Elena Kagan and Recusal–UPDATED

By Tom Goldstein
on Apr 18, 2010 at 6:20 pm

We have significant projects underway reviewing the records of the three most likely nominees to replace Justice Stevens:  Elena Kagan, Merrick Garland, and Diane Wood.  Those will be posted over the course of the next couple of weeks.  I decided to move up the publication of one part of the Kagan analysis because the issue is starting to get some attention.

Commentators have recently asserted that if Elena Kagan were nominated, she would have to recuse herself from a large proportion of the Supreme Court’s merits docket over the course of the next two to three Terms.  The issue isn’t particularly significant, and is genuinely boring as Supreme Court nomination issues go, but because it has been raised as a novel potential ground for not appointing Kagan, it is worth addressing.

This recusal issue arises from an extrapolation from the large number of recusals by the last Solicitor General to transition to the Supreme Court, Thurgood Marshall.  That extrapolation rests on an intuitive hypothesis, but part of the problem with the blogosphere is that hypothesis too often gets published as fact.  It turns out that in reality the parallel between Kagan and Marshall is not sound for a few independent reasons that combine to significant effect.

In no particular order, Kagan would have far fewer recusals principally because (i) she would be appointed earlier in the year, (ii) the Court’s docket has fewer merits cases with the United States as a party, and (iii) a substantial amount of the time Marshall recused for reasons other than his service as Solicitor General.

Here’s the summary of how the numbers work.  Marshall is reported to have recused from more than 75 merits cases in his first Term, but 57 arose from his role as the S.G.  The best estimate is that in her first Term, Kagan would recuse from 13 cases.  In his second Term, Marshall recused from 8 cases because he served as Solicitor General.  The number for Kagan would likely be 5.  Beyond the numbers, Kagan’s recusal would likely implicate only one arguably significant issue – the government’s litigation against tobacco companies.

The rest of this post provides the details.  If you’re interested, grab a Red Bull to keep yourself going and settle in.

Marshall served as a Judge on the Second Circuit from 1961 to 1965.  He was the Solicitor General from August 1965 to August 1967.  He was nominated to the Supreme Court on June 13, 1967, and served as an Associate Justice beginning August 30, 1967.

In his first Term, Marshall recused from 61 argued cases (as opposed to all decided cases), for varied reasons.  (The only other Solicitor General who was directly appointed as a Justice, Stanley Reed, appears to have recused from 29 merits cases in his first Term on the Court (which he joined in January 1938).)

The level of nonparticipation by Marshall (and Reed) is far higher than by most Justices in their first Term.  Erin Miller collected this illustrative data for ten examples:

Because the focus of commentary has been on Marshall’s recusals, I reviewed them all yesterday, both at the merits and (where applicable) cert. stages, and for many their prior history.  How was your Saturday?

Most of the recusals – 53 – relate to Marshall’s service as Solicitor General.  Marshall appears to have applied a rule that he would recuse from any case in which he had personally participated (or at least in which his name appeared on the brief).  (There are several cases in his first Term in which the United States was a party or an amicus and Marshall’s name did not appear on the government’s papers, and in which he participated as a Justice.)

In most cases in which Marshall recused based on his role as Solicitor General (48), his personal participation in the case is obvious because his name appears on the government’s briefs at the cert/probable jurisdiction stage (33) or the cert/probable jurisdiction and merits stages (15).  In 4 other cases, his recusal indicates that he played some other role – for example, authorizing an appeal.  A final case was related to another ruling in which he was otherwise recused.

5 further recusals by Marshall in argued cases appear to relate to his service on the Second Circuit.  2 appear tied to his prior work with the NAACP.  In 1 other case, I cannot figure out why he recused.

There is another, smaller body of 4 cases in which Marshall recused when the Supreme Court ruled on the merits without argument – i.e., it disposed of them summarily.  In 2, his name is on the briefs; 1 is a continuation of prior litigation in a Supreme Court case from his time as Solicitor General; and in 1 it appears he authorized the appeal.

There is a final group of 11 cases in which Marshall recused that arguably would have been regarded as merits dispositions in 1967, but wouldn’t be now (and hence have no analog for Elena Kagan).  In 7, the Court reversed the lower court’s judgment on the merits, merely by providing a citation to a prior Supreme Court ruling.  Those cases would now be handled without deciding the merits by issuing an order vacating the ruling below and remanding to the court of appeals for further consideration.  In the remaining 4, the Court summarily disposed of “appeals” – 3 times affirming and 1 time dismissing the appeal for “lack of a substantial federal question.”  This “appellate” docket has been essentially eliminated by a subsequent statute.

One quick point before moving on.  Marshall’s large number of recusals don’t seem to have had a significant effect on the outcome of the decisions.  The Court was equally divided 4-to-4 in only 2 of the cases.

The upshot of the above is that on the basis of his service as Solicitor General, Justice Marshall recused from 57 merits cases in his first Term on the Supreme Court (53 argued and 4 summary reversals), which was roughly 40 percent of the Court’s merits docket that Term.  That is a significant number, although materially less than the one commentators have been citing as the base-line from which to extrapolate how often Kagan would recuse.

Importantly, however, it does not follow that Elena Kagan would recuse from the same proportion of the docket as Marshall did.  I do expect that she would apply the same ethical principle as Marshall:  not voting in any case in which she personally participated as Solicitor General.  But for three reasons, her number of recusals (both absolute and as a proportion of the Court’s docket) would be substantially lower than his.

First, the United States is a party less often now than it was in 1967.  In the 1967 Term, the United States was a party in approximately 50 cases that were argued and decided on the merits, which was a little over 40 percent of the argument docket.  In recent Terms, the average numbers have been 28 cases and a little below 40 percent.

Although the United States participates in far more cases as an amicus curiae on the merits now than it did then, those cases would trigger far fewer recusals for Kagan.  Before the United States participates as a party on the merits, it must of course file papers at the cert. stage (signed by the Solicitor General), which is an early event that can trigger a recusal obligation.  By contrast, when the United States comes in at the merits stage as an amicus – the current trend in the docket – the Solicitor General gets involved very late.  Thus, if the Court is composed of 50 merits cases in which the United States is a party and 50 in which it is an amicus, a former Solicitor General will recuse from the party cases vastly more often, because s/he will have previously filed a brief in the party cases at the cert stage.

Second, Kagan would be nominated earlier in the year than Marshall was by roughly 1 month to 6 weeks.  That doesn’t sound like much, but it is a critical period – May to mid-June – in which the government would be filing briefs that the Court would add to its merits docket for next Term.

Third, the Court’s merits docket in the 1967 Term was disproportionately stacked in the early months with merits cases in which the United States was a party early in the Term, and thus with cases triggering recusal by Justice Marshall.  The 2011 Term does not have the same “shape.”

In fact, it isn’t necessary to hypothesize about how often Kagan would be recused.  We can identify the cases specifically.

Start with the existing merits docket for next Term.  In only 3 cases granted so far, Kagan would need to recuse:  her name is on the briefs and the United States is a party.  There are thus far no additional merits cases in which the United States participated as an amicus at the cert stage.

There is a further group of 5 cases in which cert has been granted and in which I think the United States is likely to file an amicus brief at some point in the spring, though in most of them it will not do so soon.  Whether Kagan would be recused in any one of those cases would depend on the date of her nomination.  The practice of that Office is that the Solicitor General herself ordinarily participates only less than a week before the filing of an amicus brief.  If the nomination occurred on May 1, then she would be recused from only 1 case; each further week would add roughly 1 additional case.

As an aside, the group of merits cases is a useful illustration of the much larger group of cases requiring Marshall to recuse than Kagan.  If Kagan were nominated on May 1, her name would likely appear on 4 merits briefs next Term.  By contrast, in his first Term, the Court decided 13 cases in which Marshall appeared on the merits brief of the United States.

There are 2 further cases in which the Court has called for the views of the Solicitor General in which cert is likely to be granted.  In one, the government’s brief has already been filed, and Kagan would be recused.  In the other, Kagan likely would not be recused because she presumptively would not play a role in that case until the brief was filed in late-May (the cutoff date for filing briefs at the Supreme Court), but she would have already been nominated and left the Office by that point.

There are 3 further cases in which the Solicitor General has filed a cert petition.  It is fair to assume that each of those will be granted, and Kagan would necessarily be recused from each.

Assuming a nomination on May 1, the categories above generate a total of 8 recusals:  3 merits cases; 1 merits amicus brief; 1 CVSG; 3 pending cert petitions by the Solicitor General.

Because those 8 cases can be identified with specificity, it’s worth pausing on their significance and on the question whether the recusal is likely to change the case’s outcome.  None has particularly broad jurisprudential or practical significance.  Flores-Villar v. United States could generate ideological disagreement, because it involves alleged gender discrimination.  There are multiple cases involving suits against states by prison inmates, which have ideological overtones.  One case – the government’s suit against the tobacco industry, in which the Solicitor General’s cert petition is pending – would have significant financial consequences.  (As with Marshall, there also isn’t reason to believe that more than a few of the cases would result in an equally divided court.)  None of the cases compares in importance to the Hamdan case in which John Roberts recused himself, having previously participated as a judge on the court of appeals.

Beyond the specifically identifiable cases, there are two categories of cases I haven’t yet covered.

First, there are many cert petitions to which the United States has responded but on which the Supreme Court has not acted.  (The Solicitor General often does not review briefs in opposition to certiorari, but because her name appears on the brief, I assume Kagan would regard that as sufficient participation to trigger recusal.)  There is a lead time of essentially four weeks between the date a brief in opposition is filed and the date the Court grants certiorari.  So, again assuming a May 1 nomination, there would be a group of cert petitions in which the government filed oppositions in the 40-day period between March 20 (the oldest oppositions not yet acted on) and May 1 (the final date Kagan would sign a petition) that could be granted and that could trigger a recusal obligation.  The best I can do is to extrapolate from the Court’s merits docket and say that 40 days represents one-ninth of the docket, which equates to just over 1 anticipated cert grant (with around 11 cases a Term granted with the United States as respondent).  To be conservative, I’ll estimate 2.

Second, there are cases in which Kagan would be recused because of her personal role in authorizing lower court litigation.  The Solicitor General must personally approve any appeal, rehearing request, or amicus brief by the government.  That is a significant body of cases.  But it is not a large number of cases that end up at the Supreme Court.  Here, we can conservatively extrapolate from the experience of Justice Marshall.  Marshall served as Solicitor General for almost twice as long as Kagan would.  In his first Term, Marshall seems to have recused in roughly 6 cases involving the federal government because of his approval role (or participation at some other point), without having signed a brief.  On the low end, we could extrapolate 1 recusal for Kagan based on her approval role (because her overall recusal rate is such a small fraction of Marshall’s for the reasons discussed above) and because she would have served roughly half as long as Solicitor General.  But I think a slightly higher number is appropriate because there is a large volume of federal government litigation and because the counter-point to her being nominated earlier in the Term than Marshall is that there will be more cases in which she played an approval role but did not sign the brief.  So I estimate 3 recusals on this ground.

In sum, I would expect a total of 13 recusals – 3 pending merits cases, 1 incoming merits amicus brief, 1 CVSG, 3 pending cert petitions, 2 pending briefs in opposition, and 3 appeal recommendations – if Elena Kagan were nominated on May 1.  If the nomination occurred on May 15, it would be closer to 15.  That is roughly one-fifth to one-sixth of the merits docket, nowhere near the number or proportion of cases in which Marshall recused himself.

What about the following Term?  As I mentioned, commentators have assumed that Kagan’s recusal obligations would continue to be very substantial for two to three years.  That is not correct.

Absent an unusual circumstance, the greatest body of cases triggering recusal – those in which Kagan actually participated in the Supreme Court as the Solicitor General – would all be finalized during the upcoming Term.  After that, her recusal would be triggered by cases in which she had the approval role, discussed above, which arrive at the Supreme Court far less frequently.  In Marshall’s second Term, he recused from 14 merits cases (less than one-fourth of the pace from his first Term).  6 of those involved circumstances that likely would not apply to Kagan:  3 original jurisdiction rulings in which the United States was a party (with no similar pending cases now that I know of); 1 summary affirmance (with no jurisdictional parallel now); 1 case litigated by the NAACP; and 1 that I cannot figure out why Marshall recused.  The 8 remaining cases seem to have involved Marshall’s approval responsibilities or litigation that returned to the Court from earlier Terms (which could happen with Kagan as well).  Those 8 recusals by Marshall represent roughly 5 percent of the Court’s merits docket.  If anything, the number for Kagan should be smaller because, as noted, she would have served a shorter time as Solicitor General.  But to once again be conservative, I’ll estimate 5 cases.

In sum, I estimate that if Elena Kagan is nominated on May 1, 2010, she will recuse from 13 merits cases in October Term 2010 and 5 merits cases in October Term 2011.  If correct, that level of nonparticipation would not be dramatically higher than the average in Erin’s illustrative sample above, and would be roughly equal to (or lower than) Justices Thomas and Clark.  It would therefore not seem to be a significant basis for not appointing Kagan.

Ed Whelan responds here.  Ed doesn’t advocate for the direct parallel to Justice Marshall’s experience that he did before (both in the first Term and later Terms), but raises three points that he sees as objections to my analysis above.

First, he notes that the Solicitor General may be involved in the analysis of legislation, which can trigger a recusal obligation.  The Bush Administration Solicitors General did so much more frequently.  I believe that this Solicitor General’s Office was not, for example, involved in evaluating the constitutionality of the health care legislation.  As for detainee cases, the Office has certainly played some role (Neal Katyal argued a D.C. Circuit case, for example), but it is not nearly as significant as under the Bush Administration, where first Ted Olson and then Paul Clement were as I understand it deeply involved in the formulation of policy and lower court litigation strategy.  Analyzing the constitutionality of legislation simply isn’t the job of the Solicitor General, as opposed to, for example, the Office of Legal Counsel.  At bottom, I don’t know but think that, depending on the precise case, it is possible that Kagan would recuse from some detainee-related matters, if they came to the Court.  Beyond that, I can’t identify any bodies of cases that would require recusal on this ground.

Second, Whelan objects that it is not fair to compare a second Term in which Kagan would sit as a Justice to Thurgood Marshall’s second Term, and contends that the better measure is the 40 percent of the docket in which the United States is now a party.  I confess I don’t understand this argument.  The United States was a party in a higher percentage of cases when Marshall transitioned to the Court (not lower), and Marshall was the Solicitor General for substantially longer than Kagan would be if nominated.  Whelan raises the prospect that maybe the United States loses more cases now than in 1967, so that Kagan would have personally authorized more appeals.  But he doesn’t actually contend that’s accurate.  If anything, I think the United States wins more cases now (that was a tumultuous time in which the courts were turning against the government’s position in police-related cases in particular), and that number in any event would be easily overcome by the fact that Marshall was Solicitor General for significantly longer.

Third, Whelan disagrees with my premise that Kagan would stop serving in her role as Solicitor General as of the date of her nomination.  But although he is right that Kagan would not have formally left the Office in May, I believe I am correct that she would immediately stop taking on conflicts.  In the Sotomayor nomination, the Administration debated what to do in this circumstance, because there isn’t a settled practice for nominees.  They ultimately settled on her immediately stopping participating in matters, and as a consequence she picked up no further conflicts.  I expect they would do the same thing here, so that if nominated Kagan will be instructed to recuse herself from ongoing participation in litigation during the pendency of her nomination, recognizing that for the Court’s sake it doesn’t make sense for her to generate further recusal obligations.  (The Solicitor General can announce her recusal from any case.) And Kagan would in fact be working essentially full time on her nomination, so that it would be perfectly sensible for her not to nominally participate as the Solicitor General in cases on which she in fact did no work.

In sum, I think that Whelan is right that Kagan could find herself obligated to recuse from certain detainee-related cases.  But beyond that, for the reasons in my post above, his prior assertion and that of other commentators that she would be recused from a large number of cases is deeply flawed.  This isn’t a hypothetical exercise:  generally speaking, it’s possible to identify the precise cases in which she would be, or likely could be, recused.  And the numbers are not particularly large.