Thoughts on David Remes's Definition of "Prevailing"--and of 19

In his response to my earlier post on the New York Times's 19-to-0 figure--which turns out, he says, to be his 19-to-0 figure--David Remes makes several interesting points worthy of comment. I wish to focus on two here. I will spare the reader a rehash of the underlying facts, as they are all spelled out in the earlier posts. First, Remes writes that "This is a dispute, I think, about what 'prevailing' means." That's clearly right. We are using the word in different ways, thus producing different numbers. What's less clearly right is Remes's preferred use of the word "prevail." Remes says that in his view, a detainee should only be said to prevail before the D.C. Circuit when "the D.C. Circuit’s decision led to the detainee’s transfer from Guantánamo, or at least to an order requiring the government to make best efforts to transfer him. By that measure, the D.C. Circuit has never allowed a detainee to prevail." Indeed. The trouble is that this does not strike me as the common-sense use of the word in the context of appellate litigation. When we say that someone prevails in an appeal, we generally mean that his challenge to or defense of the lower-court ruling on appeal is accepted by the appeals court. That may or may not mean that he ultimately prevails in the larger litigation. We would, after all, surely say that Salim Hamdan prevailed before the Supreme Court even though he was ultimately convicted by a military commission. It matters, when assessing whether a detainee prevailed before the D.C. Circuit, what the detainee was asking the D.C. Circuit to do. One can, after all, win the battle and lose the war--or vice versa. The result of Remes's definition of "prevail" is that the government wins in its count even when it loses. So a case like Bensayah, in which the detainee appeals a habeas denial and gets the D.C. Circuit to throw out the lower court decision and remand the case, counts as a win for the government. Why? In David's words, because,

Remands are government wins because no matter which party lives to fight another day, a remand guarantees additional years of litigation, and therefore additional years of detention. The government wins by not losing; the detainee loses by not winning. The D.C. Circuit can reverse a denial and remand for further proceedings, secure in the knowledge that the detainee will still go nowhere.

I suspect that it would surprise government counsel, detainee counsel, and the D.C. Circuit judges who decided the case alike to learn that the government had won a case whose conclusion reads as follows:

The Government argues it is authorized by the AUMF to detain Bensayah solely on the ground he was functionally a member or "part of " al Qaeda. The evidence upon which the district court relied in concluding Bensayah "supported" al Qaeda is insufficient, however, to show he was part of that organization. Accordingly, we reverse the judgment of the district court and remand the case for the district court to hear such evidence as the parties may submit and to decide in the first instance whether Bensayah was functionally part of al Qaeda.

The point here is not that it deeply matters whether the score is 19-0 or 15-0 or 20-1. It doesn't. But it does matter if we use terminology that implies that the government always wins before the D.C. Circuit if what we really mean is that no detainee whose case has been argued there has emerged with a final order of release while conceding that a several have ended up litigating in the lower court further. Remes also, I think, misinterprets my argument that we should focus on those cases where the D.C. Circuit overturns a habeas grant--and that cases in which the court affirms a habeas denial are less interesting and probative. In his account, I am assuming my premises: that the habeas cases were weak and that these were easy government wins. I don't think that's what I'm doing. Here's my thinking: Affirmances are far more likely in general than reversals of district court rulings. The appeals court owes deference to the district court on factual matters, after all, and judges agree about most things most of the time anyway. So we generally assume--without some reason to suspect a reversal is likely--that most lower court opinions will be affirmed. That is particularly true when we're dealing with habeas denials by a lower court that is (I think Remes and I agree) a far friendlier forum for detainees than the appeals court is. And it's even truer when you factor in that nearly all detainees who lose will appeal while the government, when it loses, will appeal only some of the time--with the likelihood of on appeal being at least partly dependant on its assessment of its likelihood of success. These observations rely less on assumptions about the merits of the cases than they do on assumptions about the way the courts work and the incentives facing the parties in these cases. That said, they do lead me to a different set of expectations than I think Remes has about the body of D.C. Circuit cases. They lead me to be unimpressed, for example, by relatively large numbers of D.C. Circuit affirmances of habeas denials--which is precisely the outcome I would tend to expect in those cases--and far more interested in those cases in which the D.C. Circuit disagreed with the lower court and redirected it in either direction. That universe of cases is far smaller. It consists of six cases in which the lower court would have granted the writ and the D.C. Circuit reversed and one case in which the lower court would have denied the writ and the D.C. Circuit reversed. While I would largely disagree, I think it's perfectly fair for Remes or the Times to complain that these cases constitute a betrayal of Boumediene. But that is a substantive complaint, not a numerical one--and it's a case one has to make with argument about the substance of those cases. That case does not strike me as stronger if there are many than if there are few other cases in which a district judge found against a detainee and the D.C. Circuit agreed.