Taxing Harmful Drinking

Paper for the launch of “Alcohol No Ordinary Commodity 2 ed”, 19 August, 2010.

Keywords: Health; Regulation & Taxation;

Executive Summary: This paper argues that the harm from the consumption of alcohol can be reduced by targeting the minimum of price of alcohol, but by using an excise drawback rather than setting a minimum price of alcohol.  By doing so the profits from the price hike goes to the public exchequer rather than the industry.

Excise duties on alcohol are one of the means of reducing the harm from drinking alcohol. By raising the price of a drink, consumption is reduced. Some of this avoided consumption is harmful, so harm is reduced. But not all drinking is that harmful. This paper is about how we can use excise duties to reduce harmful drinking without impacting so heavily on drinking that is not harmful.

Historically excise duty on alcohol was first used as a convenient way of raising revenue for the government. So Governor Hobson imposed duties on wine and spirits, and on sugar which was necessary for brewing beer. By the end of the nineteenth century the consumption of alcoholic beverages was seen to be a social evil, and so excise duty on it was treated as a sin tax.

Now economists have no special expertise in matters of sin, so we cannot really address this objective. However as the excise duty remains a source of revenue we maintained an interest in it as a fiscal instrument and, of course, whatever the reason – good or bad – that a tax was imposed, we were interested in the behavioural impact of the change.

By the 1970s, as welfare economics began to become elaborated and applied, economists recognised that the consumption of alcohol could impose external, or social, costs on others’ consumption, so there was a case for imposing the tax so that the actual price the drinker faced more closely covered the social (or external) costs as well as the drinker’s private costs. I wont detail this argument today. But note that there are problems of defining social costs (or harm). Sometimes it is argued that the government should only levy enough revenue to cover the costs to the public purse; others argue that the levy should take into consideration the costs that are imposed on others – such as a drunkard beating his wife – and others that the costs include the individual imposes upon her or himself, which are not taken into consideration when making the drinking decision. Calculating the social costs of alcohol misuse proves to be a complicated – indeed treacherous – exercise.

A further complication is that the social costs are not simply a matter of the absolute alcohol in the beverage. That is a reasonable approximation in the analogous case of smoking, where the quantity of tobacco consumed is near enough proportional to the damage to the smoker’s health. However, in the case of alcohol such factors as who drinks, where they drink, how many drinks they already have had and what they do after the drinking all affect the social cost. We can take other measures to moderate these costs – drink drive laws are a significant example – but in practice such measures, while having considerable effect, do not eliminate all the social costs they are targeting.

Almost in despair at the complexity of the policy conundrum of how to tax social costs, in the late 1980s the government set the excise duty on alcoholic beverages as a fixed rate of the amount of absolute alcohol irrespective of any other circumstances. There are a couple of anomalies to this rule, which I shall ignore. Today I am asking whether we can change the excise duty regime to sharpen its effectiveness in reducing harm..

The proposed approach recognises that there are some types of drinking which are more harmful than other, and some that are more amenable to price effects than others. To understand this I need to explain what we know about the impact of taxation on alcoholic beverages.

In the past (or in today’s economics courses which is often much the same thing) it was common to assume that the total outlay was proportional to the amount of absolute alcohol consumed. That is not (near enough to) true. Taste matters and so does ambience. If it did not, we could not explain why people buy $20 bottles of wine when for less than $10 they can purchase the same quantity of absolute alcohol. Nor can we explain why they may be willing to pay $40 for the bottle of wine when they are in a restaurant, when they could purchase it for $20 from a supermarket and drink it on the street.

Moreover, if an additional excise of a $1 was put on each bottle, how do we know that the $20 drinker will not just switch to a bottle of wine that previously cost $19? If they do, they drink the same amount of absolute alcohol with no change to the harmful consequences of the drinking, but one which they judge to have an inferior taste. The tax increase may be useful if you are after more public revenue or want to punish sin, but it may not do much for dealing with harmful drinking.

Actually, we know a price hike will in practice reduce the consumption of absolute alcohol to some degree. There are two (or three) groups who are particularly likely to respond more than average by reducing their drinking – and importantly, they may be among the most harm inducing drinkers. The first group is those who spend a lot on alcohol, because the excise hike reduces their effective income. Second, a price hike will have a bigger impact on those who cannot trade down quality, because they are already paying the minimum price for absolute alcohol. (Teenagers may have those characteristics of drinking cheaper alcohol and having less income; additionally those who have not got into the habit of regular drinking may also be very responsive to price.)

One implication of this analysis is that while it is generally believed that the demand for alcohol is not very price elastic, that is sensitive to changes in prices, some groups will be more sensitive than average. Unfortunately most of the econometric studies are unable to investigate this because we lack the data – although there is strong evidence that the price elasticity of teenage drinking is high. However on the basis of what we know, a good guess would be those who drink the cheapest alcohol or drink are the most price sensitive.

A nice illustration of the different elasticities is that when in 2003 the duty on fortified wines and light spirits was raised, the market for light spirits – astonishingly cheap, flavoured absolute alcohol drunk in bulk by the young and alcoholics – collapsed, and the product is no longer available, while there was little impact on sherry, presumably because elderly ladies continued to pay the little extra for their evening pleasure.

That sort of analysis has led to the proposal of a minimum retail price for liquor. Unfortunately those who have advocated it have little idea how it could be practically implemented. It turns out to be quite complicated.

But even if it was simple, who would get the financial benefit from the higher prices? You will not be surprised to learn that the financial beneficiaries from a minimum retail price would be the suppliers of the liquor – the producers and retailers – with the higher price accruing to them as higher profit. Would that be a good thing? Given that the cheapest drinks are made more profitable by a minimum retail price, has not the industry an incentive to increase the number of drinkers of the profitable cheapos, so that one might end up with exactly the opposite to the desired outcome – more people drinking more low price liquor?

So the Law Commission did not recommend a minimum retail price regime. Instead it recommended increasing the excise duty. That would put up the minimum price but the additional revenue would now accrue to the government, not to the industry.

The trouble with higher excises is that they put up the price of the drink for those further up the price hierarchy who will not be doing much harmful drinking (except for the impact on their liver and so on, which is a risk they are apparently willing to take). These drinkers see the excise duty hike as the government gouging them for no justified reason. So politically the government is reluctant to risk the wrath of the general electorate, even if the change would reduce harmful drinking.

That leads one to explore whether we can levy a higher excise duty on cheap absolute alcohol drinks relative to the more expensive ones, better targeting the harmful drinking, while ordinary drinkers would not be as heavily penalised.

A means of doing this would be to drawback the excise duty as price of the drink rose. Drinkers be purchasing more taste and ambience, are already paying paying more for their absolute alcohol so that under the drawback system they would pay less excise duty.

I illustrate this with the Law Commission’s proposal to increase the excise duty rate by 50 percent, which increases the average price of alcohol by 10 percent.

To simplify, what happens with a standard beer – a 330ml bottle of 4 percent absolute alcohol by volume (aabv)? Currently the excise rate is $25.476 per litre of absolute alcohol. The Law Commission recommended that the rate be increased to $38.214 per litre,. That would increase the excise duty on the bottle from 34 cents (including GST) to 51 cents, and the price for a standard bottle from $1.33 to $1.50 or 13 percent, which is only a little higher than the average increase of 10 percent. So the Law Commission proposal only slightly tilts the change in price against cheaper drinks.

The basic idea of an excise drawback is that, as currently, the duty is imposed on absolute alcohol, but as the purchase price of the absolute alcohol – that is the price of the drink – rises, the excise duty phases out. Thus those purchasing cheaper absolute alcohol will pay more excise duty on their absolute alcohol than those who are purchasing more expensive liquor.

Suppose we double the excise duty rather than just increase it by 50 percent as the Law Commission recommended, but we apply a drawback to generate about the same amount of revenue for the exchequer – around $430m a year. The effect of the drawback would be to tilt the levy so it raises the price of the cheapest liquor more than it does on average, so that the cheaper the absolute alcohol the higher the excise duty. .

To do this the drawback levy would be to reduce by 12.5 cents for every dollar the liquor cost to produce. Because this is a revenue neutral regime, the average price for all liquor will still increase by only 10 percent as in the case of the Law Commission proposal, but those buying the cheapest absolute alcohol will face larger price increases, while those buying the higher value drink – who are paying for taste as well as absolute alcohol – will experience smaller increases, or no increase at all.

What happens to the price of the standard beer? Under the drawback, the cost of a bottle of standard beer would be $1.55, higher than under Law Commission proposal.

What happens to bottles of wine which are currently costing $10 and $20 illustrates the tilting effect. Under the Law Commission’s proposal the price of the $10 bottle would rise to $11.08 or about 10 percent – about average. Meanwhile the price of $20 bottle would rise to $21.08 – the same dollar amount but a percentage increase more like 5 percent. So there is some tilting towards against those going for absolute alcohol rather than taste.

Under the 12.5 percent drawback regime the price of a $10.00 bottle would rise to $11.17 or by 12 percent, while price of the $20 bottle would actually fall to only $19.92 (as the drawback would exceed the hike in excise duty).

What about if the beer is purchased in a bar? The Law Commission report examples an on-license beer at $6.00, the margin above the off-licence being for the drinking environment. Under their tax proposal the price would rise to $6.17 or just 3 percent. Under the drawback regime it would rise to $6.22 or nearly 4 percent (assuming that the drawback was based on the off-licence retail price).

For completeness, what about a 330ml ready-to-drink with an aabv of 5 percent? Under the Law Commission proposal its price would increase just under 11 percent, under the drawback just under 12 percent. The small increase reflects that at $2.00 the RTD is a relatively expensive source of absolute alcohol compared to the standard bottle of beer.

With 12.5 percent Drawback

Drink Current


Cost of




Law Commission






Beer (bottle)














Beer (bar)*







Wine (bottle)







Wine Bottle










* Drawback on excise levy imposed at off-licence equivalent.

*Note that wine is levied at a lower level of excise duty than beer.

These examples, summarised in the table, illustrate the more one is purchasing absolute alcohol – rather than taste and ambience – the greater the price hike, which makes sense if the aim is to reduce harmful drinking, by penalising those whose priority is absolute alcohol rather than convivial drink.

These outcomes are constrained by the Law Commission target of increasing the average price of alcohol by only 10 percent. Had they chosen a higher increase – it is possible that a drawback scheme would have made them bolder – then the price hikes would have been greater (as would the revenue for the exchequer).

Having introduced the idea of an excise drawback, and illustrated how it tilts the price against those whose drinking is more about absolute alcohol than taste and ambience let me briefly list some of the problems.

First there is an administrative problem. (A particularly tricky area is on-licence drinking.) The beauty of the current tax regime of excise duties and GST is its administratively simplicity. A drawback would add to the complexity, even if the levy was made at the wholesale level, rather than retail level that I have assumed here.

A related complication is that for expensive drinks the drawback would exceed the excise duty – you can see that beginning to happen with the $20 bottle of wine (at about $30 there would be no excise to pay). One could put a cap on the amount of the drawback, but that would add to the complexity of the arrangement.

Third, we have not, and cannot, eliminate the criticism that excise hits the absolute alcohol drinking poor . But let’s be careful. Do we really believe that the poor’s aim is to drink absolute alcohol while the rich go more for taste and ambience? That wont prevent dull and lazy journalists and others making the cheap shot that the measure is anti-poor. They wont mention that the poor probably suffer more from alcohol induced harm, nor will they mention the extra $480m a year of additional public revenue which could be used to reduce taxes on the poor and raise their incomes.

The purpose of this paper has been to get public policy to think more clearly about what drinking causes harm and to suggest that a major source is cheap absolute alcohol. Although a minimum price regime is both complicated to administer and favourable to the profits of the alcohol suppliers, we need to think about low price alcohol. The Law Commission suggested that we should raise the excise duty even further.

What I have suggested here is that an excise duty increase with a drawback would be even more effective. But I have acknowledged that it is administratively more complex. In effect I am proposing a minimum retail price for alcohol, which is set through the excise duty. By doing so the profits from the price hike goes to the public exchequer rather than the industry.

Finally, in my view we should see the tax regime as a part of the totality of reducing policies, and we need to pursue those that are effective as vigorously as we can. Indeed in my view, we should only tax because these other policies cannot, for one reason or another, eliminate all harm. To put it the other way round, in an ideal world in which there was no harm from alcohol consumption, there would be no need for a excise duty on alcohol. That would be equivalent to assuming alcohol was just like any other product. But as the book we are launching today well demonstrates, alcohol is no ordinary commodity.

Further Reading

This paper belongs to a sequence of papers which include:

Taxing Harm Modernising Alcohol Excise Duties, report for the Alcohol Advisory Council: (;

Reviewing the Sale of Liquor Act: Tax and Pricing Consequences, report for the New Zealand Law Commission: (;

Targeting the Minimum Price of Alcohol, report for the New Zealand Law Commission: (