Australian columnist Miranda Devine recently claimed that smoking-related deaths save cost to the health system:

“Well cigarettes aren’t that bad really when you think about it,” Devine said.

“It might shorten a couple of years off the end of your life, but that’s a good thing. That actually saves money in the long run for the health system.”

Source: http://www.9news.com.au/health/2017/02/20/11/29/miranda-devine-says-smoking-isnt-that-bad

Can this claim be substantiated?

I presume she was implying the Australian public health system, and so answers should probably take into account other effects of smoking on the Australian government budget (tax revenue, state pensions etc).

2 Answers 2


TLDR: The claim does appear to have support from scientific studies.

Details: The original claimant has tweeted a citation to support her claim:

It's true! Smokers die earlier so cost the health system less than healthy people who clog up hospitals in old age http://bmjopen.bmj.com/content/2/6/e001678.full

The cited article ("The net effect of smoking on healthcare and welfare costs. A cohort study", Jari Tiihonen, Kimmo Ronkainen, Aki Kangasharju, Jussi Kauhanen) reaches the following conclusion:

Conclusions Smoking was associated with a moderate decrease in healthcare costs, and a marked decrease in pension costs due to increased mortality. However, when a monetary value for life years lost was taken into account, the beneficial net effect of non-smoking to society was about €70 000 per individual.

A review article from the Blog OpenPop that looks at the scientific literature in this area broadly agrees with these findings:

In conclusion, the weight of the evidence suggests that smoking is a positive fiscal externality. Smokers pay less in income taxes than non-smokers, yet pay considerably more in consumption taxes. Their lifetime healthcare costs are probably slightly lower than those of non-smokers. And they take far less from the pension system than non-smokers, relative to what they pay in.

They warn of several caveats to this conclusion:

First, it is tentative. At least one major study, namely Sloan et al., reaches the opposite conclusion. Second, it may only apply to the U.K. (and other similar countries), where healthcare and pensions are largely financed by the state, and taxes on cigarettes are very high. Third, as many authors have noted, the fact that smoking may have a positive fiscal impact does not imply that efforts to curtail smoking reduce social welfare. Indeed, if each QALY [quality-adjusted life year] is assumed to be worth €14,600, then Tiihonen et al.’s finding reverses, and smoking becomes a net cost for society.

Source: http://www.openpop.org/?p=646

  • "Smokers pay less in income taxes than non-smokers"? How does that work?
    – JAB
    Feb 21, 2017 at 21:09
  • 3
    @JAB, presumably, by having less income.
    – Mark
    Feb 22, 2017 at 2:42
  • 10
    @JAB The key reason is because they die earlier and so don't pay taxes for as many years. There's also the fact that smokers on average have lower incomes, but that is not assumed to be causative.
    – Ergwun
    Feb 22, 2017 at 2:59
  • Maybe you should explain "when a monetary value for life years lost"/QALY. It is not clear at all how they are getting at these numbers, or how cutting off a few retirement years is going to somehow be a drain on society.
    – Jonathon
    Mar 9, 2017 at 14:57

It's complicated

There are many different ways of modelling the cost of tobacco to society, depending on what figures are included or excluded from the total amount.

For example, Devine's described model appears to have a goal of minimising only the per person cost of health care. This is self-evidently a poor choice of metric, as it justifies absurd behaviours - like letting people die quickly - or even encouraging them to die - once they get past child-bearing age.

Some of the more substantive models, such as one I will introduce below, include the costs of losses of trained workforce labour, the costs of hiring people to support the incapacitated, the cost of increased fires and even the cost of the tobacco itself (which is not spent on other industries). They consider the costs to households, businesses and governments.

This is not even include the intangibles, such as Disability-Adjusted-Life-Years (DALY) lost to smokers and the cost to their families of their early deaths.

Also different countries have different pension and health care models, so the research must be used with care.

I have only provided a selection here - the full answer could fill a book.


This study looked at 1976 men from Finland, over a period of years. The smokers has a shorter lifespan by 8.6 years. They found that, yes, this meant that smokers had a lower health care cost, and they also missed out on 7.3 years of pension.

Overall, smokers’ average net contribution to the public finance balance was €133 800 greater per individual compared with non-smokers. However, if each lost quality adjusted life year is considered to be worth €22 200, the net effect is reversed to be €70 200 (€71.600 when adjusted with propensity score) per individual in favour of non-smoking.

Conclusions Smoking was associated with a moderate decrease in healthcare costs, and a marked decrease in pension costs due to increased mortality. However, when a monetary value for life years lost was taken into account, the beneficial net effect of non-smoking to society was about €70 000 per individual.


Miranda Devine is Australian, so the Australian research is relevant.

This entire chapter does a detailed analysis of the question, based on research from 2008 and how it affected the 2004-2005 budget year. It is worth a look.

Looking at just the tangible costs, Table 17.2.3 has a column looking at the cost to the government, which appears to be Devine's concern:

Table 17.2.3

It shows that, indeed, there is a saving of almost $140 million in Nursing homes, but the this is outweighed by hospital costs alone, let alone pharmaceuticals, medical and ambulance costs. In total, smoking adds $249.3 million to the costs to the government alone (and this pales compared to the overall tangibles costs, of over $19 billion.)

If you want to take taxes (both gained on tobacco products, and lost in productivity) into account, check out Tables 17.2.5 and 17.2.6. It is pretty dire.

Intangible costs to society are estimated to be about 150% of the tangible costs (Table 17.2.1), so you can see even this isn't the whole story.


Despite the complexity, Miranda Devine's comment is incorrect:

  • The idea that smoking "isn't that bad" is an opinion not an empirical fact, so it isn't on-topic here.
  • It shortens much more than "just a couple of years" off your life, on average (based on the Finnish data).
  • It costs the Australian health system hundreds of millions of dollars more, in the long run.
  • When you step back, and look at the broader picture than just the government-funded health system, the social cost is much, much worse.
  • I don't want to sound pro-tobacco, but the answer has some commentary that doesn't really add to it.
    – Golden Cuy
    Feb 26, 2017 at 11:59
  • @Andrew : Is it just the 'pretty horrific'? I was trying to counter the 'aren't that bad' claim but you are right - that's just countering opinion with opinion. I shall edit. Was there more?
    – Oddthinking
    Feb 26, 2017 at 13:16
  • The bit about justifying euthanasia.
    – Golden Cuy
    Feb 26, 2017 at 20:49
  • I've removed the 1st issue, now that I am in front of a real computer. I am mulling over the 2nd one. It's certainly colourful, but is it commentary? I am trying to demonstrate that the very limited criterion being used to justify smoking is not one that policy makers/society would solely use, because maximising it leads to absurd conclusions - including that smoking isn't that bad. In that sense, the claim is empty.
    – Oddthinking
    Feb 26, 2017 at 21:48

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .