Scott Sumner  

What would a scientific cigarette policy look like?

Oxfam or Oxgov?... We're Number 17...

Early studies of the impact of second hand smoke often found a statistically significant impact on heart attack rates. Or perhaps I should say, "Early published studies"; I have no idea what early unpublished studies found.

When faced with this sort of evidence, scientists will try to replicate the results with much larger data sets, accounting for omitted variables. And that's exactly what occurred. Here's

Contradictory research continued to come in. A clever study led by researchers at RAND Corp. in 2010 tested the possibility that the large reductions identified in small communities were due to chance. They assembled a massive data set that allowed them to essentially replicate studies like those in Helena, Pueblo, and Bowling Green, but on an unprecedented scale. Whereas those studies had compared just one small community to another, the RAND paper compared all possible pairings of communities affected by smoking bans to all possible controls, for a total of more than 15,000 pairings. They stratified results by age in case there were differential effects on the young, working age adults, or the elderly. And in an improvement on most other studies, they also controlled for existing trends in the rate of heart attacks.

The study found no statistically significant decrease in heart attacks among any age group. The data also suggested that fluctuations in heart attack rates were common, indicating that comparisons of small communities would frequently turn up dramatic reductions due purely to chance; large increases in heart attacks happened about as often. This explained the headline-grabbing dramatic results in places like Helena or Monroe County that eluded replication in larger jurisdictions. The conclusion of the study was blunt: "We find no evidence that legislated U.S. smoking bans were associated with short-term reductions in hospital admissions for acute myocardial infarction or other diseases in the elderly, children or working age adults."

I was directed to this article by Steve Winkler, who made this comment:

Slate explores the implications of the junk science used to ban smoking on grounds of secondhand dangers. I believe we are in an age of rising puritanism. Tobacco is the drug in the cross hairs. It is low brow. Interestingly alcohol and marijuana are higher and rising status. Once again, mood affiliation and out-group shaming guides public policy.

Let's assume that despite the RAND study, we conclude that second hand smoke has some negative health effects (which seems plausible in my view.) In that case, what would be the appropriate public policy toward second hand smoke?

It turns out that it entirely depends on the location of the second hand smoke. According to the Coase Theorem, externalities do not call for government regulation unless it is too costly to privately negotiate an efficient solution.

In the case of indoor smoke, it is almost never costly to negotiate the optimal solution. That's because in most cases the optimal policy toward indoor smoking will be the policy that maximizes the value of the property. Thus a restaurant owner will have an incentive to set a smoking policy that maximizes the value of her business. Ditto for the owners of office buildings, apartments and airplanes.

Actual public policies toward second hand smoke are almost nothing like what the science would suggestion. Conservatives are often criticized for rejecting the scientific consensus on climate change. I believe the criticism is justified. But those conservatives are only rejecting a single scientific hypothesis. Second hand smoking regulations are far harder to justify, as they represent a rejection of two very different types of science:

1. The science of how to establish statistical significance when there is publication bias in favor of rejecting the null hypothesis.

2. The science of the Coase Theorem, and particularly its implications for public policy.

Many people claim that some conservatives reject the science of global warming because they are not comfortable with the policy responses proposed by people on the other side. I prefer not to attack motives, but if that is your view, shouldn't you also be asking how many progressives reject the science of second hand smoke, and also the science of when to use government regulations, solely because they don't like the policy implications of those two types of science?

PS. I try not to let my political views influence my views on scientific questions. I was serious when I said that I believe second hand smoke does have some negative health effects. As an anecdote, my grandfather died of emphysema. He was a nonsmoker who worked in an office filled with smokers. It seems plausible to me that all of that second hand smoke was a problem for a guy who (like me) had bad lungs. Yet I oppose government regulation of smoking for Coase Theorem reasons. On the other hand, I support a carbon tax, as the transactions costs of a Coasian negotiated solution are prohibitive.

Screen Shot 2018-02-04 at 5.40.58 PM.png

Comments and Sharing

COMMENTS (27 to date)
John Hall writes:

At my company and I'm sure a lot of other ones, not only can you not smoke in the building, but you pretty much can't even smoke on the property (I think it's like 15 feet away or something, but that's basically in the street). If second hand smoke is the reason for this restriction, then shouldn't you be allowed to smoke outside?

Thomas Sewell writes:

Even more ridiculous are the bans on second-hand vapor. It's almost like it's all based on mood affiliation, entrenched bureaucratic and industry interests and a goal of signaling a politician's desired public stance.

Pierre Lemieux writes:

As @Thomas suggests (consistent with @Scott's interesting post), the vapor case suggests (strongly) that the prohibitionists are motivated not to prevent other people's diseases, but to prevent other people's liberty. My own post of today on coffee illustrates how voters (prodded by activists) have cooperated with these trends.

Scott Sumner writes:

John and Thomas, I agree.

Scott Sumner writes:

Pierre, Good point.

Mark Bahner writes:
Let's assume that despite the RAND study, we conclude that second hand smoke has some negative health effects (which seems plausible in my view.)

I can pretty much guarantee you that heavy second-hand smoke gives me *tremendous* sinus headaches. This is from many, many experiences of the effect in bowling alleys and pool halls.

The final crowning example was when a friend and I played pool in a smoky pool hall one night, then went to the Disney Epcot Center the following day. It was a very hot and humid day, and of course the pavilions were well air-conditioned. I would go into the pavilions, and get a tremendous headache, which would eventually go away. Then I'd go back outdoors, and into another pavilion, with the tremendous headaches returning.

Finally, we got to that "Incredible Journey" ride (it looks like it's no longer there). This ride was based on the movie with the shrunken submarine inside a human body. Anyway, the ride threw people back and forth, and I remember my friend laughing, but I was in absolute agony. I came out of the ride and I was drenched in sweat from the pain. My shirt was literally like I'd been in a pool, it was so wet.

The ending to the story is that in the rental car exiting Disney World for the airport, I told my friend, "Stop the car!"...and I got out of the car and violently threw up. (Interestingly, like 2-3 Disney security cars came essentially even before I even stopped throwing up.) Then after throwing up and blowing my nose thoroughly, I felt much, much better.

P.S. If I'd known about saline nasal rinses back when I had the problems, saline nasal rinses probably would have helped.

Adam writes:

Interesting points. Smoking bans seem a socially lost cause. Who has political influence? Certainly not the smoking minority. Better to apply the median voter model.

Acceptance of climate change hypothesis and choice of carbon tax policy are 2 very different items.

Policy choice is informed by incidence and opportunity costs. In my view, the global opportunity costs don't warrant a focus on CO2 in high income countries. There are just too many people killed daily by water pollution, fecal material and local air pollution. These are grave, severe and immediate problems.

Re air pollution, consider this map by Berkeley Earth:

A writes:

It's interesting to see UCSF consistently jumping the gun on weak research. They also pushed the FDA to label e-cig products as tobacco products, regardless of actual tobacco content. There might an institutional angle, wherein policy wonks who benefited from prior beliefs, must continuously refresh their authority despite changing information, or outdated knowledge in the case of e-cigs.

Hans writes:

The second hand smoke campaign (funded by tobacco consumers) is a big lie, for the purpose to regulate and eventually ban the use.

Utah, has now passed a two drinks and you are drunk bill (patrol unit will refer to you as DK), somethings which opponents said would happen. This is a case where "Conservatives" go bad. M.A.D. is happy and happy is mad.

Tom West writes:

To be honest, I find the smoking ban to a great example of status quo bias. In my local areas, restaurants were almost universally against the ban when it went into effect locally.

Then almost none intended to switch back when it looked like the ban might be repealed (it wasn't).

Not being a Libertarian, if regulations get us to a "better" equilibrium, I'm okay with that, although I do consider regulation to have a cost in and of itself.

For example, I'd probably be okay with a drop of the smoking ban now, as it might allow for some slight accommodation of smokers in the odd restaurant, but 95% would remain smoke free - a state that could never have occurred without the smoking ban in the first place.

Weir writes:

Policy first, hypothesis second. That's always the rule. Think of Google. The policy has to do with professional equality between men and women. Working backwards from that policy, the hypothesis was made to measure. Thus, millions of years of evolution have produced two sexes that are identical in every way, according to Google.

The scientific consensus within Google is that nothing happened over all that time to result in any kind of differentiation between men and women. Men and women are legally equal, and therefore they must be biologically and psychologically and chromosomally identical too. Both men and women can have two X chromosomes. Both men and women are equally likely to become serial killers. It is only discrimination that has held women back from committing an equal number of serial murders, according to Google.

I like the Google example because the consensus is enforced with such explicit, out-and-out bullying. The bullying at Google goes beyond the predictable Brahmin disdain for dalits and untouchables. That's the normal way of enforcing consensus.

At Google, however, it's explicit harassment. You could try to point out, with citations and graphs and reams of empirical data, exactly how the scientific evidence contradicts the strictly enforced scientific consensus, but you will not get a rational argument in response. You just get yelled at, and blacklisted, and sacked. This is the response from people who may have even trained as scientists, behaving like politicians.

The scientific method couldn't be more different from politics. What is politics except bullying, shaming, and ostracism? The politician stirs up hatred and contempt for the out-group, the inferior, the other. And the scientist at Google, it turns out, becomes the perfect politician in pursuit of his policies.

Michael Sandifer writes:

I fail to see the problem, on net. Smoking rates are way down, which seems to be an undeniably good development. How much taxes and regulation had to do with it is another question, and one I haven't examined.

So, a small percentage of the population is now inconvenienced for the pleasure of the overwhelming majority, on an issue that is close to last on my list in terms of importance right now. To say the least, we have a large and growing list of growing problems at the moment, including general chaos and malevolence at the highest levels of government. Democratic republicanism itself is under assault, so I really don't care about the plight of smokers very much, acknowledging some regulations have gone too far.

Hans writes:

This website is heavily _ensored.

Chris H writes:

I'm a bit curious what happens when there are asymmetric transaction costs? Consider this hypothesis: transactions costs rise with the psychological trait agreeableness (I'm referencing the big five personality traits here). This is because agreeableness is driven largely by a desire to avoid conflict and any negotiation has at least the potential for interpersonal conflict. Agreeable people are also less likely to engage in activities that are known to cause externalities to others, since again that can cause conflict with others.

So with smoking, the non-smokers would be disproportionately agreeable and are still harmed by the smoking action, but have higher preferences to avoid the complaints which could lead to a Coasian solution. The smokers would be perfectly willing to negotiate due to lower average agreeableness (since externality causing behavior is associated with low agreeableness) and therefore willingess to court potential conflict, but don't realize negotiation is even needed. Thus without third party enforcement, the non-smokers wind up with more smoke around them, simply because their personality traits disproportionately impact their transaction costs.

I haven't worked out exactly how this would work or whether the hypothesis on agreeableness is even correct, but Scott do you have any thoughts on the potentially asymmetric transaction costs and whether any research has been done on that?

The Original CC writes:

Good post, Scott re. the rejection of two different types of science.

Tom West raises an interesting question: If the smoking ban were lifted in NYC (for example), would restaurants really go back to allowing smoking? Intuitively, it seems not.

And if not, doesn't that suggest that the restaurants weren't profit-maximizing previously (back when they allowed smoking)? Did bureaucrats accidentally second-guess business owners and actually get it right?

gabe writes:

Just one little problem with the statement by Mr Sumner:
"Early studies of the impact of second hand smoke often found a statistically significant impact on heart attack rates."

"In the majority of analyses, an alpha of 0.05 is used as the cutoff for significance. If the p-value is less than 0.05, we reject the null hypothesis that there's no difference between the means and conclude that a significant difference does exist. If the p-value is larger than 0.05, we cannot conclude that a significant difference exists.

That's pretty straightforward, right? Below 0.05, significant. Over 0.05, not significant. "

DID YOU KNOW that in 22 of the 24 studies ballyhooed in the 1980's on 2nd hand smoke, the P-value was changed from 0.05 to 0.20 in order to force statistical significance.

Recalling from memory (faulty at best, I admit). one claim of these *studies* was that those exposed to second hand smoke suffered higher risk of health problems (strokes, heart attacks, etc) than *other segments* of the population.

One had to read into the details to find that the *other segments* were SMOKERS.
This is like saying that the fellow on the street corner observing a blazing conflagration in an apartment house is liable to suffer greater injury than the poor fellow trapped in the inferno.

Junk science on a par with AGW *science*!!!

Scott Sumner writes:

Tom, I see no evidence that smoking bans get us to a "better" equilibrium. Do you have any?

Michael, You said:

"I fail to see the problem, on net."

Perhaps you are a non-smoker? I know smokers who "see the problem"

The whole point of this post is that there are different groups with different interests, and that a property owner has a much better ability to come up with the optimal policy than a one-size-fits-all law. I don't see where you've addressed that point. Don't start your analysis with your preferred solution, and then look for a justification.

Chris, You said:

"So with smoking, the non-smokers would be disproportionately agreeable"

Don't take this the wrong way, but I almost started laughing when I read this. But maybe that's just my life experience. :)

Original, The optimal policy changes over time. Today there are far fewer smokers, so more restaurants would choose to go that way.

Mike Sandifer writes:

Yes, I knew smokers too, and most of them are now ex-smokers and the cost of cigarettes, along with more effective addiction treatment options, are reasons they state for quitting, along with perceived health benefits and social convenience.

But the point on the coase theorem is trivial, as it's an econ 101, which is fine. I'm not being critical so much as basically saying this is a very econ professor-type post, which makes sense on an econ blog written by professors. It's just those of us who read these blogs have to read such posts constantly. Every economics blog writer does them, instead of saying something unique about economics.

Don't get me wrong. I've enjoyed your writing for years, and you do have unique insights. Many of your posts on monetary policy, EMH/bubbles, economic development, etc. are must reads.

If you have a McCartney song catalogue with songs like Blackbird and The Long and Winding Road, I just think it's wasteful to then write Let 'Em In.

I don't doubt you're right on the theory, but there is more than just economics involved too.

Gordon writes:

Scott, now that you're a California resident, have you seen the TV commercials from Tobacco Free CA? They have these ridiculous scare ads about second hand smoke with the cigarette smoke traveling great distances without any significant dissipation to the lungs of children as if they're some sort of malevolent Star Trek cloud creature.

Tom West writes:

Tom, I see no evidence that smoking bans get us to a "better" equilibrium. Do you have any?

Only anecdotally. The people I knew (including probably 20% smokers) were probably split 60-40 pro-ban. About 2-3 years after the ban when I asked, they were 95-5 pro-ban with about 15% smokers).

Not scientific (coworkers, et al), and a sample size of about 20-30, but given they'd lived under both regimes, it was enough evidence for me to call it "better" for a blog comment.

Mike Sandifer writes:


Yes, you help illustrate my point. There's more than just economics here. We can tax merely to cover the externalities, or try to tax to a level just short of encouraging a lot of black market selling. There's nothing wrong with the goal of encouraging lower smoking rates. A libertarian might object, but I don't think a utilitarian would.

Mr. Econotaraian writes:

I agree that indoor smoking bans in public accommodations were a good equilibrium shift, and it is not going back.

I’m reminded of the “Mad Men” episode where a black secretary is moved to become the front desk receptionist, and one of the lead partners (who is not himself racist) points out that their racist customers will never stand for it. Today keeping someone out of a job because of their race would be clearly illegal, you can see how this kind of ban enables the company to say “we have no choice, my racist customer”. This breaking the equilibrium with a little bit of restriction of freedom.

I know many people with great negative airway reaction to smoke who had problems at bars and live music venues and we’re unable to go to them. Also the additional clothes cleaning by non-smokers to get the stink out was a serious cost to pay.

There are plenty of other ways to use nicotine - gum, transdermal patch, injection. All of these are without externalities for others.

But yeah, public smoking probably never killed anyone from second-hand smoke. It is just annoying.

The few exceptions I can accept from libertarianism is where the government can help to synchronize mass changes to preferable equilibriums. Eminent domain (for actually beneficial projects) is one of those. But I admit it is a very slippery slope, and equilibrium changing should be rarely done by govrrnement.

Floccina writes:

Do you have any thoughts on how the federal Government could assess a CO2 tax on imports?

I support a CO2 tax and would assume that you would estimate the amount of CO2 released in producing various imports. That would be difficult bit I assume could be done acceptably.

Pyrmonter writes:

Slightly off topic, but is that really the implication of the Coase Theorem? I thought it was that, given that transaction costs are prevalent, the allocation of initial conditions matters.

If the Coase Theorem as commonly understood (low or negligble transaction costs) 'held', we'd be indifferent to the allocation of a right to smoke; or a right to be smoke-free.

mariorossi writes:

Is heart diaseases the main health complication of second hand smoking? Is there any reason to believe that no change in heart disease would replicate in other diseases?

Is there any reason to believe that we could know that before the ban (and correlated reduction in tobacco consumption) gave us the information to conduct this study?

I really don't see how you can state that the private market could find an efficient solution to this problem. There are still small diffuse costs vs large concentrated ones. Private owners still need to aggregate preferences over a fairly large changing crowd. Just because it's not a government, it doesn't make it easy.

Plus the entire thing is path-dependent. It's far from obvious that a lift of the ban would return things to the status quo. Prohibition created speak-easy: I see no equivalent now. That information seems relevant as well...

Thaomas writes:

A limitation on "Coasian" regulation of indoor smoking is that people may not accurately value the harm caused by the smoke and so my not make it profitable for owners of indoor spaces to choose the optimal level of smoking permitted.

Peter writes:

Interesting article:

Robert H Frank: Why Even Tougher Regulations on Smoking Are Justified
Smoking harms nonsmokers, and the damage isn’t merely caused by secondhand smoke.

See also (agreeing with the claim that the risks of second-hand cigarette smoke have been greatly overstated):
Geoffrey C. Kabat: Hyping Health Risks: Environmental Hazards in Daily Life and the Science of Epidemiology. Columbia UP, 2008
Ch.6: The risks of second-hand cigarette smoke

Comments for this entry have been closed
Return to top