When a Wisconsin court deems a state statute unconstitutional, it enjoins the government from enforcing the statute. Even if the court makes that determination in the form of a temporary injunction, before full consideration of the merits or issuance of a final judgment, the government has an immediate right to appellate review.1 In such cases, the government frequently asks that the injunction be stayed—that is, prevented from taking effect—pending resolution of the appeal.2
Wisconsin law on stays pending appeal is fairly sparse. Earlier this year, however, the Wisconsin Court of Appeals (District III) and the Wisconsin Supreme Court decided preliminary motions related to stays in two cases. The resulting orders, though unsigned and unpublished, suggest a substantial shift in the law, at least in cases that involve injunctions against enforcement of an allegedly unconstitutional statute. The orders in these cases raise three questions:
First, to what extent are preliminary procedural orders by Wisconsin appellate courts precedential?
Second, how do these orders depart from prior law on stays pending appeal?
Third, do these orders essentially deprive circuit courts of the power to enjoin unconstitutional statutes?
This article explores these questions. Before doing so, it provides an overview of Wisconsin law on stays pending appeal (Section I) and a procedural history of the two cases in which the appellate courts issued the orders considered here (Section II). This article then addresses each of the above questions in turn (Sections III–V).
The seminal case on stays pending appeal is State v. Gudenschwager,3 a per curiam decision staying a circuit court’s order requiring release of a convicted sex offender who had completed his criminal sentence and was awaiting a hearing on the State’s petition for involuntary commitment.4
A. The Gudenschwager Framework to Evaluate Motions for Stays Pending Appeal
A stay pending appeal is appropriate where the moving party:
- makes a strong showing that it is likely to succeed on the merits of the appeal;
- shows that, unless a stay is granted, it will suffer irreparable injury;
- shows that no substantial harm will come to other interested parties; and
- shows that a stay will do no harm to the public interest.5
Gudenschwager did not invent these four factors. It lifted them from the court of appeals’ decision in Leggett v. Leggett,6 which in turn took them from a decision of the federal Eighth Circuit Court of Appeals.7 Gudenschwager adopted verbatim the Eighth Circuit factors, but, in applying them, relied heavily upon a case from the Sixth Circuit, without recognizing that the Sixth and Eighth Circuits articulate the factors differently.8
Before expanding upon each factor, Gudenschwager notes that the four “factors are not prerequisites but rather are interrelated considerations that must be balanced together.”9
With respect to the first factor, likelihood of success, Gudenschwager explains that “a movant need not always establish a high probability of success on the merits.”10 Instead, “the probability of success that must be demonstrated is inversely proportional to the amount of irreparable injury the [movant] will suffer absent the stay. In other words, more of one factor excuses less of the other.”11
With respect to the second factor, irreparable injury to the movant, Gudenschwager notes that any “harm alleged must be evaluated in terms of its substantiality, the likelihood of its occurrence, and the proof provided by the movant.”12 The court holds that the State satisfied those requirements in providing expert evidence supporting the circuit court’s finding “that there was probable cause to believe Gudenschwager was a sexually violent person and that it was substantially probable that Gudenschwager would engage in additional acts of sexual violence.”13
Gudenschwager has less to say about the third and fourth factors, harms to others and to the public, respectively. The court defines the public interest as protection “from further acts of sexual violence.”14 That interest favored a stay because the only way to substantially reduce the risk of Gudenschwager reoffending after release was placement in a residential sexual offender program, which had no openings at the time.15 With respect to harm to others, the court acknowledges that “Gudenschwager will suffer a loss of liberty if he is retained in custody pending appeal.”16 While there was “no doubt but that he will suffer harm if his incarceration is continued,” the court says the other factors “outweigh Gudenschwager’s interest in being released from custody pending appeal.”17
B. Uncertainties About the Gudenschwager Framework
Wisconsin appellate courts have discussed neither the extent to which Gudenschwager’s subject matter may have distorted its analysis, nor the doctrinal and practical tensions within the Gudenschwager opinion.
As a doctrinal matter, Gudenschwager says that the four factors it sets forth are “not prerequisites” but instead “must be balanced together.”18 It then goes on to explain how factors (1) and (2)—the movant’s likelihood of success on the merits of the appeal and any irreparable injury the movant will suffer absent a stay—are inversely related.19 But there is an unacknowledged tension—if not contradiction—buried within. The discussion of the inverse relationship between factors (1) and (2) is exclusive—“more of one factor excuses less of the other”20—and does not extend to factors (3) and (4). That makes sense, because Gudenschwager expressly defines factors (3) and (4)—harm to others and the public interest—as being satisfied only by a showing of “no harm.”21 Factors satisfied only by a complete lack of harm, however, are requirements, not considerations to be weighed against others. Indeed, the Sixth Circuit case Gudenschwager cited as an example of applying the factors underscores this point; it utilizes the balancing approach but accordingly eschews the “no harm” framing for the third and fourth factors.22
In practical terms, Gudenschwager’s analysis retreats from the plain language it uses in identifying the factors. Not only does the opinion teach that the factors should be balanced, but it also grants a stay after expressly recognizing that the State could not satisfy factor (3). The court acknowledges and “do[es] not mean to deprecate” Gudenschwager’s “loss of freedom.”23 Nonetheless, the court concludes, “on balance, we believe that the [other] three factors … outweigh Gudenschwager’s interest in being released from custody pending appeal.”24 All of this is inconsistent with the plain text of the factors the court has adopted.
These tensions notwithstanding, Wisconsin appellate courts have continued to follow Gudenschwager in considering requests to stay injunctions pending appeal.25
Last December, after the election of new state officers, the Legislature convened what it calls an “extraordinary session.” During that brief session, the Legislature passed—and then-Governor Scott Walker subsequently signed into law—three bills that made numerous changes to how power is allocated between the Legislature, on one hand, and the Governor, the Attorney General, and Executive Agencies on the other.26
Two of the lawsuits challenging those laws are relevant here. In one, League of Women Voters v. Evers,27 the League of Women Voters of Wisconsin (LWV) and other plaintiffs asserted that the Legislature violated two provisions of the Wisconsin Constitution by convening as it did, and, for that reason, the resulting laws were invalid.28 (Full disclosure: I represented the plaintiffs in that case.) In the other, the Service Employees International Union (SEIU) and other plaintiffs challenge specific aspects of the new laws as violating the separation-of-powers principles enshrined in the Wisconsin Constitution.29
In LWV, the Dane County Circuit Court (Niess, J.) issued a temporary injunction on March 21, 2019.30 As part of that ruling, the circuit court considered and denied the Legislature’s request to stay the injunction.31 The Legislature, which had intervened in the case to defend its actions during the December 2018 Extraordinary Session, immediately appealed and sought a stay. After expedited briefing, the court of appeals entered the stay on March 27, 2019. In entering the stay, the court of appeals issued a nine-page order that is unsigned (but not labeled per curiam), and unpublished; it is available on Westlaw.32
In the time between the circuit court’s injunctive order and the court of appeals’ stay order, Governor Evers had taken action that the Legislature wanted undone.33 It sought relief, which the court of appeals denied.34 The Legislature then invoked a panoply of procedural mechanisms to ask the supreme court for relief.35 After granting a petition to take jurisdiction over the appeal on bypass, the supreme court granted the Legislature relief under section 808.07 of the Wisconsin Statutes—a mechanism that the Legislature had not invoked—on April 30, 2019.36 The ten pages containing the majority’s rationale are unsigned.37 They are followed by a dissenting opinion, written by Justice A.W. Bradley and joined by Justices Abrahamson and Dallett.38 This order is not published, and is not available on the supreme court’s website or on Westlaw.
In SEIU, the Dane County Circuit Court (Remington, J.) granted in part and denied in part plaintiffs’ motion for a temporary injunction on March 26, 2019.39 As part of that order, the circuit court considered and denied the legislative defendants’ request to stay the injunction.40 The legislative defendants appealed and sought a stay, though they did not style their motion as one for emergency relief.41 The stay motion was fully briefed in the court of appeals when the supreme court sua sponte exercised bypass jurisdiction over the case.42 It subsequently granted a stay on June 11, 2019.43 The nine pages containing the majority’s rationale are unsigned.44 They are followed by an opinion concurring in part and dissenting in part, written by Justice Dallett and joined by Justices Abrahamson and A.W. Bradley.45 This order is not published, and is not available on the supreme court’s website or on Westlaw.
The threshold question in discussing any of these orders is whether they have precedential value. They are preliminary procedural orders, not final adjudications on the merits. They were informed by briefing, but no oral arguments were held. All three were unsigned and unpublished (though none was labeled per curiam). By contrast, Gudenschwager was issued after both briefing and oral argument. It was expressly labeled a per curiam decision, and it was published. After the Gudenschwager decision, the case went back to the court of appeals for a merits determination. In LWV and SEIU, the stay decisions were merely preludes to subsequent merits determinations at the supreme court. (LWV was decided on the merits in June;46 SEIU is scheduled for argument on October 21, 2019.)
Under section 809.23(3)(b) of the Wisconsin Rules of Appellate Procedure, the court of appeals decision granting the stay in LWV is clearly not precedential. The opinion is unpublished and unsigned. Though it is not expressly designated per curiam, it falls within the catch-all of the statutory explanation that a “memorandum opinion, summary disposition order, or other order is not an authored opinion for purposes of this subsection.”47
The status of the supreme court’s orders in LWV and SEIU is less clear, but the supreme court itself evidently understands the orders as precedential. In SEIU, the stay motion was fully briefed in the court of appeals before the supreme court took jurisdiction over the case.48 After the supreme court issued its order granting relief in LWV, it instructed the parties in SEIU to file supplemental briefs addressing the effect of that decision on the pending request for a stay in that case.49 That is, the supreme court treated the LWV Order as a precedential decision that should be considered, alongside other legal authority, in deciding whether to stay the SEIU injunction. Then, when the supreme court issued its order staying the SEIU injunction, that order relied on the reasoning of the LWV Order, treating it as a precedential decision.50
Nor is it only the majority that sees precedential value in these orders. The dissent in the LWV Order complained that “it appears that the majority substantively alters existing law.”51 So too in the SEIU Order, where the dissent asserts that “[t]he majority order appears to alter substantive law.”52 To “alter” the law, these orders must be understood as having precedential effect beyond the stays at issue. Otherwise, the dissents would complain that the majority’s approach misapplied or distorted the governing law in deciding the issue at hand. By insisting instead that the orders change the law, the dissenting Justices underscore the majority’s assumption that the orders have precedential effect.
Understanding that the supreme court sees these orders as precedential, two consequences seem to follow.
First, parties engaged in future litigation over stays pending appeals may want to cite these orders. Because the supreme court’s consideration may include the reasoning in these orders, parties wishing to persuade the court of their position should have the opportunity to address that reasoning and its application to their case. Given that the opinions are not widely available, how exactly parties can easily do so, especially in the context of expedited briefing that may be more common for briefing on stays than for other matters, is not clear.53
Second, to the extent that the reasoning in these orders departs from Gudenschwager, that seems to evidence changes in the law. What those changes might be, and what they mean, is explored in the next section.
The supreme court’s orders in LWV and SEIU depart from the Gudenschwager standard. If they have precedential value, they have sub silentio remade the law. Even if they do not have precedential value, they are useful indications of how a majority of the current Justices think about stay applications.
The first factor a party seeking to stay an injunction pending appeal must satisfy is making “a strong showing that it is likely to succeed on the merits of the appeal.”54 Almost as soon as it adopted that standard, Gudenschwager backed off, rejecting the notion that the movant must demonstrate that success is more likely than not, and explaining that, the greater the volume of irreparable harm the movant would suffer in the absence of a stay, the lower the bar for satisfying this factor.55 Gudenschwager did establish a floor: “the movant is always required to demonstrate more than the mere ‘possibility’ of success on the merits.”56
So, with respect to its chances of success on the merits of the appeal, a movant must show something between a possibility and a probability. And within whatever poorly defined zone that leaves as possibly satisfying the first factor—ranging from fifty percent (just shy of an actual probability) on the high end to, say, twenty-five percent (maybe) on the low end?—the movant’s success depends on how persuasive its argument for irreparable harm is. This seems detached from “a strong showing that it is likely to succeed on the merits of the appeal,”57 but it has the benefit of being comprehensible.
Gudenschwager then dodged application of this vague standard by invoking the presumption that enacted statutes are constitutional. The court held that the presumption itself constitutes “a strong showing that [the State] is likely to succeed on the merits of its appeal” defending a statute’s constitutionality.58
In LWV, the supreme court did not rely on the presumption of constitutionality. (Plaintiffs argued that the presumption does not apply where, as there, the challenge is to the constitutional validity of the procedure by which a law was enacted, rather than to the constitutionality of the law’s substance.59) Instead, the supreme court adopted a new bypass around factor (1)—de novo review. The majority held that the circuit court failed to “acknowledge that its determination was the first word, not the last word, on the interpretation of the relevant constitutional provisions and statutes.”60 It approvingly repeated the court of appeals’s complaint that “the circuit court never recognized that success on the merits in this case turned on questions of law that would be reviewed de novo by the appellate courts.”61 The majority portion of the order in SEIU hit these same points.62 In both orders, therefore, “the majority appears to alter the substantive law, asserting that—as a matter of law—there exists more than a ‘mere possibility’ that the Legislature will prevail on the merits.”63
To be sure, the majority in the LWV Order disclaims any change in the law.64 It argues that the point is not that de novo review satisfies factor (1) but that the applicability of de novo review underscores the circuit court’s error in declining to stay the temporary injunction it had found necessary.65 This is an odd logical move, to say the least. The majority’s reasoning insists that the circuit court should, after granting a temporary injunction based on a careful assessment of the merits, stay that injunction on the basis that the appeals court might see those merits differently. Per the majority in the SEIU Order:
The plaintiff’s likelihood of success on the ultimate merits of his/her claim is not necessarily the inverse of the appellant’s likelihood of success on appeal of a temporary injunction. In other words, the likelihood of success calculus in these two analyses is not a zero sum game. If a plaintiff has a likelihood of success on the merits of its claim, that fact does not necessarily mean that the defendant against whom a temporary injunction has been entered lacks a likelihood of success on appeal of the temporary injunction.66
This is difficult to parse. How can one court say that both sides have a likelihood of prevailing on the merits of the same question?
The majority suggests that there are actually two distinct questions, which should not be conflated. But a closer examination reveals that position to be illusory. In LWV, the court of appeals distinguished between “whether the circuit court properly denied the Legislature’s request for a stay pending appeal (taking into account an analysis of the merits of the temporary injunction)” and “whether it properly granted the Plaintiffs a temporary injunction (taking into account the likelihood of success on the declaratory judgment action).”67 This is a distinction without a difference. As Michigan Coalition of Radioactive Material Users, Inc. v. Griepentrog68—the Sixth Circuit decision upon which Gudenschwager heavily relied—notes, a stay motion requires consideration of “the same four factors that are traditionally considered in evaluating the granting of a preliminary injunction.”69 Because both analyses turn on identical considerations, there is no basis for painting them as distinct.
Perhaps the SEIU majority implicitly depends on the stay analysis requiring something less than an actual likelihood. The likelihood requirement loosens if the party seeking the stay will suffer a substantial amount of irreparable harm without a stay.70 But the problem is that the temporary-injunction analysis already accounted for the relevant harms. If the enjoined party faces harms so great as to lower the likelihood requirement in factor (1), those harms should have precluded the circuit court from granting the injunction in the first place. While there may be rare cases where the circuit court and the appellate courts weigh the harms differently, such outlier situations do not support the broad prescription offered in the LWV and SEIU Orders.
Given that there really is only one question at issue and that two opposing answers to one question cannot both be more likely than not, all that remains is the idea that the prospect of de novo review in and of itself satisfies the first factor. That is a significant change in the law from Gudenschwager—and a stunningly broad rule with expansive implications. The majority portions of the LWV and SEIU Orders do not provide any rationale for such an approach, even in the face of the dissent’s characterization of this approach as “puzzling, given that de novo review does not make the merits of a party’s arguments any stronger.”71
The second factor a party seeking to stay an injunction pending appeal must satisfy is showing that, unless a stay is entered, the movant will suffer irreparable harm.72 Such harm must be concrete.73 But in applying this factor in the LWV and SEIU Orders, the supreme court made new law by holding that “the harm that stems from refusing to stay an injunction against the enforcement of a law passed by the Legislature and signed by the Governor, regardless of the nature of the challenge to the law, is an irreparable harm of the first magnitude.”74
This is a curious idea, that the ultimate interest is in perpetuating statutory law. Perhaps that is the Legislature’s paramount concern, as they created the law. But we generally understand the Constitution to be supreme and any action, including the adoption of statutes, inconsistent with the Constitution to be void.75 Indeed, under the Wisconsin Constitution, political power originates with and belongs to the people; the government has only those powers that the people have granted it.76 It is not clear on what contrary principle the supreme court majority relies for its assertion that the interests of the Legislature—part of the government created by, for, and of the people—trump all others, including those of the people in ensuring governance that accords with our Constitution. As the dissent complains, “[t]he harm wrought by subjecting the people of Wisconsin to potentially unconstitutional ‘laws’ should be, but apparently is not, worthy of the court’s consideration.”77 Nor is it clear why such an “intangible representational injury”78 as the court imputes to the Legislature is sufficiently concrete to satisfy factor (2).
Here, too, the supreme court draws an illusory distinction between the analysis necessary for issuing a temporary injunction and that necessary for adjudicating a motion for a stay pending appeal:
When deciding a motion for a temporary injunction, a circuit court analyzes whether the party moving for an injunction has shown that it will suffer irreparable harm in the absence of a temporary injunction and that it lacks an adequate remedy at law. The circuit court also compares that showing of irreparable harm with the competing irreparable harm that the party or parties who oppose the injunction and the public will suffer if a temporary injunction is issued.
On the other hand, in the context of a subsequent motion to stay an injunction, the court must weigh the irreparable harm that the movant for a stay would face in the absence of a stay during the appeal in the event that the movant is ultimately successful in having the injunction vacated on appeal versus the irreparable harm that the party who prevailed at the circuit court would suffer without the injunction during the appeal in the event the party who prevailed at the circuit court was successful in having the temporary injunction affirmed at the end of the appeal. In other words, the analysis for a stay motion adds to the mix the ability of the respective harms to be undone or unwound by the appellate court at the end of the appeal.79
This again ignores the settled understanding that the factors to be considered in each analysis are the same.80 And it ignores the definition of irreparable harm. Some harms cannot be “undone or unwound”; that is what makes them irreparable. Given that both the temporary-injunction and the stay-pending-appeal analyses require consideration of irreparable harms, there is no basis for characterizing “the ability of the respective harms to be undone” as a new element “add[ed] to the mix.”81
Similar to the equation of de novo review with a “strong showing” of likely success on the merits to meet factor (1), the elevation of legislative harm to “irreparable harm of the first magnitude” seems to place a thumb on the scale in favor of the Legislature (and potentially to the detriment of the other, coequal branches of Wisconsin government). It is doctrinally unjustified and serves primarily as an arbitrary way for the majority to satisfy factor (2).
Gudenschwager distinguishes between harms to the movant—factor (2)—and harms to others—factors (3) and (4).82 This is important because Gudenschwager frames the factors in such a way that any harm to the party that obtained the injunction or to the public could be grounds to deny the stay, and because Gudenschwager calls for balancing in considering factors (1) and (2), but not in considering factors (3) and (4).83 Yet in the LWV and SEIU Orders, the supreme court collapses factors (2) through (4) into one big balancing exercise.
The court of appeals was explicit about this in staying the LWV injunction:
[T]he interests at stake in a particular case do not always fit squarely within one of the enumerated factors in Gudenschwager. … Flexibility as to where a particular harm is discussed does not alter a movant’s overall burden to address any facts relevant to one of the required factors in some manner, and to ultimately demonstrate that all of the combined factors favoring a stay outweigh all of the combined factors opposing a stay.84
The supreme court perpetuated this amalgamation of harms in adopting the notion that enjoining statutes imposes “a substantial harm to the Legislature and to the public.”85 The SEIU Order went even further, “acknowledg[ing] that in most cases there will be some harm to both sides,” but asserting that “does not mean, however, that the totality of the harms on each side of the issue will be of equal severity and magnitude.”86
Aggregating all harms into one analysis not only contravenes the analytical framework set forth in Gudenschwager but also creates greater opportunities for appellate courts to engage in results-oriented analyses, by opening the door to courts identifying various harms without carefully defining, categorizing, and weighing each one. For example, in LWV, the court of appeals explained that the harm to the Legislature’s “claim of an intangible representational injury is the flip side of the potential harm that the circuit court recognized that would result from enforcing an invalid law, and it is no less powerful.”87 It nonetheless granted the stay.88 But if the harm to the movant and the harm the injunction sought to prevent are in equipoise, how does that justify a stay? The court offered no rationale. The supreme court was equally, but differently, loose in its logic; after declaring that the Legislature suffered “irreparable harm of the first magnitude,” the majority asserted (contradicting the court of appeals without explanation), that “staying the injunction would not have created irreparable harm” for those parties—including the Governor—that argued the Legislature had violated constitutional strictures.89
The supreme court has authority to change the law. If the Gudenschwager factors prove insufficient or need refinement to work in subsequent cases, the court should say so. Perhaps the LWV and SEIU Orders discharged that obligation. But, if we understand those orders as precedents that changed the law as outlined above, we are left with a problem: it is far from clear whether circuit courts retain authority to issue temporary injunctions against unconstitutional laws.
The supreme court majority recognized the problem itself in the SEIU Order. It explained that a de novo standard of review must tilt factor (1) in favor of a stay, even though the circuit court found, as a prerequisite to granting the temporary injunction, that the plaintiff was likely to prevail on the merits.90 It then continued:
If the opposite were true, then no stay of a temporary injunction pending appeal would ever be entered because a circuit court must always find a reasonable likelihood of ultimate success on the merits by the party seeking an injunction in order to issue the temporary injunction in the first place.91
The majority made the same point in insisting that the irreparable-harm analysis differs in the context of motion for a stay pending appeal and in the context of a motion for temporary injunction: “If a circuit court merely conducted the same analysis of harms in deciding the stay, of course it would reach the same conclusion.”92
This argument is insufficient to show that the analyses actually differ in these two contexts, as discussed in Section IV above. That is not to say that the argument lacks a certain appeal; it may well be a reasonable basis for reconsidering the framework for adjudicating stays pending appeal so that analysis will be meaningfully distinct from the temporary-injunction analysis. Or, perhaps what the argument shows is that there is no reason to have the court that issues the temporary injunction consider a stay request: because the analyses are mirror images, “of course [the court will] reach the same conclusion” at the end of both.93
But in the LWV and SEIU Orders the supreme court took neither of those paths. The majority opinions in those orders did not alter the Gudenschwager framework in ways that distinguish it meaningfully from the temporary-injunction standard. Nor did the majority hold that motions to stay injunctions pending appeal should be considered, in the first instance, by appellate courts. Instead, the majority adopted per se rules for the first two Gudenschwager factors and significantly discounted the last two Gudenschwager factors by lumping them together with factor (2).
The per se rules the supreme court established for factors (1) and (2) strongly favor the Legislature in any case where a circuit court enjoins a law as unconstitutional. In such a circumstance, the fundamental legal question—How does a given constitutional provision or precept apply to this specific statute?—will always be subject to de novo review. And the Legislature will always benefit from the categorical rule that it “and the public suffer a substantial and irreparable harm of the first magnitude when a statute enacted by the people’s elected representatives is declared unenforceable and enjoined before any appellate review can occur.”94 Because that harm is “of the first magnitude,”95 it is difficult to conceive of a countervailing harm that would outweigh it.96
It follows from these per se rules that—after the thorough, careful analysis that concludes a temporary injunction against enforcement of an allegedly unconstitutional law is warranted—a circuit court errs if it does not immediately stay its own injunction. That is, any time a circuit court enjoins a law, it must immediately stay the injunction pending appeal. If that is the upshot, the LWV and SEIU Orders effectively divest circuit courts of the power to enjoin unconstitutional statutes. Such an understanding is in significant tension—at minimum—with Article VII, Section 8 of the Wisconsin Constitution and chapter 813 of the Wisconsin Statutes, which expressly authorize circuit courts to grant temporary injunctions.
The conclusion that the supreme court’s orders in LWV and SEIU effectively strip circuit courts of authority to enjoin enforcement of unconstitutional laws may seem stark. Yet it is not only the end of the logical chain animating those orders, but also clearly supported by the text of the SEIU Order. There, the majority expressly objects to laws being “declared unenforceable and enjoined before any appellate review can occur.”97 How else can that be understood but as a general prohibition on circuit courts issuing orders that enjoin enforcement of laws on the basis of a constitutional challenge?