I.  The Sentencing Reform Act

Before exploring the special challenges that supervised release poses to the application of § 751(a), it is necessary first to provide some background on why Congress created supervised release in the first place. This Part places supervised release in its appropriate context as one part of a wider sentencing reform scheme. It also outlines some of the specific issues that supervised release was intended to resolve.

A.    The Creation of Supervised Release

Congress passed the SRA in 1984 to enact widespread reforms to the criminal-justice system. Before the SRA, federal sentencing consisted of an indeterminate system in which power was principally divided among three authorities.15 First, Congress possessed broad authority to set sentencing ranges for specific federal crimes, typically in the form of maximum sentences.16 Second, sentencing judges had broad discretion to choose a sentence within the statutory range, based on their individualized assessments of the facts of the crime and the defendant’s personal characteristics.17 Third, the United States Parole Commission controlled sentences from the back end by selecting eligible prisoners for release, conditioned on their continued fulfillment of certain requirements.18 Theoretically, the discretion vested in parole authorities to replace incarceration with conditional release acted as an administrative check on the discretion of sentencing judges to impose prison sentences.19

This indeterminate sentencing regime fell into disrepute throughout the 1960s and 1970s.20 Commentators criticized the system for producing disparate sentences between similar defendants.21 Importantly, desire for sentencing reform was driven by dissatisfaction with the perceived arbitrariness of both sentencing judges and the Parole Commission.22 Theoretically, parole promoted rehabilitation by allowing authorities to release prisoners who made sufficient progress while incarcerated, thus incentivizing prisoners to participate in vocational and educational programs.23 By the 1980s, Congress believed that this model was a failure.24 Parole merely perpetuated sentencing disparities by conditioning release upon rehabilitation—an amorphous concept.25 The Parole Commission applied agency guidelines inconsistently, and it appeared to be unable to rationally determine when offenders were rehabilitated and therefore suited for release.26

In addition, rather than acting as a check on sentencing discretion, the availability of parole gave district court judges perverse incentives. To retain control over sentencing, judges would occasionally oversentence defendants, such that the earliest parole date would coincide with the sentence length the judge believed the defendant deserved.27 For example, a judge who believed that a defendant deserved five years in prison could exert some level of control over his release date by sentencing the defendant to a fifteen-year term without early parole eligibility. Without early release, the prisoner would become parole eligible by statute after serving one-third of his sentence.28 But this introduced a level of uncertainty to the system, as the defendant would receive the judge’s intended sentence only if the Parole Commission’s assessment of the prisoner, as well as the prisoner’s conduct while incarcerated, perfectly aligned with the sentencing judge’s prediction.

Congress passed the SRA to shift sentencing to a more rule-like and uniform regime, thus reducing perceived sentencing disparities and promoting transparency in sentencing.29 Although the Sentencing Guidelines are the most studied manifestation of this new sentencing philosophy,30 the abolition of federal parole was also a critical element of this shift. The SRA prospectively abolished parole for federal prisoners sentenced for offenses committed on or after November 1, 1987.31 Only federal prisoners sentenced prior to this date remain eligible for parole.

Congress, however, did not entirely abandon the goal of postincarceration supervision. In fact, Congress observed an anomaly in the parole system: assuming that parole truly facilitated reentry, those most in need of monitoring would not receive it.32 Specifically, better-behaved inmates would be granted parole, receiving supervision in lieu of their remaining prison time, even if they did not actually need assistance.33 Conversely, inmates who behaved poorly would not be paroled. These inmates would serve longer prison terms but would not receive any postincarceration monitoring.34

Congress created supervised release to ameliorate these problems. Supervised release would “ease the defendant’s transition into the community after the service of a long prison term” and “provide rehabilitation to a defendant who has spent a fairly short period in prison . . . but still needs supervision and training programs after release.”35 Unlike parole, supervised release would be entirely independent of the preceding prison term, helping defendants acclimate to life outside of prison without simultaneously facilitating arbitrariness in sentencing.36

B.    The Structure of Supervised Release

Supervised release is a form of postconfinement monitoring. After a defendant is sentenced to a prison term, the sentencing court may, and under certain circumstances must,37 require a term of supervised release that begins after the prison term is completed.38 Similar to probationers, individuals on supervised release are monitored by the US Probation Office.39 Releasees and probationers are subject to similar conditions.40

Supervised release differs structurally from other conditional release systems, such as parole (which is still common at the state level) and probation. Most importantly, supervised release is not an alternative to incarceration, but a separate and additional period of monitoring concerned with facilitating the reintegration of the defendant into the community. By contrast, a parolee agrees to abide by a set of conditions in exchange for release from prison. And similarly, a judge may sentence a defendant to either a probation term or a prison term, but not both.41 Supervised release must follow imprisonment; it cannot be imposed on its own.42 These differences reflect the rehabilitative character of release by fully disaggregating the punitive and transitional phases of a defendant’s sentence.43

This distinction also has important legal consequences. For example, circuits uniformly hold that supervised release terms may extend a defendant’s total time under criminal-justice supervision beyond the statutory maximum number of years permitted for imprisonment. Because release terms are independent monitoring periods rather than elements of the direct sentence, they are not subject to these statutory caps.44

These structural features may also be understood in part as rectifying procedural deficiencies that Congress identified in the parole system. Congress previously criticized the then-existing sentencing framework for permitting judges to play guessing games with parole officials, attempting to calibrate sentences to their predictions of when the prisoners would be granted parole.45 By contrast, district court judges impose supervised release terms during the initial sentencing, thus consolidating sentencing authority in the district court judge and limiting the number of individuals with discretionary control over a defendant’s sentence.

Further, supervised release is generally restricted to fixed terms not to exceed one, three, or five years, depending on the severity of the offense.46 Under the parole system, defendants were supervised for whatever length of time they had remaining on the balance of their sentence. Supervised release was intended to create a more transparent system, with each defendant receiving a fixed term of supervision at the outset of his sentence rather than through later administrative hearings.47

Defendants are required to abide by a set of conditions while on release. Some of these conditions are statutorily mandated and universal for all defendants.48 For example, all individuals on supervised release must refrain from committing further crimes and must not unlawfully possess any controlled substances.49 In addition, district courts have significant discretion to create individualized conditions, so long as they (1) are “reasonably related” to a list of sentencing factors; (2) “involve[ ] no greater deprivation of liberty than is reasonably necessary”; and (3) are “consistent with [ ] pertinent policy statements issued by the Sentencing Commission.”50 The district court may modify or eliminate release conditions throughout the term51 and may terminate the term early for good behavior.52

Several developments have called the nonpunitive nature of supervised release into question. First, Congress expanded the list of factors that judges must look at when making these decisions. These factors are drawn from the wider list of considerations that judges must look at when choosing a prison sentence.53 Originally, Congress excluded both punishment and incapacitation from the list, in line with the distinct aims of release.54 In 1987, Congress added incapacitation to the list, directing judges to consider the need “to protect the public from further crimes of the defendant” in creating release terms.55 This revision undermines the transitional aims of supervised release by permitting judges to focus on continued seclusion rather than defendants’ reintroduction into the community. Today, judges may consider incapacitation not only in choosing the appropriate prison sentence length, but also in defining the terms of the defendant’s release.

Second, Congress added specific procedures courts could use to revoke release.56 Even without revocation, judges have a great deal of flexibility in dealing with violations of supervised release conditions. If a defendant violates a term of supervised release, the sentencing judge may continue the term without alteration, impose additional conditions, or extend the term of supervised release.57 Initially, Congress did not believe that minor violations of release conditions should result in a resentencing of the defendant and thus declined to create a revocation process.58 Congress envisioned that district courts would instead use contempt proceedings to redress repeated or serious violations.59 If the defendant committed a new offense while on release, it was assumed that either a prosecution for that offense or a court order for contempt would sufficiently redress the violation.60 For technical violations, the judge could make release conditions more severe, rather than returning the defendant to prison.61

Today, in addition to sustaining, extending, or adding conditions to the original release term, judges may also revoke release entirely and sentence the defendant to imprisonment for all or part of the originally authorized supervised release term without credit for time previously served on release.62 Many commentators argue that the inclusion of revocation as a penalty reintroduced punishment into release’s aims, turning supervised release from a rehabilitative measure to a punitive form of conditional release more akin to parole.63 Many of these additional features appear to fly in the face of Congress’s initial aim of creating a system in which “a prisoner has completed his prison term when released even if he is released to serve a term of supervised release.”64 Instead of fixed prison terms followed by fixed periods of monitoring, revocation creates possibilities for further imprisonment at indeterminate points in the future. Although revocation may serve rehabilitative purposes by removing defendants from release who are likely to reoffend, revocation also allows defendants to be reimprisoned for violating conditions that proscribe otherwise-legal behaviors. Stated otherwise, revocation permits offenders to be imprisoned for failing to progress, eroding the barrier between the rehabilitative and punitive phases of a sentence.65

Importantly, while supervised release may constitute punishment as a descriptive matter, judges still consider supervised release to serve predominately rehabilitative goals.66 It is for this reason that courts have held that sentencing defendants to lifetime terms of supervised release does not violate the Eighth Amendment.67 The Second Circuit has gone a step further, observing in dicta that “[t]here is no constitutional limit on how long a supervised release term can be.68 These decisions are premised on the idea that the stakes are lower for a defendant who is subject to rehabilitative monitoring, and that supervised release should therefore be subject to reduced constitutional scrutiny.69 It is important to keep these decisions in mind when examining the penalties that defendants may permissibly incur for violating the terms of their release. If the penalties for condition violations are severe, then the stakes are in reality quite high for defendants on release, given that they may be subject to lifetime monitoring.

C.    The Uncertain Custodial Status of Supervised Release

Given the quasi-rehabilitative, quasi-punitive, and occasionally quasi-incapacitative character of supervised release, should persons on release be considered to be in federal custody? This question is quite important. A defendant’s custodial status may determine when his criminal sentence begins,70 how long his sentence lasts,71 whether various sentencing enhancements apply to him,72 and whether he may seek habeas relief.73

Similarly, a defendant’s custodial status determines whether he may be charged with escape. 18 USC § 751(a) codifies the common-law crime of escape for federal offenders. To convict a defendant, prosecutors must prove that the defendant absconded from (1) “the custody of the Attorney General or his authorized representative”; (2) “any institution or facility in which [the defendant] is confined by direction of the Attorney General”; (3) “any custody under or by virtue of any process issued under the laws of the United States by any court, judge, or magistrate judge”; or (4) the custody of a federal officer pursuant to a lawful arrest.74 Defendants on supervised release are prosecuted under category three. Strikingly, the statute defines neither “escape” nor “custody.”

Both the structure and the function of supervised release make it difficult to determine whether persons on supervised release are in custody even if one puts these statutory ambiguities to one side and proceeds from a commonsense understanding of the term. By way of example, it is intuitively easy to accept that all individuals who are currently incarcerated for committing federal crimes are in federal custody. These persons are subject to physical confinement, experience constraints on their activities, and have easily identifiable custodians. The overarching purpose of their confinement is punitive, which seems relevant as it suggests that these persons may not leave the boundaries of their confinement without facing further sanctions. Most importantly, 18 USC § 3621 clarifies that persons sentenced to prison terms are committed to the custody of the Bureau of Prisons until the expiration of that term.75

It is difficult to make the same leap for persons on supervised release, who are required to abide by a set of conditions but are not under the same extreme level of direct control. Furthermore, as discussed above, release terms are intended to form a rehabilitative phase that is separate and distinct from imprisonment. Persons on release have already completed a preceding prison term for an offense. The use of the term “release” ordinarily connotes some degree of freedom from restraint.76 Finally, there is no statutory provision clarifying whether supervised release is custodial.

To further complicate matters, release terms may take a variety of different forms, some of which resemble direct incarceration more than others. Before discussing how courts have grappled with the question whether release conditions are custodial, it is useful to first briefly outline the different forms of release.

At one end of the spectrum are forms of supervised release that replicate physical confinement. It is important to note that these conditions have historically been subject to special statutory limitations.77 For example, the SRA restricts judges’ ability to prescribe intermittent confinement (sentences that allow defendants to remain at liberty but require them to return to the custody of the Bureau of Prisons during nights, weekends, or other intervals of time).78 Intermittent confinement is not available as a general condition of supervised release; a judge may impose this sentence only to redress a condition violation.79

Other forms of release conditions approximate, but do not replicate, forms of physical confinement. These quasi-detentive conditions often restrict an individual’s freedom of movement or association. For example, courts may impose conditions requiring the defendants to reside at transitional facilities80 (colloquially known as halfway houses), to abide by curfews,81 or to submit to GPS monitoring.82 Many standard conditions of release fall into this category.83

Finally, some forms of supervised release do not approximate physical confinement at all. For example, district court judges have ordered defendants to use contraception,84 to use only their “true legal name,”85 or to refrain from driving for the duration of the release term.86 Many release conditions target behaviors that are specifically related to an individual defendant’s crimes, but there is no requirement that release conditions target only illegal behaviors. A judge may, for example, impose a condition requiring abstinence from alcohol for a defendant who used the proceeds from thefts to finance an alcohol addiction.87 Other conditions are designed to increase social productivity; for example, judges may require releasees to pay child support or seek mental health treatment.88

These diverse forms of release underscore the difficulty of determining when supervised release is custodial under § 751(a). They suggest that supervised release may look very different for different defendants. Some states with programs similar to federal supervised release have dealt with this problem by amending their escape statutes to cover such conduct. The Maine escape statute, for example, explicitly provides that a person who “intentionally violates a curfew, residence, time or travel restriction” may be charged with “escape from supervised community confinement.”89 The federal escape statute, by contrast, has not been meaningfully amended since 1948—leaving it to the courts to determine whether and when release terms are custodial under § 751(a).

II.  The Halfway House Cases

A recent series of cases all share the same basic fact pattern: a defendant completes a prison term, begins a term of supervised release requiring residence at a halfway house, leaves the facility, and is charged with escape from federal custody. These halfway house cases are the first cases in which circuits have addressed whether one can escape from supervised release.

The halfway house cases are important in that they require courts to (1) determine whether a defendant who resides at a reentry facility as a condition of supervised release is in custody for purposes of § 751(a), and, in the process, (2) articulate a workable principle for determining whether other individuals on supervised release should be considered to be in custody. This issue is settled in the context of direct imprisonment. Courts uniformly hold that incarcerated persons who are transferred90 from a penal institution to a halfway house to finish out the remainder of their sentences are in custody.91 But circuits disagree about whether the differences between supervised release and traditional incarceration militate in favor of a different rule.

Part II.A analyzes halfway house cases in the Second, Eighth, Ninth, and Tenth Circuits and traces the three rules created by the courts for determining whether an individual is in § 751(a) custody. Part II.B critiques the circuits’ approaches.

A.    Interpreting § 751(a): The Circuit Split

The Second, Eighth, and Tenth Circuits hold that a defendant who is released to a halfway house as a condition of supervised release is in custody for purposes of the escape statute. These courts rely on similarly expansive constructions of “custody” in reaching this holding. These broad constructions of § 751(a) may have interesting implications for other cases: If a defendant removes a GPS device, violates a nighttime curfew, or fails to comply with reporting requirements, has he absconded from custody within the meaning of § 751(a)?

In contrast to these circuits, the Ninth Circuit has advanced a narrower interpretation of the escape statute. The court suggests that custody is limited to those release conditions approximating traditional incarceration, but it gestures toward a rule that supervised release conditions are categorically noncustodial.

1.   The Second and Eighth Circuits: Custody is established by restraints on activities.

The Second and Eighth Circuits endorse a “restraint on activities” rule for determining custody. Under this rule, an individual is within federal custody when he is subjected to limits on his activities. The Second Circuit advanced this rule in United States v Edelman.92 Jody Edelman was convicted of a drug offense and sentenced to seven years of imprisonment followed by three years of supervised release.93 After repeated violations, the sentencing judge modified Edelman’s release to include residence in a halfway house for five months.94 Edelman left the halfway house before the end of his term and committed a series of drug offenses.95 He was arrested and charged with escape for his departure.96 Edelman appealed, arguing that his halfway house stay did not constitute custody under § 751(a).97

The Second Circuit disagreed with Edelman’s argument. The court observed that “[c]ustody may be minimal, and indeed, may be constructive.”98 The court found textual support for this reading, noting that § 751(a) refers broadly to “any custody under or by virtue of any process,” supporting an expansive reading.99 The court further observed that custody requires only “some restraint upon [an individual’s] complete freedom.”100 Conceding that Edelman’s residence at the halfway house differed substantially from traditional forms of incarceration, the court nonetheless noted that the release conditions still “subjected [the defendant] to a restraint on his activities and limited the amount of time he could spend out of the facility.”101 Residents at Edelman’s halfway house were not permitted to leave without notifying staff beforehand, were required to sign out separately before each activity, and could leave the facility for only up to twelve hours at a time.102 The Second Circuit concluded that these restraints established § 751(a) custody.103

In United States v Goad,104 the Eighth Circuit, facing similar facts, arrived at the same conclusion. According to the court, it was sufficient to sustain an escape indictment that “(1) Goad was court-ordered to reside for 120 days at the Hinzman Center and obey its rules; (2) the Hinzman Center prohibited departures without authorization; and (3) Goad left early without obtaining authorization and failed to return.”105

The defendant in Goad attempted to contrast the state of being in the attorney general’s custody with the state of being released to the supervision of a federal probation officer.106 This argument is premised on the idea that custody is relational—that custody cannot exist in a vacuum, but denotes a specific relationship between a custodian and the object of supervision. A relational understanding of custody undermines the restraints-on-activities approach by requiring courts to identify a custodian rather than simply asking about what types of activities the defendant may permissibly engage in. Various provisions of § 751(a) appear to identify a particular guardian that is the source of the custodial relationship, as they criminalize escape from the custody of the attorney general, his authorized representative, and federal officers conducting lawful arrests.107 These clauses do not identify the agency that oversees defendants on probation and supervised release as a potential custodian, even though probation existed at the federal level when the escape statute was initially passed, and probation conditions may also include forms of confinement.108 The omission of the US Probation Office is particularly surprising given that probation, unlike supervised release, is a criminal sentence in its own right that is imposed as a substitute for imprisonment.109

The Goad court quickly disposed of the relational argument, again pointing to the broad text of the relevant subsection of the escape statute.110 The particular clause used to prosecute escapes from supervised release refers to “any custody under or by virtue of any process issued under the laws of the United States.”111 Because this subsection does not identify a particular custodian, the Goad court reasoned that the defendant’s arguments were irrelevant.112 The § 751(a) requirement was met when the district court sentenced Goad to serve out his release term at a halfway house.113

2.   The Tenth Circuit: Custody is established by the right to control an individual.

The Tenth Circuit, in accord with the Second and Eighth Circuits, held in United States v Foster114 that defendants ordered to reside in halfway houses pursuant to supervised release conditions are in custody under § 751(a).115 Cheston Foster was sentenced to time served and thirty months of supervised release as a revocation penalty for violating the conditions of a previous release term.116 During the sentencing phase of the trial, Foster’s supervising officer asked the court to add a halfway house condition to the defendant’s term of release, in light of the defendant’s unstable residence and employment situations.117 Foster left the halfway house and was charged with escape.118 The district court dismissed the indictment, reasoning that the release condition was “a stop-gap measure used to prevent homelessness,” not “a custodial sentence triggering criminal liability.”119

The Tenth Circuit overruled the district court and upheld the indictment. Similar to the Second and Eighth Circuits, the Foster court looked first to the text of the statute. The court reasoned that the plain text of the statute does not distinguish between defendants in nonpunitive versus punitive forms of custody.120 The Foster court agreed with the Government’s interpretation that an individual is in custody if “another person has the legal right to control his actions or limit his freedom.”121 Because the relevant question under this analysis is whether the individual defendant is under “control sufficient to constitute custody,”122 this rule is functionally similar to the restraints-on-activities approach. Under both rules, the existence of restrictions upon the defendant establishes § 751(a) custody.

The Tenth Circuit also adheres to an interpretation that custody “may be minimal or even constructive.”123 This rule was first articulated in United States v Depew,124 a Tenth Circuit case dealing with the issue of whether a defendant was in federal or state custody during a transfer on a writ of habeas corpus ad prosequendum.125 The handcuffed defendant attempted to escape from the supervision of a county sheriff while being driven between facilities.126 The court suggested that the defendant’s status as “a federal prisoner who was due to be returned to serve time in a federal penitentiary,” as well as his being “in the direct physical presence [of] and contact with a federal law enforcement officer,” indicated that he was in federal § 751(a) custody.127

Although Depew is cited for the proposition that custody may be minimal or constructive,128 these specific facts suggest that custody is still linked to physical restraint or the prospect of confinement. Further, the cases that the Depew court offered as examples of the “wide range of circumstances” in which § 751(a) applies outside of the narrower context of prison escapes involved direct physical restraints and institutional confinement.129 The Foster court, however, denied that either was essential to the court’s conclusion, as “real physical confinement” is not a prerequisite for custody.130 The Foster court instead characterized Depew’s holding more broadly, reading the case as suggesting that a defendant is in custody when “the restrictions of life . . . are sufficiently limiting”—as when a defendant is required to live in a halfway house.131

The Tenth Circuit has also invoked Depew in circumstances bearing even more attenuated relationships to traditional incarceration. In United States v Ko,132 the Tenth Circuit sustained an escape indictment for a defendant subject to electronic monitoring.133 Michael Ko was permitted to leave his home for work, but he was required to return every evening. When he did not return one evening, he was charged with escape.134 The Tenth Circuit rejected Ko’s argument that the plain meaning of custody required physical confinement and could not encompass “leaving one’s own home.”135 The Tenth Circuit cited Depew for the proposition that the defendant’s “constant monitoring, [ ] monitoring bracelet, and spatial and temporal bounds” established custody.136 Although Ko involved a defendant who was transferred to home confinement to finish out a direct prison sentence, the Tenth Circuit has thus far given no indication that the special nature of supervised release would prevent a similar prosecution.137

3.   The Ninth Circuit: Custody is established by conditions equivalent to custodial incarceration.

In United States v Burke,138 the Ninth Circuit adopted a very different test from the other circuits. Similar to Foster, Burke involved a defendant who was “essentially homeless” and was sentenced to a halfway house stay as a condition of supervised release.139 Anthony Burke checked out of the center and did not return.140 The Ninth Circuit affirmed the district court’s dismissal of an ensuing escape indictment, holding that Burke was not in custody under § 751(a), as he was not confined “under conditions equivalent to custodial incarceration.”141

Burke is amenable to two interpretations. The first reading suggests that the court created a rule necessitating case-by-case analysis of whether a defendant on supervised release is subject to conditions equivalent to custodial incarceration.142 The Burke court did not define custodial incarceration, but instead appeared to enforce the rule by evaluating the extent of the defendant’s restraints and comparing them to imprisonment. Under this interpretation, Burke could not be charged with escape because the specific conditions of his supervised release term were too lenient to constitute custody. The Burke court observed that residents at the halfway house were “free to be employed outside the center, and to come and go during the day with permission if they logged in and out.”143 The Second Circuit relied on very similar facts in Edelman to assert that the limits placed on the defendant’s freedom indicated custody.144 The Ninth Circuit instead emphasized the extent of the liberty enjoyed by the defendant while on release.145

It is unclear at what point release conditions become custodial under this rule. The Government argued in Burke that the halfway house placed numerous restrictions on the defendant, such as “a curfew, limited visitors, assigned beds, restricted telephone use,” and requiring residents to notify staff before leaving.146 The halfway house also put residents on an “escape or abscond status” if they failed to return to the center in a timely manner; the employees put Burke on escape status after he was still absent two hours after his estimated time of return.147 In dismissing the indictment, the district court remarked that the halfway house rules were “standard” and reflected a “general understanding” that halfway houses are more lenient than custodial facilities.148 The Ninth Circuit’s application of the law to the facts of this case suggests that the equivalent-to-custodial-incarceration test may form a bright-line rule rather than a functional rule—nothing less than prison conditions may constitute custody.

As such, a second plausible reading of the opinion is that while Burke articulates an equivalent-to-custodial-incarceration test, Burke functionally creates a per se rule. Few forms of supervised release resemble traditional incarceration as closely as halfway house stays do,149 yet Burke suggests that halfway house terms are outside the reach of § 751(a). Further, several elements of the Burke court’s reasoning speak to the structure of supervised release more generally rather than to the specifics of the defendant’s release term. The Ninth Circuit observed that the conditions of Burke’s release “were much more analogous to probation than they were to imprisonment.”150 Importantly, the court highlighted the fact that condition violations may be sanctioned through revocation (as with probation) rather than an independent criminal charge of escape.151 The Burke court also found it relevant that Burke, like all releasees, had completed a preceding prison term and was thus no longer in the custody of the Bureau of Prisons; the court distinguished past precedents on this basis.152

The Burke court did not go so far as to formally create a per se rule. It was not necessary for the Ninth Circuit to determine the custodial status of other forms of supervised release in order to decide Burke. Additionally, it is difficult to reconcile sections of the opinion contrasting the specific conditions of the defendant’s confinement with direct imprisonment with a per se rule holding that supervised release is never custodial. The Burke court highlighted the similarities between supervised release and probation, but the opinion also appears to leave open the possibility that a defendant subject to a more-confining release condition (for example, a nonstandard halfway house) may be in § 751(a) custody.

At the same time, Burke nevertheless plants the seeds for a per se rule by establishing a high bar for custody at the outset. By taking a highly confining type of supervision relative to other forms of supervised release and declaring it to be insufficiently confining to generate § 751(a) custody, the Burke rule suggests that few, if any, release conditions are custodial.

B.    Critiquing the Circuits’ Approaches

There are three weaknesses in the circuits’ approaches that merit special attention. First, the reach of these holdings is unclear. Among the Second, Eighth, and Tenth Circuits, no circuit found the presence of physical or institutional confinement—in these cases, halfway houses—particularly relevant in determining whether the defendant was in custody. The Tenth Circuit directly denied the relevance of confinement, stating that “real physical confinement” is not a requirement for escape.153 Instead, the circuits created rules focusing on the presence of legal rights to control behavior or restraints on the defendant’s activities. These abstract rules suggest that many of the forms of supervised release detailed in Part I.C may be considered custodial under § 751(a), in the absence of some limitation.

Although the boundaries of these rules are unclear, there may be an implicit limiting factor within the escape statute itself. The statute requires prosecutors to prove that the defendant has “escape[d] or attempt[ed] to escape” as an element of the crime.154 Not all condition violations may be characterized as escapes. A defendant who is banned from driving as a condition of supervised release may be in custody under the restraints-on-activities rule, but if she violated the terms of her release by driving, it would be difficult to construe that violation as an escape. This limiting principle becomes more complicated when the release conditions involve location-based restrictions on movement and association. As mentioned above, the Tenth Circuit suggested in Ko that failing to return to one’s home at a specific time may still constitute an escape from federal custody. The escape element has thus far failed to provide a meaningful restriction on § 751(a) liability.

This lack of clear boundaries is problematic for three reasons. First, it makes it difficult to predict the consequences for violating release conditions. Second, it drastically expands the scope of criminal conduct by adding an overlapping criminal offense to a broad spectrum of release violations, even when those release conditions could not independently give rise to criminal liability. Even more importantly, absent a meaningful limiting factor, the circuits’ rule verges on the absurd by encompassing a broad category of individuals. Well-settled principles of statutory interpretation generally counsel against such broad and unbounded readings of criminal statutes.155 The defendants in all four cases discussed above raised arguments that the rule of lenity obligated courts to resolve ambiguities over the construction of § 751(a) in their favor.156

Third, the Second, Eighth, and Tenth Circuits’ opinions import reasoning from parallel, but inapplicable, Supreme Court jurisprudence. Although the Supreme Court has not defined § 751(a) custody, the Court has interpreted “custody” in the context of 28 USC § 2241(c)(3) habeas proceedings. Section 2241(c)(3) grants federal courts jurisdiction to hear habeas petitions from persons who are “in custody in violation of the Constitution or laws or treaties of the United States.”157 In Jones v Cunningham,158 the Court addressed the question whether a state prisoner on parole was still in custody within the meaning of the statute, such that a district court could entertain a claim that his state sentence was unconstitutional.159

The Court unanimously concluded that the defendant was in custody.160 The Court defined custody as the condition resulting from the imposition of “significant[ ] restrain[ts]” on individual liberty,161 stretching beyond “actual, physical custody.”162 The Court further observed that the defendant was confined by parole order “to keep good company and good hours, work regularly, keep away from undesirable places, and live a clean, honest, and temperate life,” and that even “a single deviation, however slight,” could result in a return to prison.163 It was not relevant to the Court’s analysis that the conditions were part of a rehabilitative process; it mattered only that the parole order prevented the petitioner from participating in activities “which in this country free men are entitled to do.”164

At first glance, Jones seems particularly instructive in that it involved a defendant on conditional release (parole) and issued an expansive reading of custody that could include forms of supervision falling short of direct imprisonment. Although the Second, Eighth, and Tenth Circuits do not explicitly cite Jones, it seems far from coincidental that their respective opinions mirror the Jones Court’s analysis. Specifically, just as Jones suggests that custody is created by restraints, regardless of whether those restraints are rehabilitative, and that custody does not require direct physical confinement for purposes of § 2241(c)(3), each circuit endorsed a similarly expansive reading of custody for purposes of § 751(a).

It is not clear that § 2241(c)(3) custody and § 751(a) custody must share the same meaning as a matter of statutory interpretation. The same word may have different meanings in different statutes.165 Words often derive precise meaning from their statutory context. In Wachovia Bank, National Association v Schmidt,166 for example, the Supreme Court declined to give the word “located” the same meaning in statutes pertaining to venue and to subject matter jurisdiction for national banks.167 Although lower courts had interpreted “located” uniformly, the Court argued that this conflation overlooked the “discrete offices” of venue and jurisdiction.168

Similarly, the arguments advanced by the Supreme Court to support an expansive reading of habeas custody do not apply with equal force to escape custody, given the differing aims of § 2241(c)(3) and § 751(a). The Court has emphasized that habeas is not “a static, narrow, formalistic remedy,”169 but a flexible device capable of “cut[ting] through barriers of form and procedural mazes.”170 A liberal interpretation of the custody requirement allows courts to hear a diverse range of claims and prevents states from evading the writ by declaring serious punishments noncustodial. In line with these aims, courts have read § 2241(c)(3) custody to include restrictions falling well short of confinement.171

The concerns that motivated the Jones Court’s broad construction of § 2241(c)(3) custody are inapplicable in the context of supervised release. While “custody” is used in § 2241(c)(3) to define courts’ habeas jurisdiction and, by extension, the availability of relief, “custody” is used in § 751(a) for a very different purpose—to define the scope of criminal liability for escape. The factors advanced in Jones in support of a flexible and vague definition of custody are not convincing as applied to statutes that define crimes. In this context, concerns regarding fair notice, due process, and the constraint of enforcement discretion are paramount.172 It thus makes little sense to mechanically import § 2241(c)(3) reasoning into the § 751(a) context. Using parallel rules may also lead to absurd results. For example, a defendant ordered to complete community service is in custody for habeas purposes.173 While a defendant who leaves a community service site is certainly disobedient, it seems incorrect to suggest that he has committed an escape, even though the legal rules defining § 2241(c)(3) and § 751(a) custody are virtually identical in some circuits.

Third, all four circuits rely on cases involving similar facts but different forms of conditional release that serve different purposes within the criminal-justice system. The Second and Eighth Circuits cited United States v Rudinsky174 as persuasive authority.175 In Rudinsky, the Sixth Circuit held that a defendant was in custody at a federal treatment center.176 The Tenth Circuit, as discussed above, cited Ko, a case involving a defendant subject to electronic monitoring.177 The defendants in both Rudinsky and Ko were completing direct prison sentences when they absconded. The defendant in Rudinsky was completing a criminal sentence for possessing stolen mail at a treatment center;178 the defendant in Ko was transferred to home confinement by the Bureau of Prisons in order to complete the last four months of a prison sentence for conspiracy to distribute methamphetamines.179 None of these opinions explains why supervised release terms should be treated as equivalent to these forms of supervision, even though there are salient differences between the two.

Further, all four circuits appear to read this legal question as revitalizing an earlier circuit split on whether halfway house stays as a condition of pretrial release could create § 751(a) custody. The three circuits holding that halfway house stays are custodial in the supervised release context all cite United States v Sack,180 a Tenth Circuit decision holding that pretrial halfway house stays are custodial.181 Conversely, the Ninth Circuit in Burke relied in part on United States v Baxley,182 a case from that circuit holding that pretrial halfway house stays are noncustodial.183

Only the Tenth Circuit offered an explanation for why Sack and Baxley might be persuasive. The Foster court suggested that pretrial detention is an appropriate analogue for supervised release, as it “also serves a non-punitive function and does not serve as a substitute for part of a prison sentence.”184 Still, the purposes of pretrial release differ substantially from the reasons for imposing supervised release. Although pretrial release is nonpunitive, it is also nonrehabilitative.185 By conflating supervised release with other forms of criminal-justice supervision, the circuits fail to sufficiently account for how the targeted aims of the SRA may bear on whether supervised release conditions are custodial.


* * *


In summary, the circuits have created unbounded rules for defining § 751(a) custody. Interestingly, none of the four circuits chose to resolve this textual ambiguity by looking at the escape statute itself and asking whether supervised release violations fall within the conduct historically proscribed by the statute. Instead, the circuits covertly imported reasoning from the Supreme Court’s habeas jurisprudence and looked to precedents involving different forms of criminal-justice supervision—thus failing to take into account the importance of supervised release as a new form of criminal-justice supervision and a critical component of sentencing reform. Part III of this Comment clarifies this area of the law by filling the gaps the circuits have left and interpreting the applicability of escape to supervised release conditions in light of both the SRA and the history of § 751(a).

III.  Supervised Release and § 751(a) Custody: The Categorical Rule

This Part argues in favor of a rule that supervised release conditions are categorically noncustodial under the escape statute. Although it may seem uncontroversial to prosecute individuals who abscond from facilities or violate court-ordered conditions, both the SRA and the history of the escape statute strongly suggest that supervised release lies outside the ambit of § 751(a). In addition, the rule is normatively desirable, as it prevents condition violations from unduly extending criminal sentences and trapping releasees in a cycle of punishment. The rule is also administratively desirable, as it prevents the need for case-by-case inquiries into whether each defendant’s form of release is custodial. Such inquiries are likely to lead to inconsistent application of the escape statute across jurisdictions. A categorical rule best gives effect to the SRA by emphasizing the barrier between retributive concerns, which must be addressed during the selection of an appropriate prison sentence, and rehabilitative concerns, which are addressed during the selection of the term of supervised release.

Part III.A traces the history of the federal escape statute. While there is no elegant definition of § 751(a) custody, several statutory developments suggest a narrow understanding of the term. Further, the statute has historically been used to rectify specific problems that do not implicate supervised release. Part III.B argues that escape prosecutions subvert the rehabilitative and discretion-centralization purposes of the SRA by adding a new layer to the penalty structure that Congress created to redress condition violations. If an individual violates a release condition, that violation can be sufficiently redressed through these preexisting mechanisms; there is no need to add additional criminal penalties in the absence of independently criminal conduct.

A.    The Categorical Rule: The Federal Escape Statute

While the circuits purport to derive their rules from the text of § 751(a), it is virtually impossible to resolve this split on a plain-meaning approach alone. Dictionary entries for “custody” contain both broad definitions that would insufficiently delineate the bounds of criminal conduct and narrow definitions that do not settle the question. Black’s Law Dictionary defines “custody” as “[t]he care and control of a thing or person for inspection, preservation, or security,” or alternatively, “[t]he detention of a person by virtue of lawful process or authority.”186 The first definition contains six subentries with qualified versions of the term (“penal custody,” “protective custody,” etc.).187 Webster’s Third similarly defines “custody” as “the act or duty of guarding and preserving,” or alternatively, “imprisonment or durance of persons or charge of things.”188 Neither source obviously compels any one of the circuits’ rules.

Courts must look beyond plain language when meaning is ambiguous or a literal interpretation would lead to an absurd result.189 Looking to the history and purposes of § 751(a) counsels against an expansive construction.

1.   The legislative and enforcement histories of § 751(a) suggest a narrow construction of “custody.”

The Second, Eighth, and Tenth Circuits quote the statutory language directly, emphasizing that § 751(a) criminalizes escape from “any custody under or by virtue of any process.”190 While the text suggests an expansive reading of “custody” at first glance, looking to the history of the statute demonstrates the extent to which broad constructions of the clause are a relatively novel phenomenon.

Congress first codified the common-law crime of escape in 1930.191 The statute was enacted as part of a program to improve the federal penal system.192 The original statute criminalized only (1) escapes from the custody of the attorney general or his authorized representative and (2) escapes from penal or correctional institutions.193 Five years later, Congress amended § 751(a) at the attorney general’s request.194 Concerned that the existing statute did not capture escapes by prisoners held prior to conviction, the attorney general sent letters to the chairmen of the House and Senate Committees on the Judiciary.195 Each letter contained a draft proposal and a list of similar state statutes proscribing escapes by inmates held in jails (as opposed to penal institutions) while awaiting trial.196

The ensuing amendment adopted the attorney general’s suggestions almost verbatim, creating two new custodial categories in addition to the two existing categories. The newly created category three relates to custody “by virtue of any process issued under the laws of the United States by any court, judge, or commissioner,” and category four accounts for “custody of [federal] officer[s] . . . pursuant to lawful arrest[s].”197 The facts that the language of the amendment mirrors that of the letter and that both of the relevant congressional reports insert the letter directly into the legislative history198 suggest that the addition of the seemingly most expansive category was intended to capture a discrete group of individuals—federal prisoners awaiting trial.

One might argue that the broad language of the amendment vitiates the significance of the attorney general’s intent. This argument would suggest that whatever his expressed intentions were, the language of the “by virtue of any process” clause (category three) sweeps broadly enough to encompass court-ordered release conditions. As the circuits have argued in the halfway house cases, the clause seems to potentially encompass many forms of criminal-justice supervision.

While the attorney general’s narrow objective is not dispositive, the history behind the 1935 amendment sheds light on the meaning of “custody.” The breadth of the “by virtue” clause is still entirely a function of the breadth of “custody,” and if the term “custody” is read narrowly, the discontinuity between the attorney general’s expressed aim (to expand § 751(a) to cover inmates who abscond from court-ordered confinement prior to trial) and the language of the “by virtue” clause disappears. Conversely, an expansive construction of category three renders category two (escapes from penal institutions) entirely superfluous. All escapes from penal institutions would be subsumed under a broad reading of three, running afoul of the interpretative canon against surplusage.199

Three developments further support a narrow construction. First, § 751(a) has never been used to prosecute violations of court-ordered parole conditions.200 Parole, which existed at the time of the amendment, may include location-based conditions and would plausibly fall within the language of the “by virtue” clause under a broad construction of custody.

This lack of enforcement history may be motivated by more than prudential concerns—at least one court has suggested in dicta that parole violations simply do not fall under the statute, observing: “[I]t has never been held, and it is not now contended, that a parole violator has escaped from custody for purposes of an escape statute.”201 This absence of case law for parole violators is significant with respect to the circuit split at issue in the halfway house cases, as parole is the direct predecessor of supervised release.202

Second, the circuits’ reliance on the statute’s reference to any custody as textual support is misguided. Although the language of category three was changed from custody “by virtue of any process” to its present form, “any custody under or by virtue of any process,”203 this appears to have been a stylistic emendation rather than a substantive edit intended to signal an expansive understanding of the term.

The 1948 act that amended the language was part of an extensive recodification of the criminal statutes intended to clean up language and eliminate redundancies in the criminal code while avoiding dramatic changes to the underlying substantive content.204 Courts have treated these edits carefully, noting that while the revisions do not require “slavish adherence to the predecessor statutes,” courts “must not attribute to the 1948 [edits] any substantial disruption in prior congressional purpose or policy without first discerning clear evidence that such a departure was intended.”205 Indeed, the relevant legislative report disavows any major alteration to § 751, describing the revisions as changes in “phraseology and arrangement.”206

Third, Congress’s handling of situations involving prisoners on furlough and work release also demonstrates a limited understanding of custody for the purposes of escape.207 If a prisoner is released to attend a funeral, but fails to return to prison for reincarceration, may he be prosecuted under the federal escape statute? If custody depends on either a restraints-on-activities or a right-to-control rule, failing to return may be prosecuted under § 751(a).

Yet in 1965, Congress passed a separate act to criminalize “[t]he willful failure of a prisoner to remain within the extended limits of his confinement, or to return within the time prescribed to an institution or facility designated by the Attorney General,” noting that such acts should be “deemed an escape.”208 The statute does not provide for any enhanced penalties that might distinguish it from an ordinary § 751(a) violation, noting only that a violation is “punishable as provided in” the escape section of the US Code.209 The fact that Congress enacted a separate statute to constructively deem such acts escapes may reflect a legislative endorsement of the view that § 751(a) custody is ordinarily linked to confinement and restraint, rather than an abstract legal right to enforce compliance. In line with this understanding, the failure-to-return statute explicitly applies to prisoners—excluding parolees and probationers from comparable liability.

2.   The purposes of § 751(a) are not served by prosecutions for supervised release violations.

The above analysis demonstrates why the circuits’ broad constructions of § 751(a) custody are inconsistent with a historical understanding of the statute, but it does not fully explain why release violations should be excluded from the statute. Given that § 751(a) predates the creation of supervised release by fifty-four years, it is important to examine the purposes underlying § 751(a) in assessing its applicability to release. Moreover, while the Supreme Court has not offered an interpretation of “custody” for purposes of § 751(a), the Court has defined the term in several other contexts with special reference to the broader purposes of the relevant provision, eschewing formal definitions for functional rules.210

In an early case, the Supreme Court observed that the escape statute is intended to reduce special risks posed to the public by a “violent” and “menacing” escape.211 The Court adopted the Government’s reasoning that escapes and attempted escapes present problems of “penal discipline,” as they endanger “the lives of guards and custodians,” and “carry in their wake other crimes attendant upon procuring money, weapons and transportation and upon resisting recapture.”212 Under this interpretation, an escaped prisoner commits an independent wrong by creating a dangerous environment. This wrong is punishable as an independent criminal offense. This reading of § 751(a) has had lasting legal effect. More recently, the Supreme Court has used this same logic to hold that escape “is a continuing offense . . . [g]iven the continuing threat to society posed by an escaped prisoner.”213         

While these concerns are certainly relevant in the context of supervised release, the harms prevented are occasionally more speculative. The “menaces” posed by a defendant who walks away from a halfway house, fails to report for a meeting, or violates a curfew are not equivalent to those posed by an escaped prisoner who immediately becomes a fugitive and must act quickly to evade detection. United States v Brown214  points to proximate custodians endangered by a prison break, but defendants on release do not have similarly close guards.215 Release violators certainly have strong incentives to avoid detection and further penalties. But because they are not under direct surveillance, this evasion may not require the dangerous dash to obtain “money, weapons and transportation” that Brown envisions.216

But more importantly, unlike a prisoner or an arrestee, a defendant on release is essentially at liberty, having completed a preceding prison term. Sentencing judges may choose only between releasing the defendant to full liberty or adding a release term and requiring the defendant to adhere to a set of conditions. In short, the public already assumes a baseline degree of recidivism risk when an individual is released from prison. It is not clear that there is any increase in risk created by requiring that individual to undergo additional monitoring, let alone an increase sufficient to justify an independent offense for escape—particularly when there are preexisting mechanisms for redressing release violations.217

One counterargument is that Congress, in authorizing judges to consider incapacitation when imposing supervised release, acknowledged potential risks posed by releasees.218 This argument is not persuasive. Sentencing judges have a corresponding statutory duty to impose “no greater deprivation of liberty than is reasonably necessary.”219 Release conditions are ordinarily quasi-incapacitative at best;220 detentive conditions such as halfway house stays still permit defendants to seek employment and leave with permission. Because the central aim of supervised release is to ease the defendant’s transition to community life, courts may not protect the public by replicating incarceration.

The transitional nature of supervised release is thus fundamentally at odds with the rationale underlying escape prosecutions. Simply put, supervised release requires judges to integrate formerly incarcerated persons into the wider community; escape is designed to punish offenders for the risks that they present to the public when they abscond from seclusion or resist lawful proceedings.

3.   The rule of lenity favors a narrow construction of § 751(a).

The rule of lenity suggests that textual ambiguities in criminal statutes should be construed narrowly.221 Lenity is rarely applied, as it is operative only in “situations in which a reasonable doubt persists about a statute’s intended scope even after resort to the language and structure, legislative history, and motivating policies of [a] statute.”222 In theory, the rule operates as a background default to advance fair notice and cabin enforcement discretion.223 Another rationale is that the rule implements the separation of powers, “strik[ing] the appropriate balance between the legislature, the prosecutor, and the court in defining criminal liability.”224 Under this view, lenity safeguards against interpretative slippage by requiring explicit legislative authorization for the expansion of criminal statutes. One commentator has gone further, suggesting that the rule of lenity is designed to “provoke legislative correction” by having judges construe statutes in favor of traditionally disfavored interest groups, causing favored groups to override the decision but clarify latent ambiguities in the process.225

The interpretative-slippage justification has special relevance in this area, in which changed circumstances may drastically expand an ambiguous criminal statute to a broad class of ex-offenders without a clear legislative mandate. These slippage concerns are more than hypothetical. Prosecutors must enforce rules even in the absence of such a legislative mandate or other similarly final resolution. Thus, expansive court constructions have a direct influence on prosecutorial policy. For example, in the escape context, the entry for § 751(a) in the U.S. Attorneys’ Manual currently provides that custody “need only be minimal.”226 The manual cites Eighth Circuit precedent in support of this rule,227 even though there is disagreement among the circuits over how broadly to construe § 751(a) custody. Although there are other strong statutory construction arguments in favor of an unequivocal finding that supervised release conditions are categorically noncustodial, in the alternative, courts may arrive at the same outcome through the application of lenity.

B.    The Categorical Rule: The Sentencing Reform Act

There are also several policy justifications for a categorical rule. Importantly, these justifications are not freestanding considerations, but are embodied within the SRA. While the circuits that have considered this issue have looked exclusively to § 751(a) in determining whether supervised release conditions are custodial, the SRA is a similarly relevant statute. As discussed in Parts I.A–B, the SRA created supervised release in order to facilitate offender rehabilitation and centralize sentencing discretion. Both of these purposes are inextricably linked with the specific penalty structure that Congress created to redress violations, a structure that is compromised by parallel prosecutions for escape.

One of the principal innovations of the SRA was to separate the punitive and rehabilitative aspects of a prison sentence. To enforce this divide, Congress has historically exerted careful control over the penalties that defendants may face for condition violations.228 The SRA originally did not contain revocation proceedings; Congress instead envisioned that contempt could redress serious violations.229 For less serious violations, releasees’ rehabilitation was best served by intermediate sanctions, rather than returning them to prison.230

Even revocation places meaningful limitations on reincarceration.231 First, revocation is not automatic. Sentencing judges have discretion in determining when to revoke, if at all.232 Second, even when judges opt for prison time, they need not require incarceration for the entire length of the term. Congress made clear that the postrevocation sentence itself may include supervised release.233 This feature was added to allow judges to carefully craft their responses to condition violations, rather than forcing judges to reimprison releasees for longer than necessary.234 In short, revocation is structured to allow judges to ensure that a defendant remains at liberty whenever appropriate.

Because sentencing judges may carefully calibrate their response to a violation, supervised release’s structure tolerates a degree of noncompliance as necessary for rehabilitation. Rather than requiring automatic sanctions for disfavored behaviors, the system allows releasees to make some errors before facing revocation. In line with this toleration, the Seventh Circuit in United States v Smith235 criticized a sentencing judge for precommitting herself to a specific sanction for a violation,236 thus “conflat[ing] criminality with disobedience.”237 Framing condition violations as independent crimes prevents rehabilitation by adding a new punitive layer to this system. Escape prosecutions blend disobedience and criminality by punishing releasees in instances in which a sentencing judge might choose a lesser sanction, or might even decline to sanction at all.238 The results are conceptually odd. Under the dominant approach among the circuits, a sentencing judge’s decision to excuse or to not excuse a condition violation has no bearing on whether the defendant may face criminal liability for that same action.

Escape prosecutions also thwart rehabilitation by drastically multiplying the consequences that a defendant faces for a location-based condition violation. For example, imagine that a defendant commits a crime and is ordered to serve the maximum  term of supervised release authorized by statute for his offense (R1) after serving his prison term (P1). The defendant is ordered to reside at a halfway house, but he leaves the facility and fails to return. The defendant may face up to five years in prison if convicted of escape (P2).239 Further, because escape is a class D felony, the sentencing judge is authorized to impose up to three years of supervised release (R2).240

One of the mandatory conditions of all release terms is that the releasee avoids committing a “Federal, State, or local crime” while under supervision.241 Thus, the defendant is now eligible for revocation of R1. The Sentencing Commission would classify an escape conviction as a Grade B violation.242 The Commission recommends that sentencing judges revoke release for a violation of this grade.243 If the judge revokes R1, the offender will be sentenced to a new prison term (P3) for up to the maximum term length that the judge could have imposed for the defendant’s first term of supervised release (R1), with no reduction for previous time spent on release.244

To further complicate matters, the postrevocation sentence P3 may itself include a supervised release term (R3), so long as the total revocation term (consisting of both imprisonment and the new release term) does not exceed the maximum term of supervised release authorized for the original offense (P3 + R3 ≤ R1).245 In total, a defendant who escapes in the second year of a three-year term of supervised release (R1 = 3) faces up to three years in prison as a revocation penalty (P3), minus any extra release (R3); up to five years in prison for escape (P2); and up to three years of a new supervised release term (R2). A single release violation may transform R1 into at most two release terms and two prison sentences.246 Even though the hypothetical defendant has already completed two years of his supervised release term, he may face a maximum of eight years of prison and a maximum of three years of supervised release.

This dual sentencing process also frustrates the discretion-centralization purposes of the SRA. A key innovation of the SRA was to consolidate sentencing authority in district court judges, rather than partitioning responsibility between judges and the United States Parole Commission.247 Escape prosecutions thwart this reform by multiplying the number of individuals with discretionary control over the defendant’s sentence from the moment that the prosecutor decides to bring an independent criminal charge of § 751 escape.

A lack of centralized control may have the effect of replicating some of the perverse incentives that Congress expressly sought to avoid by consolidating authority in district court judges. For example, both Burke and Foster involved district court judges who added halfway house conditions to give homeless releasees a place to stay at the end of their sentences.248 Tying escape liability to violations may dissuade judges from acting similarly in the future, given that they do not have complete control over the consequences if the releasee absconds and is subsequently indicted for escape. The same applies for judges who use supervised release as a vehicle for granting releasees access to mental health services or treatment programs.

These consequences suggest that courts should avoid construing release conditions as custodial conditions giving rise to § 751(a) escape liability. A categorical rule excluding supervised release conditions from § 751(a) custody best gives effect to the SRA by leaving the existing statutory penalty structure in place.


Commentators have criticized supervised release for failing to live up to its initial promise and perpetuating incarceration in instances in which “the utility of [ ] continued imprisonment is difficult to see.”249 Escape prosecutions exacerbate this effect by doubling the penalties that ex-offenders face for violating the terms of their release, trapping individuals who are not truly reoffending in a cycle of imprisonment.

States with similar supervision programs have confronted similar issues.250 These issues will continue to arise over the coming years as federal parole is phased out and the proportion of ex-offenders serving release terms increases.251 Although shielding supervised release violators from escape prosecutions may seem like a drastic step at first blush, the categorical rule is most consistent with the evolution of the federal crime of escape and best gives effect to the transitional and discretion-centralization aims of the SRA. The rule does not mean that disobedient releasees will get off scot-free—it merely requires courts to use appropriate preexisting mechanisms for sanctioning condition violations. Most importantly, the categorical rule will help clarify the divide between rehabilitation and punishment by carefully defining the consequences that may result from a condition violation.

  • 15See Alan M. Dershowitz, Indeterminate Confinement: Letting the Therapy Fit the Harm, 123 U Pa L Rev 297, 298–301 (1974) (defining indeterminate sentencing as a process in which actual time to be served is determined by both a judge and an administrative board against the background of a statutory mandate).
  • 16See Kate Stith and Steve Y. Koh, The Politics of Sentencing Reform: The Legislative History of the Federal Sentencing Guidelines, 28 Wake Forest L Rev 223, 225 & n 7 (1993).
  • 17William W. Berry III, Discretion without Guidance: The Need to Give Meaning to § 3553 after Booker and Its Progeny, 40 Conn L Rev 631, 635–36 (2008).
  • 18Parole Commission and Reorganization Act § 2, Pub L No 94-233, 90 Stat 219, 219 (1976), codified at 18 USC § 4201 et seq, repealed by the Comprehensive Crime Control Act of 1984 § 218, Pub L No 98-273, 98 Stat 1976, 2027.
  • 19See, for example, United States v Addonizio, 442 US 178, 187–90 (1979) (discussing the relationship between sentencing judges and the Parole Commission).
  • 20See, for example, Sanford H. Kadish, Legal Norm and Discretion in the Police and Sentencing Processes, 75 Harv L Rev 904, 915–29 (1962). See also generally, for example, Stanley A. Weigel, Appellate Revision of Sentences: To Make the Punishment Fit the Crime, 20 Stan L Rev 405 (1968); Marvin E. Frankel, Criminal Sentences: Law without Order (Hill & Wang 1973).
  • 21See, for example, Weigel, 20 Stan L Rev at 406–10 (cited in note 20). But see Kate Stith and José A. Cabranes, Fear of Judging: Sentencing Guidelines in the Federal Courts 106–12 (Chicago 1998) (questioning the accuracy of pre-SRA studies demonstrating sentencing disparities).
  • 22See Comprehensive Crime Control Act Report at 38–41 (cited in note 4) (criticizing the “unfettered discretion the law confers on those judges and parole authorities responsible for imposing and implementing the sentence”); William W. Wilkins Jr, Phyllis J. Newton, and John R. Steer, Competing Sentencing Policies in a “War on Drugs” Era, 28 Wake Forest L Rev 305, 308–10 (1993).
  • 23Comprehensive Crime Control Act Report at 40 (cited in note 4).
  • 24Id.
  • 25See id (“We know too little about human behavior to be able to rehabilitate individuals on a routine basis or even to determine accurately whether or when a particular prisoner has been rehabilitated.”).
  • 26See id at 38, 48. See also Jonathan D. Casper, Determinate Sentencing and Prison Crowding in Illinois, 1984 U Ill L Rev 231, 236 (noting that “[c]onservatives and law enforcement interests” supported determinate sentencing because “parole boards seemed often to release prisoners who continued to pose a danger to society,” and judges seemed “reluctant to send ‘marginal defendants’ to prison”); William J. Genego, Peter D. Goldberger, and Vicki C. Jackson, Project, Parole Release Decisionmaking and the Sentencing Process, 84 Yale L J 810, 826 & n 82 (1975) (“Extensive social science research strongly suggests that rehabilitation—defined as an increasing likelihood of successful adjustment upon release—cannot be observed, detected or measured.”).  
  • 27See Comprehensive Crime Control Act Report at 112–13 (cited in note 4). See also Stefan R. Underhill, Did the Man I Sentenced to 18 Years Deserve It? (NY Times, Jan 23, 2016), online at
    -murderer-or-a-cooperative-witness.html (visited Jan 31, 2016) (Perma archive unavailable) (offering a description of this phenomenon by a federal district judge).
  • 28See 18 USC § 4205(a), repealed by the Comprehensive Crime Control Act of 1984 § 218(a)(5), 98 Stat at 2027.
  • 29Comprehensive Crime Control Act Report at 65 (cited in note 4):

    The shameful disparity in criminal sentences is a major flaw in the existing criminal justice system. . . .

    . . . The bill’s sweeping provisions are designed to structure judicial sentencing discretion, eliminate indeterminate sentencing, phase out parole release, and make criminal sentencing fairer and more certain.

  • 30See generally, for example, Paul J. Hofer and Mark H. Allenbaugh, The Reason behind the Rules: Finding and Using the Philosophy of the Federal Sentencing Guidelines, 40 Am Crim L Rev 19 (2003); Charles J. Ogletree Jr, The Death of Discretion? Reflections on the Federal Sentencing Guidelines, 101 Harv L Rev 1938 (1988).
  • 31See Sentencing Reform Act §§ 218(a), 235(a)(1), 98 Stat at 2031, as amended by Sentencing Reform Amendments Act of 1985 § 4, Pub L No 99-217, 99 Stat 1728, 1728.
  • 32See Comprehensive Crime Control Act Report at 122–23 (cited in note 4).
  • 33Id.
  • 34Id.
  • 35Id at 124.
  • 36However, for an argument that supervised release restored indeterminate sentencing, see generally Fiona Doherty, Indeterminate Sentencing Returns: The Invention of Supervised Release, 88 NYU L Rev 958 (2013).
  • 37Supervised release is occasionally statutorily mandated. See, for example, 21 USC § 841(b) (setting out mandatory release terms for certain drug offenses).
  • 3818 USC § 3583(a).
  • 3918 USC § 3601.
  • 40See 18 USC § 3563(b) (listing permissible probation conditions); 18 USC § 3583(d) (incorporating § 3563(b) in listing valid supervised release conditions).
  • 4118 USC § 3561(a)(3).
  • 4218 USC § 3583(a) (authorizing courts to include a term of supervised release to commence “after imprisonment”). See also Doherty, 88 NYU L Rev at 998 (cited in note 36) (noting that supervised release is “a discretionary supplement to prison, not a sentence in its own right”); United States v Johnson, 529 US 53, 59 (2000) (“The objectives of supervised release would be unfulfilled if excess prison time were to offset and reduce terms of supervised release.”).
  • 43See Johnson, 529 US at 59 (“Supervised release fulfills rehabilitative ends, distinct from those served by incarceration.”).
  • 44See United States v Work, 409 F3d 484, 489–90 (1st Cir 2005) (collecting cases from the Second, Fourth, Fifth, Eighth, Ninth, Tenth, and DC Circuits).
  • 45Comprehensive Crime Control Act Report at 112–13 (cited in note 4).
  • 4618 USC § 3583(b) (setting supervised term limits, except as otherwise provided by statute). But see United States v Moriarty, 429 F3d 1012, 1023–25 (11th Cir 2005) (per curiam) (holding that a lifetime term of supervised release imposed for a sex offense did not violate the Eighth Amendment); United States v Pettus, 303 F3d 480, 487 (2d Cir 2002) (“There is no constitutionally imposed limit on how long a supervised release term can be.”). Sex offenders have the highest rate of revocation for technical violations—perhaps due to the prevalence of lengthy release terms that increase the difficulty of perfect compliance. See Federal Offenders at *68 (cited in note 5).
  • 47See Comprehensive Crime Control Act Report at 123 (cited in note 4) (“Unlike [under] current parole law, the question whether the defendant will be supervised following his term of imprisonment is dependent on whether the judge concludes that he needs supervision, rather than on the question whether a particular amount of his term of imprisonment remains.”).
  • 4818 USC § 3583(d) (listing standard release conditions).
  • 4918 USC § 3583(d).
  • 5018 USC § 3583(d).
  • 5118 USC § 3583(e)(2).
  • 5218 USC § 3583(e)(1) (authorizing a sentencing judge to terminate a release term after one year if “such action is warranted by the conduct of the defendant released and the interest of justice”).
  • 53See 18 USC § 3553(a) (listing factors for consideration such as “the nature and circumstances of the offense and the history and characteristics of the defendant”; “the need for the sentence imposed . . . to afford adequate deterrence to criminal conduct . . . [and] to provide the defendant with needed educational or vocational training, medical care, or other correctional treatment . . .”; the Sentencing Guidelines; policy statements issued by the Sentencing Commission; “the need to avoid unwarranted sentence disparities”; and “the need to provide restitution to any victims of the offense”).
  • 54See Comprehensive Crime Control Act Report at 125 (cited in note 4) (“The term of supervised release . . . may not be imposed for purposes of punishment or incapacitation since those purposes will have been served to the extent necessary by the term of imprisonment.”).
  • 55Sentencing Act of 1987 § 9, Pub L No 100-182, 101 Stat 1266, 1267, codified at 18 USC § 3583(e)(1).
  • 56See Anti-Drug Abuse Act of 1986 § 1006(a)(3)(D), Pub L No 99-570, 100 Stat 3207, 3207-6 to -7, codified as amended at 18 USC § 3583(e)(3).
  • 57See 18 USC § 3583(e)(2)–(3).
  • 58Comprehensive Crime Control Act Report at 125 (cited in note 4) (“Unlike a term of probation, however, the term of supervised release is not subject to revocation for a violation. Instead, for the usual violations, the term or condit[i]ons of supervised release may be amended pursuant to subsection (e).”).
  • 59Id.
  • 60Id.
  • 61See id at 59 (suggesting that offering sentencing options aside from imprisonment may avoid unwarranted restrictions on liberty).
  • 6218 USC § 3583(e)(3).
  • 63See, for example, Doherty, 88 NYU L Rev at 1000–04 (cited in note 36) (“[T]he possibility of revocation made supervised release indisputably about punishment, oversight, and coercion.”); Christine S. Scott-Hayward, Shadow Sentencing: The Imposition of Federal Supervised Release, 18 Berkeley J Crim L 180, 201–04 (2013) (analyzing the effect of supervised release on defendants and concluding that release does not advance public safety or rehabilitative goals).
  • 64Comprehensive Crime Control Act Report at 58 (cited in note 4).
  • 65See, for example, United States v Nolan, 109 F Supp 2d 350, 351–52 (ED Pa 2000) (imposing a revocation penalty of ten months’ imprisonment on a defendant who failed to report to five scheduled meetings with his supervising officer and failed to enter a drug-treatment program, in violation of the terms of his supervised release); United States v McCauley, 102 F Supp 2d 271, 271–72 (ED Pa 2000) (revoking release for a defendant who tested positive for cocaine and alcohol).
  • 66See, for example, Johnson, 529 US at 60 (“In the instant case, the transition assistance ordered by the trial court required respondent, among other conditions, to avoid possessing or transporting firearms and to participate in a drug dependency treatment program. These conditions illustrate that supervised release, unlike incarceration, provides individuals with postconfinement assistance.”); Tapia v United States, 131 S Ct 2382, 2390 (2011) (“[W]hen Congress wanted sentencing courts to take account of rehabilitative needs, it gave courts the authority to direct appropriate treatment for offenders. Thus, the SRA instructs courts, in deciding whether to impose probation or supervised release, to consider whether an offender could benefit from training and treatment programs.”).
  • 67See Moriarty, 429 F3d at 1023–25 (holding that a lifetime term of supervised release imposed for a sex offense did not violate the Eighth Amendment, citing the transitional aims of release); United States v Williams, 636 F3d 1229, 1232–34 (9th Cir 2011) (same).
  • 68Pettus, 303 F3d at 487.
  • 69See, for example, Williams, 636 F3d at 1232 (“[A]lthough supervised release limits a criminal’s liberty and privacy, it is a punishment far less severe than prison.”); Pettus, 303 F3d at 486 (noting that while “supervised release is technically a punishment, . . . it is primarily intended to protect the public from further crimes by easing the re-entry of a convicted defendant into society”).
  • 7018 USC § 3585(a) (“A sentence to a term of imprisonment commences on the date the defendant is received in custody awaiting transportation to, or arrives voluntarily to commence service of sentence at, the official detention facility at which the sentence is to be served.”).
  • 7118 USC § 3585(b) (listing circumstances in which a defendant may be given credit for time served in prior custody).
  • 72See, for example, USSG § 3C1.1 (providing a sentencing enhancement for obstruction of justice in cases in which the defendant absconded from federal custody).
  • 73See 28 USC § 2241(c) (authorizing habeas petitions for individuals in enumerated forms of custody). See also Part II.B.
  • 7418 USC § 751(a).
  • 7518 USC § 3621(a) (“A person who has been sentenced to a term of imprisonment . . . shall be committed to the custody of the Bureau of Prisons until the expiration of the term imposed.”).
  • 76See Johnson, 529 US at 57 (“[T]he ordinary, commonsense meaning of release is to be freed from confinement. To say respondent was released while still imprisoned diminishes the concept the word intends to convey.”).
  • 77A judge may sentence a defendant to home confinement as a condition of supervised release, but if this is done, the term of the confinement functions similarly to parole in that it is treated as an alternative, rather than as a supplement, to incarceration. That is, home confinement as a condition to supervised release acts as “an alternative to incarceration” such that the total amount of time served, in prison and in home confinement, cannot exceed the statutory maximum for the crime committed by the defendant. See USSG § 5C1.1(c)(2)–(3) (recommending that judges use home confinement to reduce a direct prison sentence); 18 USC § 3583(e)(4) (providing that judges may impose home confinement “only as an alternative to incarceration”); United States v Ferguson, 369 F3d 847, 850 (5th Cir 2004) (per curiam) (holding that “incarceration and home [confinement] are alternative punishments that may not combine in excess of the maximum statutory term of incarceration”).
  • 7818 USC § 3563(b)(10).
  • 79See 18 USC §§ 3563(b)(1), 3583(d).
  • 80See, for example, United States v Griner, 358 F3d 979, 982 (8th Cir 2004) (upholding a release condition requiring the defendant to reside at a community corrections facility).
  • 81See, for example, United States v Rivera-López, 736 F3d 633, 634, 636–37 (1st Cir 2013) (upholding conditions requiring the defendant to remain at his residence from 6:00 p.m. to 6:00 a.m. and to wear an electronic-monitoring device twenty-four hours a day).
  • 82See United States v Porter, 555 F Supp 2d 341, 343, 345 (EDNY 2008).
  • 83See, for example, USSG § 5D1.3(c)(1) (recommending a condition directing the defendant to stay within the judicial district except with permission from the probation officer). See also United States v Truscello, 168 F3d 61, 63 (2d Cir 1999) (“[T]he so-called ‘standard conditions’ [of USSG § 5D1.3(c)] . . . are basic administrative requirements. . . . [T]hey are almost uniformly imposed by the district courts and have become boilerplate.”) (quotation marks and brackets omitted).
  • 84See, for example, United States v Cason, 25 F Supp 3d 1212, 1213 (WD Mo 2014) (imposing a supervised release condition requiring the defendant to “use contraceptives . . . unless such use would violate his religious scruples or is expressly rejected by his sexual partner”). See also United States v Borski, 2013 WL 4830238, *2–3 (SD Ind) (involving a defendant who was forbidden to possess “any pornography, erotica, or nude images” as a condition of supervised release).
  • 85United States v Soltero, 510 F3d 858, 865 (9th Cir 2007) (per curiam).
  • 86See, for example, United States v Kingsley, 241 F3d 828, 837–38 (6th Cir 2001).
  • 87See United States v Thurlow, 44 F3d 46, 47 (1st Cir 1995) (per curiam).
  • 88See United States v Barajas, 331 F3d 1141, 1143 (10th Cir 2003).
  • 8917–A Me Rev Stat Ann § 755(1–B).
  • 90See 18 USC § 3624(c)(1) (directing the Bureau of Prisons, “to the extent practicable,” to ensure that prisoners spend a portion of their last year of imprisonment “under conditions that will afford that prisoner a reasonable opportunity to adjust to and prepare for the reentry of that prisoner into the community”).
  • 91See, for example, Nace v United States, 334 F2d 235, 235–36 (8th Cir 1964) (per curiam) (holding that the defendant, who was allowed to serve his sentence at a halfway house, was in custody for purposes of the escape statute notwithstanding the freedom allowed to him at the institution); McCullough v United States, 369 F2d 548, 549–50 (8th Cir 1966); Perez–Calo v United States, 757 F Supp 1, 2 (D Puerto Rico 1991).
  • 92726 F3d 305 (2d Cir 2013).
  • 93Id at 307.
  • 94Id.
  • 95Id at 307–08.
  • 96Edelman, 726 F3d at 308.
  • 97Id.
  • 98Id at 309.
  • 99Id at 309–10, quoting 18 USC § 751(a).
  • 100Edelman, 726 F3d at 309–10.
  • 101Id.
  • 102Id at 307.
  • 103Id at 309–10.
  • 104788 F3d 873 (8th Cir 2015).
  • 105Id at 876.
  • 106Id at 875.
  • 107See 18 USC § 751(a).
  • 108See, for example, United States v Dowling, 962 F2d 390, 391 (5th Cir 1992) (involving a defendant ordered to reside at a halfway house as a condition of probation).
  • 109See 18 USC § 3561(a)(3).
  • 110See Goad, 788 F3d at 876.
  • 11118 USC § 751(a) (emphases added).
  • 112See Goad, 788 F3d at 876 (suggesting that the lack of reference to any particular custodian in this clause suggests that all custodians must be treated alike).
  • 113Id.
  • 114754 F3d 1186 (10th Cir 2014).
  • 115Id at 1189–91. 
  • 116Id at 1187.
  • 117Id.
  • 118Foster, 754 F3d at 1188.
  • 119Id.
  • 120Id at 1189.
  • 121Id.
  • 122Foster, 754 F3d at 1192–93.
  • 123Id at 1190.
  • 124977 F2d 1412 (10th Cir 1992).
  • 125Id at 1413–14. Habeas corpus ad prosequendum is a writ issued to remove a prisoner from one jurisdiction to another to face prosecution. See United States v Londono, 285 F3d 348, 356 (5th Cir 2002) (per curiam).
  • 126Depew, 977 F2d at 1413.
  • 127Id at 1414.
  • 128See, for example, Foster, 754 F3d at 1189–90.
  • 129Depew, 927 F2d at 1414. The Depew court noted that the federal escape statute had been applied to offenders “serving [ ] state prison sentence[s] at [ ] federal correctional institution[s],” offenders “serv[ing] [ ] federal prison sentence[s] in [ ] state penitentiar[ies],” and offenders in transit between federal and state or local facilities. Id.
  • 130Foster, 754 F3d at 1190.
  • 131Id, quoting United States v Sack, 379 F3d 1177, 1179 n 1 (10th Cir 2004).
  • 132739 F3d 558 (10th Cir 2014).
  • 133Id at 562.
  • 134Id at 559–60.
  • 135Id at 560.
  • 136Ko, 739 F3d at 561–62.
  • 137The Tenth Circuit has suggested in dicta that a release condition requiring a defendant to reside with a family member may create § 751(a) custody. See Sack, 379 F3d at 1180 n 2:

    A defendant leaving the custody of the third party custodian . . . can menace the welfare of that custodian as easily as a defendant escaping from jail can menace corrections officers. The character of escape as “absenting oneself from custody without permission” is not altered by pre-existing relationships between the custodian and the detainee.

    But see United States v Miranda, 749 F Supp 1062, 1064 (D Colo 1990) (“[A] bond condition that defendant live in a particular place is not tantamount to ‘custody.’ If the defendant had been required to live with his parents as a condition of bond, he would not have been charged with escape from custody if he decided to leave his parent[s’] residence.”).

  • 138694 F3d 1062 (9th Cir 2012).
  • 139Id at 1063.
  • 140Id.
  • 141Id at 1064–65.
  • 142Burke, 694 F3d at 1064–65. See Foster, 754 F3d at 1191 (“In Burke, the Ninth Circuit focused only on whether Burke’s freedom was sufficiently restricted to constitute custody, looking to the fact that the conditions of his release were much more analogous to probation than they were to imprisonment.”) (quotation marks omitted).
  • 143Burke, 694 F3d at 1064.
  • 144See Edelman, 726 F3d at 307 (noting that residents at the halfway house “could not leave without notifying staff” and “could only leave the building for up to 12 hours at a time”).
  • 145See Burke, 694 F3d at 1064.
  • 146Id.
  • 147Id.
  • 148Id.
  • 149See Part I.C.
  • 150Burke, 694 F3d at 1064, quoting United States v Baxley, 982 F2d 1265, 1269 (9th Cir 1992).
  • 151See Burke, 694 F3d at 1065 (“Like an individual on probation, Burke was conditionally released from incarceration; his failure to return to [the halfway house] was a violation of his release conditions punishable by revocation of release, not an escape from ‘custody’ within the meaning of § 751(a).”); Baxley, 982 F2d at 1269–70 (stating that a probationer becomes an escapee only when, after having his probation revoked, “he thereafter fails to report for custodial incarceration”).
  • 152See Burke, 694 F3d at 1065.
  • 153Foster, 754 F3d at 1190.
  • 15418 USC § 751(a).
  • 155See Part III.A.3.
  • 156See Goad, 788 F3d at 876 n 3; Foster, 754 F3d at 1193–94; Edelman, 726 F3d at 309–10; Burke, 694 F3d at 1065.
  • 15728 USC § 2241(c)(3).
  • 158371 US 236 (1963).
  • 159Id at 236.
  • 160Id at 243.
  • 161Id at 242–43.
  • 162Jones, 371 US at 239.
  • 163Id at 242.
  • 164Id at 242–43.
  • 165See, for example, Yates v United States, 135 S Ct 1074, 1082 (2015) (“We have several times affirmed that identical language may convey varying content when used in different statutes, sometimes even in different provisions of the same statute.”); Robinson v Shell Oil Co, 519 US 337, 342–44 (1997) (noting that “employee” has different meanings in different sections of Title VII of the Civil Rights Act of 1964). But see Yates, 135 S Ct at 1096–97 (Kagan dissenting) (“[T]he normal rule of statutory construction assumes that identical words used in different parts of the same act, passed at the same time, are intended to have the same meaning. And that is especially true when the different provisions pertain to the same subject.”) (quotation marks and citation omitted).
  • 166546 US 303 (2006).
  • 167Id at 315–17.
  • 168Id at 319. See also Walter Wheeler Cook, “Substance” and “Procedure” in the Conflict of Laws, 42 Yale L J 333, 337 (1933) (“The tendency to assume that a word which appears in two or more legal rules . . . has and should have precisely the same scope in all of them, runs all through legal discussions. It has all the tenacity of original sin and must constantly be guarded against.”).
  • 169Jones, 371 US at 243.
  • 170Harris v Nelson, 394 US 286, 291 (1969).
  • 171See, for example, Hensley v Municipal Court, San Jose-Milpitas Judicial District, Santa Clara County, 411 US 345, 345, 351 (1973) (holding that a person released “on his own recognizance” after conviction satisfies the habeas custody requirement).
  • 172See Part III.A.3.
  • 173See Barry v Bergen County Probation Department, 128 F3d 152, 161 (3d Cir 1997) (holding that community service satisfies the habeas custody requirement).
  • 174439 F2d 1074 (6th Cir 1971).
  • 175See Edelman, 726 F3d at 309; Goad, 788 F3d at 875.
  • 176Rudinsky, 439 F2d at 1076–77.
  • 177See Foster, 754 F3d at 1189–92. See also Part II.A.2.
  • 178Rudinsky, 439 F2d at 1075.
  • 179Ko, 739 F3d at 559.
  • 180379 F3d 1177 (10th Cir 2004).
  • 181Id at 1182. See also Edelman, 726 F3d at 309; Goad, 788 F3d at 875–76; Foster, 754 F3d at 1190.
  • 182982 F2d 1265 (9th Cir 1992).
  • 183Id at 1269–70; Burke, 694 F3d at 1064–65.
  • 184Foster, 754 F3d at 1191 n 3 (quotation marks and emphasis omitted).
  • 185See 18 USC § 3142(c)(1)(B) (directing courts to create pretrial release conditions to “assure the appearance” of the defendant at trial).
  • 186Black’s Law Dictionary 467 (West 10th ed 2014) (“Black’s 2014”). The 1910 edition is the closest in time to the enactment of § 751(a). The entry for “custody” provides three definitions with increasing specificity: “care and keeping,” “detainer,” and “actual imprisonment.” Black’s Law Dictionary 309 (West 2d ed 1910). The entry also notes that a sheriff who enforces a custodial sentence by instead “allowing the defendant to go at large under his general watch and control” may be held criminally liable for the defendant’s escape. Id.
  • 187Black’s 2014 at 467 (cited in note 186).
  • 188Webster’s Third New International Dictionary of the English Language Unabridged 559 (Merriam 2002).
  • 189See Haggar Co v Helvering, 308 US 389, 394 (1940).
  • 190Goad, 788 F3d at 876, quoting 18 USC § 751(a). See also Foster, 754 F3d at 1189; Edelman, 726 F3d at 308.
  • 191Act of May 14, 1930 § 9, 46 Stat 325, 327, codified as amended at 18 USC § 751.
  • 192United States v Brown, 333 US 18, 21 (1948), citing generally HR Rep No 71-106, 71st Cong, 2d Sess (1930).
  • 193Act of May 14, 1930 § 9, 46 Stat at 327, codified as amended at 18 USC § 751.
  • 194See Administration of Federal Prisons, S Rep No 74-1021, 74th Cong, 1st Sess 1 (1935). See also Escape from Custody prior to Conviction, HR Rep No 74-803, 74th Cong, 1st Sess 1 (1935) (“The purpose of H. R. 3430 is to make escape or attempted escape from custody under lawful arrest before conviction a criminal offense.”).
  • 195See Administration of Federal Prisons at 1 (cited in note 194); Escape from Custody prior to Conviction at 2 (cited in note 194).
  • 196See Administration of Federal Prisons at 2 (cited in note 194) (including, for example, a Michigan statute providing that “any person lawfully imprisoned awaiting examination on trial, arraignment or sentence who shall escape shall be guilty of an offense”); Escape from Custody prior to Conviction at 1 (cited in note 194).
  • 197Act of Aug 3, 1935, 49 Stat 513, 513–14, codified as amended at 18 USC § 751.
  • 198See Administration of Federal Prisons at 1 (cited in note 194); Escape from Custody prior to Conviction at 2 (cited in note 194).
  • 199See Ratzlaf v United States, 510 US 135, 140–41 (1994) (“Judges should hesitate [ ] to treat statutory terms [as superfluous] in any setting, and resistance should be heightened when the words describe an element of a criminal offense.”).
  • 200I could not find any § 751(a) prosecutions for parole violators, even though federal parole existed when the act codifying escape was first passed and parole may include location-based conditions. Existing cases deal only with parole violators who are first arrested or recommitted and then subsequently abscond. See, for example, United States v Franklin, 313 F Supp 43, 44–47 (SD Ind 1970).
  • 201See United States v Person, 223 F Supp 982, 984 (SD Cal 1963).
  • 202See Part I.A.
  • 203Act of June 25, 1948 § 751, 62 Stat 683, 734–35, codified as amended at 18 USC § 751.
  • 204See Report from the Committee on the Judiciary, HR Rep No 80-304, 80th Cong, 1st Sess 2–9 (1947) (describing the mechanisms and goals of the 1948 revisions).
  • 205CNA Financial Corp v Donovan, 830 F2d 1132, 1146 (DC Cir 1987).
  • 206Report from the Committee on the Judiciary at A67 (cited in note 204).
  • 207Furloughs are permitted under 18 USC § 3622 (authorizing the Bureau of Prisons to release prisoners for limited periods of time in order to visit dying relatives and obtain medical treatment, among other reasons).
  • 208Act of Sept 10, 1965, Pub L No 89-176, 79 Stat 674, 675, codified at 18 USC § 4082(d). For a discussion of the two statutes, see United States v Leonard, 498 F2d 754, 757 (DC Cir 1974) (suggesting that Congress enacted the Act of September 10, 1965, in order to criminalize prisoners’ failure to return from furloughs in response to a district court opinion declaring such acts to be outside the scope of § 751(a)).
  • 20918 USC § 4082(a).
  • 210See Jones, 371 US at 241–43 (discussing the need to define § 2241 habeas custody broadly to protect individuals from wrongful restraint); Howes v Fields, 132 S Ct 1181, 1189 (2012) (defining custody for Miranda purposes as “a term of art [ ] specif[ying] circumstances that are thought generally to present a serious danger of coercion,” focusing on “all of the circumstances surrounding the interrogation”).
  • 211Brown, 333 US at 21 n 5.
  • 212Id.
  • 213United States v Bailey, 444 US 394, 413 (1980).
  • 214333 US 18 (1948).
  • 215Releasees are supervised by probation officers. Many judges impose standard conditions requiring defendants to report to their officer on a regular basis, allow visits by their officer, and notify their officer if they are arrested or questioned by law enforcement officers. See USSG § 5D1.3(c)(2), (10)–(11).
  • 216Brown, 333 US at 21 n 5.
  • 217See Part I.B.
  • 218See text accompanying note 55.
  • 21918 USC § 3583(d)(2).
  • 220Detentive release conditions are subject to special statutory limitations. See 18 USC § 3583(d)(2). See also notes 77–78 and accompanying text.
  • 221See Yates v United States, 135 S Ct 1074, 1088–89 (2015) (“[A]mbiguity concerning the ambit of criminal statutes should be resolved in favor of lenity.”).
  • 222Moskal v United States, 498 US 103, 108 (1990) (quotation marks and emphasis omitted).
  • 223See United States v Standard Oil Co, 384 US 224, 234–37 (1966) (Harlan dissenting).
  • 224Liparota v United States, 471 US 419, 427 (1985).
  • 225Einer Elhauge, Preference-Eliciting Statutory Default Rules, 102 Colum L Rev 2162, 2166, 2192–96 (2002).
  • 226See U.S. Attorneys’ Manual § 1808 (Department of Justice, Offices of the United States Attorneys), archived at
  • 227See id, citing United States v Cluck, 542 F2d 728, 731 (8th Cir 1976).
  • 228See Part I.B.
  • 229See SRA, 98 Stat at 2000, codified at 18 USC § 3585(e)(3).
  • 230See notes 58–61 and accompanying text.
  • 231See 18 USC § 3583(e)(3) (setting the upper limits for revocation imprisonment for different classes of offenders); USSG § 7B1.1 (defining three grades of release violations); USSG § 7B1.3(a) (suggesting in a policy statement that judges should revoke release for Grade A or B violations, but may choose to revoke or otherwise modify release terms for Grade C violations).
  • 232See 18 USC § 3583(e).
  • 233See 18 USC § 3583(h).
  • 234See 137 Cong Rec 14823–24 (June 13, 1991) (describing how, without postrevocation supervision, judges “may be faced with a Hobson’s choice” between “continuing a defendant on supervision who merits a more severe sanction” and “re-imprison[ing] [a defendant] for a longer period than otherwise necessary”).
  • 235770 F3d 653 (7th Cir 2014).
  • 236Id at 655.
  • 237Id at 657.  
  • 238The handbook for federal probation officers includes suggested responses for violations. These responses include formal reprimands, requests for modification of release conditions, and requests for revocations. Suggested responses to a first or second violation of home confinement include “[n]oncompliance meeting[s],” “[l]etter[s] of warning,” “[i]ntensive supervision,” and “[i]ntermittent confinement.” See Guide to Judiciary Policy: Probation and Pretrial Services; Supervision of Federal Offenders § 620.40.20(b)(1) (Dec 10, 2010), archived at
  • 23918 USC § 751(a).
  • 24018 USC § 3583(b)(2).
  • 24118 USC § 3583(d).
  • 242See USSG § 7B1.1(a)(2).
  • 243USSG § 7B1.3(a)(1).
  • 24418 USC § 3583(e)(3).
  • 245This inequality must hold because, as previously defined, R1 represents the maximum release term that a judge could statutorily authorize for the given offense. See 18 USC § 3583(h).
  • 246Terms of supervised release do not run during any period in which the defendant is imprisoned in connection with a conviction for a crime, unless the imprisonment is for less than thirty days. The defendant in this hypothetical scenario likely cannot serve P2 and R1 concurrently. See 18 USC § 3624(e). Further, while supervised release terms run concurrently, judges may run prison terms consecutively upon revocation. See United States v Deutsch, 403 F3d 915, 917–18 (7th Cir 2005) (per curiam).
  • 247See text accompanying note 45.
  • 248See Burke, 694 F3d at 1063; Foster, 754 F3d at 1187–88.
  • 249Smith, 770 F3d at 657.
  • 250See, for example, Colorado Commission on Criminal and Juvenile Justice 2008 Annual Report *164 (Colorado Department of Public Safety, Dec 2008), archived at http:// (noting that in fiscal year 2007, 1,249 individuals were charged with escape from various forms of criminal-justice supervision, including parole and community corrections programs).
  • 251In cases in which defendants were sentenced to prison terms in excess of one year and supervised release was not statutorily mandated, sentencing courts still imposed release terms 99.1 percent of the time. See Federal Offenders at *52 (cited in note 5).