Waivers of Compassionate Release in Plea Bargains: The Need for Administrative Action to Prevent Unfair Sentencing
Judge Charles Breyer of the United States District Court for the Northern District of California made headlines in May of 2020 with his decision to reject a plea agreement in United States v. Osorto (2020). As written, defendant Osorto’s plea agreement waived his future right to appeal for compassionate release, a Congressionally-created program that allows inmates to leave prison before the end of their sentences under extraordinary circumstances. Congress originally vested the right to grant compassionate release solely with the Federal Bureau of Prisons (BOP). In this case, Judge Breyer found that the plea agreement’s express waiver of the defendant’s right to file for compassionate release with the federal district court, a right newly created by the First Step Act of 2018, was invalid on two counts.
First, Judge Breyer held that a defendant’s waiver of his right to file directly in federal district court for compassionate release frustrates congressional intent. Under the original statute, compassionate release could only be obtained through a lengthy process with the inmate’s BOP warden. Few inmates received grants for compassionate release under the initial law. With the First Step Act, Congress aimed to increase the number of defendants who could receive a sentence reduction through compassionate release. By allowing inmates to petition the courts directly, the First Step Act eliminated one of the primary barriers to obtaining compassionate release: bureaucratic red tape. From 2014 to 2018, 3,182 inmates applied for compassionate release. The BOP granted 9.6% of these petitions. The average amount of time for processing petitions was 141 days for approvals and 196 days for denials; 81 inmates died while waiting for their petitions to be resolved.
The First Step Act’s impact was immediate. In 2019, 145 inmates were granted compassionate release, amounting to roughly half the number of inmates released in the preceding four years. Two-thirds of these granted petitions were filed with district courts directly by defendants who had exhausted their nonjudicial options—the First Step Act opens the judicial avenue for relief to petitioners who have waited thirty days without a decision from the BOP or have otherwise had their petitions rejected by the BOP. The United States Sentencing Guidelines have not yet been amended to reflect the First Step Act, and district court judges use the same standard for release as the BOP. That the courts were able to process and grant more petitions under the same standard, despite already-overloaded dockets, further illustrates the severity of the prior procedural barrier.
In its first year, the First Step Act clearly made compassionate release a more realistic possibility for inmates seeking sentencing relief due to extraordinary and compelling circumstances. And into its third year, spurred by COVID-19, over 3,200 inmates have been released under the extraordinary circumstances created by the virus, with an ostensibly greater number of inmates applying for release.1 Per Judge Breyer, to explicitly take away this new right under a plea agreement would frustrate Congress’s intent to reform criminal sentencing law under the First Step Act.
Second, though not the determinative basis for the decision, Judge Breyer found that a plea agreement requiring a defendant to waive his right to file for compassionate release is “appallingly cruel” and therefore an unconscionable contract. The purpose of compassionate release is to allow courts to reduce sentences based on “extraordinary and compelling” reasons. According to the United States Sentencing Guidelines § 1B1.13, such reasons could include a terminal illness or the death of a defendant’s child’s primary caretaker. Often, these circumstances are unexpected. In Judge Breyer’s view, requiring defendants to waive the right to apply for release with the court violates public policy, given the BOP’s historic ineffectiveness in processing these requests and the unpredictability of qualifying reasons for release.
Judge Breyer’s opinion is a groundbreaking statement on a novel issue created by the First Step Act, but the effects of his opinion have not yet been seen. His position as a United States Sentencing Commissioner suggest that there may be future changes to the Sentencing Guidelines consistent with his views. With half of the Commission’s voting seats still vacant, however, such changes to the Guidelines will not happen soon. Without executive or legislative action, uniform sentence reduction policy through the protection of a defendant’s right to file for compassionate release with the courts will not be achieved. Though Judge Breyer raises compelling and valid legal arguments for other district court judges to consider, they are not likely to be uniformly applied.
Consider a district court’s decision to dismiss a motion for compassionate release based on a plea agreement signed after the First Step Act. In United States v. Egebrecht (2020), decided after Osorto, the District Court for the Southern District of Indiana enforced a broad appellate plea waiver that the defendant had signed in April 2019, four months after the First Step Act was enacted. The court declined to consider the merits of her motion for compassionate release, finding that the existence of the First Step Act at the time of the plea agreement meant that Egebrecht made a “knowing and voluntary” waiver of her right to file for compassionate release with the court under the waiver’s broad terms. To be considered for compassionate release, Egebrecht must now wait for the BOP to hear her case directly, which could take over six months. And if the BOP rejects her petition, this decision is unreviewable by the district court, in accordance with the broad appellate plea waiver.
Though the judge in Egebrect reviewed a finalized plea agreement, rather than a proposed agreement as in Osorto, the judge could have determined the waiver to be void as a matter of unconscionability and frustration of congressional intent. Judge Breyer’s arguments are still relevant—if the proposed term of a plea agreement is void as a matter of public policy, then the same term in a finalized contract must have the same negative policy implications. That Egebrecht and Osorto were afforded different rights based on their respective judges’ interpretations of public policy demonstrates the need for uniform guidance on this issue. The Egebrecht decision has already formed the basis for justifying the rejection of other defendants’ attempts to receive compassionate release because they signed broad plea waivers after the First Step Act was passed.
Uniform treatment for similar offenses is an overarching goal of the justice system. As Justice Stephen Breyer wrote for the Court in United States v. Booker (2005): “Congress enacted the sentencing statutes in major part to achieve greater uniformity in sentencing . . . [and] authorized the [Sentencing] Commission to promulgate policy statements that would assist sentencing judges in determining whether to reject a plea agreement.” As a sentence-reduction mechanism, compassionate release processes must also be uniform if the justice system is to promote fair treatment.
Procedurally, uniform availability of the right to move for compassionate release before a district court is best achieved by prohibiting such waivers from becoming part of plea agreements in the first place. Judge Breyer’s unconscionability arguments are much less likely to succeed in voiding finalized contracts than they are in supporting a judge’s authority to reject a plea agreement. This is illustrated by the different rights afforded to the defendants in Osorto and Egebrecht.
These stages are governed by separate bodies of law. Rule 11 of the Federal Rules of Criminal Procedure gives district court judges broad discretion in determining whether to reject an initial plea agreement. A judge has greater flexibility to provide reasons for rejection at this stage, and prosecutors can freely negotiate alternate terms to those that have been rejected. Interpretation of finalized plea agreements, on the other hand, is governed by case law. Guidance from the Department of Justice advises courts against interpretations of broad waivers that may result in a “miscarriage of justice.” But practically speaking—as the Justice Department acknowledges—the majority of waiver provisions are still valid as long as the defendant “knowingly and voluntarily” agreed to the terms.
An attempt to use the unconscionability doctrine to void the specific waiver of the right to file for compassionate release with courts is further unlikely to succeed because the doctrine challenges the validity of broad plea waivers altogether. All plea agreements inherently waive constitutional rights: A defendant who signs a plea agreement has waived his Fifth Amendment right against self-incrimination and his Sixth Amendment right to a jury trial. It is a logical extension that a defendant is legally able to waive additional constitutional and statutory rights to secure a better bargain. If a district court judge finds that it is unconscionable to waive the right to file for compassionate release, then where should that judge draw the line with respect to other rights freely waived? A judge at the interpretation stage would be unlikely to void such a waiver on the basis of public policy without authority from Congress or the Justice Department.
Formally, Congress’s desire to increase accessibility of compassionate release can only be promoted by legislative or executive action against waiver of this right to file directly with the district court. As then-professor Stephanos Bibas observed in an article in the Harvard Law Review, and as Justice Breyer further noted in his Booker opinion, plea bargains are underregulated by the courts. Judge Breyer’s decision in Osorto to reject the waiver of First Step Act rights was notable, both for its novel legal question and for his willingness to intervene in a plea agreement more generally. Bibas further explained that a defendant’s ability to secure a fair plea agreement largely depends on the ability of his likely underfunded defense attorney to expend resources and on the disposition of the opposing prosecutor. Both parties have incentives to minimize administrative costs of the plea-bargaining process, and the terms of these agreements are often formed without court intervention. Higher-level guidance may therefore be a necessary intervention to preserve First Step Act compassionate release rights, as there are otherwise few incentives to do so in the current system.
There is precedent for prohibiting waiver of certain rights in plea agreements. This includes the right to appeal a sentence that exceeds the statutory maximum and the right to appeal a sentence that was determined on the basis of race. As scholars have noted, however, there is little doctrinal support for giving some rights greater protection from waiver than others. Why should the right to compassionate release be protected when defendants regularly waive their constitutional right to a trial by jury? Executive and legislative intervention is ultimately necessary then to avoid the arbitrary and inconsistent judicial line-drawing that would otherwise occur should the courts independently decide this issue. If the government truly intends to expand the availability of compassionate release, the legislative or executive branch must act.
Such administrative declarations are not unprecedented. In October 2014, Deputy Attorney General James Cole issued a memorandum to federal prosecutors declaring that plea agreements should no longer require defendants to waive future claims of ineffective assistance of counsel. As the Deputy Attorney General wrote, these types of waivers are legal under most case law, but the nonuniformity of application across jurisdictions necessitated an executive branch declaration.
Critics of the proposal for uniform access to compassionate release may argue that freedom of contract should always allow defendants to trade their ability to appeal in exchange for favorable concessions. The empirical data provides a mixed picture on whether defendants are truly afforded a benefit of the bargain. On one hand, the government gives more favorable terms to defendants who agree to a narrow waiver of their rights of appeal than to those who do not. On the other hand, there is no pattern of favorable terms given to defendants who sign plea agreements with broad appellate waivers. Signing a broad waiver, rather than a narrow one, provides no apparent marginal benefits. Further complicating the freedom of contract argument is that enforcement of waivers is empirically inconsistent across jurisdictions. Defendants are signing agreements that risk enforcement of an appellate waiver, rather than unequivocally giving up this right.
It is difficult to know the gravity of the issue of compassionate release waivers. Judge Breyer is the first judge to have made a public statement rejecting the unconscionability of such waivers, and no subsequent opinions have highlighted the issue specifically with regard to compassionate release. Prosecutors in other jurisdictions are either carving out exceptions for defendants to preserve their First Step Act petition rights, or the issue is an unnoticed part of routine, broad appellate plea waivers. There will be no evidence of the latter until defendants who have signed such plea agreements face their own medical or familial circumstances warranting petitions for release. But given that over 90% of all federal criminal cases end in a plea agreement, and almost 67% of these agreements include some form of appellate waiver, it is almost certain that this issue will persist. And without executive or legislative intervention, uniform sentence reductions cannot be achieved.
- 1As of April 30, 2021, the BOP reports that 3,294 inmates have been granted compassionate release since the First Step Act was passed. Since the BOP reported 145 inmates were granted release in 2019, an estimated 3,149 inmates have been granted release in 2020 and so far in 2021.