Deference, Discovery, and Making AEDPA Actually Reduce Delay

When Congress enacted the Antiterrorism and Effective Death Penalty Act of 1996, the first title was habeas corpus reform. It was intended to achieve the “effective death penalty” part by drastically cutting the delays in carrying out capital sentencing, at least the part attributable to the federal courts.

It did not work because it was not properly implemented. But 26 years later we may finally see that change. Today’s decision of the Supreme Court in Shoop v. Twyford, No. 21-511 is a large step in that direction.

There can be no doubt that delay reduction was the overarching purpose of the habeas corpus reforms of AEDPA. In the Senate debate, multiple Senators appended to their remarks long case histories of cases that had languished for decades in the federal courts with one proceeding after another, all following full litigation in the state courts.

Senators also questioned the assumption that the lower federal courts were any better at answering the key questions correctly than the state courts. That assumption was the foundation for the rule in effect since 1953 that the federal courts would re-decide from scratch questions of law already fully considered by the state courts. It is not hard to see why that was thought necessary in a case on racial discrimination from the South in 1953 (Brown v. Allen), but the continued validity of that assumption at the dawn of the twenty-first century was doubtful. The frequent and egregious errors of the U.S. Court of Appeals for the Ninth Circuit, especially the often-reversed Judge Stephen Reinhardt, loomed large. In my view, the two people most responsible for the habeas reforms of AEDPA were Timothy McVeigh and Stephen Reinhardt, in that order.

The Senate considered three options: (1) Enact reforms around the edges, such as limiting successive petitions, but keep the core rule that a federal district court could throw out a decision of a state supreme court merely because it disagreed with it on a debatable question. (2) Give the state court decision preclusive effect in federal habeas, just as Congress did with the decisions of the local D.C. courts when it set them up. (3) Provide that the state court decision stands on debatable questions but leave a safety valve for the federal courts to overturn clearly wrong decisions.

The bill that came to the Senate floor was option (3), a compromise hammered out between Senators Hatch and Specter. Senator Biden introduced an amendment with option (1). Senator Kyl introduced an amendment with option (2). He did not think that option (3) would be effective in reducing delay. Senator Hatch insisted that it would be. The Senate voted down both amendments and enacted option (3).

In theory, it should have worked. Very few state supreme court decisions are so far out of the mainstream that there is even a substantial question that they qualify as “contrary to, or involv[ing] an unreasonable application of, clearly established Federal law, as determined by the Supreme Court of the United States.” (28 U.S.C. § 2254(d)(1).) Therefore, the vast majority of claims where the inmate contends that the state courts got a question of federal law wrong should be dismissed at the threshold. The time from filing of the claim to dismissal should be measured in days or weeks, not years. The fact-based preclusion rules in paragraphs (d)(2) and (e)(2) would take somewhat longer, but they should still be threshold issues if they are to have any delay-reducing effect.

It was not to be. Immediately upon enactment of AEDPA, many judges of the lower federal courts engaged in what amounted to a “massive resistance” campaign, misconstruing and misapplying AEDPA to preserve their role as overseers of the state courts and to delay the execution of capital sentences.

It has been a long, hard slog. First, the Supreme Court had to repeatedly admonish the lower courts that AEPDA means what it says. They can’t just decide the case the way they want and then summarily declare that any state court that disagreed with them was “unreasonable.”

Eleven years ago, in Cullen v. Pinholster, the Court decided the seemingly obvious point that whether a state court decision is reasonable must be determined on the record before that court, not on additional evidence never presented to that court. The obvious implication — especially when considered in light of AEDPA’s delay-reducing purpose — is that these issues can and should be decided at the threshold. There is no need to engage in protracted discovery and hold hearings to decide facts when those facts have no bearing on the issue to be decided.

Yet many, perhaps most, federal district courts continued to fully litigate every issue and only apply the AEDPA limits on the back end as to whether they would actually grant relief. This method of proceeding completely defeats AEDPA’s purpose, but it has been common nonetheless.

Why hasn’t the Supreme Court put a stop to it until now? Part of the reason may be jurisdictional. Generally, the courts of appeals only have jurisdiction to hear appeals from final judgments of the district courts. (See 28 U.S.C. § 1291.) By the time the district court reaches final judgment, the path by which it got there is water under the bridge and not a ground for appeal. The Supreme Court’s usual jurisdiction, in turn, is over “cases in the courts of appeals” (see 28 U.S.C. § 1254), so in the normal course they have to get to the court of appeals first. There are ways around these limits, such as “extraordinary writs,” but they are, as the name says, extraordinary.

Now comes the case of Raymond Twyford. Fighting the death penalty in a case of admitted guilt, Twyford’s lawyer had a choice of two inconsistent mental claims. He chose the one more consistent with the evidence and Twyford’s own confession, but Twyford was sentence to death anyway. Now, of course, the trial lawyer is assailed in hindsight as incompetent for not going the other way. (Knowing how capital habeas goes, I expect that regardless of which one he chose, he would have accused of incompetence for not choosing the other one.) The state court held that the lawyer’s reasonable choice of the other approach was not ineffective assistance.

In federal court, Twyford’s new lawyers wanted him transported for brain testing to lay a foundation for the alternate theory they now say the trial lawyer should have pursued. But how could that possibly provide a ground for attacking the state court’s decision? Under Cullen, the reasonableness of the decision must be judged on the evidence before the state court, which does not include the now-requested test. (And the decision appears to be an eminently correct application of Strickland v. Washington.) Further, § 2254(e)(2) bars introducing a factual basis for the claim not presented to state court with certain exceptions, but Twyford made no showing that he qualified for an exception.

The state appealed to the U.S. Court of Appeals for the Sixth Circuit. That court held that the transportation order was sufficiently different from a routine discovery order that it was appealable under the controversial “collateral order” doctrine. Other courts of appeals have held the same. On the merits, the court of appeals made this remarkable statement:

At this stage, on review of Twyford’s interlocutory appeal seeking a transport order, we need not consider the admissibility of any resulting evidence. The district court is best situated in the first instance to untangle the knotty Pinholster evidentiary issues in Twyford’s case.

But there is nothing “knotty” about Pinholster. The state court decision must be evaluated on the state court record, period. It is a clear, simple rule. Its purpose, like nearly all of AEDPA’s habeas rules, is to avoid delay. The court of appeals just blew past all of that.

The Supreme Court did not. After emphasizing the purpose of reducing delay, the high court declared, quoting its decision last month in Shinn v. Martinez Ramirez:

A federal court “may never needlessly prolong a habeas case, particularly given the essential need to promote the finality of state convictions,” so a court must, before facilitating the development of new evidence, determine that it could be legally considered in the prisoner’s case.

That’s the Court’s italics. Never.

While the jurisdictional hook for appeal in this case depends on the transportation order as distinct from routine discovery, the principle in this quote does not. “Facilitating the development of new evidence” is broad enough to embrace all discovery. The Supreme Court says district courts must determine that the evidence sought can be considered before facilitating its development.

The way to comply with the Court’s directive in this case is clear and simple. Before any discovery, identify the claims subject to being blocked by § 2254 (d) or (e)(2). Resolve the subsection (d) claims first on the state court record. Limit factfinding on (e)(2) claims to the facts bearing on whether the petitioner qualifies for an exception. Dismiss the claims blocked by either provision. Only those claims that survive this process should go forward to discovery or evidentiary hearings.

Proceeding in this way will whittle down the claims at the threshold, streamlining the case and greatly reducing delay, as Congress intended.

Those district courts that proceed with business as usual will be in defiance of Supreme Court precedent. That warrants an extraordinary writ.

Five Justices joined the majority opinion. The other four disagreed on the jurisdictional point as to whether the transportation order was appealable and did not address the AEDPA question.