You know the line about a bad restaurant: “The food is terrible. And the portions are so small!” Well, this is the blog version of that restaurant. It’s not a good post. And it’s so long.
Here’s my excuse: This post is about a currently pending case in the Washington Supreme Court that seeks to set a rule of stare decisis for the state’s intermediate court of appeals. So things get abstract pretty quickly. This is a case about the law about law about law. Yup, you read that right: It’s meta-meta-law. First we’ve got the substantive law: the three-part test, the scienter requirement, the proper jury instruction, the meaning of the statutory term, &c. That’s the law. And then there’s stare decisis—i.e., the law about that law. A court might disagree with a prior decision on the substantive law. The applicable rule of stare decisis tells the court whether or when the court gets to depart from that substantive law. But who determines the relevant rule of stare decisis? And on what basis is that rule determined? That’s the law about the law about the law. And it gets a bit messy.
But first, some background…
One Court of Appeals with Three Divisions
We have an interesting Court of Appeals here in Washington. Back in 1969, our legislature “established a court of appeals.” RCW 2.06.010. The state constitution similarly authorizes the creation of a single intermediate appellate court. See Wash. Const. art. IV, sec. 30. This appellate court is “vested with all power and authority . . . necessary to carry into complete execution all of its judgments . . . according to the rules and principles of the common law and the Constitution and laws of [Washington].” RCW 2.06.030.
But this unitary court of appeals has “three divisions,” divided geographically, with Division 1 drawings its judges from the northwest part of the state (including Seattle), Division 2 drawing from the southwest part of the state (including Tacoma and the Olympic Peninsula), and Division 3 drawing from the eastern part of the state (including Spokane and Yakima). RCW 2.06.020.
As far as I can tell, nothing in the Constitution or legislative enactments actually limits the geographic jurisdiction of the Divisions of the Court of Appeals. The legislature mandates that the Divisions be “headquartered” in certain cities, see RCW 2.06.020, and that the judges in each Division come from specific counties within the Division, see id. But nothing in the enacting statutes or the Constitution actually expresses what is obviously implied: that appeals from the counties within a Division should actually go to the Division that draws judges from that county.
If anything, the legislature seems to go out of its way to downplay distinctions among the three Divisions of the unitary Court of Appeals. “Judges of the respective divisions may sit in other divisions and causes may be transferred between divisions, as directed by written order of the chief justice.” RCW 2.06.040. Without divisional distinctions, the legislature states: “The court shall sit in panels of three judges.” RCW 2.06.040. And the legislature allows that “the supreme court may . . . transfer to the appropriate division of the court for decision a case or appeal pending before the supreme court.” RCW 2.06.030. But what statute determines the “appropriate” Division?
The court-promulgated Rules of Appellate Procedure fill this vacuum. Unlike their federal counterparts, Washington rules of procedure are enacted by the judicial branch without meddling by the legislative or executive branches. We take our separation of powers (and our procedure) seriously. (Fun fact: Under Washington law, if a statute creates a rule of procedure that conflicts with a court-promulgated rule, the court-promulgated rule wins. Take that Erie doctrine!) Anyway, the RAPs expressly require a party to seek review in the Division that covers the trial court that issued the appealable order. See RAP 4.1. So in practice, the different Divisions of the Court of Appeals operate somewhat like different circuits in the federal system, albeit within a unitary court with the ability to administratively transfer cases among the Divisions, something the feds lack.
Trust me. This all matters later. (I think.)
Stare Decisis and In re Arnold
But first, the case that brought all this to a boil: Last week, the Washington Supreme Court heard arguments in In re Arnold. [Docket here (look under January 11). Video of oral argument here.] The substantive question in Arnold is whether Washington law required Mr. Arnold to register as a sex offender. But in granting review, the Supreme Court wanted to hear from the parties on a second question: Did the Court of Appeals err by applying the principle of “horizontal stare decisis” to previous decisions of the Court of Appeals?
Arnold was argued in Division 3 of the Court of Appeals. But Division 1 and Division 2 had already resolved the underlying registration question. In a remarkable decision, in which the three-judge Arnold panel issued four separate opinions, Division 3 concluded it would apply horizontal stare decisis to the prior decisions of Division 1 and Division 2. Two of the judges remained agnostic on the merits of the prior decisions, but reasoned that Division 3 should follow Division 1 and Division 2 to avoid creating problems of inconsistency in the state’s application of criminal statutes and sex-offender registration. The lone dissenting judge assumed that horizontal stare decisis compelled some allegiance to the prior decisions. But this third judge applied the Washington Supreme Court’s own test for doing away with prior precedent, concluding that the Division 3 panel should “correct” the prior Division 1 and Division 2 decisions because they were “incorrect and harmful.” See In re Arnold, 396 P.3d 375 (Wash. Ct. App. 2017).
Two points here. First, Division 3 doesn’t really have the power to “correct” prior decisions of Division 1 or Division 2. Unlike the Supreme Court’s doctrine of horizontal stare decisis, an “overruling” by Division 3 wouldn’t necessarily control future cases in Division 1 or Division 2. Judges in those other Divisions can just keep applying the prior Division precedent.
Second, all of this is a mess for trial court judges. Trial court judges are supposed to follow the Court of Appeals. And though an appeal is likely go to go the Division that includes their county, the transfer rules mean that’s not a certainty. Accordingly, disagreements among the Divisions of the Court of Appeals cause havoc on well-meaning trial judges. What’s a judge in Seattle (within Division 1) to do if a panel of three judges in Division 3 claims to “correct” or “overrule” a prior Division 1 decision? Yikes.
But wait, there’s more. In its order granting review, the Supreme Court put the question this way:
The question here is whether a geographical division of the Court of Appeals must follow a decision of another geographical division. . . . [T]he traditional rule followed by our courts, including a previous Division Three panel, is that the decision of another geographic division is not binding any other division.
That’s certainly the situation in Arnold, with Division 3 looking at precedent from Division 1 and Division 2. But it’s not immediately clear why Division 3 would give differing precedential force to a prior opinion depending on if it came from Division 3 or any other Division. The Court’s order granting review elides this issue by observing: “Nothing in the enabling statute or the constitution suggests that one of the geographic divisions of the court is bound to follow the decisions of other geographic divisions of the court.” That’s true. But of course nothing in the enabling statutes or the Constitution says anything about intra-Division stare decisis either. Nothing says anything about any kind of stare decisis.
[Side note: The government’s brief includes the following topic heading: This Court Should Hold that the Washington Courts of Appeals Is to give Deference to the Decisions of Other Panels of the Court but Is Not Strictly Bound by Horizontal Stare Decisis. Given the circumstances, I don’t even consider that a typo! We have multiple “Courts” of Appeals but they are indeed a singular court.]
Oral argument brought these uncertainties into stark relief. In response to questioning, counsel for the government asserted that the principles of stare decisis in the Court of Appeals “should be the same for intra-divisional conflict or conflict between the divisions.” (Oral Argument at 10:45.) In other words, the government’s position was that a panel of the Court of Appeals should give “deference” but not be bound by a previous panel decision, regardless of the Division that issued the previous opinion. The Court of Appeals would have a weak but uniform “deference”-based system of stare decisis among all panels of all Divisions.
Later, the Justices seemed to disagree on the obligations of trial court judges. Justice Gonzalez posited that a trial judge faced with conflicting holdings from different Divisions of the Court of Appeals need not follow the Division in which the trial court sits; instead, she can pick “the most correct one.” Counsel for the government agreed. (Oral Argument at 12:20.) But then Chief Justice Fairhurst expressed a different understanding: “I had always thought that the courts in the counties did need to follow the Court of Appeals decision of the Division in which they sat. Is that incorrect? Or is there authority which specifically says that they don’t have to?” The government had no authority either way for that proposition. The Chief Judge continued: “My own experience was that you are bound by the decisions of your own division.” (Oral Argument at 12:50.)
How Do We Get Out of This?
So you can see we have a bit of a mess. The briefing and the oral argument do a nice job of working through the various benefits of different standards of stare decisis. If you really want to get into the weeds, I highly recommend Mark DeForrest’s article, In the Groove or in a Rut? Resolving Conflicts Between the Divisions of the Washington State Court of Appeals at the Trial Court Level, 48 Gonzaga L. Rev. 455 (2013). It’s great. (Sorry for the lack of a link! The original one seems dead.)
But I want to put aside the content of the rule of stare decisis. Here what I can’t get my head around: How does the Supreme Court resolve any of this? In other words, what does it even mean for the Court of Appeals to have been “wrong” about the doctrine of inter-Division stare decisis? And what is the basis for the Supreme Court’s power to correct a supposedly wrongful application of horizontal stare decisis?
An intermediate court’s view of the precedential effect of its own prior decisions seems to me, at least at first blush, like a question the Supreme Court wouldn’t have much interest in. Imagine the following example: A panel of Division 1 gets a case. There appears to be binding Division 1 authority from fifty years ago. The current panel “distinguishes” the prior case by saying that modern technology makes it factually distinguishable. If the Supreme Court were to grant review, would the Court ever reverse or vacate on the ground that the Court of Appeals didn’t properly conceptualize the holding of the fifty-year-old Court of Appeals case? Would it ever remand with instructions to treat the old Division 1 opinion as controlling? I’d be shocked by such a decision. Instead, if the Court were to grant review, it would just reach the merits of the question.
What’s the difference in Arnold? Sure, the Court could remand with instructions to use a different concept of stare decisis for decisions from other Divisions of the Court of Appeals, but why? And, more importantly, what would the legal basis be for such a remand?
I can think of a few answers. First, perhaps the Court of Appeals’ treatment of prior precedent violated a statute. If, for example, a state statute created entirely separate courts of appeals (like the federal system), and if Division 3’s opinion indicated that the panel erroneously read the statute to require Division 3 to follow prior decisions from other Divisions, then I could perhaps see the Supreme Court simply correcting that statutory error and remanding for further proceedings.
Second, perhaps the Court of Appeals’ treatment of prior precedent violated a court-promulgated procedural rule. There, too, the Supreme Court could grant review to correct the interpretive error (this time of a rule instead of a statute) and then remand for further proceedings.
The Supreme Court’s order granting review nods at this second possibility by referencing Washington Rule of Appellate Procedure 13.4(b)(2). The rule states: “A petition for review will be accepted by the Supreme Court . . . [i]f the decision of the Court of Appeals is in conflict with a published decision of the Court of Appeals.” The rule does not mention Divisions. And though it envisions that conflicts among panels may arise, the rule does not compel any particular reason for those conflicts. Inter-divisional splits? Overly aggressive attempts to distinguish past precedent? Failure to address a relevant line of precedent? Improper overruling of prior precedent? Sparring factions within a Division? Nothing in the rule itself sheds light on, much less mandates, the proper level of deference between panels or among Divisions.
But neither the statute-based or rule-based objections seem to be what’s fueling the government’s or the Supreme Court’s ire in Arnold. The government’s brief doesn’t claim that any rule or statute was violated by the Court of Appeals’ application of stare decisis in Arnold. Instead, the government just notes that Court of Appeals precedent on interdivisional stare decisis seems a bit muddled, before providing examples of how the federal courts of appeals and some other state intermediate courts deal with the issue. The government’s brief is heavy on recommendations for what the Supreme Court “should” require and what “may be workable” for the Court of Appeals. But the brief is silent on the legal basis for the Supreme Court to require any particular system.
The Supreme Court’s order granting review is similarly sparse on the basis for its decision. Apart from the brief mention of RAP 13.4(b), the order claims that “Arnold wholly reinvented the traditional duties of a Court of Appeals division.” This “watershed departure from prior practice” apparently calls for intervention. Like the government’s brief, the Supreme Court’s order walks through prior Court of Appeals cases that demonstrate differing levels of deference to prior Court of Appeals decisions. And the order then argues that Arnold departed from past Court of Appeals practice.
[Side note: I think the Supreme Court overstates the claim that Arnold departs from past practice. Contrary to the Supreme Court’s order, the Arnold panel did not claim to be bound by prior Division 1 and Division 2 precedent. Rather, the Court of Appeals specifically recognized that it was “not required” to follow the prior decisions of other Divisions. Arnold, 396 P.3d at 378. That “mandatory” obligation of vertical stare decisis is reserved for decisions of the Supreme Court. Id. But the Court of Appeals recognized that a weaker horizontal stare decisis applied to decisions of previous panels. It specifically declined “to resolve the question of exactly how stare decisis applies” in the context of different panels from different Divisions. Id. But on the specific facts of Arnold, the Court of Appeals decided that departing from previous Division 1 and Division 2 precedent would leave the underlying law “impermissibly vague,” since Division 3 could not overrule those prior decisions. Id. According to the Court of Appeals, that vagueness was unacceptably harmful to a statewide system of sex-offender registration. Id. at 379. There’s no inconsistency between [a] the Court of Appeals asserting that it is not bound by a prior decision and therefore declining to follow it, and [b] the Court of Appeals choosing to follow a prior decision to avoid creating a potentially unconstitutionally vague and unworkable system of statewide registration.]
But in any event, the Supreme Court’s order fails to assert a clear basis for either  the Supreme Court inserting itself in the Court of Appeals’ efforts to work out a system of horizontal stare decisis, or  mandating any particular system. Sure, this issue is one of public interest. But the public is interested in a lot of issues. That doesn’t mean the Supreme Court is well situated to resolve them.
Consider the questions the Court has endeavored to answer in Arnold: What level of deference should one panel of the Court of Appeals give to the decisions of prior panels? And does that level of deference depend on whether the prior decision was from the same Division? The potential answers are many:
- Weak “deference” to prior decisions regardless of Division.
- Weak “deference” to prior decisions from other Divisions; strict “mandatory” stare decisis for prior decisions from the same Division.
- Weak “deference” to prior decisions from other Divisions; can overrule prior decisions from the same Division using the Supreme Court’s “incorrect and harmful” formulation for overruling past precedent.
- Weak “deference” to prior decisions from other Divisions; can overrule prior decisions from the same Division using some other standard for overruling.
- Strict “mandatory” stare decisis for prior decisions regardless of Division.
- Strict “mandatory” stare decisis for prior decisions from the same Division; can reject prior decisions from other Divisions using the Supreme Court’s “incorrect and harmful” formulation for overruling past precedent.
- Strict “mandatory” stare decisis for prior decisions from the same Division; can reject prior decisions from other Divisions using some other standard for overruling.
- . . . . &c.
The list goes on. Consider, also, that the Court would likely need to address accompanying rules for trial courts faced with potentially conflicting Court of Appeals decisions. Those rules would, of course, have to account for the fact that cases can be transferred among the three Divisions.
My point isn’t just that the Supreme Court has a lot of options to wade through. Rather, my point is this: No statute or court rule seems to provide any guidance for choosing the correct system from among these options. And without a basis for determining a “correct” system, I’m left searching for a principled basis for the Supreme Court to render a decision.
I think I’m left searching because this absence of clarity isn’t a legal problem; it’s a structural problem. The structure and procedures for deciding cases in the Washington Court of Appeals just aren’t conducive to any well-functioning or well-defined system of horizontal stare decisis. We have trial courts with unclear obligations to different Divisions within a unitary Court of Appeals. We have trial court decisions getting appealed generally to the Division in which they sit, but not always. We have the ability to transfer cases among the Divisions. We have no en banc panels within or among Divisions. And we have no mechanism for panels in one Division to review or overrule the decisions by panels in other Divisions. The structure calls for trial courts to follow all the Divisions (because of the unitary court and because they can’t trust where the appeal will end up) but the structure of the Divisions doesn’t allow (or perhaps permit?) them to maintain consistency with each other.
Compare the Washington system with the federal system: Like Washington, the federal system doesn’t have clear statutes or other external rules of stare decisis. But the federal system has a clearly defined court structure. There’s no overlap among the various federal circuit courts of appeals. They don’t need to follow each other because they have no interaction with each other. They are statutorily separate, not unitary. District courts absolutely know the circuit precedent they need to follow. Appeals from the Western District of Washington go to the Ninth Circuit. All the three-judge panels, regardless of location, speak for the court of appeals and therefore bind all district courts in the circuit. And the judges rotate freely among panels in different locations with all the other judges. Within a single circuit, the panels know that if they break with a previous panel decision, the other judges on the court can demand a modification or threaten to take the case en banc.
These clear structures allow the federal courts of appeals to develop their own rules of intra-circuit stare decisis. Across the circuits, each federal court of appeals can develop its own practices for en banc review. The judges can discuss these things among themselves and formulate a sort of common law of stare decisis for the circuit. These internal procedures work because the structural lines are clear and allow for the development of workable rules. And the district courts know which procedures their court of appeals follows.
Joseph Mead has written a great article on this issue: Stare Decisis in the Inferior Courts of the United States, 12 Nevada L.J. 787 (2012). Here’s what Professor Mead says about the creation of stare decisis rules in the federal courts:
Notably, circuit courts adopted their own policies without Supreme Court intervention . . . . This reflects a general principle that stare decisis is a decision that is best left to the court at issue, within broad bounds of reasonableness. Each court is uniquely qualified to decide what level of deference to give its own opinions. Each court understands the practicalities and politics present . . . and can weigh the policy considerations accordingly.
And that’s how things have worked in the federal system—not through top-down rulings from the Supreme Court or from acts of Congress, but with each court figuring out its own rules. And (perhaps surprisingly) each court of appeals has slightly different practices for following prior panel decisions or going en banc. But the system seems to work because the structure of the separate courts and their power of review over district courts are both clear.
We don’t have that clarity in Washington. And that lack of structural clarity means the Court of Appeals has not been able to create for itself clear rules of intra- and inter-Division stare decisis. Nor has the Court of Appeals developed clear rules for what opinions from which panels bind lower courts in what counties.
And that’s why I think the Supreme Court in Arnold is stuck. It can’t change the statute that creates a unitary Court of Appeals. It can’t change the rule that allows transfers among the Divisions. It can’t ensure that appeals from Seattle Superior Courts will always go to Division 1. And it can’t create a system for three-judge panels to “overrule” other three-judge panels or a procedure for en banc review. All the Supreme Court can do is create a top-down rule without a legal basis or a structural method of implementation.
So what should the Court do? My recommendation would be two-fold. First, I think the Supreme Court should decline to resolve the stare decisis question in Arnold. It should instead render that question moot by deciding the underlying substantive question. Second, the Supreme Court should ask for a statutory or rule-based change in the structure of the courts of appeals, perhaps by creating an en banc panel among the Divisions or eliminating the ability of cases to be transferred among Divisions. That job would fall to the legislature or to the Court itself in its rulemaking capacity. Whatever rules get implemented, this whole experiment with one court that’s really three courts seems to need a bit of an adjustment. But I don’t think one decision in one case can do the trick.