The Yale Law Journal

March 2014

Let the Burden Fit the Crime: Extending Proportionality Review to Sex Offenders

Draconian restrictions on the activities and privacy of convicted sex offenders are a new, and troublesome, trend. In 1994 and 2006, following a national dialogue about crimes against children sparked by several high-profile incidents, Congress passed two laws requiring states to register and regulate sex offenders residing within their borders.1 States and municipalities soon caught on, and deepened restrictions. In the last five years alone, local governments have forbidden sex offenders to live within 2,000 feet of schools;2 “be” within 500 feet of parks or movie theaters;3 enter public libraries;4 drive buses or taxis;5 photograph or film minors;6 and use social networking websites like Facebook.7 Others have required sex offenders to advertise their status on driver’s licenses8 or social networking profiles;9 wear GPS bracelets at their own expense;10 notify local police when present in any county within the state for longer than ten days;11 provide notice to all new neighbors within a roughly quarter-mile radius when they move;12 and pay up to $100 annually to maintain sex offender registries.13These burdens typically last for a decade or for life, depending on the jurisdiction and the type of crime committed.14

Some sex offender restrictions (SORs) can be harder to comply with than is apparent at first glance. Residency restrictions, for example, can create overlapping forbidden zones that bar sex offenders from living in entire cities.15 A few years ago in Miami, sex offenders were known to camp under a bridge, one of few locations outside all exclusion zones.16 Violating these restrictions can result in a felony conviction and as many as ten years in prison.17

Legislators often justify SORs on the grounds of preventing recidivism and see child molesters as the main targets.18 The theory is that if the public is notified of dangerous past offenders in their neighborhood, and those offenders are barred from entering public areas frequented by vulnerable persons, then future sexual assaults can be prevented. The argument seems to be persuasive to much of the public, for SORs are popular with constituents.19 Yet very few sex offenders—including child molesters—are recidivists.20 Of those convicted of offenses against children, even fewer pose threats to the public at large because victims are often a member of the offender’s own family.21

These restrictions often indiscriminately reach all sex offenders, irrespective of their dangerousness. Many SORs apply to first-time offenders and those convicted of minor offenses.22 An eighteen-year-old who had consensual sex with his sixteen-year-old girlfriend and was convicted of statutory rape in Arizona could be required to register as a sex offender, depending only on his and her exact birth dates.23 Or someone convicted of urinating in public in California today might be forced to live in an isolated corner of Sacramento.24

SOR statutes, however carelessly drawn, largely evade constitutional checks. They ordinarily meet the rational basis test, for legislators can always allege that sex offenders pose some threat of recidivism, and that restricting their movements or privacy reduces the threat, even when they fail to target the sex offenders who pose an actual danger. Courts have condoned SOR statutes under rational-basis review even when legislators acted on erroneous information about recidivism rates among sex offenders.25 And the Supreme Court has explicitly limited other constitutional review of SORs. In Smith v. Doe, the Court held that community notification laws for sex offenders do not violate the Ex Post Facto Clause because post-sentence restrictions on sex offenders are “civil” rather than “criminal.”26 Presumably this means SORs cannot be reviewed under the Cruel and Unusual Punishment27 or Double Jeopardy Clauses, either. Because SORs typically enroll sex offenders automatically based on prior convictions at trial, the Court held in a companion case to Smith that the Due Process Clauses do not require the government to hold individualized hearings to determine whether an offender should be subject to SORs.28

The lack of constitutional checks might be surprising given that courts review some civil sanctions for “proportionality” with the underlying wrongdoing under the Due Process Clauses. In the landmark case BMW of North America, Inc. v. Gore,29 the Supreme Court invalidated a punitive damage award in a civil lawsuit as “grossly excessive” compared to the injury done.30 The Court developed three “guideposts” for assessing proportionality that resemble the factors in cruel-and-unusual-punishment analysis under the Eighth Amendment.31

This Comment argues that the justifications the Court appealed to in the punitive damages context militate with equal or greater force for a comparable requirement for SORs, and then proceeds to show that the test the BMW Court outlined easily translates into the SOR context. While scholars have proposed constitutional review of SORs under doctrines other than the one articulated in this Comment, proportionality offers unique advantages. Unlike review under the Ex Post Facto Clause32 or Eighth Amendment,33 proportionality does not flout Smith’s judgment that SORs are civil regulations. Unlike review for violation of particular constitutional liberties, such as free speech,34 proportionality can be used to review all SORs, not just those that infringe particular substantive rights. Proportionality analysis also drives to the heart of the problem with many recent SORs: that they target more people, more harshly, than is necessary to achieve their objectives. It could therefore go a long way toward curbing SORs of great severity or duration that are imposed indiscriminately on many or all categories of sex offenders, including those convicted of non-violent crimes.

I. proportionality review of punitive damages

Courts review criminal punishments under the Eighth Amendment for proportionality, meaning that punishments must bear some reasonable relationship to the crimes that trigger them.35 While early twentieth-century precedents hinted that due process requires civil penalties, too, to be proportionate, the Court did not invalidate a single civil damages award for disproportionality from the Lochner era until 1996.36 Due process regulated only the procedures used to deprive a person of life, liberty, or property, imposing no substantive limit on the amount of that deprivation.

The early hint that proportionality might be a “substantive” due process right was confirmed in 1993 in TXO Production Corp. v. Alliance Resources Corp., when a plurality of the Court extended proportionality review to punitive damages.37 Punitive damages may be imposed in civil lawsuits, in addition to compensatory damages, to further state interests in punishment and deterrence.38 Generally they are available only for reckless, malicious, or oppressive conduct.39 TXO established that a punitive damages award violates the Fourteenth Amendment’s Due Process Clause if it is “grossly excessive” relative to the state’s legitimate purposes.40 The Court has not applied due process proportionality to any other civil penalty.41

When the Court first overturned a punitive damages award as “grossly excessive,” in Gore in 1996, it articulated three “guideposts” for excessiveness: the reprehensibility of the conduct; the disparity between the punitive and compensatory damages; and the difference between the punitive damages and other civil penalties for similar misconduct, in the same or other jurisdictions.42 This review is stricter than rational basis but falls short of heightened scrutiny.43 Using the Gore factors, the Court has invalidated punitive damages with ratios to compensatory damages of 500:144 and 145:1.45 In 2003, the Court refined the test into a near-categorical principle: “few awards exceeding a single-digit ratio between punitive and compensatory damages, to a significant degree, will satisfy due process.”46

The Court has offered several justifications, express and implied, for applying proportionality review to punitive damages. On the cursory reasoning of the early 1900s civil proportionality cases, cited by Gore, deprivations of property that are “plainly arbitrary and oppressive”47 or “grossly excessive”48 compared to what is required to achieve any government purpose are presumably issued without due process. In and since Gore, the Court has identified specific features of punitive damages that raise special due process concerns triggering proportionality review: they (1) may not give fair notice to potential offenders of the penalties for their conduct; (2) are imposed through an adjudication designed for a different purpose; and (3) are imposed with so much discretion and against such disfavored groups that the risk of arbitrary enforcement is high. Additionally, while the Court has not expressly observed it, a fourth feature unites punitive damages with other sanctions reviewed for proportionality under other constitutional provisions: they serve to punish, a function that implicates traditional retributive limiting principles like proportionality.

II. the parallel between punitive damages and sex offender restrictions

The four justifications listed above for applying a proportionality principle to punitive damages can readily be applied in other contexts. Below, I elaborate on these justifications and explain how they apply with equal or greater force to SORs. Conveniently, a 2003 opinion by Chief Justice Rehnquist suggests that a “substantive due process” challenge to SORs remains open.49

First, both punitive damages and SORs raise concerns about adequate notice. The Court’s primary reason for imposing a proportionality requirement in Gore was that, without it, potential offenders would lack “fair notice” of the legal consequences of their conduct.50 Under fair notice doctrine, due process demands that the law state explicitly and precisely what conduct is forbidden.51 But the Court has gone further in punitive damages cases, observing in Gore that “[e]lementary notions of fairness enshrined in our constitutional jurisprudence dictate that a person receive fair notice not only of the conduct that will subject him to punishment, but also of the severity of the penalty that a State may impose.”52 Amounts of punitive damages will be hard for a tortfeasor to predict, because—absent proportionality—those damages need not be anchored to any reasonably knowable facts, like the actual or foreseeable injury caused. Often juries may calculate punitive damages to deter future misconduct not just by the tortfeasor on trial, but also by other potential defendants, about whom the tortfeasor will ordinarily lack information.

SORs may create an even greater fair notice problem than punitive damages, because SORs are harder to predict. A prospective offender should be able to roughly determine the criminal sentences he could face,53 but he can only guess about SORs yet to be enacted. Legislatures may restrict new liberties of sex offenders long after their crimes and even after their prison sentences,54 and across many different aspects of their lives; legislators are growing more creative. For those convicted of minor sex offenses—such as teenagers convicted under harsh statutory rape laws—SORs may be worse than their criminal sentence of a few months in prison. If we take seriously the fair notice logic of Gore, SORs should have some limiting principle so that potential sex offenders—like tortfeasors—can predict the range of legal consequences that could result from their misconduct. Proportionality would provide that limiting principle.

The principle that criminal defendants should have notice of civil as well as criminal consequences of their crimes is already implicit in the Court’s jurisprudence. Padilla v. Kentucky recently held that criminal defendants engaged in plea bargaining have a constitutional right to be informed by counsel that a conviction might result in the civil consequence of deportation, in part because of recent expansions in deportation-eligible offenses.55 Admittedly, pleading guilty is a different decision than committing the crime, the decision with which fair notice doctrine is concerned. But Padilla implies that a person considering committing or admitting to a crime may find the legal consequences of that crime relevant, regardless of their classification as civil or criminal, because of the hardship they would inflict on that person.

Second, both punitive damages and SORs are imposed as a result of an ill-fitting process. The trial deemed “due process” for imposing each sanction is designed to make a factual determination different than the one that justifies imposing the sanction. A civil jury is primarily tasked with deciding questions of duties, causation, and valuation. Awarding punitive damages is a task more akin to penal policy-setting: how much money is necessary to punish the defendant, and to deter his and others’ misconduct?56 Yet, because liability and compensation are a trial’s focus, the jury may not receive evidence about the amount of punitive damages needed to meet these policy objectives.57 Often the only relevant evidence presented at trial concerns the defendant’s mens rea, a precondition for punitive damages. As the Court explained in a punitive damages case, “evidence of culpability warranting some punishment is not a substitute for evidence providing at least a rational basis for the particular deprivation of property imposed by the State to deter future wrongdoing.”58

The mismatch between SORs and their “due process” is greater. Almost all sex offender restrictions are justified by state legislatures as preventing crime, or allaying community fear of it.59 Legislators often cite the erroneous but common belief that sex offenders have higher recidivism rates than other types of criminals.60 Many SORs primarily aim to protect children against this recidivism. The major pieces of federal legislation that enact SORs are the Jacob Wetterling Act and the Adam Walsh Act, both named after child victims of violent crime.61 Yet a trial for a sex offense is a poor process for determining whether a sex offender is a future danger to anyone—much less a danger to children in particular. Juries are not asked to determine whether defendants are likely to re-offend, or against whom. The Court itself admitted this inadequacy of trials, suggesting that an SOR statute that imposes a severe restriction and acknowledges an aim to incapacitate “particularly dangerous” offenders might, to satisfy due process, need an additional individualized adjudication of dangerousness.62 This is a strange twist of logic, implying that a less transparent legislature that relies on generalizations or even false information that all sex offenders pose a future danger—and therefore that fails to isolate “particularly” dangerous offenders as its target—may be effectively insulated from a due process challenge.

Third, the wide discretion given juries in imposing punitive damages creates an unusual risk of arbitrary and prejudicial enforcement.63 The Court sees the disfavored and vulnerable group as big business, whose deep pockets may be tempting targets for outsized punitive damages.64 The Court has interpreted disproportionate awards as evidence of such prejudice.65 But big business is hardly a singularly vulnerable group. Sex offenders are more unpopular, and their unpopularity is often based on untenable assumptions about their crimes, their psychology, or their future danger.66 Legislators pressured to be tough on crime face few drawbacks but many rewards for restricting released sex offenders. Like juries determining punitive damages, legislators operate mostly unchecked in passing SORs; for example, they may pass them retroactively and with no process beyond the original trial. This combination of legislative discretion and offender vulnerability may have contributed to the spate of carelessly drawn SORs that sweep in minor and serious sex offenders alike.

Fourth, SORs can be sufficiently punitive in nature to invoke proportionality as a traditional retributive limiting principle of punishment. If one characteristic unites most types of sanctions—both criminal and civil67—to which the Court has applied constitutional proportionality review, it is their punitive nature. In addition to punitive damages and punishment itself, the Court has applied proportionality to “punitive” forfeitures under the Excessive Fines Clause.68 Proportionality as a retributive limiting principle on the severity of punishment means that, regardless of the state’s penal aims, it cannot dole out its punishment to an offender except in proportion to his actual wrongdoing. For example, though deterrence may be a penal aim, it is limited in any given case by proportionality.69 To the extent SORs are punitive, they should be subject to retributive proportionality.

The Court evaluates whether a sanction is “punitive” using the Kennedy v. Mendoza-Martinez factors.70 These are neither exhaustive nor definitive, but in examining Alaska’s sex offender registration and community-notification laws in 2003, the Court stressed the question “whether, in its necessary operation, the regulatory scheme: has been regarded in our history and traditions as a punishment; imposes an affirmative disability or restraint; promotes the traditional aims of punishment; has a rational connection to a nonpunitive purpose; or is excessive with respect to this purpose.”71 Applying these factors, the Court found Alaska’s scheme nonpunitive. But the Court’s analysis emphasized the “minor and indirect” nature of the “disability or restraint,” which left sex offenders “free to move where they wish and to live and work as other citizens, with no supervision.”72 Compared to the current infringements of privacy and mobility that sex offenders face, the Alaskan scheme was mild.73

If we accept that an SOR imposes an “affirmative disability or restraint,” then the key remaining Mendoza-Martinez factors will be whether the SOR has been traditionally regarded as a punishment, and whether it rationally serves a punitive purpose or is excessive vis-à-vis its nonpunitive purpose. As other scholars have discussed, the types of disabilities imposed by SORs—including near-banishment and other social exclusion—have historically been used to punish in the United States.74 Perhaps authorities will argue that their SORs serve public-safety purposes, not punitive purposes. Excess should constitute strong evidence of at least a partial punitive purpose, for the excess will be harder to justify on grounds other than the “traditional aims of punishment,” retribution and deterrence. SORs may be punitive only in certain cases, for “a civil as well as a criminal sanction constitutes punishment when the sanction as applied in the individual case serves the goals of punishment.”75 This would be the case, for example, where the SOR applies to all sex offenders even though the regulatory purpose implicates only a subset of offenders. That an SOR, as applied, is sufficiently punitive to require proportionality review—itself merely a test, not a conclusion—does not mean it will be deemed punishment subject to other constitutional restraints.

As an example, residency restrictions that, banishment-like, all but exclude sex offenders from living in a city should be found punitive. These restrictions ban sex offenders from living within a certain distance—typically 1,000 to 2,000 feet—of a school, park, or other facility frequented by children. Given the density of children’s facilities, this restriction often effectively banishes sex offenders from all city real estate. Banishment has a long history as a form of punishment in the United States.76 Ostensibly, residency restrictions serve the nonpunitive purpose of keeping children safe from sexual predators. Yet they typically apply to many sex offenders other than those who have assaulted children. Banishing a person who streaked as a high school student77 is excessively burdensome given that it does nothing to protect children but imposes a heavy cost on the delinquent minor. When a law like this fails to achieve its safety objective in many individual cases, the only conceivable objective it serves in those cases is punitive. It is retributive, inflicting suffering on a person who has inflicted suffering on others; and its severity may deter, if sex offenders are aware that their acts could gravely disrupt their lives even after they have served their sentences.

A skeptic of my argument might contend that penalties issued by juries—but not, as in the case of SORs, by legislatures—uniquely warrant due process review for excessiveness.78 Distrust of juries appears to be a principal motivation in the Court’s proportionality review of punitive damages. But distrust of all actors may be prudent, given the broader concerns underpinning proportionality review. Fair notice, the primary justification for proportionality review cited in Gore, isas important for legislative penalties as for those imposed in civil trials. The other justifications, as I’ve explained above, apply in both contexts. Indeed, several of the early twentieth-century cases hinting at proportionality review for punitive damages involved statutorily set penalties.79 Moreover, it is simply not true that legislators command more deference than juries. Juries are highly respected in the American judicial system, and their pronouncements carry the enhanced due process guarantee of individualized consideration. Jury verdicts are generally upheld if supported by any “legally sufficient evidentiary basis,”80 a standard of review hardly more demanding than the rational basis sought in legislation.

III. applying bmw v. gore to sex offender restrictions

While Gore’s proportionality test was designed for punitive damages, it can easily be adapted to the SOR context. To determine whether a deprivation of liberty is proportionate, as due process requires, the Court should inquire whether it is “grossly excessive” to achieve the legitimate state objectives of punishment.81 To reiterate, the Gore “guideposts” for evaluating the excessiveness of punitive damages are (1) the reprehensibility of the misconduct, (2) the disparity between the punitive and actual damages awards, and (3) the difference between the punitive damages and other sanctions for like misconduct, in the same and other jurisdictions.

The first factor cleanly translates to the SOR context: where Gore examines reprehensibility of the misconduct, a court reviewing an SOR examines reprehensibility of the sex crime. This factor should cast suspicion on statutes that severely penalize petty criminals, such as public urinators, or that penalize minors for life.82

The second factor presents the greater challenge, for Gore compares the monetary value of harm the defendant actually or potentially inflicted with the money he may be forced to pay as punishment. The items compared are pre-quantified for the court. For SORs, the court must instead compare the harm that the sex crime inflicted on society with the harm that society may inflict in return. This will require quantification of harm, but courts are well-acquainted with this sort of analysis as part of their sentencing function,83 and anyway it surely would not prevent detection of extreme disparities.

Where an SOR is triggered by any sex crime conviction,84 the second factor also might require aggregation that is not necessary for punitive damages: the court should consider all of the crimes of the repeat offender; and, for consistent logic, all of the restrictions imposed on the offender by states and localities.85 This aggregation aligns with courts’ approach to punitive damages, which are often awarded in greater amounts where the defendant is more likely to recidivate or has already recidivated.86 Moreover, the interactive effects of these restrictions may need to be examined. For example, where overlapping residency exclusion zones exclude an offender from living in most parts of a city, that effect should be considered in the calculus.

The third factor translates nearly as cleanly as the first. Rather than looking merely to the severity of other sanctions for comparable misconduct, the court can compare among jurisdictions both the scope and severity of civil restrictions and criminal sentences for similar sex offenses. SORs might also be compared to penalties imposed for non-sex offenses of a similar nature. For example, SORs imposed on minors convicted under laws criminalizing sex between minors might be compared to penalties imposed for other juvenile offenses. This branch of the inquiry should undermine SORs drawn with an unusual degree of carelessness or imprecision.

Finally, proportionality review of SORs should include an additional factor: a comparison between the offender’s crime and the state’s purpose for the SOR. In almost every case, the state’s purpose will match the traditional penal purposes of deterrence, incapacitation, retribution, and, less often, rehabilitation. Wide disparities between the purpose and the type of offender targeted should constitute strong evidence that a sanction is disproportionate. Some SORs, though only slightly burdensome, may do little to nothing to advance the state’s purpose when applied to particular sex offenders, or to a particular class of sex offenders. For example, regulations that restrict sex offenders’ access to areas frequented by children will do little to prevent child predation when applied to offenders unlikely to victimize children—perhaps such as those who have only ever victimized adults. While resembling rational basis review, this analysis differs because it examines whether the law’s general justification holds as applied to this sex offender. The analysis would not require legislatures to use least-restrictive means, but would only permit courts to consider the overbroad scope of an SOR as one factor suggesting its gross excess.

While this extra factor is not one of the Gore three, it finds grounding in the Court’s discussion of the second factor, comparing punitive and compensatory damages, and in the Court’s application of retributive limiting principles to punitive sanctions. The Court emphasizes that excessive punitive awards are arbitrary and fail to serve the state’s punitive interest in retribution or deterrence.87 Presumably, awards would not be excessive if their full amount served both of the state’s legitimate interests. The question of excess in the punitive damages context is necessarily only a question of magnitude. For SORs, however, the excess can take the form of suffering of an unwarranted type. In rejecting the punitive damages award in Campbell, the Court observed: “[A] more modest punishment for this reprehensible conduct could have satisfied the State’s legitimate objectives, and the [state] courts should have gone no further.”88 That is an excellent slogan for reviewing sex offender restrictions for proportionality.

Let me close by offering a fairly typical sample case. Many cities in Texas have passed ordinances restricting the areas in which registered sex offenders may live. To take just one example, the city of Burleson’s ordinance forbids any sex offender required to register in Texas, whose victim was younger than seventeen years old, to “establish a permanent residence or temporary residence within 1000 feet of any defined premise where children commonly gather, including a playground, school, day-care center, video arcade facility, public or private youth center, park, or community swimming pool.”89 Given the Texas registration scheme, the ordinance applies to, among others, teenagers convicted of having consensual sex with other teenagers.90 An eighteen-year-old still in high school could be convicted under Texas law for having sex with his fourteen-year-old freshman girlfriend. As a result, he might be banned for ten years from living in certain Texas cities with tight residency restrictions and a high density of schools and public parks, like Burleson. 91

Under the Gore analysis, a court reviewing this law would (1) consider the reprehensibility of the consensual sex in this case; (2) compare the harm inflicted on society by that act to the severity of the sanction against the offender; (3) compare the sanction against consensual sex between minors here to other sanctions for the same infraction; and (4) determine whether the sanction serves the purpose of protecting public safety, which was arguably the purpose of the residency restriction.92 Such a court should conclude that consensual sex between teenagers is one of the least reprehensible sex offenses, insofar as it involves no violence or abuse, and inflicts minimal harm on society. The decade-long residential exclusion of the SOR is a penalty far exceeding the misdemeanor penalties in other jurisdictions for consensual sex between young people three or fewer years apart in age.93 Indeed, many jurisdictions do not outlaw similar acts at all.94 Finally, the penalty is ill-designed to protect the public from sexual predators of young children, given that teenagers who have consensual sex with a girlfriend or boyfriend do not seem uniquely subject to sexual temptation by very young children from their residential neighborhood.95 All factors support the conclusion that Burleson’s residency restriction, applied to this crime, is “grossly excessive” in violation of due process.


The Supreme Court’s serious due process concerns about punitive damages awards against large corporations—regarding fair notice and an ill-fitting, potentially prejudiced implementation process—apply just as powerfully to harsh civil regulations imposed on sex offenders after their criminal sentences. And any procedural inadequacies promise to take a heavier human toll for sex offenders than they do for big business. This Comment has proposed a straightforward doctrinal solution to these dangers for sex offenders’ due process rights: extending proportionality review, already applied to punitive damages awards, to sex offender restrictions. Proportionality principles should—and practically can—be applied to stem the tide of prejudicial and ill-informed burdens imposed indiscriminately on all sex offenders.

erin miller*