Following on EDNY Judge John Gleeson’s  scathing “expose” of the United States Sentencing Guidelines as to drugs in  Dossie and Diaz, SDNY Judge Jed Rakoff  offered his views at the National Institute on White Collar Crime conference sponsored by the American Bar Association in Las Vegas.

“My modest proposal is that they should be scrapped in their entirety and in their place there should be a non-arithmetic, multi-factor test,” he said.

What?  Death to the grid?  Return judges to their historic function of being judges rather than bean counters?

Rakoff argued that the “fundamental flaw” of the guidelines is they assume every situation can be distilled into a number for the purpose of then calculating a sentence. He called the numbers assigned to various situations “arbitrary.”

For those who don’t spend enough time pondering the merits of the Guidelines, or the affect they have on framing the concept of an appropriate sentence in a federal criminal case, the numbers in cases involving financial crimes are married to loss calculations, the amount of theoretical money involved in any way, with the alleged offense. Not, as one might suppose, the actual loss, or even a fair estimate of the potential loss, but loss in a way that can only get bigger and worse, designed in such a way as to make the loss caused by the theft of a nice pen come out to $1.7 million. A bit of exaggeration, but not much.

Rakoff said he is not opposed to tough sentences, noting that in 2009 he sentenced former lawyer Marc Dreier to 20 years in prison for running a $400 million investment fraud. Prosecutors had sought 145 years.

But he was concerned with “the irrationality of the guidelines, particularly in white-collar cases.”

“This blind emphasis on the loss calculation to the exclusion of everything else leads to bizarre results in case after case after case, Adelson being just one example,” he said.

While post-Booker sentencing allows judges to factor in the legitimate purpose as expressed in the parsimony clause, there remain two problems that pervade criminal sentencing: first, that the Guidelines still influence the general perspective of what constitutes the appropriate sentence, and second, that there are now a great many judges who have never experienced sentencing without the Guidelines, and have come to believe that the numbers, and the astronomically long sentences they push on judges, reflect some sort of meaningfully appropriate length of imprisonment.

In other words, if white collar defendants whose offense involved (under the peculiar definition of the loss calculations) $10 million, then 50 years is an appropriate sentence.  Why? Because the Guidelines don’t provide for life plus cancer. That a sentence of that length is outrageously long doesn’t occur to them, as that’s the sort of sentence the Guidelines have mandated throughout their legal and judicial careers. That’s what they consider a “normal” sentence for that crime.

Instead of focusing on numbers, Rakoff said the guidelines should instead be overhauled for a list of factors judges would be required to analyze in writing. They could weigh the factors however they wanted, and the sentences would be subject to a robust appellate review.

“But it would not result in this irrational and in my view terribly dangerous situation we have now where certain factors are just singled out, given artificially inflated numbers and then imposed directly or indirectly on judges,” he said.

Not surprisingly, a spokeswoman for the United States Sentencing Commission reacted by saying

the commission reaffirmed its view that a “strong and effective” guidelines system is the best way to achieve the purpose of the Sentencing Reform Act.

The purpose of the Sentencing Reform Act is to achieve consistency in sentencing, from judge to judge, district to district.  Of course, if everybody got 100 years regardless of the nasty details that comprise the factors that otherwise go into disposing of human lives, consistency would be achieved as well. After all, does anything matter more than consistency?