Tuesday, February 17, 2009

Beer Tax Redux: Fred Thompson Weighs In

Fred Thompson, friend of the blog and resident tax and budget expert, sends along his take on the beer tax idea:

Jeff Alworth of Blue Oregon blogged very informatively the other day on the proposed increase in Oregon’s beer tax: Brewing Beer is Not a Sin. To an economist, specific excises raise two additional questions that Jeff did not address: substitution and incidence.

Let’s look at substitution first. As Patrick recently observed, one justification for taxing beer, wine and spirits is that their consumption produces bads and taxing them reduces consumption. People drink beer, wine or spirits to get drunk, which can mean anything from a pleasant buzz to total stupor. For example, although moderate alcohol consumption appears to be good for you, excessive consumption is not. Moreover, even moderate alcohol consumption can lead to negative externalities like dangerous driving. While taxes on beer, wine and spirits are a blunt instrument, the evidence is that they work to reduce the damage associated with alcohol consumption (there is some evidence, that their effects on the benefits of alcohol consumption are disproportionately high, but this evidence is by no means conclusive).

However, if reducing the damage associated with alcohol consumption is the justification for taxing booze, the tax should reflect its potential for doing damage. That means the tax should be based primarily on alcohol content and should be the same for beer, wine and spirits. That is not now the case in Oregon. The alcohol in beer and wine is implicitly taxed at lower rates than the alcohol in spirits (and the tax on spirits is based on the price of the drink not on alcohol content).

The other justification for taxing beer, wine and spirits is that booze taxes generate lots of tax revenues: a means of plucking the goose with the minimum amount of hissing. Booze is an inferior good. Its price elasticity of demand is less than 1, which means that a one percent increase in the price of alcohol will decrease its consumption by less than one percent. Of course, that’s not equally true of all populations. Price affects the consumption of occasional drinkers and teenagers much more than of alcoholics. (Teenagers tend to substitute pot and other tax-free but illegal recreational drugs for alcohol when taxes are increased on beer, which is less harmful for a variety of reasons.)

Since the flip side of price inelasticity is income inelasticity, this takes us to the incidence question. One can estimate the income elasticity of a tax using ordinary-least squares regression according to the following specification: ln(Ci) = a + b(lnYi) + e, where C is tax payment, Y is family income and i represents the ith family. Here, b shows the percentage change in tax payments as a result of a one percent change in family income, which is income elasticity. Elasticities greater than one mean that the tax is progressive; less than one, regressive. It isn’t possible to compute the income elasticity of the alcohol tax in Oregon, but we can use household accounts from the US census to estimate the income elasticity of alcoholic beverage purchases, which indicates that b is approximately equal to .6. That is, a one percent increase in income increases purchases of alcoholic beverages less than one percent, which implies that the tax is regressive.

But, wait, there’s more. There is also some evidence that most of the increase in alcoholic beverage purchases as income increases is not reflected in greater alcohol consumption, but a preference for up-market drinks – Beefeaters instead of Potters gin, fine wines rather than cheap plonk, and designer brews rather than Bud. Consequently, it is likely that a tax on alcohol is not only regressive, but highly regressive.

Is this bad? Hard to say. However, it is somewhat anomalous that the same people who want to increase taxes on alcohol are also big fans of minimum wage increases. One hand gives; the other takes away.

2 comments:

Jeff Alworth said...

Fred, over at Beervana, I have posted the per-capita consumption by state alongside the state beer taxes. There appears to be no correlation between the two. Given that state taxes vary rather substantially, I was surprised by this--though it's not clear how it affects the debate. I wonder also how it relates to your price elasticity argument.

(I'm just a poor beer drinker, ignorant of economics, so bear with me.)

Fred Thompson said...

That the demand for alcoholic drinks is price inelastic means that very large price inceases lead to small decreases in alcohol consumption. This claim implies that differences in state beer taxes wouldn't explain very much of the variance in per-capita consumption, if factors other than price affect one's taste for suds. That's especially likely where state taxes on other drinks are not relatively equal, as they are not in most states.

Nevertheless, the evidence seem pretty persuasive that some drinkers are more responsive
to price than others, e.g, teenagers.

The leading expert is a guy at CUNY named MICHAEL GROSSMAN.