National Drug Strategy
National Drug Strategy

The costs of tobacco, alcohol and illicit drug abuse to Australian Society in 2004/05

2.1 to 2.7

prev pageTOC |next page

2.1 The definition of costs

The definition of the economic costs of drug abuse used in the three previous Collins and Lapsley studies, and again for the purposes of the present study, is:

This cost concept, which applies to what has been labelled the demographic approach, is based on the calculation of the size and structure of a hypothetical population in which no drug abuse had occurred. The hypothetical population in this counterfactual situation is then compared with the actual population size and structure, as a basis for estimating drug abuse costs.

Most of the literature on drug abuse cost estimation pays virtually no attention to the implicit counterfactual situation against which the costs of abuse are estimated. However, if a study's assumed counterfactual situation is not made explicit, interpretation of its results becomes difficult. For example, a statement that "the social cost of smoking in year X was Y million dollars" is not particularly informative unless we know what was the alternative situation assumed for the purposes of the calculation. The study might be referring to:
To examine the first concept would not be useful since a high proportion of smoking-related morbidity or mortality in a given year results from smoking in earlier periods. If the social costs of smoking in a given year were calculated on this basis they would be minimal. Clearly, then, we should be examining the impact of smoking over an extended period of time and this implies comparison with the counterfactual situation in which there was no smoking over this extended period. The comparison is being made between the actual smoking situation over an extended period and a hypothetical alternative situation of no past or present smoking.

Thus, for the purposes of this study, the counterfactual situation which is compared with the actual 2004/05 drug abuse situation is one in which there has been no abuse of the drug in question for an extended period of time. In this study that period is assumed to be at least 40 years.

To postulate a situation of no past or present drug abuse is not to suggest that such a situation is necessarily achievable. It is, almost certainly, not achievable. Thus the costs attributed to drug abuse will exceed by a considerable margin the potential reduction in costs available to public policies designed to reduce drug abuse. Identification of the potential reduction in costs, given the implementation of a set of appropriate public policies, involves the estimation of avoidable costs, which is not undertaken in this study. The issues involved in estimating avoidable costs are dealt with extensively in Collins, Lapsley et al. (2006).

Top of Page

2.2 The concept of drug abuse

The definition of tangible abuse cost used in this study refers, as discussed above, to the extra resources which would have been available if there had been no past or present abuse. This implies that, had there been no abusive consumption, the resources devoted to satisfying those consumption demands would have been released for other consumption or investment purposes. This cost concept, therefore, implies the need for both definition and measurement of abusive consumption.

There is no problem in using the term "abuse" when referring to the consumption of tobacco or illicit drugs. In the case of tobacco, virtually all consumption is harmful to the smoker and/or to others. In the case of illicit drugs, by definition, society has decided to proscribe their consumption, with the implication that any consumption is abuse. However, the use of the term "drug abuse" in relation to alcohol is problematic. There is no concept of alcohol abuse in the National Alcohol Strategy (Commonwealth Department of Health and Aged Care, 2001), only alcohol misuse. On the other hand, a new National Alcohol Strategy which has recently been published (see www.alcohol.gov.au) offers no preference, separation, or definition for the terms alcohol misuse or abuse.

For convenience, drug abuse is defined for the purposes of this study as consisting of tobacco abuse, illicit drug abuse and alcohol abuse/misuse.

A definition of abuse which is meaningful in medical terms is that it occurs when a relevant aetiological fraction is greater than zero, i.e. when drug abuse adversely affects the health of the user or of any other individual. This is a very narrow medical definition which ignores many relevant abuse costs. A more comprehensive economic definition, which encompasses non-medical costs such as accidents and policing, is that drug abuse exists when drug use involves a net social cost additional to the resource costs of the provision of that drug. Abuse occurs if the community incurs net costs as a result of drug use.

As discussed below, this study concentrates on social costs because these are the costs that are relevant for the development of public policy. However, the consumption of alcohol and, to a much lesser extent, tobacco can yield social benefits and such social benefits (in effect, negative social costs) are equally relevant for the development of public policy. Accordingly, this study estimates the net social costs (social costs less social benefits) of drug use.

2.3 Abusive and addictive drug use

Identification of consumption due to addiction requires separate consideration for tobacco, alcohol, and illicit drugs. For tobacco it is concluded that there is a small proportion of tobacco consumed which is not addictive. However, from a health perspective, all tobacco consumption is abusive, i.e. there is no safe level of consumption. Even those aetiological fractions which are negative for tobacco do not negate the cancers, heart disease and other conditions caused by smoking.

Alcohol presents different challenges from the perspective of the determination of social costs. In Collins and Lapsley (1996) it was concluded that 20 per cent of alcohol consumption was by addicted drinkers. (It should be emphasised that this does not suggest that 20 per cent of alcohol consumers are addicted, as the average alcohol consumption of addicted drinkers will be much higher than that of other drinkers.) This does not address the issue of misused alcohol which is consumed by non-addicted drinkers, the results of which can include costs of illness, road accidents, violence (including domestic violence), reduced productivity, crime and drug-induced accidents. Accordingly, the proportion of abusive alcohol consumption is assumed to be higher, at 30 per cent.

It is recognised that this proportion represents an educated guess and it has been suggested that the true proportion may well be much higher—at least 50 per cent. This figure affects only one component of the alcohol cost estimates—the resources used in abusive consumption. The estimates of all the other components of the social costs of alcohol are made independently of this figure. The figure estimated for the resources used in abusive alcohol consumption, on the basis of the 30 per cent assumption, is $1,689m. If a proportion of 50 per cent were adopted, this figure would rise to $2,815m.

As indicated earlier, the consumption of all illicit drugs is treated here as abusive because Australian society has decided that use of these drugs is illegal and that their consumption represents abuse. In effect, this determines the counterfactual scenario as a situation in which there is no consumption of illicit drugs. The degree to which such unauthorised use is abuse, as discussed elsewhere in this report, is a question outside the scope of this study.

Top of Page

2.4 Demographic and human capital approaches

The approach to cost estimation adopted in previous Collins and Lapsley studies and in the present study has been labelled the "demographic" approach, rather than the "human capital" approach of some other studies. The brief comparison of the two approaches presented below is based on that written by the present authors for inclusion in the International Guidelines for Estimating the Costs of Substance Abuse, produced by the World Health Organization (Single et al., 2003). These guidelines acknowledge the legitimacy of both approaches.

Both approaches relate to the valuation of the loss of production arising from the abuse-related deaths of otherwise productive members of society. Both approaches compare production and abuse costs in the actual situation with those in the hypothetical alternative situation which would have existed had there been no past or present substance abuse.

The difference between the two approaches relates to the way in which the production costs of premature mortality are treated.

The human capital approach is to estimate the value of the worker's future production stream, brought back to present day values by the use of an appropriate discount rate. A thousand dollars received this year is worth more than a thousand dollars received next year (even if there is no inflation) because this year's resources become available for consumption or investment purposes a year earlier and so produce interest receipts or profits a year earlier. The use of a discount rate acknowledges this fact and adjusts for the difference between present and future values. Two major problems arise in the human capital approach—how to forecast future production levels and how to choose the appropriate discount rate.

The demographic approach compares the actual population size and structure with the size and structure of the hypothetical alternative no-abuse population. From this comparison the actual and hypothetical outputs are compared to yield the production costs in the year of study of past and present substance abuse. The major difficulty in this approach is the estimation of the alternative population structure.

The essential difference between the two approaches can be summarised in the following way. The human capital approach calculates the present and future production costs of abuse-induced deaths which occur in the present year. The demographic approach calculates the present production costs of abuse-induced deaths which have occurred in past and present years. Which approach should be adopted depends, therefore, upon which type of information is needed and upon the precise nature of the counterfactual scenario. The two approaches are complementary rather than competitive.

The two approaches are likely to yield different aggregate cost relativities for the three drug categories. This arises from the fact that the age at which an Australian drug-attributable death occurs is lowest for illicit drugs, relatively low for alcohol and highest for tobacco. The average potential years of life lost (PYLL) for all illicit drug-attributable deaths (for the ages 0–74) is 39, for alcohol it is 32, and for tobacco it is only 16. Thus, the human capital approach will place a higher emphasis on costs which involve the discounting back to present day values.

The human capital approach is necessarily always adopted in benefit–cost analysis (BCA) where the nature of the task is to compare, on a common basis, time streams of costs and benefits. Abuse cost estimates can have a different objective. In the case of the present study it is to estimate the costs of drug abuse which are borne in a given year. It can be argued that for the purpose of estimating abuse costs, this is a more comprehensible and useful concept of cost than that delivered by the human capital approach.

If abuse cost studies are to be extended into the BCA of proposed drug programs, the human capital approach is indispensable. The data sets used in the two approaches are largely overlapping and it would be perfectly feasible to extend the present study to produce human capital-based estimates.

A major problem of both approaches is the valuation of life. The interpretations to be placed on the life valuations differ between the two approaches. The human capital approach estimates the value of the loss of a life. The demographic approach estimates the value of the loss of a year's living.

Top of Page

2.5 General equilibrium impacts of drug abuse

The abuse cost concept adopted here, by referring to the resources which would be released for consumption or investment purposes, explicitly avoids the problem of what economists refer to as the general equilibrium impact of drug abuse.

It is often argued that, if an industry producing abused substances ceased to exist, there would be substantial loss of employment, output and income. Thus this employment, output and income are represented to be benefits of drug abuse. The difficulty with this analysis is the implicit assumption that the opportunity cost of resources used in the drug industry (that is, the value of the sacrificed alternatives, which is their most productive alternative use) is zero. This is an assumption that would appear impossible to justify. If such logic were pursued, there would be no benefit from microeconomic reform, since all resources released as a result of that process would have no alternative uses. It is difficult to imagine that agricultural resources used in the production of abused substances would have zero opportunity cost and it is impossible to imagine that manufacturing and distributive resources would have no alternative use.

A similar problem arises in valuing production losses resulting from substance abuse. If there were high levels of unemployment, the loss of production might be small or zero (because the prematurely deceased could be replaced by workers who otherwise would be unemployed). The costs of drug abuse borne by society would apparently be much lower in periods of high unemployment than in periods of low unemployment. Increases in unemployment would apparently reduce drug abuse costs.

These types of issues can only be settled by the use of an appropriately specified econometric model and an appropriate counterfactual. Even if such models existed, and we are unaware of their existence (although Richter and Gori (1980) made such an attempt), there would still remain, with the human capital approach, the problem of forecasting future rates of unemployment, growth and productivity over the remaining normal lifetime of the prematurely dead. It is not possible to produce robust estimates of the opportunity cost over extended periods of time of resources used in the production of abused substances.

The following point should once again be made explicit. The approach of this paper is to estimate the value of the resources which would have been made available had there been no past or present drug abuse. The alternative uses to which these resources would have been put would be largely determined by government macroeconomic and microeconomic policies and to forecast these uses would therefore be speculative.

2.6 Private and social costs of drug abuse and their policy significance

A fundamental issue of abuse cost estimation is whether the estimates should incorporate the private costs and benefits of drug consumption and production. In the classic work which explores this issue, Markandya and Pearce (1989) define the total costs of drug abuse as the private costs plus the social costs. "To the extent that the costs are knowingly and freely borne by the consumer or producer himself, they are referred to as private costs but to the extent that they are not so borne but fall on the rest of society they are referred to as social costs". Thus, according to Markandya and Pearce, total costs equal private costs plus social costs. What Markandya and Pearce call social costs are often called negative externalities. (Negative externalities occur when individuals or firms undertake actions which impose costs upon other individuals or firms, while providing no, or insufficient, compensation to those who bear these extra costs).

An important issue, as the two authors point out, is "the extent to which the consumer is aware of the costs that he bears. If his actions are determined by a perceived cost that is in fact less than his actual cost, the difference between the two can be viewed as a social cost". This is because "the individual himself has not adjusted his behaviour to reflect these higher costs and they are, therefore, unaccounted for".

In these circumstances individuals are not necessarily behaving irrationally. They are simply adjusting their behaviour according to the best available, relevant, information. As Markandya and Pearce implicitly accept, costs borne by the individual drug abuser can be social costs even if that individual is rational, if those costs have not been knowingly incurred. This point merits further elucidation.

In a private market transaction the consumer is assumed to make a comparison between the costs of purchase and the benefits received as a result of that purchase. If the consumer has proceeded with that purchase it can be assumed that the private benefits exceed the private costs (that is, there is some consumer surplus). But what if, as a result of misperceptions about the level of private costs or benefits, actual private costs exceed private benefits? The community will then be worse off than it would have been had the purchase not gone ahead. The community as a whole (through the purchaser) has borne a cost because of the lack of appropriate information on the part of the purchaser. The purchaser has borne that cost, a social cost.

As an example of this analysis, assume that a motorist purchases a new car in ignorance of the fact that this particular model is liable to burst into flames in relatively minor rear-end collisions, with catastrophic implications for its occupants. In this theoretical example the fire risk had not been revealed to the population of motor vehicle buyers so that the purchaser had no way of taking it into account in the purchase decision. Had the risk been known, the purchaser would either have been unwilling to buy that model at all or only willing to buy it at a significantly lower price. In this circumstance the purchaser is worse off because the benefits of ownership of that model are, in the purchaser's own fully-informed estimation, less than the purchase price. The real wealth of society has been reduced by this transaction even though the cost is borne by the private purchaser (see Collins and Lapsley, 2002).

Thus, the crucial issues in relation to the estimation of the social (external) costs of abuse are:
If any one of these conditions is not satisfied the resultant costs are social costs. Only if all three conditions are simultaneously satisfied will the resultant costs be private costs. Since the objective of this study is to estimate the social cost of drug abuse, purely private costs are not relevant to, and are not counted in, this study.

Why is so much attention paid to the distinction between private and social costs and benefits? As the Productivity Commission (1999, p. 4.3) says in its report on gambling, it is not because private costs are unimportant: "… in fact, often they are far more significant than the social benefits and costs of an activity. Rather, they generally do not justify government action on the basis that:
Thus the existence of private benefits and costs does not normally provide a justification for government intervention, unless the distribution of private benefits and costs is seen to be in conflict with society's concept of fairness.

If the objective were to estimate the total costs of abuse, both private costs/benefits and social costs/benefits should be incorporated in the estimates. When only social costs (according to Markandya and Pearce terminology) are estimated, private benefits and costs should be ignored. The present study is concerned with the social costs of drug abuse and so does not estimate the value of purely private benefits and costs. It is social costs which are relevant to the formulation of public policies.

Top of Page

Being fully informed about the private costs of abuse requires the abuser to have access to, and have the ability to process and evaluate, epidemiological information on the effects of drug use. It also requires the drug user to be able to evaluate the probable future health and other costs resulting from the drug use. It is difficult to believe that drug users, by their nature, are fully-informed, or even well-informed, about the costs of their abuse (Courtwright et al., 1989). Abusers are likely to be less well-informed than non-users, since well-informed users are much more likely to have ceased or avoided abuse.

Potential major sources of information for drug abusers are public health campaigns, advertising by manufacturers and information disseminated by the media. Public health campaigns and media information are highly useful but clearly are not perfect vehicles for conveying relevant health information. Advertising of tobacco (which was legal until relatively recently) and alcohol in Australia has generally provided at best little information other than price and at worst impressionistic images totally at odds with the actual effects of abuse of these products.

The question of rationality also raises interesting issues. Rationality, as defined in the paper by Becker and Murphy (1988) on the theory of rational addiction, implies utility maximisation over time. Stevenson (1994) says that the theory of rational addiction "assumes that drug users are rational, forward looking utility maximisers who base consumption decisions on full knowledge of the consequences of addiction."

It should be noted that rationality here implies full knowledge. The theory of rational addiction, which has been widely quoted by industry groups as supporting their case, does not merely demand rationality; it demands both rationality and full knowledge. Furthermore, it demands rational behaviour in a situation of full knowledge at the time at which the addiction was acquired. A high proportion of addictions are acquired in the early- or mid-teens when it would seem that the presence of both rationality and full information is unlikely.

The notion of rationality as maximisation of utility over time is itself an interesting one. The comparison by an individual of benefits and costs accruing over time can only be undertaken by using some concept of a time preference rate. Are very high time preference rates, which place a very high value on current benefits and a very low value on future costs, rational? Is there any time preference rate which is not consistent with the notion of rationality? If not, rationality seems to lose any significance since any behaviour pattern can be seen to be consistent with utility maximisation. But society itself is clearly unwilling to accept all behaviour patterns (for example, self-destructive behaviour even when it does not impose social costs).

The present research assumes that addicts do not satisfy the rationality and information requirements discussed above. Ellemann-Jensen (1991) takes issue with this analysis, pointing out that total addiction has been assumed to imply that the smoker enjoys no utility at all from smoking but continues to smoke. He suggests that "such behaviour is clearly in contrast to the hypothesis of utility-maximisation in standard economic theory". In fact, this is not necessarily the case. A 1991 editorial in the British Journal of Addiction suggests that addiction involves, inter alia:
In these circumstances, the objective of drug consumption may well be to avoid highly unpleasant effects of withdrawal rather than to gain any positive benefits. Since the withdrawal effects result from previous consumption of the addictive drug, avoidance of these effects can hardly be viewed as a benefit of drug consumption. Short-run utility maximisation need not necessarily imply long-term positive overall benefits from drug use. Rational behaviour of an addict is not the same as rational behaviour of a person contemplating acquiring an addiction, and the two cannot be equated.

Top of Page

2.7 Costs of consumption of abused drugs

If all drug abuse ceased to exist, the consequent reduction in consumption would release resources which could be used for other consumption or investment uses. Thus, on the basis of the definition of tangible cost adopted in this study and earlier studies, the resources used in abusive consumption represent one of the costs of drug abuse. The correct measure of these resources is the value of consumption rather than the value of production since the latter fails to take into account imports or exports of the abused substances. Data on consumption at market prices must be adjusted to a basis of factor cost by deducting taxes less subsidies.

Estimating the turnover of illicit drugs provides serious difficulties. Because the market is illegal, there are no national accounts data on consumption. The street value of illicit drugs is not a useful measure from the point of view of this study because a high proportion of the street value represents a return for the risks involved in drug dealing. The resources used in illegal drug activities would, in their alternative uses, undoubtedly command much lower rates of return. If dealing in or use of the drugs in question did not attract legal sanctions, their prices would be very much lower, although presumably if illicit drugs were legalised they would be taxed.

prev pageTOC |next page