The Benefits and Costs Of Gambling: Some Policy Implications

Commissioned Report: April 2010

Keywords: Health; Regulation & Taxation; Social Policy;

Executive Summary

The following is not the usual executive summary. Rather it draws out the policy implications of the known costs and benefits of gambling, thereby summarising the main report, which consists of two parts. Appendix I is a general account about how economists think about social benefits and costs set around the standard social cost evaluation approach. This provides a context for the main report which considers the benefits and costs of gambling.

The economics approach does not judge gambling in moral terms.

A key element of the economists approach is that if individuals voluntarily choses to participate in an activity there is an increase in social benefit, providing they act rationally, and the prices they face reflect the true social cost of the resources they use. In such cases the social costs of the used resources are offset by the benefit that the decision-maker obtains from the consumption. (if the benefits were less, the consumption would not proceed).

When a resource usage is not taken into account in the individual consumption decision there may be ‘externalities’ in which the social cost of the resources is not included in the benefit decision and hence is not offset by any benefits. It is thought that these resource externalities are not large in the case of gambling, especially if the gaming duties are seen as a contribution to the resource costs of its regulation.

Externalities include crime. According to a survey of those involved in illegal activities, about a quarter said they would not have committed a crime had they not been gambling. This amounts to about 1100 criminals a year. While gambling is associated with crime in this way, it is not obvious that the regulation of gambling can (or should) be adapted to minimise such crime. Thus while a common form of gambling crime is embezzlement, it is better dealt with by measures that deal with embezzlement generally. (It is accepted that there is a need to prevent parts of the gambling industry being taken over by criminal elements and this is best dealt with by direct industry regulation.)

Social and benefits costs are not only resource costs. There may be human (or intangible) costs and benefits arising from gambling affecting individual’s welfare. The direct intangible benefits from gambling offset the social costs which are taken into account during the consumption decision. The evidence is that many people would have lower personal welfare if they had not the opportunity to gamble. (This includes housing and lotteries as well as more complicated forms of gambling.)

However there are gamblers who, on reflection, think they would be better off were they not to gamble. This seems to contradict the economists’ assumption that people make rational decisions in their own self-interest. Economists explain this by the rational decisions being ‘inconsistent over time’. More simply such gamblers are ‘addicts’, at the extreme they are ‘problem gamblers’.

A second group who may be worse off are associates of heavy gamblers. There are people who suffer a loss of welfare from other’s gambling. It is unlikely that their loss of welfare has been taken into account when the gambler was deciding to thinking gamble (although there may be individuals who limit their gambling because it affects associates). The evidence is that a considerable number of people are in this category.

The regulation of gambling therefore faces a tradeoff between

– providing opportunities to participate in gambling for the population as a whole, who decide rationally (or not too un-rationally) and as a result of their decision are better off;


– protecting those who are addicts (including ‘problem gamblers’) and do not make rational decisions which take into account the personal detriment of their gambling nor of the impact of their gambling on their associates.

This is an extremely difficult task; interventions are likely to limit the benefits to some rational decision-makers, and will still leave some irrational decision-makers exposed to opportunities to gambling.

The policy area which has most struggled with this sort of tradeoff is the regulation of alcohol. It has a public policy objective of harm minimisation. It is not obvious that the specific lessons of alcohol regulation are always useful for gambling regulation, but there may be some general lessons to be learned from it.

A major difference between alcohol and gambling regulation is that there seem greater differences in potential social detriment among the various modes of gambling than there is for types of alcohol modes where the amount of absolute alcohol is treated as the primary trigger for detriment irrespective of the form in which it is drunk.

But drinking circumstances matter; that may have some parallel with modes of gambling.

There is some evidence that electronic gaming machines are the single largest source of net social detriment from gambling. However it is possible that some other modes have higher unit detriment but fewer people involved.

The report notes there is little policy rationality in gaming taxation. (Excise duties on alcohol are the residual policy intervention when all effective targeted interventions have been put in place.) Little is known about the impact of ‘prices’ on gambling behaviour, and hence the effect of changing gaming duties.

Implicit in the report is that although there is probably insufficient data to carry out a full social evaluation of the costs and benefits, of gambling, further quantitative progress could be made. It may be particularly valuable in identifying the more problematic modes of gambling.


The Economists’ Approach to Gambling

Many people have strong moral views on the activity of gambling. Among them are economists who, like the population, have a wide variety of personal views. However the professional economics account of gambling does not take such views into account. This is not because economists are personally amoral, but rather the profession does not try to impose a morality on the public.

Thus, when two adequately informed carry out a voluntary exchange between them which does not affect others, an economist observes that each considers themselves better off (otherwise the exchange would not have been voluntary) and so each experiences an increase in personal welfare. Such exchanges are usually considered a ‘good thing’ (even if a third person may have reservations about the transactions).

Not all transactions in an economy meet these conditions . The participants may not be adequately informed, the transaction may not be voluntary (as in the payment of tax, but we shall meet other examples) or involve others, or after the event the person may regret the decision – a complication which we shall discuss in due course as ‘time-inconsistency’; which is the economists approach to dealing with (some forms of) addiction.

Probably most gambling transactions meet the conditions of adequate information (assuming that each gambler has a reasonable knowledge of the ‘true’ odds), and the transaction is voluntary and does not impact on others. In such cases the economic analysis considers that there has been a net gain in welfare. Note that the economic analysis does not distinguish between the mode of gambling. Others may want to distinguish between housie or betting on horses, say, from playing at the casino or poker machines.

Parallels with Alcohol Consumption

There are parallels between the economists’ treatment of gambling and the treatment of alcohol consumption. In terms of the preceding analysis, most cases of alcohol consumption are beneficial (or any external costs – to be explained below – are small). Voluntary transactions are therefore not of a great public policy interest.

However, some transactions are not socially beneficial – as when someone else is required to bear costs which were not taken into account when the individual considered her of his consumption decision. In some cases the social cost is very high – as when a drunk driver causes road deaths. As a result the community attempts to regulate the consumption of alcohol in order to minimise the social costs of alcohol consumption, without reducing the benefits from the consumption. (Part of the reason that much alcohol consumption is socially beneficial is because the regulatory framework is reasonably effective.) Even so there is a tradeoff. Typically reducing the social damage from alcohol reduces some social benefits because it limits some voluntary transactions which are not socially damaging.

The same applies to gambling. Because, as we shall see, some gambling is social damaging, there is seen to be a need to regulate it. Doing that may also reduce the benefits from participating in gambling.

Economists recognise that some regulatory frameworks are more efficient than others, so that while tradeoffs cannot be avoided, they can be minimised. Designing an effective framework – one that minimises the tradeoffs and minimises the damage – is a major policy challenge.

Optimal design requires the measurements of costs and benefits under different assumptions. This is extremely difficult. A first step is to consider the costs and benefits.

Tangible Social Costs Associated With Gambling

The tangible (resource) social costs of gambling are those resources which are utilised relative to some counterfactual scenario – perhaps no gambling) which are not offset by some benefit. In essence it is these net costs which are not taken into account when the gambler (or whoever) makes the decision to gamble.

While they have not been measured, the indications are that those resource costs are not large (once any benefit is offset). The most substantial net costs may be the regulatory costs which are not charged to the gambling providers or passed onto the gamblers.

Does the Additional Economic Activity Generate a Benefit?

Insofar as people choses to participate in gambling activities and have a higher standard of well-being as a result, then gambling activities represent a social benefit.

However, sometimes it is claimed that the extra industry activity including employment is also a benefit. But such a claim ignores that if there was no gambling, individuals would spend their income on other activities which would also generate employment. Thus the net gains in resource terms from introducing gambling activities are small and may be largely ignored.

The introduction of a gambling facility may be beneficial to a region if it encourages gambling (and possibly other) activity in the region by diverting the expenditure from other regions (as when outsiders visit the region to participate in gambling or locals remain rather than go outside in order to participate). However the gains to one region are a loss to another. This also applies for the nation as a whole if the gambling facility attracts tourists from overseas (or discourages nationals from going overseas to gamble). This means a region may strongly lobby for a new facility – say a casino– because it will benefit (say from employment generation and such like) to the detriment of other regions.

Crime Effects

Costs of crime. invole some caveats to this analysis. It is only if resources are used rather than transferred that crime appears as a cost in the economists’s assessment. Thus the resources used for policing, justice and corrections are all costs which, typically, have no offsetting benfit in the individual’s gambling decision (unless there is a tax to cover them)

But money transfers from the gambling criminal to someone else are not treated as a social cost. This arises because the method does not treat resources which are transferred as a social cost. This approach avoids comparing the benefit the person who gains from the transferred resources (perhaps after a number of transactions) with the loss the person who involuntarily loses the resource suffers.

Even so, crime is a social issue, and it threatens the integrity of the property rights which underpin a market economy. On a wider assessment of the social impact of gambling, there in no question that insofar as gambling generates crime that is a downside of gambling.

A survey of the population found that 1.3 percent of the population admitted they had engaged in illegal activities (mainly stealing and fraud) during the last 12 months. High participation gamblers were more likely to be involved in illegal activities. Of those involved in illegal activities, 25% (i.e. about 0.3 percent of the survey population, or around 11,000 adults) said they would not have committed a crime had they not been gambling.[1]

Intangible (Human) Social Costs Associated With Gambling

While the (tangible) impact of gambling on resource usage seems small, gambling may also impact – positively and negatively – on the welfare of many people including those associated with gamblers.

There is little epidemiological evidence on gambling’s impact on the mortality and morbidity, presumably because the quantum on individual health is small and difficult to measure. It seems likely that the main intangible effect is on the quality of life of gamblers and their associates.

This section now reports the results of a survey in which people were asked about their personal well-being on a number of dimensions.[2] The data was then used to assess the impact of gambling on individuals.

The method compares those who are involved in gambling with those who are not. After controlling for a wide range of socioeconomic variables – such as age, gender, marital status, employment, education and income the study – the study found that those who gamble heavily report lower personal well-being that those who do not.

There is methodological problem here. because as a rule the cross-sectional data cannot identify cause. Thus the method does not discriminate between those whose personal well-being is reduced by going gambling and those of a melancholic disposition who choose to go gambling as a result.

However the survey also asked respondents whether people considered there would have been a change in their well-being had they not been gambling in the last 12 months. Those who were more heavily involved in gambling were more likely to think their well-being would have been improved along the following dimensions: physical health, mental well-being, relationships with families and friends, quality of life, overall satisfaction of life, financial situation, housing situation, material standard of living, study performance and care giving of their children. All these were statistical very strong differences relative to those less involved in gambling. (The only life dimensions where there was not a strong effect were work performance and care-giving to the elderly.) [3]

We will return to the paradox that many individuals voluntarily engaging in gambling appear to be less well-off as a result, and here note that this result seems to suggest that there is a strong causal path from heavy gambling to the lowering of well-being. This does not rule out there is still the reverse path, but it seems likely that it is not as strong.

In the following paragraphs we assume that the strong gambling to poorer well-being causal path explains much of the correlation. However precise quantities are not reported here, just the general tenor of the research findings.

Satisfaction with Life

First, suppose there was no gambling (i.e. the counterfactual is that all current gamblers do not gamble). According to the econometric investigation there would be a reduction in satisfaction of life if there was no gambling. We take this to mean that gambling improves their quality of life of many people, as we might expect given they choose to pursue the activity. There appears to be more gamblers gainers that there are losers. (Note this is numbers only; it is possible that the losses of the losers from gambling – including associates of gambling – exceed the gains of the more numerous ‘recreational’ gamblers.)

A second counterfactual scenario considered those people who had associates (usually family and friends) who were gambling. It appears that if many gamblers stopped gambling the people associated with them would have been better off – many moving from being ‘satisfied with life’ to being ‘very satisfied with life’. The econometric evidence concludes that there can be a major detrimental impact on the lives of those associated with heavy gambling, a conclusion supported by the rest of the survey and by other studies.

A third counterfactual scenario was the assumption that there were no electronic gaming machines, and there was no displacement of gambling activity (that is, people gave up gambling and did not move to other modes such as casinos and the TAB). The evidence from the survey is that EGMs are a major associate with poor life satisfaction and by far the most important in aggregate. (That does not mean that other forms can necessarily be ignored – even if there is no displacement. The two largest ‘substance abuse’ problems are tobacco use and alcohol misuse. Even so considerable effort is put into narcotic use because while numbers involved are much smaller the costs to the individuals and their associates can be very high.)

In summary,

1. Many people get an improvement in their life satisfaction from recreational gambling.

2. However there are a group of heavy gamblers who experience deterioration in their life satisfaction from gambling. The largest group are those who use electronic gaming machines.

3. There is also a substantial group of people who are associated with heavy gamblers who would have a higher satisfaction of life if their associates did not gamble.

There are similar results for other dimension of life. The impact on the mental well-being and physical gamblers is much larger than for life satisfaction; however the impact on associate’s mental and physical health is slightly smaller.

Resolving the Paradox

It is easy to understand how the quality of life of associates of gamblers may be worse off as a result the gambling for it seems likely that many gamblers do not take into consideration the impact of their behaviour on others, just as heavier drinkers ignore the impact of their drinking on others.

But additionally, there appears to be a group of who chose to gamble at – according to their own judgement – a detriment to their personal well-being on many measures. Since their involvement in gambling is by choice this suggests a paradox in terms of the standard theory that people behave rationally making decisions to make themselves better off.

The best explanation comes from behavioural economics – as elaborated in the Appendix.[4] It may be gamblers exhibit time-inconsistent behaviour in which they maximise decisions at each point in time, but the decisions do not cohere together over time. In particular they decide to participate in gambling activities with which with hindsight they regret. In effect they are people who are unable to control their urge to gamble, even though in the long run they know they will be worse off. They are ‘serious addicts’ or ‘problem gamblers’. (Such behaviour leads to Self- Exclusion Programs where a person who wants to control their gambling asks to be denied use of gambling facilities.)


While there are many people whose life satisfaction benefits from the availability of the facilities for gambling, there are a minority who are worse off if they, or an associate, gamble. This deterioration in the quality of life due to excessive gambling seems to be the most serious social cost of gambling.

There are also some resource usage from regulation of gambling from enforcement following crimes committed to facilitate gambling, and for treatment of problem gambling. In value terms these are probably not as large as the equivalent reduction in intangible benefits.

There is some preliminary evidence that Electronic Gaming Machines may be the largest source of social costs of gambling. What it not known is the cost per user. It may be higher in some lesser used gambling modes.

Prices, Taxation and Charitable Donations

Implicit in this discussion is the notion that the prices that individuals pay for a product (good or service) reflect the social costs of the resources that contribute to that product. This is generally true in the economy providing ‘social’ cost is understood in to have a technical economics meaning reflecting the values arising from private market decisions.[5]

There are specific taxes on gambling. Gaming duties amount to $240m a year. They are of historical origin, vary from by mode of gambling and do not obviously reflect any overall rational policy.

There is also a quasi-tax insofar is that some profits do not go to the entrepreneur or facility owner but are used to fund various community charities (and problem gambling programs). Again the rational of this is not clear. The profits are probably supernormal, arising from the restrictions on entry. They could be replaced by either the central government raising gaming taxes or by local authorities charging an annual licence fee for each gaming facility. These are issues not directly relevant to those covered in this report.


The theory which underlies the use of social costs evaluations, is the same theory used in cost benefit analysis (CBA). The principles in WHO International Guidelines for Estimating the Costs of Substance Abuse were consciously based on the CBA. (The author on this report was one of its authors.) Whenever a measurement issue of how to the treat social cost arises, the standard should be that of the CBA.

The Economist’s Notion of Costs

The purpose of this appendix is to convey the intricacies of the concepts, including drawing attention to various limitations. It will seem tedious to some, insufficient to others. Economists deal with costs and benefits with the same forensic care as other professions do of their key concepts

Economics always measures costs (and benefits) as an opportunity cost, that is relative to some alternative. While sometimes the alternative is not stated (it may be obvious), especially in public policy analyses it is almost always wise to state an explicit ‘counterfactual scenario’ (the standard term for the alternative).

For example, evaluations of the social cost of alcohol consumption sometimes have a counterfactual scenario in which there is no alcohol consumption, and sometimes one in which only heavy or unsafe drinking does not occur. That matters because low level consumption may be mildly beneficial – and so the aggregate social cost will depend on which of the two counterfactual scenarios is evaluated.

Having established a counterfactual scenario, the next step is to list all the differences between that and the actual scenario. Many things will be the same; only those where there is a difference are of interest since the ‘sames’ cancel out.

Such lists underpin the all the social costs studies. However the number of items in the list, it is likely to be difficult to appreciate their overall significance, and so there is a need to aggregate each to a single index, or for comparison purposes a single number which is the difference between the two indices.

The most common way of aggregating the items on the list is to calculate the social costs of the resources being used. The logic behind this is that standard economic theory attributes a social value to each resource. In essence this is the market price (although sometimes there may be adjustments for the impact of taxation). A heuristic indication of the underlying theory is that individuals value the resources in their consumption decisions at the market prices they pay for them.

So the list of items are aggregated by adding them together weighted (valued) by each’s market price. The meaning of each aggregate is that it is the (market) value of the resources used for the items on the list. The difference between the two items is the difference in the market value of the resources in the two lists, and therefore the resources used in the two scenarios.

If the aggregate for the actual exceeds that for the counterfactual, then the actual scenario uses more resources than the counterfactual one. (And vice versa.)

Some Complications

The Population

Implicit in an analysis is the relevant population. Often one does not have to be explicitly identified, but sometimes it matters. In particular the evaluation of an activity which involves (temporary) migration – such as a gambling facility which attracts patronage from outside the region or discourages locals from going outside the region – may have a different conclusions depending on whether the relevant population perspective is the region, or the whole country (or world).


A caution is necessary of the approach in regard to the impact of crime. Suppose something is stolen – that is its ownership is transferred involuntarily (at first, because it may be transferred after). It will appear in a different places in the the list of the counterfactual and the actual lists but it will be (typically) valued equally in both, so they cancel each other out. This is an illustration of a more general phenomenon – cost-benefit analyses ignores transfers of resources. Briefly economics has been unable to offer a means of incorporating such involuntary transfers in a cost benefit analysis – a CBA is about resource usage not about resource ownership. It follows that social cost evaluation has the same limitation. When crime is involved it is best to add a footnote explaining any transfers do not appear in any aggregate.[6]

How Benefits Fit In

In order to understand how consumer benefits relate to these costs, consider a situation where the differences are due only to voluntary exchanges.

Suppose the counterfactual scenario was there was no potatoes. Let us assume that when there are no potatoes that the land used for growing potatoes and all the associated production resources are used for kumera which is consumed instead.

The lists of the two scenarios will reduce to individuals consuming potatoes in the actual and more kumera in the counterfactual. When they are valued at social cost the aggregates will have exactly the same value (since we assumed that the resources to produce them were exactly the same).

(Of course the true story is more complicated than that. Perhaps the potatoes require less ground to grow or fewer workers, so other things can be produced in the actual relative to the counterfactual scenario.)

Now there is a problem here for while the aggregate values of the lists will be the same (that is the social costs for producing the outputs for the two scenarios are the same) the scenario with potatoes is evidently better than the scenario without, since consumers choose to consume some potatoes rather than only kumera.

The reason for this paradox is that items are valued at the resource costs used to produce them. This is not the same as the benefits from the consumption. The market prices used in the valuation reflect the value of the last quantity consumed (consumption at the margin), not the value of every item consumed which in aggregate will be more than the outlay. This excess is called ‘consumer surplus’.

(Consumer surplus, is the benefit a person gets from consumption above that the cost that was used to purchase it – and hence the resources used to provide it. The theory says that where there is a number of units purchased the first consumed are of greater benefit than the later ones. In principle the consumer surplus can be measured if one knows the demand schedule for the product. Usually we dont – and often we dont for all consumers either.)

The convention is to ignore this consumer surplus, partly because it is difficult to measure and partly – because of this difficulty – there is a tendency to make exaggerated claims for (or against) each consumed good by advocates (or detractors). What is implicitly assumed is that the consumer surpluses are proportional to the outlays for all resources and products involved in the transaction.

There is perhaps a more subtle interpretation. When there are voluntary transactions then it does not matter, as we saw with the with and without potatoes scenarios. The answer is clear anyway. The introduction of a new product – such as potatoes or a new mode of gambling – will lead to some benefits above the cost of the resources insofar as individuals voluntarily choose to take up the opportunities it presents (and assuming that there are no externalities).[7]

Where there are involuntary transactions the comparison becomes more complex.

When Social Costs Differ From Private Costs

Involuntary transactions greatly complicate the story. Suppose the purchaser converted some of the potatoes into vodka. In many circumstances this would be no different from converting them into mash or chips or whatever. However, suppose as a result of the vodka consumption the consumer had an accident while driving a car. Various resource-using activities would result. Some in the accident might end up in Accident and Emergency, perhaps the car (or cars) would require repairs (or be written off), police resources would probably be involved; later the report discusses what if there is any damage to the quality of life of those involved through injury or death.

The critical point here is that the outcome is an involuntary one, unplanned by anyone involved. and not taken into account when the initial private consumption decision was made. Yet it results in resources being diverted from a preferred use to deal with the consequences of the unintended decision.

It does not matter that many of the subsequent actions are voluntary (like friends visiting the victims in hospital). What is important is that when the vodka was being drunk, the consumer did not take into account these other resource uses (costs). Once these external costs are included, the value of consumption as determined by the consumer no longer offsets all the resources that society ultimately used in producing the product

In terms of the two scenarios, the potato scenario will have the higher resource aggregate indicating that more resources have to be used to attain it than for the other scenario. Those resources will have to have been diverted from other (presumably worthwhile) uses.

From the perspective of social costs then the potato scenario with these additional external costs is inferior to the non-potato scenario because it takes more resources to produce it. However, as we have already noted there is a sense in that the potato scenario is superior to the non-potato scenario because it gives more people choice, and they take up the options.

Suppose that the vodka induced accident were to occur only very rarely (and so the social cost difference between the two scenarios was tiny). We might well make the social judgement that the net effect is that the potato scenario is superior to the non-potato scenario, despite the off accident.

However, this is not a conclusion that can come totally from the economics, for the economist qua economist does not claim to make such judgements. A social judgement is necessary to assess the tradeoff.

Externalities and Taxation

An alternative way of thinking – an earlier approach by economists – about involuntary transactions is externalities, which are those resources involved in a transaction are not taken into account when the decisions are made. For example the drinker of the vodka is unlikely to take into consideration the impact of any accidents which might be generated from their consumption decision.

What is to be done about such externalities? The basic economists advice is to internalise them, that is get arrange the price signals so that the decision takes the additional resource costs it generates into consideration.

This is often difficult – or impossible – to implement. In the simple case of tobacco smoking the additional social costs from externalities are probably closely proportional to the amount of tobacco used. A tax related to the average social cost has the effect of signalling to the purchaser-consumer of the additional social costs that will be triggered by the smoking of the tobacco, so they indirectly take into consideration these external social costs.

More typically, the social costs are not so simply related to average consumption. For instance the likelihood of having a traffic accident following the imbibing of alcohol will depend on whether the drinker is going to drive, and the circumstances in which the drinking takes place (a drinker is less likely to use a car if the consumption is at home). An even greater divergence arises because the social costs associated with each drink is a function of recent levels of the drinker’s consumption. The seventh drink in a session will typically do much more social damage than the first drink.

These are examples of circumstances of aligning, by a tax or levy, the purchase price of a good with the exact total social costs its generates is not practical. That is why where the externalities are great but their incidence is erratic it is usual to use a range of interventions, each of which is (or should be) targeted towards particular behaviour which generates social costs. Public policy towards alcohol consumption is an example of this multi-intervention approach.

Where it is impossible to target precisely the interventions and eliminate all social costs from externalities it is not unusual to also impose a tax or levy on all consumption, with the purpose of reducing the externalities by discouraging average consumption. (Road usage is another example.) Its incidence relative to social costs will be erratic and some consumption which does not generate externalities will be inhibited (which would a social detriment) while some consumption which generates external costs will continue (again another social detriment).

Prohibition is also an option which is used in some cases (almost) totally as for narcotics or partially as in the case of prohibiting sales of alcohol to inebriated persons.

Implicit in the analysis is that the consumer is well informed about any costs involved in the consumption decision, although the price is all they may need to know in regard to the physical resources that are using (since in principle in a well-run market economy market price reflects social costs). However, they may not be aware of the impacts on, say, their personal well-being. Thus the need for educational campaigns to inform people of the implications to their health of tobacco consumption or of excessive alcohol consumption.

Intangible (Human) Costs

Thus far the analysis has focussed on tangible resource costs. However a transaction is likely to involve a change to the quality of life of those involved. Cost benefit analysis, on which this methodology is based, recognises this by incorporating an evaluation of any changes.

If there was no allowance for changes to the quality of life from a treatment (or intervention), the CBA would lead to some peculiar decisions. For instance health care would be (mainly) an expense and the CBA conclusion would be not to provide any health care, missing the point that tangible resources were being used to improve the intangible quality of life.

Including the quality of life in the calculations poses some difficulties. One is how to quantify the quality of life. The standard way is to use a measure called the QALY (quality adjusted life year). There has been considerable work done on QALYs but they are better dealing with physical limitations and not as successful at assessing psychological states on the quality of life. The latter is important when the impact on gambling is being assessed.[8]

Once a QALY measure has been established it is often necessary to put a dollar value on each QALY. (This is avoided in cost-effectiveness analysis which assesses the cost of different treatments with the same ultimate goal but that is more limited than an evaltion.) Derivation of the dollar value of a QALY is problematic – and there is not a really robust estimate for New Zealand.

Moreover, such estimates of the dollar value there are can involve very large values, so that the intangible costs of the loss of quality of life swamp the tangible losses from unintended use of resources. This may well be true – people may value life far more than resources. But there is a tendency to compare the aggregate net social costs with, say, GDP. The comparison is invalid because it compares a resource and human cost (the tangibles and the intangibles) with a resource cost (tangible GDP only); nevertheless implicitly or explicitly people do.[9]

Are losses of quality of life taken into consideration by the consumer? If they are then they are not an externality and should not appear in aggregate net social costs. Suppose someone decides to go on a binge drinking episode, knowing that the following day they will have a low quality of life. That reduction should not appear in the aggregate net social costs, since it has already been taken into account by the decision-maker. But do drinkers consciously take into consideration that their probability of cirrhosis of the liver will be elevated by their imbibing? (Do they even know what this probability is?) Social costs studies tend to assume that long term effects on health and mortality are not included in the consumer’s calculations. [10]

An important non-tangible impact may be on the quality of life of others. This may range from persons unknown to the consumer (such as victims of a car accident) to close associates (such as a partner, children and other close family members who has to deal with the consequences of the consumption – such as assault).

Are Consumers Rational?

Another problematic assumption is that consumers are always rational and act in their own best interests. If they are not, a voluntary transaction may not take into consideration some of the costs which are normally assumed to be offset by the benefits from transaction.

This issue is a contentious one in economics. Some economists see rationality as a key element in the discipline’s methodological foundation, although the emphasis on strong rationality is relatively recent – less than four decades old. Some economists are unwilling to contemplate the possibility of irrationality because they see it leading to the conclusion that someone else can judge better a (normal) person’s best interests. This does not follow at all; if X cannot judge their best interests well, it does not follow that Y can do it any better.

There are a host of psychological and economic studies which demonstrate that sometimes individuals (mature adults) do not appear to make irrational decisions this. In 2002 the Economics Prize in honour of Alfred Nobel was awarded to Daniel Kahneman and Vernon L. Smith for their work in behavioural economics. (Kahneman is a psychologist, Smith’s work is in experimental economics, indicating that the relevant economics has taken considerable notice of psychology, and followed some of its approach – this represents a new development in economics.)

Partial theories abound. The difficulty has been how to incorporate this sort of phenomena into comprehensive economic analysis. For instance that social cost evaluations assume individuals are rational enables a netting out of private costs against private benefits. What happens if they are not rational? Any answer requires some theory of decision-making.

Where the evaluation involves such activities as alcohol, narcotics, tobacco or gambling it focuses on addiction. In 1988 Kevin Murphy and Gary Becker (who is also in a Nobel laureate but not particularly for this work) proposed a theory of ‘rational addiction’. It has been subsequently elaborated but it is also heavily criticised. As far as I know it has not been taken seriously in the social cost evaluations studies, in part because there is a simpler and more elegant solution which can be readily applied (if there is the data).

The essence of time-inconsistent decision-making (which economists use as an alterative to the notion of ‘addiction’) is that people make rational decisions at any point in time but they make inconsistent decisions over time and may regret past decisions (even though there is no subsequent changes in the assumed circumstances since the decision as when meeting a checkpoint after a night drinking which will increase the regret).

Consider a person who lives only for the moment. It is not hard to tell stories where a moment later they regret the decision they have just made. A more general – but more complicated – example is that when people are planning for the future they may not discount time at the same rate throughout the future. One formal version of this is ‘hyperbolic discounting’ but a simpler version is that immediate returns are given a greater weight than they would be under a conventional rational behaviour. What this means is that individuals will later regret a previous consumption decision.

A simple example is the person who decides to go into a bar to have a couple of drinks, in there consumes far more than they planned before they went in, and later regrets the additional decision. In each of the ‘before’, ‘during’ and ‘after’ situations they make rational consumption decisions, but the decisions are not consistent over time.

As far as social evaluation is concerned, one may want to argue that the regretted excess is not a benefit to the consumer who made the decision. In which case there is no benefit to offset the resource costs involved. If so, they should be explicitly incorporated in the evaluation.

Note that this approach does not designate a small minority as addicts, whose consumption can be treated differently from the great majority. Rather, anyone can be vulnerable to time-inconsistent decision-making. Often the inconsistency may be small – say eating one chocolate more than planned – but sometimes the issue may be large enough to require some social attention. As like as not, such cases are subject to social cost evaluations.

This solution is elegant and simple, but note it has slipped in the assumption that the evaluation should be from a long term perspective. Even if this approach is adopted, it is unusual, however, to have good estimates of the retrospectively regretted consumption.

(Another consequence of the time-inconsistent decision-making is that an excise tax on the to-be-regretted consumption can be of benefit to the individual insofar as it reduces regretted decisions. [11])


[1]  Casswell, S. et al (2008).

[2]  Casswell, S. et al (2008); Easton & You (2009).

[3]  Among the heavy gamblers 2.1% said they would have done some study or employment related training had they not been gambling in the last 12 months and 8.3% said they would have been in paid employment.

[4]  Another possibility is that there are many dimensions of life and that for each person the gambling raises some dimensions but not other, and by some sort of fluke, the heterogeneity of dimensions leads to ana average reduction on the dimensions. This seems unlikely.

[5]  The most important exceptions are that not all environmental resources are properly priced – but they are not particularly important in gambling decisions – and the impact of specific taxes.

[6]  Policing, justice and correction costs are able to be allowed for as they are resource usages as a consequence of crime. But that does not cover the effects of resource transfers which involve not resources usage.

[7]  There is an alternative way of interpreting the data as follows. It asks what would happen (at the margin) if the society moved (slightly) in the direction of the counterfactual scenario. Would resources usage go up or down? This is not a common interpretation although it is formally correct.

[8]  Other measures include the DALY – disability adjusted life years which have proved not to be as robust as QALYs, and deaths but mortality ignores morbidity (injury).

[9]  The SHORE study on the social costs of gambling measured self-rated well-being – more loosely described as ‘happiness’. This is a burgeoning research area but as yet there are no valuations placed upon well-being so the SHORE study did not produce a single aggregate dollar value for social costs of gambling.

[10]  This cannot always be correct – given the Achilles syndrome, of choosing a a short life and an exciting one over a long dull one. But probably that is an infrequent enough phenomenon not to require an adjustment in the calculations.

[11].  O’Donoghue & Rabin (2006)

Selected Bibliography

Australian Institute for Gambling Research (2001) Social and Economic Impacts of Gambling in New Zealand, Final Report.

Banks, G. (2002) The Productivity Commission’s Gambling Inquiry 3 years On.

Casswell, S. et al (2008) Assessment of the Social Impacts of Gambling In New Zealand, Centre for Social and Health Outcomes Research and Evaluation & Te Ropu Whariki. Report for the Ministry of Health,.

Collins, D. & H. Lapsley (2003) ‘The Social Costs and Benefits of Gambling: an Introduction to the Economic Issues,’ Journal of Gambling Studies, 19(2):123-147.

Curtis B. (ed.) (2002) Gambling in New Zealand, Dunmore Press.

Easton, B. H. (2002) ‘Gambling in New Zealand: And Economic Overview’, in B. Curtis (ed).

Easton, B.H. & R.Q. You (2009) Measuring the Impact of Gambling, Paper to the Wellington Statistical Group (

O’Donoghue, T. & M. Rabin (2006) ‘Optimal Sin Taxes,’ Journal of Public Finance, November 2006, pp 1825-1849.

Rankine, J. & D. Haigh (2003) Social Impacts of Gambling in Manukau City: A report for Manuaku City Council, July 2003.

Single, E., D. Collins, B. Easton, H. Harward, & H. Lapsley (2002) International Guidelines for Estimating the Costs of Substance Abuse: Second Edition, WHO.

Walker, D. M. (2003) ‘Methodological Issues in the Social Cost of Gambling Studies,’ Journal of Gambling Studies, 19(2):149-183.

Wynne, H. J. & M. Anielski, (2000) The Whistler Symposium Report. The first international symposium on the economic and social impact of gambling.