BA-Thesis Jordy Lindeloo

Document Sample
BA-Thesis Jordy Lindeloo Powered By Docstoc

Addictive Substances
  Substitutes or Complements?
             Jordy A. Lindeloo

As a result of mankind’s relentless desire to establish order in a universe of chaos, economists have -

since the origin of economics as a science - classified the myriad of existing economic goods or

products into different subgroups. In an average economics textbook one is likely to encounter terms

such as normal goods, inferior goods, luxury goods, Giffen goods, Veblen goods et cetera. However,

the term ‘addictive good’ has emerged comparatively recently in the context of economics. This

obviously does not imply that such goods and their consumers (possibly addicts) did not exist before.

As a matter of fact, economists basically ignored the issue of addiction, regarding it as a psychological

rather than economic topic. The most important pioneers of incorporating addiction in the science of

economics were economists Kevin Murphy and Gary Becker, who formulated the highly controversial

theory of Rational Addiction. According to this theory, the consumption of addictive goods is

perfectly rational behaviour for the individual who engages in such activity. This statement is

defended by the claim that although a certain addict’s life might seem miserable, he or she would

have been even more unhappy without the consumption of the addictive substance. However, the

rational addiction model mainly emphasises the fact that addicts respond rationally to present and

future price changes, thereby largely ignoring the non-pecuniary costs that addictive substances may


          Current consumption of addictive goods is influenced by future prices as a result of two

characteristics that differentiate such goods from normal goods. First of all, the consumption of an

addictive product causes tolerance, which implies that the higher the amount of total past

consumption, or so-called “addictive capital”, the lower the utility derived by an individual at a

specific point in time. This indicates that consumption of an addictive good results in harmful

consequences for health and therefore this only applies to unhealthy addictive goods. In case of

healthy addictions such as exercising, a larger quantity of addictive stock leads to higher level of

utility. In addition, reinforcement results from using an addictive good.

This means that greater past consumption causes a greater desire for consuming a certain addictive

good (Becker, Grossman, Murphy, 1991). In other words, the marginal utility derived from consuming

an additional unit of addictive substance increases when the stock of past consumption of the

substance increases. This entails the existence of so-called “adjacent complementarity”: past

consumption positively affects present consumption; hence present consumption positively affects

future consumption. Graph 1 illustrates the combined effects of tolerance and reinforcement: both

utility functions (with and without current consumption) are downward-sloping with respect to the

stock of past consumption, an indication of tolerance; in addition, the vertical gap between the two

utility functions widens as the stock of past consumption rises, implying reinforcement.

Graph 1: The lower line depicts individual utility without current consumption of the addictive good;

the upper line demonstrates utility with current consumption

Utility                                                        Marginal Utility of Consumption

          Low                                  High                    Stock of Past Consumption

Taking the reinforcement effect (or adjacent complementarity) into account, rational addicts react to

announced future price increases by curtailing their present consumption of an addictive substance,

realising that this will reduce their future desire of consumption and hence their future consumption

itself. This implies that addicts also have a downward sloping demand curve and can thus be

considered ‘rational’. This prediction has been supported by empirical evidence. However, being

responsive to price changes does not necessarily make an individual rational. This is claimed by

supporters of the counterpart of the rational addiction model, the so-called myopic addiction model.

They believe that individuals do not always behave in their own best interest. One prominent pitfall

of the rational addiction model is the assumption that agents are in the possession of perfect

information. Needless to say, this assumption does not always hold in the real world. For example, in

the case of drug use, consumers are occasionally unaware of the harmful consequences that drug

usage imposes on them. This possibly explains the empirical findings showing that drug use is

significantly more prevalent amongst individuals having a relatively low socioeconomic status

compared to those having a high socioeconomic status. Those having a lower educational

background and thus having a lower socioeconomic position may lack the essential knowledge

regarding the negative consequences of drug utilisation, thereby being more likely to consume

unhealthy (and possibly addictive) products.

        Furthermore, there are several other explanations of why an individual might not behave in

his or her own best interest. Even if a person is provided with complete information concerning the

consequences of utilising a specific drug or another addictive good, he or she might still irrationally

decide to consume the good. One explanation is the concept called ‘time inconsistency’, which

implies that an individual initially intends to behave in a certain way, after which he or she decides to

behave differently. In Behavioural Economics, time inconsistency is modelled by the use of

hyperbolic discounting, in which case future costs and benefits are severely underestimated.

Therefore, individuals behaving in a time inconsistent manner are called ‘present-biased’.

When this reasoning is applied to the consumption of addictive goods, a probable possibility exists

that a student decides to consume just two beers at a housewarming taking place two weeks from

now, taking all consequences of drinking too much beer into consideration, while ending up totally

wasted at the housewarming, since he was in a present-biased state of mind at the moment the first

sip of beer was taken. In this case, the rational decision made by his ‘farsighted planner’ was

replaced by the irrational decision made by his ‘myopic doer’, implying that the same person can

have different ‘selves’ behaving differently (Thaler and Shefrin, 1981). Therefore, individuals should

be aware of the fact that they can only make truly rational decisions when they are in the so-called

‘cold’ state and that they have to stick to these decisions when in the ‘hot’ state.

        Nevertheless, awareness alone is not sufficient, since weakness of will can still induce an

individual to behave irrationally, despite the fact that he or she is aware of it. This human trait is

clearly demonstrated by the fact that there are countless individuals who desperately want to quit

smoking, but are somehow incapable of doing so. Another example is the too regular occurrence of

obesity amongst inhabitants of developed countries, even though many obese individuals realise

they would be better off in a more healthy condition.

        Yet another reason of behaving sub-optimally regarding the consumption of addictive

substances is the ‘status quo bias’. Behavioural economists have discovered the phenomenon that

people in general excessively retain the current situation, in spite of other existing options that

would have made someone better off. With respect to the consumption of addictive goods, one

could argue that once a certain addiction has become a habit, addicts are reluctant to give up this

habit because they do not want to alter the status quo.

        Finally, empirical evidence has indicated that many individuals are too confident about their

own abilities, which is called overconfidence or over-optimism.

As a result, an individual might have perfect information regarding the addictiveness of a certain

good, but might simultaneously believe that this information does not apply to him or her, since this

individual is overconfident about his or her ability to avoid becoming addicted. Considering the

multitude of explanations of irrational behaviour, claiming that addiction is perfectly rational would

certainly be a bold statement.

         Besides, the nature of addictive products is tediously complex. Since a wide array of addictive

substances exists, the effects caused by these substances differ significantly. This implies that some

of these goods have more detrimental effects on health than other goods, whereas yet other goods

impose no harm on health at all. It is no wonder that cocaine usage has more severe consequences

on health than drinking too much coffee. And an individual who is addicted to exercising in the gym

might actually improve rather than deteriorate his or her health. Secondly, addiction is a rather

subjective concept, meaning that a specific substance might be addictive to one person, whilst not to

another. For example, many people in non-Islamic countries consume alcoholic beverages, but only a

fraction of these people is truly addicted. Moreover, individuals can be addicted to products that are

not officially classified as addictive substances, such as an addiction to working, which is called ‘being

a workaholic’. However, these kinds of highly ‘customised’ addictions will not be discussed in this


         Another feature of addictive substances is that some are always maleficent to one’s health,

whereas other substances can be either beneficial, neutral or harmful with respect to health,

depending on the quantity consumed. Alcohol use for instance is not necessarily harmful when

consumed in relatively small quantities, just like food, which is even a necessity, although both are

detrimental to health when consumed in excessive quantities. Other addictive goods on the other

hand, such as cigarettes and cocaine, are always harmful to one’s health, regardless of the amount

consumed. In addition, the consumption of addictive substances can give rise to non-pecuniary costs

that are not directly related to health. Exorbitant alcohol consumption for instance causes damage to

brain cells, thereby diminishing the user’s intelligence.

The same holds for a selection of other drugs, such as, among others, cocaine, amphetamines and

heroin. Another example is that cigarette smoking, besides causing harmful effects on health, also

deteriorates the quality of the skin, which can also be considered as a cost to the user. Finally, many

addictive goods not only cause damage to their users, but also affect society at large negatively.

Examples of these so-called negative external effects or externalities are plentiful. For example, being

a passive smoker by inhaling another individual’s cigarette smoke is also detrimental to one’s health.

Driving under influence also imposes a great threat to other persons’ lives. Alcohol abuse is one of

the main contributors to the occurrence of violence in night life areas. These are just three examples

from a virtually endless list of negative externalities that accompany the consumption of addictive


        Taking all these characteristics of addictive goods into account, this paper will focus on the

relationships amongst the variety of addictive substances. So far several researchers have tried to

unravel these relationships, analysing empirical data or conducting experimental research in order to

discover whether addictive goods are substitutes or complements of one another. According to the

rational addiction model, such goods should be complements of each other, since an addict who is

behaving in a rational manner when consuming an addictive substance apparently does not care

much about his or her health and should therefore also consume other unhealthy addictive goods in

order to behave rationally, assuming that all addictive goods provide (at least initially) some sort of

pleasure for the user. However, not caring about one’s health might potentially induce irrational

behaviour as empirical research (for example in the field of Happiness Economics) has proved that

health is a vital determinant of someone ’s happiness in which case a positive relationship between

health and happiness exists.

        Despite the attempt of multiple researchers to uncover whether addictive goods are

substitutes or complements, the results are heavily mixed. In order to clarify the findings, potential

explanations for both substitutability and complementarity have to be analysed thoroughly.

If this can be accomplished successfully, the empirical evidence will be supported by underlying

theory, which will greatly increase the validity and reliability of the empirical findings. This will be the

extension of this paper regarding the economics of addictive products. There are mainly two reasons

why a study such as this one is of great importance. First of all, from a positive perspective, it will

provide us with a better comprehension of the world around us, which is the goal of any scientific

study. Secondly, from a normative point of view, the results of this study have the potential to be

useful policy implications. For example, if it is proved that alcohol consumption has a positive causal

effect on cigarette consumption, then increasing the governmental taxes on alcoholic beverages

might simultaneously reduce the consumption of alcohol and cigarettes, thereby killing two birds

with one stone. If evidence is provided that alcohol and cannabis consumption are economic

substitutes, then legalisation of cannabis use could lead to an increase of marijuana consumption

and hence to a decrease in the consumption of alcohol. As a result, additional harm caused by

increased utilisation of cannabis could be nullified by decreased alcohol consumption, hence

producing a net effect of zero on population health after legalisation of the cannabis market.

        The remainder of this paper will be structured in the following manner: section II will discuss

evidence for complementarity amongst addictive substances; section III will focus on proof of

substitutability amongst addictive goods; in section IV evidence from behavioural economic studies

for both the existence of economic substitutes and complements will be examined; in section V

potential explanations for a complementary relationship will be analysed; in section VI reasons for a

substitute relationship will be examined; section VII will consist of concluding remarks and the final

section will consist of a discussion concerning the political and legislative relevance of this study.

Statistical evidence for complementarity

Over the years numerous economists have investigated whether addictive substances are economic

complements or substitutes of each other. The two most commonly utilised substances in their

research have been alcohol and cigarettes. The primary reason for this is the fact that these two

drugs are sold legally in most countries (strict Islamic countries may be an exception) and thus there

is transparency regarding their prices. This information is vital in suchlike research since a price

increase of one good results in a decreased demand for another good in case these two are

complements, while it results in an increased demand for the other good if the two goods are

economic substitutes. These tendencies are called cross-price elasticities. However, an overall

consensus regarding the economic relationship between alcohol and cigarette consumption has so

far not been achieved. Even single papers produce opposing results. In this section multiple papers

will be discussed which conclude that addictive substances are complements of one another.

        It is important to note that complementarity between two addictive goods is not necessarily

symmetric (Pacula, 1998a). For example, it could be perfectly possible that smoking cigarettes does

not coincide with drinking alcoholic beverages, but that drinking on the other hand is accompanied

by cigarette smoking. The former relationship either implies that the two goods are economic

substitutes or that they are unrelated, while the latter one means that they are economic

complements. Even if we would assume that those who smoke cigarettes never combine this with

drinking alcohol, we cannot say that the two are therefore economic substitutes. Two goods are

substitutes of each other when they more or less satisfy the same needs. Thus, never combining

salmon with red wine in a restaurant for instance does not entail that the two goods are economic


        Using micro-level data from the National Longitudinal Survey of Youth (USA) to estimate

individual demand equations for alcohol and cannabis, Pacula (1998a) concludes that increasing

taxes on beer not only results in a lower demand for beer itself, but also in a reduced demand for

marijuana in an even more sensitive manner.

Furthermore, Pacula found that in decriminalised states – states in which consumption of cannabis is

legalised to a certain degree – the prevalence of alcohol consumption is statistically significantly

higher compared to other states. This implies that the greater demand for cannabis in the

decriminalised states has resulted in a greater demand for alcoholic beverages as well and hence

complementarity between the two goods. These findings are reinforced by results produced by Thies

and Register (1993), who used combined data on youth drug use, drinking age laws and

decriminalisation indicators in order to capture the full price of both alcohol and cannabis and

concluded that individuals inhabiting states where cannabis consumption has been decriminalised

were more prone to consuming alcohol as well. Combining the conclusions of these two papers

would suggest that the complementary relationship between alcohol and marijuana consumption

has a symmetric nature. In addition, Saffer and Chaloupka (1998), using nationally representative

household surveys (USA) on drug use, found a negative relationship between the price of alcohol and

cannabis use and consequently concluded that the two are complements.

       Evidence for complementarity also exists with respect to the consumption of cannabis and

cigarettes. According to a study conducted by Cameron and Williams (2001), using individual level

data obtained from the National Drug Strategy Household Surveys (Australia), the price of cannabis

has a negative and significant effect on the probability of being a cigarette smoker. A 10 per cent

increase in the price of marijuana decreases the probability of being a smoker by 1.32 percentage

points (Cameron and Williams, 2001). This propensity could result from the fact that cannabis and

tobacco are often consumed simultaneously, known as a joint. When the price of cannabis increases,

the demand for cannabis declines, which is in accordance to the law of demand. Consequently, the

reduced demand for marijuana requires a smaller amount of tobacco, correspondingly resulting in a

lower demand for cigarettes as well. Moreover, they also found a complementary relationship

between the consumption of alcohol and cigarettes.

Although the price of alcohol was found to have an insignificant effect on cigarette smoking

participation, the probability of participation in alcohol consumption is negatively and significantly

affected by the price of cigarettes. Therefore, the authors concluded that, on average, alcoholic

drinks and cigarettes are economic complements.

        Since most addictive substances are illicit drugs, the availability of price data regarding these

substances is extremely limited. This is problematic for estimating the cross-price elasticities

between two addictive goods, which requires the use of price data. In order to remedy this

impediment, the researchers use proxies as a substitute for the market price. Examples of such

proxies regarding marijuana use include the probability of being arrested for cannabis possession and

the severity of fines and jail sentences after being caught. A higher probability of being arrested can

be translated to a higher indirect price of marijuana in case the consumer is arrested. If, on the other

hand, the probability of the producer or seller of cannabis being arrested is increased, the costs of

producing or selling marijuana will increase. As a result, the producer will charge a higher market

price (direct price) for cannabis. The same reasoning can be applied to the strictness of fines and jail

sentences regarding the production and consumption of marijuana. Hence, decriminalisation status

in a particular state or country is a frequently used proxy for the price of cannabis as legalisation

reduces the indirect costs of producing and/or consuming marijuana and thus its full price.

        Compared to studies focusing on the relationship between alcohol and cannabis

consumption, those examining the interdependencies between marijuana and tobacco consumption

are relatively rare. Pecula (1998b), utilising the National Longitudinal Study of Youth, discovered that

both current and past prices of cigarettes have a negative impact on the probability of current

marijuana use, indicating a complementary relationship between the two products. This conclusion is

supported by the results acquired by Chaloupka et al. (1999), who examined cross-sectional data

regarding 8th , 10th and 12th graders from the 1992 – 1994 Monitoring the Future Surveys.

They found that an increase in cigarette prices results in a decrease in both the probability – albeit

not statistically significantly so – and frequency of cannabis consumption, whilst the own price

elasticity of cigarettes is less strong than the cross-price effect of cigarettes with respect to marijuana

use. A similar tendency has been found by Pacula (1998a) concerning the cross-price elasticity of

beer with respect to cannabis use, which has been discussed earlier. However, the reliability of the

results indicating that tobacco and marijuana are complements of each other can be disputed. Not

only do the studies rely on a rather limited number of years of time series cross-sectional data,

thereby curbing the variation in cigarette prices, also the number of cross-sections themselves are

too limited to control for unobserved, state-specific characteristics.

        Using data on seven years of pooled cross-sections and a more extensive database regarding

marijuana policies, Farrelly et al. (2000) claim that they have controlled for state-specific

characteristics by incorporating state-fixed effects. An example of state-specific characteristics that

potentially bias the results is the fact that state policies could be endogenous. This implies for

instance that states in which tobacco consumption is more prevalent are more likely to impose

higher cigarette taxes. Nevertheless, these authors also conclude that cigarettes and cannabis are

economic complements.

        It is important to note that a certain relationship between two addictive products might

differ among different age groups and between gender. This further impedes the implementation of

government policies aimed at hampering addictive substances. For example, alcohol and tobacco

might be complements for teenagers, but substitutes for adults. This means that imposing higher

taxes on alcoholic beverages leads to a diminished amount of cigarettes smoked by teenagers, but a

larger amount smoked by adults. In order to distinguish these different effects, some papers have

either focused on one age group only or examined different age groups and genders independently.

In an elaborate paper produced by Williams, Pecula, Chaloupka and Wechsler (2001), the relationship

between alcohol and marijuana use among US college students has been analysed.

The fact that the implementation of stricter alcohol policies aimed at discouraging binge drinking on

college campuses has coincided with an increase of marijuana use amongst college students by 22

per cent between 1993 and 1999 might indicate that the two substances are economic substitutes.

However, the paper provides “overwhelming evidence” that alcohol and cannabis are complements.

In addition to the fact that a higher price of marijuana results in a lower probability of both marijuana

and alcohol consumption, hindering access to alcohol resulted in a reduced amount of marijuana

consumed. Instead, the rise in cannabis consumption amongst college students was caused by a

substantial decrease in the price of marijuana during the period under investigation. Considering that

youths are particularly price responsive due to their relatively low incomes compared to adults, and

thus face a relatively high price elasticity of demand, means that it is not surprising that their

demand for cannabis has increased. The subsequent section will be devoted to existing evidence

regarding substitution amongst addictive products.

Statistical evidence for substitution

It is not surprising that there are also economists who believe that a substitution effect exists

amongst the various kinds of addictive substances. Several of such substances, as a matter of fact,

satisfy the same desire: being in an altered state of consciousness. As a result, literature regarding

the substitutability amongst a number of addictive products has accumulated over the last decades,

in which researchers have reinforced their belief in a substitution effect by empirical evidence.

Besides, several economists analysing this subject have found inconclusive or mixed results, implying

that either no clear relationship between two addictive goods has been found or that the empirical

data support both the hypothesis of complementarity and substitutability, depending on the

methods applied and the substances taken into consideration.

        According to research conducted by DiNardo and Limieux in 1992, based on annual state

level data measures of cannabis and alcohol use from the 1980 through 1989 Monitoring the Future

(MTF) surveys to examine the effects of higher legal drinking ages, marijuana and alcohol are

economic substitutes. As expected, the higher legal drinking ages resulted in a reduction of alcohol

consumption by youths; in addition, this tendency was accompanied by an almost proportional

increase in the consumption of cannabis. This relationship between the two substances was regarded

as persuasive evidence of an economic substitution effect. The same conclusion has been obtained

by Model, who studied the effects of marijuana decriminalisation on drug related emergency room

hospitalisations and violent crime. She uncovered that states in which cannabis had been

decriminalised faced higher numbers of emergency room hospitalisations resulting from the

consumption of marijuana, whilst those related to alcohol and other illicit drugs were in fact lower.

Furthermore, she found that lower violent crime rates coincide with marijuana decriminalisation. As

many of such crimes result from alcohol abuse, she concluded that the lower crime rates were the

consequence of substation away from alcohol to cannabis in the decriminalised states.

Supplementing these findings is research conducted by Chaloupka and Laixuthai (1997).

Utilising data obtained from the 1982 and 1989 Monitoring the Future surveys and the Fatal Accident

Reporting System, they found a positive relationship between the price of cannabis and the number

of fatal and non-fatal driving accidents. They applied the ensuing reasoning in order to explain this

relationship: although a decrease in the full price of marijuana, for example resulting from reduced

penalties for possession and/or use after the implementation of decriminalisation policies,

unsurprisingly leads to a higher demand for marijuana, it simultaneously leads to a lesser demand for

alcoholic drinks and other illegal substances (besides cannabis), thus exhibiting a substitution effect

between marijuana on one hand and alcohol and other illicit drugs on the other. Apparently, this

substitution from one type (alcohol) of drugs to the other (cannabis) causes a decline in the number

of driving accidents. There are several explanations that address this issue. First of all, individuals

under the influence of marijuana, thereby performing so-called “stoned driving”, are potentially

more careful and are driving more slowly than their drunk counterparts (Stein 1983; Robbe and

O’Hanlon, 1993). Secondly, youths tend to be less likely to drive after consuming cannabis, resulting

from the fact that alcohol, unlike marijuana, stimulates risky or reckless behaviours. Thirdly, it is very

probable that the consumption of cannabis does not impair someone’s driving skills as heavily as

alcohol does (Hansteen et al. 1976; Moskowitz et al. 1976). In short, substitution from alcohol to

marijuana leads to a net decrease in driving accidents, insinuating that the higher number of

cannabis induced car accidents is more than offset by the lower amount of driving accidents resulting

from alcohol consumption.

        Mixed results are obtained from a study conducted by Cameron and Williams (2001) focusing

on the interrelatedness of cannabis, alcohol and cigarette consumption, which has been discussed in

the previous section. Utilising individual level data from the National Drug Strategy Household

Surveys (Australia), one of their findings is that a 10 percent increase in the full price of alcoholic

drinks increases the probability of marijuana use (not the frequency) by 4.17 percentage points,

implying that the two goods are economic substitutes.

However, this is contradicted by the fact that an individual residing in the state of South Australia,

which was the only decriminalised state concerning cannabis use in Australia, is 2.8 percentage

points more likely to consume alcohol. This provides evidence for complementarity instead of

substitutability between the two substances. Nevertheless, the latter finding might be biased as a

result of a so-called ‘South Australia’ effect, which means that inhabitants of this particular state are

more liberal towards alcohol and cannabis use, thereby not necessarily proving evidence of a

complementary relationship between the two types of drugs.

        An example of a study that supports substitutability between alcohol and cigarette

consumption is the one conducted by Goel and Morey (1995). Using panel data on the US state-level

for the period 1959 – 1982 the authors estimated both own and cross price effects regarding the

demand for cigarettes and alcoholic beverages. They found cross price elasticities of 0.1 and 0.332

with respect to the demand for cigarettes and liquor respectively, of which the positive signs indicate

economic substitution. However, according to the researchers Decker and Schwartz, their study has

several drawbacks. Firstly, the own price elasticity concerning the demand for alcoholic drinks

reported in their study deviates substantially from those reported in other, closely related studies.

Their figure of -0.13 is much smaller compared to figures obtained in alternative studies, implying a

smaller degree of price sensitivity with respect to alcohol consumption. Furthermore, their estimated

income elasticity of the demand for cigarettes is 0.312, of which its positive sign indicates that

cigarettes are a normal good (an income elasticity larger than + 1 is a luxury or superior good),

thereby opposing the results found elsewhere in the literature which indicate that cigarettes are an

inferior good. Cigarettes being an inferior good means that cigarette consumption decreases as one’s

income rises, thus implying a negative relationship between income and tobacco consumption.

Decker and Schwartz on the other hand, conducting a similar research of their own, used individual

level data rather than state level data concerning approximately half a million individuals from 45

states surveyed between 1985 and 1993 from the Behavioral Risk Factor Surveillance System.

The advantage of using individual level data over state-level data is the avoidance of biased results

arising from the discrepancy between state sales levels and actual state consumption, since

smuggling and bootlegging of cigarettes among different states account for higher sales levels than

consumption levels in states with relatively low excise taxes on cigarettes and vice versa. In addition,

the use of individual level data capacitates one to estimate price effects on participation and quantity

demanded separately and to execute separate subgroup analyses for men and women (Decker and

Schwartz, 2000). The overall cross price elasticity of alcohol consumption with respect to the price of

cigarettes found by the authors is 0.5, implying that alcohol and cigarettes are substitutes, since 0.5

constitutes a positive number. The main contributor to this positive cross price elasticity is the

change in participation in drinking rather than the change in consumption level amongst current

drinkers, since the separate cross price elasticities are 0.39 and 0.12 for the effect of cigarette price

changes on drinking participation and consumption of alcohol respectively. Moreover, it seems that

women’s drinking participation is significantly more price responsive than men’s, as the cross price

elasticity of participation is 0.52 for women, while it is 0.27 in case of men (Decker and Schwartz,

2000). Surprisingly, the cross price elasticity in the opposite direction, the effect of the price of beer

on cigarette consumption, is the figure -0.14, which is negative and therefore indicates a

complementary instead of a substitution relationship. Although the cross price elasticity of cigarette

consumption amongst current smokers has been estimated at 0.04, implying that an increase in the

price of beer results in an increase in the amount of cigarettes consumed, this positive relationship is

nonetheless reversed by the significantly more powerful and negative relationship between smoking

participation and beer prices. Besides, cigarette consumption amongst women has a greater price

responsiveness than consumption amongst men, which contradicts the results obtained by Decker

and Schwartz on own price elasticity of cigarette consumption.

        Summarising, Decker and Schwartz have found asymmetric cross price elasticities of alcohol

and cigarette consumption: the relationship between the price of cigarettes and the demand for

alcoholic beverages is positive (substitutability), whereas the demand for cigarettes has a negative

cross price elasticity with respect to the price of beer (complementarity).

Behavioural economic evidence regarding complementarity and substitutability

Thus far evidence for the existence of either substitutability or complementarity between two

addictive commodities has been attained by examining actual, real life data. However, this procedure

potentially lacks reliability when the interrelatedness of several addictive substances is being

analysed. The main reason is the limited transparency and availability regarding the actual prices of

the various kinds of illicit drugs. It goes without saying that this is not a surprising fact, considering

that the sales of such drugs exclusively occur in the black market or underground economy. Another

drawback of statistical research in this area is the probable discrepancy between the total amount of

sales and actual consumption in a particular geographical location, which has been described

previously. In order to avoid these issues, some researchers have applied tools from behavioural

economics to the topic of addictive goods.

        An example of such research is a study conducted by Sumnall et al. (2004), in which cross

price elasticities amongst alcohol, cocaine, ecstasy and amphetamine (more familiar by its informal

name ‘speed’) have been estimated. In the conducted experiment forty-three participants, all of

which were polysubstance users (using at least three different kinds of drugs) were allotted £40 and

a price list indicating the respective prices of the four types of drugs, which reflected the street prices

when the study was designed. In order to uncover the cross price elasticities, the price of one type of

drug was altered whereas the other kinds remained constant. By applying the same strategy to each

type of drug, the following results were obtained: alcohol was substituted for amphetamine as the

price of alcohol rose, cocaine was a complement drug with respect to alcohol while ecstasy use was

independent; when altering the price of amphetamine, alcohol was found to be a substitute while

the other forms of drugs were independent of the price of amphetamine; in case of a change in the

price of cocaine, the results showed that alcohol and ecstasy can be regarded as substitute drugs,

whilst the purchase of amphetamine was not affected by the price of cocaine; finally, only the

purchase of cocaine was influenced by a modification of the price of ecstasy, in such a manner that

the two drugs can be thought of as substitutes of one another.

Taking this study into account, one could argue that different relationships exist between alternative

kinds of drugs. However, due to the fact that this study is a laboratory simulation, the authors

recommend to interpret their findings with caution.

        A similar kind of study is a behavioural economic analysis performed by Petry (2000), in

which fifty-three subjects participated. Only those having a history of alcohol abuse or dependence

and having used at least three other types of drugs were considered eligible for participation. As a

form of compensation for their efforts, all participants received $50 for participation. The types of

drugs used in this experiment were alcohol, cocaine, heroin, marijuana, nicotine and Valium (a form

of tranquilliser which is a licit prescription drug) and their respective prices were the actual street

prices for the amounts generally required for a successful ‘hit’. Three different experiments were

executed: one in which the price of alcohol varied whereas the other drug prices remained constant;

in the second experiment the price of cocaine fluctuated; in the final experiment Valium was

assigned several different prices while the other prices remained unchanged. In each experiment, the

subject was able to spend $35 of imitation money during each of the four different price scenarios

with respect to the price of alcohol, cocaine and Valium respectively. When the price of alcohol

increased, only the number of cocaine purchases plummeted significantly, whilst the purchased

quantities of the other drugs remained unaffected as the price of alcohol increased, indicating that

cocaine can be regarded as a complement to alcohol and that the demand for the other types of

drugs is independent of alcohol prices. Moreover, the price of cocaine was found to significantly

influence the demand for alcohol, since more alcohol was purchased when the price of cocaine was

relatively high, implying that alcohol was a substitute for cocaine amongst the participants of this

experimental study. Apart from alcohol, no other drug purchases were modified as a result of

changes in the price of cocaine, and thus these drug purchases are independent of cocaine price.

Finally, shifts in the price of Valium did not affect the quantities of drugs purchased besides the

quantity of Valium itself. Therefore, neither a complementary nor a substitute relationship seems to

exist between Valium and alternative kinds of drugs involved in the experiments.

The most interesting finding that this study offers is the fact that there is asymmetrical substitution

between alcohol and cocaine, since cocaine seems to be a complement to alcohol while on the other

hand alcohol seems to be a substitute for cocaine. In order to test the validity of the results obtained

from the simulation, the hypothetical drug purchases of individuals submitting urine samples which

were positive for the respective drugs were compared to those of subjects submitting urine samples

which were negative for the respective drugs. One may expect that participants who have recently

consumed a specific type of drug are quite likely to purchase an ostensibly greater amount of this

drug during the experiment. This difference in quantities purchased between subjects tested either

positive or negative was statistically significant regarding cocaine and heroin purchases, but

insignificant with respect to cannabis purchases. In addition, the hypothetical drug purchases of all

types of drugs were positively correlated with self-reported years of lifetime use. However, the

author stresses that the study must be interpreted cautiously. The foremost reason for this

recommendation of hers is the fact that the results of this research cannot be generalised to

everyone, considering that exclusively primary alcoholics with rather prolonged histories of polydrug

use were eligible for participation in the experiments.

        Another, quite recent behavioural economic analysis concerning the interrelatedness of

various types of drugs is an Australian study conducted by Chalmers et al. which was published in

2010. The subjects consisted of one hundred and one individuals being at least 18 years of age who

stated that they were methamphetamine (a psychoactive drug known for its highly stimulant effects)

users. Besides, the subjects were predominantly male characterised by a relatively low socio-

economic status. The design of this study resembles that of the two behavioural studies discussed

previously. Each participant was assigned an amount of $200 of artificial money which could be spent

freely on several types of drugs (heroin, methamphetamine, cannabis, cocaine, non-prescription

benzodiazepines, alcohol and non-prescription pharmaceutical opioids), the baseline prices of which

were obtained from a 2008 survey.

A novelty of this research is the fact that the authors, amongst the subjects, distinguished those

dependent on neither drug, dependent on methamphetamine only, dependent on heroin only and

those dependent on both drugs. Regarding cross-price elasticity, a change in the price of

methamphetamine did not significantly induce a modification of the demand for other types of

drugs. Exceptions to this tendency were the fact that subjects not dependent on either drug decided

to consume a larger amount of pharmaceutical opioids as the price of methamphetamine increased,

whilst the participants dependent on both drugs simultaneously purchased more heroin and those

only dependent on heroin demanded more cocaine. A change in the price of heroin on the other

hand affected the purchases of other types of drugs more significantly. First, all participants,

regardless of the group to which they belonged, decided to buy more pharmaceutical opioids as the

price of heroin increased. The estimated cross price elasticity concerning these opioids was three to

four times larger for the non-dependent users compared to those for the other three groups.

Second, the subjects dependent on both drugs also substituted benzodiazepines (sleep-inducing

sedatives) for heroin. Conversely, those dependent on heroin purchased less cocaine and those only

dependent on heroin decided to consume a smaller amount of cannabis, both of which imply a

complementary relationship. Finally, the participants dependent on both drugs purchased more

methamphetamine as the price of heroin rose, indicating that the substitution between

methamphetamine and heroin was symmetric. (Chalmers et al. 2010).¹

       ¹ In the original paper, the authors stated that “those dependent on both drugs purchased more heroin” as the price of heroin

       rose. Since this statement contradicts standard economic theory concerning economic substitutes, one may assume that the

       authors actually meant that purchases of methamphetamine increased, but that they mistakenly substituted the word

       “methamphetamine” for “heroin”.

The main policy implication stressed by the authors is the conclusion that although some evidence

for substitution into other drugs – as the price of either methamphetamine or heroin increased – has

been found, the decrease of either methamphetamine or heroin use resulting from respective price

increases evidently outweighs the rise in consumption of substitute substances. However, the

authors caution that their behavioural economic research is subject to limitations and that their

findings should therefore be interpreted carefully. The two foremost shortcomings of this study are

the fixed budget of $200 received by the subjects in each experiment, thereby restraining their

purchasing decisions unrealistically, and the lack of a truly representative sample of Australian

methamphetamine users, since snowball sampling (a sampling technique in which case current

subjects themselves recruit additional subjects amongst their own acquaintances) was used instead

of the more reliable technique of random sampling.

Reasons for complementarity amongst addictive substances

In the previous three chapters, the statistical and experimental evidence for the existence of

economic substitutes and complements regarding addictive commodities has been examined. It has

become evident that the relationships amongst such commodities are far from straightforward.

Some types of drugs appear to be complements to one another, whereas other types seem to serve

as economic substitutes for each other. Even more confusing is the fact that the findings of many

scientific studies devoted to this subject are contradictory. In an endeavour to create order from

chaos, this section will focus on the underlying potential explanations for complementarity amongst

addictive substances. Having a clearer insight with respect to the possible causes of either

complementarity or substitutability will be useful in determining which statistical and experimental

findings are likely to be correct and which are more dubious.

        Despite some discrepancies, most studies confirm the existence of a complementary

relationship (although asymmetrical) between alcohol and cigarettes, in which case the use of

alcohol affects the consumption of cigarettes. This implies that an increase in the price of alcohol –

for example due to additional excise taxes – results not only in a decrease in alcohol consumption

itself, but also in the demand for cigarettes. What could be the reasons for this phenomenon? First,

alcohol is renowned for the fact that it tends to encourage more risky behaviours and actions by

users, since the ability to evaluate the consequences of their behaviours is impaired. Therefore, it is

quite probable that individuals who would abstain from smoking when in a sober state of mind are

more tempted to consume cigarettes after a few drinks, since the hazardous effects on health

induced by tobacco use are recognised to a lesser extent when under the influence of alcohol.

Another potential reason concerns the effects generated by the consumption of cigarettes. In case of

many complementary goods, the purchase of only one of the two goods is utterly useless unless its

complement is purchased as well.

For example, buying ink cartridges is utterly futile if one has not bought a printer first; procuring

DVDs does not make sense without owning a DVD player; buying antivirus software will not be

beneficial in any respect when you are not in the possession of a computer and so forth. In case of

cigarettes, a similar kind of reasoning may be applied. One could argue that the effects of smoking a

cigarette are more pleasurable after consuming alcohol. Under normal circumstances, the benefits of

smoking tobacco may not outweigh the costs (both monetary and non-pecuniary costs), while

consuming cigarettes when tipsy or drunk does induce a greater amount of pleasure of one kind or

another that exceeds the costs of smoking, which remain unchanged. Furthermore, the decision to

smoke might be influenced by situational factors that are in turn affected by the price of alcoholic

drinks (Decker and Schwartz, 2000). For instance, the price of alcohol is one of the determinants of

the willingness to go out tonight: if the price of alcohol is too excessive, going out may not be an

option anymore. If you do decide to go out, you are provided with a more suitable setting in which to

smoke. Thus, going out may result in smoking activities that would not have occurred in another


           Among others, Pacula (1998a) has found a complementary relationship between alcohol and

cannabis. This relationship could be attributed to the so-called ‘gateway effect’, which implies that

using a relatively mild or easily available type of drugs (alcohol) results in an increased probability of

subsequently trying more hazardous or illicit kinds of drugs (cannabis is illicit in most countries).

However, this gateway effect is subject to several criticisms, the most important of which is the fact

that the positive correlation between alcohol and marijuana consumption may not occur as a result

of a gateway effect. Alternatively, the mere fact that licit drugs are available at an earlier age

compared to illicit drugs, could be the reason for the fact that experimenting teenagers and

adolescents start with the use of the former type of drugs, after which they are also eager to use the

latter kind of drugs as soon as this type becomes more easily available.

If the gateway theory does not hold for the relationship between alcohol and cannabis, then it goes

without saying that this theory will not suffice as a potential explanation for a complementary

relationship between alcohol and marijuana. According to Pacula, an increase in the taxes on beer

not only results in a reduced quantity of beer demanded, but also in a reduction of both the quantity

and probability of cannabis consumption. Similar reasons that applied to the complementary

relationship between alcoholic beverages and cigarettes might also apply to the positive relationship

between alcohol and marijuana consumption. Nevertheless, it has to be taken into account that this

relationship is contradicted by other studies providing evidence for economic substitutability

between the two substances.

       Contrary to the asymmetric interrelatedness between alcohol and cigarette use, there is

evidence that a greater cannabis use results in higher quantities of alcohol consumption as well. For

example, purchases of alcoholic drinks seem to be more prevalent in states which have

decriminalised marijuana usage. If we assume that this is indeed the result of a complementary

relationship in this direction rather than the result of a spurious effect due to the existence of

confounding factors, such as unobserved characteristics regarding the states that have decided to

legitimate cannabis use, there must be an explanation for this relationship. One possible explanation

is the fact that consumers of marijuana have the tendency to strengthen or extend their pleasurable

experience by drinking alcohol simultaneously, in which case cannabis serves as their base drug. The

opposite is equally likely, in which case marijuana use supplements alcohol consumption in order to

enhance the feelings of relaxation. It should be noted that this form of reasoning is interchangeable

with the idea that smoking cigarettes is more palatable after the consumption of alcohol, as

mentioned earlier: namely, one could state that tobacco smoking while in an intoxicated state of

mind enhances the effects induced by alcohol, thereby making the overall experience temporarily

more agreeable.

       Furthermore, several papers conclude that a complementary relationship exists between

marijuana and cigarettes.

For two main reasons, this connection between the two substances is modestly predictable. First, in

most countries, among which the US is the main exception, cannabis is blended with tobacco in order

to smoke and inhale the drug more conveniently. Therefore, the purchase of marijuana necessarily

requires the purchase of tobacco (cigarettes). This implies that the price of cannabis negatively

affects the amount of cigarettes purchased. Besides, evidence has also been found of cigarette prices

influencing marijuana consumption. One could argue that the consumption of cigarettes serves as an

ideal steppingstone for using cannabis, as consuming cigarettes familiarises the user with the

principles of smoking, which are also applied when consuming marijuana. On the other hand, it could

also be possible that a person who initiates smoking tobacco is doing so for the fact that, among

other reasons, he or she is more comfortable at inhaling smoke. This means that such an individual

has a higher probability of becoming a smoker, both of cigarettes and cannabis; the fact that

cigarette smoking itself causes the tendency to try marijuana as well may be a mere illusion as

cigarettes are simply more easily available at a younger age and thus consumed before cannabis.

        The final combination of substances of which evidence of concurrent use has been found is

the joint consumption of alcohol and cocaine, in which case cocaine has a complementary role with

respect to the consumption of alcohol. However, this relationship appears to be an asymmetric one,

since evidence obtained from the same studies suggests that cocaine is substituted in exchange for

alcohol when the price of cocaine increases sufficiently. Since this chapter covers complementary

relationships only, for the time being only the former rather than the latter relationship between

alcohol and cocaine use will be clarified by potential explanations. Combined use of these two drugs

may result in an enhanced or extended high (McCance et al., 1993) and is therefore preferable over

the situation in which either drug is consumed separately. In addition, there is a possibility that the

two drugs are mutually consumed as a result of their opposing drug type characteristics, since

alcohol is regarded as a depressant (or downer) while cocaine is a stimulant (or upper). Hence, the

effects of these two drugs when consumed simultaneously balance each other out.

For example, being heavily intoxicated by alcohol and thus having diminished bodily functions such

as limitations of motion and reaction might evoke an individual’s desire to use cocaine in order to

remedy the adverse effects induced by alcohol consumption. Another reason for the apparently

concurrent consumption of alcohol and cocaine could be the attenuation of the negative effects of

cocaine such as nervousness and insomnia (Gawin and Kleber, 1986). Which of these possible

explanations for economic complementarity between alcohol and cocaine is prevailing in reality is a

question that requires additional research before it can be answered properly, which is beyond the

scope of this paper. In the subsequent section, potential explanations for substitutability amongst

several addictive commodities will be scrutinised.

Reasons for substitutability amongst addictive substances

In short, two goods are considered economic substitutes of one another when they fulfil

approximately the same need. This implies that a change in the price of a product affects not only its

own quantity demanded, but also that of its substitute(s). In real life, many varieties of substitute

goods exist, among which different brands of automobiles is just one example. In this case, the need

that is satisfied by all the various brands is transportation from one place to another. A characteristic

that all the different kinds of drugs have in common is the mind altering effect they have upon the

user. Therefore, rather broadly speaking, one could argue that drugs in general gratify the same

need, namely the desire to be in an altered state of consciousness. Solely based on this reasoning,

different types of drugs can be regarded as economic substitutes. However, the scientific evidence

on this subject that has been accumulated over the years does not corroborate this statement. As

discussed previously, some types of drugs appear to be economic complements instead of

substitutes. Thus, one might question whether it is actually true that various types of drugs basically

fulfil the same desire. One could just as easily argue that the effects resulting from the use of the

various types of drugs are so divergent that the same need is not satisfied by all kinds drugs at all.

Nonetheless, evidence of substitutability amongst drugs also exists and this section will therefore

attempt to examine possible explanations for such a relationship.

        Albeit the preceding chapter focussed on potential causes of a complementary relationship

between alcohol and cannabis, scientific verification of a negative correlation between alcohol and

marijuana consumption exists as well. The most straightforward explanation has been discussed

before, namely the fact that consumers are to some extent indifferent regarding the preference for

alcohol and marijuana, since both substances more of less fulfil the same need for intoxication.

However, in case of narcotics there might also be additional reasons for economic substitution that

do not apply to normal substitute goods. One alternative reason why an increase in cannabis

consumption occurs at the expense of alcohol use could be the fact that certain combinations of

drugs have severely adverse and acute consequences.

This eliminates the possibility for simultaneous use of different kinds of drugs and hence the

consumer has to choose which drug to utilise. Therefore, an increase in the price of cocaine for

example will tempt some individuals to switch entirely to ecstasy pills, since partial substitution away

from cocaine to ecstasy consumption is not an option due to the disadvantageous effects of

combined use. Besides, combinations of certain types of drugs exist of which their respective effects

nullify each other, which renders their combined use utterly useless. Among others, an example of

such a combination is the concurrent use of different kinds of psychedelic drugs (e.g. psilocybin

mushrooms and LSD), which are known to be cross-tolerant to one another (Appel and Freedman,

1968). This also coerces an individual to choose which drug to take and hence these drugs can be

considered substitutes. Since the contemporaneous consumption of alcohol and cannabis has neither

dangerous adverse consequences nor effects of cross-tolerance, these two potential reasons for the

reluctance to combine two kinds of drugs are not applicable to marijuana and alcoholic drinks. The

fact that both drugs are depressants and hence fulfil approximately the same need, on the other

hand, could explain the evidence of substitutability between the two drugs.

        Likewise, the apparent substitutability between cocaine and ecstasy, found by Sumnall et al.

(2004), may result from the fact that these two compounds have equivalent, stimulant effects.

Individuals having a desire for temporarily enhanced mental and physical functions are possibly to a

large extent indifferent regarding which stimulant drug to consume (Boys et al.). Therefore, the

respective market prices of such drugs play quite a prominent role in determining the drug of choice.

Naturally, this implies that a price increase of a stimulant drug, e.g. cocaine, affects the demand for

other ‘uppers’ such as ecstasy in a positive manner. By the same token, it is quite possible that the

apparent substitute relationship between heroin and pharmaceutical opioids is caused by the

similarity of their effects (Chalmers et al.). However, the evidence only supports the notion of

asymmetric substitutability, in the sense that an increase in the price of heroin positively affects the

quantity of pharmaceutical opioids demanded (but not vice-versa).

        A rather intriguing finding is the fact that, although cocaine seems to be a complement to

alcohol, cocaine is substituted for alcohol when the price of cocaine increases. One explanation could

be that an increase in the price of cocaine, which is already one of the most expensive types of drugs

considering the short duration of the ‘high’, makes the drug unaffordable, thereby forcing consumers

to abandon cocaine use and instead resort to alcohol as a substitute drug.

        Finally, the observed economic substitutability amongst a range of different kinds of drugs

could be elucidated by the propensity of individuals to restrain their unhealthy behaviours in order to

preserve an adequate level of health. This idea implies that individuals will limit their drug use to a

certain extent and will thus have to decide which is their drug of choice. When determining which

drug to consume, prices play a considerable role. An increase in price of a certain kind of drug may

induce consumers to substitute away from that drug into other kinds. The concept of ‘an adequate

level of health’ is demonstrated by graph depicted below.


                                                                Level of Health

Graph 2: Relationship between individual health and utility.

The graph illustrates that although an increase in health enhances one’s utility, it does so at a

decreasing rate. Stated differently, the marginal utility of health is positive though decreasing as an

individual’s level of health improves, until it eventually reaches the number zero. When an individual

has attained the level of health beyond which additional improvements in health no longer yield an

amelioration in utility, further enhancements in health are rendered fruitless. This means that an

individual who occasionally consumes unhealthy commodities such as drugs and thereby reduces his

or her level of health does not necessarily deteriorate his or her well-being, as long as he or she

remains on that part of the graph where the slope is equal to zero. Technically speaking, a person

maximises his or her well-being by consuming a specific amount of drugs in which case his or her

level of health is the lowest one possible whilst still having a marginal utility of zero². This level is the

represented in the graph by the intersection between the vertical red line and the horizontal axis.

Since the amount of drugs an individual can afford to consume without diminishing his or her level of

health below the threshold is (strictly) limited, economic substitution amongst different kinds of

drugs will inevitably occur.

          Unfortunately, there is no universal explanation for the seeming substitutability among the

various addictive commodities. Which of the possible explanations discussed above prevail(s) in

reality, whether or not in combination with each other, remains tentative and therefore requires

further examination.

² This reasoning only holds under the assumption that this person derives pleasure from the consumption of drugs that exceeds the

monetary costs associated with purchasing drugs.


This paper endeavoured to shed more light on the issue of economic complementarity and

substitutability amongst the class of commodities that could be named ‘addictive substances’.

Despite the fact that most economists have a some kind of fetish for straightforward – and thus

potentially overconfident – answers to economic problems, this paper has done its utmost to

prevent being tempted by the appeal of aligning its conclusions to its preconceptions. Hence, this

paper does not provide a definite answer to the question it has addressed, which would thus

undoubtedly be most disturbing to policy makers, who always crave clear-cut information on which

to base their decisions. However, it has become evident that addictive substances cannot simply be

placed in one category; instead, some of them could be considered economic complements of each

other, whereas others economic substitutes. This means that policy makers should be very cautious

when implementing policies aimed at reducing drug abuse. For example, if the government of a

particular country decides to increase the excise taxes on alcoholic beverages, the result may not just

be an expected decline in alcohol abuse, but also an unexpected rise in the use of alternative kinds of

drugs that are substituted for alcohol. On the other hand, if only complementary drugs with respect

to alcohol exist, then the government is able to kill two birds with one stone: the reduction in alcohol

consumption leads to a simultaneous decline in the use of the complementary drugs. Thus, policy

makers are recommended to obtain relevant information about the specific drug towards which

policy implementations are directed.

          In short, based on the evidence, we can assume that cigarettes are a complement to alcohol,

whereas this does not hold vice-versa; no general consensus has been achieved regarding the

relationship between alcohol and cannabis; while cocaine appears to be a complement to alcohol,

alcohol seems to be a substitute with respect to cocaine; ecstasy and cocaine are likely to be

economic substitutes and finally, pharmaceutical opioids seem to serve as substitutes with respect to


The classification of the different types of drugs based on their respective substitutability or

complementarity with alternative kinds of addictive commodities is recapitulated in the table below.

Results obtained from less reliable studies have been omitted from the table. The names of the drugs

expressed in bold letters can be considered as the ‘reference’ drug (the drug of which the price is

altered). For example, one can observe that cigarettes, cannabis and cocaine are considered

complements with respect to alcohol (written in bold characters).

Types of drugs          Alc.   Tob.    Can.    Coke     XTC   Her.    Amph.      Meth.   Pharmaceutical


Alcohol                        C       C/S     C                      S

Tobacco                                C

Cannabis                C/S

Cocaine                 S                               S

Ecstasy                                        S

Heroin                                                                           S       S

Amphetamine             S

Methamphetamine                                               S                          S

                  Table 1: The abbreviations ‘S’ and ‘C’ utilised in the table denote ‘substitute

                  relationship’ and ‘complementary relationship’ respectively.

          Furthermore, is has become obvious that there are several potential explanations for both

economic complementarity and substitution amongst the various types of addictive goods.

Therefore, it is not unreasonable to assume that a certain combination of the aforementioned

reasons for the existence economic substitutes and complements among addictive products is in

effect in real life.

The most important potential reason for complementarity is the increased pleasure derived from

combined drug usage; the most probable explanations for substitutability are the fact that all drugs

more or less fulfil the same desire for an altered state of mind and that most individuals tend to

refrain themselves from excessive drug use, thus opting for their drug of choice.

        It is important to note that the possibility exists that a certain relationship between two

types of drugs differs among individuals: cannabis might be a substitute for alcohol for one person

but simultaneously a complement to alcohol for another individual. In addition, even the same

individual might regard alcohol a substitute for marijuana at a specific moment in time while a

complement to cannabis at a different moment. An individual’s mood or the setting in which he

finds himself could influence such a tendency.

        Nevertheless, this issue still requires additional study in order to uncover which precise

combination of factors is responsible for either a positive or negative relationship between

whichever two addictive commodities.


Currently the Dutch demissionary cabinet wishes to change its policies regarding cannabis

consumption. In order to reduce the negative externalities associated with drug tourists from

neighbouring countries – such as the disturbance caused by these tourists in the vicinity of coffee

shops or damage caused by these tourists while driving under the influence of cannabis – the

government has decided to implement a law which coerces owners of coffee shops to register their

customers by distributing a limited number of so-called ‘weed card’ per shop, in which case only

Dutch citizens are eligible for receiving such a card. This means that entry in a coffee shop is only

permitted after identifying oneself with the appropriate ‘weed card’, thereby fending off non-Dutch

drug tourists. If everything continues as planned, the law will be implemented on 1 January 2013 in

all Dutch provinces. Several criticisms towards this new policy have emerged from different corners

of Dutch society, the most notable of which is the expected surge in black market activities

concerning marijuana purchases by foreigners and citizens alike. This research might shed additional

light on the consequences of the implementation of the ‘weed card’ regarding possible changes in

drug use.

       Focusing on Dutch civilians only, different scenarios could occur. If we assume that being in

the possession of a ‘weed card’ does not coincide with additional ‘costs’ to the customer of a certain

coffee shop, then we should not expect a shift in drug consumption. However, it is not unreasonable

to assume that there are also individuals who are quite reluctant to hold a ‘weed card’. Particularly

youths would consider the possession of such a card risky as their parents might more easily discover

their drug use. Moreover, since the ‘weed card’ will be linked to the holder’s social security number,

sensitive information might fall into the wrong hands. The customs at the airport for example have

the ability to check whether you are a card holder or not and could accordingly deny you access.

Therefore, a cannabis user will have an incentive to either purchase their marijuana from illegal

street dealers or give up using cannabis in exchange for a substitute drug. According to the studies

analysed in this paper, two alternative consequences may occur.

        If alcohol is considered a complement to cannabis, then a decline in marijuana use as result

of the implementation of the ‘weed card’ (without taking partial substitution to illegally purchased

cannabis into account) will coincide with a decrease in alcohol consumption. If, on the other hand,

alcohol is regarded as a substitute for cannabis, marijuana users will be encouraged to shift their

consumption away from cannabis towards alcohol; hence, alcohol consumption will increase. Since

the introduction of the card will oblige an adequate amount of individuals to relinquish cannabis use

altogether, it would not be surprising that they will fill the ‘pleasure vacuum’ with the consumption

of alcohol instead. Thus, the effect of substitutability will clearly outweigh the effect of

complementarity. One could expect that an insignificant drop in marijuana consumption by a certain

individual will result in a decrease of alcohol consumption as well, which would indicate a

complementary relationship. A substantial decline in cannabis use by someone resulting from

external factors, however, will deprive this specific individual from his or her base drug and hence

stimulate to find an alternative base drug (alcohol).

        In conclusion, it can be said that the Dutch government should also consider the potential

shift in drug use that will result from the introduction of the mandatory possession of a ‘weed card’,

apart from the other disadvantages of such a policy which have already been expressed by critics

from a wide array of backgrounds.


Appel and Freedman (1968) “Tolerance and Cross-tolerance among psychotomimetic drugs”
Psychopharmacology ,Vol. 13, No. 3

Bask and Melkersson (2001) “Rationally addicted to drinking and smoking?” Applied Economics, Vol.
36, Issue 4.

Becker, Grossman and Murphy (1991) “Rational addiction and the effect of price on consumption”
The American Economic Review, Vol. 81, No. 2

Cameron and Williams (2001) “Cannabis, Alcohol and Cigarettes: substitutes or complements?”
Economic Record, Vol. 77, Issue 236, pp. 19 – 34

Cawley and Ruhm (2011)”The Economics of Risky Heath Behaviors” Handbook of Health Economics,
Vol. 2, pp. 95-199

Chalmers, Bradford and Jones (2010)” The effect of methamphetamine and heroin price on polydrug
use: A behavioural economics analysis in Sydney, Australia” International Journal of Drug Policy,
Vol. 21, Issue 5, pp. 381 – 389

Chaloupka and Laixuthai (1997) ”Do Youths Substitute Alcohol for Marijuana?” Eastern Economic
Journal, Vol. 23, No. 3, pp. 253 – 276

Chaloupka, Pacula, Farrelly, Johnston and O’Malley (1999) “Do higher cigarette prices encourage
youth to use marijuana?” National Bureau of Economic Research, Working Paper No. 6939

Chaloupka, Grossman and Tauras (1999) “ The demand for cocaine and marijuana by youth”
National Bureau of Economic Research, pp. 133 – 156

Cole, Goudie, Field, Loverseed, Charlton and Sumnall (2007) “The effects of perceived quality on the
behavioural economics of alcohol, amphetamine, cannabis, cocaine, and ecstasy purchases” Drug
and Alcohol Dependence, Vol. 94, Issues 1 – 3, pp. 183 – 190

Cutler and Lleras-Muney (2010) “Understanding differences in health behaviors by education”
Journal of Health Economics, Vol. 29, Issue 1, pp. 1 – 28

Cutler and Glaeser (2005) ”What Explains Differences in Smoking, Drinking, and Other Health
Related Behaviors” American Economic Review, Vol. 95, pp. 238 – 242

Decker and Schwartz (2000) “Cigarettes and Alcohol: Substitutes or Complements?” National Bureau
of Economic Research, Working Paper No. 7535

DiNardo and Limieux (2001) “Alcohol, Marijuana and the American Youth: the unintended
consequences of government regulation” Journal of Health Economics, Vol. 20, Issue 6, pp. 991 –

Farrelly, Bray, Zarkin and Wendling (2000) “The joint demand for cigarettes and marijuana:
evidence from the National Household Surveys on Drug Abuse” Journal of Health Economics, Vol. 20,
Issue 1, pp. 51 – 68

Gala ma and van Kippersluis (2010) “The Theory of Socioeconomic Disparities in Health over the Life
Cycle” Tinbergen Institute Discussion Paper 2010-079/3

Galama, Pacula, van Kippersluis, Buchinsky, van Ours, Angrisani, Avendano and van Ourti “Explicit
modeling of the trade-off between addiction and health, longevity”

Goel and Morey (1995) “The Interdependence of Cigarette and Liquor Demand” Southern Economic
Journal, Vol. 62, No. 2, pp. 451 – 459

Grossman, Chaloupka and Brown (1999) “The demand for cocaine by young adults: a rational
addiction approach” Journal of Health Economics, Vol. 17, No 4, pp. 427 – 474

Gruber (2000) “Risky behaviour among Youths: An Economic Analysis” National Bureau of Economic
Research, Working Paper No. 7781

Loewenstein, Brennan and Volpp (2010) “Asymmetric Paternalism to improve health behaviours”
Journal of the American Medical Association, Vol. 298, No. 20

McCance-Katz, Price, McDougle, Kosten, Black, Jatlow (1993) “Concurrent cocaine-ethanol ingestion
in humans: pharmacology, physiology, behavior and the role cocaethylene” Psychopharmacology,
Vol. 111, No. 1

Measham (2004) “The decline of ecstasy, the rise of ‘binge’ drinking and the persistence of
pleasure” Probation Journal, Vol. 51, No. 4, pp. 309 – 326

Model (1993) “The Effect of Marijuana Decriminalisation on Hospital Emergency Room Drug
Episodes: 1975 – 1978” Journal of the American Statistical Association, Vol. 88, No. 423, pp. 737 –

O’Donoghue and Rabin (2001) “Risky Behavior among Youths: Some Issues from Behavioral
Economics” National Bureau of Economic Research, Vol. ISBN: 0-226-31013-2

Pacula (1998a) “Does increasing the beer tax reduce marijuana consumption?” Journal of Health
Economics, Vol. 17, Issue 5, pp. 557 – 585

Pacula (1998b) “Adolescent alcohol and marijuana consumption: is there really a gateway effect?”
National Bureau of Economic Research, Working Paper No. 6348

Pacula, Chirqui and King (2003) “Marijuana Decriminalization: What does it mean in the United
States?” National Bureau of Economic Research, Working Paper No. 9690

Petry and Bickel (1998) “Polydrug abuse in heroin addicts: a behavioral economic analysis”
Addiction, Vol. 93, No. 3, pp. 321 – 335

Petry (2000) “A behavioral economic analysis of polydrug abuse in alcoholics asymmetrical
substitution of alcohol and cocaine” Drug and Alcohol Dependence, Vol. 62, Issue 1, pp. 31 – 39

Reiman (2009) “Cannabis as a substitute for alcohol and other drugs” Harm Reduction Journal, 6:35

Rogeberg (2003 ) ”Taking absurd theories seriously: Economics and the case of rational addiction
theories” Philosophy of Science, Vol. 71, No. 3

Sumnall, Tyler, Graham, Wagstaff, Cole (2004) ”A behavioural economic analysis of alcohol.
amphetamine, cocaine and ecstasy purchases by polysubstance misusers” Drug and Alcohol
Dependence, Vol. 76, Issue 1, pp. 93 – 99

Thaler and Shefrin (1981) “An Economic Theory of Self-Control” Journal of Political Economy, Vol. 89,
No. 2, pp. 392 – 406

Van Ours (2003) “A pint a day raises a man’s pay; but smoking blows that gain away” Journal of
Health Economics, Vol. 23, Issue 5, pp. 863 – 886

Viscusi (1992a) “The value of Risk to Life and Health” Journal of Economic Literature, Vol. 31, No. 4,
pp. 1912 – 1946

Williams, Pacula, Chaloupka and Wechsler (2001) “Alcohol and Marijuana Use Among College
Students: Economic Complements or Substitutes?” National Bureau of Economic Research, Health
Economics, Vol. 13, No. 9, pp. 825 – 843


Shared By: