That racial gap has widened since the Supreme Court restored judicial discretion in sentencing in 2005, according to the Sentencing Commission's findings, which were submitted to Congress last month and released publicly this week.
In its report, the commission recommended that federal judges give sentencing guidelines more weight, and that appeals courts more closely scrutinize sentences that fall beyond them.
The commission, which is part of the judicial branch, was careful to avoid the implication of racism among federal judges, acknowledging that they "make sentencing decisions based on many legitimate considerations that are not or cannot be measured."
Still, the findings drew criticism from advocacy groups and researchers, who said the commission's focus on the very end of the criminal-justice process ignored possible bias at earlier stages, such as when a person is arrested and charged, or enters into a plea deal with prosecutors.
"They've only got data on this final slice of the process, but they are still missing crucial parts of the criminal-justice process," said Sonja Starr, a law professor at the University of Michigan, who has analyzed sentencing and arrest data and found no marked increase in racial disparity since 2005.
Douglas A. Berman, a law professor at the Ohio State University who studies sentencing, said, "It's not surprising that the commission that's in charge of both monitoring and amending the guidelines has a general affinity for the guidelines."
The Sentencing Commission didn't return requests for comment.
The Supreme Court, in the 2005 case U.S. v. Booker, struck down a 1984 law that required federal district judges to impose a sentence within the range of the federal sentencing guidelines, which are set by the commission.
The law was meant to alleviate the disparity in federal sentences, but critics say placing restrictions on judges can exacerbate the problem by rendering them powerless to deviate from guidelines and laws that are inherently biased. An often-cited example is a federal law that created steeper penalties for crack-cocaine offenses, which are committed by blacks more frequently than whites, than for powder-cocaine offenses. Congress reduced the disparity in 2010.
In the two years after the Booker ruling, sentences of blacks were on average 15.2% longer than the sentences of similarly situated whites, according to the Sentencing Commission report. Between December 2007 and September 2011, the most recent period covered in the report, sentences of black males were 19.5% longer than those for whites. The analysis also found that black males were 25% less likely than whites in the same period to receive a sentence below the guidelines' range.
The Sentencing Commission released a similar report in 2010. Researchers criticized its analysis for including sentences of probation, which they argued amplified the demographic differences.
In the new study, the Sentencing Commission conducted a separate analysis that excluded sentences of probation. It yielded the same pattern, but the racial disparity was less pronounced. Sentences of black males were 14.5% longer than whites, rather than nearly 20%.
Jeff Ulmer, a sociology professor at Pennsylvania State University, described the commission's latest report as an improvement but said it was "a long way from proving that [judicial discretion] has caused greater black-white federal sentencing disparity."