Tuesday, 15 December 2009

Game theory and cigarettes

Watching QI last night, I was reminded of John Nash and game theory. Nash was the troubled mathematical genius portrayed by Russell Crowe in the film A Beautiful Mind. He didn't invent game theory per se, but he did invent an off-shoot of it - the Nash equilibrium - which, to put it simply, is where both players in a game are aware of the other player's ability and motives, and therefore play in a predictable way to maximise their own advantage. 

In game theory, neither player will do better by changing their strategy - even if both would do better by collaborating - and, therefore, neither player will change their strategy even if it means losing out if the other player plays the same way

In popular culture, game theory can be seen in TV shows such as Goldenballs (or its earlier incarnation Shafted), where players can choose to take or share the winnings. If both players share, they share. If they both choose to take, neither gets anything. If one takes and the other shares, the taker gets it all.

The Cold War is the classic example of game theory in action, which brings me to Adam Curtis's fascinating documentary The Trap: What Happened to Our Dreams of Freedom. In it, Curtis describes a game Nash invented called Fuck You Buddy, which could only be won if the player betrayed his partner. The whole documentary is on Youtube. This is the most relevant part...

Nash believed that humans were inherently suspicious of each other, hardly surprising since he was a paranoid schizophrenic. His was a gloomy and pessimistic view of the world, but it did not apply to the mass of humanity. In reality, we are more likely to collaborate than Nash thought. We do not always act in naked self-interest.

So what's this got to do with tobacco? Well, as Stephen Fry pointed out on QI, game theory can be applied to advertising. If one company is advertising, all the rest must do the same. But if no companies advertise, they can save money. And in the case of tobacco advertising, the state intervened to remove the competition. In effect, the state forced the tobacco companies to collaborate. 

As Fry said, it should have been predictable who the real winners would be:

If neither of them advertised, they could keep the money and the market would remain the same. So what it resulted in is the bizarre situation where they banned tobacco advertising, and it was to the benefit of the tobacco firms because they were suddenly saved money which they were otherwise wasting.

It is almost heretical to say this, but the purpose of tobacco advertising is to make people switch brands (or to stick with the brand they are already smoking). Anti-smoking groups have long disputed this, saying that tobacco advertising exists to make people start smoking, but they have never produced any credible evidence to show that this is the case.

It is likely that the anti-smoking groups are wrong. The US has a lower smoking rate than the UK, for example, despite cigarette advertising (in print) being legal in the former but not the latter. Cigarette sales also happened to go up in both countries after cigarette commercials were banned on TV.

In truth, cigarette advertising does little more than give different brands a chance to compete. As I said in Velvet Glove, Iron Fist, the biggest winner from the clamp-down on cigarette advertising in the 1970s was Philip Morris, because they happened to be top dog at a time when competing brands were prevented from gaining market share. It is no coincidence that Marlboro has been the world's top-selling brand since the early 1970s. Other brands simply cannot compete on anything other than price (eg. Lambert & Butler). 

There are cigarette companies that would love to see tobacco advertising reintroduced, but not the big ones. As so often happens, the anti-smoking groups have inadvertently done Big Tobacco's work for them.


Curmudgeon said...

And, as I point out here, exactly the same would happen in the alcohol market if advertising were banned or severely restricted. The market share of the existing big brands would be set in stone, and the opportunities for new entrants choked off.

Anonymous said...

Interesting piece. Yesterday I was reading Armchair Economics by Steven Landsberg, which devotes a section to businesses favouring restrictions on their trading opportunities. Butchers in Chicago once supported a proposal to limit shop opening hours because they wanted more leisure time.

Blueblackjack said...


Never been overly convinced on the 'but advertising is only to compete for a fixed market share, not to attract new customers' argument. It implies the existence of an exclusive selective effect for advertising that is contrary to marketing theory and pretty implausible in effect.

Nintendo don't go mad on the big pre-xmas media ad-rush to exclusively attempt to convert x-box and playstation owners to buy a wii. Of course they try to get Microsoft and Sony's customers and also to expand the market by creating new games console owners (ensuring of course that they enter the market by getting a Nintendo).

Anybody who's had their kid moaning on endlessly about the next new thing they want because of a well-placed targetted TV ad in the middle of an afternoon kid's flick will tell you that.

Also, a brief pubmed search turns up quite few studies on cigarette advertising that are longitudinal in design - evidence that advertising encourages uptake?

Brian, follower of Deornoth said...

Interestingly, the vast levels of taxation on cigarettes means that competing on price is becoming increasingly difficult, thus cementing the market share of the main brands.

Anonymous said...

A Nintendo or a Wii offers an easily identifiable benefit; having fun playing games. Food advertising would undoubtedly be an exclusive matter of getting people to switch between particular brands of food. Food advertising obviously doesn't make people take up eating.

The situation is obviously somewhat different when it comes to smoking. Non-smokers are not particularly interested in tobacco products at all, and probably pay little attention to the advertising. The reasons that non-smokers become smokers seems to have little to do with advertising. Rather, people are more likely to try smoking because people they know smoke, and there's a social benefit because smokers commune over smoking. Non-smokers are largely left out of that, and it doesn't escape their notice that they've chosen not to partake in smoking while smokers seem to enjoy the experience quite a bit.

So, it seems to me that people who take up smoking are likely to smoke the brand that is most readily available to them via other smokers. This is also likely to be the brand most heavily advertised in a particular area. After they've started smoking, smokers will sometimes switch brands in response to advertising.

This would account, in part, for the prevalence of menthol brand use amongst blacks in America. The idea that a person's skin color makes it more likely that they'll like menthol seems implicitly racist to me, as does the call for menthol to be removed from cigarettes by anti-tobacco activists. I would speculate that menthol cigarettes more popular amongst blacks because of a past price preference. In other words, when the tobacco companies tried to market menthols, they were unpopular at large, but the promotional prices were attractive to people living in poorer areas. Then excess supply allowed for the lower prices to continue as the menthols were sold at markdown. Also, the Newport brand used to give away promotional items like T-shirts with the purchase of a carton of cigarettes. Since blacks make disproportionately lower incomes than whites, blacks would be the most likely to buy the cheapest cigarettes. In this case, menthols. And once menthol brands became the most readily available from existing smokers, they also became the first brands tried by new smokers.

So, where would the tobacco companies then most heavily market menthol cigarettes? In poor and black areas of course. So, it would be hard to distinguish between the effect of advertising and the prevalence of brand use, and one would seem to cause the other when no such thing is really happening.

Of course, I have no study to reference to prove this. There might be some internal tobacco company info on it somewhere. Nonetheless, doesn't this make a whole lot more sense than thinking people with darker skin have an unexplainable preference for menthol tastes? WS.

Carl V Phillips said...


The assessment that the cigarette companies benefited from a ban on expensive broadcast advertising is compelling. They also seem to have benefited from warning labels (as you have noted) and probably even the MSA. Of course, any of these might have been good for public health too, so the only people who are really bothered are probably those more interested in hurting the industry than in helping people.

A more specific reference for the game theory phenomenon you cite is to the game called the Prisoners' Dilemma, which there is an extensive literature about. The Nash equilibrium of that game is the one you describe, where both players are worse off than they would be if they could collude/cooperate.

Also, the game theory of the cold war is best attributed to Thomas Schelling and the inventor of modern game theory was John von Neumann.

As for the hope that some had that reducing advertising would reduce sales, I think the best explanation is the fallacy of composition that Krugman blogged about just today (http://krugman.blogs.nytimes.com/2009/12/16/would-cutting-the-minimum-wage-raise-employment/?8ty&emc=ty): It is definitely true that if one company stops advertising it will lose sales. What is not necessarily true is that if all companies stop advertising that they will lose sales in total.

--Carl V Phillips

Anonymous said...

I'm not able to watch the episode of QI here in the US. However, there is a good example of game theory and the Nash Equilibrium in Richard Dawkins' documentary "Nice Guys Finish First" at 10:30 at the link below.WS

Anonymous said...

Here's the link. WS


Mark Wadsworth said...

Excellent post, saved me the trouble of writing it.

Funnily enough, a few years ago, spirits manufacturers declared a similar truce, saying that they would no longer advertise on TV, they appear to have abandoned that recently, as there are loads of adverts for spirits on TV again.