"Prediction is very difficult, especially if it's about the future," said Nobel Prize-winning physicist Niels Bohr, quoting a fellow Dane.
Earlier this year, Attorney General Michael Mukasey predicted that if Congress allowed new guidelines granting retroactive application of lower prison sentences to go into effect on March 3, up to "1,600 convicted crack dealers, many of them violent gang members, will be eligible for immediate release," with 3,800 eligible within the first year. Proponents of retroactivity accused the Attorney General of trying to scare the public into thinking the new law would be a "get-out-of-jail-free card" for all crack convicts, including career criminals. They cited Sentencing Commission projections that fewer prisoners-almost none of them repeat offenders-would be eligible for immediate early release.
The actual statistics may prove everyone wrong. As of April 2, the Federal Bureau of Prisons (BOP) had received 3,107 judicial orders for early release of crack convicts. Every workday since March 3, 135 felons, on average, have received sentence reductions from federal judges under the new guidelines.
So who is getting out of jail early? Are they first-time and nonviolent offenders, whom the Attorney General and others argued should be the sole beneficiaries of retroactivity? Are federal judges protecting public safety, as predicted by proponents of blanket retroactivity, and keeping violent and career criminals locked up? Have any of those just released committed new drug-related offenses-or other violent crimes-and been re-charged by state or federal authorities? Will any of the predictions prove correct?
No one knows the answers to these and other important questions, because no one is keeping track of the statistics. But dramatic changes in public policy, such as these new sentencing guidelines, need to be evaluated and studied to inform future policymaking, and any such study must be based on facts.
To inform future Sentencing Commission proceedings, deliberation by Congress, and the public debate, the Department of Justice should collect and regularly publish facts on the effect of the retroactivity provision, particularly as regards prison releases and recidivism. Further, Congress should require the department to provide these regular reports to the appropriate congressional committees to ensure that reporting does not lapse after a change in Administration or departmental priorities.
Until the Sentencing Commission's recent reduction in penalties related to crack possession, the sentencing guidelines contained a hundred-to-one disparity in penalty thresholds relating to crack versus those relating to powder cocaine. This disparity led, in some cases, to unjust results.
The disparity dates back to the Anti-Drug Abuse Act of 1986, which created a two-tiered system of minimum sentences for those convicted of drug offenses under federal law. Whether a drug dealer received a lower-tier sentence (five years) or an upper-tier sentence (10 years) depended on the weight of the drugs in the dealer's possession, based on the reasonable assumption that a major distributor will have access to greater quantities than a low-level dealer will have.
Cheap crack was swamping many American inner cities by 1986, bringing with it violence, addiction, and family disintegration. Congress took a hard line with people who distributed or possessed crack. Defendants caught with five grams of crack faced a five-year mandatory minimum sentence, and those with 50 grams received a 10-year sentence. Those caught possessing powder cocaine-a drug less associated with violence-received five- and 10-year sentences with 50 and 500 grams, respectively. At the time, the tough penalties for crack won wide support among politicians from both parties and the public at large.
At the same time, the U.S. Sentencing Commission was formulating its first set of federal sentencing guidelines. Congress had created the commission in 1984 to establish uniform mandatory sentences for crimes in hopes of deterring crime and ensuring equal justice in the federal criminal-justice system.The commission adopted Congress's hundred-to-one ratio as the basis for its 1987 guidelines on penalties relating to crack.
Over time, opinions on the crack-powder disparity shifted. The commission pushed Congress for lower crack sentences-though still significantly higher than those for powder cocaine-in 1995, 1997, and 2002. Congress refused to act.
In April of last year, the commission revised the sentencing guidelines for crack, and because Congress did not act to overturn the revised guidelines, they became law last November. Then, late last year, the Sentencing Commission further amended its guidelines to make the lower sentences retroactive. That change went into effect on March 3.
Earlier this year, Attorney General Mukasey urged Congress to pass legislation barring retroactivity for most individuals convicted under the old guidelines, warning that failure to act would make 1,600 convicted crack dealers eligible for immediate release and could lead to lower sentences for more than 20,000 crack dealers overall, threatening public safety. These estimates were calculated not by Mukasey or the Department of Justice, but by the Sentencing Commission.
Mukasey's testimony was roundly criticized. Editorial boards accused him of trying to "scare" Congress into acting and dismissed his alarm as "unwarranted." As these critics pointed out, any sentence reductions would have to be approved by a federal judge. They argued that judicial review would ensure that dangerous criminals wouldn't get out early.
It is inevitable, though, that some of the prisoners released early due to retroactivity will engage in the same kinds of conduct that led to their arrest, conviction, and subsequent incarceration. Many of the first-time and nonviolent offenders who get out early will likely be "scared straight" and focus on putting their lives back together by getting a job and staying out of trouble, but some of the criminals released early will return to their communities, take up arms, and commit more crimes of violence.
Certain factors can be useful in evaluating the likelihood that those who are released from prison will commit subsequent crimes, and it is often these kinds of statistics that are the basis for subsequent changes in policy. For example, it would be useful to know how many of those benefiting from retroactivity took advantage of educational opportunities while incarcerated. How many entered into job training programs to ease their way back into their communities? How many were turned away from re-entry programs because the system was not ready to handle a wave of releases?
No one knows the answers to these or other related questions, because the federal government is not keeping track. When a felon who benefited from retroactivity commits a violent crime-an eventuality that is inevitable-neither policymakers nor local law-enforcement officials will have the statistics at their disposal to evaluate the threat of retroactive releases to public safety and to respond to public concern in affected communities.
Justice in Practice
To understand the full effect of retroactivity, it is not enough to focus on the hyper-technical details of the sentencing guidelines without considering how prosecutors do their jobs and how retroactivity throws a wrench into the system.
Prosecutors exercise great discretion in developing their cases. Before they even begin an investigation, they know which crimes are "easy" to prove, which crimes take more time and resources to develop, and the sentences associated with every crime. Proving distribution or possession of crack is relatively easy. Thus, when choosing how to prosecute a defendant involved in a drug case, the prosecutor will often take the path of least resistance: prosecuting the drug charge without developing the gun or violent crime aspects of the case.
Prosecutors make deals with defendants because it is in the interest of fairness and prosecutorial economy. Over 90 percent of criminal cases are resolved by plea bargain. Defendants who plead guilty usually do so in exchange for the prosecution's dropping other charges. Once those charges are dropped, the government cannot go back and re-litigate them.
This complicates things. For example, a prisoner who received a 10-year sentence as part of a plea agreement for a crack offense but was not charged with a related violent crime or weapons offense due to prosecutorial discretion might get a few years knocked off of his sentence under the new guidelines, while a prisoner with a 10-year sentence for the same weapons offense alone would get nothing.
Knowing that the Sentencing Commission and Congress are willing to change the rules of the game after cases have been resolved introduces an element of uncertainty into the criminal justice system. With federal prosecutors already overstretched and some crime rates inching upwards, making their jobs more difficult will not help public safety. It also will not help prosecutors reach just outcomes by exercising their discretion instead of throwing the book at every defendant.
To understand how retroactivity will play out in practice, it is also important to consider the strategies of the criminal defense bar. All attorneys have a solemn obligation to represent their clients zealously, in an ethical and competent manner. A well-informed and zealous defense bar is one of the greatest strengths of the American criminal justice system. It is also innovative and able to build quickly on its legal successes, which could greatly increase the impact of retroactivity.
In January, the Sentencing Resource Counsel of the federal Office of Defender Services sent a memorandum of law to all Assistant Federal Defenders explaining how to exploit alleged loopholes in the new guidelines, such as by arguing that the law requires a full re-sentencing hearing for each eligible defendant. Despite the fact that the commission's new rules state that proceedings "do not constitute a full re-sentencing on the defendant," the memorandum makes a persuasive argument that the new guidelines are merely advisory and perhaps in violation of Supreme Court precedent.
The memorandum also urges counsel to ask the court for lower sentences for convicts who received sentences pursuant to the "career offender" or "armed career offender" enhancements. The memorandum presents legal arguments for treating "career offenders and armed career criminals" exactly the same as first-time, nonviolent offenders for the purposes of resentencing under the lowered guidelines. Indeed, it goes on to argue that these offenders may have a stronger argument for sentencing reductions because their sentences are likely to depart more from the sentencing range for nonviolent, non-career offenders. In addition, the memorandum suggests that all offenders may be eligible for sentence reductions far larger than those approved by the Sentencing Commission.
Some of these arguments have merit and will prevail before some judges. The new guidelines, a delicate blend of law and public policy, will be and should be tested in the courts. As a result, some career and violent offenders will be released, and the total number of convicts released in the near future could be much higher than the Sentencing Commission or anyone else predicted. This could have an impact on criminal activity in many communities.
The matter of greatest concern to legislators and the public is how retroactivity affects crime. Data on this particular question are sparse.
The Department of Justice should use its existing internal authorities to collect and report detailed statistics on felons who are affected by the retroactive application of sentencing reductions. Recidivism studies are expensive, however, and this study would be particularly expensive, according to a senior Justice Department official, because these felons will be released from federal prisons across the United States. Additionally, some crack felons are deportable because they are in the United States illegally and so will be difficult or impossible to track once deported. Congress should insist that the Department dedicate the appropriate funds to this important study.
The department should regularly report how many felons are resentenced due to retroactivity, the amounts of the sentence reductions, and how these new sentences comport with both the Sentencing Commission's guidelines on retroactivity and its new crack-related guidelines. It should also keep track of the pertinent characteristics of those who are released early, such as their statuses as first-time offenders, career criminals, or armed felons. The department should also record and report statistics on the criminal profiles of the individuals who receive reduced sentences, including any other offenses for which they were charged or convicted. Without these kinds of statistics, Congress and the public will not be able to evaluate whether the criminal-justice system has, in practice, followed the directions of the Sentencing Commission and whether courts have fulfilled their duty to give resentencing petitions careful scrutiny and deliberation.
Next, the department should collect and report statistics on what happens to felons who have been released early due to retroactivity. How many offenders who were released early got a job? How many took advantage of free educational programs while incarcerated? How many entered transitional job training programs once they were released?
Most important in evaluating the impact of retroactivity will be data on recidivism. Determining whether resentenced offenders are more or less likely to commit crimes, what kinds of subsequent crimes they are likely to commit, and the violent or nonviolent nature of those crimes would provide much guidance in determining whether and how to implement retroactivity in the future.
All of this information should be of special interest to Members of Congress-especially those who serve on the House and Senate Judiciary Committees. Congress should require the department to collect this information.
Without good statistics, the public and policymakers will have no way to evaluate the impact of this drastic policy change and make informed decisions about similar matters in the future. It is crucial that the effect of the retroactivity experiment on which the nation has embarked be recorded, analyzed, and reported.
Despite some opposition in Congress and substantial concerns by the Department of Justice, Congress allowed the Sentencing Commission to apply its reductions in sentencing guidelines for crack-related offenses to those who have already been convicted, including violent and repeat offenders. Congress and the public must have the facts in order to evaluate whether this change in sentencing policy was sound public policy.
The Department of Justice should collect and regularly publish facts on the effect of the retroactivity provision, particularly as regards prison releases and recidivism. Predicting the future is very difficult; judging policy choices in the past is easier when you have all the facts.
Charles D. Stimson, a former prosecutor and defense attorney, is Senior Legal Fellow and Andrew M. Grossman is Senior Legal Policy Analyst in the Center for Legal and Judicial Studies at The Heritage Foundation.
 See Letter to the Editor, Future Imperfect, Economist, Jun. 28, 2007. The quote has been attributed to, among others, Samuel Goldwyn and Yogi Berra. Id. Bohr, who apparently uttered the line in response to a question about the Heisenberg Uncertainty Principle, attributed it to the Danish artist and writer Robert Storm Petersen. The Perils of Prediction, Economist.com, July 15, 2007, at http://www.economist.com/blogs/theinbox/2007/07/
 Darryl Fears, Crack-Sentencing Reductions Decried, Washington Post, Feb. 7, 2008, p. A2.
 See, e.g., Editorial, Toward Drug Case Justice, N.Y. Times, Feb. 9, 2008.
 Anti-Drug Abuse Act of 1986, Pub. L. No. 99-570, 100 Stat. 3207 (codified as amended at 21 U.S.C. § 841(b) (2000)).
 See Drug Enforcement Agency, U.S. Dept. of Justice, DEA History: 1985-1990, at http://www.usdoj.gov/dea/pubs/history/1985-1990.html (April 1, 2008).
 Anti-Drug Abuse Act of 1986 Pub. L. No. 99-570, 100 Stat. 3207 (1986).
 Sentencing Reform Act of 1984, Pub. L. No. 98-473, 98 Stat. 1837 (1984).
 U.S. Sentencing Comm'n, supra note 9.
See Michael Coyle, Race and Class Penalties in Crack
Cocaine Sentencing, The Sentencing Project, Aug. 2003, at http://www.sentencingproject.org/Admin/Documents/
 U.S. Sentencing Comm'n, U.S. Sentencing Commission Votes To Amend Guidelines for Terrorism, Sex Offenses, Intellectual Property Offenses, and Crack Cocaine Offenses, Press Release, April 27, 2007, at http://www.ussc.gov/PRESS/rel0407.htm.
 U.S. Sentencing Comm'n, U.S. Sentencing Commission Votes Unanimously To Apply Amendment Retroactively for Crack Cocaine Offenses, Press Release, Dec. 11, 2007, at http://www.ussc.gov/PRESS/rel121107.htm.
 Fears, Crack-Sentencing Reductions Decried, supra note 2.
 N.Y. Times, supra note 3.
 See Timothy Lynch, The Case Against Plea Bargaining, Regulation, Fall 2003, p. 24, at /static/reportimages/21B8A1C45C533DA5866B6BCD410EFDC4.pdf.
 Memorandum from the Sentencing Resource Counsel, Office of Defender Services, to All Defenders regarding Sentence Reductions Under the Retroactive Crack Amendment (Jan. 2, 2008), at /static/reportimages/8ED0BCDC3904D3F648264B2056E333A5.pdf.