Smokonomics

Hayek goes barcrawling

In the coming months, Washington, D.C., will decide whether to join California, New York, and a host of other towns and municipalities in banning the combustion and inhalation of tobacco products in its bars and restaurants. To the delight of the self-appointed guardians of public health, my local watering hole may soon have the freewheeling ambiance of a dentist's waiting room.

The putative logic of these laws is to protect bar and restaurant workers. In reality, this is something of a fig leaf: The primary support for smoking bans seems to come from folks who don't like having to dry clean their clothes after a night out, and who are therefore asserting their God-given right to visit any bar or restaurant they like on their own terms, owners and other customers be damned.

But the anti-smoking movement's official spokespeople still seem to have the minimal decency required to be ashamed of such a naked appeal to self-interest. Instead, they've been forced to resort to the argument that workers who don't mind smoke must be protected from making stupid choices about their own bodies, while those who do mind it must be protected from the hassle of either making a trade-off or seeking a more congenial job over the course of the many years it takes for environmental tobacco smoke to increase health risks. To the extent that it does, that is—the data on the precise degree of risk involved is notoriously ambiguous.

Ban opponents have countered that those same workers may suffer as smoking patrons opt to walk a mile—home—for a Camel. Despite some evidence that this has not happened to the extent that some feared in California or New York, ban boosters seem determined to prove that, even when they seem to have a legitimate argument, it's more fun to run ads made of ecologically friendly 100 percent straw. Nobody, after all, ever claimed that smoking bans would mean that "no one will ever" go out carousing, "ever," or that it would "wipe out all" bars. And there's something more than a little disingenuous in the claim that people who continue to use non-smoking elevators won't perhaps behave differently when it comes to the decision about how long to spend at the neighborhood pub.

But putting aside this sort of intellectual dishonesty, it's nevertheless true that a welter of studies have not found a dramatic decline in aggregate hospitality sales in many areas that have enacted smoking bans. In some, business even appears to be up.

Yet aggregate data can be misleading. It may be that aggregate revenue doesn't show a drop in business, but that doesn't mean that particular restaurants aren't hurting. In fact, you can count them. The pattern emerging in nearby Montgomery County is that business is steady or improving in big chains like Ruby Tuesdays, while small independent businesses are taking a hit. In an industry with paper-thin profit margins, those places may eventually have to close—a danger that's not visible if you lump together all bar and restaurant revenues, treating them as a mega-business, rather than a collection of separate businesses.

But pretend you don't care about the diversity of nightlife: Students of Mises, Hayek, and their intellectual descendants should spot another problem with the kind of analysis that yields the conclusion that smoking bans aren't economically harmful.

I've never been comfortable with the most extreme forms hyper-a priorism advanced by some adherents of Austrian economics. As the failure of Euclidian geometry to describe our spatially curved universe proves, the internal consistency of a formal system is no guarantee that it models reality accurately. An extreme anti-empiricism can easily become a lazy way of ensuring that being an Austrian economist means never having to say you're sorry. But this seems like a case where the Austrian objection to over-reliance on econometrics seems like good counsel.

Reality, as social scientists frequently lament, is messy. Laboratory experiments can be rigorously controlled; real life is less accommodating. You can measure which way restaurant revenues trend before and after some new legislation, but establishing causality is a trickier matter. The niggling question always remains: Can we be sure we've isolated the independent variable? Have we controlled for every potential explanatory factor?

But these methodological quibbles are secondary. The Austrian insight that's most important to bear in mind when considering the wisdom of such laws is that, contrary to neoclassical assumptions, real world markets are never in equilibrium. Rather, market activity is a constant discovery process.

This is important because, absent this observation, it may well seem as though smoking bans have successfully corrected a market failure. If, in the wake of a ban, aggregate business remains constant, or even picks up, and people report satisfaction with their new smoke-free carousing environments, it seems natural to regard the new status quo as an improvement. And if smoking policies never changed, that conclusion would be right.

In reality, though, these policies change over time: Owners change their mind about which rule will be most profitable, and new non-smoking establishments open while smoke-friendly ones close down. Economists sometimes forget that markets are not lightning calculators. In an old joke, two Chicago school economists are walking down the street when they notice a five dollar bill on the sidewalk. One stops to retrieve it, and the other warns: "Don't bother; if it were really there, someone would have picked it up already." But of course, in the real world, sometimes there is money on the sidewalk.

It seems fairly clear that, at present, the market mix of smoking and non-smoking establishments is suboptimal: It does not yet fully reflect the public's growing preference for smoke free dining and carousing. If we imagine an inverted-U curve, with the relative proportion of smoking and non-smoking establishments on the X-axis and restaurant revenue (and, presumably, customer satisfaction with the available options) on the Y-axis, it's fair to suppose that, at present, localities that permit smoking are well to the left of the optimum, with the mix biased too heavily in favor of smoke-friendly joints. Smoking bans effectively jump the curve, apparently landing at a higher point far to the right of the optimum.

The problem is that, while the market process seeks to approach that optimum over time, uniform prohibition locks in an extreme, almost certainly suboptimal mix. Ironically, an excellent argument for the repeal of these restrictions is provided by the fact that some bar owners have reported with delight that, contrary to their expectations, business picked up in the wake of the ban. Those businesses now know that, given their clientele, a non-smoking policy is optimal for them. The unlucky losers, the bars whose chimneyesque customers have deserted them, discovered the opposite.

It's also worth stepping back to ask whether the debate really is—or should be—really about restaurant revenues at all. Opponents of pork-barrel spending have long cited Frederic Bastiat's famous "broken window" fallacy to explain why government spending doesn't really "create jobs," as its advocates sometimes like to claim. The money injected into the economy at one point had to first be removed somewhere else: The "seen" effect of state spending is perfectly balanced by the "unseen" loss of whatever enterprise would have been supported by the same dollars had they been left in private hands.

Yet the same holds for jobs lost to regulation. A smoker who spends less of her time and money at the local caf? because she can't puff a clove with her latte isn't just going to use the cash for toilet paper: She'll spend it somewhere else. There are significant deadweight losses and transaction costs while the economy adapts, of course, but it remains the case that less spending at the bar will mean more spending elsewhere.

The obvious rejoinder is that tampering with the natural market outcome makes the reallocation of spending inefficient. But "efficient" here isn't used in the colloquial sense of putting out more widgets per pound of aluminum. Efficiency in the economic sense is about using resources in the way that maximizes the satisfaction of subjective preferences. The problem with regulation isn't that it decreases the net amount of money floating around the economy—it doesn't. Rather, the problem is that it forces that money to be channeled suboptimally, into sectors less productive of customer satisfaction.

Economic arguments, arguments that look at the flow of cold hard cash, have a comfortingly scientific ring to them. You can appeal to objective-looking graphs and figures. That's why restrictions on, say, the color you're permitted to paint your house, are typically justified by an appeal to preserving "property values," rather than the simple majoritarian argument that most of your neighbors think that fuchsia with a puce trim is ugly. But, of course, all it means to say that such a color scheme would lower property values is that one expects a majority of potential buyers would find it ugly.

A smoking ban may or may not reduce revenue at D.C. bars relative to the status quo—that will depend in part on the willingness of smokers to leave the district for smoke-friendlier parts, an easier task than abandoning LA for a bar in Nevada. It will almost certainly be worse, economically, than the mix that the market would eventually achieve. But all that is just shorthand for saying that if we leave owners and customers to make their own choices, everyone can have the kind of experience that they most prefer. And while that's not the kind of argument that translates well to a PowerPoint slide, it has the virtue of being true.