Introduction
The drug policy reform debate has implications for a
wide range of issues – from crime and health, to the
economy and even the environment. But for many,
the overriding concern is the effect that any policy
change would have on levels of drug consumption (and,
implicitly, the negative consequences that stem from it).
Indeed, the fear that drug use will increase following
any move away from a punitive approach is the most
frequently raised and politically potent of all the
objections to reform, appealing to worries about things
such as the health and welfare of loved ones, increased
public disorder and drugged driving.
Support for maintaining such an approach is typically
predicated on three assumptions:
• That criminalising (or otherwise punishing) drug
users is necessary in order to deter people from
using drugs;
• that enforcement against the supply of drugs
restricts their availability (and, in turn, their use) to
a sufficiently worthwhile extent; and
• that levels of use are a good proxy measure for
levels of harm, both to drug users themselves and
wider society.
This report argues that this rationale for continuing
with an enforcement-led approach to drugs is poorly
supported by empirical research, and that alternative
policies – in particular the decriminalisation of personal
drug possession or the introduction of legally regulated
drug markets – can produce better outcomes while also
avoiding dramatic increases in use. It also makes the
case that overall levels of drug use are not an accurate
indicator of levels of drug-related harm, and should not
be considered as such.
Given the pace at which real-world alternatives to
criminalisation and prohibition have taken hold in
recent years, and the extent to which drug policy
reform has become a mainstream issue, there is now
no shortage of claims and counterclaims about how
different policies affect levels of drug consumption.
This report attempts to cut through this debate,
addressing the widely held concern about increased use
by reviewing the evidence acquired since the modern
international drug control framework was established
in the 1960s, and looking at what is known about
other approaches to managing a range of substances.
Ultimately, the intention is to provide a representative
overview of what is known about the relationship
between drug policy, drug use, and related harms.
But there is also a need to be frank about what is
currently uncertain: the more far-reaching the
departure from the status quo, the more scope there is
for unpredictability. So while there is now a significant
body of research into drug use under prohibition, there
is still no perfect counterfactual to it. Although there
is direct evidence of how decriminalising drugs affects
levels of consumption, there is a much more limited
empirical basis for inferring what the consequences
would be of full-scale legalisation (which in any case
can take many forms). That aside, taken together, the
evidence presented in this report suggests the following
broad conclusions:
• The importance of prevalence of use as an indicator
of the success of drug policy is often overstated,
at the expense of equally or more important
indicators, such as problematic use or drug-related
deaths;
• levels of drug use can be a poor proxy measure
for levels of drug-related harm, and since such
harm is the more important policy priority, more
accurate, alternative indicators should be employed
to measure it;
• the decriminalisation of drug possession for
personal use has, at most, only a marginal impact
on levels of drug use;
• legal drug markets that are highly commercialised
and loosely regulated are likely to lead to significant
increases in drug use compared with levels of use
under prohibition with illicit markets;
• it is possible to create legal drug markets that
are sufficiently regulated and taxed so as to
avoid dramatic increases in drug use (including
problematic use) compared with levels of use under
6
prohibition with illicit markets; and
• the effect of legalisation and regulation (however it
is designed and implemented) on levels of drug use
is likely to vary significantly depending on which
drug or drugs are made legally available.
- Drug use: the issues
- a) Drug use and harm
The academic debate on drug policy reform is most
commonly framed in utilitarian terms (although
politicians and the public often view it through a
moral lens). So when participants in this debate make
their case for a particular approach, they tend not to
appeal to the intrinsic rightness or wrongness of drugtaking,
but to the magnitude of positive and negative
consequences of drug-taking. Of course, different people
with different values weight these various outcomes
differently. So questions such as “Would legalisation be
better than prohibition?” cannot, in the strictest sense,
be answered scientifically. Nevertheless, the academic
framing of the debate can and should be informed by
empirical evidence.
Taking an archetypal consequentialist approach,
MacCoun and Reuter (2001; 2011) recommend that
different drug policy regimes be evaluated by their
effect on the overall level of harm caused by drug use.
They identify three factors that determine most of the
harms caused by drug use: the number of users, the
average number of doses per user, and the average
harm produced per incident of drug use. Expressed
more clearly, the relationship is:
Total Harm = Prevalence x Intensity x Harmfulness
Those who support legalising drugs (or other reforms
that reduce or eliminate penalties for possession)
typically ascribe most importance to the last element
of this equation, and include not just harms to the drug
user, but costs to wider society as well in their definition
of harmfulness.
So supporters of reform highlight how prohibition
and criminalisation can dramatically increase the
risks associated with drug use itself, as the primary
risks of many illicit drugs are a product not of their
pharmacology, but of their being produced and
supplied by an unregulated criminal market, with users
directed to the criminal justice system, rather than the
healthcare system. Street heroin mixed with potent
adulterants such as fentanyl and used with shared
needles in unhygienic environments, for example,
carries far greater risks than pure, pharmaceutical
heroin (diamorphine) used in a supervised clinical
setting.
Added to this, the broader social costs stemming from
drug use that are created or exacerbated by prohibition
can include, among other things, the stigma and limited
life chances that stem from a criminal conviction for
drug possession, racial disparities in the enforcement
of punitive drug laws, and the violence and conflict
generated by street dealers and drug trafficking
organisations. Such harms, it is argued, could be
significantly reduced if the drug trade was moved above
ground and legally regulated, or if users did not run the
risk of being punished.
The logical extension of such a focus on harmfulness is
that if an incident of drug use does not cause significant
harm to the user or others (or lead to harmful use later),
then it should not be a concern of public policy. So, it is
argued, policy should seek to reduce overall harms –
whether to the user or wider society – from problematic
use, not necessarily to eliminate use per se. (This is
discussed in more detail in section 2.c.)
By contrast, supporters of prohibition focus mostly on
the first, and to a lesser extent the second, component
of the equation above, rejecting changes to the status
quo out of concern that drug use would increase if
sanctions were reduced and/or a legal supply of drugs
was established. After all, decriminalisation or legal
regulation might cause total harm to rise if, despite
reducing the average harm per incident of use, this
was outweighed by a sufficiently large increase in
the number of such incidents. And depending on the
7
extent of such an increase in use, that could ultimately
mean higher levels of drug dependence, greater
numbers of intoxication-related accidents, and/or
drug-related deaths.
Both sides of the drug policy debate are therefore
sometimes guilty of neglecting key factors that
contribute to the overall amount of drug-related harm.
Reformers are often reluctant to engage with the
issue of whether, under a less punitive approach, use
could rise to what would be, on net, more damaging
levels; and advocates of prohibition rarely consider the
proposition that an increase in use could be justified by
a sufficiently large reduction in average harm.
As alluded to above, use reduction has tended to be the
dominant aim of drug policy, with the priority usually
a reduction in the number of people who use drugs
(rather than a reduction in the intensity of their use).
In the US, for example, the national strategy of the
Office of National Drug Control Policy (2007: 1) has
historically been framed almost exclusively in terms of
reducing the prevalence of use (although this has begun
to change in recent years). - b) The dominance of prevalence
The prevalence of drug use (the number of people
using drugs in a given population) is clearly a legitimate
concern under any policy model, but its status as the
key indicator of the success or failure of drug policy
stems in part from a simple lack of more accurate drug
data systems. The production and use of illegal drugs is
not something that market participants happily report
to the authorities. Hence there is nothing like the kind
of administrative data that exists for tracking activity in
other, legal markets.
Drug arrests and drug-related deaths were originally
used to monitor trends in drug supply and use, but
these measures are highly imprecise: variations in the
number of drug arrests, rather than signifying changes
in consumption or production, can simply indicate
changes in enforcement practices; and drug-related
deaths are often multiply determined – an intoxicated
driver killing a pedestrian may be recorded as a traffic
death, rather than as a consequence of drug use.
So when drug use grew to the point of being a
central concern in the 1960s and 1970s, there was a
concerted effort to devise better data systems. This
led to the development of school-based surveys of
drug use by youth, and surveys of drug use in the
general population (often called “household population
surveys”).
Data limitations therefore played a major role in the
prevalence of drug use assuming such importance
in the evaluation of drug policy. But this situation is
also partly a function of a culture that has, whether
explicitly or implicitly, defined the consumption of
certain drugs as intrinsically wrong and damaging.
Public perceptions of illicit drug use (and those who
engage in it) have been distorted by more than a
century of moral panics, early examples of which were
driven by racial prejudice and, in the United States in
particular, by an influential Temperance movement
that considered the use of all psychoactive drugs –
including alcohol – as a root cause of social decay
(Berridge and Edwards, 1981; Jay, 2002; Musto, 1999;
Courtwright, 2005).
But while attitudes towards alcohol use changed
following America’s failed experiment with alcohol
prohibition, the use of other, more “foreign” drugs
has remained highly stigmatised. Prohibition-era
rhetoric was (and often still is) simply applied to a
different range of substances, as can be seen in the UN
convention that underpins today’s international drug
control regime. It describes drug addiction as a “serious
evil … fraught with social and economic danger to
mankind”, one that the international community has
a “duty to … combat” (United Nations, 1961: 1). This set
the tone for the “war on drugs” declared by President
Richard Nixon in the 1970s.
Against this backdrop, in which the use of certain
drugs is conceived of as a threat to the very fabric of
society, it is perhaps unsurprising that drug policy has
8
overwhelmingly focused on prevalence reduction.
“Harm reduction” approaches – which seek to make
drug use safer, rather than just reduce or eliminate
it – have faced significant political obstacles, arguably
because they focus less on the simple fact of whether
a person uses drugs, and more on whether a person’s
drug use is having negative consequences (Harm
Reduction International, 2016). - c) Prevalence of use as a measure of
total harm
Clearly, drug use can, in itself, cause substantial harm to
individuals, but it is by no means an inevitability. There
are different types of drugs and drug-using behaviours,
motivated by different priorities, which have different
outcomes. It is far from the case that everyone who
takes drugs becomes a chronic, dependent user, whose
consumption is high-risk and likely to cause themselves
and/or others harm.
The prevalence of drug use can be estimated by a
number of methods, including wastewater-testing,
or testing of arrestees, but is usually calculated
from survey data. This is then augmented by
population-specific research, such as school-based
surveys (although these are less consistent across
jurisdictions, complicating national comparisons).
Many countries have established surveys that focus
specifically on drug-taking behaviour, while others
incorporate questions about drug use into general
health surveys, or, as in the case of England and
Wales, into national crime surveys. The context of
the survey inevitably influences the type of questions
that are asked and who responds to it (EMCDDA,
2009: 15) .
The European Model Questionnaire recommends that
countries ask questions on a minimum of six drugs:
cannabis, ecstasy, cocaine, heroin, amphetamines
and LSD. However, some drug use is rarely captured;
certain surveys now include questions on some
new psychoactive substances (NPS), but not all (the
bewildering array of such products making it all but
impossible), while the use of diverted prescription
drugs is generally poorly monitored, leaving a messy
and incomplete picture of drug use trends.
Moreover, the very nature of these surveys leads to
underestimates of the true extent of drug use: people
are generally reluctant to admit to illegal activity,
and added to this, the most high-risk, problematic
drug users are unlikely to be represented in surveys
of households, given that they often live chaotic
lives (Home Office, 2006). The most commonly used
types of surveys therefore bias prevalence estimates
towards conservatism, even if they do so consistently,
in a way that can reliably reveal trends.
Three indicators are typically used to measure
the prevalence of drug use in a given country or
jurisdiction: lifetime prevalence, last-year prevalence
and last-month prevalence. Of these three indicators,
lifetime prevalence – the number of people who
have ever used drugs – is the least useful, because
by definition, such use cannot be reversed: even if
people stop using drugs, they will remain lifetime
users. Lifetime prevalence therefore tends to rise
consistently over time. (This measure is, however,
considered more useful for schoolchildren, as
initiation into most drug use occurs in teenage years,
so lifetime prevalence can provide a better snapshot
of drug-use trends among this group.) Last-year and
last-month prevalence are better proxy measures of
current levels of drug use, and should be used to track
emerging consumption trends (EMCDDA, 2002).
Measuring the prevalence of drug use
9
In fact, the vast majority of people who use drugs would
not fit this description. The United Nations Office
on Drugs and Crime (UNODC, 2015) – the agency
that oversees the international prohibitionist drug
control system – estimates that approximately 90%
of people who use illicit drugs worldwide do so nonproblematically.1
Hence prevalence of use alone is not
a particularly useful or accurate indicator of harm: it
conflates both problematic and non-problematic drug
consumption, even though the risks associated with
each, and the responses that they call for, differ greatly.
A 17-year-old heavy user of crystal methamphetamine
who is supplied by a criminal market and commits
crimes to feed their habit, risks causing far greater
harm to themselves and others than does a middleaged,
occasional cannabis user who grows their own
plants. If the latter ceased their drug-taking while
the former did not, the prevalence of drug use in this
sample would fall by half, and yet would produce little,
if any, change in total levels of health and social harm.
While this distinction between the relative harms
associated with different types of drug use is
acknowledged in the academic debate on reform, it is
frequently overlooked in the public debate – politicians’
press releases and media soundbites talk all too often
about the prevalence of drug use as it were a single,
homogenous phenomenon.
That overall prevalence of use is a poor proxy measure
for aggregate harm can be seen from the situation in
the UK. In 2014, Prime Minister David Cameron said:
“We have a policy which actually is working in Britain:
drug use is coming down” (quoted in Wallis Simons,
2014). The second part of this claim was essentially
accurate at that time: the long-term picture was of a
decline in the overall prevalence of drug use in the
UK (although, since around 2010, use had actually - The UNODC defines “problem drug users” as “people who engage
in the high-risk consumption of drugs, for example people
who inject drugs, people who use drugs on a daily basis and/or
people diagnosed with drug use disorders or as drug-dependent
…” http://www.unodc.org/documents/wdr2014/World_Drug_
Report_2014_web.pdf xvii
stabilised [Home Office, 2015]). Yet the picture is more
complicated – and less impressive – when the broader
context is examined.
The decline in overall levels of drug use that occurred
in the previous decade was driven mostly by a
reduction in the number of young people (aged 16-24)
using cannabis, the most widely used illicit drug (see
box, p. 10, for speculation on the causes of this trend).
The use of other drugs has, however, proven more
resilient. Between 2012/13 and 2014/15, the number of
young people who took ecstasy in the past year almost
doubled,to levels not seen since 2003, and young
people’s cocaine use also sharply increased over the
same period (Home Office, 2015).
But despite this, overall prevalence was either in
decline or stable. So does this consumption trend
indicate that UK drug policy is “working”? A more
useful indicator suggests not. Fewer people using drugs
does not mean fewer people being harmed by drugs.
The drug-induced mortality rate among adults in the
UK was 55.9 deaths per million in 2013, almost three
times the most recent European average of 19.2 deaths
per million (EMCDDA, 2016). Official 2014 estimates
for the whole of the UK are not yet available, but in
England and Wales, drug-induced deaths reached the
highest levels ever recorded (ONS, 2016). This trend is
being driven by a complex mixture of factors, mostly
related to opioid consumption. Nationally and locally,
an ageing population of users, changes in available
treatment options, and batches of particularly strong or
contaminated opioids are behind the rise. These deaths,
and the reasons for them, underline the importance
of not taking headline figures on use as a simple
determinant of policy success or failure.
It is also notable that, historically, governments tend
to cite any change in levels of drug use as evidence to
support their position: when use falls, it is heralded
as a triumph that renders any debate about reform
irrelevant; but when use rises, calls are made for
enforcement efforts to be intensified. As such, a fixation
on levels of use can often shut down vital discussions
about policies that could deliver better outcomes.
10
From 2000-14, levels of cannabis use among 15- to
34-year-olds remained stable or increased in many
European countries, including France, Sweden,
Denmark and Finland (EMCDDA, 2015b). In England
and Wales, however, consumption among this group
fell by almost half over the same period (EMCDDA,
2015a). Listed below are several possible explanations
that have been put forward for this trend. (It should
be noted that these are all speculative explanations,
reflected by the references given for them, most of
which are to news articles that feature interviews
with drug policy experts, rather than peer-reviewed
research. These proposed explanations should
therefore be treated with caution, although they
usefully illustrate how drug use is influenced by
myriad factors, rather than policy alone.)
• The decline in tobacco smoking in the UK. The
fall in the number of tobacco smokers, which
predates the fall in cannabis use, may have
helped de-normalise smoking more generally.
And since Britons commonly consume cannabis
in a mix with tobacco, fewer non-smokers may
be taking up the drug (The Economist, 2015).
• The rise of high-strength “skunk”. The UK
cannabis market is increasingly dominated
by more potent varieties of the drug that are
unappealing to novice users given the higher
probability of negative experiences (McVeigh and
O’Neill, 2012).
• A cultural shift. The decline in young people’s
cannabis use is mirrored by similar, but less
pronounced, declines in alcohol consumption and
the use of other drugs, pointing to the possible
emergence of a more abstemious and risk-averse
youth culture (McVeigh and O’Neill, 2012;
Benedictus, 2011).
• The rise of the internet and social media. Smart
phones, the internet and gaming consoles may
be having an effect on young people’s behaviour
in relation to drugs (Goldhill, 2014; Barnes, 2012).
The rise of such technology may be reducing
boredom or the amount of “dead time” that might
otherwise be filled by casual drug use, or it may
be reducing real-world interactions in which
drug sharing or peer pressure come into play
(Cabinet Office and Department of Health, 2015).
Online photo sharing may also be increasing
image and body consciousness, turning people
off more unhealthy or unattractive drug-using
behaviours.
• The rise of new psychoactive substances. Some
young people may be switching from real
cannabis to synthetic cannabinoids, such as
“Spice”, which mimic its effects and were, until
a blanket ban in 2016, relatively cheaply and
legally available via high-street “head shops”.
Comparatively little is known about levels of use
of such substances (EMCDDA, 2015b).
• Societal changes. More young people are living
at their parents’ home than in previous years
(ONS, 2011), which may be limiting opportunities
for illicit behaviour such as cannabis use. The
employment rate for 16-17-year-olds has also
declined substantially over the past decade (ONS,
2013), potentially meaning fewer teenagers have
the money to purchase cannabis.
Declining cannabis use in England and Wales: possible causes
11 - d) Beneficial drug use?
The use of illicit drugs, even when moderate, controlled,
and recognised as being relatively low-risk, is still
typically characterised as antisocial or having net
adverse effects. There is a general reluctance to accept
or admit that currently illicit substances may confer
any benefits to users or wider society.
In contrast, this point is often acknowledged with
regard to the use of licit drugs. The UK government’s
2012 Alcohol Strategy, for example, says:
“In moderation, alcohol consumption can have a
positive impact on adults’ wellbeing, especially where
this encourages sociability. Well-run community
pubs and other businesses form a key part of the
fabric of neighbourhoods, providing employment
and social venues in our local communities. And
a profitable alcohol industry enhances the UK
economy” (HM Government, 2012: 3).
It is difficult to imagine any government making a
similar statement in reference to illicit drugs, but the
logic is sound: most currently illegal substances can
be – and most often are – consumed responsibly, with
negligible harm resulting either to users or wider
society. (And where harm to wider society does occur,
it is often a product of the criminal nature of the drug
trade, rather than drug use itself.)
All of this is not to say that policy should not attempt
to deter people (particularly young people or other
vulnerable populations) from using drugs – especially in
high-risk ways; it is simply to highlight the distinction
between drug use and drug harm, and place this
distinction within the context of the reasons why
people take drugs in the first place.
Pleasure is the “great unmentionable”(Hunt and Evans,
- in drug policy research and the public debate
on reform (Moore, 2008; Holt and Treloar, 2008) but
given the central role it plays in motivating various
forms of drug use, it must be factored into thinking
around policy responses to changing levels of use. That
pleasure – or any quasi-medical or lifestyle “benefit”,
such as relaxation, stress relief, or enhanced social
experiences – is the primary desired outcome of drug
use is indisputable, and for most drug users suggests
a willingness to bear at least some degree of risk to
achieve that outcome (Ritter, 2014).
Casual / non-problematic use
Recreational, casual or other use that has
negligible harmful health or social effects, e.g.,
moderate cannabis, cocaine or MDMA use in
social settings
Chronic dependence
Use that has become habitual and compulsive
despite negative health and social effects, e.g.,
long-term opioid dependence, which is funded
through acquisitive crime
Beneficial use
Use that has positive health, spiritual or
social effects, e.g., medical pharmaceuticals,
stimulants – such as coffee or tea – to increase
alertness, sacramental use of ayahuasca,
therapeutic use of MDMA
Problematic use
Use that begins to have negative
consequences for individual, friends / family,
or wider society, e.g., use leading to impaired
driving, binge consumption, harmful
methods of administration
Spectrum of psychoactive substance use
Adapted from: British Columbia Ministry of Health Services (2004)
12
In fact, there are countless pleasure-seeking activities
– be it sport, sex, sunbathing, or the consumption of
sugar or fatty foods – that people are willing to engage
in despite their risks. Society therefore accepts that a
certain level of risk is permissible in order to achieve
certain pleasures. Sometimes – in the case of, say,
motorbike racing or some extreme sports – the risks are
very high, and may not only be tolerated, but viewed as
an intrinsic part of the experience.
But with drugs, the overriding concern with reducing
use has marginalised a more pragmatic and constructive
debate about what motivates consumption, what
level of drug-related risk or harm should be tolerated,
and which policy approaches can help manage and
moderate those risks. Instead, unlike with other risky
pleasure-seeking activities, there is a range of groups
and institutions dedicated to exaggerating potential
harms and denying benefits. Rugby or American
football, as played by the typical school student, is likely
to carry far more risk of damaging the developing
brain than cannabis, but there is no agency devoted to
publicising those risks in the media.
- The impact of different
policy models on the
prevalence of drug use
There are various approaches that can be taken in
response to the use of currently illicit drugs, many of
which – such as the introduction of harm reduction
services or the intensification of enforcement efforts –
can take place under a range of policy models. Rather
than consider the impact of these kinds of relatively
more incremental reforms on the prevalence of drug
use, this report focuses on how, if at all, drug use differs
under the three most commonly discussed overarching
policy regimes. It will first look at drug use under
prohibition, then under two options for reform – the
decriminalisation of personal drug possession and the
legal regulation of drug markets. It should be noted
that there can be significant variation in the design and
implementation of each of these policy models (drug
prohibition in China is far more punitive than in the
UK, for example). There is not space here to explore
all the possible forms these policies can take, so this
As discussed, care must be taken when inferring
levels of drug-related harm from levels of drug use.
The two are far from equivalent; that is why other
indicators must be considered in order to determine
whether drug policy is promoting public health and
safety effectively.
Evidence from studies of alcohol use illustrate this
point further. As one cross-sectional survey of the
effects of alcohol in Russia, the Czech Republic and
Poland found: “Overall alcohol consumption does
not suffice as an estimate of alcohol related problems
at the population level” (Bobak et al., 2004). This
was because the Czech Republic, despite having
significantly higher levels of alcohol use than
Russia, experienced far less alcohol-related harm, as
measured by the proportion of people whose drinking
causes difficulties with, among other things, their
relationships, work, physical, psychological or mental
health and financial circumstances.
Emphasising the way in which culture and patterns
of consumption are important determinants of harm,
this result was attributed to the fact that Czechs,
while drinking more often and more as a whole,
consume relatively small amounts per occasion,
whereas Russians’ drinking is more concentrated
– they drink less frequently, but consume large
amounts of alcohol when they do (and such patterns
of binge use tend to be associated with higher health
and social harms). Studies of other countries have also
found that low overall levels of alcohol use are by no
means a guarantee of low levels of alcohol-related
harm (Rossow, 2001; Ramstedt, 2001; Christie, 1965;
Poikolainen, 1977; Room, 1974; Norström, 2001).