Thursday, March 06, 2008

Price Elasticity of Booze in Allegheny County

Warning --- this post is pure geekery --- beware

From a policy analyst perspective, figuring out elasticities and policy impacts is a pain in the ass, especially at low levels of analysis as the substitution effects, non-compliance impacts, and border skipping can be a pain in the ass to fully disaggegate in your data. Furthermore, policies are seldom enacted in isolation and without other incentives being generated. Good analysis can and frequently does occur, but it costs money, takes time, and involves significant head-wall interactions.

I have to love the Allegheny County's 10% poured drink tax that went into effect on Jan. 1, 2008 as this is a great little experiment that naturally isolates a lot of confounding and correlating influences. It only taxes the booze at the point of sale, it is not a tax on the purchase of booze from the wholesaler so there is no buyer substitution effects going on by seeing restaurants buying booze in Butler County instead of Allegheny County for instance. No other major policy changes have occurred in the past three months impacting Allegheny County or any other surrounding counties. This is about as clean as an experiment as you can get if you want a quick and dirty elasticity of demand for restaurant served booze in Allegheny County.

And if you really want it, we have it, via the Tribune Review ---

The state sold $3.74 million of alcohol to liquor license holders in the
county in January, a nearly 3.5-percent drop from $3.87 million in January 2007,
said Francesca Chapman, a Liquor Control Board spokeswoman. At the same time,
sales to license holders statewide climbed about 4 percent, from $28.54 million
to $29.69 million.

If we assume that restauranteers have a good ability to project demand into the near future and that storage is expensive, we would expect a just-in time liquor purchase system for most restaurants, or at least just this week/month systems. They want to buy about the amount of booze that they will sell.

So at a wholesale level the crude elasticity of demand is about .35 --- a 10% increase in retail prices will produce a 3.5% reduction in demand. A slightly more sophisticated measure would that if the tax was not enacted, we would expect Allegheny County to pretty much follow statewide trend, so the elasticity of demand is roughly .70 --- a 10% increase in retail prices will produce a 7% reduction in quantity demanded.

Now there are a couple of provisos in here. The first is that it is a single data point, and a better analysis would look at some trend information to narrow the error band. Secondly, this is only for hard alcohol; there is no data to suggest if restaurants, taverns, and bars sold less hard booze (which the state has a monopoly on) but saw some substitution to cheaper beer. Third and most noticably, very few bars raised their prices by exactly 10% to cover the new tax. I know at a downtown bar my coworkers and I occassionally visit for happy hour that my $3.00 Penn Dark is now a $3.50 pint. That is a 16% increase in prices.

But this gives us a quick and dirty maximal number of the short run elasticity of demand for hard booze in the county is no more than .70. Since the elasticity is less than 1.0, the county could probably raise the rate a bit more and still take in more revenue.

No comments: