court has become increasingly solicitous of defendants claiming qualified immunity (notwithstanding some recent notable exceptions). But the court’s decisions are a small fraction of the work of the federal courts—to call them the tip of the iceberg, even, would exponentially overstate the extent to which they are representative of all federal court adjudication. It should come as no surprise, then, that the Supreme Court’s treatment of qualified immunity has had an overriding impact on perceptions of the importance of the defense, both among academics and practitioners (in other work, Joanna Schwartz and I have separately presented data confirming this perception and its role in how attorneys select cases).
Recent empirical work has undermined some of these assumptions. Schwartz’s pathbreaking study of district courts suggests that qualified immunity is rarely dispositive in Section 1983 litigation brought for alleged Fourth Amendment violations. And in earlier work studying Bivens litigation, I reported data along the same lines about the role of qualified immunity in litigated cases.
There is a disjunction, then, between empirical work regarding the role of qualified immunity in trial courts and the resolution of qualified immunity in the handful of cases that reach the Supreme Court. Missing from the discussion, and critical to understanding the role of qualified immunity in the resolution of litigated cases, is an empirical examination of appellate decision-making. My recent article “Qualified Immunity on Appeal: An Empirical Assessment” fills this significant gap in the literature by providing the most comprehensive study to date of the resolution of qualified immunity appeals in federal court.
Prior studies, while informative, were limited in scope and focused on different questions. Some studies have covered only published decisions—a significant problem, as my data show that there is a substantial difference in the outcomes of published and unpublished decisions, confirming that published decisions are not a representative sample. Other studies cover only a random selection of decisions issued in a given time frame. And no prior study had evaluated more than 850 total opinions. Moreover, none of the prior studies was designed to evaluate the basic, but central, question addressed in my analysis here: who succeeds on qualified immunity arguments in the courts of appeals, and what variables are correlated with success. And none of the prior studies provided the detailed certiorari-stage analysis that I provide in the paper, following each appellate decision to determine whether any party sought cert, whether it was granted, and what the outcome was in the Supreme Court. (Reuters published a qualified immunity study that included cert-related data, but it focused only on about 120 petitions involving police excessive force, and its data analysis was sparse.)
By contrast, “Qualified Immunity on Appeal” analyzes the results of 4,054 decisions in the federal courts of appeals, encompassing every appellate opinion issued regarding qualified immunity in the years 2004-2008 and 2010-2015. I omitted 2009 so that I could better evaluate whether the Supreme Court’s 2009 decision in Pearson v. Callahan—a ruling that gave lower courts the freedom to resolve qualified immunity without deciding the predicate question of whether a defendant’s conduct even violated the Constitution—had any impact on appellate outcomes. The results, provided in summary form here, provide several insights that have not been addressed in past empirical work and that should speak to scholars, advocates and policymakers alike.
Appellate Court Analysis
Even if scholars have helped to show that qualified immunity plays a limited role in the resolution of litigated cases in federal district court, “Qualified Immunity on Appeal” shows that when the defense is deployed, it has a significant impact. Over the course of the study period, defendants prevailed on appeal in cases involving qualified immunity much more than plaintiffs—about twice as often, in fact. Relatedly, key to defendants’ overall success in the courts of appeals was asymmetric treatment of district court decisions granting or denying qualified immunity. This dynamic is referred to as “asymmetric review”—district court decisions denying qualified immunity were reversed far more often than decisions granting qualified immunity. In other words, appellate courts have tended to grant qualified immunity in cases where district courts had previously denied it. Figure 1 illustrates this dynamic, in which district court decisions denying qualified immunity were reversed twice as often—almost 46 percent of the time—compared to district court decisions granting qualified immunity—about 22 percent of the time.
Figure 1 shows how each district court (DCT) outcome was resolved in the courts of appeals (CTA).
This asymmetrical treatment of qualified immunity depending on the outcome in district court is persistent. Notwithstanding other variables that appeared relevant to overall outcomes in the courts of appeals, asymmetrical treatment of district court decisions granting or denying qualified immunity was ever-present. It also is unexpected, given that district court decisions resolving qualified immunity at the summary judgment or motion to dismiss stage are reviewed de novo, meaning that appellate courts review district court decisions with no deference to the district court’s decision. One might expect reversal rates to be higher under de novo review than deferential review, but all else being equal, assuming that district and appellate courts are applying the same law to the problem of qualified immunity, one should not expect reversal rates to depend on which party prevailed in the first instance.
The results also show that, even though the courts of appeals as a whole have been relatively consistent in their treatment of qualified immunity from 2004 through 2015, some variables were significantly correlated with how the defense was resolved. Two stand out: circuit court identity and the political party of the president who appointed appellate judges.
As to the first, it will not surprise any experienced civil rights practitioner to learn that some circuits were more generous to defendants in cases involving qualified immunity than others, as different circuits have established different reputations for being more or less favorable to civil rights claimants. In other words, just as Supreme Court justices are perceived as “liberal” or “conservative,” so too are some circuit courts, a perception driven by precedent and the current composition of each court. It is not uncommon for close observers of the circuit courts to speak of presidents having the chance to “swing” a circuit in one direction or another, based on judicial nominations, but one should not overstate this dynamic. Nonetheless, in the qualified immunity dataset, there was evidence that some circuits are quite different in their treatment of qualified immunity.
In the decisions I analyzed, the First, Third, Fifth, and Eighth Circuits were the friendliest to defendants on appeal, with numerically and statistically significant evidence of asymmetric review of district court decisions. Eighth Circuit panels were more than four times as likely to reverse a district court decision denying qualified immunity compared to decisions granting the defense, First and Fifth Circuit panels were three times as likely, and Third Circuit panels were more than twice as likely.
By contrast, in the D.C., Fourth, and Ninth Circuits, there were no statistically significant differences in the reversal rate of district court decisions granting or denying qualified immunity. In the other circuits, reversal rates for denials were higher and statistically significant, but the differences were, in general, not as stark as in the First, Third, Fifth, and Eighth Circuits. In the full paper, I explain why these data are likely not simply a reflection of different strategic approaches taken by defendants and plaintiffs when deciding whether and when to appeal adverse decisions.
The other variable that was significantly related to reversal rates and overall success of the parties was the political party of the president who appointed the judges on each appellate panel, a rough proxy for judicial ideology. As Figure 2 shows, in counseled cases, plaintiffs tended to prevail more often in the cohort of cases decided by panels with more judges appointed by a Democratic president. In other words, Democrat-appointed appellate judges are more likely to deny the qualified immunity defense than are their Republican-appointed counterparts, and Republican-appointed judges are more likely to grant the defense.
Panels composed entirely of Republican-appointed judges issued decisions in which plaintiffs prevailed about 29 percent of the time, compared to panels composed entirely of Democrat-appointed judges, in which plaintiffs prevailed about 50 percent of the time. Further, the relationship appears linear, which means that with each substitution of a Democrat-appointed judge for a Republican-appointed judge, the plaintiffs’ overall success rate steadily increased, about 7 percent each time.
(There was no apparent attitudinal effect in pro se cases, in which litigants represent themselves rather than opting for a lawyer. In these cases, plaintiffs prevailed in the neighborhood of 20-25 percent of the time, regardless of the makeup of the panels. Almost none of the prior studies distinguished between appeals litigated by counsel and those litigated by pro se plaintiffs, a remarkable omission given the evidence that representation is significantly correlated with success in other contexts.)
In Figure 2, one can see that party success on appeal appears related, in a nearly linear fashion, to the number of judges appointed by a Democratic or Republican president.
Part of the explanation for the observed difference in treatment of qualified immunity appeals can be found in asymmetric review of district court decisions, but that asymmetry, like overall outcome, appears closely connected to the ideological makeup of the appellate court. As Figures 3 and 4 show, the data reveal a nearly linear relationship between the number of judges appointed by the president of a particular party and asymmetric review—as the number of Democrat-appointed judges on a panel decreased and the number of Republican-appointed judges increased, defendants were more likely to win reversal of decisions denying qualified immunity and plaintiffs less likely to obtain a reversal of decisions granting qualified immunity.
And for almost every iteration of ideological makeup of the panels, the variation in reversal rates is statistically significant. Panels with one, two or three Republican-appointed judges were more likely to reverse plaintiff wins than defendant wins, with statistically significant differences; panels with one, two, or three Democrat-appointed judges were more likely to reverse defendant wins than plaintiff wins, with the differences statistically significant for all but the three-Democrat-appointed panels. In the full article, I provide additional data to explain why these differences are likely not simply a product of governing law or strategic differences between plaintiffs and defendants. Rather, the disparities are better explained by ideological differences in the resolution of the immunity defense.
Figures 3 and 4 show how the reversal rate on appeal for a grant or denial of qualified immunity is related to the ideological composition of the appellate panel. Panels with more Republican-appointed judges reverse denials of qualified immunity more often than they reverse grants of qualified immunity, and vice versa.
These appellate observations were remarkably stable over time, despite the fact that during the entirety of the study period, the Supreme Court consistently signaled its view that qualified immunity should be granted early and often. This finding suggests that decision-making in the courts of appeals has not been affected by the Supreme Court’s consistent drumbeat of decisions finding that qualified immunity is appropriate.
In the article, I also examine data related to certiorari practice, the process through which the Supreme Court selects its cases (in a given year, about 10,000 petitions are filed with the court seeking review of a lower court decision, but only about 100 or fewer of these petitions are granted). This analysis relating to certiorari practice and outcomes provides additional novel insights on how courts treat qualified immunity. Overall, certiorari was granted in about 8.5 percent of cases that the study examined, a figure significant on its own, given that just about 1 percent of all petitions are granted in modern times.
Here, I will touch on three interesting features of certiorari review in cases related to qualified immunity. (Additional aspects are discussed at length in the full article.)
First, the asymmetric review that was present in the courts of appeals was even stronger in Supreme Court certiorari practice. Although plaintiffs sought certiorari at a slightly higher rate than defendants, the Supreme Court was about six times as likely to grant certiorari, or decide to hear a case, when requested by a defendant than by a plaintiff.
Second, the ideological hue of asymmetric review also was reflected in certiorari practice. When defendants sought certiorari from an appellate decision denying qualified immunity, they had a much higher rate of success when seeking review of decisions issued by appellate panels with more Democrat-appointed judges. The court was much less likely to grant certiorari when all-Republican-appointed panels denied qualified immunity. And the difference between these rates of success were stark: The Court granted cert 25 percent of the time when defendants were petitioning to review a decision issued by three judges appointed by a Democratic president, and granted cert less than 5 percent of the time when defendants sought cert from a decision issued by three judges appointed by a Republican president.
Third, the data suggest that plaintiffs seek certiorari more often than defendants, in all categories of cases. This is a significant finding, as the phenomenon has occurred despite the evidence that a writ of certiorari was much more likely to be granted when seeking to review a decision favoring the plaintiff, and despite the evidence that, when granted, defendants almost always prevailed in the Supreme Court, particularly in Fourth Amendment cases. Plaintiffs sought certiorari more often in every category of case (the cases were coded based on whether they presented Fourth Amendment claims, Eighth Amendment claims, First Amendment claims or due process claims), although the difference was significant (both in a statistical sense and numerically) in only some of the categories. And there was large variation in the rates at which defendants sought certiorari, depending on the case type. Perhaps the variation in certiorari practice reflects strategic thinking about which cases are likely to result in success in the Supreme Court. But just as with the overall grant rates, for every category of case, petitions for a writ of certiorari were far more likely to be granted when seeking to reverse a decision that favored the plaintiff.
These data are significant for what they suggest about the resolution of qualified immunity in the courts of appeals and the Supreme Court. At the appellate level, the data suggest that the outcome of an appeal is influenced by who prevailed in the district court, which circuit decided the case, and the political party of the president who appointed the district court judge who heard the case in the first instance and the appellate judges who heard the appeal. The evidence from the Supreme Court’s certiorari practice mirrors much of what was revealed in the data from the courts of appeals. Evidence of asymmetric review was strong: Defendants were more successful at obtaining review of, and ultimately obtaining relief from, unfavorable appellate decisions. And judicial ideology appeared relevant to the Supreme Court’s resolution of qualified immunity: The court exercised its jurisdiction to review decisions more often from appellate panels that comprised Democrat-appointed judges when those panels issued plaintiff-friendly opinions on qualified immunity. These data thus have implications in multiple arenas.
For one, the findings offer additional evidence that, when raised, the immunity is a powerful defense—which is especially important information for those who have argued to revisit or eliminate qualified immunity. Overall, defendants experienced far greater success than plaintiffs in the courts of appeals, in terms of both overall outcomes and their success in defending successful outcomes on appeal. Qualified immunity was granted at a higher rate than it was denied, and both the courts of appeals and the Supreme Court were more likely to reverse denial of qualified immunity than to reverse a grant of qualified immunity. In short, appellate courts appear to have supported the doctrine’s protections much more than the district courts whose decisions they review.
Moreover, these data undermine the presumption regarding qualified immunity that “clearly established” law has an objectively verifiable content. Rather, the data suggest that ideological priors (with the caveat that the party of the appointing president is a rough proxy for ideology) filter to a significant degree the perceptions of when law is so clear that it is obvious to any reasonable officer. While legal realists will not be surprised by this outcome, the data expose the possibility that qualified immunity serves as a means to obscure the role that judicial ideology plays in adjudicating civil rights litigation. This may offer further reasons to restrict or eliminate access to the defense.
In addition, more broadly, these data cast doubt on whether the Supreme Court is exercising its certiorari jurisdiction in qualified immunity cases to achieve uniformity in the application of federal law. One of the principal justifications for the exercise of certiorari jurisdiction is to ensure that federal law is uniformly applied in federal and state litigation. Other than cases involving the death penalty, the exercise of certiorari jurisdiction is not thought to be used for error correction. The data gathered here suggest that there is persistent disuniformity in the application of qualified immunity, despite the Supreme Court’s consistent attention to the doctrine over the past two decades. There is considerable variation between circuits and based on panel composition within each circuit, to name just two variables.
Perhaps one explanation for the Supreme Court’s failure to secure more uniform application of qualified immunity lies in the evolution of the doctrine itself. As I describe in more detail in my article, the court has made the doctrine more powerful by narrowing what counts as clearly established law and by providing mechanisms whereby courts never answer the predicate question of what rights exist, as opposed to whether a right is clearly established. Scholars have worried about the implication of this dynamic for recognizing new rights. But if deciding whether a right is clearly established is as fact intensive as the court has suggested, it also means decisions holding that a particular right was not clearly established have less relevance for the next qualified immunity case. The court may have laid the groundwork for an essentially lawless qualified immunity jurisprudence. Alternatively, and relatedly, perhaps courts of appeals perceive the Supreme Court as engaging in error correction, muting any signal the court may intend to be sending.
Finally, these data offer avenues for further research into both the resolution of qualified immunity and the litigation of the issue by attorneys. As with any empirical study, the analysis raises as many questions as it answers. The data suggest, but cannot evaluate, the possibility that the plaintiffs’ and defendants’ bars undertake different strategic considerations when deciding whether to appeal or seek certiorari. Nor can they evaluate whether the plaintiffs’ or defendants’ bar is better at assessing the likelihood of success on appeal or certiorari. And while the data suggest that the Supreme Court’s gradual strengthening of the power of qualified immunity has not had an impact on the resolution of the defense in the courts of appeals, further research is necessary to confirm that conclusion. It is possible, after all, that as the court has raised the bar for plaintiffs in civil rights cases, counsel have shifted their strategies in qualified immunity cases, with plaintiffs selecting different cases and defendants attempting to invoke the defense in more marginal cases. If this were the case, even as the court changes the legal landscape, the relative success of plaintiffs and defendants may remain unchanged. These and other avenues for future research are discussed in greater detail in the full-length article.
Qualified immunity is considered to be one of the most significant barriers to civil rights enforcement, and it is currently under assault from multiple perspectives. The Supreme Court has made clear that it views qualified immunity as an increasingly important tool in protecting defendants from damages liability. The evidence suggests that district courts are less sanguine on the doctrine, using it far less often to resolve civil rights cases.
This study unlocks one key to understanding the disjunction between district courts and the Supreme Court. Qualified immunity is a powerful defense when deployed in the courts of appeals, but if the Supreme Court means to be sending a signal that appellate courts should be strengthening the defense, this analysis suggests that it is muddled. At the same time, however, this study also indicates that resolution of qualified immunity appears to be linked in a significant way to variables that undermine a core premise of the immunity defense: that “clearly established” constitutional law is ascertainable to the reasonable government official.