James A. Woods
Administrative Office of the U.S. Courts
Data and Methodology
The Sentencing Reform Act
Growth of the Gap
THE FEDERAL JUDICIARY’S request to Congress for resources for postconviction supervision of offenders is based on forecasts generated by the Statistics Division of the Administrative Office of the U.S. Courts (AO). These forecasts, in turn, are aided by an understanding of the population dynamics of those under postconviction supervision.1 Because criminal defendants convicted in federal courts constitute the population in postconviction supervision, a straightforward causal model implies that increases or decreases in criminal defendants sentenced to federal prison will produce, with a suitable lag, corresponding changes in the numbers of persons entering the system. In reality, this relationship is more complicated, being affected by factors that, from 1997 to 1999, resulted in a widening convicted-to-received for supervision gap, a separation that has persisted to the present.
This paper examines the relationship between the number of persons sentenced to federal prison and the number received into postconviction supervision. In this study, we hope to explain past deviations from a simple causal model of the defendant-tosupervised- release process and, by gaining a better understanding, improve the AO’s annual forecast of individuals under supervised release.
The data for this study come from the U.S. federal district courts and are compiled and maintained by the Administrative Office of the United States Courts (AO). Each of the 94 district courts creates a criminal data file at the end of each month and sends it to the Statistics Division (SD) of the AO. Using this file, the AO updates the national criminal database. This analysis used 12-month fiscal year defendant data sets from 1987-2005.2
Data on postconviction supervision used in this study come from the Federal Probation Service Information System (FPSIS), which was maintained by SD.3 Under this system, supervision data were collected by the probation office in each district. Every month each district probation office created an extract file containing 12 months of probation data. This analysis used 12-month fiscal year data sets from 1987-2005. The population studied consisted of persons convicted and sentenced in U.S. federal district courts,4 excluding duplicates and transfers,5 and those receiving a sentence of life, death, or other.6
To analyze the movement of these two time series, this paper focuses on offense types associated with relatively large proportions of both convictions and persons received into the postconviction supervision system. These offenses include drug crimes, embezzlement, firearms, forgery and counterfeiting, fraud, immigration, larceny and theft, robbery, and traffic offenses. Each year from 1987-2005, these offenses accounted on average for 61 percent of those sentenced to prison and 83 percent of persons received into the postconviction supervision system.
The Sentencing Reform Act (SRA) of 1984 was part of the broader Comprehensive Crime Control Act of 1984.7 The SRA abolished indeterminate sentencing and replaced it with a system of determinate sentences without the possibility of parole.8 Through this act, Congress established the United States Sentencing Commission and charged it with setting a narrow sentencing range for each federal crime. If a sentence handed down by a judge fell outside these guidelines, the judge was required to give reasons for that departure in writing. This encouraged more uniform sentences and eliminated wide variations in sentences for similar offenders who committed similar crimes.9 The law was challenged, but the Supreme Court declared the guidelines constitutional.10 Only defendants charged with crimes committed on or after November 1, 1987, were subject to this law, and for several years offenders convicted of crimes committed before this date were sentenced under the previous system.11 By 1994, nearly all defendants convicted in federal court were sentenced under the new system.12
Figure 1 presents defendants sentenced to prison and persons received into postconviction supervision from 1987 through 2005. This figure shows that the two series generally track each other. However, an interesting trend evident in Figure 1 is the increasing size of the gap between the two series—that is, the number of criminal defendants sentenced to prison grew at a faster rate over the period than did the number of persons received into postconviction supervision. Three factors discussed below account for this widening gap: 1) the hiring freeze imposed by the Department of Justice in the early 1990s, 2) the federal sentencing guidelines that took effect in 1987, and 3) the increase in persons charged with and sentenced to prison for immigration offenses since 1994, who never enter post-conviction supervision because they were deported when released from prison.
From 1993 to 1995, the Department of Justice imposed a hiring freeze that affected assistant U.S. Attorney positions (AUSA), as well as other law enforcement positions in the Federal Bureau of Investigation (FBI), the Drug Enforcement Administration (DEA), Customs, and the Immigration and Naturalization Service (INS).13 This freeze led to fewer prosecutions and, ultimately, to fewer offenders being sentenced to prison.14 The decrease occurred in the number of defendants sentenced to prison for all of the offenses examined in this paper except those involving immigration, fraud, and traffic offenses. Overall, defendants sentenced to prison fell by 4,940 offenders (down 9.8 percent).
As depicted in Figure 1, the number of persons received into postconviction supervision generally tracks the series for those sentenced to prison with a delay of about two years. Data for 1996 through 1998 show the only deviation from this pattern.15 During that period, the number of persons sentenced to prison increased 15 percent, but the number of persons entering postconviction supervision dropped 5 percent. This difference can be accounted for by noting that the median prison sentence for persons sentenced to prison in any year is about three years. Therefore, half of those offenders sentenced to prison will enter postconviction supervision by the third year after sentencing. During the three-year period from 1993 through 1995, the number of offenders sentenced to prison fell because of the hiring freeze for AUSAs and law enforcement personnel,16 a reduction that corresponds with the decline in the number of persons released from prison into postconviction supervision from 1996 through 1998.17
Under the influence of the guidelines, the yearly averages for sentence length in the early 1990s (69.9 months in 1991, 65.6 in 1992, 63.0 in 1993, 64.1 in 1994, and 66.4 in 1995), were longer, which meant that the dates on which offenders were released into terms of supervised released were pushed farther into the future.18 Lengthier sentences, particularly for those convicted of offenses involving drugs, firearms, and robbery, delayed the release of these offenders into postconviction supervision, thereby causing the received time series to be flat for several years. During that time, numbers of convicted persons sentenced to prison continued rising. Because one series (offenders sentenced to prison) was growing while the other (received into PSC) remained essentially stable, the distance between the two series increased. When those who had received longer prison sentences in the early 90s completed their respective periods of incarceration, the numbers of persons received for postconviction supervision began to rise again and the gap between the two series stabilized.
Over the entire period, the average growth per year in defendants sentenced is slightly more than twice as great as that for persons received (up 1,748 and 790, respectively). However, after 1998, the numbers for both series are virtually the same (average defendants sentenced = 2,153 per year, average persons received = 2,003 per year). Thus, by 1998, the two-year deviation beginning in 1996 had ended, and both series behaved similarly from 1998 through 2005, albeit with a substantially larger gap between them.
The number of persons sentenced to prison for immigration offenses increased over 1,800 percent from 1987 to 2005, rising from 791 to 15,068. Increases in this category occurred in every year during the period except 1992 (down 139 cases) and 2001 (down 1,073 cases). Figure 2 illustrates this trend and shows the acceleration. From 1994 to 1998, defendants convicted of immigration offenses rose 223 percent from 2,215 to 7,148. That trend continued after 1998. In 2005, 15,068 immigration defendants were convicted and sentenced to prison, an increase of 111 percent since 1998 and 580 percent since 1994.
For any other crime, particularly one with a very short prison sentence,19 such growth would have produced, after a two-year lag, an influx into postconviction supervision. But immigration is unlike any other offense category. During the period under discussion, the number of persons received into postconviction supervision for immigration offenses was flat; with the average increase only 69 per year from 1987 through 2005, and 170 per year from 1994 through 2005. Examining only the data for persons convicted of illegal entry and re-entry reveals a similar pattern.
The offenders in these categories are mostly illegal aliens who, when they are released from prison, are deported rather than placed under supervision. Therefore, while the number of offenders entering prison for immigration offenses was increasing, and was skyrocketing for the crimes of illegal entry and illegal re-entry, the number of persons received into postconviction supervision for these offenses was very low. This accounts for the steeper increase in the time series for those sentenced to prison as compared to the series for those received for supervision.20
The ratio of immigration offenders sentenced to prison to those sentenced directly to probation is 13.8, meaning that on average between 1990 and 2005, almost 14 times more offenders were sentenced to prison than sentenced to probation.21 The ratio of those sentenced to prison to those received into the probation system is 4.9.
Beginning in the Clinton administration and continuing into the Bush administration, the federal government directed significant resources to the Southwest border districts22 to attack the problem of illegal immigration. The first of these initiatives were Operation Gatekeeper in San Diego, California, and Operation Hold the Line, in El Paso, Texas, both implemented in 1994. These original initiatives were followed by Operation Rio Grande in McAllen, Texas, and Operation Safeguard in Tucson, Arizona.23 The Illegal Immigration Reform and Immigrant Responsibility Act of 199624 authorized a staffing increase of 5,000 positions in the U.S. Immigration and Naturalization Service, two-thirds of them in the Border Patrol.
One consequence of these policies was an increase in the number of immigration cases filed in the federal district courts in the Southwest border districts.25 In 1995, the number of defendants charged with violating immigration laws in the Southwest border districts was 1,292. In 2005, that number was 13,080, a 912 percent increase.26 Nationally (see Appendix Table 6), immigration offenders sentenced to prison jumped 375 percent (from 3,169 to 15,068) from 1995 to 2005.
The number of offenders sentenced to prison began to show marked growth around the mid-1990s, just when the increasingly rigorous government initiatives began. The ratio of those sentenced to prison to those sentenced directly to probation began a dramatic increase from 7.17 in 1994 to a high of 30.25 in 2004; i.e., in 1994, there were 7.17 times as many offenders sentenced to prison than sentenced directly to probation, but by 2004, that had increased to 30.25 times as many. During this time, the impact of the policy of deportation rather than postconviction supervision can be seen in the ratio of those sentenced to prison to those received into postconviction supervision. This ratio increased from 3.14 in 1994 to a high of 8.01 in 1998, then fell to 5.84 in 2005.
Shorter sentences have a more immediate impact on the postconviction supervision series, as entry occurs only a year or two into the future. An offender who serves a relatively short prison sentence is admitted into the probation system fairly soon after conviction. The crimes with relatively shorter sentences are larceny (see Appendix Table 1), embezzlement (see Appendix Table 2), forgery (see Appendix Table 3), fraud (see Appendix Table 4), and traffic offenses (see Appendix Table 5). Each of these offenses has an average sentence length of about two years or less. Each has a lower persons sentenced to prison/persons sentenced to probation ratio than do offenses with longer average sentences. The number of persons sentenced for each of these offenses generally remained fairly stable over the period. Fraud is the single exception, as the number of fraud offenders sentenced to prison more than doubled from 2,376 in 1987 to 5,336 in 2005, an average increase of 164 individuals per year. The number of persons received into the probation system for each of these offenses generally has shown a modest decline, with fraud the exception. The number of individuals received into postconviction supervision for fraud increased from 6,367 in 1987 to 8,123 in 2005, an average increase of 98 per year.
In contrast to the above-mentioned offenses, the following three crimes fall into the longer-sentence group: drug offenses (see Appendix Table 7), firearms offenses (see Appendix Table 8), and robbery (see Appendix Table 9). Each of these crimes had an average length of sentence during the study period of more than six years. These offenses also have the highest ratio of defendants sentenced to prison to those sentenced directly to probation and the highest average ratio of those sentenced to prison to those received into probation.
For drug offenses, the average sentence for the period was 80.6 months. For the last year in the study, 2005, that average was 85.7 months. Convictions for drug offenses thus took nearly seven years on average to affect the time series for entry into postconviction supervision. For drug offenses, the ratio of those sentenced to prison to those entering postconviction supervision was 1.18. The ratio of those sentenced to prison to those sentenced directly to probation was 9.53, indicating that most people found guilty of this offense served time in prison.
Firearms offenses also generally produced longer prison sentences. The average length of a prison term from 1987 through 2005 for firearms offenders was 79 months. Those sentenced to prison for firearms offenses began to increase beginning in 1998 (see Appendix Table 8). From 1987 to 1998, this number jumped 208 percent. From 1998 to 2005, the number rose 174 percent. The overall increase from 1990 to 2005 was 374 percent, most of which came after 1999. The growth in actual cases from 1987 to 1999 was 2,153 defendants, and from 1999 to 2005 was 4,884 defendants. The time series for those received into the probation system tracks this growth steadily with an increase of 194 percent from 1999 to 2005.
For firearms crimes, the average ratio of persons sentenced to prison to persons sentenced directly to probation with no prison time over the period 1987 through 2005 was 8.6. Thus, nine times as many offenders were sentenced to prison terms as were sentenced to probation only.
The average sentence for robbery from 1987 through 2005 was 103 months, the highest for any offense examined in this paper. The number of persons sentenced to prison for robbery began to decline in 1994. Appendix Table 9 depicts this drop, a subsequent rise, a longerterm leveling off, and a slight downturn in the 2000s. The number of persons sentenced to prison for robbery, on average, increased by 19 persons per year from 1987 through the end of the period. During that period, persons received into postconviction supervision who had been convicted of robbery offenses increased steadily. A spike occurred in 2004 due to the duplicate BOP cases, but over the period the average yearly increase was 49 individuals.
For robbery, the ratio of persons sentenced to prison to persons sentenced directly to probation was 37 from 1987 through 2005, indicating that on average, 37 times as many robbery offenders were sentenced to prison terms as were sentenced directly to probation. Thus, only rarely is someone convicted of robbery sentenced to probation rather than to prison and then to supervised release.
Figure 3 compares the numbers of short-sentence offenders (excluding those convicted of immigration crimes) sentenced to prison and received into postconviction supervision from 1990 through 2005. Both of these series are well-behaved and move almost as one throughout the entire period.
Figure 4 compares those sentenced to prison and received into postconviction supervision for the three long-sentence offenses of drugs, firearms, and robbery. These two series also generally follow each other closely.
The impact of the sentencing guidelines was felt on postconviction supervision around 1998, at which time the gap between these two series became considerably wider, but at this point the series were moving together, increasing at almost the same rate per year.
Figure 5 adds the immigration numbers to the long-sentence offenses. This causes data for both series to move upward, but shows that offenders sentenced to prison grew at a faster rate than did those received into probation. The number of immigration crime offenders sentenced to prison clearly accounted for the distance between the two series. Since the initial federal crackdown on immigration through Operations Gatekeeper and Hold the Line in 1994,27 the number of immigration offenders caught and sentenced to prison steadily increased, but the number of immigration offenders received into probation has not grown at the same rate, reflecting the government’s policy of deporting these offenders rather than putting them into post-conviction supervision.
This paper examines the relationship of the number of persons sentenced to prison and those received into postconviction supervision between 1987 and 2005. The study shows that the two time series move in the same direction, except for a two-year period, 1997 through 1998, yet exhibit an increasing divergence for several years after 1998. Although the gap between the two series stabilized after 2000, it has remained large. The anomalous two-year period which opened the gap reflects the hiring freeze affecting assistant U.S. attorneys and other federal law enforcement positions within the FBI, the DEA, Customs, and the INS from 1993 to 1995.
Two factors are examined here: 1) implementation of the federal sentencing guidelines and determinate sentencing for offenses committed after November 1, 1987, and 2) policies initiated in 1994 to curb the entry of illegal aliens into the United States along its South-western border. Both of these factors played a role in the widening gap between the numbers of persons sentenced to prison and the numbers received into PCS.
For persons convicted of the shorter-sentence offenses involving larceny, embezzlement, and traffic, the number received into postconviction supervision declined over the period studied, whereas the number sentenced to prison remained stable. Fraud offenders showed upward trends for both series beginning well before 1994 and continuing until 2003 for the number sentenced and 2004 for persons received into PCS.
Those received into postconviction supervision after serving terms in prison for immigration offenses declined from 1996 to 1998, going from 988 to 892. However, immigration offenders sentenced to prison increased from 1993 to 1995, going from 1,957 to 3,169. Thus, immigration convictions, which produced an average sentence of slightly over two years, played a part in the divergence of the two series between 1996 and 1998. For non-immigration offenses, persons who went to prison in 1993 (or 1994 or 1995), would on average be received into the probation system in 1996 (or 1997 or 1998). However, most immigration offenses are committed by illegal aliens who never enter the probation system, but are deported after they serve prison sentences.28 Their effect on the probation series was downward—that is, the series did not experience the influx of approximately 1,200 to 2,000 new individuals who had been imprisoned.
The ratio of those sentenced to prison to those sentenced to probation for immigration is similar for longer-sentence offenses. Immigration is the only short-sentence offense for which virtually everyone convicted receives a prison sentence rather than probation. The number of immigration offenders sentenced to prison is almost 14 times greater than those sentenced directly to probation.
The longer-sentence offenders convicted in federal courts of crimes involving drugs, firearms, and robbery had average sentences over the period ranging from about 80 months to 100 months and served an average prison term of 8 to 10 years before they were released into PCS. By 1998, when these long-sentence offenders started to be released into postconviction supervision, the numbers begin to assume the same slope as the count of those sentenced to prison. The initial divergence between the two series (sentenced to prison and received for supervision), which began in 1996 as a result of the preceding DOJ hiring freeze, was perpetuated and enlarged by this influx of longer-sentence offenders.
The two offenses responsible for the greatest percentage increase over the period involve immigration and firearms. As the federal offense category with the greatest number of offenders, almost all of whom serve long prison sentences, drug offenses had a profound effect on both the persons sentenced to prison time series and the number received into postconviction supervision time series. The numbers for drug and firearms crimes resumed their long-term growth in 1999. This, coupled with the continued increase in immigration cases, has driven up the number of persons sentenced to prison and the lengths of their sentences, as well as the number received into PCS. As the number of drug and firearms convictions increased, the postconviction supervision series recovered from the slight drop it showed from 1996 through 1998. As the number of immigration offenders sentenced to prison and then deported increased, the gap between the two series grew.29
Congress passes laws that affect the numbers of persons arrested for different offenses and the lengths of prison sentences. Executive branch agencies develop programs to implement these laws. Some of these policies affect prison populations, but may not impact post-conviction supervision. Decisions regarding the level of assistant U.S. attorney staffing, as well as personnel levels for the border patrol, customs, FBI, DEA, and other police-type entities, affect the numbers of persons eventually prosecuted and sent to prison, and, later, the number entering postconviction supervision or deported. National priorities, such as initiatives to close the Southwest border to illegal entry, also affect the relationship between the number sentenced to prison and the number received for supervision. All these factors must be taken into account when projecting the numbers of those brought under the supervision of the federal probation system. One defendant found guilty at trial reveals little of when—or if—he will enter postconviction supervision.