In this paper I wrote with David Prentice we say yes though it is difficult to say by how much because there have been no previous trials of this policy device. The policy is however essentially a tightening of opportunities to market cigarettes. We therefore use the industrial organisation literature and evidence from two previous phases of advertising restrictions – on, in turn, advertising in electronic and in the print media – to deduce that (i) cigarette prices will plausibly fall and (ii) incentives for entry of new legal suppliers will weaken because of reduced profitability. Price falls can and should be offset by increases in the tobacco excise so that these mechanisms suggest that, there is no potential for the plain packaging reforms to increase cigarette consumption. At worst the effect in reducing smoking will be small and this seems unlikely given the key role that branding occupies in the marketing efforts of tobacco producers. Likewise the incentives offered to illegal producers will diminish and the possibility of an expansion of illegal sales seems to have been overstated given worsening general industry returns. Issues of counterfeiting illegal brands also seem to have been overstated since plain packaging can be drab and unattractive but still be complicated to reproduce.
Australia’s plain packaging package seems at worse to be a costless experiment. Indeed it can be viewed as providing an international public good at almost zero public cost. The policy insight obtained from this experiment about the extent to which plain packaging is effective is highly useful information. (1054)
1) Are there opinions on the effects of this on the real crucial market segment, say 11-18-year-olds?
2) When reading Golden Holocaust, (did your copy come yet?)I found that cigarette filters have included amazing substances.
I found that from 1952-156, Kents (marketed as milder/less dangerous, etc) had “micronite” filters, which had asbestos.
I’ve read it but knew much of the discussion. I think “Cigarette Century:” is the classic but Golden Holocaust is good.
I won’t worry about arguing which of the two books is #1, both authors praise each others’ work, good enough for me.
I figure they overlap a lot, but get more detailed in different areas, and it is nice that there are 2 really good books on the topic with different views and emphases.
Doesn’t all this suggest that running randomized control trials on tobacco interventions make an awful lot more sense than just implementing things like plain packaging across the country? You can’t really do it with excise as that very plausibly leads to cross-treatment arbitrage opportunities. But if Oz and NZ are both going for plain packaging, why not set treatment and control communities to see whether the policy is as good as you expect? What’s the worst that happens, if you’re right – a couple years’ delay in implementing plain packaging in the control communities, but a better evidence base for promoting the policy elsewhere. And there’s always the chance that plain packaging doesn’t really buy the country anything except for lawsuits – it just might be worth knowing whether the game’s worth the candle.
Why delay Eric? It costs almost nothing to do it and the effects are at worst weak – given the importance of branding the effects are likely to be significant.
I think there companies will lose the lawsuits. Trademarks are not intended to defend those who sell poisons in attractive packages.
There is a lot of tobacco company research on the effects of packaging now freely available on the web.
A random trial is an interesting idea but have thought a little bit about the feasibility of a random trial. The treatment groups would have to be reasonably isolated. If for example, it was done in an urban area it would be easy enough for people to pick up packets with packaging – am not sure if we would get the full pricing effects either for similar reasons. There may also be political issues if, for example, all of Tasmania (or the South Island) were required to have plain packaging and not other areas. Would be interested to know about random trials tried with broadly similar types of policies.
@hc: Don’t we want to know whether any changes in consumption are due to plain packaging or due to time effects?
@davidp: You’d want to randomly assign cities into treatment and control areas. I can easily imagine that differences in prices lead to arbitrage across cities; I have a harder time believing anybody would mule branded packs from Melbourne to Sydney (or whatever). There would almost certainly have to be compensation to the tobacco companies for increased distribution costs with two packaging lines; there’d also have to be some consideration given to existing distribution networks when setting treatment and control cities.
Eric,
I can see how assigning the cities would work and agree that the problems with spillovers would be much less than within urban areas. Rural cities might be even better. It would be interesting to see the results. The main issue is probably going to be political. My guess is that this would probably be opposed by the tobacco companies as well. I believe there were different point of sale restrictions at different times in different states – if it were possible to get transaction prices in each state before and after this would be pretty interesting to see if there were differences in prices there as a result. We might also expect to see more effects on brands sold at supermarkets than tobacconists – this would be feasible to do.
If Australia and NZ are going to do something that hasn’t been tried anywhere else, it seems almost criminal not to set it up as a proper policy trial so we can tell what it actually did and let other folks know.
Eric “almost criminal”? It is a low cost policy that might save some lives and, indeed, might save many. It will certainly have no harmful effects given our analysis. What is the downside – that it might only avert a few hundred deaths rather than several thousand?
I’d describe as “almost criminal” the marketing of a product that its producers have known – on the basis of their own research over 50 years – kills people when consumed as intended.
I know you’ve got strong views about the evils of tobacco, Harry. But surely you’d prefer actually knowing the effects of the policy to not knowing, right? If you’re right, and it has at least some positive effect in reducing smoking, you’ll be a better example unto others if you can point to something like “Hey, we did this randomized control trial and towns and cities getting plain packaging got a blah% bigger drop in smoking than other places, and it was significant…” No?
Eric,
No? Even if the policy has a small beneficial effect that is still useful because it is costless. If it has a strong disincentive effect on smoking that is a bonus.
Its not my strong views on tobacco that describe it as a poison being knowingly fostered on youth for over 50 years. Its a fact – see Golden Holocaust as cited above by John Mashey or Allan Brandt’s wonderful ‘The Cigarette Century’. The carcinogenic character and addictive characteristics of cigarettes were known to the companies from the late 1950s on the basis of their own internal experiments – e.g. painting tar on backs of mice. They lied at major public hearings up until the 1990s about both of these issues. Of course I feel revulsion at the activities of such firms but the revulsion reflects their history – it does not create it.
You mention the “evils of tobacco” as if you are caricaturing the views of a fundamentalist evangelical. That is inaccurate. Cigarettes led to the premature deaths of 100 million people during the twentieth century. I value life and don’t see anything positive in this experience at all. The libertarian line that it is all an issue of “freedom of choice” is likewise a campaign launched and maintained by the tobacco companies – this campaign is likewise comprehensively documented. Australian kids start smoking on average at 13.9 years of age and we know they systematically underestimate the difficulty of quitting. Agents who exercise rational choices?
But it’s not costless! It takes political capital to push through a policy like plain packagaing. If it works, maybe it’s worth it, or maybe other policies that are as politically costly are even more effective. If it doesn’t work, anti-tobacco people in other countries would do better to use that capital for other policies.
Where is the cost? Both sides of politics in Australia endorse the policy and the electorate like it. The international attention has favoured Australia’s actions – including New Zealand. The only groups who oppose the legislation are the tobacco companies and their supporters at groups such as the IPA.
I guess I’d taken the lawsuit as signifying too much then. I still think it’s better to implement policy in ways that let us learn whether it’s effective when the cost of running a trial seems relatively low.
I’ve never heard of doing randomized trials like this in US. Has Oz ever done something like this?
Of course, we get some of the effect in US as policies can differ on some things by:
- state
- county
- town
Here are numbers for US states, which varied from 11% to 29% rates. See green-tinted map – mouse-over gives rates. There are enough states to allow for some reasonably similar pairs for comparison. Is there an Oz equivalent map with rates?
Generally, as best as I can tell, in some cases, big differences between nearby states sometimes come from policy, sometimes from other reasons. Utah is lowest, at 11%, given Mormonism. CA is next, at 15%, because it pushes hard. A slight surprise is North Dakota, @ 17%, compared to nearby SD’s 22%. Maybe MN (17%, historically a leader in tobacco reduction) has rubbed off on ND.
But it is hard to imagine any way to set up a really well-controlled experiment that would be cheap, and as HC says, it’s easy enough to do the experiment and it can hardly hurt.
In any case, the real interesting experiment is not between existing packages and plain, but between plain packages and horrific pictures. The latter is supposed to happen in US, but of course, tobacco companies are fighting hard.
If an adult is already addicted, I doubt the packaging makes much difference.
The real issue is whether or not packaging changes lessen the rate addiction in the 11-18-year age range (“younger adults” is a common phrase), the only thing that really matters, I think.
Tobacco companies fight to the death to avoid such lessening … well, it’s not their death, it’s the kids’ deaths, slowly.
Here’s a fantasy thought experiment:
a) Somehow guarantee that no one smokes before age 22.
b) But from 22 on, people can start if they want, and cigarette taxes are slashed to make them a lot cheaper, and they can have any packages they want, and no hassles on nicotine content or anything else.
Item b) sounds attractive … but would tobacco companies take that deal?
I doubt it. Maybe 22 is too high, how about 19? I still doubt it.
The companies know perfectly well they *must* addict children to have a continuing market, given that most adult smokers woudl rather stop if they could, and people who start later find it easier to stop.
Sorry, but I don’t see the point. I think cigarette companies already have limited marketing possibilities, and frankly I don’t see how this plain packaging would help people smoke less. I never got why Governments don’t just make tobacco illegal if they want it gone so badly. I’m not saying that would be right, and I’m sure they like all that money coming from tobacco companies, but these measures are ridiculous
[...] one news item I could not pass over at least without comment. The paper I wrote with David Prentice on the intellectual case for plain packaging did not discuss the constitutionality of the issue – that is now confirmed. John Quiggin [...]