Changing What Judges Do

Toby J. Heytens

82 U Chi L Rev Dialogue 151 (2015) [Essay PDF]

A Response to Matthew Tokson, Judicial Resistance and Legal Change, 82 U Chi L Rev 901 (2015) [Article PDF]


I wish I had found a way to write about this first. Professor Matthew Tokson starts with a great topic: the possibility that the very judges charged with implementing new doctrines or other legal rules will sometimes end up resisting them instead.1 Tokson’s article brings together a bunch of interesting examples, and it offers sensible and by-and-large compelling explanations for what unites those examples, as well as criteria for identifying further ones. It even has a neat title.

This being a response, I will, of course, be doing some responding. My first point involves the source and nature of the challenge. Tokson has identified a variety of broadly applicable and not obviously “political” preferences that may slow the pace at which judges implement new legal requirements. At the same time, however, I suggest that using the word “nonideological” may not be the best way of capturing at least some of the preferences that Tokson identifies (for example, a bias in favor of the status quo) and that the strength of some of those preferences (for example, a preference for familiar rules over unfamiliar standards) may vary substantially from judge to judge. Second, although I think that Tokson is onto something when he suggests that “scholars looking for noncompliance in controversial, highly publicized constitutional law cases have been looking in exactly the wrong place,”2 it seems to me that Tokson’s own map is missing at least one of the most important markers: for example, the extent to which a particular judicial ruling will be subject to meaningful appellate review. Third, and most broadly, I think that the term “judicial resistance” is surprisingly hard to pin down and perhaps should be abandoned in favor of a broader approach focused on potential barriers to effectuating legal change.

I.  Why Might Judges Resist Legal Change?

It is sometimes easy to forget, but “[j]udicial policies do not implement themselves.”3 Professor Tokson is thus surely right that “[c]hanging the law that actually resolves legal disputes is not as simple, or as immediate, as it appears.”4

It seems unsurprising that judges may resist (or, less pejoratively, may have a difficult time faithfully implementing) new legal regimes that go against their deeply held moral or political beliefs.5 But Tokson points out that judges may hinder the direction or rate of on-the-ground legal change for other, less obvious reasons. In particular, as Tokson explains, judges are vulnerable “to many of the same unconscious biases, aversions to costs, and preferences for the familiar status quo as the rest of us.”6 These limitations may, in turn, cause judges to seek to do less work rather than more, favor less mentally taxing modes of decisionmaking over more taxing ones, and prefer the familiar and the routine over the unfamiliar and the new—and to do so regardless of whether a new law tells them to do otherwise.7

There is not much here about which I disagree with Tokson, so instead I’ll make two limited points. First, although I agree that many of the preferences that (may) lead (some) judges to resist implementing new legal rules are not necessarily “political” in nature, I am not certain that it is accurate to label all of them “nonideological.”8 “Nonideological” may be a useful label for things like a desire to avoid work or mental fatigue, or the difficulty of overcoming anchoring effects.9 But an all-other-things-equal bias in favor of keeping things the way they have been in the past10 starts to sound a lot like a certain type of conservatism. And conservatism is most definitely an ideology—one that the sorts of people who first become lawyers and then become judges are particularly likely to share.

Second, it seems likely that these nonpolitical preferences, like political preferences, may vary a lot from person to person. Take an example that Tokson uses: the shift from a bright-line and easy-to-apply rule to an all-things-considered standard.11 That sort of switch undoubtedly makes the judge’s decisionmaking process more complicated, and I find it entirely plausible that some (perhaps even most) judges would be inclined to resist it for that reason, among others.12 But what about the opposite—the switch from a highly discretionary standard to a hard-and-fast rule? Count me among those who suspect that at least some judges would be inclined to resist that shift as much (or even more) than the opposite type of change, either because they prefer to have more discretion as a general matter or because they think that this particular new rule leaves out too many relevant considerations.13 Whether lower court judges are likely to resist a new legal regime may, in short, depend on whether those judges’ preferences for rules versus standards more closely resemble those of Justice Antonin Scalia or Justice Stephen Breyer.

II.  When Is Resistance Most Likely?

For now, let’s bracket the challenges of defining “judicial resistance” and the appropriateness of using that particular term. Instead, let’s use a rough-and-ready “I know it when I see it” approach that defines “resistance” as any time the law has changed and it appears a judge got the new law “wrong” in a way that seems to reflect the influence of the old law. When is that sort of thing most likely to happen?

As I read Professor Tokson, he sensibly suggests that there are two general factors that tug in opposite directions: reasons not to follow the new law weighed against reasons to follow the new law.14 My main concern is with the second part of the analysis: the weight of the judge’s reasons for following the new law. It’s not that I disagree with the things that Tokson says. In particular, I agree that socialization, norm internalization, and a desire to avoid being reversed will tend to push judges toward compliance with new legal rules, as will the prospects for public criticism or high-profile sanctions for noncompliance.15

At the same time, I think that something is missing from Tokson’s analysis, something that could help explain at least some of his own examples: the severe procedural and practical limits on the availability of appellate review of many judicial decisions.

When I teach Civil Procedure, I tell the students that there may be few lawyers on earth with more unconstrained power than federal district court judges. That often strikes my students as counterintuitive. After all, trial judges are at the bottom of the Article III judicial hierarchy and, unlike appellate judges, their decisions are subject to appeals as of right rather than discretionary petitions for writs of certiorari.16

But trial courts actually have a lot of things going for them “in any struggle with their judicial superiors,”17 and some of those things seem directly relevant here. First, the final-judgment rules mean that appellate review is not even theoretically available in many cases, because the case settles before the trial court enters a final decision in favor of one party.18 Second, even if there is a final judgment in a case, only the losing party can appeal and may do so only if it is aggrieved by the decision it seeks to appeal.19 Third, even if an appeal is allowed, deferential standards of review make it difficult to overturn front-end factual findings or back-end exercises of discretion.20 Fourth, even if an appellate court agrees that an appealing party is right about the law, there is always the possibility that the reviewing court will deny relief on the ground that the error was harmless.21 Fifth, appeals require time and money, and any sensible litigant must weigh the costs of taking an appeal against the uncertain benefits of doing so. For all those reasons, many trial court rulings are, for all intents and purposes, effectively unreviewable.

So why does that matter? It matters because any sensible trial judge knows everything I just said and her knowledge of whether a particular action is likely to be subject to meaningful appellate review may well be a critical factor in determining whether and how vigorously that judge tries to resist a new legal command.

To see why, let’s walk through some of Tokson’s examples. If a party to a securities case believes that her opponent violated Federal Rule of Civil Procedure 11, she can move for sanctions. If she does not do so, is she even aggrieved for purposes of appealing if the trial court fails to comply with the provisions of the Private Securities Litigation Reform Act of 199522 (PSLRA), which require the trial judge to determine whether each party and each lawyer in the case have complied with Rule 11?23 Besides, how many lawyers who just won (or lost) a case on the merits have any incentive to ask an appellate court to order the trial court to make findings that will also result in a determination about whether they violated Rule 11 themselves?24 What are the odds, to use two of Tokson’s other examples,25 that giving the jury instructions at the wrong time or violating a rule that bars jurors from taking notes would not be held to be harmless error on appeal, or that a federal criminal sentence that is within (or near) the now-advisory Sentencing Guidelines would be held to be an abuse of discretion?26

Or take Tokson’s last and most developed example: the sequencing of qualified immunity determinations.27 From 2001 until 2009, lower courts were told that they had to decide whether there was a constitutional violation before determining whether the constitutional right in question was clearly established at the time the official acted.28 But what actually happened if the district court or the appellate court didn’t follow that sequence? What if the court instead assumed for the sake of argument that there was a constitutional violation but then ruled for the defendant on the ground that the right in question was not clearly established? The prevailing defendant would have little incentive to appeal, and it is unlikely that she would even be permitted to do so because of the adversity requirement.29 The losing plaintiff, in contrast, would be permitted to appeal. But she might well choose not to appeal if the lower court’s holding as to the “clearly established” requirement seemed unlikely to be reversed on appeal. And even if the plaintiff did appeal, there is a strong possibility that the trial court’s bottom-line judgment would be affirmed anyway.30

All of this, in turn, suggests a further point: The examples Tokson gives in Part III of his article strike me as fairly atypical in at least one respect. Nearly all of them involve situations in which a lower court ruling is likely to generate an appealable decision, either because the ruling in question is something that happens at or shortly before the entry of a final judgment, because of the special rules governing appeals of orders granting or denying injunctions, or because the ruling in question is subject to an immediate appeal under the collateral-order doctrine.31

If we want to find places where lower court judges are especially likely to resist legal change, maybe the places to look are those in which the initial decisions are unlikely to generate a reviewable decision in the first place. Take, for example, changes to the discovery rules or to the rules governing pretrial case management. Imagine a district judge who still follows the adage “No Spittin,’ No Cussin’ and No Summary Judgment.”32 For that matter, take one of the examples that Tokson discusses: the recent Supreme Court–initiated changes that aim to make it easier for defendants to win motions to dismiss for failure to state a claim.33 These sorts of changes seem to fit Tokson’s description to a T (low-visibility changes that could require judges to change deeply entrenched habits while creating more work and greater mental fatigue) while also occurring in situations in which effective appellate monitoring is particularly unlikely.

III.  What Exactly Is “Judicial Resistance” Anyway?

Professor Tokson uses several different words to label the phenomena that he examines. These are not technical terms. Rather, they are commonly used words that carry heavy normative baggage and generally connote at least some level of mens rea. Words like “resistance.”34 “Noncompliance.”35 Even “defiance.”36

In some situations, precise definitions may not matter very much. When a federal judge “attempted to reverse Brown [v Board of Education of Topeka] on the basis that blacks were not intelligent enough to attend school with whites,”37 we all can probably agree that the judge was engaged in resistance.

But other situations will be harder. Imagine a busy trial court judge who simply forgets that state law was recently changed to require courts to give juries written copies of all jury instructions,38 or a situation in which one party’s brief recites a since-superseded legal standard and the opponent’s brief does not call it out.39 There is a good chance that the new law is not going to be followed in either case, but is it really accurate to label what happened as “judicial resistance”?

Or what about situations in which there is room for good faith disagreement about what (if anything) the new law even requires? In 2004, for example, the Supreme Court abandoned its previous balancing approach to the Sixth Amendment’s Confrontation Clause and replaced it with a bright-line rule that asks whether the out-of-court statement that the prosecution seeks to admit is “testimonial” in nature.40 Not surprisingly, lower court judges disagreed about whether particular types of statements satisfied that standard, and the Supreme Court eventually held that some lower courts had given “testimonial” an unduly narrow construction.41 The lower courts were reversed, but was it because they were “resisting,” much less “defying,” the Supreme Court?

Finally, what about situations in which the whole point of the new regime is to give judges discretion about how to act in a particular situation? Return to Tokson’s qualified immunity example. Before 1991, the law was unclear about the desirability of following a particular procedure.42 From 1991 until 2001, the law “was not entirely clear” about whether or when that procedure was required.43 In 2001, the Supreme Court made it “unmistakable” that the procedure was mandatory in every case,44 before reversing course in 2009 and holding that judges “should . . . exercise their sound discretion in deciding” whether to follow the procedure.45 Tokson finds that, post-2009, lower court judges are still following the procedure more often than they were before 1991 or even between 1991 and 2001.46 But again, can we really say that the reason is “resistance” to the post-2009 rule—a rule that, by its terms, tells judges to do what they think is best in a particular case?

I don’t think this is just quibbling about semantics. If we are going to catalogue past examples of “judicial resistance”—let alone develop criteria for predicting and reducing future instances—we need to at least come up with a working definition of what “resistance” is. Yet, so far as I can tell, Tokson’s article offers no single definition of that all-important term.

Nor am I suggesting that defining “resistance” or coming up with an acceptable alternative term will be easy. Far from it. To the contrary, my own inability to do so has been one of the biggest reasons why, despite my own interests in the challenges posed by legal change and the relationship between courts at different levels of the judicial hierarchy,47 I have never attempted the task at which Tokson makes such a great start here.

A major challenge with defining “judicial resistance” is that our two most plausible options for doing so are deeply problematic, and they remain problematic whether we use them alone or in combination. Option 1: a subjective standard that focuses on intent. That is, we might define “judicial resistance” as encompassing situations in which a judge is aware that her conduct does not conform to the best understanding of current law but proceeds with that conduct anyway because of a preference for the old law. Option 2: an objective standard that focuses on outcomes. In other words, create a procedure for determining the “right” answer or method of analysis under the new law and then label as “judicial resistance” any situation in which (a) the new law isn’t followed and (b) what actually happens looks at least somewhat more like what used to happen under the old law.48

Both of these options have serious drawbacks. Option 1 (the subjective standard) has crippling problems of proof that render it useless in all but the most flagrant cases. Judges almost never admit that they are attempting to evade current law, and neither reviewing courts nor academics have ways of seeing into judges’ heads or making them admit what really motivated their decisions.49

But the problems that plague Option 2 (the objective standard) may actually be even worse. Judges may get the new law “wrong” for all sorts of reasons, ranging from conscious and deliberate defiance to genuine and entirely reasonable disagreement. If we are going to disclaim any attempt to figure out why a particular outcome happened, it seems that we should consider using a word whose connotations are a bit less loaded than “resistance,” much less “defiance.”

More fundamentally, however, a purely objective definition of resistance risks falling prey to the same culprit that haunts many attempts to study judicial behavior empirically: an inability to determine an unambiguously “correct” answer to most interesting legal problems, especially when the legal questions are (as they often are) tied up with a need to determine the underlying facts. To be sure, there will be exceptions, such as those in which the new rule says “Always do clearly defined thing X” and we are able to determine that judges are not in fact always doing X. Some of Tokson’s examples may fit that description, such as jury improvement rules about when jurors are to be instructed or whether they are permitted to take notes,50 or the provisions of the PSLRA that require district court judges to make certain findings in all covered cases.51 But Tokson is not content to limit his examples of “resistance” to what he labels “[d]irect noncompliance”52 or “overt noncompliance.”53 And one need not be a legal realist of the what-the-judge-ate-for-breakfast variety to worry that there will be many situations in which it will be difficult (if not impossible) to achieve universal agreement about the “right” answer or analysis for a given case.54

I’m not sure there is a way to untangle this Gordian knot. So let’s propose a more radical solution, one that never occurred to me until I read Tokson’s article: abandon use of the term “resistance” or any attempt to determine whether a particular situation is an example of it.55 As Tokson reminds us, the end goal here isn’t to crush resistance for its own sake but rather to change what happens in the real world by “chang[ing] lower court practices.”56 So the real question, it seems to me, is how to identify both the things that may make changing actual on-the-ground practices unusually difficult and the strategies for how we deal with those difficulties. Tokson’s article does both things, and I think that his analysis takes a huge step in the right direction.


Judicial Resistance and Legal Change tackles a topic that is too often overlooked. It was fun to read. It made me think. And it made me want to have a long conversation with the author. What’s not to like?

           David H. Ibbeken ’71 Research Professor of Law, University of Virginia School of Law.

      1     See Matthew Tokson, Judicial Resistance and Legal Change, 82 U Chi L Rev 901, 901 (2015).

      2     Id at 929.

      3     Bradley C. Canon and Charles A. Johnson, Judicial Policies: Implementation and Impact 1 (CQ 2d ed 1998).

      4     Tokson, 82 U Chi L Rev at 973 (cited in note 1).

      5     This all, of course, requires a method for assessing whether judges are “correctly” implementing a new legal regime. I discuss this point in Part III.

      6     Tokson, 82 U Chi L Rev at 903 (cited in note 1).

      7     See id at 925.

      8     Id at 902.

      9     See id at 911–16.

      10   See Tokson, 82 U Chi L Rev at 923–24 (cited in note 1).

      11   See id at 929.

      12   See id at 912.

      13   See id at 967 (discussing scholarship that contends that judges prefer more discretion to less).

      14   See Tokson, 82 U Chi L Rev at 927–30 (cited in note 1).

      15   See id.

      16   Compare 28 USC § 1291, with 28 USC § 1257. Of course, a sensible appellate judge knows that, except in the rare case, the Supreme Court is almost entirely uninterested in fact-bound error correction. See, for example, US S Ct Rule 10 (“A petition for a writ of certiorari is rarely granted when the asserted error consists of erroneous factual findings or the misapplication of a properly stated rule of law.”).

      17   Toby J. Heytens, Doctrine Formulation and Distrust, 83 Notre Dame L Rev 2045, 2072 (2008).

      18   See id at 2068.

      19   See id.

      20   See id. See also Toby J. Heytens, Reassignment, 66 Stan L Rev 1, 34 (2014).

      21   See Chapman v California, 386 US 18, 22 (1967).

      22   Pub L No 104-67, 109 Stat 737, codified as amended in various sections of Title 15.

      23   See Tokson, 82 U Chi L Rev at 936–40 (cited in note 1).

      24   See id at 937–38 (observing that the PSLRA requires district courts to make Rule 11 compliance findings for “every attorney . . . for every pleading or dispositive motion that the parties file”).

      25   See id at 935.

      26   See id at 950 (noting that “appeals courts have signaled that they will affirm nearly all sentences that fall within a Guidelines range”). See also id at 944–52 (discussing the rise and fall of the Sentencing Guidelines and the standard of appellate review for decisions made pursuant to them).

      27   See Tokson, 82 U Chi L Rev at 952–61 (cited in note 1).

      28   See id at 955–56.

      29   In Camreta v Greene, the Supreme Court held that it has the power to hear a defendant’s challenge to an appellate court decision that finds a constitutional violation but grants judgment to the defendant on the ground that the law was not clearly established. Camreta v Greene, 131 S Ct 2020, 2031–32 (2011). However, the Court specifically reserved the question whether a court of appeals may do the same. See id at 2033. Further, the Court did not address the situation in which a lower court simply assumes without deciding that there was a constitutional violation.

      30   See Theodore Eisenberg, Appeal Rates and Outcomes in Tried and Nontried Cases: Further Exploration of Anti-plaintiff Appellate Outcomes, 1 J Empirical Legal Stud 659, 664 (2004).

      31   See 28 USC § 1291 (providing for appeals from a final decision); 28 USC § 1291(a)(2) (providing for appeals from the grant or denial of injunctions); Mitchell v Forsyth, 472 US 511, 524–30 (1985) (holding that a district court’s denial of qualified immunity is immediately appealable).

      32   Susan Taylor Wall, “No Spittin,’ No Cussin’ and No Summary Judgment:” Rethinking Motion Practice, 8 SC Law 29, 29 (May/June 1997) (quoting “[a] sign prominently displayed in [ ] a rural Alabama courthouse”).

      33   See Tokson, 82 U Chi L Rev at 970 (cited in note 1).

      34   Id at 904.

      35   Id at 903.

      36   Id at 904.

      37   Tokson, 82 U Chi L Rev at 910 (cited in note 1).

      38   See id at 936.

      39   See id at 932–34.

      40   See Crawford v Washington, 541 US 36, 68 (2004).

      41   See Melendez-Diaz v Massachusetts, 557 US 305, 307, 310 (2009) (holding that “affidavits reporting the results of a forensic analysis which showed that material seized by the police and connected to the defendant was cocaine” were “testimonial” and thus covered by the Confrontation Clause).

      42   See Tokson, 82 U Chi L Rev at 953 (cited in note 1).

      43   Id (discussing Siegert v Gilley, 500 US 226 (1991)).

      44   Tokson, 82 U Chi L Rev at 954 (cited in note 1) (discussing Saucier v Katz, 533 US 194 (2001)).

      45   Pearson v Callahan, 555 US 223, 236 (2009).

      46   Tokson, 82 U Chi L Rev at 957–59 (cited in note 1).

      47   See generally Heytens, 66 Stan L Rev 1 (cited in note 20); Toby J. Heytens, The Framework(s) of Legal Change, 97 Cornell L Rev 595 (2012); Heytens, 83 Notre Dame L Rev 2045 (cited in note 17); Toby J. Heytens, Managing Transitional Moments in Criminal Cases, 115 Yale L J 922 (2006).

      48   Tokson seems to come close to embracing this conception of resistance when he discusses reasons why judges may inadvertently or even unknowingly reach the “wrong” result in a particular case. See Tokson, 82 U Chi L Rev at 919–20 (cited in note 1).

      49   See id at 930–31 (noting that “judicial noncompliance is most likely to occur in situations in which it is difficult to detect”).

      50   See id at 935–36.

      51   See id at 936–40.

      52   Tokson, 82 U Chi L Rev at 909 (cited in note 1).

      53   Id.

      54  The problem gets worse, not better, if we decide to reserve the label “judicial resistance” for those decisions that are “clearly wrong” under current law. At that point, we don’t just need to figure out whether a particular decision was “wrong”; we also need a way of identifying whether the “clearly” standard is satisfied in a particular case.

      55   Bracket, for a moment, what to do about cases in which a judge admits that she is refusing to follow her own best understanding of current law.

      56   Tokson, 82 U Chi L Rev at 969 (cited in note 1).