« A morning full of sentencing stories | Main | Blakely's lawyer gets more props »

December 20, 2004

Debating sentencing data

A few weeks ago, as noted here, the Washington Post ran this op-ed entitled "Mandatory Madness" in which law professor and NACDL president Barry Scheck calls for reform of harsh federal mandatory drug sentences.  In the middle of his wide-ranging and powerful critique of federal sentencing laws, Professor Scheck stated: "According to the Bureau of Prisons, more than half of the 180,000-plus people in federal institutions are there for drug law violations. Most are low-level, small-time and nonviolent offenders."

This weekend, the Washington Post ran this letter in response from Dan Bryant, who is identified as "assistant attorney general for legal policy at the Justice Department."  The letter asserts that Scheck's claim about low-level, federal drug offenders "is inaccurate," and then rattles off the following statistics:

Justice Department data show that 91 percent of all prisoners (state and federal) are either recidivists or violent offenders. Of those in state prisons, 76 percent are multiple offenders and 62 percent have a history of violence, while a full 66 percent of federal offenders have been convicted of multiple or violent crimes.

Furthermore, most nonviolent criminals are neither low-level nor small-time: 84 percent of these "nonviolent" offenders in state prison have prior criminal records, averaging more than nine arrests and four convictions apiece. In fact, a third of these nonviolent offenders could even be classified as "previously violent," as they have previous arrests for violent crimes. Federal nonviolent inmates have only marginally less criminal backgrounds than their state counterparts: 79 percent have prior records, averaging more than six arrests and two convictions. The notion that our prisons are filled with nonviolent, first-time offenders is simply not true.

This letter concludes: "We agree that there should be a healthy debate about sentencing, but we insist that this requires equipping Congress and the American people with the facts, not misleading rhetoric."  (The use of the "we" hints that this letter may represent a semi-official Justice Department response, rather than Dan Bryant's personal views.  Indeed, the letter echoes points and phrases used by Assistant AG Christopher Wray, in his official testimony to the US Sentencing Commission last month.)

Over the weekend, this letter and its data-based rebuttal of Professor Scheck's assertions have been the buzz of a listserve to which I subscribe.  In the dialogue, I noted that the letter makes heavy use of state statistics (or combined state/federal statistics) in response to an op-ed which was focused exclusively on federal sentencing.  Another person spotlighted that the letter makes a "rhetorical slip from 'low-level' to 'first-time' offenders."  Others noted that even some minor federal drug offenses are statistically categorized as "crimes of violence."  For a letter espousing the importance of facts over misleading rhetoric, the letter does a mighty good job stressing facts which could mislead.

Putting aside dickering over rhetorical use of facts, the data stressed in the Bryant letter actually prove Scheck's chief points.  The statement that "66 percent of federal offenders have been convicted of multiple or violent crimes" in turn means that 34% (more than 1/3) of all federal offenders are one-time, nonviolent offenders (and I suspect the percentage of low-level, nonviolent drug offenders may be even higher).  With a federal prison population of over 180,000, this suggests that in excess of 60,000 persons are serving time in a federal prison as a result of a one-time, nonviolent offense.   It seems our federal prisons are in fact filled with nonviolent, first-time offenders.  (Notably, the 60,000 persons now serving federal time as a result of a one-time, nonviolent offense is more than double the total federal prison population 25 years ago.)

I am glad to see from the Bryant letter that the Justice Department welcomes "healthy debate about sentencing," and I am also glad to see an emphasis on offenders "convicted of multiple or violent crimes."  The states have generally been effective at focusing long sentences on repeat and violent offenders, and federal law should follow their lead.  Indeed, based on the themes and claims in the Bryant letter, it would seem that DOJ would and should be against all mandatory sentencing except for serious recidivist or violent criminals.  That was the main thrust of Scheck's op-ed, and a careful analysis of the Bryant letter perhaps reveals more harmony than discord in views about sensible federal sentencing policy.

December 20, 2004 at 12:45 PM | Permalink


TrackBack URL for this entry:

Listed below are links to weblogs that reference Debating sentencing data:


I am a retired U.S. Probation Officer and now consult defense attorneys on sentencing issues. What everyone seems to forget is that a number of those "First time, non violent offenders" are white collar criminals, child pornography cases, tax protesters. etc..I can't speak for the whole of the sentencing judges, but I can tell you in my 20 years of writing presentences and supervising federal offenders, the ones who were young, first time offenders of non-violent offenses rarely went to prison.
And I wish the Supreme Court would hurry up, my business has fallen off considerably!

Posted by: Sharon Valenti | Dec 20, 2004 12:58:39 PM

I, too am a retired probation officer. I have always had a problem with categorizing drug offenses as "violent crimes." I have the same problem with offenses involving the presence of, but not the use of, a firearm as a crime of violence. I think those determinations need to be fact specific. Commissioner O'Neil, one of the present US Sentencing Commissisioners wrote an interesting article for one of the Boston schools (Boston COllege or Boston University, I forget which), talking about the number of first time and low level offenders in federal prison. It makes the same point that a large percentage of federal offenders are low level. I just wonder when the Commission or Congress will do something about that problem. I'm not holding my breath.

Posted by: Bryan Unites | Dec 20, 2004 2:59:44 PM

I am glad to see the debate on sentencing data involve experienced probation officers. In addition, shouldn't this "debate" on sentencing data also involve the Sentencing Commission? Understandably, DOJ and the defense bar have their view on sentencing, but USSC was formed partly with the intent to analyze this data. Does the 15-year report shed any light onto this debate?

Posted by: DEJ | Dec 20, 2004 3:41:47 PM

Is there some DOJ dictionary that is available out there in the interest of transparent and open communication of "facts" to the American people? Maybe something that freshman prosecutors get so they can learn DOJ-speak in 10 quick easy lessons? I am particularly interested in the DOJ-specific definitions for "cooperation", "substantial assistance", "violent v non violent", "truth" and of course "justice".

Posted by: NL | Dec 20, 2004 5:43:26 PM

It seems to me that repeat nonviolent drug offenders are included in the "recidivist/repeat offender" count." Doesn't this slant the statistic irrationally?

Posted by: Robert Berman | Dec 20, 2004 8:27:31 PM

"Healthy Debate"? I don't think so. I have worked as a research specialist in the Federal PD's office since 1989, and I have watched district judges sentence drug offenders since 1971 when I started work in the Federal Court Clerk's Office. There has been a monumental change in attitude toward drug offenders during the last 34 years. I used to see routinely a "second-chance" attitude where those involved in drugs were given low sentences and/or probation. Since the advent of the Guidelines, and even before, there has been a "one strike and you're out" mentality that has gripped Congress and the Executive Branch.

When I first read Blakely, I thought of the adage, "Be careful when you wish for, you just may get it." I'm afraid Congress will take yet another opportunity to lay scorn on the Judiciary by increasing penalties after Booker and Fanfan. Congress reflects the general ignorance of its constituency and the Executive Branch by its nature will amass power unto itself until Congress acts to restrain it.

The Executive Branch wants power, not debate. A stupid catch phrase like the "war on drugs" is enough to justify any outlandish governmental action. Unless the citizenry intervenes, the balance of power will remain in the Executive and the prison population will continue to escalate with little concern or public outcry.

Posted by: joe phelan | Dec 21, 2004 12:32:19 PM

Look at the debate from the Federal Bureau of Prison's perspective. It operates prisons at four
security levels (minimum, low, medium, and high) and has one maximum-security prison (less than 1 %of the prisoners). If you figure out the criteria for security level assignment (prisoner classification) and the % of inmates at each level, you'll get another (and I think interesting) perspective on the debate.

Posted by: Gregory Siegfried | Dec 21, 2004 4:29:33 PM

Mr. Siegfried,

Yours is an important and telling perspective. Please remember, though, that examining the data from this suggested point of view (prisoner security level designation), can be nuanced. By that I mean that a prisoner with a long sentence may be deemed a high security level on that basis alone. So sentence inflation may also directly cause security level inflation.

Note that from the guards' perspective, this may be reasonable. A prisoner who must serve an unwarrentedly long sentence wrt a nonviolent offense may lose all hope of recovering his life, and as a result, become violent while in prison under the inhumane strain.

Posted by: Jeannie | Dec 22, 2004 11:55:37 AM

I am not familiar with the federal classification system, although I would be surprised if sentence length meaningfully affects security level. I do have one reseach paper indicating initial federal classification no longer relies on sentence length, ironically because a study showed it was irrelevant to predicting risk of violence in prison.

My general point is prediction of misbehavior versus other justifications for criminal sentencing. Prisoner classification says what it says. You cannot equate risk of violence in a controlled environment with risk on main street. But it gives you another perspective on the assertion that persons in prison should be in prison because they are a menance to society. And I believe more intelligent sentencing policy would come from data collection and analysis of offender behavior in and out of prison. I hold far less faith in the political process and frankly constitutional law.

Posted by: Greg Siegfried | Dec 29, 2004 10:29:15 AM

I just saw your post. Maybe the BOP should get a copy of the studies you are referring to. If you are interested in researching security classifications, it is in black and white. It is in the BOP policies on the BOP website- www.bop.gov. You will see very clearly that length of time to serve can be the sole determinant of a higher security level. If you have been researching this issue, you may want to look at this basic resource. Apparently the BOP hasn't fully digested that it's not based on anything logical- or maybe some special interest lobby influence? Higher security level creates more jobs, costs more tax dollars and someone is getting that money-- considering the economics and special interests of this system is pretty disgusting if you care to look at it.

Posted by: NL | Jan 10, 2005 1:04:50 PM

Post a comment

In the body of your email, please indicate if you are a professor, student, prosecutor, defense attorney, etc. so I can gain a sense of who is reading my blog. Thank you, DAB