Concept
Shady Grove in Depth
Federal rule or statute on point
Shady Grove in Depth
Shady Grove Orthopedic Associates v. Allstate Insurance Co. runs the two-step framework from earlier in this chapter on one set of facts, and it shows how divided the Court remains over the second step. The Court held that Federal Rule 23 governs whether a diversity suit may proceed as a class action, so the Rule displaced a New York statute that would have denied class status to certain claims for statutory interest.1 The first step asks whether the Rule answers the question in dispute. The second asks whether the Rule is valid under the Rules Enabling Act, which requires that it deal with practice or procedure and not abridge, enlarge, or modify a substantive right.2
Five Justices agreed that Rule 23 controls, and they could not agree on why it is valid. Justice Scalia wrote for only four Justices, a plurality, meaning an opinion short of a majority that does not state the Court’s rule.3 Justice Stevens supplied the fifth vote in a separate opinion concurring in the judgment, and a dissent found no collision between the Rule and the statute at all.4
The subsections below take the opinions in the order a court meets them: the collision, the plurality’s validity test, Justice Stevens’s concurrence, the dissent, and what lower courts have done with a decision that has no majority rationale on validity. An answer built on this case should say which opinion carries which step. Agreement that the Rule covers the question is firm. Agreement on validity is not.
The Conflict With Section 901(b)
The first step turned on one question, whether Shady Grove’s suit could proceed as a class action, and the facts show why that question was the whole case.
Shady Grove Orthopedic Associates, a medical provider, treated Sonia Galvez for injuries from an automobile accident, and she assigned to it her right to benefits under a policy that Allstate had issued in New York. Allstate had 30 days under New York law to act on the claim, and the statute set a high interest rate on overdue benefits, evidently to encourage prompt payment.5 Shady Grove sued for the statutory interest on overdue payments of benefits.6
Its own claim was worth about $500. No provider would litigate that alone, because the cost would exceed the recovery, and the sum falls far below the amount in controversy for an individual diversity suit under 28 U.S.C. § 1332(a).7 A class action for all similarly situated providers was another matter. The total might reach $5 million, and many lawyers would gladly pursue it for a percentage of the recovery.8 Shady Grove invoked the class-action diversity provision, 28 U.S.C. § 1332(d)(2)(A).9 Whichever rule governed the availability of a class action therefore also decided whether the suit could stay in federal court.
Section 901(b) of New York’s Civil Practice Law and Rules prohibits class actions to recover a statutory penalty or minimum recovery unless the statute creating the penalty explicitly authorizes them.10 The district court held that § 901(b) applies in diversity suits despite Rule 23, that statutory interest is a penalty under New York law, and that without the class the $500 claim could not satisfy the jurisdictional amount, so it dismissed for lack of jurisdiction.11 The Second Circuit affirmed.12
The Supreme Court began where Hanna says to begin. A court “must first determine whether Rule 23 answers the question in dispute,” and if it does, the Rule governs “unless it exceeds statutory authorization or Congress’s rulemaking power.”13 Rule 23 provides that when its prescribed preconditions are met, “a class action may be maintained.” The majority read that as a categorical rule entitling a plaintiff whose suit meets the criteria to proceed as a class, and it noted that the Federal Rules regularly use “may” to confer categorical permission.14 Courts do not maintain actions; litigants do, so the discretion in “may” belongs to the plaintiff.15 Section 901(b) gives the opposite answer to the same question, since it says a suit seeking a penalty “may not be maintained as a class action.”16 A court can apply only one of the two. That is the direct collision described earlier, and it means § 901(b) “cannot apply in diversity suits unless Rule 23 is ultra vires,” a phrase meaning beyond the authority the Rules Enabling Act confers.17
The Second Circuit had avoided the collision by dividing the field. Rule 23, it reasoned, sets only the criteria for deciding whether a given class should be certified, while § 901(b) asks an antecedent question, whether the type of claim is eligible for class treatment at all, on which the Rule is silent. Allstate embraced that analysis.18 The majority replied that the distinction is made to order, because relabeling Rule 23(a)’s prerequisites “eligibility criteria” would erase Allstate’s objection.19 A label cannot decide whether two provisions collide, since a state could restate any limit on a Rule’s permission as a threshold question the Rule never reached.
A related argument points to the exceptions Congress has written into Rule 23’s reach. Congress, unlike New York, holds ultimate authority over the Federal Rules and may create exceptions by amending a Rule or enacting a separate statute. Those exceptions do not show that Rule 23 fails to apply generally. They show the opposite, because they would be unnecessary if Rule 23 did not authorize class actions across the board.20
Purpose supplied the next argument. The dissent’s theory was that § 901(b) has nothing to do with whether the suit may be maintained as a class action and affects only the remedy a plaintiff may obtain if it wins. The majority answered that by its terms the provision precludes a plaintiff from maintaining a class action seeking statutory penalties.21 An inquiry into legislative purpose would also be hard to run. Even a state supreme court’s statement of a law’s purpose would not settle the issue, since the factual predicate for avoiding federal preemption is ultimately a federal question, and federal judges would be left to pore through state legislative history that may be less easily obtained, less thorough, and less familiar than its federal counterpart.22
Nor could the Court read Rule 23 narrowly to escape the collision. It would have done so had the Rule been susceptible of two meanings, one that would violate § 2072(b) and one that would not, but Rule 23 “unambiguously authorizes any plaintiff, in any federal civil proceeding, to maintain a class action if the Rule’s prerequisites are met.”23 What remains is the argument that applying the Rule would change a substantive right. That argument goes to validity, and the next subsections take it up.
The majority also marked the edge of its holding. It did not decide whether a state law that limits the remedies available in an existing class action would conflict with Rule 23, because § 901(b) does something else: it precludes a plaintiff from maintaining a class action seeking statutory penalties.24 That a plaintiff could drop the penalty request and proceed as a class does not make the statute a remedies provision. A state ban on class actions for fraud claims would let a plaintiff drop the fraud counts and proceed on the rest, yet the ban would not mean the law provides no remedy for fraud. It would affect only the procedural means by which the remedy may be pursued.25
In Shady Grove both provisions answered whether this suit could proceed as a class action, and they answered in opposite ways, so the labels “procedural” and “substantive” and the state’s reasons never entered the collision question. The cap answers a different question, how much a class that is already proceeding may collect. State the question each provision answers before saying anything about the legislature’s goals. If a collision exists, the analysis moves to whether Rule 23 is valid, and there the Justices divided.
The Plurality’s Validity Test
Once the collision was established, the question became whether Rule 23 is valid under the Rules Enabling Act, and on that question Justice Scalia spoke for only four Justices.27 The plurality looked at the Rule and nothing else: “it is not the substantive or procedural nature or purpose of the affected state law that matters, but the substantive or procedural nature of the Federal Rule.”28 A Rule that regulates procedure is authorized by § 2072 and is valid in all jurisdictions, with respect to all claims, regardless of its incidental effect on state-created rights.29
The plurality drew the line the Court has drawn since Sibbach, which asks whether a rule “really regulates procedure.”30 A Rule that governs only “the manner and the means” by which litigants’ rights are “enforced” is valid. A Rule that alters “the rules of decision by which [the] court will adjudicate [those] rights” is not.31 The test cannot ask whether a Rule affects the parties, because valid Rules do. Each Rule the plurality discussed had some practical effect on the parties’ rights, yet each regulated only the process of enforcing those rights and altered neither the rights themselves, nor the available remedies, nor the rules of decision.32 A single question about the Rule also has the virtue commentators credit to Hanna itself, a simple and clear test that protects the Federal Rules from encroachment by conflicting state practices.33
Rule 23 passes that test. A class action, no less than traditional joinder, of which it is a species, merely enables a federal court to adjudicate the claims of multiple parties at once instead of in separate suits. Like joinder, it leaves the parties’ legal rights and duties intact and the rules of decision unchanged.34
Allstate objected that a class changes the stakes. The plurality granted that some plaintiffs who would not bring individual suits for small sums will join a class, but that has no bearing on Allstate’s or the plaintiffs’ legal rights. The likelihood that a class action will induce some plaintiffs to sue is the sort of “incidental effect” that does not violate § 2072(b).35 The device changes how many claims arrive in one courtroom. It does not change what each claim is worth.
Both the amendment and the actual Rule may raise a defendant’s total exposure, and the plurality’s test asks where that exposure comes from. Under Rule 23 as written it comes from the number of claims that reach court. Under the amendment it would come from a changed entitlement on each claim.
The plurality accepted that excluding certain class actions may cap the damages a defendant faces in a single suit, but it held that the law itself alters only procedure, much as a state law forbidding simple joinder would.36 Allstate’s fallback was that § 901(b), even if procedural, was enacted “for substantive reasons.”37 The plurality held that purpose does not matter, because the validity of a Rule depends entirely on whether it regulates procedure.38
Two considerations support looking only at the Rule. First, a test that varied with each state’s law could make one Rule valid in some places and invalid in others, and the plurality read Sibbach’s exclusive focus on the Rule as driven by the real concern that Federal Rules which vary from State to State would be chaos.39 Sibbach itself refused to measure a Rule by the importance of the right it touches, warning that such a criterion would “invite endless litigation and confusion worse confounded.”40 Second, looking to the state side multiplies the work. Instead of a single hard question whether a Federal Rule regulates substance or procedure, courts would face hundreds of hard questions about the character of every state rule that might conflict with that Rule.41
Justice Stevens proposed a different second step, asking whether a state rule is so bound up with a state right or remedy that it defines the scope of that right, and the plurality’s reply shows how firmly it held to the single-criterion test. It conceded that Sibbach’s exclusive focus on the challenged Rule is hard to square with the terms of § 2072(b).42 But Sibbach has been settled law for nearly seven decades, setting aside precedent requires a “special justification” beyond a belief that it was wrong, and the burden is greater for a statutory precedent because Congress remains free to correct the Court.43 Allstate had not even asked the Court to overrule it.44
The plurality also accepted what its rule produces. Divergence from state law, with the forum shopping that follows, is “the inevitable (indeed, one might say the intended) result of a uniform system of federal procedure,” because Congress itself has created the possibility that the same case may follow a different course in federal court.45 A Federal Rule governing procedure is valid whether or not it alters the outcome in a way that induces forum shopping.46 That applies Hanna’s division of labor, in which the twin aims of Erie guide the choice only when no Rule or statute is on point, a point a later section takes up in full.
The weight of this opinion depends on the count. The plurality’s validity test drew four votes. Justice Stevens reached the same judgment by another route, and lower courts have largely treated his concurrence as the controlling approach, which leaves the plurality’s test the more hospitable to Federal Rules and the less hospitable to state limits on class actions.47
Justice Stevens’s Concurrence
Justice Stevens supplied the fifth vote by agreeing that Rule 23 controls while rejecting the plurality’s account of why. He concurred in part and in the judgment.49 He also agreed with Justice Ginsburg that some state procedural rules must be applied in diversity cases because they “function as a part of the State’s definition of substantive rights and remedies.”50 The concurrence therefore accepts the majority’s reading of the Rule and the dissent’s concern for a state’s substantive framework, and it uses the Enabling Act’s second sentence to reconcile them.
His framework runs the two steps of this chapter. A court first asks whether the Federal Rule is sufficiently broad to control the issue, leaving no room for the state law to operate.51 In the second step, a Rule cannot govern a case in which it would displace a state law that is procedural in the ordinary sense but so intertwined with a state right or remedy that it functions to define the scope of the state-created right. Without a governing Rule, the court conducts the traditional Rules of Decision Act inquiry under Erie.52 The two steps can also bleed together, because when a Rule appears to abridge, enlarge, or modify a substantive right, a court must ask whether the Rule can reasonably be read to avoid that result.53
Justice Stevens rejected the plurality’s test because it reads only half of § 2072. Asking whether a Rule “really regulates procedure” tracks the first limitation, to “general rules of practice and procedure,” but it ignores the second, that Rules “not abridge, enlarge or modify any substantive right.” In doing so it ignores the balance Congress struck between uniform federal procedure and respect for a State’s construction of its own rights and remedies, along with the separation-of-powers and federalism presumptions against judicially created rules that displace state substantive law.54 A state’s choice of vehicle matters too. When a State uses a traditionally procedural vehicle to define the scope of substantive rights or remedies, federal courts must recognize and respect that choice, or Federal Rules would limit the ways sovereign States may define their rights and remedies.55
Such laws exist. A rule requiring plaintiffs to post a bond before suing, as in Cohen, or a statute of limitations, as in Guaranty Trust, makes a claim harder to bring or prove and so limits its scope, and a state procedure for reviewing jury verdicts, as in Gasperini, can cap the remedy.56 They are the exception. Most state rules that bear on the litigation process are adopted for some policy reason, but few define the scope of a substantive right or remedy.57 On this view Sibbach stays consistent. The medical-examination rule there reflected competing judgments about privacy, which did not pertain to the scope of any state right at issue, and the analysis might have differed had the Rule displaced a state rule sufficiently intertwined with a state right or remedy.58
Having widened the plurality’s test, Justice Stevens at once narrowed it. The bar for finding an Enabling Act problem is a high one. A state law designed as a procedural rule suggests a judgment about how state courts ought to operate, not about the scope of state-created rights, and trying to discover the true nature of every state procedural rule imposes costs on a federal court system. “The mere possibility that a federal rule would alter a state-created right is not sufficient. There must be little doubt.”59
Section 901(b) did not clear that bar. Its text expressly and unambiguously applies to claims under New York law, federal law, or the law of any other state, and no New York court had read it otherwise, so it is hard to see it as defining New York’s own right or remedy.60 It also sits in New York’s procedural code as a rule about when to certify class actions brought under any source of law.61 The history cut both ways. Statements that the provision answered fears of “annihilating punishment of the defendant” are weak evidence that it defines who may obtain a statutory penalty, since any device that makes litigation easier makes it easier for plaintiffs to recover damages.62 With two plausible competing narratives, Justice Stevens would respect the plain textual reading of § 901(b) and Congress’s decision that Rule 23 governs class certification in federal court.63
The facts that went against the state in New York’s case are the ones the hypothetical reverses: the reach of the text, the placement of the provision, and the state courts’ reading of it. Section 901(b) fell on one side of the concurrence’s line because each fact pointed toward a rule about how courts operate. The bar in the hypothetical falls on the other side because each points toward a rule that defines what the plaintiff can recover. The same Federal Rule can therefore displace one state class-action bar and yield to another, depending on what the state rule actually does, and that is why the concurrence is called case-specific.
The concurrence also shows where Justice Stevens would protect state interests: inside the Enabling Act. He accepted that forum-shopping considerations would matter if the Rule did not govern the question or could fairly be read not to, “for precisely the reasons given by the dissent,” but he thought Rule 23 plainly governs class certification, so those considerations did not arise.64 His challenge to the dissenters was direct. If they feel strongly that § 901(b) is substantive, they should argue within the Enabling Act’s framework, because otherwise “the Federal Rule applies regardless of contrary state law.”65
This opinion matters for what follows because the plurality’s four votes do not make a majority rule on validity. Courts applying the decision have largely treated the concurrence as controlling, since the plurality’s test is significantly less hospitable to state laws that curtail class actions in federal court while the concurrence’s is more case-specific and nuanced.66 A diversity problem involving a state limit on class actions should therefore be analyzed under the concurrence’s question, whether the state rule functions to define the scope of a state right or remedy, with the plurality’s test offered as the alternative.
The Dissent’s Approach
Justice Ginsburg’s dissent for four Justices ran the same two steps and parted company at the first. She agreed that a Federal Rule that controls an issue and directly conflicts with state law applies in diversity so long as it is consonant with the Rules Enabling Act, and that the Rules of Decision Act and Erie govern when no Rule or statute reaches the issue.68 She disagreed that Rule 23 reaches the issue at all. The Court, she wrote, “finds conflict where none is necessary,” and she concluded that Rule 23 does not collide with § 901(b).69 Having found no Federal Rule on point, the dissenters never asked whether Rule 23 is valid. They moved to the Erie analysis, where they concluded that the state rule should apply.70
Her reasoning started from what each provision is for. Rule 23 prescribes the considerations relevant to class certification and postcertification proceedings, but it does not command that any particular remedy be available when a party sues in a representative capacity. Put another way, Rule 23 “describes a method of enforcing a claim for relief, while § 901(b) defines the dimensions of the claim itself.”71 It is immaterial that § 901(b) bars statutory penalties in wholesale rather than retail fashion, because the legislature could have written the same limit into every provision creating a penalty, and § 901(b) operates as shorthand to that effect.72 The majority asked whether a suit “may” or “may not” be maintained as a class action. The dissent asked why. Rule 23 authorizes class treatment because the class mechanism generally affords a fair and efficient way to aggregate claims, while § 901(b) denies statutory damages to class members because the New York Legislature considered the result of adjudicating such claims en masse to be exorbitant.73 The fair and efficient conduct of class litigation is the concern of the federal rulemakers, and the remedy for an infraction of state law is the concern of the State’s lawmakers.74
She then showed that a court could honor both provisions in the very case before it. Plaintiffs seeking to vindicate claims that carry a statutory penalty may pursue a class action if they forgo statutory damages and seek actual damages or injunctive or declaratory relief, and any class member who objects can opt out and pursue the penalty in an individual action.75 As the Second Circuit explained, “Rule 23’s procedural requirements for class actions can be applied along with the substantive requirement of CPLR 901(b).”76 Although § 901(b) is phrased as a statement about whether certain class actions may begin, it is “unmistakably aimed at controlling how those actions must end,” and on that remedial issue Rule 23 is silent.77 The solution also has practical appeal. Because Rule 23 expresses no policy on the availability of statutory penalties in class actions, applying the state limit would nullify nothing in the Rule, and it would keep plaintiffs from nullifying New York’s policy simply by filing in federal court.78
The plurality refused that accommodation, and its reasons show how far apart the camps stood. The text came first: Rule 23 says a class action “may be maintained” when its criteria are met, and § 901(b) says a suit seeking a penalty “may not be maintained as a class action,” so the two speak to the same thing.79 More fundamentally, the plurality rejected the premise beneath the dissent, that a reluctance to displace a state’s substantive policy should inform whether a Federal Rule collides with a state rule.80 The dissent in turn doubted the majority’s reservation for laws that limit remedies in an existing class action. It saw no real difference in purpose or intended effect between a bar on class treatment and a statute prescribing that “no more than $1 million may be recovered in a class action,” and it said that treating one as inside Rule 23’s domain and the other as outside “fundamentally misperceives the office of Rule 23.”81
Precedent supplied her second support. A Rule displaces state law only if it, “when fairly construed,” is sufficiently broad to control the issue and leave no room for the state law,82 and interpretation of a Federal Rule can and should reflect an awareness of legitimate state interests, as the Hanna majority itself recognized.83 The Enabling Act enters her reasoning here as a reason for restraint, since the Court has counseled against an “adventurous application” of Rule 23.84 She also noted that New York is not alone in containing penalties and minimum recoveries by disallowing class relief, since Congress has precluded class treatment for certain claims seeking a statutory minimum recovery.85
Finding no Rule on point, the dissent turned to Erie and asked whether applying the state rule would have “so important an effect upon the fortunes of one or both of the litigants that failure to [apply] it would be likely to cause a plaintiff to choose the federal court.”86 A statute can be substantive for Erie purposes even when it has a procedural thrust. In Gasperini the Court found that New York’s standard for reviewing the excessiveness of a jury verdict was designed as a control analogous to a damages cap; the statute was framed as “a procedural instruction,” but “the State’s objective [wa]s manifestly substantive.”87 The dissent applied the same observation here. By barring statutory damages in a class action, § 901(b) controls a defendant’s maximum liability, and the legislature’s choice to write the limit as a class bar and not as an explicit cap should be inconsequential.88 The statute’s breadth did not make it procedural either, since the absence of limiting language is best explained by the fact that New York legislators make law with New York plaintiffs and defendants in mind.89 The history pointed the same way. Blocking class proceedings for statutory damages makes little sense for the fair conduct or efficiency of litigation, so it serves a manifestly substantive end, limiting a defendant’s liability in a single lawsuit to prevent the exorbitant inflation of penalties.90
The variation between the two courtrooms was stark. Shady Grove sought class relief ten thousand times greater than the individual remedy available in state court, and, as the plurality acknowledged, forum shopping will undoubtedly result if a plaintiff need only file in federal court to seek a massive award that state law bars.91 The dissent invoked Guaranty Trust: a federal court sitting in diversity “cannot afford recovery if the right to recover is made unavailable by the State.” It would have held that New York’s limit on statutory damages applies and would have affirmed.92
The dissent closed with an irony about the jurisdictional door. Shady Grove could sue in federal court only through the Class Action Fairness Act, yet Congress envisioned fewer class actions overall, not more. It never anticipated that the Act would make federal courts a mecca for class actions seeking state-created penalties for claims that would be barred from class treatment in the State’s own courts.93
The dissent did not carry the Court, and the reason connects it to the concurrence. Justice Stevens shared its premise that some state procedural rules function as part of a State’s definition of its rights and remedies,94 but he placed that concern in the Enabling Act’s substantive-rights proviso instead of in the reading of Rule 23, and he found the high bar for invalidity unmet.95 The dissent’s outcome-determinative step comes from Guaranty Trust, which involved a federal judge-made rule of equity.96 Hanna held that a case controlled by the Rules Enabling Act, as the majority viewed this one, does not call for that test.97 On a diversity problem, the dissent supplies the answer to the first step: no direct collision, because the Rule governs class procedure and the statute governs the remedy, so the court turns to Erie. It is the minority position.
After Shady Grove
No rationale for the Rule’s validity won five votes. The plurality tested the Rule alone, Justice Stevens added a question about the state rule, and the dissenters found no collision at all. A fractured decision leaves lower courts two tasks: they must choose which opinion to follow, and they must decide whether the state class-action restriction before them survives under that choice.99
On the first task, courts have largely treated Justice Stevens’s concurrence as the controlling approach.100 The plurality would apply any Rule that really regulates procedure. The concurrence would hold a Rule inapplicable only where the state provision is so intertwined with a state right or remedy that it defines the scope of that right, which is the narrower route to the same judgment. The choice has consequences, because the plurality’s approach is significantly less hospitable to state laws that curtail class actions in federal court, and the concurrence’s approach is more case-specific and nuanced.101
The second task produces divergent results, and the divergence follows from the case-specific question the concurrence asks. A court that applies it enforces a state class-action restriction in federal court only if there is little doubt that the restriction defines the scope of a state right or remedy. Without that showing, Rule 23 governs.102 In McKinney v. Bayer Corp. the court distinguished Shady Grove and applied a state class-action prohibition in federal court, while In re OnStar Contract Litigation held that Rule 23 displaced a state statute limiting class recovery to state residents, so a nationwide class could proceed. Several courts have upheld class-action restrictions found in state consumer protection laws, but no clear consensus has emerged. An answer should therefore say that lower-court treatment is unsettled and then work through the concurrence’s question on the facts given.103
The decision also signals that the Hanna sequence will be run firmly at the first step. The plurality rejected Justice Ginsburg’s proposal to apply Rule 23 to the mechanics of certification while applying the state rule to the available remedy, and it declined to let reluctance to displace a state’s substantive policy affect whether the two provisions collide.104 Hanna suggests that forum-shopping concerns enter only when the Rule is not validly promulgated. One may wonder after Shady Grove whether the Court should give those concerns more weight in deciding whether a direct collision exists, but for now Hanna is likely to be applied in a strict, or in Justice Ginsburg’s word “wooden,” way.105 That strictness preserves what commentators praise in Hanna, a simple and clear test, in contrast to inquiries into what law governs a state’s primary conduct, which raise problems of vagueness and nonobjectivity.106 The majority and the dissent parted ways at the first step. The majority found a Federal Rule on point, while the dissent, reading Rule 23 narrowly, moved to the Erie inquiry, where the effect on forum choice favored the state’s limit.107
The Court has never announced a test for when a Federal Rule collides with state law. It sometimes reads Rules in implausibly narrow ways to avoid a collision and sometimes seems to treat displacement as self-evident, so predictions rest on the pattern of its decisions and not on a stated standard.108 Commentary also describes the validity of a Federal Rule as an “all or nothing” matter, akin to a facial challenge to a statute, so a Rule is assessed once and does not hold in some suits and fail in others. The dissent did not contend that Rule 23 must yield to state class-action rules in every circumstance, only under the peculiar circumstances of the case.109 Justice Stevens’s concurrence adds the one case-specific limit by asking about the particular state provision, and it sets the bar high.110 A diversity problem built on a state class-action limit therefore runs in a firm order: find the collision, which a categorical Rule such as Rule 23 will usually produce; then analyze validity under the concurrence’s question; and offer the plurality’s test as the alternative.