Sunday, 25 February 2018

The minimum pricing bait-and-switch.

Minimum pricing doesn't start in Scotland until May, but phase two of the campaign has already begun.

From the Sunday Times...

Minimum drink price of 50p ‘just for starters’

The cost of wine, spirits and beer in Scotland could rise by far more than the 50p minimum unit price (MUP) taking effect in May, after politicians and experts warned that the measure will not do enough to safeguard public health.

The temperance beast is never satiated. The battle over minimum pricing isn't about 50p units. It's about whether the price mechanism should be in the grubby hands of the 'public health' racket. Once the government has conceded the point that charging more for a unit of alcohol 'saves lives', there is no obvious reason not to raise it to 'save' more lives.

The junk Sheffield model claims that there would be more health benefits if the price was set at 60p, 70p or 80p. The higher it goes, the greater the supposed benefit.

The idea that a 50p minimum price is 'evidence-based' has always been absurd, and not just because the model itself is worthless. In what sense does the evidence dictate a 50p price rather than a 75p price or a £3 price? It doesn't and it can't. There is no evidence to tell us what the correct price is, but by the logic of 'public health' and the Sheffield model, it should always be higher.

As such, politicians can always be accused of allowing x number of people to die by failing to increase the unit price. Once this Pandora's Box has been opened, it cannot be closed.

The Royal Society of Edinburgh (RSE) has advised Nicola Sturgeon’s administration that the price now could be 64p, based on the consumer price index for alcoholic products in 2011-17. It wants ministers to consider a 60p rate, as proposed by Willie Rennie MSP, the Scottish Lib Dem leader — or even a 70p rate.

The RSE said: “A rise to a rate of 70p would reflect a greater degree of ambition, might also be supported, and would have a larger effect on the consumption of alcohol and on inequalities of outcomes."

Adjusted for inflation, 50p in 2011 is actually closer to 60p than 64p, let alone 70p, but it is true that 50p today is worth less today than it was then. If the intention is to get the putative benefits laid out in the Sheffield report of 2012, a case could be made for setting the price at 60p. But that is not what the public have been sold. The claim that minimum pricing has little or no effect on moderate drinkers and people on low incomes is a lie, but it is a lie based on modelling a 50p unit. At 60p or 70p, the impact is impossible to disguise.

When Sheffield's guns for hire started spinning minimum pricing as a policy that would have a negligible impact on moderate drinkers, they based their assumptions around a 45p unit price. That was in 2014 when nobody was seriously talking about a minimum price below 50p. They focused on the 45p scenario precisely because it was less damaging than 50p.

They have since abandoned the 45p ruse but all their subsequent claims, including the risible lie that minimum pricing 'very specifically affects the alcohol that's only purchased, really, by heavier drinkers', is based on a 50p unit.

Now we see that even 50p may have been unrealistic. The Scottish public has been sold a pig in a poke.

Some politicians and experts want radical action following research indicating that alcohol is 60% more affordable in the UK than it was in 1980, and that it is possible to exceed new lower-risk guidelines for alcohol of 14 units a week for less than £2.50.

Will they be happy when it is possible to exceed the new, fictitious guidelines for less than £7.50? Of course not. There is no theoretical limit to the unit price, but we can be sure that above-inflation rises will be demanded regardless of whether the policy is seen to have worked or not.

In theory, a £1 minimum price would be still more effective. A £3 price would be even better. The only reason nobody is calling for a £3 unit price is that it would raise the cost of living and create all sorts of unintended consequences.

But so will a 50p unit. The only difference is that politicians and medics would be personally affected by a £3 unit whereas they can afford it at 50p.

Thursday, 22 February 2018

Minimum pricing in Wales - the final PR push

I was on BBC Radio Wales this morning talking about minimum pricing (you can listen from 38 minutes in). This was prompted by the thrilling news that the people at Sheffield University who have spent the last ten year pushing minimum pricing are still in favour of it.

I gather that their latest government-commissioned report claims that 66 lives will be extended in Wales (population: 3 million) if drinkers respond to a 50p minimum price in the implausible ways that the Sheffield team assume. This is up from 53 in their previous report for reasons that are almost certainly not worth bothering with. Since there are around 1,500 alcohol-related deaths in Wales each year, it's doubtful whether such a small change can even be measured.

The media haven't shown much interest in the 66 theoretical lives that will be theoretically saved at a cost of millions of pounds to consumers. Instead they have focused on this statistic from the latest Sheffield publication:

75% of all alcohol consumed in Wales is drunk by the 22% of the adult population who are hazardous or harmful drinkers, according to a new report.

Within this – the 3% of the population who are harmful drinkers, account for 27% of all alcohol consumed.

This is just the Pareto Principle. It's no surprise to hear that most alcohol is drunk by people who drink the most. A fifth of Welsh adults don't drink at all so they obviously account for 0% of sales. 58% drink little (less than 14 units per week) and so account for relatively few sales. It says a lot about the decline in drinking in the UK that only 22% of Welsh adults consume more than 14 units a week and only 3 per cent are 'harmful drinkers' (defined as >35 units per week for women and >50 units per week for men.

As I discovered when I gave evidence in Cardiff - and as was confirmed in my interview today - the Welsh government has accepted that minimum pricing won't help the 'harmful drinkers' who have a dependency on alcohol, but they are hopeful of changing the ways of people like me who laugh in the face of the 14 units target.

It looks as if the Welsh government is going to go ahead with the policy because they're already starting to manage expectations. A spokesman says: 'Minimum unit pricing is not intended to work in isolation' and Alcohol Concern Wales says: 'Minimum pricing is one way to solve it but there's no cure.' It's a sure sign that an underwhelming 'public health' policy is on the way when campaigners start playing the 'no silver bullet' card.

Just before my interview on BBC Wales, listeners were treated to an audio clip of Sheffield University bullshit vendor Colin Angus turning the lies up to eleven. He said:

'Minimum unit pricing is a very well targeted policy because it very specifically affects the alcohol that's only purchased, really, by heavier drinkers.'

This is totally untrue. As the IFS and others have shown, 70 per cent of all off-trade sales will be affected and Angus's own report says that 22 per cent of the alcohol bought by moderate drinkers is currently bought at less than 50p.

Moderate drinking here is defined as 14 units a week or less. Under the previous definition of 21 units or less for men, minimum pricing would affect moderate consumers even more. This shows why the neo-temperance lobby were so keen to change the guidelines. It has a material and significant effect on the statistics, inflating the number of hazardous drinkers and reducing adverse effects of regressive policies on 'moderate' drinkers.

As regular readers will recall, the guidelines were changed largely as a result of another implausible model created by... the Sheffield Alcohol Research Group. That model initially supported a threshold of 21 units but the team made unjustifiable changes to their methodology at the eleventh hour in order to support a lowering of the limits after Public Health England told them to.

In any serious field of academia, being exposed as a gun for hire who is prepared to change your model if your funder writes you an additional cheque would be career-ending. It would surely be enough for the government contracts to dry up. In the world of 'public health', however, it is no barrier at all. In fact, it probably helps.

Tuesday, 20 February 2018

Talking Killjoys

I did a podcast with Cameron English recently which has been uploaded here. It's always a pleasure to talk to Cameron. He lives in California, God help him. Mainly we talked about my book Killjoys and the corruption of the 'public health' movement. Have a listen.

By the way, I am aware that the comment facility seems to have disappeared from this blog for the time being. I haven't disabled it, so if you have any idea why it's gone and how I can get it back, let me know via e-mail (scroll to the bottom). Cheers.

Monday, 19 February 2018

Another neo-temperance flop

Remember the great 'public health' initiative of removing high strength beer and cider from shops in Suffolk? These are the 'cheap' drinks that are supposedly targeted by minimum pricing (this is a shameless lie). Minimum pricing is going to effectively remove these drinks from sale - why buy them when you can get the same number of units from vodka for the same price? - but outlets in Ipswich removed them from sale back in 2012.

How did this 'Reduce the Strength' (RtS) scheme go? Well, it must have gone well because the Local Government Association has a briefing document telling councils that they should do likewise:

Suffolk – perhaps more than any other area – has led the way with tackling the consumption of high-strength, cheap alcohol... The campaign has proved incredibly successful. Ninety out of 138 off-licence premises have signed up, including big chains like Tesco, Sainsbury’s and Morrisons.

Note that the scheme is 'incredibly successful' because it had a large take-up, not because it reduced alcohol consumption or alcohol-related problems. It has been claimed that the number of street drinkers declined in Ipswich, but it seems more likely that these people (most, if not all, of whom are homeless) moved away rather than that they suddenly sobered up.

(As a brief digression, what are the chances of Scotland's homeless moving to England en masse in even greater numbers once minimum pricing is in place? Will anybody be studying this as part of the SNP's rinky dink evaluation?)

Since Ipswich councillors were primarily interested in shooing away street drinkers, they may consider the policy to be a success, but from a 'public health' perspective, the aim was to reduce alcohol consumption. Even if reducing overall consumption wasn't the explicit aim, consumption should decline if the number of the street drinkers and heavy drinkers declines because these people consume a non-trivial proportion of the booze sold. Aren't we told that 4 per cent of the population consume a third of all alcohol?

So did that happen? A new study published last week suggests that it did not. The authors got hold of sales data from the East of England Co-op which got rid of all its cheap high strength beer and cider (>7.5% ABV) from its 53 Suffolk branches in September 2012. Its Norfolk and Essex branches followed suit in September 2013.

The authors describe the roll out in Suffolk as Wave 1 (W1) and the roll out in Norfolk and Essex as Wave 2 (W2). Superstrength booze prior to removal made up 6.5% and 3.6% of total alcohol units sold in W1 and W2 respectively.

Let's see what happened when these drinks were removed...

Our analysis indicates that the impact of RtS on units of alcohol sold for beer/lager and cider was not significant in the two waves. More specifically, following RtS implementation, W1 stores experienced a non-significant increase (3.7%, 95% confidence intervals (CI): −11.2 to 21.0, P = 0.647) whereas W2 stores experienced a non-significant decrease (−6.8%, 95% CI: −20.5 to 9.4, P = 0.390).

Awkward.

The researchers missed a trick by not using Norfolk and Essex as the control group when W1 went into effect, but the results are interesting nonetheless. There was no significant change in the number of units sold and so far as the Co-op is concerned, unit sales went up in one area and down in the other.* As the authors say:

...the changes observed in the two waves were not consistent and so the overall findings showed no intervention attributable impact.

Although the results were non-significant in both places, the Suffolk experience is of most interest because it was there that the Reduce the Strength initiative extended to the most shops, and because Suffolk branches of the Co-op were selling more high strength products before the quasi-ban came into effect (6.5% of alcohol unit sales). It should therefore have seen the biggest decline in sales.

As the authors note, the RtS scheme did not make it impossible for people to buy high strength products in Suffolk. Not all retailers signed up. But the fact that alcohol sales rose in the Suffolk branches does not immediately suggest much switching to other shops. It seems that people in the area did not cut down their alcohol consumption, although some of them must have switched to different drinks.

This evidence goes against neo-temperance beliefs. The 'public health' lobby claims that some people drink too much because high strength booze is being sold at pocket money prices. The drinks taken off the shelves in Suffolk offer the cheapest units of alcohol money can buy in Britain. High strength beer and cider were still available in outlets like the Co-op after the RtS scheme began, but they were more expensive varieties, such as craft beer and scrumpy cider. Good, healthy, bourgeois rocket fuel, in other words.

The logic - and computer models - behind minimum pricing suggest that overall alcohol consumption should have declined as a result of high strength drinks becoming much more expensive, but that doesn't seem to have happened. The authors don't even seem to be very surprised by this. They've seen it all before. Under the heading 'What is already known about this topic', the authors write:

Evaluation of the Scottish Alcohol Act 2010 showed that banning alcohol multi-buy promotions did not reduce alcohol purchasing at the household level, and the introduction of the Alcohol Act was not associated with any changes in off-trade beer sales. 

I'd almost forgotten about the ban on multi-buys. Remember when that was an urgent 'public health' policy that was costing lives for every day politicians delayed? All water under the bridge now. It didn't work, but nobody's going to repeal it. Failure goes with the territory.

In 'public health', the only response to failure is to do more of the same but more vigorously. The authors of the study pounce on the fact that the RtS scheme was 'voluntary', as if it had been the retailers who had come up with the policy when it was actually the brainchild of Ipswich Borough Council, the Police, Suffolk County Council and the NHS. They conclude with a call for coercive, but unspecified, policies to do what this microcosm of minimum pricing couldn't...

Our findings suggest that voluntary RtS initiatives, have little or no impact on reducing alcohol availability and purchase amongst a broader population of customers. The research literature suggests that more effective regulatory public health interventions will be required to achieve substantial population health benefits in reducing alcohol consumption and alcohol-related harms.

Next time will be different, eh?


* Statistical significance is used in this instance to see whether the sales outcomes in Co-op shops reflects the change in sales cross the county.  

Friday, 16 February 2018

The sugar conspiracy debunked

There's been a lot of 'Big Whatever is the new Big Tobacco' rhetoric recently. It's all part of following the anti-smoking playbook which, oddly enough, involves accusing the food/alcohol/soda/gambling industry of following the tobacco industry playbook. It's all so meta, but it's an effective public relations exercise because ad hominem attacks work (read this fascinating study for proof).

The list of businesses that are accused of using 'tobacco industry tactics' is almost endless. Just in the last week, we have seen the booze industry, the baby food industry, the agrochemical industry, the food industry and Facebook all accused of using these mysterious tactics.

Most of the time, the tactics amount to no more than manufacturing a product and trying to sell it; in the final analysis, that is what the single-issue fanatics really object to. The 'tactics' can also often involve lobbying, or rather counter-lobbying as these companies are usually defending themselves from the unprovoked aggression. It is somehow seen as shocking that businesses affected by radical government action tries to have their say (see this 'scoop' from Ireland, for example).

If tobacco industry tactics have a unique meaning in the public's mind, they probably involve covering up evidence and sowing doubt. Of that, tobacco companies have undoubtedly been guilty in the past and the idea that other industries are selling us products that they know to be dangerous is a powerful narrative in these paranoid times. This is the fear to which Stanton Glantz appealed in 2016 when he switched his attention from smoking to sugar. And the global media fell for it. Here is the New York Times, for example...

The sugar industry paid scientists in the 1960s to play down the link between sugar and heart disease and promote saturated fat as the culprit instead, newly released historical documents show.

The internal sugar industry documents, recently discovered by a researcher at the University of California, San Francisco, and published Monday in JAMA Internal Medicine, suggest that five decades of research into the role of nutrition and heart disease, including many of today’s dietary recommendations, may have been largely shaped by the sugar industry.

“They were able to derail the discussion about sugar for decades,” said Stanton Glantz, a professor of medicine at U.C.S.F. and an author of the JAMA Internal Medicine paper.

Glantz's claims were entirely based on an obscure evidence review written in the 1960s by three people, one of whom had undisclosed links to the sugar industry. In those days, researchers did not have to list their interests in journals and it was normal for nutritional scientist to work with the food industry (it still is). The review concluded that fat was a bigger risk factor for cardiovascular disease than sugar. This was an orthodox view then and is the orthodox view now. Glantz did not attempt to challenge this conclusion with empirical evidence. In his mind, one of the authors had received money from food companies so he must be lying.

If you have encountered the anti-carb/anti-sugar crowd on social media, you won't be surprised to hear that this narrative has gone down well. They have a penchant for conspiracy theories and ad hominems, and Gary Taubes has embraced it in his recent book The Case Against Sugar. Nina Teicholz, who borrows liberally from Taubes, created a version of history in The Big Fat Surprise in which John Yudkin's anti-sugar message was silenced by scientists and businesses who wanted people to consume carbohydrates.

There are all sorts of variations of this theory, but the basic theme is that good honest scientists always knew that sugar was the devil, but they were thwarted by devious proponents of the saturated fat hypothesis who shouted them down with the help of lavish funding from Big Carb.

I don't take sides in the fat versus sugar debate. For what it's worth, I think the claims of both factions are overblown. But the narrative presented by Glantz is historically illiterate nonsense and a new study published in Science takes it apart:

Building on a newly popular narrative holding that the low-fat campaign of the 1980s was not based on solid science, these allegations have suggested that if not for the machinations of the sugar industry and its cadre of sponsored researchers, the history of U.S. dietary policy might have unfolded very differently. In this article, we argue that the historical evidence does not support these claims.

Although we do not defend the sugar industry and cannot address every aspect of this history, we believe recent high-profile claims come from researchers who have overextended the analogy of the tobacco industry playbook and failed to assess historical actors by the norms and standards of their time.

Our analysis illustrates how conspiratorial narratives in science can distort the past in the service of contemporary causes and obscure genuine uncertainty that surrounds aspects of research, impairing efforts to formulate good evidence-informed policies. In the absence of very strong evidence, there is a serious danger in interpreting the inevitable twists and turns of research and policy as the product of malevolent playbooks and historical derailments. Like scientists, historians must focus on the evidence and follow the data where they lead.

The article is not behind a pay wall and I highly recommend it. Here are a few more choice cuts...

As we have shown, by the 1960s the paradigm that dietary fat was a likely risk factor for heart disease prevailed among a coalition of scientists closely linked with NIH and AHA and was based on extensive research. By contrast, the sugar theory was developed by a small number of researchers, was supported by limited evidence, and was not accepted by key authorities. Normal science is a social project in which a community of scientists develops consensus over theory. Heart disease epidemiology, in adopting a multifactorial model, could plausibly have accommodated sugar if the theory had withstood testing. But Yudkin's claims were seen as weak and antagonistic, and his signature finding could not be replicated. Moreover, sugar did not appear to meaningfully affect serum cholesterol—the only then-accepted lipid pathway to coronary disease.

As we have also shown, the sugar industry approached Hegsted only after learning of the results of his dairy industry-backed study suggesting that fat and not sugar was a factor in heart disease. “There was no, ‘We'll get money from them and make the results come out this way,’” recalled Lown, who worked in the department. “It didn't happen that way,” he said.

.. We do not claim the sugar industry had no influence on nutrition work at Harvard, nor on the field in general. But we believe that there is no good reason to conclude that SRF's sponsorship of a literature review meaningfully shaped the course of dietary science and policy. Moreover, we think it is an error to demonize, almost as a reflex, scientists and their research when there is evidence of private funding.

.. Our history also underscores the fallacy of emphasizing the machinations of one commodity sector when multiple food industries were deploying similar techniques of influence in the battle for market share. It is notable that during the low-fat era of the 1980s, when suspicion fell heavily on the meat and dairy industries, it was argued that, “The ‘fat lobby’ has not only influenced our nation's food and nutrition policies, it has determined those policies” [emphasis original]. Nearly 40 years later, at a moment when some have said “butter is back” and sugar is toxic, “Big Sugar” is the behemoth accorded these dramatic powers. Caught in the cross fire of these “diet wars” have been the reputations of historical nutrition scientists, whose statures have risen or fallen based on the extent of their contribution to current theories.

If the personal attacks on the authors of this study have not already begun, I'm sure it's only a matter of time.


Thursday, 15 February 2018

Childhood obesity figures are worthless

No news story about obesity in Britain is complete without the factoid about one in three children being overweight or obese. This statistic can easily be found on the Office for National Statistics website, but understanding the methodology behind it requires a little more digging.

I have hesitated to write about this for some time because I thought I must be missing something. I couldn't believe that such important statistics could be estimated in such a ridiculous way.

But they are, and the scale of child obesity in this country has been grossly inflated as a result. Read my article at Spectator Health and all will be explained.

There's more to say about this so I'll probably write a follow up piece. If you have questions or comments, leave them under the Spectator article and I'll try to respond.





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.