I haven’t seen any research on this (if someone knows of any, let me know in comments), but my best guess in the absence of good evidence would be that the success of the ban reflected instabilities in previously existing informal norms about where people could or could not smoke. Laws that work against prevailing social norms face an uphill battle in implementation – unless people come to a general belief that non-compliers are highly likely to be sanctioned by the public authorities, they are likely to carry on doing what they always do. Hence, for example, the continued failure of the RIAA etc to stop file-sharing – file-sharers who both (a) think that there is nothing wrong with swapping music and movies, and (b) that the chance that they are going to be punished is low, are going to go on sharing files (current US law tries to counterbalance this problem by applying relatively draconian penalties to the few file sharers who are caught, but this strategy carries its own problems). Laws that broadly fit with prevailing informal norms, will, obviously, have few implementation problems.
But what we may have seen (if my guess is right) with smoking bans is an unusual case in which prevailing norms (that Irish people can smoke in pubs to their hearts’ content, and that others will just have to put up with it) were much more fragile than they appeared to be, and that the change in law made it easier for those disadvantaged by the prevailing norms to challenge smokers and to shame them into stopping smoking in certain places, hence creating a new set of robust norms.
For the record: I smoke, but hate the smell of stale smoke and generally have a preference for it being done outdoors (below-freezing Christchurch nights excepted). New Zealand banned smoking in all indoor workplaces a few years ago. I think this was very bad policy – I oppose coercive solutions to minor or nonexistent problems – but don’t feel particularly put out myself.
It seems pretty obvious that the optimal market outcome would involve some smoking and some non-smoking bars. This is what the market was moving towards, with a small but growing number of smokefree bars here in New Zealand before the ban. Due to the stickiness of social norms, though, this movement might have been slower than we might prefer. I have no idea what the optimal mix would be, but I’m fairly comfortable saying there were too few non-smoking bars in New Zealand before the ban.
I’ll put on my vulgar utilitarian hat for a second and offer some thoughts. My feeling is that the situation in which all bars ban smoking would be preferable to the situation in which all bars allow it. I can even buy the idea that the new blanket ban increases welfare relative to the old situation with few non-smoking bars. To simplify things horrendously, I suspect there’s a relationship between utility and the proportion of non-smoking bars which looks something like this:
Before the ban, we were somewhere near the far left of the curve. With the ban, we were pushed to the extreme right. This is welfare improving in the short run, but if we think we were slowly moving towards the centre anyway – which seems undeniable – the ban precludes the optimal long-run equilibrium. Whether the ban is welfare-improving depends on the relative magnitudes of the short-term and long-term welfare effects, as well as the appropriate discount rate. As a vulgar utilitarian, I have no way of knowing which dominates without some sort of revealed preference mechanism. The aggregate welfare implications of the ban are not obvious. What is obvious, though, is that a removal of the ban would improve things, as Jonathan Adler argues:
What would happen were such bans to be repealed? My best guess is that relatively little would change. When I think about my favorite local restaurants, I cannot see any of them allowing patrons to smoke even if the law were changed. There are one or two local bars, however, that I suspect might allow smoking on the premises, but they would be the exception. So whereas before the smoking ban here in Ohio, most restaurants and bars allowed smoking in a separate room or at the bar, were the ban repealed today I would be willing to bet that most restaurants and bars would remain entirely smoke-free.
What does this all mean? On the one hand, if most restaurants and bars would remain smoke-free, it seems to me the argument for allowing some establishments to adopt different rules is that much stronger. Remove the bans and us libertarian-types can still toast to the free market system in a smoke-free pub. But it is important to acknowledge that this state of affairs exists today because of the initial government intervention. The smoking ban appears to have helped solve a collective action problem that had kept a suboptimal norm in place. So even if a ban limited the ability of business owners to set the rules for their own businesses, it may have also helped them shift toward preferable business practices. Non-governmental efforts may have produced the same result eventually, but it would almost certainly have taken longer. So smoking bans have been beneficial, but it may also be the case that the maintenance of such bans is unnecessary to retain most of their benefits.
The initial ban shifted us to the extreme right of the graph, which was preferable to the old situation. A removal of the ban would allow us to move towards the centre as some bars would allow smoking, presumably right up until the optimal point. This would be the best of all possible worlds. Unfortunately, I can’t see the New Zealand smoking ban being repealed anytime soon. The normally-dominant angry libertarian in me still thinks the anti-smoking movement is all bigotry, by the way. At worst, I think smoking in bars bothering others is a small problem. Social norms are only going to stop people voting with their feet for smoke-free bars when they are pretty close to indifference.