Skip Navigation
Expert Brief

What Is the First Step Act — And What’s Happening With It?

While much of the law is operating as intended, a secret change is keeping more people in federal prison during the pandemic.

Published: June 23, 2020

The First Step Act is inten­ded to do two things: cut unne­ces­sar­ily long federal sentences and improve condi­tions in federal prison.

More than a year after it was enacted in 2018, key parts of the law are work­ing as prom­ised, restor­ing a modicum of fair­ness to federal senten­cing and help­ing to reduce the coun­try’s uncon­scion­ably large federal prison popu­la­tion. But other parts are not, demon­strat­ing the need for contin­ued advocacy and more congres­sional over­sight. The way the Justice Depart­ment has been hand­ling pris­oner releases during the coronavirus pandemic gives some insight into what’s going wrong.

Pres­id­ent Trump has bragged about sign­ing the law, which was the first crim­inal justice reform bill passed in nearly a decade. But simply sign­ing it is not enough. He needs to see it through.

Fixing a broken system

The First Step Act is the product of years of advocacy by people across the polit­ical spec­trum. Indeed, a very similar bipar­tisan bill nearly passed in 2015, but was dragged down by elec­tion-year polit­ics. The Trump admin­is­tra­tion began work­ing on its own crim­inal justice bill in early 2018, and an initial deal was cata­lyzed by a core group of bipar­tisan legis­lat­ors. It was then refined through a series of comprom­ises and, once the Senate decided to pick up the bill, sailed through both houses of Congress with super­ma­jor­ity support.

The law we now know as the First Step Act accom­plishes two discrete things, both aimed at making the federal justice system fairer and more focused on rehab­il­it­a­tion.

Its senten­cing reform compon­ents shorten federal prison sentences and give people addi­tional chances to avoid mandat­ory minimum penal­ties by expand­ing a “safety valve” that allows a judge to impose a sentence lower than the stat­utory minimum in some cases. These parts of the First Step Act are almost auto­matic: once the act was signed, judges imme­di­ately began senten­cing people to shorter prison terms in cases came before them. Simil­arly, people in federal prison for pre-2010 crack cocaine offenses imme­di­ately became eligible to apply for resen­ten­cing to a shorter prison term.

The law’s prison reform elements are designed to improve condi­tions in federal prison in two ways. One is by curb­ing inhu­mane prac­tices, such as elim­in­at­ing the use of restraints on preg­nant women and encour­aging placing people in pris­ons that are closer to their famil­ies. The other is by reori­ent­ing pris­ons around rehab­il­it­a­tion rather than punish­ment. That is no small task. Success­fully expand­ing rehab­il­it­at­ive program­ming in federal prison will require signi­fic­ant follow-through from Congress and the Depart­ment of Justice. It’s no surprise, then, that these distinct parts of the act are func­tion­ing differ­ently.

Relief from harsh sentences

First, the good news. The First Step Act’s senten­cing reforms are off to a strong start.

Before 2010, an offense involving 5 grams of crack cocaine, a form of the drug more common in the Black community, was punished as severely as one involving 500 grams of powder. The Fair Senten­cing Act of 2010 changed that, redu­cing this 100:1 dispar­ity to 18:1 — but only on a forward-going basis. People convicted under the now-outdated crack laws were stuck serving the very sentences that Congress had just repu­di­ated.

The First Step Act fixed that by making the Fair Senten­cing Act retro­act­ive. Accord­ing to the Justice Depart­ment, as of May, roughly 3,000 people serving outdated sentences for crack cocaine crimes had already been resen­tenced to shorter prison terms. Shortened or bypassed mandat­ory minim­ums mean that every year another estim­ated 2,000 people will receive prison sentences 20 percent shorter than they would have.

Another 3,100 people were released in July 2019, anywhere from a few days to a few months early, when the First Step Act’s “good time credit fix” went into effect. This simple change allowed people in federal prison to earn approx­im­ately an extra week off their sentence per year — and it applied retro­act­ively, too. For some people it meant seeing their friends and family months earlier than expec­ted.

Taken together, these changes repres­ent an import­ant decrease in incar­cer­a­tion. One year after the First Step Act was signed, the federal prison popu­la­tion was around 5,000 people smal­ler, continu­ing several years of declines. It has contin­ued to shrink amidst the coronavirus pandemic.

To be sure, there’s still a long way to go: the federal prison popu­la­tion remains sky-high. And the Depart­ment of Justice does not appear to be in complete lock­step with the White House’s celeb­ra­tion of the law. In some old crack cocaine cases, federal prosec­utors are oppos­ing resen­ten­cing motions or seek­ing to rein­car­cer­ate people who have just been released. Prosec­utors in these cases argue that any motions for resen­ten­cing must also consider the (often higher) amount of the drug the applic­ant possessed accord­ing to presen­tence reports. Other tech­nical disputes are also crop­ping up, with the Depart­ment of Justice often arguing for a narrow inter­pret­a­tion of the First Step Act.  

Toward a better prison system

When it comes to improv­ing program­ming within federal prison, even more work remains to be done. Federal pris­ons offer some oppor­tun­it­ies for people in prison to parti­cip­ate in services that either address their indi­vidual needs, or help prepare them for life after release. Drug treat­ment and drug educa­tion are two examples; others include ESL and educa­tional classes.

The First Step Act calls for the Bureau of Pris­ons to signi­fic­antly expand these oppor­tun­it­ies. Within a few years, the BOP must have “evid­ence-based recidiv­ism reduc­tion programs and product­ive activ­it­ies” avail­able for all people in prison.  That means voca­tional train­ing, educa­tional classes, and beha­vi­oral ther­apy (to name a few options recom­men­ded in the act) should be staffed and broadly avail­able. Parti­cip­at­ing in these programs will in turn enable imprisoned people to earn “time cred­its” that they can put toward a trans­fer to prerelease custody — that is, a halfway house or even home confine­ment — theor­et­ic­ally allow­ing them to finish their sentence outside of a prison.

Rolling out this system as inten­ded will be a chal­lenge, however, in part because BOP programs are already under­staffed and under­fun­ded. Around 25 percent of people spend­ing more than a year in federal prison have completed zero programs. A recent BOP budget docu­ment described a lengthy wait­ing list for basic liter­acy programs. And accord­ing to the Federal Defend­ers of New York, the true extent of the BOP’s program­ming short­fall isn’t even known, because the BOP won’t share inform­a­tion on program­ming avail­ab­il­ity or capa­city. The bureau offered little insight into its exist­ing capa­city and needs at a pair of congres­sional over­sight hear­ings last fall. Worse, a recent report from the First Step Act’s Inde­pend­ent Review Commit­tee, made up of outside experts to advise and assist the govern­ment with imple­ment­ing the law, casts doubt on the qual­ity of the BOP’s exist­ing program­ming.

The BOP’s lack of trans­par­ency also makes it hard to know how “time cred­its” are being awar­ded — and thus, whether the BOP is permit­ting people to make progress toward prerelease custody as Congress inten­ded. The First Step Act provides that most imprisoned people may earn 10 to 15 days of “time cred­its” for every 30 days of “success­ful parti­cip­a­tion” in recidiv­ism-reduc­tion program­ming. But a recent report suggests the BOP will award a specific number of “hours” for each program. How many “hours” make up a “day,” for the purpose of award­ing cred­its? This highly tech­nical issue may appear trivial, but could signi­fic­antly affect the reach of the First Step Act.

People with inside know­ledge of the system point to other concerns, too. The law excludes some imprisoned people from earn­ing cred­its based on the crime that led to their incar­cer­a­tion or their role in the offense, and some advoc­ates report that the BOP has applied those exclu­sions broadly, disqual­i­fy­ing a much larger part of the imprisoned popu­la­tion than Congress inten­ded. Advoc­ates also have heard that program avail­ab­il­ity in pris­ons is much spot­tier than the BOP has sugges­ted, render­ing illus­ory the DOJ’s claim that people are already being assigned to programs and “product­ive activ­it­ies” tailored to their needs.

While some progress is being made, all of these concerns point to a need for more congres­sional over­sight — narrowly focused on the avail­ab­il­ity and capa­city of “recidiv­ism reduc­tion” program­ming in the BOP.

Fund­ing Insec­ur­ity

Any expan­sion of program­ming also won’t be free. Know­ing that, the First Step Act author­ized $75 million per year for five years for imple­ment­a­tion. But author­iz­a­tion is only the begin­ning of the budget­ing process. Congress must also form­ally appro­pri­ate money to the BOP to fund the First Step Act for each year that it is author­ized.

Thank­fully, in Decem­ber 2019, Congress passed and the pres­id­ent signed the Consol­id­ated Appro­pri­ations Act, which included full fund­ing for the First Step Act through the end of the current fiscal year. That makes up for a bumpy start: last year, Congress failed to appro­pri­ate anything at all, forcing the DOJ to use $75 million from else­where in its budget to cover the tempor­ary short­fall.

While this new fund­ing is good news, it still might not be enough. Accord­ing to a member of the Inde­pend­ent Review Commit­tee, full imple­ment­a­tion may cost closer to $300 million. After factor­ing in “train­ing, staff­ing, [and] build­ing things like classrooms,” he says, $75 million simply may not be enough. But any more substan­tial fund­ing seems unlikely to mater­i­al­ize: in a recent budget, the White House sought nearly $300 million to fund improve­ments related to the law, of which only $23 million was earmarked for program­ming.

Under­valu­ing rehab­il­it­a­tion and entrench­ing bias

In late March, the Justice Depart­ment appeared to finally acknow­ledge the need to trans­fer people in federal prison to home confine­ment to keep them safe from the coronavirus. But trans­fers have been slow, and a recent report from ProP­ub­lica shows one reason why.

The DOJ prior­it­ized trans­fers for people deemed to pose a “minimum” risk of recidiv­ism under a new system developed for the First Step Act. But that system, called “PATTERN,” has never been perfect, and appears to have been quietly revised to make it more diffi­cult to reach a “minimum” score — and by exten­sion, that much harder to win a trans­fer to home confine­ment. This revel­a­tion offers the latest example of how the First Step Act’s imple­ment­a­tion has proceeded in fits and starts.

Before imprisoned people can use “time cred­its” they’ve earned from prison program­ming, two condi­tions must be met. First, if they’re seek­ing a trans­fer to a halfway house, there must be a bed avail­able. That’s not a sure thing, espe­cially after a round of budget cuts by the Trump admin­is­tra­tion. Second, an incar­cer­ated person must demon­strate that their risk of commit­ting a new crime is low, as calcu­lated by PATTERN.

The process of devel­op­ing that tool has not gone smoothly. The first version unveiled by the DOJ, in July 2019, appeared to use a method for calcu­lat­ing “risk” that over­stated the actual risk of re-offend­ing among formerly incar­cer­ated people, exag­ger­ated racial dispar­it­ies, and gave people only marginal credit for complet­ing educa­tion, coun­sel­ing, and other programs while in prison. Taken together, the tool gave short shrift to the idea that people can change while in prison — the very premise of the First Step Act.

Early this year, the DOJ shared what seemed like good news: PATTERN, it said, had been improved to partially address some of the concerns raised by the Bren­nan Center and others. But the tool contin­ued to use an overly broad defin­i­tion of recidiv­ism. And while the DOJ claimed it intro­duced changes to reduce racial dispar­it­ies, it did not release data on the revised tool’s actual effect on racial dispar­it­ies. As a result, imprisoned people, famil­ies, lawyers, and advoc­ates were all concerned when the DOJ announced it would use PATTERN to help determ­ine who would be trans­ferred out of federal prison during the pandemic.

Those concerns were justi­fied. It now seems that the BOP changed PATTERN more signi­fic­antly than they initially disclosed to make it much harder to qual­ify as a “low” or “minimum” risk. (A DOJ report, released a week after ProP­ub­lica’s story, suggests that the impact of these changes was relat­ively small, but offers limited data, and — still — provides no analysis of racial dispar­it­ies.) This change will certainly narrow the effect of the First Step Act. But more urgently, it seems likely to keep more people in federal prison, exposed to a heightened risk of catch­ing a deadly disease.

It’s tempt­ing to write off these tech­nical imple­ment­a­tion issues, but they carry real consequences for imprisoned people and their famil­ies.

Trump has repeatedly claimed credit for passing the First Step Act. He did the same in a Super Bowl ad. Now his admin­is­tra­tion needs to do the hard work of ensur­ing that the law lives up to its prom­ise.