In the interesting debate on libertarian paternalism at Cato Unbound both sides have mentioned the issue of smoking bans. In the anti-paternalism lead essay, Glen Whitman uses smoking bans as an example of a slippery regulatory slope: first it was just the airplanes, then it was restaurants, what’s next? Richard Thaler counters in the pro-paternalism essay that smoking bans aren’t even really nudges:

First, most of the anti-smoking laws are based on externalities, not paternalism. People do not want to fly, eat, or work in smoke-filled environments. Indeed, many smokers favor such laws. Note that while smoking bans are not nudges, they are shoves…

Scott Sumner chimes in, disagreeing with Thaler that smoking is an externality at all:

Doesn’t this argument violate the Coase Theorem? For example, let’s take the ban on smoking in the workplace. Where is the externality argument? Doesn’t the employer already have an incentive to put in place the smoking rules that minimize his productivity-adjusted wage bill?

One important problem not discussed is that if we are nudging (or shoving) people away from a behavior that is (or could be) an externality, what are we nudging them towards? The unintended consequences are potentially serious here. For instance, the literature on smoking bans suggests that they increase drunk driving and exposure to secondhand smoke for the children of smokers, both of which are more obvious and more egregious externalities than the one we were trying to get rid of in the first place.

Negative unintended consequences may occur even when there is no externality. For instance, a new paper in the current edition of AEJ: Applied Economics uses a field experiment at a fast food chain to show that if not properly designed, a nudge that successfully reduces an individual’s caloric intake of sandwiches may be offset by an increase in their caloric intake of soda and side orders. Depending on the kind of sandwich, side order, and soda involved, this may make the individual worse off. Shifting from less of a turkey wrap to more french fries with trans fats would clearly make the person worse off.

Both examples show that when weighing nudges (or shoves) we should consider the potential unintended consequences, otherwise we may end up pushing people in a bad direction. When externalities are potentially involved, this is especially important. The smoking bans also provide a warning that even if we identify a very serious unintended consequence down the road, the legislation won’t necessarily be undone. We should be extremely cautious about passing these laws if we are uncertain about our ability to identify unintended consequences in advance; by the time we know for sure it could be too late.