Smoke gets in your eyes

To those of you prone to apoplexia gravis, a word of caution: this post does not advocate smoking anything, other than possibly herring.  Nor does it represent a defence of tobacco companies or other enterprises dedicated to the challenge of profiting from the sale of highly addictive toxic substances.  It is instead a plea not to abandon reason and the careful use of language when writing about stuff we strongly disapprove of.  Overstating a strong case often hurts it. It is also dishonest.

I am fortunate in that my kids, when they were mere tots, bullied me into giving up smoking.  As soon as I lit up in their vicinity, they would cry out “daddy, you are going to die!”.  Worse than that, they used to rat me out to my wife when I snuck outside for a quick smoke behind the shed.  It was a battle I could not win, so I quit.  Filthy habit.

You must have seen headlines stating something like “Smoking ‘kills five million a year’” (the year in question was 2000).  What does this mean?  Is this a bad thing or a good thing? Does it mean that five million people who died in 2000 would not have died when they did?  That they would not have died ever, if only they had not smoked? That they would have died later (if so, by how many years or months), and that the manner of their dying would have been more comfortable that their smoking-caused deaths?

The headline in question really ought to have read: “Smoking-related illness and disease caused the premature deaths of five million people worldwide in the year 2000.  Average life spans were -reduced by N years. If they had not smoked, the five million would not have died of smoking-related illnesses and diseases – cancer (lung, throat, mouth, larynx oesophagus, lung, kidney, bladder, pancreas, stomach, blood and cervix), diseases of the cardiovascular system (atherosclerosis, stroke, heart disease, aneurysms of the aorta and peripheral vascular disease), diseases of the respiratory system (emphysema, bronchitis and pneumonia), increased health risk and risk of death to the unborn from smoking pregnant women, periodontal disease, brittle bones, cataracts, ulcers.  Instead they would have died, had they not smoked, at some later date, of cardiovascular diseases, infectious and parasitic diseases, ischemic heart disease, assorted cancers, strokes, lower respiratory tract infections, respiratory infections, respiratory diseases, unintentional injuries, HIV/AIDS, chronic obstructive pulmonary disease, perinatal conditions, digestive diseases, diarrheal diseases, intentional injuries (suicide, violence, war, etc), tuberculosis, malaria, road traffic accidents, neuropsychiatric disorders, diseases of the genitor-urinary system, cirrhosis of the liver, nephritis/nephropathy, Alzheimer’s disease and other dementias, musculoskeletal diseases, hepatitis B, Parkinson’s disease, alcohol use, drug use, upper respiratory infections, skin diseases, hepatitis C, Huntington’s disease, multiple sclerosis, motor neurone disease or some other condition.

Many of the smoking-related deaths are horrible.  All are premature.  But the usual announcements of deaths caused by smoking are framed as if the poor unfortunate smokers would, if only they had not succumbed to the evil weed, lived for ever or, at least, would have lived longer, with a good quality of life and would have experienced a ‘good death’.

Living through the deaths of friends, family and loved ones, I am as sure as a person with only one life to live can be, that there are deaths that are even worse than the worst smoking-related deaths.  Having watched non-smokers die of bone marrow cancer, of a long sequence of minor strokes, of the result of a fall causing a major breakage of brittle bones, of a slow brain tumour, of Parkinson’s disease and of senile dementia has cured me of the notion of a gentle death or good death.

So please, when there is a public announcement that X number of people died during 2009 as a result of smoking-related illnesses, please provide the public with two bits of information.  First, how much longer the victims of smoking would have lived if they had not smoked and, second, what they would have been expected to die of instead.

I was prompted to write this post because of a report that a study, published in the journal Tobacco Control, reported that smoking costs the UK National Health Service at least £5.17 billion in 2005/06.  The study also calculated that almost one in five deaths in 2005 were due to smoking. “We estimate that 109,164 deaths (18.6% of all deaths) in the UK in 2005 can be attributed to smoking (27.2% of male deaths and 10.5% of female deaths,”.

The full text of the article (“The burden of smoking related ill health in the UK”, by Steven Allender, Ravikumar Balakrishnan, Peter Scarborough, Premila Webster and Mike Rayner of the University of Oxford, United Kingdom) is unfortunately only accessible by subscription or payment from

The paper is the kind of publication that gives the social sciences a bad name.  In 2006/7, UK government revenues from taxes on tobacco products (excise duties and VAT) were around £10bn.  But let’s leave that aside.  The authors of the study are apparently unaware of the fact that the concept of cost relevant from the perspective of the allocation of scarce resources is opportunity cost, not financial cost or outlays.  Even as regards the financial costs imposed by smoking on the NHS, the paper is a spectacular failure.  It strongly conveys the impression that the NHS is at least £5bn worse off because people smoke. Let me quote from the abstract of the paper:

“Methods: A systematic literature review to identify previous estimates of National Health Service (NHS) costs attributable to smoking was undertaken. Information from the World Health Organization’s Global Burden of Disease Project and routinely collected mortality data were used to calculate mortality due to smoking in the UK. Population attributable fractions for smoking related diseases from the Global Burden of Disease Project were applied to NHS cost data to estimate direct financial costs.”

The paper does not point out the key fact that what the authors calls the direct financial costs to the NHS attributable to smoking exclude the (present discounted value of the) savings in financial costs to the NHS caused by people smoking.  One fact everyone agrees on is that smoking causes people to die younger.  Because people die younger, they will make fewer demands on the services on the NHS.  A big human loss, but a financial gain to the NHS.

Since the adverse health effects of smoking, and its effect on mortality rise with the length of the period during which people smoke, the reduction in conditional life expectancy is more significant for older people.  In other words, heavy smokers are unlikely to live to a ripe old age.  Health care needs are highest in the early years of life, during child-bearing years for women, for the very old and in the very last months of life.

Smoking costs the NHS heavily as regards the cost of treating cardiovascular and respiratory diseases and assorted cancers.  It saves the NHS and other parts of the government budget (the state pension, care for the elderly etc. ) massively by causing people to die younger.  Even if age and financial demands make on the NHS were uncorrelated, this would means financial savings to the NHS.  In addition, it is likely, although hard evidence is hard to come by, that while rising life-expectancy has been accompanied by a rise in the number of years during which people enjoy a reasonably good health, once people live to a very advanced age (90-plus, say), their demands on the health services increase markedly.  Smoking causes fewer people to reach that very advanced age at which health expenditures escalate.

The way to save the NHS and the government budget untold amounts of money would be to impose involuntary euthanasia on all those reaching pensionable age.  No doubt some future paternalistic-libertarian government will do a cost-benefit analysis which demonstrates that a Logan’s Run scenario (in which everyone over the age of 30 is vaporised) is socially optimal.  Smoking is a decentralised, voluntary mechanism for achieving a weaker version of this public finance Nirvana, though voluntary/addiction driven euthanasia of part of the population.

The true opportunity cost of smoking is not measured by the direct financial costs of smoking to the NHS or by the total financial cost of smoking to the NHS, which allows for the financial costs imposed on the NHS by the smokers in the counterfactual scenario where they did not take up smoking.  The total financial cost of smoking to the NHS could well be negative.  The total financial cost to the state budget of smoking is almost certainly negative.

The true opportunity cost of smoking would be the monetary measure of the effect on the present and future well-being of all those affected by smoking (directly or indirectly).  It includes the income and production lost by active and passive smokers as a result of their exposure to the noxious and poisonous weed, but goes beyond that to include the effects on their quality of life.  It also would include the lifetime real resource costs of medical care with smoking and without smoking.

Smoking is a terrible addiction.  Taxation, regulation and education should be used to minimise its incidence. Still, if I ever were diagnosed with Huntington’s disease or some other terrible and terminal affliction, I would light up, and not just to save the NHS some money.

Maverecon: Willem Buiter

Willem Buiter's blog ran until December 2009. This blog is no longer active but it remains open as an archive.

Professor of European Political Economy, London School of Economics and Political Science; former chief economist of the EBRD, former external member of the MPC; adviser to international organisations, governments, central banks and private financial institutions.

Willem Buiter's website