One might look with suspicion upon a court that initially issues a unanimous decision in favor of one litigant, and then reverses course and renders a unanimous decision for the other.

Yet that’s exactly what the U.S. Court of Appeals for the Fifth Circuit did in In re Katrina Litigation.

In March, a three-judge panel ruled — unanimously — that the U.S. Army Corps of Engineers was not entitled to immunity under the Federal Tort Claims Act with respect to damages from flooding in the wake of Hurricane Katrina allegedly arising out of negligent construction and maintenance of the Mississippi River Gulf Outlet.

In September, the same three-judge panel issued a new opinion on rehearing, ruling instead — again unanimously — that the corps was entitled to immunity.

To be sure, courts sometimes reverse themselves. In the context of courts that hear cases in panels, that usually happens upon en banc review by the full court. But that’s not what happened here; indeed, the government petitioned for en banc review, a petition that the three-judge panel treated as a simple motion for rehearing (which the government didn’t make).

One also might be less surprised if a court reverses course based on a change in vote of a single judge. However, what we see in this case — a change in the vote of all three judges — begs for an explanation.

A comparison of the two opinions divulges no obvious explanation. On the issue in question, both opinions set out essentially the same legal standard: The government is entitled to immunity under the statute’s discretionary-function exception if the decision is “susceptible to policy analysis” — the existence of a law or regulation conferring discretionary authority erects “a strong presumption” of entitlement to immunity. The two opinions diverge only in the application of that standard.

The original opinion explains that there was “ample record evidence indicating that policy played no role in the government’s decision to delay armoring” the gulf outlet. But the second opinion explains that “there is ample record evidence indicating the public-policy character of the corps’ various decisions contributing to the delay in armoring.”

Did the panel simply miss the ample evidence pointing the other way during its first encounter with the case? The only hint of an explanation lies in a footnote, where the court grants that “[m]uch, if not all,” of the evidence relied on by the plaintiffs to show that the corps simply made errors and was not making policy judgments “suggests negligence in the original design” and “does not support the theory that the corps’ decision to delay armoring was grounded in a purely scientific misjudgment.”

If the original panel made an error of law — the strength of the presumption of immunity — one would have hoped one of the judges would have caught it the first time. Moreover, judges on the court other than those on the panel in theory read court opinions before they are issued.

While in practice this may not always happen, one would have hoped that an opinion on the government’s liability for damages arising out of Katrina would have attracted the attention of all the judges.

If the original panel made an error of fact — misapprehending evidence of the corps’ negligent construction for evidence of the corps’ negligent maintenance — one might understand more why judges not on the panel might not have caught the error. But one is still at a loss to explain why none on the panel saw the problem until a petition for rehearing.

The panel’s switch is hard to accept. While few would have batted an eye had the court issued the second opinion in the first instance, the fact that the first opinion predated it puts the onus on the panel to explain why it changed its view so fundamentally.

That the final outcome turns on the applicability of a presumption — which the first panel opinion found rebutted — that the corps’ decisions were grounded in public policy further heightens the problematic absence of explanation. It would be far easier to understand a panel initially accepting the presumption, but then realizing on reflection that the presumption was rebutted. It’s much harder to accept the sequence that sees a court first finding the presumption rebutted, and then backtracking to find that it controls.

At the end of the day, the court’s actions undermine its own legitimacy and more generally people’s confidence in the courts. Absent some explanation for the change, the panel, and indeed the entire court, would have been better off simply granting en banc review by the entire court and summarily reversing the original panel decision. •