Sin Taxes and the Welfare State

In the comments to E.D.’s post on a hypothetical tax on alcohol, Michael Drew writes:

A significant portion of the negative externalities come in the form of medical costs from DUI and long-term disease, in other words health-care costs. These are costs that will simply be rolled into the general health care system, whichever way. Health care is simply government’s one overriding cost liability going forward. In theory, yes, one could do the accounting necessary to try to guide these dollars to addressing alcohol-related health-care costs, but why bother? the point is just to discourage excessive use, and raise revenue. If alcohol abuse prevention programs are shown cost-effective, okay, fund those first with this money, but if they are proven cost effective, we should fund them anyway. Money is fungible. the one thing that would seem important would be to see that the revenues raised go to the levels of government that bear the related costs. If that is primarily states, the feds could offer matching funds to those states that raise the tax to X.

Again, people are completely within their rights to object to this as nanny statism creeping into our lives, and they will, but health costs public and private are really the issue, and this is a rather obvious way to help those on both counts. The claim that the revenues will be targeted only to the specific negative externalities of this specific behavior is really kind of a neat formalistic way to sell it — even if it’s done in practice. Net-net, it’s just a good move, on the assumption that revenues will be tight, and government will be being forced to prioritize in response to public demands in any case.

Michael is absolutely, 100% correct in making this argument,* at least in the context he is making it.  I recall that he has also made a similar argument in the context of New York’s proposed sugar tax.  Once we have embarked on a particular social program, it becomes fiscally responsible to minimize the costs of that program, whether by placing strings on participation in the program (e.g., here) , or by prohibiting activities that increase the program’s cost, or – as in this case – by taxing the activity that increases the program’s cost.  In short – anything that increases the program’s cost becomes a negative externality that is fair game for regulation, prohibition, or taxation. 

When this happens, the impact of the underlying social program on liberty is very real.   It is no longer just a matter of citizens seeing a few dollars less on their paycheck, but instead becomes a matter of citizens being unable to make their own decisions about their personal lives without some form of direct government intrusion.  Now, admittedly, when the intrusion takes the form of a tax, it is a much smaller intrusion on liberty than when it takes the form of a prohibition or regulation.  But it is nonetheless an intrusion. And, importantly, for the very people the social program is most supposed to benefit, this minimal intrusion can be quite significant, since they are the ones who will be least able to pay the taxes. 

Therein lies the crux of the problem.  Social programs are supposed to improve the quality of life of the poor.  But the restrictions that we wind up imposing on those programs undermine those improvements to some degree or another.  Now, if by “quality of life,” we all mean a particular thing – say a particular lifestyle, or the maximization of life span –  this is not a real problem.  Unfortunately, one’s idea of “quality of life” can vary tremendously from person to person and locale to locale.  For instance, one can rationally – and correctly – conclude that the happiness brought about by the occasional overindulgence in alcohol or the daily glass of soda one drinks makes life worth living in a way that a marginal increase in life expectancy does not.   If one so concludes, then the benefit to them of the relevant social program has been reduced in a material way.  If the restrictions or taxes wind up covering a wide range of happiness-inducing behaviors or are particularly heavy-handed, then they can conceivably even so reduce the benefit of the relevant program as to make it a net loss in quality of life for the program’s beneficiaries – and this says nothing of unseen jobs that are lost or never created because of the restrictions.  A sufficient number of these restrictions can wind up being a very real way of using social programs (even if unintentionally) to control the behavior of beneficiaries thereof.  The beneficiaries, in this sense at least, become “Well-Fed Slaves.”

The more we make everyone responsible for helping everyone else by law, the more we make everyone else’s otherwise entirely personal behavior an externality.  There is a tendency, I think, to realize this only after we’ve made everyone else responsible for helping everyone else by law in a particular arena.  So we often don’t internalize the full costs of a program until long after it’s passed.  Program X may well be able to obtain majority support for passage when it is rationally claimed that it will help millions of people escape from Problem Y at a cost of Z dollars; but would it still be able to obtain such support if it was also mentioned that several years down the road, people will be unable or less able to engage freely in activities A, B, and C in order to obtain marginal savings in Z?  Maybe, maybe not – but the point is that the tradeoff doesn’t get included in the initial debate, nor realistically could it.   

So where does that leave me?  I’ve gone on record as saying that I support the concept of a welfare state and, indeed, of some form of social insurance, even if in a significantly different form than they currently exist.  But in many ways, almost any welfare system or social insurance system is going to be susceptible to these problems.  I think a negative income tax system would, in addition to being a near-ideal way of providing a social safety net, be significantly less susceptible to the “negative externality of increased negative externalities” I describe above.  I think this because everyone would at least in theory be receiving an identical subsidy – one’s personal decisions will not affect the baseline one receives, only the amount of taxes they pay in return, and the argument in favor of restrictions on welfare/social insurance recipients isn’t usually that they ought to be making boatloads of money, it’s that they ought to be deterred from acting in ways that make it more likely that they’ll continue to be eligible for welfare or that make the costs of their social insurance higher.  But even still, I would not expect a negative income tax to be completely immune to these pressues – a rationale could still be created that there is a societal duty to avoid behaviors that restrict one’s potential income. 

Exit question: Assuming that a welfare state should and must exist, is it possible to prevent or mitigate the response to the “negative externality of increased negative externalities” that results in increased regulation of personal behaviors through prohibition, taxation, or regulation? If so, how?  Constitutional amendment?  Differently structured social programs and safety nets?

*I’m actually less opposed to sin taxes than one would expect me to be, at least under the right circumstances.  In the context that Mark Kleiman is proposing the liquor tax, for instance, I think a liquor tax makes a decent amount of sense as a legitimate Pigou tax that need not be dependent on an “increased-cost-to-social-programs” rationale.  In other words, the negative externalities would exist independent of any social programs.  In the context that Matt Yglesias proposes such a tax, however, I firmly disagree with the rationale for the reasons stated above.