When does the US Supreme Court ask for information from the federal government and bypass OSG? When does the Court do its own research into factual data underlying a party’s constitutional theory?

According to the Blog of Legal Times, both those things happened in Graham v. Florida, the case decided earlier this month holding that the Eighth Amendment prohibits a life-without-parole sentence for a juvenile whose crime did not involve homicide.

As part of its search for “objective indicia of national consensus,” the Court sought to fill in the gaps in a study that the inmate had offered showing that only 109 juveniles were serving such sentences nationwide.

The Court’s opinion said:

Although in the first instance it is for the litigants to provide data to aid the Court, we have been able to supplement the study’s findings. The study’s authors were not able to obtain a definitive tally for Nevada, Utah, or Virginia. … Our research shows that Nevada has five juvenile nonhomicide offenders serving life without parole sentences, Utah has none, and Virginia has eight. See Letter from Alejandra Livingston, Offender Management Division, Nevada Dept. of Corrections, to Supreme Court Library (Mar. 26, 2010) (available in Clerk of Court’s case file); Letter from Steve Gehrke, Utah Dept. of Corrections, to Supreme Court Library (Mar. 29, 2010) (same); Letter from Dr. Tama S. Celi, Virginia Dept. of Corrections, to Supreme Court Library (Mar. 30, 2010) (same). The study [submitted by the litigants] also did not note that there are six convicts in the federal prison system serving life without parole offenses for nonhomicide crimes. See Letter and Attachment from Judith Simon Garrett, U. S. Dept. of Justice, Federal Bureau of Prisons, to Supreme Court Library (Apr. 12, 2010) (available in Clerk of Court’s case file).

This last statement apparently caught the Office of the Solicitor General off guard. The request for information did not go through OSG.

Deputy SG Neal Katyal, writing a letter to the Court soon after the decision, said his office learned of the prison system’s response to the Court “only upon the release of the Court’s decision, [because it] was submitted in response to a confidential request from Court personnel.” (Katyal’s letter is here. The emphasis is mine.)

Having learned (too late) of the request, OSG oversaw a more careful review and concluded that the federal data was wrong. None of those six federal convicts fit the category. Five of them had committed their acts as adults; the other one committed a crime that did involve homicide.

Of course, a correction in this direction — that the federal count was zero instead of six — would only have made the Graham Court’s point stronger. It did not, unlike last year’s similar error in Kennedy v. Louisiana, call the majority’s reasoning about a “national consensus” into doubt.

But for appellate advocates, it’s how the US Supreme Court gathered this data that is worth pause. The method of court officials making “confidential requests” to third parties for factual data is, (one hopes) unusual for the adversary system. A public request would have ensured that interested parties had a chance to evaluate or explain the data.

Hat tip to Volokh Conspiracy for the initial link.