CHAPTER V THE IMPACTS OF CANNABIS POLICY REFORMS WITHIN THE CURRENT DRUG CONTROL REGIME INTRODUCTION
This chapter reviews the existing evidence on the impact on cannabis use and other indicators of the alternative regimes of cannabis control which have been implemented within the current constraints of the international drug conventions. As noted in the preceding chapter, in a number of countries, at both a national and sub-national level, governments have introduced changes to the policies and laws applying to cannabis.
The reforms that are generally agreed to have been conducted within the bounds of the existing international drug treaties and conventions (e.g. Krajewski, 1999), can be broadly seen as moving away from prohibition with strict criminal penalties on the user (full prohibition) to some form of depenalisation. As described in detail in the previous chapter, the reform typologies include: prohibition with cautioning or diversion; prohibition with civil penalties (often termed ‘decriminalisation’); partial prohibition, including both ‘de facto’ and ‘de jure’ legalisation; and regulated availability of cannabis as a medicine, often referred to as ‘medical marijuana’. While not all of these reforms have been extensively evaluated, there is a small but growing evidence base, and this is the focus of the present chapter.
CAVEATS AND DATA LIMITATIONS
Caution needs to be exercised in interpreting the published evidence evaluating the impact of cannabis policies for a number of reasons.
Firstly, the policy environment is a dynamic one where effects decay and the policy that is implemented changes over time (see Pacula et al., 2005). This is important from the point of view of single longitudinal studies, as results may be affected by formal legislative and procedural changes as well as informal law enforcement practices (see Kilmer, 2002). When trying to make sense of the results of
comparative reviews, one needs to clarify the cannabis law and its implementation in a location at the time that the review was conducted.
Secondly, international comparisons are difficult and results can be confounded by cultural, political, geographic and climatic differences.
Thirdly, cannabis law reform often occurs in locations with high rates of use. This means that pre-post or longitudinal designs with ‘matched’ control locations are needed to identify true impacts on rates of cannabis use. Simply creating a dummy ‘decriminalised’ variable’ and comparing rates of use in so-called ‘decriminalised’ versus ‘non-decriminalised’ states, without adjusting for pre-change rates of cannabis use, will run the risk of erroneously concluding that ‘decriminalised’ states have higher rates of cannabis use because of the policy change. Causation may in fact go in the other direction: policy may have been liberalised precisely because of high rates of use and experiences of people being prosecuted under the law.
Fourthly, evaluations rarely take into account the level of knowledge in the community about the laws which apply to cannabis (Pacula et al., 2005). Levels of knowledge may vary in the general community, in occasional and in regular cannabis users in ways that may influence how changes to the law affect rates of cannabis use.
Finally, any research evidence may not predict the effects of new reforms in other locations, because the impacts of future cannabis policy reforms may depend on contextual factors and on how the reforms are implemented. Therefore it remains important that any changes to cannabis policy are evaluated, and their effects monitored and used to review their performance.
To date, research has focused on three outcome domains:
(i) General deterrence effects – the impact of changing the law on rates of cannabis use in (a) the general community; and (b) among the young - who are seen as the most vulnerable to any adverse health (and particularly mental health) effects of regular cannabis use (Arseneault et al., 2002; Coffey et al., 2002; Lubman et al., 2007; Smit et al., 2004; Winters & Lee, 2008).
(ii) Specific deterrence effects - the impacts on the cannabis use of those who have been apprehended.
(iii) Adverse social impacts of the system of control on apprehended users. Few studies have looked at: impacts on harms, the drug market or sentinel groups such as regular users, who may be more likely to show adverse effects that are less evident in general population studies.
In this short review we have refrained from going to original sources of data such as prevalence statistics and attempting to draw conclusions about what they can tell us about the impact of cannabis reforms or otherwise. Given the heterogeneous patterns of reform, the difficulties in identifying implementation practices, and the existence and impact of extraneous non-legal factors, we have limited ourselves to considering those examples of cannabis law reforms where detailed studies have been conducted by authors with knowledge and insights into the range of country- and region-specific issues which need to be taken into account when trying to understand the impact of such reforms.
EVIDENCE ON IMPACTS OF REFORMS
Most of the published research has evaluated the impact of moving from strict prohibition to prohibition with civil penalties. This has primarily been by comparing outcomes of policy changes in different states in the United States and Australia. A few studies have evaluated the impact of the Netherlands’ prohibition with an expediency principle scheme. There are at least four studies on the economic impacts of such changes and one or two on the wider impacts of medical marijuana initiatives in North America.
POLICY IMPACT STUDIES United States
Impacts On Rates Of Use
At least 4 uncontrolled studies and 4 studies employing matched control states were conducted to measure the impact on rates of use of the changes in state cannabis law that occurred in the USA from 1973 to 1978. Although data from the uncontrolled studies were used to support the changes, interpretation of the results of these studies was problematic and for that reason we focus here on the controlled studies. Taken together, these four studies indicated that states which introduced reforms did not experience greater increases in cannabis use among adults or adolescents. Nor did surveys in these states show more favourable attitudes towards cannabis use than those states which maintained strict prohibition with criminal penalties (Single, 1989; Single et al., 2000; Theis & Register, 1993).
In a study not designed to measure the impact of cannabis ‘decriminalisation’, Stuart and colleagues (Stuart et al., 1976) conducted a policy impact study of change in cannabis use sanctions in Ann Arbor, Michigan. They found that cannabis use was not affected by several changes in the municipal cannabis laws, including imposing a maximum penalty of $5 fine for cannabis use, compared to three neighbouring communities in Michigan which did not reduce penalties for use of the drug.
Saveland & Bray (1981) conducted secondary analyses of four national drug use surveys between 1972 and 1977. They found that cannabis use was higher in the ‘decriminalised’ states, both before and after the changes in law. Those that did moderate their law had increasing rates of use among adolescents and adults, but this was greater in the ‘non-decriminalisation’ states, and the greatest proportional increase in use was in the states with the most severe penalties. However, the ‘decriminalised’ states employed by the authors were California, Maine, Minnesota, Ohio, and Rhode Island. The last of these was not strictly one of the 11 ‘decriminalised’ states because, while it reduced the penalties for cannabis use, it still allowed the possibility of imprisonment (Single, 1989). However, its small population size meant that it was unlikely to affect the results.
The third controlled study used data from the Monitoring the Future national surveys of high school students. Oversampling high school students in the decriminalized states and comparing trends in them with trends elsewhere, Johnson and colleagues (e.g. Johnston et al., 1981) concluded that decriminalisation had no effect on either the rates of cannabis use, or related attitudes and beliefs about cannabis use, among this age group.
In a more recent study of the effects of the legal status imposed by individual states for possession of a small amount of cannabis, Theis and Register (1993) conducted a logistic regression analysis of a sample of 3,913 males in the National Longitudinal Survey of Youth who were between the ages of 14 and 21 in 1979 and were reinterviewed in 1984 and 1988. Controlling for a range of factors including age, education, marital status, ethnicity, urbanisation, parents' education, and religious participation, income and wealth, and ‘wellbeing’, they found ‘no strong evidence’ that cannabis ‘decriminalisation’ affected the choice, frequency, or use of alcohol, cannabis or cocaine.
These controlled studies are very dated but, for reasons already noted, there are problems with conducting analyses of more contemporary data to determine the ‘impact’ on current cannabis use of changes to cannabis laws which occurred decades earlier in these states. Pacula and colleagues (Pacula et al., 2003; Pacula et al., 2005) concluded, after reviewing state laws from 1989 to 1999, that some (Alaska, California, North Carolina and Arizona) of the so-called decriminalised states could no longer be classed as ‘decriminalised’, and a number of previously ‘non-decriminalised’ states had effectively removed criminal penalties (Louisiana, Massachusetts, New Jersey, Vermont, Wisconsin, and West Virginia). According to Pacula et al., there were 15 ‘decriminalised’ states in 1999, and a further 13 of the so-called ‘non-decriminalised’ states had introduced cautioning and diversion schemes. Thus, analyses of rates of cannabis use a decade or more after the changes in the original 11 ‘decriminalisation’ states will be confounded by these effects.
Impacts on other indicators of harm
A small number of studies have tried to measure impacts of the ‘decriminalisation’ of cannabis on indicators of harm. For example, Model (1993),
who conducted a study of hospital emergency room data from 1975 to 1978, suggested that ‘decriminalisation’ of cannabis in the 12 States was accompanied by a significant decrease in emergency room episodes involving drugs other than cannabis, and an increase in cannabis episodes. Her interpretation was that when cannabis was depenalised, illicit drug users tended to stay with the use of cannabis, and move away from the use of the other more severely punished illicit drugs.
Evaluations have been conducted on the prohibition with cautioning and prohibition with civil penalties schemes that have been in place in the eight Australian states and territories. However, evaluations of the cautioning and diversion schemes have been exclusively process evaluations (e.g. Hales et al., 2004). They have not looked at the impact of these changes on rates of cannabis use, and for that reason are not considered further here. A number of studies of impacts on rates of use have been conducted comparing Australian jurisdictions which have introduced civil penalties for minor cannabis offences and those which have not. The South Australian Cannabis Expiation Notice (CEN) system, which commenced in 1987, is the longest running and most researched Australian scheme.
Impacts on rates of use
An analysis of national population survey data conducted as part of an evaluation of the South Australian Scheme indicated that over the 10-year period from 1985 there had been an increase in self-reported lifetime cannabis use among persons aged 14 years and over, in all states and territories, with a greater degree of increase in South Australia than in the average of the other Australian jurisdictions (Donnelly et al., 1999, 2000). Between 1985 and 1995, the adjusted prevalence rates of ever having used cannabis increased in South Australia from 26% to 36%. There were also significant increases in Victoria (from 26% to 32%), Tasmania (from 21% to 33%) and New South Wales (from 26% to 33%), all of which had maintained strict prohibition of cannabis use. However, jurisdictions also differed in rates of change, with Victoria and Tasmania having similar rates of increase to South Australia (Donnelly et al., 1999, 2000). The analysis failed to find a statistically significant
difference between South Australia and the rest of Australia in the rate of increase in weekly cannabis use. This suggested that even if South Australians were slightly more likely to have ever tried cannabis than Australians in other jurisdictions, this did not result in higher rates of regular use in that state (Donnelly et al., 1999, 2000).
Although the introduction of the CEN scheme in South Australia did not apply to juveniles, there was some apprehension that a liberalisation of the laws might encourage cannabis use by young people. To examine this possibility, two groups examined data from three thousand South Australian students aged 11 to 16 years surveyed in 1986, 1987, 1988 and 1989 (Donnelly et al., 1992; Neill et al., 1991). Cannabis consumption levels in this age group remained stable between 1986 and 1989, with 20% saying that they had ever tried cannabis and 6% having used within the last week (Neill et al., 1991). On the basis of this analysis, it did not appear that changes to the cannabis laws impacted on cannabis use by secondary school students. Unfortunately, there were no data available concerning the prevalence of cannabis use amongst teenagers not attending school, who may have been more likely to engage in cannabis use compared to those remaining in formal education.
Donnelly, Hall and Christie (1999, 2000) also analysed trends in cannabis use among 14-to 29-year olds, the age group with the highest rates of initiation and regular cannabis use (Donnelly & Hall, 1994). Their analysis concluded that there had been an Australia-wide increase in rates of lifetime cannabis use between 1985 and 1995 in this age group, but that the rate of increase in lifetime cannabis use in South Australia did not differ from that in the rest of Australia. Indeed, the SA rate was in the middle of the range of rates found among the different jurisdictions in the 1995 survey (Donnelly et al., 1999, 2000). Regarding recent use, these authors also failed to find any consistent trends in rates of weekly cannabis use among 14-to 29- year olds in any jurisdictions from 1988 to 1995. However, they concluded that much larger samples would be needed definitively to rule out the possibility that there had been small differences in rates of increase among young people between jurisdictions (Donnelly et al., 1999).
Overall, Donnelly and colleagues concluded that the increase in lifetime cannabis use in South Australia was unlikely to be due to the implementation of the
cannabis infringement notice system, because: (1) similar increases occurred in Tasmania and Victoria, where the legal status of cannabis use did not change during the same period; (2) weekly cannabis use in South Australia did not increase at faster rate than the rest of Australia; and (3) there was no greater increase in cannabis use among young adults aged 14 to 29 years, which is the age group with the highest rates of initiation of cannabis use, in South Australia (Donnelly et al., 1999, 2000).
In a study of the impact of the Simple Cannabis Offence Notice (SCON) scheme which came in to effect in the Australian Capital Territory (ACT) in 1992, McGeorge and Aitken (1997) compared pre (1992) and post (1994) rates of cannabis use among 221 university students in the ACT with a control group of 246 university students in Victoria, which had maintained a strict criminal penalty scheme. Prevalence of lifetime cannabis use remained stable at about 53% over time, and frequency of use was indistinguishable between the two sites. The authors concluded that decriminalisation of cannabis had no effect on rates of cannabis use.
Surveys investigating the impact of the Western Australian Cannabis Infringement Notice (CIN) scheme (which came into effect in March 2004, and is the most recent of the Australian civil penalty schemes) suggested that cannabis use in that state has not increased since the scheme’s inception. Use in the past year among those aged 14 to 70 decreased from 18% in 2002 to 12% in 2007. The authors do not claim that these reductions were due to the CIN scheme itself, because similar declines have occurred in other Australian jurisdictions and this trend began a number of years before the CIN scheme came into effect (Fetherston & Lenton, 2007).
Williams and colleagues (Cameron & Williams, 2001; Williams, 2004; Williams & Mahmoudi, 2004) conducted a small number of omnibus studies of the impact of the Australian civil penalty schemes. They used data from the five national drug household surveys conducted in Australia from 1988 to 1998, along with police seizure data and other sources, to determine the impact of criminal and non-criminal sanctions on decisions to use cannabis. For example, Williams (2004) concluded after controlling for cannabis price, level of police enforcement, propensities to use and other factors, that the decision to use cannabis by males over the age of 25 was affected by changes in legal penalties in South Australia, the Australian Capital
Territory and the Northern Territory. But there was no evidence that that ‘decriminalisation’ significantly increased initiation to cannabis use by young males or females, or the frequency of use among existing users. However, a problem with these studies is that, in comparing cannabis prevalence and frequency rates in ‘decriminalised’ vs ‘non-decriminalised states’, they failed to take into account that rates of cannabis use are typically higher in jurisdictions which introduce civil penalties well before the reforms are put in place. Thus, the limited finding that prevalence of ‘ever’ using cannabis was higher in those states that had introduced civil penalty schemes does not mean that this was due to the legal reforms. It is also open to the alternative interpretation: that older adults are less likely to deny earlier use of cannabis if criminal penalties for use have been removed.
Social impacts of civil versus criminal penalties
Lenton and colleagues (Lenton et al., 1999; Lenton et al., 2000) compared the social impacts of a conviction under strict cannabis prohibition in place at the time in Western Australia, with that of an infringement notice under the CEN system in South Australia. They did so by comparing the experiences of 68 matched first-time apprehended cannabis users from each of these states. Although based on self-selected convenience samples rather than randomly selected cannabis offenders, this study remains one of few to have documented the social impacts of conviction versus civil penalty for a minor cannabis offence. Those in the WA convicted group were significantly more likely than the South Australian infringement notice group to report: adverse employment consequences (32% vs 2%); further contact with the criminal justice system (32% vs 0%); relationship problems (20% vs 5%), and accommodation difficulties (16% vs 0%) that could be attributed to their apprehension for the cannabis offence (Lenton et al., 1999; Lenton et al., 2000). The study failed to find a significant difference between the groups in terms of negative travel effects of conviction or infringement notice (0% vs 7%). This was possibly because the time from apprehension to interview (average 38 months) may not have been long enough for any effects on travel to be evident in a large enough number of the convicted sample to result in a significant finding (Lenton et al., 1999; Lenton et al., 2000).
Lenton and colleagues (Lenton et al., 1999; Lenton et al., 2000) also found that neither the infringement notice nor the cannabis conviction appeared to have much impact on subsequent cannabis use. Some 91% of the South Australian infringement notice group and 71% of the Western Australian convicted group said that their cannabis use one month afterwards was ‘not at all’ affected by their apprehension. Rates of post-apprehension cannabis use were highly correlated with rates of use prior to apprehension, consistent with other research (e.g. Erickson, 1976; Erickson, 1980) .
Cannabis users arrested and convicted for the first time in Western Australia were more likely to report negative attitudes to police and the justice system than their South Australian counterparts who received an infringement notice. Thus, 49% of the Western Australian group, compared with only 18% of the expiators, said that they had become less trusting of police, and 43% of the Western Australian convicted group, compared to 15% of the South Australian expiators, were more fearful of police as a result. The greater loss of trust in the Western Australian sample appeared in part due to the greater number of that group who were apprehended in a private residence, but did not appear to be due to other possible confounders. (Lenton et al., 1999; Lenton et al., 2000).
Research has also been conducted in Australia on the impact of introducing prohibition with civil penalty schemes on public attitudes and knowledge. Three main conclusions can be drawn.
(1) In Australian states which have prohibition with civil penalty schemes, even after these schemes have been in place for a number of years, the majority of the public support the use of non-criminal penalties for minor cannabis users, even if there may be a small reduction of support for the measures from pre-change levels. For example, in South Australia a public attitude sample conducted 10 years after the introduction of the CEN scheme found that 57% thought the laws should remain as they are or become less strict, compared to 38% who thought they should be more strict (Heale et al., 2000). In Western Australia the proportion of the public who thought the Cannabis Infringement Notice Scheme introduced in 2004 was ‘a good idea’ fell
significantly from 79% in 2002 to 66% in 2007, but, despite this drop, two-thirds still supported the scheme (Fetherston & Lenton, 2007).
(2) In Australian states where such schemes have been introduced, a larger proportion of the population than in states where criminal penalties apply mistakenly believe that cannabis use has been legalised (Fitzsimmons & Cooper-Stanbury, 2000).
(3) There is no evidence that cannabis is seen as less harmful in those states once civil penalties were introduced. Indeed, in Western Australia, Fetherston & Lenton (2007) found that the WA public saw cannabis use as more harmful to health in 2007 than they did in 2002 before the scheme was introduced. Again, this was not to say that the more negative views towards cannabis occurred because of the legal changes, as attitudes to cannabis had been becoming progressively more negative since well before the legal changes among both adults (Draper & Serafino, 2006) and school children (Miller & Lang, 2007). Rather, there was no evidence of cannabis being seen as more benign.
Problems with civil penalty schemes – low expiation rate, netwidening, exploitation by criminal elements and disproportionate impact on the socially disadvantaged.
There were three major problems associated with the cannabis infringement notice system that was implemented in South Australia. Firstly, only 45% of people paid their fines by the due date (Christie & Ali, 2000). Secondly, there was 2.5-fold increase in the number of people who were issued notices from 1987/88 to 1993/94. This so called “net-widening” appeared largely due to the ease with which notices could be issued by police. Its effect was to increase the numbers at risk of criminal sanction for non-payment of fines, an outcome that can particularly disadvantage those with limited financial means (Christie & Ali, 2000). Thirdly, there was evidence that criminal gangs were syndicating cannabis cultivation by aggregating for sale multiple crops of cannabis plants that individuals were allowed to grow for personal use under the original 10-plant limit (Sutton & McMillan, 2000).
Recent evaluation of the first 3 years of the Cannabis Infringement Notice (CIN) scheme in Western Australia suggests that features of this scheme may have addressed these problems. While only 43% of people expiate their notice by the due date, a further 25% were expiated in the 1-2 months afterwards, because non-payers risked the sanction of losing their driver’s licence if the notice wasn’t expiated. The overall expiation rate, allowing for delayed payment beyond the 1-2 month period, is estimated at up to 75% (Swensen, 2007). It appears that the additional time to pay and further penalty provided through this administrative sanction may contribute to the higher overall expiation rate. While there has been some net-widening in the scheme, this only resulted in an overall increase in total consequences by some 14% (compared to 250% in the SA case), 90% of which was due to the inclusion of implement offences (possession of bongs, etc.) within the CIN scheme (Swensen, 2007). Furthermore, under the WA scheme, only 2 plants per primary residence are eligible for an infringement notice. This, in addition to other aspects of the scheme (such as providing support for police to charge people they believe are using the infringement levels as a cover for dealing activities, and reducing the cut-off for being deemed a supplier from 25 to 10 plants) seems to have prevented the exploitation of the plant limit by syndicating growing that was seen under the SA scheme (see Sutton & McMillan, 2000).
Another problem with civil penalty schemes is that they can cause disproportionate hardship for already socially and economically disadvantaged members of society. Early work on the South Australian scheme indicated that socially disadvantaged persons were over-represented among those failing to pay expiation fines (Sarre et al., 1989). Studies conducted since that time show that attempts to reduce the financial burden of fines have had some success (Christie & Ali, 2000; Swensen, 2007). However, data from the first 3 years of operation of the Western Australian scheme show that indigenous people continue to be disadvantaged by new penalty options, and are 6 times less likely to expiate their notice than their non-indigenous counterparts. Special effort needs to be made to monitor and address this problem by those implementing civil penalty schemes (Lenton, 2007).
Hughes and Stevens (2007) reviewed available literature and statistics and conducted stakeholder interviews to understand the impacts of the Portuguese reforms introduced in 2001. As described in the previous chapter, these reforms removed criminal sanctions for personal possession, use and acquisition of all drugs including cannabis, and introduced a system of diversion to Commissions for the Dissuasion of Drug Addiction (CDTs). Hughes and Stevens (2007) concluded that: (i) changes in reported patterns of cannabis, and other drug use patterns were apparent, but it was unclear whether these changes were real and, if so, whether they were due to the policy reforms or to more reflective of Europe-wide trends in drug use, such as those that occurred in Spain and Italy, which did not change the law; (ii) there had been increases in numbers of people attending for cannabis treatment; (iii) there had been difficulties with the operation of the Community Drug Treatment centres (CDTs); and (iv) there were views at a political level about changes needed to the reforms which ranged from scrapping them to streamlining the operation of the CDTs and increasing resources for treatment services.
General population surveys of the Portuguese public only commenced in 2001, the same year the reforms were introduced, and there are no recent data series at the population level to monitor changes since that time. However, surveys of school students have been conducted as part of the European School Survey Project on Alcohol and Other Drugs (ESPAD), the most recent of which was conducted in 2003. These data indicate an increase in self-reported cannabis use among 16 to 18 year olds in Portugal from 9.4% in 1999 to 15.1% in 2003. However, Hughes and Stevens note that school surveys such as these are often susceptible to reporting bias, and they thought that decriminalisation may have led young people to be more willing to admit to cannabis use in such surveys (Hughes & Stevens, 2007).
Official treatment statistics show a reduction in heroin-related treatment presentations and an increase in cannabis presentations. Thus, referrals to the CDTs for cannabis increased from 47% in 2001 to 67% in 2003, and stayed at around that level in the two years thereafter. Heroin presentations, by contrast, dropped from 33% in 3001 to 17% in 2001 and have remained fairly stable since. Importantly, while
the CDTs have been seen as crucial to the success of the reform process, they have also been criticised as being too administratively cumbersome and resource intensive. The Portuguese reforms have highlighted the need for a well-operating referral and treatment system (Hughes & Stevens, 2007).
In conclusion, these authors state:
The Portuguese experience cannot provide a definitive guide to the effects of decriminalization of drugs, but only indications of the results of decriminalization in the specific Portuguese context. It is not possible to tell the extent to which changes were caused by decriminalization or the wider drug strategy. The extent to which difficulties in implementation impeded the impacts from the reform remains unclear. (Hughes & Stevens, 2007, 9)
The UK government reclassified cannabis as a Class C drug on 29 January 2004. In July 2007 the question of reclassifying cannabis as a Class B drug was referred to the Advisory Council on the Misuse of Drugs (ACMD). Despite the fact that the ACMD, having reviewed the available evidence, believed that cannabis should remain a Class C drug (Advisory Council on the Misuse of Drugs, 2008), the British Government rejected this advice and, consistent with a number of statements by Prime Minister Gordon Brown, decided to reclassify cannabis as a Class B drug (Castle, 2008).
We can find no evidence of the 2004 reclassification leading to an increase in cannabis use. Indeed, data from the latest British Crime Survey suggests that while the number of police contacts for cannabis use increased, cannabis use decreased. Police contacts for possession of cannabis increased from 88,263 in 04/05 to 119,917 in 05/06 and 130,406 in 06/07 (Nicholas et al., 2007, 40), which the authors note may reflect changes such as ‘the use of formal warnings for cannabis, rather than real changes in its incidence’ (p.21). The outcome is consistent with the ‘net-widening’ reported in South Australia, because there has not been any accompanying increase in reported rates of cannabis use in the general community. Self-reported cannabis use among the public aged 16 to 59 declined from between 10.3% to 10.8% in the 5
years prior to the change, to 9.7% in 04/05, 8.7% in 05/06 and 8.2% in 06/07 (Nicholas et al., 2007, 43). Similarly, the proportion of 16 to 24 year olds who report using cannabis in the last year has decreased from over 26% in the 5 years prior to reclassification to 25.3% in 03/04, 23.6% in 04/05, 21.4% in 05/06, 20.9% in 06/07 (Nicholas et al., 2007, 44).
May and colleagues (2007) conducted an internet-based survey of 749 respondents in England and Wales between June and November 2005. Although the authors acknowledge that the sample could not be said to be representative of general population, they found 77% supported the reclassification from B to C. With regards to knowledge of the change, 98% knew that cannabis remained illegal under the changes and 78% believed that they had a fairly good understanding of the changes, 74% understood that street warnings could be issued and 69% that the cannabis user could be arrested. However there was some confusion as to how juveniles (as against adults) were treated under the provisions, and many incorrectly thought that police ‘turning a blind eye’ to cannabis was legitimate under the reforms (May et al., 2007). A small interview study of 61 young people between that ages of 14 and 21 who had either used cannabis or been stopped by police regarding drugs found that, while 69% supported reclassification to Class C, 52% did not understand that nothing had changed for those under 18 and they would still be arrested for cannabis possession. Interviews with 150 police officers indicated that 59% believed the government were wrong to reclassify cannabis as a Class C drug, and 93% of officers surveyed stated that they had come across people who believed—or more likely claimed to believe— that cannabis had been legalised (May et al., 2007).
An analysis of cannabis seizure data from London and a number of other boroughs in the UK suggests that the potency of sinsemilla has increased from a mean of 6% THC in 1996 to 13% in 2005, and that increasingly in many boroughs sinsemilla dominates the market over resin (hash) and herbal cannabis (Potter et al., 2008). However, these changes are not a result of reclassification, but rather reflect increased domestic cannabis production in the last 5-7 years (Personal communication May, 2007), primarily by hydroponic cultivation. Similar trends have
been seen in a number of other countries around the world (United Nations Office of Drugs and Crime, 2006).
As mentioned in Chapter 2, trends in both cannabis-related drug treatment and mental health presentations in the UK since classification were examined in the UK Drug Policy Commission submission to the ACMD Cannabis Classification Review 2008 (UK Drug Policy Commission, 2008). They found that there had been increases in the number of people, particularly those under 18, in treatment for problematic cannabis use. Since reclassification, the number of hospital admissions for mental and behavioural disorders diagnosed as having been caused by cannabis use had also increased. They noted that, although these increases could be seen as concerning, this was not necessarily an indication that harmful use had increased. Rather, better detection of problem users or more treatment places could also explain the increase, particularly as indications were that rates of cannabis use had declined since reclassification. They noted that other economic, social and cultural factors were likely to be stronger drivers of cannabis use than its classification status (UK Drug Policy Commission, 2008). In addition, there is typically a 5- to 10-year lag between initiation to cannabis use and presentation to treatment for problem use, so people presenting for treatment in the present probably initiated use 5 or more years ago.
Impacts on rates of use
The evaluation literature on the effect of the cannabis coffee shop system in the Netherlands has been controversial. In two publications, MacCoun and Reuter (1997, 2001b) compared available data on the effects of drug laws on cannabis use across the Netherlands, USA, Denmark and Germany. In the second of these papers, the authors analysed data from 28 separate studies that employed adequate controls. They concluded that reductions in criminal penalties in the Netherlands from 1976 to 1992 had little effect on cannabis use, but they suggested that the increase in commercial access in the Netherlands from 1992 to 1996 with the growth in numbers of cannabis coffee shops (termed by them de facto legalisation) was associated with growth in the cannabis-using population, including young people. For example, they
claimed that lifetime prevalence of cannabis use among Dutch aged 18 to 20 increased consistently and sharply from 15% in 1984 to 44% in 1996, and that past-month prevalence in this group increased from 8.5% to 18.5% over the same period (MacCoun & Reuter, 2001b). They hypothesized that the rapid increase in cannabis use in the Netherlands in the mid-1980s may have been the consequence of the gradual progression from a system of passive depenalisation to a commercialisation era which, up until 1995 at least, allowed for greater access and increasing promotion of cannabis (MacCoun & Reuter, 2001b). This included indirect effects such as heightened salience and glamorisation of cannabis because of visible promotion especially in counter-culture media advertising, and through postcards and posters (MacCoun & Reuter, 2001b). However, MacCoun and Reuter (2001b) note that the increases in prevalence in the Netherlands from 1992 to 1996 were similar to trends that occurred in Norway, the USA, the UK and Canada. The latter points to the important role of poorly understood social and cultural influences on the prevalence of cannabis use. They also noted that, even with this more liberal policy, rates of cannabis use in the Netherlands were no greater than those in the USA, which had a consistently more punitive policy towards cannabis use over the same period.
MacCoun and Reuter’s glamorisation hypothesis has been judged as lacking credibility by others (e.g. Abraham et al., 2001; de Zwart & van Laar, 2001; Korf, 2002). For example, Abraham and colleagues (2001) criticised the MacCoun and Reuter analysis for its comparison of data from cities with statistics for whole nations, and questioned the validity of some of the comparative data sets. Both points have been acknowledged by the authors (MacCoun & Reuter, 2001a, 2001b).
In a subsequent paper, Korf (2002) looked at trends in current, rather than lifetime, cannabis use and showed that, across the Netherlands, other European countries and the US, there had been wave-like trends in cannabis use. He noted that it was ‘striking’ that these trends in cannabis use among young people in the Netherlands mirrored the four stages in the availability of cannabis identified by MacCoun and Reuter (2001b). Thus, the number of cannabis users peaked in the 1960s and early 1970s, when cannabis distribution occurred through an underground market. Recent use of cannabis by Dutch youth increased during the 1970s, when
house dealers were taking over from the underground market. It increased again during the 1980s, when coffee shops established themselves as the sale points for cannabis. In the late 1990s, rates of recent use among Dutch youth stabilised. Rates decreased by the end of the 1990s, after the age of access to coffee shops was increased from 16 to 18 years in 1996 and the number of coffee shops was reduced (Korf, 2002).
Korf (2002) concludes nonetheless that it is doubtful that changes in cannabis use were causally related to changes in cannabis policy. More recently, an analysis of the changes in cannabis prevalence among Dutch secondary students aged 12-17 found that these changes were paralleled by changes in the age of first use of cannabis. Average age of onset of cannabis use decreased from 15 years in 1992 to 14 years in 1996. After the legal age of buying cannabis from coffee shops was raised in 1996 the age of onset stabilized through to 2003 (Monshouwer et al., 2005). However, as Korf (2002) noted, after raising the age limit in coffee shops, Dutch students were more likely to obtain their cannabis from outside coffee shops, mainly from friends. This could have increased the likelihood of them being offered other drugs or getting involved in other crime (Korf, 2002; Monshouwer et al., 2005).
Impacts on patterns of use and market access
A project commenced by Peter Cohen, Craig Reinarman, Stephan Quensal and Lorenz Bollinger has involved using a standardised survey instrument to compare representative samples of experienced cannabis users (used 25 times or more) in three cities with different cannabis policies. The surveys began in Amsterdam in 1995, Bremen in 1998 and San Francisco in 1999 (Borchers-Tempel & Kolte, 2002). The study found that the prevalence of different ‘consumer types’ varied across the cities. Thus ‘moderate’ users — long-term users of the drug who use regularly and tend to limit their use to reach a desired level of intoxication - were more prevalent in Amsterdam (19%) and Bremen (19%), but were less likely to be found in San Francisco (10%) (Borchers-Tempel & Kolte, 2002). Conversely, San Francisco had higher rates of ‘leisure-oriented occasional consumers’ (29%) and heavy (mostly dependent) users (22%). The study concluded that “repressive policies do not reduce consumption, but do produce problematic consumption patterns
among many of those who defy authority” (Borchers-Tempel & Kolte, 2002, p. 411). However, this conclusion seems to over-state the impact of policy on rates of use, and does not seem to entertain that there are likely to be other social and cultural differences between the three cities which were important.
Nonetheless, the work confirms that the Dutch have shown that a system of cannabis supply can be established which largely separates the cannabis market from that for other illicit and potentially more harmful substances. Some 87% of the Amsterdam sample said that they usually brought their cannabis from cannabis coffee shops, whereas more than 80% of the Bremen sample and 95% of the San Franciscan sample said they usually brought their cannabis from friends who knew a dealer, or from ‘known dealers’. Furthermore, 85% of the Amsterdam sample, compared to 51% of the Bremen sample and 49% of the San Francisco sample, said that no other drugs were available from the source where they obtained their cannabis (Borchers-Tempel & Kolte, 2002, pp. 399 - 400).
A more recent paper from the same study comparing rates of use in Amsterdam and San Francisco found no evidence that ‘criminalisation’ (in San Francisco) reduced use or that ‘decriminalisation’ (in Amsterdam) increased use. Rather, except for higher rates of cannabis use in San Francisco, they found strong similarities across both cities (Reinarman, Cohen & Kaal 2004). In contrast to the conclusions of the Borches-Temple and Kolte (2002) paper, they concluded that regular cannabis users organised their use around sub-cultural norms about when, where, why, with whom and how to use, rather than around laws and policies (Reinarman et al., 2004).
Conclusions on Dutch policy
So what can we say about the impact of the Dutch cannabis policy on prevalence and patterns of cannabis use and harms? Firstly, there is a consensus that depenalisation in the Netherlands did not, in itself, lead to increases in population levels of cannabis use among adults nor among young people. This finding is consistent with results in the Australia and the USA. Secondly, the Dutch system does appear to have successfully separated the market for cannabis from other substances. The majority of cannabis users who buy their cannabis from the
regulated environment of coffee shops do not need to have contact with other illicit sources (dealers or friends who have contact with dealers), where they may be exposed to other drug use and criminality.
Thirdly, there are competing views of the impact of the ‘commercialisation’ of cannabis sales from the mid-1980s to mid-1990s. On the one hand, the increases in prevalence of cannabis use among youth appear to mirror changes in other countries that began prior to this period. On the other hand, the prevalence changes correlate with changes in policy, increasing as access and availibility increased. On balance we would say that the case is still open about whether de facto legalisation led to more use by youth and an earlier age of onset; it cannot be ruled out that increases in youth prevalence may have been associated with increasing de facto legalisation, and subsequent decreases with tightening up of this policy. The critical point here is that, while international research shows that there is little relationship between rates of cannabis use and cannabis policy when use remains illegal in both a de facto and de jure sense, the Dutch experience raises the question about whether going beyond depenalisation to de facto legalisation may increase rates of cannabis use among the young, who are most vulnerable to the adverse effects of cannabis. Some will disagree with this analysis, but we believe at this stage a cautious conclusion is warranted, pending further research.
Changes in Italian policies towards cannabis and other drugs should ideally provide a perfect case to examine the impact of depenalisation using an ABA type research design (penalisation-depenalisation-repenalisation). The popular view is that there has not been any relationship between the legal changes and prevalence of cannabis (and other drug use). An often cited paper by Solivetti concludes:
The first and most impressive fact that emerges from the Italian history of drug policy is the lack of visible impact of the various legislative actions in this field. What is particularly impressive is the lack of visible impact as regards the – in most cases increasingly – repressive actions. The harsh and undifferentiating repressive sanctions of the 1950s did not prevent the drug diffusion boom of the 1960s. The increased
criminal sanctions against drug traffickers, provided by the 1975 Act, did not result in curbing drug diffusion; on the contrary, the latter grew more and more. The reintroduction of sanctions against drug use in 1990 had no visible impact on the phenomenon. (Solivetti, 2001, p. 51)
One issue is that Solivetti only analysed data series on cannabis and other drug use up to 1998. Now, 10 years later, a longer data set should have accrued to investigate impacts of the last depenalisation, which came into effect in Italy in 1993. The EMCDDA has noted that rates of cannabis use in a number of European countries including France, Spain and Italy have increased in recent years and that they are developing methodologies to investigate more frequent (daily) use patterns, which are likely to be indicative of cannabis dependence and other harms (EMCDDA, 2007). So, consistent with our conclusions for other countries, while there is no evidence supporting the view that cannabis depenalisation in Italy resulted in increase rates of use and problems, it needs to be noted that further research would be useful to confirm the rather limited evidence examined to date.
As noted in the previous chapter, while the consumption of cannabis is prohibited by law in Switzerland, police forces in the 26 Swiss Cantons exercise considerable discretion in how the law is applied. Capitalising on this and using data from the World Health Organization (WHO) European Health Behavior in School-Aged Children survey (HBSC) study, Schmid (2001) employed hierarchical linear modelling to examine the impact of peers, urbanisation and the severity of enforcement of cannabis laws on rates of cannabis use among 3,107 15-year-olds. The study suggested that while living in urban areas affected the extent to which cannabis use was affected by peers, whether the canton in which one lived had repressive or more lenient cannabis policy did not have any effect on cannabis use by individuals (Schmid, 2001).
There have been at least four economic analyses of the impact of introducing prohibition with civil penalty schemes on law enforcement and other costs. Civil penalty
schemes can in principle be far cheaper than criminal or diversion schemes in terms of enforcement and justice resources, because civil offences can usually be dealt with more expeditiously and with fewer procedural protections (Pacula et al., 2005). Economic analyses have found that the magnitude of potential criminal justice savings depends on the size of the jurisdiction and the cost of the existing mechanisms for control of minor cannabis offences. Annual savings were estimated at $US100 million in California (Aldrich & Mikuriya, 1988), at $US24.3million in Massachusetts (Miron, 2002), but at less than $US1 million in South Australia (Brooks et al., 1999). However, as Hall and Pacula (2003) have suggested, given the way the budget process in democracies works, it is not clear where these savings would in fact be directed – and doubtful that it would be towards such “social dividends” as tackling more serious crime or otherwise reducing drug-related harm (Lenton, 2005).
Studies of the impact of introduction of medical marijuana schemes on rates of cannabis use in the general community are difficult to locate. While the US General Accounting office conducted a thorough investigation of such initiatives in Oregon, Alasks, Hawaii and California in 2002, this did not include impacts on rates of cannabis use generally (US House of Representatives United States General Accounting Office, 2002). Khatapoush and Hallfors (2004) analysed 3 waves of telephone survey data in 1995, 1997 and 1999 to investigate the impact of Proposition 215, the Compassionate Use Act, which came into effect in California in late 1996. They compared attitudes and use rates among 16- to 25-year-olds in selected communities in California and 10 control states. They concluded that medical marijuana policy had little impact on youth and young adult marijuana-related attitudes and use in the selected communities in California and beyond. After Proposition 215 came into effect, Californians were less likely to see occasional marijuana use as risky, and were more approving of marijuana legalization and use for medicinal purposes, but were less approving of (recreational) marijuana use. Rates of recent cannabis use (past month and past year) were higher in California compared to control states, both before and after Proposition 215 came into effect.
CONCLUSIONS AND IMPLICATIONS
This chapter reviewed the available literature on the impact of cannabis reforms which have been undertaken within the provisions of the existing international drug treaties and conventions. It is apparent that only a minority of the examples of reform identified in the previous chapter have been subjected to any evaluation that can assess the impacts of the reforms on rates of cannabis use and cannabis-related harm.
Furthermore, two caveats need to be made. Firstly, as we have done in this chapter, one needs to take into account that much of the research which has been conducted is compromised by methodological flaws. Secondly, caution also needs to be exercised in drawing conclusions from one reform example in one country and trying to implement similar changes in another. Geographical, social, cultural, political, legislative and other contextual factors are likely to influence reforms and their impacts. Thus, it is important that any reforms which are implemented are subject to rigorous evaluation. Ideally the results of any such evaluations should allow the reforms to be modified to minimise any unintended adverse effects, such as net widening. Nevertheless, despite these caveats, the following conclusions can be drawn on the basis of the available literature.
Impacts On Prevalence Of Use
There does not appear to have been any large increases in cannabis use in countries that have maintained the de jure illegality of cannabis but implemented reforms which, either at a national or subnational level, have reduced the penalties to civil or administrative sanctions. Among these policy reforms, the shift from strict prohibition to prohibition with civil penalties is the best researched. The methodologically stronger studies fail to find that these changes in penalties have large impacts on the prevalence of cannabis use at a population level or among school children. Although the results here may be limited by low statistical power and limited sample size to detect small increases, the available evidence suggests that if increases in use are found at general population level, they are likely to be small.
The evidence on the impact of depenalisation in the Netherlands suggests that it has not resulted in increased prevalence of cannabis use at a community level, and that it has been successful at separating the cannabis market from other drug markets. Questions remain open about the extent to which the increased commercialisation of cannabis in the Netherlands in the period 1992-1996 may have resulted in more cannabis use by the young and in an earlier age of onset. If it did, then the age restrictions and other changes introduced in the coffee shops in the mid 1990’s appear to have arrested this trend.
Regarding impacts of reforms on prevalence of cannabis use, it is also apparent from a number of the studies that, at least as long as the illegality of cannabis is maintained, the laws and sanctions which apply seem to have, at most, a relatively modest impact on rates of cannabis use. In a number of examples, trends in cannabis use appear to be independent of the penalties which apply. It is likely that, as far as reforms under the existing international treaties are concerned, other non-legal factors such as social, economic and cultural trends, some of which exert their influence across state and national boundaries, have a far greater impact on cannabis use than the penalties which apply in a particular jurisdiction to that very small proportion of users who are ever arrested.
Reducing The Adverse Consequences Of Prohibition
The research suggests that those reforms which have been undertaken under the existing international drug conventions have reduced, but not eliminated, some of the adverse social impacts of prohibition on individuals. They also appear to reduce the costs to the criminal justice system of prosecuting cannabis use offences. However, these benefits can be undercut by police practices that increase the number of users who are penalized, or enforce the law in a discriminatory way. The costs to individuals apprehended can be substantially reduced by civil rather than criminal sanctions for many users. Nonetheless these schemes can have a disproportionate impact on those of limited financial means and the socially disadvantaged, who may still end up being processed by the criminal courts because they are unable to pay fines. Although savings in criminal justice resources have been noted and appear to
be proportional to the size of the jurisdiction, it is unlikely that these savings are in practice strategically diverted to address other more serious crime problems, but rather ‘disappear’ into general police budgets or central revenue.
Whilst there is debate about whether the de facto legalisation system in parts of the Netherlands contributed to increasing rates of use, there is good evidence that the Dutch coffee shop system has removed the risk of legal penalties for most users and effectively separated the market for cannabis from other illicit drugs. It may be the case, however, that the age restrictions introduced in the coffee shops in the mid 1990s may have paradoxically increased adolescent exposure to the illicit market and thereby increased associated risks and harms.