Statutory procedure cases are strange animals.  Procedural matters are low political salience, so there’s rarely clean legislative history.  And we often lack ironclad empirical evidence about the consequences of rule choices before the Court fully adopts them.  As a result, the Court is effectively free to choose its own course.  Presupposition, speculation, and thinly veiled judgments about the social value of litigation often carry the day.  CalPERS v. ANZ Securities, decided on the last day of the Court’s term, is a case in point.

The question CalPERS presented was arcane even by lawyer standards.  Does the filing of a class action lawsuit asserting claims under Section 11 of the Securities Act toll the statutory three-year limitations period for a class member who seeks to proceed alone when class certification is denied?  The Court, in a 5-4 decision that cleanly split along conservative and liberal lines, said no.

Let’s unpack this for a moment.  Virtually all federal securities laws have a two-tiered time bar. First, they have a shorter statute of limitations that starts to run when the violation is or should have been discovered.  Second, they have a longer limitations period (often called a statute of repose) that runs from the actual violation.  The two limits are worded differently.  Statutes of limitations typically say that a plaintiff “may file an action no more than five years after she knew or should have known about the injury.”  Statutes of repose say “in no event shall a plaintiff file an action more than five years after the violation.”  The CalPERS case concerned the second type of limitations period.

For statutes of limitations, at least, the Court long ago held in American Pipe Construction v. Utah that a court can use its equitable powers to toll (or pause) the limitations period for putative class members while the court is making its decision whether to certify a class.  The point is that the time a court takes to decide such a motion shouldn’t count against an unnamed plaintiff in the event class certification is denied.

A key question in CalPERS was whether this same logic should apply to statutes of repose, given their different wording.  Writing for a bare majority, Justice Kennedy said no.  Statutes of repose, he reasoned, are designed to give “complete peace” to defendants so they can continue their lives, and arrange their affairs, without the threat of continued litigation.  By choosing a slightly different wording for statutes of repose, Congress meant to foreclose courts’ “residual authority” to apply their own tolling rules.

Here’s the problem:  The text is far from clear, and there is very little other evidence that Congress meant any such thing.  Indeed, while Justice Kennedy insisted that “the text, purpose, structure, and history of the statute” supported his reading of it, his opinion cited surprisingly little evidence apart from the statutory text itself.  As a matter of statutory interpretation, the case was a push.  One could easily imagine it going the other way.

Given the lack of clear text or clean legislative history, the practical concerns raised by tolling rules were likely to loom large.  Some of these concerns should be easy to see.  Without tolling, we risk cutting off rights for no other reason than a plaintiff’s bad luck that a court took a long time to decide a class certification motion.

But the American Pipe rule hedges against another kind of risk, too.  If a statute of repose is about to expire and the court still hasn’t ruled on class certification, then a savvy investor who is a putative class member but not a named plaintiff would be wise to file what’s called a “protective action.”  Filing such an action will, in effect, put a time stamp on the investor’s claim and thus prevent it from being time-barred if class certification is denied.  Without tolling, in other words, we not only risk cutting off rights.  We also risk spurring lots of unnecessary filings that burden federal courts—and that might not be made at all in cases where the court ultimately grants class certification.

I tried to help the Court assess this latter concern by filing an amicus brief, joined by a dozen of the nation’s leading civil procedure and securities law scholars.  The main contribution was an empirical study that used data on securities cases filed between 2002 and 2009 to show that roughly one-third to one-half of all securities class actions did not reach a certification decision by the time the statute of repose had run as to one or more class members.  Extrapolating out to all securities class actions filed over the past couple of decades, that’s thousands of cases that could have produced unnecessary, wasteful filings without American Pipe’s protective rule.

My brief also showed that examining cases from 2002 to 2009—not cases filed since the Second Circuit first limited American Pipe’s reach in 2013—was the best gauge of the practical stakes in the case.  One reason is that not enough time has elapsed since 2013 to give us a good look, especially because most federal securities claims—those brought under Section 10(b) of the Securities Exchange Act and SEC Rule 10b-5—are subject to a five-year statute of repose, not the three-year one at issue in CalPERS.  Perhaps more importantly, until the CalPERS Court made American Pipe inapplicable across the land, the circuits were either split or undecided on the issue.  As a result, a putative class member who sought to go it alone after class certification was denied could often just file her lawsuit in a court that was friendlier than the Second Circuit on tolling.

But Justice Kennedy’s majority opinion ignored both points.  Instead, he bluntly observed that the petitioner had “not offered evidence of any recent influx of protective filings in the Second Circuit, where the rule affirmed here has been the law since 2013.”  “This is not surprising,” he opined, because the class action’s “very premise” is that “small recoveries” do not provide litigants with sufficient incentive to go it alone, and so many class members “may have no interest in protecting their right to litigate on an individual basis.”

Justice Kennedy’s highly motivated reading of the legislative history, in combination with his willingness to substitute impressionistic observations for the empirics at his disposal, captures the painful reality of too many of the Court’s recent procedure decisions.  Undisciplined by clean legislative history or a felt need to grapple with complex empirics, the conservative majority instead simply read its own anti-litigation bent into the statute.

This anti-litigation bent is of a piece with a more general conservative skepticism about the social value of litigation that has fueled a broad-ranging attack on our civil justice system in recent decades.  The leaders of that assault have sought to cast litigation as excessive and efficiency-killing, and litigation regimes as awash in frivolous claims and bad-faith actors, even where the empirics suggest otherwise.  In cases like CalPERS, the more measured view—that litigation is an imperfect but necessary regulatory tool that compensates for the limited resources of enforcement agencies and prosecutor’s offices—has mostly faded from view, along with any empirics that might support it.  The Court’s procedural jurisprudence, in my view, is worse for it.

David Freeman Engstrom is a Professor of Law and the Bernard D. Bergreen Faculty Scholar at Stanford Law School.