Document Type


Publication Date



Narcotics violations, Drug prosecutions, Federal drug sentences, Federal sentencing guidelines


Criminal Law | Criminal Procedure


This is the first of two articles, the second of which will appear in January 2002 edition of the Iowa Law Review, in which we seek an explanation for the little-noticed and hitherto unexamined fact that the average length of prison sentences imposed in federal court for narcotics violations has been declining steadily since 1991-92.

According to figures maintained by the Administrative Office of the United States Courts, in the eight years between 1991 and 1999, the average federal drug sentence decreased from 95.7 months to 75.2 months, a drop of 22%, or nearly two years, per defendant. United States Sentencing Commission statistics report a less precipitous, but still unmistakable, decline in average drug sentence from 88.2 months in 1992 to 75.2 months in 1999, a drop of 14.7%. Preliminary figures from the Sentencing Commission show the decline continuing through FY 2000.

The article examines national sentencing data in an effort to determine whether the decline in federal drug sentences is real (rather than a statistical anomaly), and to identify and analyze possible causes of the decline. We focus particularly on the question of whether the decline has resulted primarily from trends in discretionary decision-making among the front line actors in the sentencing system (prosecutors, defense counsel, probation officers, and district judges) when dealing with individual cases, or whether the decline was caused primarily by non-discretionary factors such as changes in statutory or guidelines law, alterations in the mix of criminal cases brought to the federal system, and so forth. (We recognize, of course, that the categories "discretionary" and "non-discretionary" are necessarily somewhat imprecise and that some causal phenomena have both discretionary and non-discretionary aspects - a point the article discusses in detail.) Based on our review of national data, we arrive at four conclusions:

First, the downward trend in federal drug sentence length is real.

Second, at least some of the decrease is attributable to non-discretionary factors, such as the passage in 1994 of the so-called "safety valve" measures that allowed a reduced sentence for certain first-time drug offenders.

Third, the decrease in drug sentences since 1991-92 cannot be entirely explained by non-discretionary causes. Rather, the continuing downward movement over nearly a decade is, to a significant degree, the product of an array of discretionary choices by judges, prosecutors, defense counsel, and probation officers involved in sentencing individual defendants. Put plainly, the national statistics show: (1) at virtually every point in the Guidelines sentencing process where prosecutors and judges can exercise discretionary authority to reduce drug sentences, they have done so; and (2) where we can measure trends, the trend since roughly 1992 has always been toward exercising discretion in favor of leniency with increasing frequency.

Finally, we suggest, albeit far more tentatively, that these discretionary choices are, at least in part, a product of a widespread perception among the judges, lawyers, and probation officers of the federal criminal justice system that drug sentences are often too high, or are at the very least often higher than necessary to achieve the personal or institutional objectives of these front line actors of the federal criminal system.

In the second half of our project, we will examine some regional and local federal sentencing data to gain additional perspective on national drug sentencing trends.


This article predates Prof. Heise's affiliation with Cornell Law School.

Publication Citation

Published in: Iowa Law Review, vol. 86, no. 4 (May 2001).