Professional Documents
Culture Documents
One of the core objectives of supply-side drug law enforcement is to reduce drug use by raising the cost of buying drugs.
The effectiveness of this strategy depends on how illicit drug users respond to the rise in costs. The aim of the current study
was to estimate how methamphetamine users would respond to changes in the prices of methamphetamine and heroin,
using hypothetical drug purchasing scenarios. A sample of 101 people who reported using methamphetamine in the past
month was recruited from primary healthcare facilities in Sydney, Wollongong and Newcastle. Participants were given a
hypothetical fixed drug budget, presented with a range of drug ‘pricelists’ and asked how many units of each drug on the
pricelist they would buy with their drug budget. The prices of methamphetamine and heroin were varied independently
across successive trials and the quantity of each drug purchased at each methamphetamine and heroin price was recorded.
Results revealed that methamphetamine purchases decreased significantly as the price of methamphetamine increased
(a 10% price increase led to an 18%-19% fall), as did heroin purchases in response to heroin price increases (a 10% price
increase led to a 16%-27% fall). Among methamphetamine users, increases in methamphetamine prices produced some
substitution into heroin. Additionally, dependent methamphetamine users purchased more pharmaceutical opioids while
the non-dependent group purchased more cocaine. Dependent and non-dependent heroin users also responded differently
to changes in the price of heroin. Dependent heroin users reacted to increased heroin prices by significantly increasing
their purchases of methamphetamine, pharmaceutical opioids and benzodiazepines. At the same time they purchased
less cocaine. Non-dependent heroin users responded simply by increasing their purchases of pharmaceutical opioids. In
most instances where substitution occurred, the fall in consumption of amphetamine (or heroin) was considerably greater
than the increase in consumption of other drugs.
KEYWORDS: price elasticity, cross-price elasticity, supply-side drug law enforcement, methamphetamine, ice.
drug use (e.g. school-based education elasticity of demand. For a particular drug, surveys or from under-cover police drug
programs) or treatment for dependent own-price elasticity of demand measures buying operations. Consumption on
users (e.g. cognitive behavioural the change in consumption of that drug the other hand, is typically measured
counselling or pharmacotherapies). as a consequence of variation in its own either directly via self-report surveys
Not all demand reduction activities are price. The more consumption falls in (e.g. Becker, Grossman & Murphy 1991;
effective but there is some evidence in response to a price increase, for example, Grossman & Chaloupka 1998; Saffer
support of programs such as school- the more elastic is demand. Price elastic & Chaloupka 1999; see Grossman et
based drug prevention (e.g. Caulkins demand describes the situation where al. 2002 for a review) or indirectly from
et al. 2002). While pharmacotherapies consumption responds to price changes administrative data sources such as
have been well established for heroin to the extent that an x per cent increase drug-related emergency department
dependence for many years, there in price leads to at least an x per cent admissions or the proportion of arrestees
has typically been much less success reduction in consumption. In contrast, testing positive for drugs (Silverman &
in developing such treatments for price inelastic demand is said to occur Spruill, 1977; Caulkins, 1995; 2001; Dave,
methamphetamine and other psycho- when an x per cent price increase results 2004; 2006). Manski, Pepper and Petrie
stimulants. Nevertheless, recent clinical in a less than x per cent reduction in (2001) reviewed a large number of studies
trials have provided some promising signs consumption. If demand is price elastic, that used both direct and indirect methods
of progress (e.g. Shearer et al. 2009). pushing up the price of an illicit drug and estimated that the own-price elasticity
will produce a reduction in its use and of demand for cocaine ranged from -.59 to
Harm reduction strategies are not intended
a reduction in overall expenditure on -2.5 (i.e. ranging from relatively inelastic
to reduce drug use per se but instead
the drug. If demand is inelastic, pushing to highly elastic).
aim to reduce the harms associated
up the price of a drug may produce
with substance abuse (e.g. needle There is strong evidence from studies that
some reduction in its use but overall
and syringe programs and supervised use direct measures of use that own-price
expenditure on the drug will increase.
injecting facilities are intended to reduce elasticity of demand for heroin is relatively
the transmission of blood borne viruses). The way in which consumption of a elastic. Saffer and Chaloupka (1999), for
Again, some harm reduction efforts will drug changes in response to variations example, estimated the price of cocaine
be more effective than others but there is in its price is only part of the picture. and heroin from the US Department of
certainly evidence that strategies such as Methamphetamine users, like most drug Justice’s database of undercover drug
needle and syringe programs have been users, tend to use a range of substances. purchases and measured consumption
effective in limiting the harms associated Some drugs are used in concert with based on the pooled 1988, 1990 and
with injecting use of illicit drugs such methamphetamine (i.e. as complement 1991 national household drug use
as methamphetamine (Commonwealth drugs) while others may be used in place surveys. Saffer and Chaloupka estimated
Department of Health and Ageing 2002; of methamphetamine (i.e. as substitute that a 10 per cent increase in the price
Hurley, Jolley & Kaldor 1997). drugs). Consequently, increases in of cocaine and heroin would result in
the price of methamphetamine may 2.8 and a 9.4 per cent decrease in
Supply reduction efforts are those
result in decreased consumption of participation in cocaine and heroin,
that aim to intercept illicit drugs before
methamphetamine and some other respectively. While the studies based on
they get onto the market (e.g. through
(complement) drugs but could lead direct measures of use provide important
cross-border interdiction or dismantling
to increases in consumption of other insights into the elasticity of demand for
domestic drug laboratories). One of the
just as harmful (substitute) drugs. A drugs, the self-reported drug use data
core rationales underpinning supply side
more complete understanding of the tend to be collected via household or
law enforcement is that, by increasing
effectiveness of an increase in the price telephone surveys which by their very
the risks associated with trading in illicit
of methamphetamine requires knowledge nature exclude marginalised populations,
drugs, dealers will be forced to increase
of how consumption of all drugs responds including those who might be at greatest
the price of drugs to compensate
to that price increase. For a particular risk of being dependent on illicit drugs.
themselves for the risks they take. Some
drug, cross-price elasticity measures the This can be a problem when estimating
economists argue that, like most other
change in consumption of another drug price elasticity because dependent users
commodities, an increase in the price of
in response to variations in the price of might be less able than non-dependent
a drug should reduce consumption of that
that drug. When the cross-price elasticity users to curb their consumption in
drug. Reductions in use should, in turn,
is positive the drugs are considered to be response to price increases.
bring reductions in drug-related harm.
substitutes and when it is negative they
The relative efficacy of supply reduction Indirect measures of drug use might
are considered to be complements.
efforts therefore depends critically on better reflect the experiences of
the extent to which consumption of illicit Economists have tended to estimate the dependent users because those who use
drugs responds to changes in drug responsiveness of demand for illicit drugs drugs most frequently are also most likely
prices. Economists quantify the consumer using secondary data sources. Price is to come into contact with public agencies
reaction to price change in terms of price typically either estimated from self-report such as hospitals or police. While
B U R E A U O F C R I M E S T A T I S T I C S A N D R E S E A R C H
most well-conducted studies that have about the implications of price changes different drugs and are subsequently
employed indirect measures of use tend in one drug for the consumption of other asked how much of each drug they would
to also find some responsiveness to price drugs. Analyses of drug use among purchase given a fixed drug budget. The
changes, the elasticity estimates have the general population tend to find that researcher then changes the prices of the
varied considerably. For example, using alcohol is a complement for illicit drugs various drugs over successive trials and
data derived from arrestees and from (e.g. Saffer and Chaloupka, 1999; records the participants’ drug purchases
hospital emergency room admissions, Pacula 1998). Furthermore, Saffer and at each of the drug prices. The resulting
Caulkins (1995; 2001) estimated the own- Chaloupka (1999) found complementarity purchase data is used to determine own-
price elasticity of demand for cocaine to between illicit drugs such as heroin, and cross-price elasticities. The theory
be between -1.3 and -2.5 and own-price cocaine and cannabis. Studies using behind this methodological approach is
elasticity of demand for heroin/morphine indirect measures of illicit drug use also that, by virtue of making these purchases
to be between -.84 and -1.5. On the tend to find complementarity between in a controlled environment, any observed
other hand, Dave (2004; 2006) used heroin and cocaine use (e.g. Dave 2004; changes in hypothetical consumption
similar data but a different analytical 2006). Research into the effects of the should be attributed entirely to changes
methodology and estimated much more Australian heroin shortage found that a in price. When purchased in a natural
modest short-run elasticities. In fact, Dave proportion of heroin users substituted setting the quality of the drug can, and
concluded that demand falls within the their use – at least in the short term often does, vary with price.
inelastic range (-.15 to -.27 for cocaine – with other drugs such as cocaine and
In a classic study of this kind conducted
and -.08 to -.10 for heroin). methamphetamine (Maher et al. 2007;
in the United States, Petry and Bickel
Most of the studies reviewed above Weatherburn et al. 2003). It is important
(1998) found that the decrease in
examined the impact of the declining to point out, however, that these studies
consumption of heroin was roughly
heroin prices in post-Vietnam war America observed drug substitution among those
proportional to the increase in its price
and, in particular, the crack cocaine users who were still by definition in
among a group of current and former
epidemic that took hold in America in the the market and this substitution does
opioid-dependent participants. In contrast
1980s. In Australia, the heroin shortage not appear to have occurred at a more
to the results of studies using secondary
that manifested itself around Christmas aggregate level (Snowball et al. 2008).
data sources (Dave, 2004; 2006), Petry
2000 provided compelling support for Nevertheless, this research does suggest
and Bickel found evidence of substitution
the notion that demand for heroin might that heroin and psycho-stimulant drugs
between illicit drugs. Increases in the
be relatively elastic. When the real price might act as short-term substitutes for
price of heroin induced relatively strong
of heroin more than doubled in a very some frequent substance users.
substitution into valium and cocaine
short space of time, consumption of the The literature reviewed above clearly (cross-price elasticities of 1.02 and
drug reduced dramatically, as did other shows that estimates of price elasticity 0.8, respectively) with more moderate
indicators of drug related harm such vary markedly depending on the substitution into cannabis and alcohol
as fatal and non-fatal overdoses and population under study, the nature of the (elasticities for both were 0.5).4 In a more
drug-related property offences (Moffatt, drug in question and the methodology recent study, Jofre-Bonet and Petry
Weatherburn & Donnelly 2005; Smithson employed to correlate measures of price (2008) also found that the own-price
et al. 2004; Weatherburn et al. 2003). and consumption. Manski, Pepper and elasticities of demand for cocaine and
While specific price elasticities were Petrie (2001) outline several reasons for heroin were close to one, regardless of
not estimated in relation to the heroin these varying estimates, including the levels of dependence on one or other of
shortage, Weatherburn and colleagues lack of uniform price data, the variation in the drugs. Heroin-dependent people were
(2003) estimated from their survey of 167 prices charged (even by the same dealer) found to complement their heroin use with
injecting drug users that a one per cent and the heterogeneity of drug users. For cocaine, alcohol and cannabis. The only
increase in the price per pure gram of this reason, a growing body of overseas substitute for heroin was valium. Cocaine-
heroin resulted in a 0.32 per cent fall in
literature has employed behavioural dependent people complemented their
expenditure on the drug.
economics techniques to explore cocaine use with heroin and alcohol but
In summary, the prevailing view, based responsiveness of demand in a more substituted into cannabis and valium. In
on direct and indirect observations of controlled environment (Cole et al. 2008; the only study to consider amphetamine
drug consumption measures, is that Goudie et al. 2007; Petry 2000; Petry & use, Sumnall and colleagues (2004) found
consumption of illicit drugs such as Bickel 1998; Sumnall et al. 2004). These that demand for amphetamine, cocaine
heroin and cocaine is responsive to studies make use of experimental data, and ecstasy was quite elastic among
changes in price. However, the actual whereby current or former drug users poly-substance recreational users in the
elasticity estimates vary depending typically make hypothetical purchases of UK. That is, purchases of each of these
on the populations, data sources and a range of drugs using imitation money drugs decreased at a proportionately
methodologies employed. The economic provided by the researcher. Participants greater rate than the associated increase
research community is far less certain are given a price list outlining several in price. Further, this study revealed
B U R E A U O F C R I M E S T A T I S T I C S A N D R E S E A R C H
that for this population, alcohol was a Method “Think of a typical week. We are going
substitute for amphetamine, while ecstasy to use a price list and some pretend
money to play a sort of game. Assume
and cocaine purchases were independent
Procedure that you have $200 that you can use to
of amphetamine prices.
buy drugs for that week or the length
The methodology, in large part, of time it would normally take you to
There is clearly still much to be learned
follows that of Petry and Bickel (1998). spend $200 on drugs. The drugs that
about how methamphetamine users
Interviews were conducted between are available to you and the price of
respond when faced with significant
August and November 2008 by six those drugs are listed on this sheet.”
changes in price. We cannot be sure
trained interviewers. One hundred The amount of each participant’s drug
that samples of Australian drug users
and one people who were at least 18 budget was held constant at $200 during
would respond in the same way as those
years of age and who reported using
studied internationally. Furthermore, it the experiment.5 The size of the drug
methamphetamine in the previous
may be the case that responsiveness budget approximates the weekly income
month were recruited into the study.
to changes in price varies according support received by single unemployed
Potential participants were identified
to the characteristics of those who Australians at the time of the interviews.
through their attendance at one of
use drugs (e.g. contingent on their As the majority of the study participants
four cooperating agencies in Sydney,
levels of dependence on heroin and/or relied on income support as their primary
Newcastle or Wollongong. Each of
methamphetamine) and the drug market source of income, this budget was
the agencies included a needle and
inhabited by the drug user. This is an intended to provide a realistic frame of
syringe programme. In the first instance,
important knowledge gap in light of the reference for the hypothetical purchases.
participants were directly approached by
considerable investment Australian law Participants were told that they did not
the researchers and invited to undertake
enforcement agencies make in attempting have to spend all of the $200 on the listed
a face-to-face interview. Snowball
to curb illicit drug supply. The current drugs and were advised to hypothetically
sampling was also used, whereby
study used behavioural economics participants who had completed the purchase drugs at the rate they normally
techniques to examine the elasticity of survey informed their friends about the would. As a result, the timeframe over
demand for methamphetamine among a study and these potential participants which participants could spend the $200
sample of Australian methamphetamine approached the researchers. There budget was allowed to vary. For some
users. Because it is well established was no intent to sample a representative participants, this budget was sufficient
that there is a high degree of poly-drug group of methamphetamine users. to cover their actual weekly usage, while
use among populations of injecting Prior to participating in the interview, it would be clearly insufficient for other
drug users (Darke & Hall 1995), we the nature of the study was explained respondents.
also examined how this group would and participants were advised that any Each price list included the following
respond to changes in the price of information provided would be treated drugs: heroin, methamphetamine (base,
heroin. Participants were presented with confidentially. All survey participants then
powder or crystal), cannabis, cocaine,
hypothetical drug buying scenarios and signed written consent forms.
non-prescription benzodiazepines,
the price of methamphetamine and heroin
Most interviews were conducted on site at alcohol, and non-prescription
were systematically varied to determine
one of the four agencies, although a small pharmaceutical opioids (i.e. oxycodone
how consumption would likely respond to
number of interviews were conducted at and morphine). The price of each drug
changes in drug price.
a public housing estate where the local reflects the average price from the 2008
The specific research questions that the needle and syringe program provided a Illicit Drug Reporting System (IDRS)
current study aimed to address were: mobile service. During each interview, survey (Phillips and Burns 2009). The
(a) How much would consumption the interviewer read out all questions baseline prices were $50 per cap of
of methamphetamine change in and recorded participant responses heroin, $50 per point of base, powder or
response to changes in its price? on a paper copy of the questionnaire. crystal methamphetamine, $20 per gram
Each interview ran for 30 to 45 minutes of cannabis, $50 per cap of cocaine,
(b) How much would consumption
and participants were reimbursed $30 $2 per benzodiazepine pill, $30 per
of heroin change in response to
as compensation for their time at the pharmaceutical opioid pill6 and $5 per unit
changes in its price?
completion of the survey. of alcohol. Only one quantity per drug
(c) How much would consumption of
other drugs change in response was employed to avoid issues of quantity
to changes in the price of Questionnaire discounting.
methamphetamine or heroin? The critical part of the questionnaire The interviewer then explained the “rules”
(d) Does responsiveness of involved a series of questions about of the game, as follows:
consumption vary according to hypothetical drug purchases across a “You may buy any drugs that you like with
whether the participant is dependent range of drug prices. The interviewer this money, but you can only spend $200
on methamphetamine or heroin? read aloud the following instructions: on drugs. No other drugs are available
B U R E A U O F C R I M E S T A T I S T I C S A N D R E S E A R C H
to you aside from those you buy with dependence on methamphetamine and techniques were employed to estimate
this money. For example, you cannot severity of dependence on heroin (as the elasticities reported in this bulletin.
use drugs given to you by a friend and
indexed by the Severity of Dependence Several models were considered but
you cannot buy any more drugs once
Scale; Gossop et al. 1995). ultimately zero-inflated negative binomial
you have spent the $200. Even if you
have drugs stashed away, you cannot and zero-inflated poisson models were
use them. You can’t bargain with your Analysis found to provide the best fit to the data.7
dealer or buy in bulk to reduce the price. All analyses were undertaken using
Also, you cannot use the $200 to buy a The behavioural economics literature Stata/SE 10.1. Stata’s cluster estimator of
drug not listed. The drugs that you buy has typically measured price elasticity the variance covariance matrix was used
can only be used by you and cannot be in terms of the bivariate relationship
given away or sold. Finally, the prices of
to account for the non-independent nature
between hypothetical purchases of the outcome variable (Baum 2006, pp.
some of the drugs will change across
the different price lists but the quality of
(or consumption) and drug prices, 138-139).
the drugs does not change with price operationalised as the percentage change
changes. The drugs are of the same in hypothetical consumption in relation to
Calculating price elasticity
quality that you’d expect to get from the percentage change in price. The own-
your regular supplier and when the drug
from count data analysis
price elasticities for methamphetamine
prices change I want you to imagine The estimated coefficient for a continuous
and heroin and cross-price elasticities
that the prices will stay that way for the
for all the drugs in relation to the prices explanatory variable (such as price)
foreseeable future. Please purchase
the drugs you would like from the list.” of both methamphetamine and heroin represents the proportionate change in
are measured by the slope of the plot hypothetical purchases brought about by
Illicit drug market adjustments do not a $1 change in price, with all explanatory
of the price and consumption on log-
only manifest as changes in the price of variables set to their mean values. The
log coordinates calculated by linear
a drug. More often than not it is the purity estimated elasticity with respect to price
regression (Petry and Bickel 1998; Petry
or quality of the drug that fluctuates and is simply the mean price ($51) multiplied
2000; Goudie et al. 2007). We follow this
price changes tend to be accompanied by the estimated coefficient (Cameron et
methodology by estimating the impact of
by variations in the purity or quality of al. 1998).
methamphetamine price on the amount of
the drug (Caulkins 2007). Hence, it was
each drug purchased.
important to impress upon participants
that the quality did not vary with price. The data obtained from this survey Results
contain two features that complicate the
Following the presentation of the base
analysis of changes in the quantity of
price list, eight further price lists were Participants
drugs purchased. Firstly, all of the 101
presented to each participant in a The characteristics of the 101 participants
participants provided multiple purchase
randomised order and the prices of are summarised in Table 1. Most
amounts in response to the nine unique
methamphetamine and heroin were participants were men (n=58). The
price lists. It is problematic to analyse
varied one at a time with each successive mean age of participants was 38.6 years
the data as though the hypothetical
presentation. In addition to the baseline (range=18-58) and 84 had not completed
purchases represent a simple random
prices of $50 per cap of heroin or point any formal education beyond years 10
sample because responses are correlated
of methamphetamine the following prices or 11 of school. A little over half of the
within individuals. In other words,
were presented for both drugs: $10, $20, participants (n=52) reported that their
the amount any individual is willing
$75, and $100. This range of prices was primary source of income was income
to purchase at a given price is highly
selected because it was within the bounds support. All but five of the respondents
related to the amount they are willing
of what street-based samples of illicit indicated that they usually purchased the
to purchase at another drug price. The
drug users in Sydney reported paying for illicit drugs that they used. On average,
second complicating feature of these
heroin and methamphetamine (Phillips participants reported that they were 16
data is that the hypothetical units of
and Burns 2009). years of age when they first used any
a drug purchased are non-negative
illicit drug and 21 years old when they first
A number of other measures were integers, or counts (e.g. 0, 1, 2, 3, … 20
used methamphetamine.
also collected from participants points of methamphetamine). Because
both to contextualise the patterns of these count data are bounded by zero Heroin was the most often cited main
drug use among this sample and to and concentrated around a few discrete drug of choice at the time of the interview
observe how these characteristics values – often close to zero – the (n=42) with a third of respondents
interacted with willingness to purchase distribution is highly skewed. The discrete, citing methamphetamine as their main
methamphetamine and other drugs. skewed and non-independent nature drug of choice (n=33). One quarter
These characteristics included a range of the data violates the assumptions (n=26) of the respondents reported that
of demographic characteristics (e.g. age, that underpin ordinary least squares methamphetamine was the last drug used
sex), patterns of drug use, severity of regression and alternative regression while 39 participants last used heroin.
B U R E A U O F C R I M E S T A T I S T I C S A N D R E S E A R C H
B U R E A U O F C R I M E S T A T I S T I C S A N D R E S E A R C H
B U R E A U O F C R I M E S T A T I S T I C S A N D R E S E A R C H
B U R E A U O F C R I M E S T A T I S T I C S A N D R E S E A R C H
Table 2. Own-price and cross-price elasticity of demand when methamphetamine and heroin price change
Pharm.
Meth. Heroin Alcohol Cann. Cocaine opioids Benzos
Methamphetamine price
Methamphetamine dependent -1.766*** 0.200*** -0.011 0.050 0.121 0.333*** 0.070
Not methamphetamine dependent -1.866*** 0.054** -0.046 0.036 0.705*** 0.137 -0.152
Heroin price
Heroin dependent 0.255*** -1.553*** 0.099 -0.094 -0.538*** 0.395*** 0.308***
Not heroin dependent 0.013 -2.674*** -0.009 0.038 -0.202 0.764* 0.054
*** p<=.01, ** p<=.05, *p<=.10
B U R E A U O F C R I M E S T A T I S T I C S A N D R E S E A R C H
10
B U R E A U O F C R I M E S T A T I S T I C S A N D R E S E A R C H
11
B U R E A U O F C R I M E S T A T I S T I C S A N D R E S E A R C H
Cannabis (grams) 45(-10) 41(-18) 50 47(-6) 45(-10) In short, the results suggest that demand
Cocaine (caps) 9(13) 8(0) 8 6(-25) 7(-13) for both methamphetamine and heroin
appears to be in the elastic range. We
Pharmaceutical opioids (pills) 3(-77) 4(-69) 13 19(46) 17(31)
estimated that the own-price elasticities
for these two drugs were greater than
one, which suggests that any increase in
Discussion dependence but non-dependent the price of these drugs would result in a
heroin users were found to be decrease in consumption and a decrease
The results of the current study can be significantly more responsive to own- in overall expenditure on these drugs.
summarised as follows: price changes than dependent users. However, there was also evidence of
• Demand for both methamphetamine This greater responsiveness among substitution into other drugs when the
and heroin was estimated to be non-dependent heroin users was price of heroin and methamphetamine
relatively elastic (a 10% increase largely attributed to the low baseline rose. This substitution must be viewed
in methamphetamine price was rates of heroin use among this group; with concern. Here, it is important to
estimated to result in a 18 - 19% acknowledge that these findings are
• Increases in the price of
decrease in the quantity of specific to the study participants, who
methamphetamine were estimated
methamphetamine purchased and a should not be taken as representative of
to lead to significant substitution
corresponding increase in the price methamphetamine users as a whole, and
into heroin (cross-price elasticity
of heroin would lead to a 16% - 27% who operate in particular drug markets.
[CPE]=0.200 among dependent users
reduction in the quantity of heroin Cocaine, for example, has limited
and 0.054 among non-dependent
purchased); availability outside of Sydney.
users), pharmaceutical opioids
• Own-price elasticity of demand for (CPE=0.333 among dependent users) Weighing up the net effect of these price
methamphetamine did not vary and cocaine (CPE=0.705 among non- changes in light of this substitution is
according to methamphetamine dependent users). Substitution was a difficult task. Because we were not
12
B U R E A U O F C R I M E S T A T I S T I C S A N D R E S E A R C H
interested in testing the income elasticity over the longer term (Snowball et al. prices was also consistent with the
of demand, the study design involved 2008). range reported among illicit drug users
giving participants a fixed budget and a in the Sydney market (Phillips and
Like Saffer’s and Chapoulka’s (1999)
variable amount of time in which to spend Burns 2009). Ajzen and Fishbein (1980)
study, we found complementarity between
that budget. As a result, most participants proposed in their often-cited ‘theory of
heroin, cocaine and cannabis, at least
spent their entire drug budget on each reasoned action’ framework that, in such
among the heroin dependent group.
trial and overall expenditure varied little circumstances, people’s behavioural
On the other hand, the current findings
with drug prices. Based on the elasticity intentions usually closely reflect their
on cross-price responsiveness are not
estimates and baseline purchases of the behaviour. Indeed, a good example of
consistent with the only behavioural
various drugs, we suspect that the overall the close alignment of intention and
economics study to have explored
decrease in heroin and methamphetamine behaviour can be found from observing
poly-drug use involving amphetamines.
purchases would outweigh the observed electoral opinion polls. If conducted with
While Sumnall et al. (2004) established
substitution into other drugs. When methodological rigour, telephone polls
that demand for amphetamine was
methamphetamine prices were varied, of usually very closely predict electoral
relatively elastic, increases in the price
most concern was substitution into heroin outcomes. Nevertheless, that there is a
of amphetamine were found to have no
use among those who met the criteria close association between intentions and
impact on purchases of cocaine and
for methamphetamine dependence. behaviour remains an assumption of the
ecstasy although they did result in some
However, in response to a $10 increase methodology employed here that we are
substitution into alcohol. These discrepant
in the price of methamphetamine from not able to test.
findings are likely to be a function of
$50 (20 per cent) the estimated increase
the populations of drug users sampled. While every effort was made to ensure
in heroin purchases from baseline would
be in the order of two caps of heroin Sumnall et al. (2004) investigated a that the scenarios reflected real-world
(from 56 caps) for the entire group, younger group of drug users that reported drug market characteristics, the complex
compared with a decrease of 30 points a preference for ecstasy, whereas the characteristics of real world drug-markets
of methamphetamine (from 84 points).11 current study focused on a comparatively cannot be perfectly simulated in an
For heroin dependent respondents, the older population of drug users, of which experimental setting. The experimental
biggest concerns were substitution into a significant portion regarded heroin or design provided respondents with a fixed
benzodiazepines and methamphetamine methamphetamine as their drug of choice. budget, which is often not the case in
among heroin-dependent participants. Furthermore, in the United Kingdom, reality. Income can vary depending on
Again, however, the analogous price there is some evidence that amphetamine availability of legitimate employment or
increase from base rate, would translate use is decreasing amongst poly-drug through willingness to commit criminal
into nine additional benzodiazepine pills users (Sumnall et al. 2004), whereas in activity to raise income to purchase
(from 152 pills) and three extra points of Australia, as reviewed in the introduction, illicit drugs. While the majority of the
methamphetamine (from 57 points). This methamphetamines remain one of the respondents were reliant on a fixed
compares with a decrease of 34 caps of most commonly used illicit drugs. income – income support – a notable
heroin (from 110 caps). proportion received most of their income
Based as they are on hypothetical drug
from criminal activity or sex work, which
The current findings support the findings purchases, the applicability of the results
are more flexible income flows. As the
of similar behavioural-economics studies depends on the degree to which stated
price of a drug increases, one real world
and economic studies of secondary data preferences would translate into actual
response would be to undertake more
that demand for an illicit drug is relatively behaviour. However, there are prima
crime (although there is, of course, an
elastic with respect to its own-price (e.g. facie reasons to have confidence that
upper limit to the amount of crime or sex
Goudie et al. 2007; Grossman et al., people would behave in the way they
work one person can do).
2002). They are also consistent with say they would. We followed a well-
observations from the Australian heroin respected experimental approach. The A further limitation of the current
shortage that took hold around Christmas scenario conditions were defined so that findings is that we did not attempt
2000. As the real price of heroin doubled they reflected real-world drug markets to obtain a representative sample of
in a short space of time, indicators as closely as possible. The fixed income methamphetamine users, nor did we
of heroin use declined rapidly. While was set at a level that approximated make any attempt to control for the
ethnographic research revealed that weekly levels of income support – on characteristics of our sample other
many primary heroin users supplemented which most people in the sample received than their level of dependence. While
their heroin use with methamphetamine as their primary source of income – and we cannot, therefore, generalise
while heroin supply was low (Maher et drug prices were set at the median levels the results of the current study to all
al., 2007), the supplementation does not estimated from sentinel groups of illicit methamphetamine users, it was not our
appear to have been large enough to drug users such as the sample obtained intention to do so here. Our primary aim
manifest itself in population-level data here. The range of methamphetamine was to determine how responsive sentinel
13
B U R E A U O F C R I M E S T A T I S T I C S A N D R E S E A R C H
groups of frequent methamphetamine Acknowledgements mean and the negative binomial model
users might be to changes in its price and is more appropriate since, by allowing for
that of heroin. Nevertheless, it is possible This research was funded by NSW heteroskedasticity, the variance can differ
that other groups of methamphetamine Health with in-kind support from the Drug from the mean (Greene 2008; Kennedy
Policy Modelling Program (DPMP). The 2008). However, the standard negative
users might respond differently than the
authors would like to thank the agency binomial model was also found to provide
group studied here.
a poor fit to the data here due the large
management and staff for facilitating
Perhaps the most critical limitation of number of zero counts. Zero-inflated
recruitment and interviews, Rebecca
the approach used here was that illicit regression models (poisson and negative
Mcketin for sharing her wisdom about all
drug market adjustments are more likely binomial) were adopted to account for the
things “methamphetamine”, Erin Kelly
to manifest as changes in the purity high number of zero responses. The zero-
and Natasha Sindicich for assistance in inflated models differentiate between two
and/or quantity of drug sold for a given
recruiting agencies, Colleen Faes and groups of respondents - those who would
price (Caulkins 2007). Concerned about
other DPMP colleagues for undertaking not purchase the drug at any price and
the difficulties associated with varying
the interviews, Tim Slade and Barbara those who would be willing to purchase the
quality in an experimental setting we, in
Toson for statistical advice, Michael drug at some price (Greene 2008, pp.922-
a sense, proxy such quality changes as
Lodge for helpful comments and Don 924).
price changes. Again we impressed upon
Weatherburn for valuable feedback. 8. The Friedman two-way analysis of variance
the respondents the need to assume that
We also thank Florence Sin for desktop by ranks test was applied to determine
quality, purity and quantity were constant
publishing this report. But most of all we whether the median of the six groupings
over the life of the experiment. However,
owe a debt of gratitude to the participants of hypothetical purchases was equal for all
some respondents were reluctant to buy
for their patience and enthusiasm for “the prices (all participants, methamphetamine
methamphetamine when its price was $10
game”. dependent and not methamphetamine
on the basis that this price would normally dependent vis-à-vis methamphetamine
be associated with poor quality product. price and all participants, heroin dependent
Gaining an understanding of the complex Notes and not heroin dependent vis-à-vis heroin
interplay between drug price, drug purity price) (Siegel and Castellan, 1988). The
and drug consumption should be at the 1. Drug Policy Modelling Program, National tests indicated that in each of the groupings
Drug and Alcohol Research Centre, UNSW
forefront of future attempts to understand there was a significant difference between
consumer responsiveness to changes in 2. NSW Bureau of Crime Statistics and the median purchases for at least two
drug market characteristics. Research of the prices at the 1 per cent level of
significance. The Page test, a distribution
3. In this paper methamphetamine includes
free rank-test for ordered alternatives
amphetamine and methamphetamine but
Conclusion suitable for repeated measure data, was
excludes ecstasy.
then used to determine whether price was
In summary, the current study found 4. However, purchases of heroin were inversely related to hypothetical purchases
that, for this group of methamphetamine independent of changes in the price of (Siegel and Castellan, 1988). For each
users, consumption of heroin and valium, a finding later confirmed by Petry grouping the Page test suggests that
methamphetamine fell substantially in (2001) in a study examining drug use there is a statistically significant inverse
response to increases in their price. among alcoholics. relationship between purchases and the
While there was evidence of substitution relevant price; at the one per cent level
5. Behavioural economic analyses and
into other drugs, this substitution proved of significance for the three groupings
economic analyses have both found that
pertaining to methamphetamine price, at
nowhere near large enough to offset the purchases of a drug are contingent on
the one percent level for all participants and
reductions in methamphetamine and available income (e.g. Petry and Bickel
heroin dependent participants in relation
heroin use. In determining the net effect 1998; Bretteville-Jensen 2006) and income
to heroin price, but only at the five per cent
on drug use this analysis illustrates the elasticity of demand was not the focus of
level for non-heroin dependent participants
importance of distinguishing between this study.
in relation to heroin price.
the behavioural responses of those 6. One tablet represented 100mg of morphine
9. The Friedman tests revealed that purchases
dependent on the drug in question and and 80mg of oxycodone.
of cannabis and pharmaceutical opioids
those who are not dependent, and taking
7. Count data is typically analysed with a were significantly related to the heroin price
account of the rates of drug use prior to
Poisson regression model where the at the one per cent level in both groups.
any price change. These findings lend
expected hypothetical units purchased is In the dependent group, so too were
support to one of the primary levers by modelled as a function of the explanatory cocaine purchases. Methamphetamine and
which supply side drug law enforcement variables. However, the Poisson model benzodiazepine purchases also showed
policy aims to limit drug-related harm, makes the strong assumption that the a significant relationship at the five per
which is to put upward pressure on drug expected count (mean of units purchased) cent level. In the non-dependent group
prices and, in turn, downward pressure on is identical to its variance. Often, as was alcohol was significantly related to price at
drug consumption. the case here, the variance exceeds the the five per-cent level. Visual observation
14
B U R E A U O F C R I M E S T A T I S T I C S A N D R E S E A R C H
suggests that there is substitution into Department of Justice, National Institute Degenhardt, L, Roxburgh, A, Black,
pharmaceutical opioids in both groups with of Justice, NCJ 173823, Washington DC. E, Bruno, R, Campbell, G, Kinner, S &
the dependent group also substituting into Fetherston, J 2008, ‘The epidemiology
methamphetamine and benzodiazepines. Caulkins, JP 2001, ‘Drug prices and
of methamphetamine use and harm in
Members of this group also appear to emergency department mentions for
Australia’, Drug and Alcohol Review, vol.
purchase cocaine with heroin. Page tests cocaine and heroin’, American Journal of
27, pp. 243-252.
revealed that when all participants were Public Health, vol. 91, no. 9, 1446-1448.
considered there was a significant inverse Degenhardt, L & Topp, L 2003, ‘’Crystal
Caulkins, JP 2007, ‘Price and purity
relationship between the heroin price meth’ use among polydrug users in
and cocaine at the 10 per cent level and analysis for illicit drug: data and
Sydney’s dance party subculture:
significant positive relationships between conceptual issues’, Drug and Alcohol
characteristics, use patterns and
heroin price and both methamphetamine Dependence, vol. 90, pp. S61-S68
associated harms’, International Journal
and pharmaceutical opioids at the five per
Caulkins, JP, Pacula, RL, Paddock, S of Drug Policy, vol. 14, pp. 17-24.
cent and one per cent levels respectively.
& Chiesa, J 2002, School-based drug
However, when grouped by level of González-Sáiz, F, Domingo-Salvany, A,
dependence only the relationship between
prevention: what kind of drug use does it
Barrio, G, Sánchez-Niubó, A, Brugal,
heroin price and methamphetamine in the prevent? Drug Police Research Center,
MT, de la Fuente, L & Alonso, J 2009,
dependent group maintained significance. Rand Corporation, Santa Monica,
‘Severity of Dependence Scale as a
California.
10. We thank our anonymous reviewer for diagnostic tool for heroin and cocaine
bringing this to our attention. Cho, AK & Melega, WP 2002, ‘Patterns dependence’, European Addiction
of methamphetamine abuse and their Research, vol. 15, pp. 87-93.
11.These estimates are based on the
elasticities shown in Table 2. consequences’, Journal of Addictive
Gossop, M, Darke, S, Griffiths, P, Hando,
Disorders, vol. 21, pp. 21-34.
J, Powis, B, Hall, W & Strang, J 1995,
Cole, JC, Goudie, AJ, Field, M, ‘The Severity of Dependence Scale
References
Loverseed, A, Charton, S 2008, ‘The (SDS): psychometric properties of the
Ajzen, I & Fishbein, M (Eds) 1980, effects of perceived quality on the SDS in English and Australian samples of
Understanding attitudes and predicting behavioural economics of alcohol, heroin, cocaine and methamphetamine
social behaviour, Prentice-Hall, New methamphetamine, cannabis, cocaine, users’, Addiction, vol. 90, no. 5, pp. 607-
Jersey. and ecstasy purchases’, Drug and Alcohol 614.
Dependence, vol. 94, pp. 183-190. Goudie, AJ, Sumnall, HR, Field, M,
Australian Institute of Health and Welfare
2008, 2007 National drug strategy Commonwealth Department of Health and Clayton, H & Cole, JC 2007, ‘The effects
household survey: detailed findings, Drug Ageing 2002, Return on investment in of price and perceived quality on the
Statistics Series no. 22, cat. no. PHE 107, needle and syringe programs in Australia: behavioural economics of alcohol,
AIHW, Canberra. report, Commonwealth Department of methamphetamine, cannabis, cocaine
Health and Ageing, Canberra. and ecstasy purchases’, Drug and Alcohol
Baum, CF 2006, An introduction to
Dependence, vol. 89, pp. 107-115.
modern econometrics using Stata, Stata Darke, S & Hall, W 1995, ‘Levels and
Press, College Station, Texas. correlates of polydrug use among heroin Greene, W. 2008, Econometric analysis,
users and regular methamphetamine 6th edition, Prentice Hall, Upper Saddle
Becker, GS, Grossman, M & Murphy, KM
users’, Drug and Alcohol Dependence, River, NJ.
1991, ‘Rational addiction and the effect
of price on consumption’, Economics of vol. 39, pp. 231-235. Grossman, M & Chaloupka, FJ 1998, ‘The
Drugs, vol. 81, no. 2, pp. 237-241. Darke, S, Kaye, S, McKetin, R & Duflou, demand for cocaine by young adults: a
J 2008, ‘Major physical and psychological rational addiction approach’, Journal of
Bretteville-Jensen, AL 2006, ‘Drug
harms of methamphetamine use’, Drug Health Economics, vol. 17, pp. 427-474.
demand - initiation, continuation and
quitting’, De Economist, vol. 154, pp. and Alcohol Review, vol. 27, pp. 253-262. Grossman, M, Chaloupka, F & Shim, K
491-516. Dave, D 2004, ‘Illicit drug use among 2002, ‘Illegal drug use and public policy’,
Health Affairs, vol. 21, pp. 134-145.
Cameron, A, Trivedi, P, Hammond, P & arrestees and drug prices’, NBER
Holly, A 1998, Regression analysis of Working Paper 10648, National Bureau Hurley, SF, Jolley, DJ & Kaldor, JM
count data, Cambridge University Press, of Economic Research, Cambridge, 1997, ‘Effectiveness of needle-exchange
Cambridge. Massachusetts. programmes for prevention of HIV
infection’, The Lancet, vol. 349, pp. 1797-
Caulkins, JP 1995, ‘Estimating elasticities Dave, D 2006, ‘The effects of cocaine and
1800.
of demand for cocaine and heroin with heroin price on drug-related emergency
data from the Drug Use Forecasting department visits’, Journal of Health Jofre-Bonet, M & Petry, NM 2008,
System’, Final Report, United States Economics, vol. 25, pp. 311-333. ‘Trading apples for oranges? Results of
15
B U R E A U O F C R I M E S T A T I S T I C S A N D R E S E A R C H
an experiment on the effects of heroin Moffatt, S, Weatherburn, D & Donnelly, Smithson, M, McFadden, M, Mwesigye,
and cocaine price changes on addicts’ N 2005, What caused the recent drop S & Casey, T 2004, ‘The impact of illicit
polydrug use’, Journal of Economic in property crime?, Crime and Justice drug supply reduction on health and
Behavior & Organization, vol. 66, Bulletin no. 85, NSW Bureau of Crime social outcomes: the heroin shortage in
pp. 281-311. Statistics and Research, Sydney. the Australian Capital Territory’, Addiction,
vol. 98, pp. 340-348.
Kennedy, P 2008, A guide to Pacula, LR 1998, ‘Does increasing the
econometrics, 6th edition, MIT Press, beer tax reduce marijuana consumption?’ Snowball, L, Moffatt, S, Weatherburn, D &
Cambridge, Massachusetts. Journal of Health Economics, vol. 17, Burgess, M 2008, Did the heroin shortage
no. 5, pp. 557-585. increase methamphetamine use?,
McKetin, R, McLaren, J & Kelly, E 2005b,
Crime and Justice Bulletin no. 114, NSW
The Sydney methamphetamine market: Petry, NM 2000, ‘Effects of increasing Bureau of Crime Statistics and Research,
patterns of supply, use, personal harms income on polydrug use: a comparison Sydney.
and social consequences, National of heroin, cocaine and alcohol abusers’,
Drug Law Enforcement Research Fund Addiction, vol. 95, pp. 705-717. Sumnall, HR, Tyler, E, Wagstaff, GF &
Monograph no. 13, National Drug Law Cole, JC 2004, ‘A behavioural economic
Enforcement Research Fund, Canberra. Petry, NM 2001, ‘A behavioural economic analysis of alcohol, amphetamine,
analysis of polydrug abuse in alcoholics: cocaine and ecstasy purchases by
McKetin, R, McLaren, J, Kelly, E, Hall,
asymmetrical substitution of alcohol and polysubstance misusers’, Drug and
W & Hickman, M 2005a, Estimating
cocaine’, Drug and Alcohol Dependence, Alcohol Dependence, vol. 76, pp. 93-99.
the number of regular and dependent
vol. 62, pp. 31-39.
methamphetamine users in Australia, Topp, L & Mattick, RP 1997, ‘Choosing
NDARC Technical Report no. 230, Petry, NM & Bickel, WK 1998, ‘Polydrug a cut-off on the Severity of Dependence
National Drug and Alcohol Research abuse in heroin addicts: a behavioural Scale (SDS) for methamphetamine
Centre, Sydney. economic analysis’, Addiction, vol. 93, users’, Addiction, vol. 92, no. 7, pp. 839-
no. 3, pp. 321-335. 845.
McKetin, R, McLaren, J, Lubman,
DI & Hides, L 2006, ‘The prevalence Phillips, B & Burns, L 2009, New South Vincent, N, Shoobridge, J, Ask, A, Allsop,
of psychotic symptoms among Wales drug trends 2008: findings from S & Ali, R 1998, ‘Physical and mental
methamphetamine users’, Addiction, vol. the Illicit Drug Reporting System (IDRS), health problems in methamphetamine
101, pp. 1473-1478. Australian Drug Trends Series, no. 20, users from metropolitan Adelaide,
National Drug and Alcohol Research Australia’, Drug and Alcohol Review, vol.
McKetin, R, Ross, J, Kelly, E, Baker, E,
Centre, Sydney. 17, pp. 187-195.
Lee, N, Lubman, D & Mattick, R 2008,
‘A comparison of harms associated with Saffer, H & Chaloupka, FJ 1999, ‘The Weatherburn, D, Jones, C, Freeman, K &
injecting vs smoking methamphetamine demand for illicit drugs’, Economic Makkai, T 2003, ‘Supply control and harm
among methamphetamine treatment Enquiry, vol. 37, no. 3, pp. 401-411. reduction: lessons from the Australian
entrants in Australia’, Drug and Alcohol heroin ‘drought’’, Addiction, vol. 98, pp.
Review, vol. 27, pp. 277-285. Shearer, J, Darke, S, Rodgers, C, 83-91.
Slade, T, van Beek, I, Lewis, J, Brady,
Maher, L, Li, J, Jalaludin, B, Wand, H, White, B, Breen, C & Degenhardt, L
D, McKetin, R, Mattick, R & Wodak,
Jayasuriya, R, Dixon, D & Laldor, JM 2003, NSW party drug trends 2003:
A 2009, ‘A double-blind, placebo-
2007, ‘Impact of a reduction in heroin findings from the Party Drugs Initiative
controlled trial of modafinil (200 ms/day)
availability on patterns of drug use, risk (PDI), NDARC Technical Report no. 182,
for methamphetamine dependence’,
behaviour and incidence of hepatitis C National Drug and Alcohol Research
Addiction, vol. 104, no. 2, pp. 224-233.
virus infection in injecting drug users in Centre, Sydney.
New South Wales, Australia’, Drug and Siegel, S and Castellan, N (1988)
Alcohol Dependence, vol. 89, pp. 244-250. Nonparametric statistics for the
behavioural sciences, McGraw-Hill., New
Manski, CF, Pepper, JV & Petrie, CV
York.
2001, Informing America’s policy on
illegal drugs: what we don’t know keeps Silverman, LP & Spruill, NL 1977, ‘Urban
hurting us, National Academy Press, crime and the price of heroin’, Journal of
Washington DC. Urban Economics, vol. 4, pp. 80-103.
NSW Bureau of Crime Statistics and Research - Level 8, St James Centre, 111 Elizabeth Street, Sydney 2000
bcsr@agd.nsw.gov.au • www.bocsar.nsw.gov.au • Ph: (02) 9231 9190 • Fax: (02) 9231 9187
ISSN 1030 - 1046 • ISBN 978-1-921626-24-1
© State of New South Wales through the Department of Justice and Attorney General NSW 2009. You may copy, distribute, display, download and otherwise freely
deal with this work for any purpose, provided that you attribute the Attorney General’s Department of NSW as the owner. However, you must obtain permission if you
wish to (a) charge others for access to the work (other than at cost), (b) include the work in advertising or a product for sale, or (c) modify the work.