A criminal register that never expires is not a public safety instrument. It is a permanent placement. And a permanent placement is the state’s way of saying that the question of who a person is has already been answered.
The principle is simple, and it is the principle every serious legal system outside the American default already accepts: apart from the most serious crimes with an identifiable victim—murder, rape with violence, modern slavery, weapons of mass destruction, terrorist violence, robbery with violence—the register should expire. Not automatically, and not on a timetable alone. It should expire when two conditions are met. First, the person has served the sentence. Second, an independent expert, requested by a competent authority, has assessed the likelihood of recidivism as low and has confirmed that the person understands why the act was an offense and has demonstrated a commitment to non-recidivism. The assessment is not a character reference. It is a risk determination, made by someone with no stake in the outcome except accuracy.
This is not leniency. It is the only arrangement consistent with what the evidence actually says about desistance.
The Evidence of Desistance
Redemption research—the study of the point at which a criminal record stops predicting future offending—is not a marginal literature. It is one of the most consistent findings in criminology. The risk of reoffending declines steadily with time since the last contact with the criminal justice system. U.S.-based research funded by the National Institute of Justice has established that an individual’s propensity to commit a future violent crime decreases as crime-free duration increases; an individual with a prior violent conviction who has been crime-free for twenty years is less likely to commit a future crime than one who has been crime-free for ten (Moses, 2015, citing Blumstein). Redemption scholars estimate that after an average of seven to ten years without a new arrest or conviction, a person’s criminal record essentially loses its predictive value (Moses, 2015).
The same pattern holds for sexual offenses in U.S. data. For registrants, recidivism risk drops by roughly fifty percent every five years of remaining offense-free in the community, regardless of initial risk level; if the risk was ten percent at release, it is five percent after five years and 2.5 percent after ten (Hanson, cited in ACLU of Michigan, n.d.). A large-scale study of nearly 9.5 million people with convictions across seven U.S. states—including Texas and North Carolina—found that in five of those states, close to four in five people had not been convicted again within a decade; in the two highest-reconviction states, the figure was still two-thirds. More than half of the people in the dataset had a five percent or smaller chance of being convicted of anything in the coming year (RAND, 2026, describing the study published in Statistics and Public Policy).
These are not abstractions. They are the empirical basis for a policy: the register should be keyed to risk, and risk is dynamic. A register that does not expire is a register that has decided, in advance, that the evidence of desistance does not matter.
The Registry as the Exception, Not the Rule
The American sex offender registry is the clearest case of a system that has abandoned the risk principle. It includes everyone a U.S. court has convicted—whether factually guilty or later found innocent—because it is keyed to the conviction, not the truth. It includes people who are not at risk of reoffending. It includes people who committed the offense while themselves a victim, while underage, or as an aggregate of circumstances they had no control over. And it does not include those not yet convicted who are at risk of offending, nor those who have used their status and connections to escape judgment.
The registry’s defenders sometimes say it is a public safety tool. The evidence says otherwise. Studies verify that between 95 and 96 percent of new sexual crimes are committed by persons with no previous sexual crime conviction and therefore not on a registry (Sandler, Freeman, & Socia, 2008; Prescott & Rockoff, 2008). Law enforcement cannot effectively monitor high-risk registrants because the registry is bloated with low-to-no-risk registrants. A registry that pins the label on people for decades, and usually for life, endorses the mistaken view that they are irredeemably defective—a view that has no empirical support. The registry is not a risk instrument. It is a placement instrument. It sorts by legibility, not by danger. And the placement is permanent because permanence is the point.
The European systems start from the opposite default. In France, judicial rehabilitation (réhabilitation judiciaire) erases convictions from the criminal record, and legal rehabilitation operates automatically after specified periods depending on the sentence imposed (Service-Public.fr, 2025; Légifrance, Articles L5513-3 and L5533-1). In Germany, the Federal Constitutional Court’s Lebach judgment of 1973 grounded Resozialisierung—resocialization—in the constitutional protection of human dignity; the Court held that broadcasting a documentary identifying a perpetrator after his release would endanger his reintegration, and that the protectability of the perpetrator’s personality right increases with the passage of time since the offense (Bundesverfassungsgericht, 1973). The penal system is oriented toward reintegration, not permanent exclusion. Permanence is the exception, and it must be named, justified, and legislated.
The principle is not that every offense disappears. The principle is that the register expires when the risk expires, and that the risk is assessed by someone competent to assess it.
Immigration and the Character Test
Immigration law is where the permanent register does its most visible damage, because it converts a spent conviction into an exclusion that the ordinary legal system has already closed.
Australia’s Migration Act contains a character test under section 501. The test is discretionary, and Ministerial Direction No. 110 explicitly requires decision-makers to consider the risk of reoffending and evidence of rehabilitation achieved since the last known act (Direction No. 110, paras. 8.1.2, 8.1.2(2)(b)). Canadian law is even clearer: rehabilitation is “forward-looking,” and the “most important” or “primary” inquiry is the likelihood that the individual will reoffend (Pjetrushi v. Canada, 2026 FC 673, para. 2, citing Hadad). The Canadian Federal Court has held that a decision-maker acts unreasonably when it dismisses evidence of rehabilitation as “self-serving” without justification and ignores the forward-looking question (Pjetrushi v. Canada, 2026 FC 673, paras. 2, 6).
In practice, the discretion is often overridden by the freeze. Australia’s Direction 110 acknowledges the relevance of ties, time in the community, and rehabilitation, but the structure of the direction places the protection of the Australian community as the first primary consideration, and the seriousness of the conduct can outweigh mitigating factors (Direction No. 110, para. 8.1). When the Minister acts personally, the direction need not be followed at all, and there is no merits review. Spent convictions, which the ordinary law would ignore, are treated as still relevant. A person who has led an upstanding life for fifteen or twenty years can have a decades-old conviction cited as evidence of failing the character test.
This is the same mechanism described throughout this series: the law contains a rehabilitation logic, and the logic is real for the ordinary person, but for the white crow the law is read through the frozen image rather than the image being corrected by the law. The character test is not applied to the person who has served the sentence and demonstrated desistance. It is applied to the person as they were at the moment of the offense. The test’s own forward-looking structure is bypassed by a backward-looking placement.
The Protectionist Alibi
If a state wants to restrict immigration for protectionist reasons—to shield domestic labor, to manage adjustment costs, to ration access to public goods—it should say so. Protectionism is a legitimate policy choice, and it can be debated on its merits. What is not legitimate is to dress protectionism in the vocabulary of criminal law.
The economics is straightforward. International labor mobility confers gains by improving the global allocation of resources. Immigration quotas and entry limits are calibrated to manage adjustment costs, and those costs are real: labor market integration, congestion of public goods, and social adjustment. A state that wants to limit immigration for these reasons is making a protectionist decision. The honest framing is protectionist.
The dishonest framing is the one that uses a criminal record as a proxy for risk when the risk has already been assessed and found low. The trade and immigration literature has long noted the paradox: labor-scarce economies have generally maintained open trade and restricted immigration, even though “importing labor-intensive products is pretty much like importing labor” (Hatton & Williamson, 2006). By the same logic, an immigration policy designed to protect domestic labor will be undone by free trade; the desired effect will only be achieved by restricting both trade and immigration (Hatton & Williamson, 2006). This is a debate about economic policy, not about criminal danger. When a state excludes a rehabilitated person on character grounds, it is not protecting the community from a risk. It is using the criminal law as an alibi for a protectionist choice it does not want to defend on its own terms.
The result is a foreign and trade policy that is incoherent. If a state wants to restrict labor mobility, it should restrict it through quotas and labor-market rules that are transparent and debatable. If it wants to liberalize trade, it should recognize that the free movement of goods and the free movement of people are governed by the same underlying logic of allocation and adjustment. Using criminal records as an arbitrary barrier does neither. It signals that the state’s immigration policy is not about risk, and it signals that the state’s trade policy is not about efficiency. It signals that both are about the placement of the unplaceable.
The Constructive Alternative
The constructive alternative is a register that expires on evidence, and an immigration system that separates risk from protectionism.
On the register: the default is expiry. For the most serious crimes, permanence is an exception that the legislature must name and justify. For everything else, the register expires when the sentence is served and an independent expert has assessed the risk as low and the understanding as genuine. The expert is requested by a competent authority, not by the person seeking expungement. The assessment is forward-looking. It asks whether this person is likely to reoffend, not whether the original offense was serious. The seriousness of the offense is relevant to the sentence, not to the permanent placement.
On immigration: the character test should be a risk test, and the risk test should be satisfied by evidence of desistance and rehabilitation. If a state wants to exclude a person for protectionist reasons, it should do so under a protectionist provision, with the economic rationale stated openly. The trade and immigration policy should be coordinated, not because coordination is a virtue in itself, but because the current arrangement—open trade and closed immigration, with criminal records doing the work that protectionism will not admit—is incoherent and unjust.
The debt that cannot be discharged is not a debt to the victim. It is a debt to the category system. The category system needs the white crow to remain placed, and the permanent register is the mechanism that keeps the placement in force. The alternative is to let the evidence do the work. Risk declines. Desistance is real. The question is whether the state is willing to be governed by the evidence, or whether it prefers the permanent placement because the placement is the point.
References
ACLU of Michigan. (n.d.). Plaintiffs’ statement of material facts (No. 123-1), ¶¶ 179–180. United States District Court, Eastern District of Michigan.
Bundesverfassungsgericht. (1973). Lebach, BVerfGE 35, 202 (1 BvR 536/72).
Direction No. 110. (n.d.). Visa refusal and cancellation under section 501. Australian Government, Department of Home Affairs.
Hatton, T. J., & Williamson, J. G. (2006). A dual policy paradox: Why have trade and immigration policies always differed in labor-scarce economies? Weatherhead Center for International Affairs, Harvard University.
Moses, M. C. (2015). Research notes: Redemption research and offender employability. National Institute of Justice, U.S. Department of Justice.
Pjetrushi v. Canada (Citizenship and Immigration), 2026 FC 673 (Federal Court of Canada).
Prescott, J. J., & Rockoff, J. E. (2008). Do sex offender registration and notification laws affect criminal behavior? 54 Journal of Law and Economics 161.
RAND. (2026, October 15). A better way to use criminal records in hiring. RAND Corporation.
Sandler, J. C., Freeman, N. J., & Socia, K. M. (2008). Does a watched pot boil? A time-series analysis of New York State’s sex offender registration and notification law. Psychology, Public Policy, and Law, 14(4), 284–302.
Service-Public.fr. (2025, October 20). Can a conviction be erased from the criminal record? République Française.
Statistics and Public Policy. (2026). Study of nearly 9.5 million people with convictions across seven U.S. states.