How to decriminalize drugs: the design features of a non-criminal response to the personal possession of drugs

Internationally, stakeholders across jurisdictions are considering alternative, non-criminal responses to the personal possession of drugs. Alternative responses include various decriminalization, diversion, and depenalization schemes. However, what details need to be considered? In this paper, we present the design features of non-criminal responses to simple possession and consider the implications of the choices available. The features include the reform architecture (objectives, decriminalization options, de jure or de facto approach), eligibility criteria (person-, place-, and drug-based criteria), and the actions taken (deterrence, therapeutic, and enforcement strategies). From this examination, it is clear that details matter – some can undermine the goals of reform altogether. We also show that the notion of ‘decriminalization’ is not a simple, unified framework. Rather, there are meaningful differences in policies and options available within a non-criminal response. The design features covered in this paper offer a framework that can be used to design future reforms and operationalize research on non-criminal responses to personal drug possession.


Introduction
For over 50 years, dozens of countries, states, and cities have implemented varying alternatives to criminal penalties for the personal possession of drugs. The Portugal model is perhaps the most well-known example, although 48 other countries and jurisdictions have removed and/or reduced criminal sanctions related to drug possession, including Mexico, Uruguay, Germany, Lithuania, Australia, Costa Rica, the Czech Republic, and Oregon, USA (Talking Drugs, 2020). Core to these reforms is a greater emphasis on public health, social justice, and human rights, alongside revenue, political, and pragmatic considerations (Global Commission on Drug Policy, 2014;Transform Drug Policy Foundation, 2015). Several reviews examine the impact of removing or downgrading criminal penalties across jurisdictions, highlighting "tremendous variation in the laws and regulations surrounding so-called decriminalization policies adopted by Western countries" (Pacula et al. (2005) p. 347) and internationally (Eastwood et al., 2016;Hughes et al., 2018;Scheim et al., 2020;Stevens et al., 2019;Unlu et al., 2020). In fact, 'decriminalization' is not a singular, uniform approach, but is instead a term that describes myriad legislative models. To simplify the legislative and/or regulatory schemes, Stevens et al. (2019) sort each model into one of three primary classes: decriminalization, depenalization, or diversion. Even so, these scholars and others point out considerable heterogeneity and a lack of articulation regarding the definitions and details under each class (Fischer et al., 2021;Stevens et al., 2019). Kilmer (2019)  These can be in terms of the number of people intervened upon (net widening, Cohen 1985) and the intrusiveness of the intervention imposed (net deepening, and/or mesh thinning, Cohen 1985). Net widening happens where a broader population comes under state control under the new regime; net deepening occurs when sanctions are unintentionally more severe for 'low or shallow end' offenders who, under the previous regime, would have received lesser punishment; mesh thinning is where alternatives make it relatively harder to get out of the criminal justice system. These consequences are demonstrated globally (Hughes et al., 2016;Roberts & Indermaur, 2006;Stevens et al., 2019). For instance, in Mexico, contact between police and people who use drugs increased as officers had greater leeway to arrest and charge individuals with administrative sanctions (Arredondo et al., 2018;Beletsky et al., 2016). Such consequences demonstrate the importance of good policy design.
The purpose of the current paper is to describe and examine the key design features of non-criminal responses to personal drug possession. We present the features, define the options, and consider their implications. The variety of options available are organized into three categories, below: 1) Reform architecture; 2) Eligibility criteria to which the scheme applies; and 3) Actions taken upon detection of drugs for personal use (see Table 1).

Designing a non-criminal response for simple drug possession
[ Table 1 here]

Reform architecture
The design of a non-criminal response to simple possession will be crucially shaped and/or constrained by its reform architecture: the reform objectives or goals, the legal framework, and the approach taken.
Reform objectives: What are the goals of the change?
In theory, changes to drug policy should start with a clear set of objectives (Reuter, 2013;UK Drug Policy Commission, 2012). Well-defined objectives promote evaluation, accountability, and implementation. The goals of reform can be numerous: to prevent new or compounded criminal records; to limit contact with the criminal justice system, including police officers; and/or to minimize the resources and cost to the criminal justice system and society. Another goal may be to connect people who use drugs with healthcare, treatment, social services, and/or education. In any case, defining reform objectives is a crucial first step so reform decisions can logically and pragmatically follow.

Legal framework: Will the offence remain criminal?
Depending on the objective, a key design consideration is whether or not simple possession will remain a criminal offence in law. This legislative decision will determine the necessity for and availability of most subsequent policy options.
If simple possession remains in criminal legislation but removal of punishment is a goal, then model features must be specified, including eligibility criteria, actions taken, role of police, and potential penalties. Given that criminal penalties are still available for use (but not applied in some cases), this response is known as depenalization (Stevens et al., 2019;Stevens et al., in press).
Alternatively, if simple possession is removed from criminal legislation, new penalties can be added (or not). In some cases, civil penalties can be used in lieu of criminal ones. If neither a criminal nor civil offence applies, features such as prosecutorial and/or police discretion or diversion, are similarly inapplicable. This latter model is a gold standard amongst organizations of people who use drugs and many health professionals as it eliminates punitive drug policies and permits a more health-and rights-based approach (International Association of People Who Use Drugs (INPUD), 2021; Scheim et al., 2020;Shane, 2020). A model that does not introduce new penalties or sanctions is considered full decriminalization (Stevens et al., 2019).

De jure or de facto: Will the model be in law or procedure only?
The distinction between decriminalization and depenalization underscores differences between reforms that occur in law (de jure), or changes that occur only in practice or procedure (de facto). These two approaches differ significantly in terms of how reforms unfold.
De jure reforms tend to be more entrenched or longer lasting as they occur in law.
They usually allow for less discretion, although 'street-level bureaucrats' (i.e. police) can be adept at continuing old practices despite legal changes (Lipsky, 1980). De jure reforms can also be more conservative and less responsive to people who use drugs' needs and experiences, in part because the legislative process is slow-moving, rigid, and averse to change.
De facto approaches are relatively more adaptable and easier to implement as they rely on practical and procedural non-enforcement of the criminal law. For instance, policing policies may advise officers not to arrest or charge for possession, such as in Denmark from 1969to 2004(Houborg et al., 2020, or prosecutorial guidelines may advise limiting the pursuit of criminal charges, as in Canada and the Netherlands (5.13 Prosecution of Possession of Controlled Substances Contrary to s. 4(1) of the Controlled Drugs and Substances Act, 2020; Talking Drugs, 2020). However, de facto approaches leave considerable interpretation and discretion to various actors, creating vulnerability to inequitable application and uncertain outcomes for people who use drugs.

Eligibility criteria to which the scheme applies
In any model, decision-makers must determine eligibility criteria to delineate the circumstances in which the new regime applies. Criteria include details relating to the person (age, population, previous and concurrent offending), place, and drug (drug type, threshold quantities).

Age: What age group(s) does the model apply to?
In most non-criminal schemes, there is a difference in the treatment of adults and minors. In lieu of criminal penalties, responses to drug possession among minors include enforced confiscation of drugs or a therapeutic response (McClean et al., 2020). For example, in the US Virgin Islands, minors found to possess under two ounces of cannabis must attend an education class and failing to complete it can result in criminal penalties. Internationally, some schemes have explicitly excluded minors from a non-criminal response, whereas others specify circumstances where policies apply to them. Up until 2004 in Denmark, for example, several police forces enacted de facto schemes to avoid arresting and charging young people for simple drug possession, even where these criminal justice responses were still applied to adults (Houborg et al., 2020;Hughes et al., 2018).

Population: Which specific groups does the model apply to (or not)?
In some jurisdictions, reforms have only applied to certain groups. For instance, in 2013, Costa Rica amended Law 8204 to reduce and deliver alternative sentences for women in prison who are experiencing poverty, are elderly, disabled, and/or caregivers (Eastwood et al., 2016). In contrast, prosecutorial guidelines in Canada (reflecting a de facto depenalization model) generally limit prosecution to "the most serious manifestation of the offence", for example where offenders are a public or peace officer (5.13 Prosecution of Possession of Controlled Substances Contrary to s. 4(1) of the Controlled Drugs and Substances Act, 2020, p. 1). Police policies may also encourage officers to use greater discretion with certain groups, such as Indigenous peoples in Canada.
In any case, it is critical to consider how reform decisions may affect some groups differently. For instance, even small criminal citations and low-level drug arrests can lead to deportation and/or compromise immigration status (Drug Policy Alliance, 2020).

Previous and concurrent offending: How will previous offences be handled?
Individual responses to simple possession may depend on whether the person has previous offences and/or been found in possession of a drug more than a specified number of times. Pacula et al. (2005) call this 'partial depenalization' -an approach adopted by several countries, including Germany. In some Australian and US states, first time offences may accrue civil penalties whereas repeat offences may incur criminal charges. For instance, under the English cannabis and khat de facto warning schemes, first-time possession is met with a police warning, second-time with a fixed on-street fine (a 'Penalty Notice for Disorder'), and third-time with a criminal justice response (formal caution or charge) (Hughes et al., 2018).
If drug possession is no longer a criminal offence, another consideration is whether previous convictions will be expunged retroactively, as occurs in Costa Rica and some US states. In many places, expunging criminal records requires a legislative process, including defining the parameters of previous offences to which the reform will apply (Quinton, 2017).
Another decision is whether possession will be considered a criminal offence when it occurs alongside concurrent offending. For example, when drug possession occurs alongside other offices such as driving, weapon, or assault offences (which in turn may or may not be related to possession of drugs).

Place: What locations/context does the model apply to?
Another design detail is to specify geographic locations and contexts that the model will apply to. For example, drug possession may face criminal penalties in the context of prisons, schools, and certain workplaces, or within a defined proximity to these locations.
Similar to some legal psychoactive substances such as alcohol, public drug use is commonly prohibited, even under non-criminal regimes, as in Jamaica, Spain, and Washington, DC (Eastwood et al., 2016;Stevens et al., 2019). Retaining the prohibition in specified places can perpetuate the criminalization of drug use (and its adverse effects), since police are enabled to arrest for low-level drug offences (Levine, 2009)a policy that can inequitably impact people using drugs in public spaces, such as people living in poverty and/or experiencing homelessness.

Drug type: Which drugs are included?
A significant decision is which drugs are subject to a non-criminal response. Countries such as Portugal, Germany, and the Czech Republic have removed criminal penalties for small quantities of all drugs, whereas several US and Australian states only include cannabis in their reforms. It is possible to define applicable drugs based on the class of drug, such as opioids, amphetamines or benzodiazepines, or specify the drugs themselves, such as crack and powder cocaine, or fentanyl, heroin, and opium.
Applying reforms to some drugs and not others can create a complex system. Some countries have looser definitions, such as the Czech Republic which codifies cannabis alongside 'plant drugs' (Belackova & Stefunkova, 2018). Countries typically rely on drug schedules to define which drugs are included or excluded. The presence of fillers or cutting agents in certain illegal markets (for instance fentanyl or fentanyl analogues in most North American street markets) may also impact the ability to define or categorize applicable drugs.
Another decision is whether the model applies to the presumed drug, actual chemical substance, or precursor of the drug, found in possession. In most jurisdictions that take a non-criminal response to simple possession, the drug in possession is based on officer judgement rather than laboratory testing. However, it is unclear how the law applies for multiple or unknown substances, particularly when the drug content is unknown to the possessor or officer assessing it.
Other countries have treated so-called 'hard' drugs differently from 'soft', as in the Netherlands and Italy (Eastwood et al., 2016). This superficial distinction perpetuates inequities. People with higher risk profiles, who are most in need of support, are more likely to be criminalized.
There may be other unintended consequences to limiting the scope of drugs/drug classes included in the model. For instance, consequences may arise in the drug market such as changes in use and/or supply and distribution. Differences in the treatment of various drugs may encourage people to switch from one to another.
British prisoners switched from using cannabis to synthetic cannabinoid receptor agonists (known as 'Spice' or 'Mamba') during a period when cannabis use was detected in urine tests and punished, while use of synthetic cannabinoids was not (Ralphs et al., 2017).
Threshold quantity: What is/are the amount(s) of drugs that defines personal possession?
Many, but not all, non-criminal responses define the drug quantity (e.g., weight of powder, number of pills) of which possession is deemed a non-criminal offence. The specified amount is commonly referred to as a threshold quantity (TQ) or limit. TQs can be based on the total weight of the substance, including any non-psychoactive substances in that weight, or only include the pure drug. Nearly all jurisdictions that employ a TQ use the former, weight-based approach.
International examples of TQs for personal possession vary dramatically (European Monitoring Center for Drugs and Drug Addiction, 2015). For example, the TQ for cocaine in Lithuania is ten times higher than that in Norway. In Germany and the USA, TQs differ between states. In other countries such as Uruguay, Denmark, and Spain, no TQs exist; rather, a supply offence only exists where there is evidence of intent to supply. In some jurisdictions, TQs are ambiguous, defined as 'more than small amounts', as in the Czech Republic, Costa Rica, and Poland (Belackova & Stefunkova, 2018;Talking Drugs, 2020). Undefined TQs are vulnerable to interpretation, discretion, and bias from police officers and prosecutors.
It is possible to have multiple TQs that differentiate between low-and high-level personal possession. As proposed in one Australian jurisdiction, neither are supply offences but the former is a non-criminal response whereas the latter a criminal one (see Ritter et al., 2021).
TQs may also be specified as binding or as indicative: binding TQs create a strict application of the law or penalty whereas indicative TQs are merely a guideline (Talking Drugs, 2020). In the latter case, officers or prosecutors may use their discretion to not charge. Where possession is no longer a criminal offence, officers do not have discretion to charge below TQs.
Setting TQs is a critical decision point. If appropriately set and applied, TQs can limit biased application of police discretion. If set too low, however, TQs can heighten enforcement and result in net widening. Low TQs can disproportionately impact marginalized groups, including people who are racialized and people experiencing poverty or psychiatric distress (International Network of People who Use Drugs, 2021).
Other potential harms associated with low TQs include incentivising interactions with the unregulated market as people try to avoid criminalization by frequently purchasing smaller (below-TQ) amounts. As well, the concentration or potency of drugs may increase to keep them below a threshold weight -potentially increasing overdose risk for certain drugs.
One option is to use an 'average' amount consumed for personal use, as in Portugal, where the criminal law applies to people who possess amounts above ten times the median daily dose (Hughes & Stevens, 2010;Rêgo et al., 2021). However, 'averaging' assumes drug consumption is relatively homogenous across people. The very nature of setting TQs does not enable consideration of diverse patterns of personal use, mobility, body composition, socioeconomic conditions, and geography. For example, people who live in rural/remote communities may need to possess a large quantity of drugs given the inaccessibility of drugs in their location.

Actions taken upon detection of drugs for personal use
Being found with illegal drugs can be met with a number of actions -or not. In the absence of criminal penalties, governments may choose to not define additional actions towards personal possession. Alternatively, they may introduce administrative or civil sanctions/penalties (Stevens et al., 2019). Globally, most reforms have produced policies that elicit actions under a deterrence, therapeutic, or educative logic. Several options, in any combination, are available for reform (Table 2).

Non-criminal disciplinary actions: What sanctions can be imposed on the possessor?
Some models include non-criminal disciplinary action for simple possession even if the criminal offence is removed. Notably, although these 'deterrent' actions are not criminal penalties, they can be punitive, discriminatory, and serve as a gateway to future engagement with the criminal justice system. Jurisdictions that employ a deterrence regime of administrative sanctions include Portugal, the Czech Republic, Australia, the USA, and Jamaica (Hughes et al., 2018;Stevens et al., 2019).
There are several potential consequences to deterrence strategies. Administrative

Diversion and referral programs: Will individuals be re-directed or diverted into other systems?
Non-criminal diversion programs redirect individuals away from criminal justice processes into social, educational, or health programs. If the reform objectives include providing greater health support, treatment, and social care, then diversion actions may be preferred.
A key decision is defining the program that people who use drugs are diverted to.
Programs include healthcare and treatment, education, and social services. Decisions about who may be diverted where, and their degree of agency, is key. The decision may be entirely voluntary and with the consent of the possessor, or rest with healthcare providers, police officers, or committees.
Another important design feature for diversionary approaches is the level of compulsion or coercion attached to such diversion. In places such as Norway, failure to comply with a police order for mandatory counselling results in a fine or financial penalty (Government of Norway, 2021). Even seemingly voluntary programming may include elements of coercion or be constrained, as in jurisdictions where 'choosing' the program is a mitigating factor for sentencing. Others note that coercive and punitive diversion schemes can be as harmful as criminalization (Eastwood et al., 2016).
Nevertheless, the absence of a criminal record can reduce reoffending and improve future employment prospects, even when diversionary measures are similar to probation following criminal conviction (Mueller-Smith & Schnepel, 2021).
In some diversion models, police officers are both the first point of contact with people who use drugs and also act as a broker for health and social services, which they may not be trained to do (Goetz & Mitchell, 2006). Over-reliance on police can result in avoidance strategies among people who use drugs, as seen in Mexico where police diversion programs negatively impacted health and social outcomes (Beletsky et al., 2016). Alternatively, police may refer people to another triage system, such as the dissuasion commissions in Portugal which include lawyers, psychologists, and social workers. However, this model too has been criticized by people who use drugs (INPUD, 2021).
Diversion can result in net widening by increasing the number of people who are brought into the criminal justice system who may otherwise have been ignored by police officers and who may not need or want treatment (Hughes et al., 2018;Roberts & Indermaur, 2006). For instance, in the UK, large numbers of people who used cocaine recreationally were referred for mandatory assessment following the 'testing on arrest' policy (Connor et al., 2020).
While there are a number of limitations and potential adverse consequences of diversion programs, the common aim of diversion programs is to reduce the burden on the criminal justice system and the criminalization of people who use drugs (Hughes et al., 2018). With an adequate design and triage system, diversion programs can provide additional pathways into health and social supports for people who need it.
Internationally, diversion is a widely used intervention for personal possession offences (Stevens et al., 2019).

Police discretion: How flexible is enforcement of laws and/or administrative sanctions?
Decisions must be made about the discretion or flexibility that police officers have in enforcing laws or sanctions. If the objective of designing an alternative model is to reduce the burden on the criminal justice system and remove criminal penalties, police discretion is highly relevant. Schemes that do not formally remove criminal penalties for simple possession often must rely on police discretion and/or officer interpretation of depenalization policies. Alternatively, if criminal penalties are removed from the law, there is less scope for police discretion -although, even in several US states that formally decriminalized drug possession, police retained considerable autonomy in their decision to arrest for drug offences (Logan, 2014;Pacula et al., 2005). This outcome may be partly because officers can use other offences (e.g., possession in public view) to target people who use drugs.
Police discretion can be viewed from two perspectives. On the one hand, discretion creates an opportunity for leniency and decreases penalties and burden on the criminal justice system. On the other hand, the subjective nature of discretion can produce inequities. Police discretion is vulnerable to inconsistent, unpredictable, and discriminatory application (Greer & Ritter, 2021). In some cases, it can stunt implementation of the de facto regime altogetherparticularly if there is no top-down leadership or guidance on the purpose and benefits of non-enforcement . Consequently, eliminating or minimizing police discretion may abate the problems associated with it: racialized profiling, inequitable application of the law, and the stress and negative impact that discretion can have on both people who use drugs and police officers.

Confiscation and destruction of drugs: Can drugs be seized and destroyed by the police?
One key detail in the design of non-criminal responses for simple possession is whether police will retain powers to seize and destroy drugs. In most jurisdictions with non-criminal models for simple possession, seizing and destroying drugs occurs irrespective of whether the offence remains in law (Talking Drugs, 2020). Some schemes, such as the UK Psychoactive Substances Act, allow for the retention and destruction of substances, even though simple possession (outside prison) is not a crime. Confiscation of drugs is similarly routine in several other jurisdictions where personal possession is no longer a criminal offence, such as Costa Rica, Croatia, Switzerland, Italy, and several US states. However, if there is no administrative or criminal penalty for possession, seizing and/or destroying personal drugs by police or others may not be legal.
Confiscation and destruction of drugs can produce unintended consequencestherefore, it is an important decision point. As many people who use drugs have pointed out, drug confiscation can produce harms, including increased violence and crimes associated with replacing destroyed or confiscated drugs (INPUD, 2021;Shane, 2020).

Responses to non-compliance: What happens if individuals do not comply with mandated actions?
If possession is no longer a criminal offence and alternative sanctions are not introduced, compliance is not a design issue. However, if personal possession is still subject to administrative or criminal penalties, a key model feature is the response to non-compliance with these sanctions. Details include whether the original criminal offence can be re-instated due to noncompliance, or if other penalties can apply, including alternative criminal penalties (non-drug offences) or actions such as fines.
Non-compliance measures for administrative or civil penalties can reflect a deterrence aim with the view that consequences for non-compliance upholds the administration of justice and act as a disincentive to use or possess drugs. However, non-compliance measures may introduce inequities and negatively impact marginalized groups, such as those experiencing homelessness or financial insecurity. Mandatory fines pose distinct harms for marginalized offenders, including disproportionate financial consequences, a pronounced threat of incarceration, targeting by collections efforts, and an "indefinite sentence" for offenders who are unlikely to ever be able to pay. (R. v. Boudreault, 2018). The Australian expiation system negatively impacted people who could not comply with the scheme due to financial difficulty (Humeniuk & Drug and Alcohol Services Council, 1999;Single et al., 2000).

Discussion
As detailed, there are a myriad of crucial decisions and options in the design of noncriminal responses to simple possession. Given the substantial policy attention that these alternative models (such as decriminalization, depenalization, and diversion) have received, it behooves decision-makers to carefully consider the design options.
The task of reform is not achieved by mere support for removing criminal penalties.
The real work is in the process of designing actual reforms, which starts with setting clear reform objectives, along with eligibility criteria, specifying the actions taken (or not), and deciding whether a deterrence, therapeutic, and/or enforcement strategy is involved. Importantly, the features in each of these categories are not neutral or benign decisions but can have unintended consequences. At a time that is pivotal for legal reform in many jurisdictions, it is imperative to progress a reform agenda that takes these details into consideration.
In outlining the options available in the design of a non-criminal response to criminal penalties for simple possession, we have offered a range of examples that highlight the importance of clearly defined and carefully designed models. Vague or confusing reforms can result in poor implementation and uptake. Poorly designed reforms can also influence the interventions that follow and associated positive or negative consequences. Importantly, under all options, there are a number of equity issues to consider, including the disproportionate impacts on people experiencing homelessness, racialized people, and immigrants (Drug Policy Alliance, 2020; INPUD, 2021).
Clearly designed and defined reforms are paramount for understanding the impacts of the policy changes. Pacula et al. (2005) notes that "'decriminalization' does not mean what researchers and policy analysts think it means" (p. 25), pointing to the impossibility of truly understanding the differences between models when the nuance and details within models is obscured yet required to evaluate them. Importantly, the design considerations provided in this paper can be used to operationalize future reforms and study various decriminalization, depenalization, and diversion models. As well, we hope to have demystified the ubiquitous use of the word 'decriminalization'.
This term is not a simple, unified framework; rather there are meaningful differences in the policies, options available, and nuances between non-criminal responses, including decriminalization.
In this paper, we have briefly addressed the multitude of design features and lessons learned from jurisdictions that have introduced varying degrees of non-criminal responses to personal possession. Those examined in this paper are not necessarily exhaustive and the lessons learned from decriminalization, depenalization, and diversion schemes are not limited to those covered here (for example, see: Stevens et al., 2019;Unlu et al., 2020). We also do not suggest one 'ideal' model to adopt from the key design features presented here. The sociolegal context in which policies and laws are situated are unique; they each come with unique barriers and circumstances (Unlu et al., 2020). This context-specific nature of policy reform and implementation means that there will always be some degree of uncertainty of the outcomes; "policy expectations define the means, the context mainly shapes the ends" (Unlu et al., 2020, p. 5). Given that evidence is still limited in many of these decision areas, evidence of the effectiveness of some interventions remains limited. What remains clear, however, are the inequitable harms that exist from current punitive models.
Finally, given that the aim of this paper was to only outline model design decisions, we did not cover the process of policymaking itself but need to underscore its relevance and importance. Key stakeholders and experts in any drug reform are the people directly impacted by subsequent policies and interventions -people who use drugs.
However, the drug policy design process often happens without significant engagement with the affected community, which can result in decisions and interventions that lack relevance, applicability, acceptability, and equity (Greer et al., 2016;Greer & Ritter, 2021;INPUD, 2021). By ensuring the inclusion and involvement of many people who use drugs in the design, development, and implementation phases of reform, it is possible to mitigate some of the issues that can arise from poorly designed policy.