The question the statute answers

The First Step Act answers a question Congress had asked in different ways for more than three decades: what should be done about a federal sentencing system that imposes the longest drug sentences in the industrialized world while housing only a fraction of the people imprisoned in the United States. The statute’s answer was not a single reform but a bundle of them, and the most consequential part of the bundle was not the part its sponsors discussed most. The public debate centered on mandatory minimums for drug offenses, but the provision that ultimately released the most people was a quiet procedural amendment tucked into Title VI, which allowed federal prisoners to ask a court for compassionate release directly rather than waiting for the prison bureau to ask on their behalf.

That asymmetry is the key to reading the act. The sentencing changes were carefully bounded, prospective in nearly every respect, and drafted to reach only defined categories of offenders. The credit system that rewards program participation reaches further but carries a long list of exclusions that determines its real scope. The good conduct time correction touched every federal prisoner immediately. And the compassionate release change, which attracted almost no attention during the legislative debate, became the principal legal mechanism for early release during the public health emergency that began about fifteen months after the statute took effect. To understand the act is to understand all five of these mechanisms and the different groups of people each one touches.

This article carries the full treatment in one place because the statute profile framework requires it. It covers the act’s identity, the conditions that produced it, how it moved through Congress, the architecture of the text title by title, who it covers and who it exempts, how the executive branch put it into practice, the litigation that has construed it, what changed in practice, and the honest account of what it did not do. Readers who want the comparison with the 1994 crime statute will find it in the companion comparison article; the provision made retroactive is treated at length in the article on the 2010 Fair Sentencing Act.

The formal identity of the First Step Act

The short title is the First Step Act of 2018. The public law number is 115-391. It was signed on December 21, 2018, by the 115th Congress, passed as S. 756, with the Senate concurring in the House amendment on December 18, 2018 by 87 votes to 12 and the House agreeing to the Senate amendment on December 20, 2018 by 358 votes to 36. The official short title in the enrolled text is the “First Step Act of 2018,” and it is the first and to date the only statute of its kind to pair sentencing reduction with a programmatic recidivism-reduction system in federal prisons. The pairing is the statute’s signature: earlier reforms had addressed sentencing or prisons, but none had combined the two in a single legislative package, and the combination is what made the coalition possible.

The statute runs to six titles. Title I establishes the risk and needs assessment system and the recidivism reduction programming framework, including the earned time credit system. Title II addresses secure firearms storage for correctional officers. Title III restricts restraints on pregnant prisoners. Title IV carries the sentencing reforms: reduced and restricted enhanced mandatory minimums for repeat drug offenses, a broadened safety valve, the clarification of the firearm stacking provision, and the retroactive application of the 2010 Fair Sentencing Act. Title V reauthorizes the Second Chance Act of 2007. Title VI carries miscellaneous provisions, including the home confinement expansion, the elderly offender pilot reauthorization, data collection, health care products, and the modification of imposed terms of imprisonment that rewrote compassionate release. The cluster of sentencing provisions is concentrated in Title IV, while the provisions that changed daily life inside federal prisons are concentrated in Title I and Title VI.

The enrolled bill also carries two short titles for portions of the act: the Lieutenant Osvaldo Albarati Correctional Officer Self-Protection Act of 2018 for Title II, named for a federal correctional officer killed in the line of duty, and the Second Chance Reauthorization Act of 2018 for Title V. The presence of named short titles inside the larger act reflects the legislative reality that the statute is a vehicle: provisions that could not have passed on their own rode inside a package that could. The vehicle metaphor also explains the act’s uneven drafting. The sentencing provisions are precise and heavily negotiated, the product of years of committee work. The prison reform provisions are broader and more delegating, leaving the Department of Justice and the Bureau of Prisons to fill in the operational details. The miscellaneous title is a collection of provisions that various members wanted and that found a home in the only criminal justice vehicle moving that Congress.

What problem was the good time correction solving?

The good conduct time system illustrates how a small drafting ambiguity can compound over decades. The Sentencing Reform Act of 1984 provided that a prisoner serving more than a year could earn up to 54 days of credit per year for exemplary compliance with institutional rules. The Bureau of Prisons read “per year” as per year of time actually served, which meant that as the credit accumulated and the release date moved earlier, the base on which the credit was calculated shrank, producing the roughly 47-day effective rate. Prisoners litigated the reading for years, and the Supreme Court sustained the bureau in Barber v. Thomas in 2010, with the majority finding the statutory language ambiguous and the bureau’s interpretation reasonable.

The 2018 act’s correction is a single sentence in section 102(b), but it resolved a dispute that had consumed thousands of hours of litigation and administrative calculation. The bureau’s recalculation applied to the entire population at once, which required reprogramming the sentence computation system that tracks every federal prisoner’s release date. The episode is a useful illustration of the statute’s theory of change: where the bureau’s discretion had produced a restrictive reading, Congress did not order the bureau to reconsider but changed the underlying text so that the restrictive reading was no longer available.

The codification touches dozens of sections of the United States Code, with the most heavily worked provisions being 18 U.S.C. section 3582(c), the compassionate release provision; 18 U.S.C. section 3624(b), the good conduct time computation; 18 U.S.C. section 924(c), the firearm offense and its consecutive sentences; 18 U.S.C. section 3553(f), the safety valve; and the controlled substances mandatory minimums at 21 U.S.C. section 841. The new risk and needs assessment system was added to chapter 229 of title 18 as a new subchapter.

What Congress was legislating against

The federal prison population in December 2018 stood at roughly 180,000 people, down from a peak of about 219,000 in 2013 but still nearly eight times the level of 1980, when the federal system held fewer than 25,000. That growth had been driven overwhelmingly by drug sentences. Nearly half of all federal prisoners were serving time for drug offenses, and the mandatory minimums enacted in the 1980s guaranteed long fixed sentences that judges could not reduce regardless of the circumstances of the offender. The Controlled Substances Act, as amended by the Anti-Drug Abuse Acts of 1986 and 1988, set mandatory minimums of five, ten, and twenty years, with life for a third conviction, and the sentencing guidelines translated drug quantities into additional severity through the drug quantity table.

The federalization of drug enforcement is itself part of the story. Before the 1980s, drug prosecution was overwhelmingly a state matter, and the federal system handled a small number of large trafficking cases. The mandatory minimums changed the incentives: federal prosecutors could guarantee long sentences that state courts could not, and the resulting shift of cases into federal court drove the population growth that the 2018 act was written against. The racial dimension compounded the effect. Federal drug enforcement concentrated on crack cocaine markets in Black neighborhoods, while powder cocaine, pharmacologically identical, was policed less aggressively in whiter markets. The Sentencing Commission’s demographic data showed that the federal crack docket was overwhelmingly Black, and the 100-to-1 ratio converted that enforcement pattern into a sentencing disparity that the commission called the most visible racial injustice in the federal system.

The fiscal dimension gave the reform coalition its conservative wing. Housing a federal prisoner cost roughly $36,000 per year by the late 2010s, and the Bureau of Prisons consumed a growing share of the Justice Department’s budget, crowding out funding for prosecutors, agents, and the courts themselves. The argument that mass incarceration was fiscally unsustainable did not persuade everyone, but it gave Republican members a reason to support reform that had nothing to do with the racial justice arguments that persuaded Democrats. The coalition’s durability depended on both arguments operating at once, and the statute’s design reflects the compromise between them: sentencing relief for the cases both sides agreed were unjust, programming and incentives that both sides could describe as public safety policy.

The enhancement structure for prior convictions had grown harsher through interpretation rather than legislation. Under 21 U.S.C. section 841 as it stood before the act, a prior “felony drug offense” triggered the enhancements, and the courts had read that phrase to include low-level state drug convictions, sometimes decades old, that bore little resemblance to the serious trafficking offenses the enhancement was designed to reach. The result was that a defendant with two minor state drug convictions from the distant past could face a mandatory life sentence for a subsequent federal trafficking offense, a outcome that had produced some of the most widely publicized life sentences in the federal system and that judges across the ideological spectrum had criticized from the bench.

The firearm provision had grown through the same dynamic of accumulation. Section 924(c) imposes consecutive mandatory minimums for possessing, brandishing, or discharging a firearm during and in relation to a crime of violence or drug trafficking crime: five years for possession, seven for brandishing, ten for discharging. In Deal v. United States, decided in 1993, the Supreme Court held that a second or subsequent 924(c) conviction could be charged in the same indictment as the first, so that a defendant facing multiple counts in a single case would receive consecutive 25-year enhancements on top of the first count. The practice became known as stacking, and it produced sentences of fifty, sixty, or even more than a hundred years for defendants who had never been convicted of anything before the case at bar. The Sentencing Commission, the Judicial Conference, and the American Bar Association had all identified stacking as the single most distorting feature of the federal mandatory minimum structure.

The safety valve, enacted in 1994 at 18 U.S.C. section 3553(f), allowed judges to sentence below a mandatory minimum for certain low-level drug defendants who met five criteria: limited criminal history, no violence or weapon, no serious injury, no leadership role, and truthful disclosure. But the criminal history criterion was drawn so narrowly that defendants with modest records could not use it, and the federal defender community had documented cases in which a defendant’s history of minor offenses excluded him while the underlying conduct was minor.

The Fair Sentencing Act of 2010 had reduced the crack-to-powder-cocaine sentencing ratio from 100-to-1 to 18-to-1, but Congress had made the change prospective only. Thousands of prisoners sentenced before August 3, 2010 continued to serve under the old ratio, a group whose sentences had been widely acknowledged as racially disparate in effect. The African American share of federal crack offenders had been overwhelming under the old ratio, and the refusal to make the 2010 change retroactive had been a standing item on the reform agenda for eight years.

The prison reform side had its own separate history. The Second Chance Act of 2007 had created reentry programming and a pilot for elderly offenders, and the federal prison system had built an extensive but uneven patchwork of vocational, educational, and drug treatment programs. Good conduct time, which the Sentencing Reform Act of 1984 had set at up to 54 days per year, was being calculated by the Bureau of Prisons at 54 days per year of time actually served rather than per year of the sentence imposed, a method that reduced the credit to roughly 47 days per year of the stated sentence. Litigation had upheld the bureau’s reading in Barber v. Thomas in 2010, so only a statute could change it. Compassionate release, at 18 U.S.C. section 3582(c)(1)(A), existed on paper but required the Bureau of Prisons to file the motion; the bureau had filed a handful per year, and a 2013 Justice Department Inspector General report had documented that the bureau frequently failed to inform prisoners of the program or to respond to requests.

How did the mandatory minimum architecture get built?

The mandatory minimums the act narrowed were constructed in a four-year burst of legislation between 1986 and 1990. The Anti-Drug Abuse Act of 1986 established the quantity-triggered structure that still governs federal drug sentencing: five grams of crack cocaine or 500 grams of powder cocaine triggered a five-year mandatory minimum, and fifty grams of crack or five kilograms of powder triggered ten years. The act also created the 100-to-1 quantity ratio between crack and powder cocaine, a ratio that had no pharmacological basis and that fell overwhelmingly on Black defendants, because crack was the form of cocaine most available in the low-income urban markets that federal enforcement targeted. The Anti-Drug Abuse Act of 1988 extended mandatory minimums to simple possession of crack, the only drug for which federal law imposed a mandatory minimum for possession without intent to distribute, and added the death penalty for certain large-scale drug offenses.

The 1986 act passed the House by voice vote and the Senate by 97 to 2 in the weeks before a midterm election, a legislative context that helps explain the severity of the penalties and the thinness of the deliberation. Members competed to appear tougher on drugs, amendments ratcheted the quantities downward and the penalties upward, and the conference process produced a bill more severe than either chamber’s original version. The speed of the process meant that basic questions, such as whether the 100-to-1 ratio had any empirical foundation, were never seriously examined. The ratio became the single most consequential number in federal sentencing for the next quarter century, and undoing it took until 2010.

What did the Sentencing Reform Act leave behind?

The Sentencing Reform Act of 1984, which created the United States Sentencing Commission and the federal sentencing guidelines, is the framework the First Step Act amends, and its history explains why the 2018 act looks the way it does. The 1984 act abolished parole in the federal system and replaced it with the 85 percent truth-in-sentencing rule, required the guidelines to be followed except in specified circumstances, and created the good conduct time system the 2018 act later corrected. The guidelines translated drug quantities into offense levels through the drug quantity table, so that a defendant’s sentence was driven primarily by the weight of drugs attributed to him, often including quantities from relevant conduct that were never charged or proven to a jury.

The guidelines were mandatory until the Supreme Court’s 2005 decision in United States v. Booker, which made them advisory while preserving the mandatory minimums enacted by Congress. The result was a hybrid system: judges gained discretion over the guidelines range but remained bound by the statutory floors, which meant that the mandatory minimums, rather than the guidelines, became the binding constraint in the most serious cases. The First Step Act operates inside this hybrid system. It adjusts the statutory floors and the mechanisms for earning time off, but it does not restore parole, does not make the guidelines mandatory again, and does not give judges general authority to revisit old sentences.

Why did the crack ratio survive for so long?

The Sentencing Commission recommended eliminating or sharply reducing the crack-to-powder disparity four times between 1995 and 2007, and Congress rejected or ignored each recommendation. The 1995 recommendation, which would have equalized the ratio, was affirmatively disapproved by Congress, the first and only time Congress has rejected a guidelines amendment. The commission’s 2002 report documented the racial disparity in exhaustive detail and recommended a 20-to-1 ratio. The 2007 report recommended further reductions. Each report moved the consensus without moving the statute, because changing the ratio required an act of Congress and because the politics of appearing soft on crack cocaine remained toxic.

The Fair Sentencing Act of 2010 broke the stalemate by reducing the ratio to 18-to-1 and eliminating the five-year mandatory minimum for simple possession of crack. It passed with bipartisan support and the President’s signature, but the compromise that secured its passage was prospectivity: the act applied only to sentences imposed after August 3, 2010. The thousands of prisoners serving under the old ratio were left in place, and the reform community made retroactivity its central demand for the next eight years. Section 404 of the First Step Act is the answer to that demand, and its drafting reflects the compromise: courts may reduce sentences but are not required to, and defendants already resentenced under the 2010 act are excluded.

Where did the prison reform half come from?

The prison reform provisions of the act have a separate lineage from the sentencing provisions, and they arrived in the statute through a separate bill. The Second Chance Act of 2007, signed by President George W. Bush, created federal reentry programming, mentoring grants, and the elderly offender early release pilot, and it established the principle that the federal government should invest in reducing recidivism rather than merely warehousing prisoners. The research base for that principle had been building for decades: studies of correctional programming consistently found that education, vocational training, and drug treatment reduced reoffending, while idleness and overcrowding increased it.

The state reform wave of the 2000s and 2010s provided the political proof of concept. Texas, under a Republican legislature and governor, invested in treatment and diversion rather than new prison construction and saw its incarceration rate fall while crime continued to decline. Georgia, South Carolina, and Mississippi enacted similar reforms with similar results. Conservative organizations, notably Right on Crime, framed these reforms as fiscally responsible and consistent with redemption values, which gave Republican members of Congress political cover to support federal reform. The House’s 2018 prison reform bill drew directly on this state experience, and its 360-to-59 passage margin showed that prison reform, unlike sentencing reform, commanded a genuine bipartisan consensus.

Why did the coalition hold together this time?

The coalition that carried the act was unusual: conservative advocacy organizations that framed the bill as redemption policy, progressive civil rights groups that framed it as racial justice policy, prosecutors who supported the recidivism reduction machinery while resisting sentencing reduction, and a White House that made the bill a personal legislative priority of the President. The coalition held because each side got something and because the sentencing provisions were bounded enough that law enforcement opposition, which had killed earlier versions, was blunted rather than inflamed. The critical drafting choice was to keep the sentencing changes narrow and prospective while making the prison reform provisions broad, a trade that secured the votes of senators who would not have supported a sentencing reform bill standing alone.

How it moved through Congress

The bill that became the act began in the House as H.R. 5682, the FIRST STEP Act, introduced by Representatives Doug Collins and Hakeem Jeffries and focused almost entirely on the prison reform side: the risk assessment system, programming, and reentry provisions. The House passed that version on May 22, 2018 by 360 votes to 59, an overwhelming margin that nevertheless left the sentencing provisions out entirely. The Senate’s criminal justice negotiators, led by Chuck Grassley, Dick Durbin, Cory Booker, and Mike Lee, had been working on a companion sentencing bill, and the decisive legislative maneuver of the year was the decision to merge the two: the Senate took up S. 756, which had begun life as an unrelated marine debris bill, substituted the combined prison reform and sentencing text as a House amendment, and drove it through the Senate in the lame duck session.

The Senate’s procedural path was contested. Majority Leader Mitch McConnell had been reluctant to schedule floor time, and the bill’s supporters forced the question through a combination of White House pressure and an organized whip effort that produced public commitments from a filibuster-proof majority. On December 17, 2018, the Senate invoked cloture on the motion to concur by 82 votes to 12. Two amendments were then considered that would have narrowed the act’s reach: amendments sponsored by Tom Cotton and John Kennedy that would have expanded the exclusion list and required victim notification of release dates. The Durbin statement on the floor described the notification requirements as duplicative of the Crime Victim Rights Act, and the Senate rejected the amendments by 62 to 37. A Cruz amendment expanding the list of offenses excluded from time credits was agreed to by voice vote, which modestly lengthened the exclusion list in the final text.

The Senate concurred in the House amendment with its own amendment on December 18, 2018 by 87 votes to 12, Record Vote Number 271. The House, which had to accept the Senate’s added sentencing provisions, took up the bill under suspension of the rules, which required a two-thirds majority, and agreed on December 20, 2018 by 358 votes to 36, Roll Call Number 448. The President signed the bill on December 21, 2018, and it became Public Law 115-391. The entire sequence from cloture to signature took four days, a speed that reflected how thoroughly the negotiations had been completed before the floor action began.

What were the two bills that became one?

The statute that emerged in December was a merger of two separate legislative efforts that had been running on parallel tracks all year. The first was the House prison reform bill, H.R. 5682, introduced by Doug Collins, a Georgia Republican, and Hakeem Jeffries, a New York Democrat. It contained the risk assessment system, the programming expansion, the earned time credits, and the reentry provisions, but no sentencing changes at all. The House Judiciary Committee reported it with broad bipartisan support, and the full House passed it on May 22, 2018 by 360 votes to 59. The margin demonstrated that prison reform was the easy half of the coalition’s agenda.

The second effort was the Senate sentencing bill, which had been introduced in various forms since 2015 and which carried the mandatory minimum reductions, the safety valve expansion, the stacking fix, and the Fair Sentencing Act retroactivity. Its sponsors, Grassley, Durbin, Booker, and Lee, had negotiated the text through multiple congresses, and the bill had repeatedly stalled on the question of whether the Senate would ever give it floor time. The decisive move came in the lame duck session, when the negotiators and the White House agreed to combine the two: the Senate would take S. 756, a shell bill that had passed the Senate the previous year on an unrelated subject, substitute the combined text as a House amendment, and drive the package through both chambers before the 115th Congress adjourned.

How was the floor fight won?

The Senate’s procedural path ran through the reluctance of the Majority Leader, who had not scheduled the bill and who faced pressure from members who wanted it buried. The bill’s supporters countered with a whip count showing a filibuster-proof majority and with direct White House engagement: the President had endorsed the bill publicly and pressed for a vote, which made it difficult for the leadership to keep it off the floor without an open break. On December 17, 2018, the Senate invoked cloture on the motion to concur by 82 votes to 12, Record Vote Number 267, a margin that ended any doubt about the outcome.

Two amendments tested the coalition’s discipline. The Cotton-Kennedy amendments would have expanded the list of offenses excluded from earned time credits and imposed new victim notification requirements for release dates. Durbin’s floor statement argued that the notification provisions duplicated the Crime Victim Rights Act and that the expanded exclusions contradicted positions the amendments’ own supporters had taken earlier in the process. The Senate rejected the amendments by 62 to 37. A Cruz amendment expanding the exclusion list in a narrower way was agreed to by voice vote, which modestly lengthened the ineligibility catalogue in the final text without altering the bill’s structure. The Senate then concurred in the House amendment with its own amendment on December 18 by 87 votes to 12, Record Vote Number 271.

The House faced a different procedural problem: it had to accept sentencing provisions it had never voted on, and the suspension calendar, which the leadership chose for speed, required a two-thirds majority. The vote on December 20, 2018 cleared that threshold comfortably at 358 to 36, Roll Call Number 448. Members who had voted for the prison-only bill in May were now voting for the combined package, and the lopsided margin showed that the sentencing additions cost the bill almost no House support.

What did the signing set in motion?

The President signed the bill on December 21, 2018, and it became Public Law 115-391. The signature triggered a series of statutory deadlines that determined the pace of implementation. The Attorney General had 210 days to develop and release the risk and needs assessment system, which produced PATTERN in July 2019. The Bureau of Prisons had two years to implement the system at its facilities, with the earned time credit provisions phased in as the programming expanded. The good conduct time correction took effect immediately, and the bureau began recalculating release dates within weeks. The compassionate release amendment took effect on enactment, and the first prisoner-filed motions reached the courts in early 2019, months before anyone understood how large that docket would become.

The architecture of the text

What does the sentencing title actually change?

Title IV reduced the mandatory minimums for repeat drug offenses from twenty years to fifteen and from life to twenty-five, narrowed the prior convictions that trigger them, expanded the safety valve for low-level defendants, ended the stacking of firearm enhancements in a single indictment, and made the 2010 crack sentencing changes retroactive. All but the last apply prospectively.

Title IV of the act carries four sentencing provisions, and each one works differently. Section 401 reduces and restricts the enhanced mandatory minimums for repeat drug offenders. Before the act, a defendant with one prior felony drug conviction faced a mandatory minimum of 20 years for a high-level drug offense, and a defendant with two or more such convictions faced mandatory life. Section 401 reduced the 20-year minimum to 15 years and the mandatory life minimum to 25 years, and just as importantly it narrowed the triggering priors: the enhancement applies only where the prior convictions are for a “serious drug felony” or a “serious violent felony,” as defined in the new section 851 note, rather than for any felony drug offense. The practical effect was to exclude the low-level state drug convictions that had been driving the most extreme enhancements, and the change was prospective, applying to sentences imposed after the date of enactment.

Section 402 broadened the safety valve at 18 U.S.C. section 3553(f). Under the old rule, a defendant could not use the safety valve if he had more than one criminal history point. Section 402 raised that ceiling to four criminal history points, with a further refinement: a defendant with four points is excluded if the points arise from a prior three-point offense or a prior two-point violent offense. The effect was to make the safety valve available to a substantially larger group of low-level drug defendants, particularly defendants whose records consist of minor offenses that accumulated points without reflecting serious criminality. The Sentencing Commission’s analyses of safety valve usage before and after the change documented a marked expansion in the share of drug defendants who qualified.

Section 403 clarified 18 U.S.C. section 924(c) to end stacking. Under the amendment, the enhanced consecutive term for a second or subsequent 924(c) conviction applies only where a prior 924(c) conviction has become final before the second offense was committed. Multiple counts charged in the same indictment can no longer generate consecutive 25-year enhancements. This was the provision with the largest effect on the longest federal sentences: defendants who had received fifty-year or longer sentences for stacked counts became eligible, going forward, for sentences that reflected one enhanced term rather than several. Like section 401, the change was prospective and did not reopen sentences already imposed.

Section 404 applied the Fair Sentencing Act of 2010 retroactively. A defendant convicted of a crack cocaine offense and sentenced before August 3, 2010 may petition the court for resentencing as if the 2010 act had been in effect at the time of the original sentencing. This was the only retroactive sentencing provision in the statute, and it produced several thousand sentence reductions in the years after enactment, a figure documented by the Sentencing Commission’s retroactivity reporting. The provision was written to exclude defendants whose sentences had already been reduced under the 2010 act and to leave the district court’s discretion intact: the court may reduce the sentence but is not required to do so.

How do the earned time credits work?

Eligible prisoners earn ten days of credit for every thirty days of completed recidivism reduction programming, or fifteen days if classified as minimum or low risk on two consecutive assessments. Credits apply toward transfer to prerelease custody or up to twelve months of early supervised release. Prisoners convicted of listed offenses cannot earn credits at all.

The recidivism reduction framework sits in Title I. Section 101 directs the Attorney General to develop a risk and needs assessment system, which the Department of Justice named PATTERN, the Prisoner Assessment Tool Targeting Estimated Risk and Needs. The system assigns each federal prisoner a risk category of minimum, low, medium, or high, and identifies programming needs in areas such as education, vocational training, substance abuse treatment, and mental health. Section 102 requires the Bureau of Prisons to implement the system and to expand recidivism reduction programming, defined as evidence-based programming demonstrated to reduce recidivism, and productive activities, defined as activities designed to keep prisoners occupied and to reduce the likelihood of misconduct.

The incentive that gives the system its force is the earned time credit. Under 18 U.S.C. section 3632(d)(4), a prisoner who successfully completes recidivism reduction programming or productive activities earns 10 days of time credits for every 30 days of successful participation. A prisoner classified as minimum or low risk who has maintained that classification over two consecutive assessments earns an additional 5 days per 30 days, for a total of 15. The credits are applied toward prerelease custody, meaning transfer to a residential reentry center or home confinement, or toward up to 12 months of early supervised release. Credits cannot be earned retroactively for programming completed before the act’s enactment, and credits cannot be earned for programming completed while a prisoner was detained pending adjudication of the case.

The exclusions determine the provision’s real scope. Section 3632(d)(4)(D) lists the convictions that make a prisoner ineligible to earn time credits at all: a long catalogue that includes terrorism offenses, murder and assault, kidnapping, robbery, firearms offenses including repeat felon in possession, sex offenses and sexual exploitation offenses, human trafficking, espionage, and specified high-level drug offenses. Prisoners subject to a final order of removal under the immigration laws are also ineligible. The list runs to dozens of categories, and the practical result is that the credit system reaches most strongly the nonviolent drug and property offenders who make up the largest programming-eligible population in federal custody, while excluding the violent and high-risk categories entirely.

What did the good conduct time fix change?

The good conduct time correction is one of the shortest provisions in the statute and one of the most broadly felt. Section 102(b) amended 18 U.S.C. section 3624(b) to clarify that a prisoner may earn up to 54 days of good conduct time per year of the sentence imposed, rather than per year of the sentence actually served. The Bureau of Prisons had been calculating the credit on time served, a method that yielded roughly 47 days per year of the stated sentence, and the Supreme Court had upheld that method in Barber v. Thomas in 2010. The amendment changed the computation for every federal prisoner serving a sentence of more than one year, effective immediately, and the bureau applied the corrected calculation across the population. For a prisoner serving a ten-year sentence, the difference amounts to approximately 70 additional days of credit, and across the federal population the change accelerated the release dates of tens of thousands of people.

What did the safety valve look like before the expansion?

The safety valve was created by the Violent Crime Control and Law Enforcement Act of 1994, the same statute this series covers in the 1994 article, as a limited exception to the mandatory minimums for low-level drug defendants. Under 18 U.S.C. section 3553(f), a judge may sentence below the mandatory minimum if the defendant meets five criteria: no more than one criminal history point, no violence or credible threats of violence, no death or serious bodily injury, no leadership role in the offense, and full and truthful disclosure of the defendant’s knowledge of the offense. The criminal history criterion was the binding constraint: a defendant with even a modest record was disqualified.

Section 402 of the First Step Act raised the ceiling from one criminal history point to four, with two refinements: the defendant may not have a prior three-point offense, and may not have a prior two-point violent offense. The refinements target the provision’s intended beneficiaries: defendants whose records reflect repeated low-level involvement rather than serious or violent criminality. The Sentencing Commission’s data showed that the expansion roughly doubled the number of defendants eligible for the valve, and the provision’s prospective application means its full effect appears in the sentencing data year by year as new cases are sentenced under the broader rule. The safety valve is the provision that most directly affects the ordinary drug defendant, the courier or street-level seller with a modest record, rather than the kingpin the mandatory minimums were nominally aimed at.

What programming counts for the credits?

The statute distinguishes two categories of activity. Recidivism reduction programming is evidence-based programming demonstrated to reduce recidivism or to be effective in reducing crime or juvenile delinquency, and it is this category that earns time credits. Productive activities are activities designed to keep prisoners productively occupied, which may reduce misconduct and idleness but which have not been shown to reduce reoffending; productive activities do not earn credits but can affect other aspects of a prisoner’s classification and placement. The distinction matters because the bureau’s program catalogue is large and uneven: education, vocational training, and drug treatment programs have the strongest evidence base, while work assignments and some enrichment activities fall into the productive category.

The bureau maintains a list of approved evidence-based programs, and the statute requires the bureau to expand the availability of such programming so that every eligible prisoner has the opportunity to participate. The expansion requirement is the system’s soft underbelly: the bureau’s staffing and facility constraints have produced waitlists, and prisoners who cannot get into programming cannot earn credits regardless of their willingness. The Government Accountability Office’s audits have tracked the bureau’s progress in expanding capacity, and the bureau’s public reporting on program completion has become one of the principal measures of the statute’s implementation.

What changed in compassionate release?

Before the act, only the Bureau of Prisons could file a compassionate release motion, and the bureau’s refusal was unreviewable. Section 603 amended 18 U.S.C. section 3582(c)(1)(A) so that prisoners may petition the court directly after exhausting administrative remedies or after thirty days of warden inaction. That procedural change produced thousands of court-ordered releases.

Section 603 of the act rewrote 18 U.S.C. section 3582(c)(1)(A), the compassionate release provision, in a way that attracted little attention during the legislative debate and then mattered enormously. Before the act, only the Bureau of Prisons could file a motion for compassionate release; the prisoner could request that the bureau file, but the bureau’s refusal was effectively unreviewable, and the bureau granted a small number of motions per year. Section 603 added a provision allowing the prisoner to file the motion directly with the court after exhausting administrative remedies or after 30 days have passed without action on a request to the warden.

The change converted compassionate release from an administrative favor into a judicial remedy. Courts began to receive motions directly from prisoners and to decide them on the statutory standard of extraordinary and compelling reasons, with the Sentencing Commission’s policy statement at section 1B1.13 providing guidance on what those reasons encompass. The timing of the change proved decisive: the public health emergency that began in March 2020, about fifteen months after enactment, made medical vulnerability a compelling reason for release in thousands of cases, and the courts, not the bureau, became the forum in which those cases were decided. The provision that the sponsors had described as a modest transparency improvement became the principal legal mechanism for early release during the emergency.

What do the remaining titles carry?

Title I’s remaining sections build the infrastructure around the credit system: the Government Accountability Office is directed to audit the risk assessment system’s use, appropriations are authorized for fiscal years 2019 through 2023 with 80 percent reserved for bureau implementation, an Independent Review Committee is established to validate the system and review its effectiveness, and section 105 provides that the earned time provisions do not authorize prerelease custody or supervised release for anyone serving a state sentence, a rule of construction that states the federal-only scope in the text itself.

Title II, the Lieutenant Osvaldo Albarati Correctional Officer Self-Protection Act of 2018, requires the Bureau of Prisons to allow correctional officers to securely store personal firearms on bureau premises outside the security perimeter. Title III prohibits the use of restraints on federal prisoners during pregnancy, labor, and postpartum recovery, with exceptions for immediate flight or safety risks. Title V reauthorizes the Second Chance Act of 2007 through fiscal year 2023 and improves its programs, including reentry courts, career training, and the elderly offender pilot.

Title VI’s provisions include section 601, which directs the bureau to place prisoners as close as practicable to their primary residence and within 500 driving miles where practicable; section 602, which expands home confinement for low-risk prisoners; section 604, which requires the bureau to help prisoners obtain identification documents, including a Social Security card, driver’s license or other photo identification, and a birth certificate, before release; section 605, which expands the markets in which Federal Prison Industries may sell its products; section 606 on de-escalation training; section 607 on evidence-based treatment for opioid and heroin abuse, including medication-assisted treatment; section 610 on data collection for the National Prisoner Statistics program; section 611 requiring feminine hygiene products to be provided without charge; and section 613 restricting the use of solitary confinement for juveniles in federal custody.

What did the Dorsey decision leave for section 404?

The retroactivity provision did not arrive without precedent. In Dorsey v. United States, decided in 2012, the Supreme Court held that the Fair Sentencing Act’s reduced penalties applied to defendants who committed their offenses before August 3, 2010 but were sentenced after that date. The decision resolved the pipeline cases, the defendants whose offenses straddled the effective date, but it left untouched the much larger population of defendants who had been both sentenced and imprisoned under the old ratio. Section 404 is the legislative answer to the question Dorsey left open: whether the old ratio should continue to determine the sentences of people already serving them.

The drafting of the covered-offense definition reflects the Dorsey experience. Congress defined the eligible class as defendants convicted of an offense the statutory penalties for which were modified by the 2010 act, which captures the crack offenses without requiring courts to reconstruct what sentence each defendant would have received under the new ratio. The provision’s exclusions, for defendants already resentenced under the 2010 act and for defendants whose prior motions were denied on the merits, prevent relitigation of cases the courts have already decided. The result is a provision that is narrower than full retroactivity but broader than the pipeline rule in Dorsey, and the several thousand grants it produced are the measure of the gap between those two rules.

What counts as a serious drug felony?

Section 401’s narrowing of the triggering priors turns on two new defined terms, and the definitions are where the provision’s real work happens. A “serious drug felony” is an offense described in 21 U.S.C. section 841, 848, or 960, or a comparable state offense, for which the offender served more than 12 months of imprisonment and for which the maximum authorized term is at least 10 years. A “serious violent felony” is defined by reference to 18 U.S.C. section 3559(c)(2), the federal three-strikes definition, which lists the qualifying violent offenses. The 12-months-served requirement is the provision’s most important filter: it excludes the low-level state drug convictions, often resulting in probation or short jail terms, that had been triggering the old enhancements.

The effect is easiest to see through the cases the old rule produced. Under the pre-act reading of “felony drug offense,” courts had applied the 20-year and life enhancements on the basis of decades-old state convictions for small-quantity possession, including convictions that the states themselves treated as minor. The new definitions require both seriousness of the offense and seriousness of the punishment actually served, which means the enhancement tracks the trafficking offenses Congress originally meant to reach. The Sentencing Commission’s data on the provision’s early application showed the expected pattern: fewer defendants qualifying for the enhancements, and the enhancements that remained applying to defendants with genuinely serious criminal histories.

Why did Deal v. United States matter so much?

The stacking practice that section 403 ended traces to a single Supreme Court decision. In Deal v. United States, decided in 1993, the Court held that a “second or subsequent conviction” under 18 U.S.C. section 924(c) does not require the first conviction to have become final before the second offense was committed. The case involved a defendant convicted of six armed robberies in a single proceeding, with six 924(c) counts charged together; the Court’s reading produced a sentence of 105 years, with 100 of those years coming from the stacked firearm counts. The decision resolved a circuit split in favor of the government’s broader reading, and prosecutors used the stacking threat as leverage in plea negotiations for the next quarter century.

The sentences stacking produced became some of the most criticized in the federal system. Reported cases included a 55-year sentence for a first-time marijuana trafficker whose offense involved a firearm, and multi-decade sentences for defendants whose criminal history consisted entirely of the counts in the single indictment before the court. The Judicial Conference, the Sentencing Commission, and the American Bar Association all recommended ending the practice, and the trial judges who imposed the sentences frequently stated on the record that the sentences were unjust but required by law. Section 403’s clarification, that the enhanced consecutive term applies only where a prior 924(c) conviction is final, restored the provision to the meaning its drafters had intended: a recidivist enhancement for repeat offenders, not a multiplier for first-time defendants facing multiple counts.

How are the credits actually applied?

The earned time credit system has a set of application rules that determine how theoretical credits become actual early release. Credits are applied toward prerelease custody under 18 U.S.C. section 3624(c), which means transfer to a residential reentry center or to home confinement, or toward early supervised release, capped at 12 months. The bureau’s implementation rules, developed through notice and rulemaking, set the mechanics: credits accrue as programming is completed, the bureau periodically calculates each prisoner’s balance, and the credits reduce the date on which the prisoner becomes eligible for prerelease placement.

Several limitations shape the system’s operation. Credits cannot be earned for programming completed before the act’s enactment, which means the large population of prisoners who had completed programs in prior years started from zero. Credits cannot be earned for programming completed while detained pending adjudication, which excludes the pretrial period even for prisoners who participated in programs at detention facilities. And the credits of prisoners who are later found to have committed disciplinary violations can be reduced, which gives the bureau a mechanism for policing the incentive structure. The practical result is a system that rewards sustained participation over the course of a sentence rather than a one-time burst of program completion.

What did Title III change for pregnant prisoners?

Title III prohibits the use of restraints on federal prisoners during pregnancy, labor, and postpartum recovery, with narrow exceptions for immediate and credible flight risks or threats of harm. The provision requires the bureau to report on its compliance and establishes a right of action for violations. It was among the least debated provisions of the act and one of the most straightforward in its operation: the practice it prohibits had been condemned by medical organizations and correctional standards bodies for years, and the statute simply imposed a uniform federal rule where bureau policy had been uneven.

The provision’s significance lies less in its direct effect on the population, which is small, than in what it reveals about the statute’s method. Title III is a conditions-of-confinement reform, not a sentencing or release reform, and it sits in the act because the vehicle was moving and the provision’s supporters could attach it. The same is true of the hygiene products provision and the juvenile solitary restriction. The act is in part a collection of long-sought correctional reforms that had waited years for a legislative vehicle, and their presence in the statute is a reminder that omnibus legislation aggregates as much as it designs.

What did the placement provision require?

Section 601 directed the Bureau of Prisons to place prisoners in facilities as close as practicable to their primary residences, and to transfer prisoners to facilities closer to home when practicable. The provision addressed a longstanding complaint of prisoners’ families: the bureau’s placement decisions often sent prisoners to facilities hundreds of miles from home, which made visitation difficult and weakened the family ties that the reentry literature identifies as protective against recidivism. The statute did not create an enforceable right to a particular placement, but it imposed a statutory preference that the bureau must consider and document.

The placement preference interacts with the programming provisions. A prisoner transferred closer to home may lose access to a specialized program available only at the original facility, and the bureau must balance the placement preference against the programming needs identified by the risk and needs assessment. The statute does not resolve the tradeoff, and the bureau’s placement decisions continue to reflect operational constraints, including bed space, security levels, and program availability. The provision’s effect is therefore modest but real: it moved family proximity from an administrative afterthought to a statutory factor, and the bureau’s placement data became subject to the same oversight as the rest of the act’s implementation.

What happened to the elderly offender pilot?

Section 604 reauthorized the elderly offender early release pilot program, which had been created by the Second Chance Act of 2007 and extended by subsequent appropriations. The pilot allows the Bureau of Prisons to place elderly prisoners on home detention before the end of their sentences, subject to criteria including age, time served, and the absence of a history of violence. The original pilot had been limited in scope and in the number of prisoners it reached, and the reauthorization expanded its availability while preserving the bureau’s discretion over placement decisions.

The pilot’s history illustrates the act’s incremental method. The 2007 act created the program as an experiment, the appropriations process kept it alive, and the 2018 act made it a permanent feature of the bureau’s authorities. The population it serves is small: elderly prisoners who meet the criteria are a fraction of the federal population, and the bureau’s use of the authority has been cautious. But the pilot’s existence matters for the prisoners it does reach, and its reauthorization signaled that Congress viewed early release for the elderly as a settled policy rather than an experiment. The provision is one of several in the act whose significance is best measured one prisoner at a time.

What did the prison industries provision change?

Section 605 expanded the markets in which Federal Prison Industries, the government corporation that employs federal prisoners in manufacturing and services, may sell its products. Before the act, the corporation’s sales were largely restricted to the federal government. The amendment added the government of the District of Columbia and nonprofit organizations to the list of authorized customers, which allowed the corporation to expand its operations and, in principle, to employ more prisoners in its programs.

The provision’s significance is economic as well as penological. Prison labor programs are credited with reducing idleness and teaching work skills, but they have also been criticized for paying wages far below market rates and for competing with private businesses. The expansion of the customer base addressed the competition concern by directing the new sales toward government and nonprofit buyers rather than the private market, while the employment effect supported the statute’s broader reentry theory: prisoners who work during their sentences are better positioned to find work after release. The provision is one of the act’s quieter economic interventions, and its effects are measured in the corporation’s sales and employment figures rather than in the sentencing data.

What do the remaining titles carry in detail?

Title V reauthorizes the Second Chance Act of 2007 through fiscal year 2023 and improves its programs. The reauthorization covers the reentry grant programs for state and local governments and nonprofit organizations, the career training programs, and the audit and accountability provisions for grantees. Section 504 improves federal reentry planning, section 505 coordinates interagency reentry efforts, and section 507 requires evaluation of the Second Chance programs, building an evidence base for what works. The title is less dramatic than the sentencing and credit provisions, but it funds the infrastructure on which the recidivism reduction strategy depends.

Title VI’s miscellaneous provisions include several with significant practical effects. Section 605 expands the markets in which Federal Prison Industries, the federal prison labor program, may sell its products, adding the District of Columbia government and nonprofit organizations to the authorized customers. Section 607 requires the bureau to report on its capacity to provide evidence-based treatment for opioid and heroin abuse, including medication-assisted treatment, and requires the Administrative Office of the United States Courts to report on treatment capacity for prisoners on supervised release. Section 610 expands the data collection requirements of the National Prisoner Statistics program, which improves the statistical foundation for future policy. Section 611 requires the bureau to provide tampons and sanitary napkins without charge. Section 613 restricts the use of solitary confinement for juveniles in federal custody. Each of these provisions is narrow, but together they represent the most extensive congressional direction of bureau operations since the 1980s.

Who the statute reaches and who it does not

The act reaches federal prisoners only. Section 105 of the act states the point directly, and the entire credit and programming structure is built on the Bureau of Prisons population. The federal system held roughly 180,000 people at enactment, out of a total incarcerated population in the United States of more than two million. The federal share sits below ten percent of the total, which means that every release figure the act produces must be read against that structural ceiling. A reform that reduces federal sentences substantially still leaves the state systems, which hold the overwhelming majority of prisoners, untouched.

Within the federal population, the exclusions sort prisoners into those who can earn time credits and those who cannot. The exclusion list at 18 U.S.C. section 3632(d)(4)(D) is organized around categories of offense rather than risk scores: terrorism, espionage, murder, assault, kidnapping, robbery, weapons offenses, sex offenses, human trafficking, and certain drug offenses all disqualify. Prisoners with immigration detainers, meaning a final order of removal, are excluded as well. The practical shape of the exclusion list is that the credit system rewards programming among the nonviolent drug and white-collar offenders who already made up the bulk of minimum and low security federal custody, while the violent offenders whom the public most associates with long federal sentences remain outside the system entirely.

The sentencing provisions sort defendants by time rather than by offense. Sections 401 through 403 apply only to sentences imposed after enactment; a defendant sentenced the day before the signing received no benefit from them. Only section 404 reaches backward, and it reaches only crack cocaine defendants sentenced before August 3, 2010. The compassionate release change applies to motions filed after enactment regardless of when the sentence was imposed, which is why it could affect prisoners whose sentences long predated the act. The good conduct time correction applied immediately to the entire population serving sentences of more than a year.

How does the exclusion list work in practice?

The exclusion list at 18 U.S.C. section 3632(d)(4)(D) operates as a conviction-based filter applied at the threshold: if the prisoner’s current sentence includes any listed offense, the prisoner cannot earn time credits at all, regardless of risk score or program participation. The filter is categorical rather than individualized, which means a prisoner classified as minimum risk by PATTERN is still excluded if the offense of conviction appears on the list. The Cruz amendment, agreed to by voice vote during the Senate’s consideration, added offenses to the list, reflecting the political reality that the exclusion catalogue was the price of the broader bill.

The immigration detainer exclusion operates alongside the offense list and has been one of the provision’s most contested features. Prisoners subject to a final order of removal are ineligible for earned time credits, which means noncitizen prisoners, who make up a significant share of the federal population because of the federal government’s immigration enforcement role, are categorically excluded from the credit system even when their offenses are nonviolent. Advocates have argued that the exclusion creates a two-tier system in which otherwise identical prisoners receive different treatment based on immigration status. The statute’s defenders respond that the exclusion reflects the reality that such prisoners will be transferred to immigration custody rather than released to the community, so prerelease credits serve no reentry purpose. The dispute is unresolved and the exclusion remains in the text.

Why does prospectivity matter so much?

The choice to make sections 401 through 403 prospective only is the single most important limit on the sentencing title, and it deserves emphasis because it is easy to miss. A defendant sentenced on December 20, 2018, the day before the signing, serves under the old enhancements. A defendant sentenced on December 22 serves under the new ones. The difference can be decades. The drafters made this choice deliberately: retroactivity would have required resentencing thousands of prisoners, would have provoked far stronger law enforcement opposition, and would have complicated the bill’s passage beyond what the coalition could bear. Section 404’s retroactivity was the exception the coalition could agree on, because the crack ratio’s injustice was the most widely acknowledged and because the population it reached was defined and finite.

The prospectivity line has produced the statute’s sharpest equities problem. Prisoners serving stacked 924(c) sentences imposed before enactment, some of them the very cases that motivated the reform, received no relief from the provision written to end the practice. Clemency became the only avenue for such prisoners, and the executive’s use of commutation for stacked-sentence cases has been episodic rather than systematic. The statute’s defenders describe this as the cost of legislative compromise; its critics describe it as the compromise swallowing the reform. Both descriptions are accurate, and the tension between them is the central moral fact of the sentencing title.

How the bureau put the machinery into practice

The Department of Justice released PATTERN in July 2019, within the 210-day statutory deadline for the system’s development. The tool produces two scores: a general recidivism risk score and a violent recidivism risk score, each placing the prisoner in one of four categories, and it feeds a needs assessment that matches prisoners to programming. The Independent Review Committee, established under section 107, was charged with reviewing and validating the system, and its reviews, along with outside analyses by research organizations, produced the central implementation controversy of the act’s early years: whether the tool’s risk categories were racially neutral in their construction and whether the needs assessment was directing programming to the prisoners who would benefit most.

The critiques followed a familiar pattern from the literature on actuarial risk assessment. Researchers pointed to differences in risk classification rates across racial groups and argued that the tool’s reliance on criminal history variables reproduced disparities in policing and prosecution. The department responded with a revalidation and recalibration that adjusted the tool’s scoring, and the Independent Review Committee’s subsequent reviews found the tool’s predictive validity acceptable while recommending continued monitoring. The dispute is a genuine methodological disagreement rather than a settled question, and the statute’s own design anticipates it: section 103 requires the Government Accountability Office to audit the system’s use, and the audit mandate has produced periodic public reporting on how the bureau applies the tool.

The bureau’s implementation of earned time credits moved more slowly than the statute’s supporters had hoped. The credits required the bureau to build a tracking system for program participation and to determine how credits would be applied to prerelease custody and supervised release, and the first large-scale application of credits followed a bureau rulemaking that set the operational rules. The bureau’s public reporting has documented tens of thousands of prisoners earning credits and thousands moving to prerelease custody or early supervised release as a result, though the precise figures have moved with each reporting cycle and with subsequent adjustments to the credit calculation rules.

The home confinement expansion under section 602, which directed the bureau to place low-risk prisoners on home confinement for the maximum time permitted, gained prominence during the public health emergency, when the Attorney General issued memoranda directing the bureau to prioritize home confinement transfers for vulnerable prisoners. The bureau’s use of home confinement authority expanded sharply during that period, and the question of whether prisoners transferred under emergency authorities would be returned to custody afterward became one of the act’s most contested implementation episodes, resolved ultimately by a Justice Department legal opinion and subsequent legislation that allowed the transfers to continue.

Is the risk assessment tool fair?

The question has no consensus answer, and the honest treatment presents both sides. The department’s position is that PATTERN is an actuarial instrument validated on federal data, that its risk categories predict recidivism at rates consistent with accepted standards for such tools, and that revalidation has addressed the measurable biases researchers identified. The critics’ position, advanced by organizations including the Leadership Conference on Civil and Human Rights and by academic researchers, is that any tool trained on criminal history data will carry forward the disparities in the underlying system, and that a tool used to allocate liberty should be held to a higher standard than one used to allocate programming. The statute does not resolve the dispute; it creates the audit and review machinery through which the dispute is conducted.

What does PATTERN actually measure?

PATTERN scores each prisoner on general and violent recidivism using criminal history, age, institutional behavior, education, and program participation, then assigns one of four risk categories. The classification drives programming priority and credit eligibility. The department revalidated the tool after finding racial classification disparities, and independent review continues.

PATTERN draws on variables that the criminological literature has associated with recidivism: criminal history, age at first offense and at assessment, history of violence, escapes and institutional infractions, education level, and participation in programming. The tool produces separate scores for general recidivism and violent recidivism, and each score maps to one of four categories: minimum, low, medium, or high risk. The category assignment drives two things: priority for programming placement and, for minimum and low risk prisoners, eligibility for the additional 5 days of earned time credits per 30 days of programming.

The department’s revalidation, conducted after the initial release, adjusted the tool’s scoring in response to the finding that classification rates differed across racial groups. The Independent Review Committee’s subsequent assessments found the tool’s predictive accuracy within accepted ranges for actuarial instruments while continuing to recommend monitoring for disparate impact. The honest summary is that PATTERN is a typical second-generation risk assessment tool: it predicts better than chance and worse than certainty, it carries the biases of its training data, and the statute’s audit requirements mean its performance is more transparent than that of the proprietary tools many states use. The critics’ objection is not that the tool fails by the standards of its genre but that the genre’s standards are too low for decisions about liberty.

How did the bureau build the credit tracking system?

The earned time credit system required the Bureau of Prisons to build something it had never had: a prisoner-level accounting system that tracks program participation day by day and converts it into credits. The bureau developed the system through internal policy and then through formal rulemaking, with the key operational rule setting out how credits accrue, how they are calculated, and how they are applied to prerelease custody and supervised release. The rulemaking drew extensive public comment, much of it from prisoners’ advocates arguing that the bureau’s proposed rules were narrower than the statute allowed, particularly on the question of how credits apply to prisoners with immigration detainers and on the calculation of the 12-month supervised release cap.

The bureau’s public reporting on the system has documented the scale of its operation: tens of thousands of prisoners earning credits, thousands applying them toward prerelease placement or early supervised release, and a steady increase in the program participation the credits were designed to incentivize. The figures move with each reporting cycle as the bureau refines its calculations, and advocates have disputed particular bureau interpretations, but the direction is consistent with the statute’s design. The credit system became the largest structured incentive program in the history of the federal prison system.

What happened with home confinement?

Section 602 of the act directed the bureau to place low-risk prisoners on home confinement for the maximum time permitted under 18 U.S.C. section 3624(c)(2), which authorizes up to six months or ten percent of the sentence, whichever is less. The provision attracted modest attention at enactment and then became one of the act’s most contested implementation episodes. In March and April 2020, the Attorney General issued memoranda directing the bureau to prioritize home confinement transfers for medically vulnerable prisoners as the public health emergency spread through federal facilities, and the CARES Act, enacted that March, expanded the bureau’s home confinement authority by removing the statutory time limits during the emergency period.

The bureau transferred thousands of prisoners to home confinement under these authorities. The controversy arose when the emergency period approached its end: a Justice Department Office of Legal Counsel opinion concluded that prisoners transferred under the CARES Act authority would have to return to custody when the emergency ended, a reading that would have reincarcerated thousands of people who had been living at home, working, and complying with supervision. The opinion produced an intense legislative and advocacy response, and Congress ultimately enacted a provision allowing the transfers to continue, resolving the question in favor of the prisoners. The episode illustrates the act’s implementation dynamic: broad statutory language, aggressive executive use during an emergency, and a legislative fix when the executive’s reading threatened the statute’s purpose.

What does the name mean?

The acronym FIRST STEP stands for Formerly Incarcerated Reenter Society Transformed Safely Transitioning Every Person, a backronym constructed to carry the bill’s reentry theme. The name was attached to the House prison reform bill, H.R. 5682, and it survived the merger with the sentencing provisions even though the combined statute is as much about sentencing as about reentry. The enrolled bill also carries two short titles for portions of the act: the Lieutenant Osvaldo Albarati Correctional Officer Self-Protection Act of 2018 for Title II, named for a federal correctional officer killed in the line of duty, and the Second Chance Reauthorization Act of 2018 for Title V. The presence of named short titles inside the larger act reflects the legislative reality that the statute is a vehicle: provisions that could not have passed on their own rode inside a package that could.

Who watches the machinery?

The statute builds oversight into its own implementation. Section 101 establishes an Independent Review Committee, composed of experts in risk assessment and recidivism reduction, to advise the Attorney General on the development and use of the risk and needs assessment system. The committee’s reports are public, and its assessments of PATTERN’s accuracy and fairness have been the most authoritative independent evaluations of the tool. The statute also requires the Attorney General to report annually to Congress on the system’s operation, and the Government Accountability Office has conducted audits of the bureau’s implementation of the programming and credit provisions.

The oversight machinery matters because the statute delegates so much. The risk assessment tool, the programming catalogue, the credit calculations, and the prerelease placement decisions are all bureau operations, and without the reporting and audit requirements, the statute’s promise would depend entirely on the bureau’s good faith. The committee and the audits do not eliminate the bureau’s discretion, but they make its exercise visible, and the visibility has shaped the bureau’s behavior: the public reporting on credit calculations and program availability has been a check on the restrictive readings the bureau might otherwise have adopted.

What training and pilot provisions did the act carry?

Title VI includes several provisions aimed at the bureau’s own operations. Section 606 requires the bureau to provide de-escalation training for its correctional officers and employees, including training in the use of force alternatives and in interacting with prisoners with mental illness. Section 608 establishes pilot programs for youth mentorship and for the use of service animals in prison programming. Section 609 requires the bureau to report on the availability of medication-assisted treatment for opioid use disorder. These provisions are small individually, but together they represent the most extensive congressional direction of bureau training and operations since the 1980s, and they reflect the reform coalition’s view that the conditions of confinement are a legitimate subject of federal legislation.

How small is the federal share?

The scope limitation deserves a final emphasis, because it is the fact that most often goes missing from public discussion of the act. The federal prison system held roughly 180,000 people at enactment, out of a total incarcerated population of more than two million. The state systems, which hold the overwhelming majority, are governed by state sentencing laws that the act does not touch. The act’s sentencing provisions affect only federal prosecutions, its credit system only federal prisoners, and its compassionate release changes only federal courts.

The federal share is also distinctive in composition. Federal prisoners are disproportionately drug offenders, because the federal government’s criminal jurisdiction is concentrated in drug trafficking, immigration, firearms, and fraud, while the state systems hold the violent offenders who make up the majority of the national incarcerated population. The act’s design reflects this composition: its sentencing provisions target drug and firearm offenses, its exclusion list removes violent and serious offenses from the credit system, and its overall effect is concentrated in the drug docket. A reader who understands the federal share understands both the act’s real achievements and its structural limits.

The 115th Congress that passed the act was divided government: a Republican Senate, a Republican House for the first session and a Democratic House after the 2018 midterms, and a Republican President. The criminal justice package moved through this divided terrain because its coalition crossed the partisan divide in both directions. The lame duck session, the period between the November election and the January seating of the new Congress, provided the window: outgoing members were free from electoral pressure, the leadership wanted a bipartisan achievement to close the Congress, and the White House wanted a signing ceremony. The timing was not incidental to the outcome. The same bill, introduced in the same form in a non-lame-duck session, might never have reached the floor.

What did the victims’ provisions do?

The Crime Victims’ Rights Act of 2004 gives victims of federal crimes the right to reasonable notice of court proceedings and of the defendant’s release or escape. The First Step Act’s release mechanisms raised the question of how those rights apply to prerelease transfers and credit-driven release dates, which are bureau administrative decisions rather than court proceedings. The Cotton-Kennedy amendments would have imposed new victim notification requirements tied to release dates; their defeat left the existing framework in place, with the bureau’s victim notification program continuing to operate under the 2004 act’s requirements.

The debate over the amendments illustrated the statute’s underlying tension. The reform coalition argued that the notification provisions duplicated existing law and would have created administrative burdens that slowed the release mechanisms. The amendments’ supporters argued that victims deserved specific notice of credit-driven releases. The Senate’s 62-to-37 rejection of the amendments was a substantive decision about the statute’s priorities: the bill’s release mechanisms would operate through the existing victim-notification framework rather than through new bespoke requirements. The Cruz amendment that was agreed to by voice vote added offenses to the exclusion list without touching the notification question.

How did the Second Chance grants work?

The Second Chance Act’s grant programs, reauthorized by Title V, fund state and local reentry programs: employment training, substance abuse treatment, mentoring, and housing assistance for people leaving prison. The grants are administered by the Department of Justice’s Office of Justice Programs, and the reauthorization through fiscal year 2023 provided continuity for programs that had been operating since 2009. The evaluation requirement in section 507 built an evidence base for which grant-funded interventions reduce recidivism, and the bureau’s own programming expansion under Title I drew on the same research literature.

The grants’ significance is easy to overlook because they operate outside the federal prison system, in the state and local systems that hold the vast majority of prisoners. The reentry grants are the one part of the act that reaches beyond the federal share, and they do so through the spending power rather than through direct regulation. The amounts are modest relative to the scale of the state systems, but the grants have funded some of the most evaluated reentry programs in the country, and the evaluation requirement means the funded programs contribute to the knowledge base on which future reforms will draw.

What came after?

The act’s passage did not end the legislative debate over federal sentencing. Later congresses considered proposals to expand the act’s reach: bills to make the sentencing provisions retroactive, to narrow the exclusion list, to expand the earned time credit system, and to address the 85 percent truth-in-sentencing rule. None of these proposals has been enacted, and the act’s boundaries have remained where the 115th Congress drew them. The Department of Justice’s implementation choices have been the principal source of change since enactment: the bureau’s rulemaking on credit calculations, the department’s revalidation of PATTERN, and the executive’s use of the home confinement and compassionate release authorities have all shaped the statute’s practical effect more than any subsequent legislation.

The courts have also continued to construe the statute. The Supreme Court’s decisions on the act’s provisions have resolved the major interpretive disputes, and the lower courts’ compassionate release case law has matured into a stable body of precedent. The statute entered the phase of its life that most major legislation reaches: the text is fixed, the implementing institutions have built their systems around it, and the remaining questions are operational rather than interpretive. The act’s legacy will be measured in the data the oversight machinery continues to produce.

What is the 85 percent rule and why did the act leave it alone?

The Sentencing Reform Act of 1984 requires federal prisoners to serve at least 85 percent of their imposed sentences, with good conduct time capped so that the maximum reduction is 15 percent. The rule is the federal system’s substitute for parole: where the old system released prisoners to parole supervision after a portion of the sentence, the current system requires service of nearly the entire term, with good conduct time as the only general reduction. The First Step Act’s correction of the good time calculation operates within this rule rather than changing it. The 54 days per year of the sentence imposed is the mechanism by which the 85 percent floor is implemented, and the act’s correction restored the calculation Congress intended without touching the floor itself.

Proposals to reduce the 85 percent requirement, or to restore parole in the federal system, have been introduced in later congresses but have not been enacted. The 85 percent rule is the structural reason the federal system cannot replicate the release rates of state systems with parole: even the earned time credits, which can produce up to 12 months of early supervised release, operate as a supplement to good conduct time rather than as a replacement for the truth-in-sentencing framework. The act’s designers understood this constraint and worked within it. The statute’s theory is that incentives and programming can change outcomes even inside a truth-in-sentencing system, and the credit data will eventually show whether that theory holds.

How did the act affect the District of Columbia?

The federal prison system includes offenders sentenced under the District of Columbia Code, who are housed in federal facilities and supervised by federal agencies under the National Capital Revitalization and Self-Government Improvement Act of 1997. The First Step Act’s provisions apply to these offenders on the same terms as to other federal prisoners, because their sentences are served in the Bureau of Prisons. The DC population is a reminder that the federal system is not only the system for federal crimes: it is also, by historical accident, the prison system for the nation’s capital. The act’s reach therefore includes a population whose offenses were prosecuted under local law but whose sentences are administered federally, a quirk that the statute’s drafters did not need to address because the existing jurisdictional framework already placed these prisoners inside the bureau’s custody.

The courts and the statute

The litigation over the act has concentrated in three areas. The first is the scope of section 404’s retroactivity. The statute allows resentencing for defendants convicted of a “covered offense,” defined as a violation of a federal criminal statute whose statutory penalties were modified by the Fair Sentencing Act. Lower courts divided over whether defendants convicted of crack offenses that also involved other drugs, or whose sentences were driven by career offender guidelines rather than the crack penalties, qualified. The courts of appeals produced a body of case law working through these questions offense by offense, and the general direction of the holdings was that eligibility turns on the statute of conviction rather than on the guidelines calculation that produced the sentence, a reading that expanded the provision’s reach beyond the narrowest interpretation while still leaving some categories of defendants outside it.

The second area is the prospectivity of sections 401 through 403. Defendants sentenced before enactment argued that the new, more lenient provisions should apply to their cases, and the courts uniformly held that the provisions apply only to sentences imposed after the date of enactment. The statute’s text states the effective date plainly, and no court has found a basis for retroactive application of the stacking fix or the enhancement reductions. The contrast with section 404, the one provision Congress made retroactive, has been cited by courts as evidence that the prospectivity of the other provisions was deliberate rather than inadvertent.

The third area is compassionate release, and it is by far the largest. The courts received tens of thousands of motions under the amended section 3582(c)(1)(A), first in a steady stream after enactment and then in a flood during the public health emergency. The litigation addressed the exhaustion requirement, with courts dividing over whether the 30-day provision could be waived in extraordinary circumstances; the scope of “extraordinary and compelling reasons,” with courts developing standards for medical vulnerability, age, and family circumstances; and the relationship between the statutory standard and the Sentencing Commission’s policy statement, which had not been updated to reflect the new direct-filing procedure and which the courts treated as guidance rather than binding constraint in the interim. The compassionate release docket became one of the largest bodies of federal criminal motion practice in the country’s history, and its decisions, numbering in the thousands, form the most extensive judicial gloss on any provision of the act.

The decisions construing the act’s relationship to the broader sentencing framework are treated in the series’ article on sentencing law in the Supreme Court, which carries the constitutional and guidelines cases that shape how these provisions operate. This article’s account is confined to the act’s own provisions and the cases that have construed them directly.

Did the courts expand or narrow the act?

The net effect of the litigation has been expansionary at the margins but bounded by the statute’s text. The section 404 case law widened eligibility beyond the narrowest reading of “covered offense.” The compassionate release case law built a substantial body of judge-made standards for a provision Congress had left largely undefined. But the prospectivity holdings kept the sentencing provisions within their enacted bounds, and no court has rewritten the exclusion list or the credit calculation. The statute that emerged from the courts is recognizably the statute Congress passed, with the rough edges of its drafting smoothed rather than reshaped.

How did courts handle the exhaustion requirement?

The compassionate release statute requires the prisoner to exhaust administrative remedies or to wait 30 days after presenting a request to the warden before filing with the court. During the public health emergency, prisoners argued that the exhaustion requirement should be excused where the bureau was not responding to requests or where delay would defeat the purpose of the motion. The courts divided. Some held the requirement jurisdictional and unwaivable, dismissing motions filed before the 30 days ran. Others held it a claim-processing rule subject to equitable exceptions, reaching the merits where the prisoner showed that waiting would cause irreparable harm.

The division reflected a genuine ambiguity in the statutory text, which states the requirement without stating its consequences. The practical effect was a period of uneven access: prisoners in some districts could obtain emergency consideration while prisoners in others had to wait out the clock. Over time, the bureau improved its response times and the courts converged on treating the 30-day provision as the operative rule, with the understanding that the warden’s failure to act within 30 days satisfies the requirement regardless of whether administrative appeals have been completed. The exhaustion litigation has largely settled, but it determined who could get into court during the months when timing mattered most.

What counts as extraordinary and compelling?

Courts recognize terminal illness, serious medical conditions preventing self-care in custody, age-related deterioration, the loss of a child’s only caregiver, and, during the emergency period, medical vulnerability combined with unprotectable conditions of confinement. Rehabilitation alone is generally insufficient. The Sentencing Commission has since updated its policy statement to reflect the case law.

The statute uses the phrase “extraordinary and compelling reasons” without defining it, and the Sentencing Commission’s policy statement at section 1B1.13, which lists qualifying circumstances, had not been updated to reflect the new direct-filing procedure. The courts therefore built the standard case by case, and the resulting body of law is the most detailed judicial treatment of compassionate release in American history. The categories that emerged include terminal illness and serious medical conditions that substantially diminish the ability to provide self-care in custody; age-related deterioration, particularly for prisoners over 65 with serious health conditions; family circumstances, such as the death or incapacitation of the only available caregiver for a prisoner’s minor child; and, during the emergency period, medical vulnerability to the virus combined with conditions of confinement that prevented effective protection.

The courts also addressed what does not qualify. Rehabilitation alone, however impressive, was generally held insufficient without an additional extraordinary circumstance. Disparities between the sentence imposed and the sentence that would be imposed under current law were treated as a factor rather than as an independent basis in most districts, though some courts gave such disparities greater weight. The Commission eventually updated its policy statement to reflect the direct-filing procedure and to incorporate the categories the courts had developed, which brought the guidance into alignment with the case law and reduced the variation across districts.

How did the covered-offense disputes resolve?

The section 404 litigation turned on the meaning of “covered offense,” defined as a violation of a federal criminal statute the statutory penalties for which were modified by the Fair Sentencing Act of 2010. The disputes fell into three categories. First, defendants convicted of crack offenses that also involved other controlled substances argued that their offenses were covered because the crack penalties were among those modified; the courts generally agreed where the statute of conviction was one the 2010 act had changed. Second, defendants whose sentences were driven by the career offender guideline rather than by the crack penalties argued that they were nevertheless convicted of covered offenses; the courts generally held that eligibility turns on the statute of conviction rather than on the guidelines calculation, which expanded eligibility beyond the narrowest reading. Third, defendants convicted of attempt or conspiracy argued that their offenses qualified; the courts divided, with the better-reasoned decisions holding that the inchoate offenses fall within the definition where the underlying substantive offense does.

The net effect of the covered-offense case law was to widen section 404’s reach beyond the interpretation the government initially advanced, while keeping it within the bounds of the statutory definition. The litigation has largely run its course, and the provision’s work is substantially done: the population of prisoners sentenced under the old ratio has been reduced through the grants the courts approved.

What changed in practice

The measured effects of the act are unusually well documented for a criminal justice statute, because the Bureau of Prisons, the Sentencing Commission, and the Government Accountability Office all report on the relevant populations. The section 404 retroactivity provision produced several thousand sentence reductions, with the Sentencing Commission’s retroactivity reports documenting the grants offense by offense and the average reduction in months. The good conduct time correction applied to the entire federal population serving sentences of more than a year, accelerating release dates across the board and producing the act’s broadest immediate effect on time served.

The compassionate release provision produced the largest number of individual releases of any single provision. The courts granted thousands of motions during the public health emergency, and the provision’s transformation from a rarely used administrative mechanism into a routine judicial remedy is the act’s most significant structural change to the federal release system. The bureau’s own compassionate release grants, which had numbered in the dozens per year before the act, were supplemented by thousands of court-ordered releases, and the procedural right to file directly became a permanent feature of federal sentencing law by the amendment.

The earned time credit system has produced a smaller but growing number of early transitions to prerelease custody and supervised release, with the bureau reporting tens of thousands of prisoners earning credits and thousands applying them. The recidivism effects of the programming itself are harder to measure: the bureau’s programs are being evaluated on the schedules the statute requires, and the honest assessment is that the evidence on whether the programming reduces reoffending is still developing. The statute’s theory, that structured programming plus incentives reduces recidivism, is plausible and supported by the general literature on correctional programming, but the act-specific evidence remains incomplete.

The federal prison population continued the decline that had begun in 2013, falling further in the years after enactment, though the decline reflected multiple forces, including charging and sentencing trends that predated the act, and no serious analyst attributes the population change to the statute alone. The act’s contribution is best measured provision by provision: several thousand section 404 reductions, tens of thousands of good-time recalculations, thousands of compassionate release grants, and a growing number of credit-driven transitions, each documented by an official source for its own period.

The section 404 grants were concentrated in the districts with the largest historical crack caseloads, and the average reduction ran to several years, reflecting the scale of the old ratio’s distortion. The good conduct time correction, by contrast, was diffuse: it shortened every sentence of more than a year by a matter of weeks or months, and its aggregate effect on the population exceeded that of any other single provision in terms of total days of imprisonment avoided. The compassionate release grants were concentrated in the emergency period and among medically vulnerable prisoners, and their distribution reflected the courts’ development of the extraordinary-and-compelling standard in real time.

The recidivism evidence remains the least developed part of the record. The bureau’s programming evaluations proceed on the schedules the statute requires, and the Government Accountability Office’s audits have documented the system’s operation without yet producing a definitive assessment of its effect on reoffending. The general literature on correctional programming supports the statute’s theory, and the bureau’s internal data on program completion and misconduct show movement in the expected directions, but the act-specific causal evidence is still being built. The honest statement is that the statute created the conditions for measuring its own success and that the measurement is underway.

The Sentencing Commission’s implementation data provides the most systematic picture of the sentencing title’s reach. The commission’s reports on the act’s early years documented thousands of section 404 motions decided, with grant rates varying by district but averaging well above half, and average sentence reductions measured in years rather than months. The commission’s data on sections 401 and 402 showed the provisions’ prospective effect accumulating as new cases were sentenced: fewer defendants subject to the enhanced minimums, more defendants qualifying for the safety valve, and average drug sentences for the affected categories declining. The commission’s annual sourcebooks carry the act’s effects as a permanent feature of federal sentencing statistics, which means the statute’s measurement will continue as long as the guidelines data is collected.

The compassionate release data tells a different story, one of a provision whose use was concentrated in an emergency no one anticipated. The Administrative Office of the United States Courts reported thousands of motions filed in the emergency period, with grant rates that varied dramatically by district and by judge. The variation reflected the absence of an updated policy statement: with the commission’s guidance stale, each judge built the standard from the statutory text and the developing case law, and reasonable judges disagreed about what counted as extraordinary and compelling. The eventual update of the policy statement narrowed the variation, but the emergency-period grants remain a distinct episode in the data, a spike that will not recur unless another emergency produces the same conditions.

The earned time credit data is the most operationally complex. The bureau’s reporting tracks credits earned, credits applied, and prisoners transferred to prerelease custody or released to supervision, and the figures have grown as the bureau’s tracking system matured. The bureau’s data also reveals the exclusion list’s effect: a substantial share of the population is ineligible for credits, and the eligible population’s credit balances are concentrated among prisoners with longer sentences and sustained program participation. The credit system’s full effect on time served will not be visible until the cohorts that entered programming after enactment have completed their sentences, which means the definitive assessment of the system’s effect on release dates is still years away.

The honest account of what it did not do

The act did not change the federal mandatory minimums themselves for first-time offenders. The five-year and ten-year minimums for drug quantities remain exactly as they were. It did not alter the 85 percent truth-in-sentencing requirement that the Sentencing Reform Act of 1984 imposed on federal prisoners, meaning that even with good conduct time and earned time credits, federal prisoners serve the great majority of their stated sentences. It did not make any sentencing provision other than section 404 retroactive, so the defendants serving the stacked 924(c) sentences and the enhanced drug sentences imposed before December 2018 received no relief from the act’s sentencing title.

It did not reach the states. The federal system holds below ten percent of the incarcerated population, and the act’s ceiling was therefore structural: even a statute that released every eligible federal prisoner would leave the American incarceration rate among the highest in the world, because the rate is driven by state systems the statute does not touch. The reform organizations that supported the act described it as a first step, and the short title says the same thing, but the description is sometimes read as modesty when it is actually precision. The statute was designed to be the first step in federal reform, not a step in state reform, and its provisions reflect that boundary.

It did not resolve the risk assessment debate, the question of whether earned time credits reach the prisoners who need programming most, or the empirical question of whether the programming reduces recidivism. Those questions were assigned by the statute to the audit and review machinery, and they remain open. The act built a system for answering them rather than answering them itself, which is either the statute’s wisdom or its evasion depending on the observer’s view of what legislation can accomplish.

It did not change the structure of federal prosecution that produces the sentences in the first place. The mandatory minimums remain available to prosecutors as charging leverage, and the decision whether to charge quantities that trigger the minimums, whether to file the enhancement notice under 21 U.S.C. section 851, and whether to stack 924(c) counts within the new rules all remain with the executive. The act changed what happens after conviction; it did not change the incentives that shape charging. Reformers who hoped the statute would reduce prosecutorial leverage over plea bargaining have been disappointed, while prosecutors who feared the statute would disarm them have found their core tools intact. The stability of the charging structure is one reason the act’s effects, while real, have been smaller than either its supporters or its opponents predicted.

It did not address the Bureau of Prisons’ staffing and capacity constraints, which limit how much programming the bureau can actually deliver. The earned time credit system assumes that programming is available to every eligible prisoner who wants it, but the bureau has faced chronic staffing shortages, and program waitlists have been a persistent implementation problem. A credit system without available programming is a promise without a mechanism, and the bureau’s ability to expand programming has been the binding constraint on the system’s reach. The appropriations the act authorized for implementation were a start, but the bureau’s operational capacity remains the variable that determines whether the statute’s theory works in practice.

The provision and reach table

Provision	Who it affects	Retroactive	People reached where documented	Exclusion or limit
Section 401: reduced and restricted enhanced minimums	Federal drug defendants with prior convictions	No	Prospective only; affects new sentences	Narrowed to serious drug and serious violent felonies
Section 402: broadened safety valve	Low-level federal drug defendants	No	Prospective only; expanded eligibility	Criminal history ceiling of four points with refinements
Section 403: ended 924(c) stacking	Federal defendants with multiple firearm counts	No	Prospective only; largest effect on longest sentences	Requires prior conviction to be final before the new offense
Section 404: Fair Sentencing Act retroactivity	Crack defendants sentenced before August 3, 2010	Yes, the only retroactive provision	Several thousand sentence reductions documented	Court retains discretion; excludes already-reduced sentences
Earned time credits	Federal prisoners completing programming	No, and not for pre-enactment programming	Tens of thousands earning; thousands applying	Long offense exclusion list; immigration detainers excluded
Good conduct time correction	All federal prisoners serving over one year	Immediate recalculation	Entire population; roughly 70 extra days on a ten-year sentence	Applies only to sentences exceeding one year
Compassionate release direct filing	Federal prisoners with extraordinary and compelling reasons	Applies to motions filed after enactment	Thousands of court-ordered releases	Exhaustion or 30-day warden inaction required
Home confinement expansion	Low-risk federal prisoners	Ongoing bureau implementation	Expanded sharply during the emergency period	Bureau discretion within statutory maximums

Closing

The First Step Act is best read as five statutes in one: a sentencing reform that narrowed the harshest enhancements while leaving the minimums themselves intact, a retroactivity provision that corrected the oldest acknowledged disparity in federal drug sentencing, a programming and credit system that rewards participation while excluding the most serious offenders, a technical correction that restored the good conduct time Congress had intended in 1984, and a procedural amendment to compassionate release that became the statute’s most used provision. Each of these does different work, reaches different people, and is limited in different ways, and the article’s claim is that the least discussed of the five mattered most.

The statute also rewards a particular way of reading legislation. The public debate centered on the sentencing provisions, which were the subject of the floor speeches and the advocacy campaigns. The policy analysis centered on the credit system, which was the subject of the implementation reports and the audits. The provision that changed the most lives was a paragraph about who may file a motion, a paragraph that almost no one discussed on the Senate floor and that few advocates listed among their priorities. The lesson is not that the debated provisions were unimportant: the sentencing changes and the credit system are real and durable. The lesson is that procedural provisions, which determine who gets into court and who decides, can matter more than substantive provisions, which determine what the court may do once the case is there. Access turned out to be worth more than entitlement, and that is the frame in which the act deserves to be remembered.

The statute’s significance depends on the framework around it. The comparison with the 1994 crime statute, which this act was in several respects written against, is carried in the comparison of the 1994 crime bill and the First Step Act, which sets the two legislative moments against each other. The provision made retroactive is treated in full in the guide to the Fair Sentencing Act of 2010, which explains the ratio the 2018 act finally applied to old sentences. The minimums the act narrowed are explained in the article on mandatory minimum sentencing law, which covers the architecture built in the 1980s. The decisions that construe the act are carried in the article on sentencing law in the Supreme Court, and the framework the act amends is laid out in the guide to the Sentencing Reform Act of 1984. Readers who keep notes on statutes, citations, and case chronologies can keep them together free on VaultBook with the VaultBook legislation study notebook.

The broader lesson is the one the statute keeps teaching. The sentencing changes were the subject of the public debate, and they were carefully bounded and prospective. The credit system was the subject of the policy analysis, and it was carefully hedged with exclusions. The provision that changed the most lives was a paragraph about who may file a motion, a paragraph that almost no one discussed on the Senate floor. Access to a court turned out to be worth more than every substantive entitlement in the bill combined, and that is the frame in which the act deserves to be remembered.

Frequently Asked Questions

Q: What did the First Step Act change?

The act changed federal sentencing and federal prison administration in five distinct ways. It reduced the enhanced mandatory minimums for repeat drug offenders and narrowed the prior convictions that trigger them. It broadened the safety valve so more low-level drug defendants can be sentenced below a mandatory minimum. It ended the practice of stacking multiple firearm enhancements in a single indictment. It made the 2010 Fair Sentencing Act retroactive, the only retroactive sentencing provision. And on the prison side, it created an earned time credit system for completing recidivism reduction programming, corrected the good conduct time calculation to 54 days per year of the sentence imposed, and allowed prisoners to petition courts directly for compassionate release. The act reached only the federal system, which holds below ten percent of incarcerated people in the United States.

Q: Does the First Step Act apply to state prisoners?

No. The act applies exclusively to the federal system. Section 105 of the statute states the point directly: the earned time provisions do not authorize prerelease custody or supervised release for anyone serving a sentence for a state offense. The risk assessment system, the programming, the credit calculations, and the compassionate release changes all operate inside the Bureau of Prisons. The federal system held roughly 180,000 people at enactment out of more than two million incarcerated people nationwide, so the act’s structural ceiling was set by jurisdiction rather than by policy. State prisoners are governed by their own states’ sentencing and good time laws, which vary widely and which this statute does not touch.

Q: What is 924(c) stacking under the First Step Act?

Stacking was the practice of charging multiple violations of 18 U.S.C. section 924(c), the firearm offense, in a single indictment and applying the enhanced consecutive term to each count. Under the Supreme Court’s 1993 decision in Deal v. United States, a second or subsequent 924(c) conviction did not require a prior conviction to be final, so a first-time defendant facing several counts in one case could receive consecutive 25-year enhancements on top of the first count, producing sentences of fifty years or more. Section 403 of the First Step Act clarified the statute so that the enhanced term for a second or subsequent conviction applies only where a prior 924(c) conviction became final before the new offense was committed. The change was the largest single fix to the longest federal sentences, but it was prospective only.

Q: How do earned time credits work in the First Step Act?

Eligible federal prisoners earn 10 days of time credits for every 30 days of successful participation in recidivism reduction programming or productive activities. Prisoners classified as minimum or low risk under the PATTERN assessment who maintain that classification over two consecutive assessments earn an additional 5 days per 30 days, for a total of 15. The credits are applied toward prerelease custody, meaning transfer to a halfway house or home confinement, or toward up to 12 months of early supervised release. Credits cannot be earned for programming completed before the act’s enactment or for programs completed while detained pending adjudication. A long list of offense categories, including terrorism, murder, robbery, sex offenses, and certain drug offenses, makes a prisoner ineligible to earn credits at all.

Q: Did the First Step Act change compassionate release?

Yes, and the change became the statute’s most consequential provision. Before the act, only the Bureau of Prisons could file a compassionate release motion under 18 U.S.C. section 3582(c)(1)(A), and the bureau filed a small number per year. Section 603 allowed prisoners to file the motion directly with the court after exhausting administrative remedies or after 30 days pass without action on a request to the warden. Courts decide the motions on the statutory standard of extraordinary and compelling reasons. During the public health emergency that began about fifteen months after enactment, courts granted thousands of such motions, converting a rarely used administrative mechanism into one of the largest bodies of federal criminal motion practice. The direct-filing right was made a permanent feature of federal sentencing law by the amendment.

Q: What did the First Step Act do to the good conduct time calculation?

The act corrected how the Bureau of Prisons computes good conduct time under 18 U.S.C. section 3624(b). The statute had long provided for up to 54 days of credit per year, but the bureau calculated it at 54 days per year of time actually served rather than per year of the sentence imposed, which reduced the effective credit to roughly 47 days per year of the stated sentence. The Supreme Court upheld the bureau’s method in Barber v. Thomas in 2010, so only Congress could change it. Section 102(b) of the act amended the provision to state the credit as 54 days per year of the sentence imposed. The bureau applied the corrected calculation immediately to every federal prisoner serving a sentence of more than one year, accelerating release dates across the population by roughly 70 days on a ten-year sentence.

Q: Which offenses cannot earn First Step Act time credits?

The exclusion list at 18 U.S.C. section 3632(d)(4)(D) covers dozens of offense categories. They include terrorism and related national security offenses, espionage, murder, manslaughter, assault, kidnapping, robbery, firearms offenses including being a felon in possession of a firearm, sex offenses and sexual exploitation offenses, human trafficking, and specified high-level drug offenses. Prisoners subject to a final order of removal under the immigration laws are also ineligible. The practical effect is that the credit system reaches most strongly the nonviolent drug and property offenders who make up the bulk of minimum and low security federal custody, while violent offenders remain outside it entirely. Excluded prisoners may still participate in programming and earn the act’s non-credit incentives, such as additional phone privileges and visitation.

Q: Was the First Step Act bipartisan?

Overwhelmingly so. The Senate concurred in the House amendment on December 18, 2018 by 87 votes to 12, and the House agreed to the Senate amendment on December 20, 2018 by 358 votes to 36, with the two-thirds suspension threshold cleared easily. The coalition was ideologically unusual: conservative organizations framed the bill as redemption and public safety policy, progressive groups framed it as racial justice policy, and the White House made it a personal legislative priority. The coalition held because the sentencing provisions were narrow and prospective while the prison reform provisions were broad, a trade that secured senators who would not have supported a sentencing reform bill alone. The Cotton-Kennedy amendments that would have narrowed the bill were rejected 62 to 37, and a Cruz amendment expanding the exclusion list passed by voice vote.

Q: How does the PATTERN risk assessment assign categories?

PATTERN, the Prisoner Assessment Tool Targeting Estimated Risk and Needs, produces a general recidivism score and a violent recidivism score from variables including criminal history, age, education, and program participation. Each score places the prisoner in one of four categories: minimum, low, medium, or high risk. The category determines programming priority and, for minimum and low risk prisoners, eligibility for the additional 5 days of earned time credits per 30 days of programming. The Department of Justice released the tool in July 2019 and has revalidated and recalibrated it in response to critiques about racial disparities in classification rates. An Independent Review Committee reviews the system’s validity, and the Government Accountability Office audits its use, so the tool operates under more oversight than most actuarial instruments in criminal justice.

Q: What is the difference between earned time credits and good conduct time?

They are separate systems created by separate authorities. Good conduct time, under 18 U.S.C. section 3624(b) as corrected by the act, is earned automatically by avoiding disciplinary infractions and is available to every federal prisoner serving more than a year, at up to 54 days per year of the sentence imposed. Earned time credits, under 18 U.S.C. section 3632(d)(4), are earned only by successfully completing recidivism reduction programming or productive activities, at 10 days per 30 days of participation with an additional 5 for minimum and low risk prisoners, and they are unavailable to prisoners convicted of the excluded offenses. Good time reduces the sentence itself; earned time credits are applied toward prerelease custody or early supervised release. A prisoner may benefit from both.

Q: Can a prisoner earn credits for programs completed before the law passed?

No. The statute provides that prisoners cannot retroactively earn time credits for programs completed before the act’s enactment, and they cannot earn credits for programs completed while detained pending adjudication of their cases. This was a deliberate drafting choice: the credit system was designed as a forward-looking incentive for participation in the new programming framework rather than as a reward for past conduct. The distinction matters because federal prisoners had been completing vocational, educational, and drug treatment programs for years before the act, and a retroactive credit would have produced an immediate large-scale release effect that the bill’s drafters did not intend. Only the Fair Sentencing Act retroactivity provision, section 404, reaches backward, and it operates through resentencing rather than credits.

Q: What happens to earned time credits at the end of a sentence?

Earned time credits are applied toward prerelease custody or supervised release, not as a direct reduction of the sentence the way good conduct time is. A prisoner applies accumulated credits to transfer earlier to a residential reentry center or to home confinement, or to begin up to 12 months of supervised release early. If a prisoner does not have enough remaining sentence time for the credits to be applied, or if the prisoner is ineligible for prerelease placement for security reasons, the credits may go unused. The bureau’s implementation rules determine the mechanics of application, and prisoners near the end of short sentences may find that credits earned late in the term have limited practical value. The design reflects the statute’s purpose: the credits are an incentive for programming, not a general sentence reduction.

Q: Who sponsored the First Step Act and who opposed it?

The Senate effort was led by Chuck Grassley, Dick Durbin, Cory Booker, and Mike Lee, with the House prison reform bill introduced by Doug Collins and Hakeem Jeffries. The White House, through the President and his senior adviser on the issue, made the bill a personal priority. Opposition came from two directions. Some senators, led by Tom Cotton, argued the bill was too lenient and offered amendments to expand the exclusion list and require victim notification; those amendments were rejected 62 to 37. Some progressive advocates argued the bill was too modest, leaving mandatory minimums for first-time offenders intact and excluding too many offense categories from the credit system. The overwhelming final votes show that both sets of objections were minority positions in each chamber.

Q: How many people received sentence reductions under the retroactivity provision?

The Sentencing Commission’s retroactivity reporting documented several thousand sentence reductions under section 404, with the grants recorded offense by offense and the average reduction measured in months. The provision allowed defendants convicted of crack cocaine offenses and sentenced before August 3, 2010 to petition for resentencing as if the Fair Sentencing Act of 2010 had applied at the time. Courts retained discretion to deny the motions, and defendants whose sentences had already been reduced under the 2010 act were excluded. The reductions were concentrated among defendants serving long sentences imposed under the old 100-to-1 crack-to-powder ratio, and the provision’s reach was limited to that defined group. It remains the only retroactive sentencing provision in the statute.

Q: What is the safety valve and how did the act expand it?

The safety valve, at 18 U.S.C. section 3553(f), allows a judge to sentence a drug defendant below an otherwise applicable mandatory minimum when five criteria are met: limited criminal history, no violence or weapon, no serious bodily injury, no leadership role in the offense, and truthful disclosure of information about the offense. Before the act, a defendant with more than one criminal history point was ineligible. Section 402 raised the ceiling to four criminal history points, with a refinement excluding defendants whose four points arise from a prior three-point offense or a prior two-point violent offense. The expansion made the valve available to a substantially larger group of low-level drug defendants whose records consist of minor offenses. The Sentencing Commission documented a marked increase in the share of drug defendants qualifying after the change.

Q: Does the act change the 85 percent truth in sentencing rule?

No. The requirement that federal prisoners serve at least 85 percent of their stated sentences, imposed by the Sentencing Reform Act of 1984, remains intact. Good conduct time can reduce time served to roughly 85 percent of the sentence, and the act’s correction of the good time calculation restored the full 54 days per year that the 1984 act had intended, but it did not change the 85 percent floor. Earned time credits do not reduce the sentence itself; they are applied toward prerelease custody or early supervised release. The distinction matters for understanding the act’s limits: even a prisoner who earns the maximum credits and the maximum good time still serves the great majority of the stated sentence, and the act’s release effects operate at the margins of the sentence rather than on its core.

Q: What did the act do about pregnant prisoners?

Title III of the act prohibits the use of restraints, including shackles and handcuffs, on federal prisoners during pregnancy, labor, delivery, and postpartum recovery, with exceptions where the prisoner presents an immediate risk of flight or a threat to safety. The provision was one of the least contested in the statute and attracted broad support. The Bureau of Prisons was also required under Title VI to provide feminine hygiene products without charge. These provisions reflect the portion of the act concerned with conditions of confinement rather than with sentencing or release, and they apply to the federal women’s population, a small but growing share of federal custody. The restraint prohibition is stated as a direct command to the bureau rather than as a standard for courts to enforce case by case.

Q: How did the act affect elderly prisoners?

The act reauthorized and modified the elderly offender early release pilot program created by the Second Chance Act of 2007. Under the reauthorized pilot, the Bureau of Prisons may place elderly offenders on home confinement if they meet age, time-served, and risk criteria, and the act expanded eligibility to include terminally ill offenders. The provision operates as a pilot rather than as a general entitlement, and the bureau retains discretion over individual placements. The reauthorization was part of Title V’s broader renewal of Second Chance Act programs, which also covered reentry courts, career training, and substance abuse treatment. The elderly pilot has produced a modest number of home confinement placements relative to the act’s other release mechanisms, and its implementation has been less documented than the credit and compassionate release provisions.

Q: What is the Independent Review Committee?

The Independent Review Committee is a body established by section 107 of the act to assist the Department of Justice in developing and validating the risk and needs assessment system. Its statutory duties include reviewing existing prisoner risk assessment tools, recommending evidence-based recidivism reduction programs and productive activities, researching the effectiveness of such programs, and reviewing and validating the PATTERN system the department adopted. The committee’s reviews have been the principal public venue for the methodological debate over the tool’s fairness and predictive validity. The statute also requires the Government Accountability Office to audit the system’s use at bureau facilities, creating two parallel oversight mechanisms: the committee’s expert review of the tool itself and the audit office’s review of how the tool is applied.

Q: Did the act reduce the federal prison population?

The federal prison population continued a decline that had begun in 2013, falling further in the years after enactment, but the decline reflected multiple forces and no serious analyst attributes it to the statute alone. Charging decisions by federal prosecutors, sentencing trends, and the good time recalculation all contributed, and the act’s provisions operated alongside them. The statute’s contribution is best measured provision by provision: several thousand section 404 sentence reductions, the good conduct time correction applied to the entire population, thousands of compassionate release grants, and a growing number of credit-driven transitions to prerelease custody and early supervised release. Each figure comes from an official source for its own period. The population effect is real but bounded by the federal-only scope: the federal system holds below ten percent of all incarcerated people in the United States.