Monday 12 February 2018

John Stuart Mill spins in his grave

Robert H. Frank offered an unusual justification for clamping down on smokers in the New York Times last month. As it touches on some of the themes in Killjoys I want to discuss it.

One the main points in Killjoys is that paternalists are always looking for ways to disguise their paternalism. Even in these puritanical times, it is still not socially acceptable to ask the government to interfere in someone's private life just because you don't like what they're doing.

In the field of tobacco control, three justifications have been put forward which broadly fall under the category of 'protecting harm to others'.

The first is that anti-smoking policies prevent the tobacco industry harming smokers (by selling them cigarettes).

The second is that smokers put a cost on nonsmokers by developing expensive diseases.

The third is that smoking bans are necessary because secondhand smoke harms nonsmokers.

The tobacco control lobby has got a lot of mileage out of these arguments over the years despite none of them being very strong. The idea that anti-smokers are protecting people from industry ignores the fact that industry has no means of coercion. The economic literature shows that smokers do not have higher healthcare costs over a lifetime. And the epidemiological evidence on secondhand smoke is all over the place.

Crucially, though, the power of these arguments wanes as more and more anti-smoking policies are implemented. Campaigners can present an advertising ban, for example, as an 'anti-industry' measure, but it difficult to pretend that a ban on people smoking in their own home, or mandatory nicotine testing of employees, is an attack on the tobacco industry.

As tobacco taxes rise, the idea that smokers are not 'paying their way' becomes increasingly difficult to maintain. And claims about secondhand smoke lose their political significance once smoking has been banned in all enclosed public places.

Frank's article is an attempt to find a new justification for coercive anti-smoking policies in an era where the old excuses are losing their power. He starts by acknowledging that smoking in New York has been banned virtually everywhere, including in some outdoor spaces, and says...

Given the longstanding American hostility to social engineering, each of these steps faced heavy pushback. When called on to justify them, regulators have offered their traditional response: Restricting individual freedom is often the only way to prevent undue harm to innocent bystanders.

The specific harm cited has almost always been well-documented health hazards caused by secondhand smoke. This rationale is similar to the one for requiring catalytic converters on cars: We need them to prevent pollution that would otherwise cause undue harm to others.
But unless you work in a crowded bar with no ventilation, the health risks from secondhand smoke are small compared with those from being a smoker. For example, more than 85 percent of American deaths from lung cancer are attributable to smoking...

To digress briefly, this widely cited statistic is wrong. It may be true that 85 per cent of people with lung cancer are smokers, but that does not mean that 85 per cent of cases are caused by smoking. Half the population are smokers by the epidemiological definition (ie. have smoked more than 100 cigarettes in their life). Since non-smokers also get lung cancer, albeit in much smaller numbers, the proportion of lung cancer cases that are caused by smoking is more like 70 per cent. I can explain this in more detail in a future post if the logic is unclear.

...with fewer than one-third of the remainder linked to passive smoke exposure. Regulators may insist that their aim is not to protect smokers from themselves, but our regulations do vastly more to protect smokers (by inducing them to quit) than to protect bystanders.

This, of course, is the unspoken reason why smoking bans are introduced. They are not about barworkers or passive smoking, they are about making it difficult for smokers to smoke.

But they have run their course in places such as New York, and so a new justification is required...

In fact, smoking also harms bystanders in a more important way: Each person who becomes a smoker makes it more likely that others will become smokers as well. This additional effect outweighs the harm caused by secondhand smoke by enough to suggest that our efforts to discourage smoking, strict as they seem, may not be nearly strict enough.

He is arguing that smoking is, in effect, contagious.

By far the most powerful predictor of whether a person will smoke is the percentage of her closest friends who smoke. If the share of smokers in someone’s peer group rises to 30 percent from 20 percent, for example, the probability that she will smoke rises by about 25 percent. Whereas most of my teenage friends were smokers, relatively few of my sons’ friends were. In 2016, only about 19 percent of American men were smokers, and only about 14 percent of women.

This is a rather naive interpretation of the data. People will tend to associate with others who have similar interests and backgrounds. Drinkers will tend to spend time with drinkers, vegans will spend time with vegans, heavy metal fans will spend time with heavy metal fans. This is not contagion.

On the other hand, people do have an influence over their peers. If you were born in a different country, you would probably have different tastes in sport, music and food.

But you would still have free will - and this is the aspect that Frank overlooks. He gives an example of two peer groups, one with a smoking rate of 30%, the other with a smoking rate of 20%. Whatever the 'probability' of someone smoking in each group (which is somewhat tautologous anyway), the majority are still nonsmokers. Influence is not coercion.

Today’s environment is different mostly because of the taxes and other regulatory measures we have taken to discourage smoking. Well and good, but does anyone think that still having more than one smoker in six people is a desirable population ratio?

What is the desirable ratio? To Frank, it appears to be zero, but that is clearly not the desirable ratio for people who enjoy smoking. So who gets to decide? Someone like Frank, who doesn't smoke and who doesn't think other people should either, or the people who like smoking?

Why should anyone decide what percentage of the adult population smokes? Surely the 'desirable population ratio' is that everybody who likes smoking smokes and everybody who doesn't like it doesn't smoke. Whatever happened to 'life, liberty and the pursuit of happiness'?

Our stated rationale for discouraging smoking — to prevent harm caused by secondhand smoke — greatly understates the amount of harm that these actions prevent. When a regulation results in one smoker fewer, every friend of that person will have one smoker fewer in her peer group. Every member of every one of those peer groups will then become less likely to smoke. And that, in turn, will make others less likely to smoke, and so on.

Frank doesn't make any specific policy recommendations, but implications of his argument are sinister, bordering on totalitarian. It suggests that individuals should be forced to sacrifice their pleasures in case they unwittingly influence their friends into adopting the same way of life.

Most people don’t like being regulated, but even strict libertarians concede the legitimacy of regulations to prevent undue harm to others. As John Stuart Mill memorably wrote in “On Liberty”: “The only purpose for which power can be rightfully exercised over any member of a civilized community, against his will, is to prevent harm to others. His own good, either physical or moral, is not sufficient warrant.”
Causing someone to be more likely to smoke clearly inflicts substantial harm on that person. 

Not 'forcing' someone to do something. Not even 'causing' them to do something. But 'causing them to be more likely' to do something!

This is one of the grossest distortions of John Stuart Mill's harm principle that I have ever come across. It is inconceivable that Mill would have supported coercive legislation to force individuals to set a good example to those around them. Indeed, he explicitly rejected the idea and was clearly exasperated when the nineteenth century temperance movement made a similar argument about alcohol. In On Liberty, he described it as a 'monstrous' idea that would make 'all mankind a vested interest in each other’s moral, intellectual, and even physical perfection, to be defined by each claimant according to his own standard'. In practice, he wrote, it would mean ‘no right to any freedom whatever, except perhaps to that of holding opinions in secret’.

Today’s regulations to discourage smoking are strict, yes. But without violating libertarian sensibilities, we could adopt even stricter measures.

This is laughably untrue. It is hard to think of anything more likely to violate libertarian sensibilities than the idea that individuals must be coerced into changing their behaviour so that they can be a walking advertisement for a government-approved lifestyle.

Incidentally, I have encountered Frank's strange theories before in a different context. You can read about one of them from page 110 in my book Selfishness, Greed and Capitalism.


2 comments:

Vova said...

The same idea is influential in 'alcohol studies' - all kicked off by
http://onlinelibrary.wiley.com/doi/10.1111/j.1360-0443.1985.tb05294.x/abstract

oldngrumpy said...

Has anyone come across these articles.

https://www.medscape.com/viewarticle/850708

https://www.medpagetoday.com/meetingcoverage/iaslc/53476

It appears the incidence of Lung Cancer in NON smokers is on the rise. According to a study by Dr Lim LC incidence among NON smokers has been “increasing steadily” since 2008 (13% in 2008; 28% in 2014). This is mainly caused an increase in absolute numbers and not relative changes due to fewer smokers.

These results are not only puzzling but slightly bizarre given that the UK Smoking ban came into force in 2007. I’ve always been sceptical of any passive smoking/LC link but expected a smoking ban to produce a broadly neutral effect. I’m not suggesting that the smoking ban is a cause of the increase but it does suggest health agencies have been spending far too long pursuing non-existent risks.