How Do First Step Act (FSA) Earned Time Credits Work?

First Step Act earned time credits are days taken off the back end of a federal sentence, earned at 10 days for every 30 days of successful participation in approved programming, or 15 days for every 30 at minimum or low risk held across two consecutive assessments. Up to 365 of those days can move you onto supervised release early, and credits past that buy time in a halfway house or on home confinement instead of a prison.

Almost everything that goes wrong happens in the gap between earning a credit and getting it applied. The statute and the Bureau treat those as two separate questions with two separate sets of rules.

How you earn credits, and what counts as programming

The rate comes from 18 U.S.C. 3632(d)(4)(A). Ten days per 30 days of successful participation, plus an additional five days per 30 for a prisoner at minimum or low risk who, in the words of the statute, “over 2 consecutive assessments, has not increased their risk of recidivism.” The regulation states that second condition differently, as having “maintained a consistent minimum or low risk of recidivism” across the two most recent assessments (28 C.F.R. 523.42(c)(2)), and the Bureau applies its own wording.

What counts is broader than most people expect. The list of qualifying activity types at 28 C.F.R. 523.41(a) runs to thirteen items, among them academic classes, vocational training, substance abuse treatment, parenting, mentoring, victim impact classes, and trauma counseling. Inmate work and employment opportunities sit on that list at 523.41(a)(11), which is how a prison job assignment can count. Productive Activities are defined separately, at 523.41(b), as an activity that lets an inmate remain productive and thereby maintain or work toward a minimum or low risk of recidivating. Attorney Peter Tomasek walked our group through the calculation in his October 2022 Speaker Series talk.

The counting method was fought over and then changed. The proposed rule defined a “day” of participation as one eight-hour period of programming, so four days at two hours a day would have earned one day of credit. More than 150 commenters objected, and the final rule at 87 Fed. Reg. 2705 dropped that in favor of a flat thirty-day-period model. The program statement counts those periods as days in earning status.

Credits are auto-calculated in 30-day increments and partial credit is not awarded, so days short of 30 carry into the next cycle instead of posting. A time credit worksheet goes into your central file at each auto-calculation, and you get a copy at your regularly scheduled program review.

The two assessments that sit underneath every credit

Two instruments run underneath the credit, and they answer different questions. Mixing them up is why people assume any completed class earns something.

PATTERN, the Prisoner Assessment Tool Targeting Estimated Risk and Needs, answers the risk question and sorts a person into minimum, low, medium, or high. The Bureau publishes separate scoring instruments for men and women, the violent-offense codes that feed the score, and the cut points that turn a raw score into a category. Version 1.3 is the one in current use, built, the Bureau says, so that emphasis falls on “a system that accurately measures an inmate’s change during incarceration, and provides opportunities for inmates to reduce their risk scores during periodic reassessments.” The result does two jobs: it sets whether you earn 10 or 15 days per 30, and it gates whether credits can be applied at all.

SPARC-13, the Standardized Prisoner Assessment for Reduction in Criminality, answers the needs question across thirteen areas, and staff recommend programs targeting the ones flagged for that individual.

The thirteen SPARC-13 need areas
Anger/Hostility	Family/Parenting	Substance Use
Antisocial Peers	Finance/Poverty	Trauma
Cognitions	Medical	Work
Dyslexia	Mental Health
Education	Recreation/Leisure/Fitness

The link between the two decides whether a class counts. Under 28 C.F.R. 523.41(c)(2), successful participation means participating in the programs “that the Bureau has recommended based on the inmate’s individualized risk and needs assessment.” A course signed up for independently, outside your assessed needs, is not what the regulation describes.

Reassessment is not optional either. Under 18 U.S.C. 3632(d)(5) someone who successfully participates is reassessed at least annually, more often at medium or high risk within five years of a projected release date. Because the rate and the ability to apply credits both key off the two most recent assessments, the release estimate moves when the score moves.

The programs that actually generate the credit

The Bureau’s First Step Act Approved Programs Guide is the catalog unit teams work from. Its September 2023 edition lists 52 Evidence-Based Recidivism Reduction programs plus structured, curriculum-based Productive Activities, spread across 122 institutions and mapped to the thirteen need areas. Availability varies enormously by program.

Program	Need area it is coded to	Where it is offered
Bureau Literacy Program (GED)	Education	All institutions
Vocational and Certification Course Training	Work	All institutions
Apprenticeship Training	Work	All institutions except those exempt by policy
Money Smart for Adults	Finance/Poverty	All institutions
Federal Prison Industries (UNICOR)	Work	51 named sites
Post-Secondary Education	Work	15 named sites
Residential Drug Abuse Program	Antisocial Peers, Cognitions, Substance Use, and Mental Health (Dual Diagnosis and FIT)	Designated RDAP sites

The right-hand column is the part that decides a real timeline. UNICOR work counts as programming in its own right, and the guide gives hiring priority to people within three years of release, but post-secondary coursework runs at fifteen institutions out of 122.

How the credits actually move your date

Earned credits have two destinations, with a different ceiling on each.

Where the credits go	Ceiling	Authority
Early transfer to supervised release	12 months, applied by the Bureau as up to 365 days	18 U.S.C. 3624(g)(3)
Time in a halfway house or on home confinement	No maximum stated in 18 U.S.C. 3624(g); drawn from credits not used on the 365 days	18 U.S.C. 3624(g)(2); Program Statement 5410.01

The order matters. Good conduct time is applied first, then up to 365 days of time credits move the transfer to supervised release, and only what is left over reaches prerelease custody. That sequence is why two people with identical credit balances can end up with very different amounts of halfway house time.

It also separates this pathway from the Bureau’s older prerelease authority, where 18 U.S.C. 3624(c)(2) caps home confinement at “the shorter of 10 percent of the term of imprisonment of that prisoner or 6 months.” Prerelease custody funded by time credits under 3624(g) has no such cap. It has a floor instead: under 3624(g)(2)(A)(iv) that prisoner “shall remain in home confinement until the prisoner has served not less than 85 percent of the prisoner’s imposed term of imprisonment.”

Before anything applies, 18 U.S.C. 3624(g)(1) and 28 C.F.R. 523.44(b) require credits equal to the remainder of your imposed term plus reassessments showing a demonstrated risk reduction or a maintained minimum or low risk. Prerelease custody adds minimum or low risk on your last two assessments, or a warden-approved petition finding you are not a danger to society, made a good faith effort to lower your risk, and are unlikely to recidivate. Early supervised release needs minimum or low risk on your last assessment plus a term of supervised release imposed by the court, and the Bureau applies up to 365 days automatically once that holds with no detainer, unresolved pending charge, final order of removal, or opt-out. People at medium or high risk earn credits but get no estimated First Step Act release date.

Halfway house and home confinement referrals ordinarily go to the Residential Reentry Management office 12 months before the projected release date, or at least 60 days before the projected placement date, whichever is greater. The date you get depends on your release residence, program requirements, and contract bed space. Sam Mangel and Craig Rothfeld covered placement in their Speaker Series conversation on prison consulting.

Who cannot earn credits, and who can earn but cannot use them

These are two different lists, and conflating them is the most common error in secondhand explanations.

Ineligible to earn at all: anyone serving a sentence for an offense named in 18 U.S.C. 3632(d)(4)(D). The list runs from clause (i) to clause (lxviii) and is dominated by homicide, kidnapping, sexual offenses, terrorism, espionage, weapons, and drug offenses causing death or serious injury. For a white collar case the notable thing is what is absent: wire fraud, mail fraud, securities fraud, bank fraud, tax offenses and false statements appear nowhere on it. A few white collar adjacent offenses do appear: computer offenses under 18 U.S.C. 1030(a)(1), sanctions violations under section 206 of the International Emergency Economic Powers Act, and Export Administration Act offenses. The Bureau publishes its own list of disqualifying offenses.

Able to earn but barred from applying: under 18 U.S.C. 3632(d)(4)(E), someone subject to a final order of removal accumulates credits but cannot use them, and the regulation adds that credits cannot be applied to a sentence under laws other than the U.S. Code, which covers state sentences. That paragraph carves out D.C. Code sentences, but the program statement says D.C. Code inmates can neither earn nor apply credits unless and until the D.C. Council defines eligibility. Bureau policy reaches further than the statute here: Program Statement 5410.01 provides that people with unresolved pending charges or detainers, including unresolved immigration status, may earn credits but cannot apply them until those resolve. That bar is agency policy rather than statutory text, because 3632(d)(4)(E) names only a final order of removal.

One disqualifying count can take down an entire aggregated sentence

The trap in a multi-count case appears nowhere in the First Step Act itself. It comes from 18 U.S.C. 3584(c), which provides that multiple terms of imprisonment ordered to run consecutively or concurrently “shall be treated for administrative purposes as a single, aggregate term of imprisonment.”

The Bureau reads that into credit eligibility, and the courts of appeals have agreed. In Giovinco v. Pullen, 118 F.4th 527 (2d Cir. 2024), the Second Circuit held that “pursuant to the aggregation provision, a prisoner ‘is serving a sentence for’ any offense that is part of his aggregated term of imprisonment.” The Fourth Circuit reached the same result, over a dissent, in Bonnie v. Dunbar, No. 24-6665 (4th Cir. Nov. 5, 2025), where a consecutive 24-month term tied to an 18 U.S.C. 924(c) conviction wiped out credits across a 144-month aggregate. The Sentencing Commission describes the rule the same way: credits can be earned only by someone not serving a sentence for a disqualifying conviction “alone or as part of an aggregated sentence.”

Three consequences follow. The disqualifying count does not have to be the longest. It does not have to be the count currently being served, because aggregation collapses that distinction. And finishing the portion attributable to it does not restore eligibility, which is the argument the Second Circuit rejected.

What stops the clock, and how earned credits get taken away

Being eligible is not the same as being in earning status on a given day. Under 28 C.F.R. 523.41(c)(4) and Section 7 of the program statement, you are generally not earning while designated outside the institution on a medical or escorted trip or full-day furlough, while in transit or in another jurisdiction’s custody on a writ, while on a mental health hold, while held as a material witness or in civil contempt or commitment, or while opted out. The texts differ on segregation: the regulation says a Special Housing Unit, the program statement says disciplinary segregation status, and the Bureau’s FAQ says “SHU disciplinary segregation.”

Opt-out status is the other half, and the Approved Programs Guide lists the triggers alongside what reverses each one.

What puts you in opt-out status	What returns you to opt-in
Refusing an initial needs assessment, in TRULINCS or with staff	Completing the initial assessments
Declining recommended programming tied to an identified need	Removal or change of the program decline code
Refusing the Financial Responsibility Program	Participating in the program
Placement in disciplinary segregation	Automatic on removal from that status
Leaving a designated institution for a full calendar day or more	Automatic on return to a designated institution

Days spent in opt-out status are not reinstated retroactively, and opting out is documented but is not itself a disciplinary violation.

Credits already earned can be taken only through the disciplinary process. Loss of time credits is a sanction imposed by a Discipline Hearing Officer on a finding that you committed a prohibited act, it cannot be suspended, and future credits are not forfeitable. Restoration is decided case by case by the warden, only after clear conduct across two consecutive assessments, with a six-month wait after a denial.

Good conduct time and RDAP are different benefits

Good conduct time comes from 18 U.S.C. 3624(b), up to 54 days for each year of the sentence imposed, and it vests at release. Time credits are separate, and 18 U.S.C. 3632(d)(6) says the First Step Act incentives are in addition to any other reward a prisoner may be eligible for.

The Residential Drug Abuse Program has its own sequence. Its unit-based component must run at least six months under 28 C.F.R. 550.53(a)(1), the Bureau describes the program as typically nine months, and completing every component including community treatment can support a reduction of up to 12 months under 18 U.S.C. 3621(e). Under the program statement that benefit is applied first and time credits after it; if there is not enough time for both, the credit days applied are reduced so the required 120-day community treatment component still fits. Sam Mangel, who sponsors the support group and spoke to our members about prison consulting, keeps a practitioner walkthrough of RDAP eligibility, the interview and how the time off is calculated that follows the same regulations from the inside of a case.

A third authority sits outside both, and unlike credits it turns on age rather than programming: home detention under 34 U.S.C. 60541(g), for someone at least 60 years old who has served two-thirds of the term with no conviction for a crime of violence, sex offense or terrorism offense.

What the Bureau’s own numbers show

The Sentencing Commission published a fact sheet in December 2024 on everyone released from federal custody in calendar year 2023 after earning and applying time credits. It is the clearest available picture of what credits are worth in practice rather than on paper.

Group	People	Average credits earned	Share of imposed sentence
Earned and applied credits	17,465	10.3 months	16.9%
Applied to early supervised release only	10,607	5.1 months	16.0%
Applied to both supervised release and prerelease custody	6,858	18.5 months	18.4%

Where those people were living on their last day in custody says as much as the totals. Of the 18,009 in the dataset, 40.0 percent went to supervised release straight from secure custody, 33.5 percent from home confinement, and 26.5 percent from a Residential Reentry Center. The 365-day cap drives that split, because credits beyond it can only buy community placement.

The offense breakdown is where a white collar reader will find themselves. Fraud accounted for 13.9 percent of the group that earned and applied credits and 0.0 percent of the 9,328 people scheduled for release that year who were statutorily ineligible. Money laundering ran 4.1 percent against 0.2 percent. Women were 19.9 percent of the credit-applying group against 4.6 percent of the ineligible group, a gap that makes these rules unusually consequential for women in federal prison.

Where the law is still unsettled

Two questions are open, and the answer depends on which circuit you are in.

The first is whether unused credits can shorten a term of supervised release or only start it early. The Ninth Circuit held in Gonzalez v. Herrera, No. 24-2371 (9th Cir. Aug. 19, 2025), that credits may reduce the length of time spent on supervised release. The Sixth Circuit went the other way, over a dissent, in Hargrove v. Healy, No. 24-3809 (6th Cir. Sept. 10, 2025). In March 2026 the Second Circuit joined the Sixth in Rivera-Perez v. Stover, No. 25-149, holding that credits “do not reduce a post-incarceration term of supervised release.” That is a live circuit split, and neither answer is settled federal law.

A separate route to a shorter term exists no matter how that split resolves, and the Bureau plays no part in it. Under 18 U.S.C. 3583(e)(1), the sentencing court may end supervision any time after one year has been served if warranted by the person’s conduct and the interest of justice. That is early termination, a judicial decision rather than a credit calculation.

The second open question is when earning starts. The Bureau’s regulation says credits begin accruing on arrival or voluntary surrender at the designated facility, so nothing accrues in pretrial detention or in Marshals Service custody before designation. In April 2026 the First Circuit held in Miles v. Bowers, No. 25-1291, that the statute instead ties eligibility to commencement of the sentence under 18 U.S.C. 3585(a), and that the regulation “plainly conflicts with the text of the FSA.” The same opinion held that completing the risk and needs assessment is not a prerequisite to earning credit. Neither the regulation nor the program statement has changed.

When your worksheet and your release date disagree

That mismatch is documented at the agency level. In a report published January 27, 2026, the Government Accountability Office found that the Bureau completed initial risk assessments on time for about 75 percent of the 57,902 people who entered its facilities between June 2022 and March 2024, that staff recorded program declines inconsistently, and that credits reached supervised release reliably but prerelease custody with gaps. The Bureau agreed to all six recommendations.

Disputes about a calculation run through the Administrative Remedy Program at 28 C.F.R. Part 542, the process the Bureau directs inmates to use and the one courts ordinarily expect to be completed before they will take up the same question on a habeas petition.

Step	Form	Deadline to file	Time to respond
Informal resolution with staff	Local form, commonly called a BP-8	Before the BP-9	Set locally
Request to the warden	BP-9	20 calendar days from the date the basis occurred	20 calendar days
Appeal to the Regional Director	BP-10	20 calendar days from the warden’s signed response	30 calendar days
Appeal to the General Counsel	BP-11	30 calendar days from the Regional Director’s signed response	40 calendar days

Each response deadline can be extended once, and if no response arrives in the allotted time, including any extension, the absence of a response may be treated as a denial at that level. Our explainer on the BOP administrative remedy process walks through that filing sequence in detail. Program Statement 5410.01 and the Approved Programs Guide, listed below, are what your unit team works from.