This is a guest post by Craig Roche, a data scientist and artisanal landlord.
Whiskey is very easy to make. Farmers used to make it at home using their crops, and Henry Ford designed the Model T to run on home-distilled ethanol. George Washington distilled 55,000 bottles/year when he retired from being President. Even the mutineers from the Bounty set up a still on Pitcairn Island, and proceeded to get rip-roaringly drunk for weeks at a time. Whiskey is also very cheap to produce; a bushel of corn ($5 or so), plus 60 cents worth of natural gas can produce 11 liters of automobile-grade ethanol, which, when suitably diluted and aged for drinking purposes, can fill 35 bottles. Whiskey for human consumption requires higher-quality inputs, more energy for multiple distillations, and additional handling, but even so, decent hooch can be produced for less than $2/bottle. In the 1830s, the equivalent of a bottle of whiskey went for about $5, and Americans responded by guzzling roughly one each week per capita; as young children generally abstained, actual drinkers drank substantially more, all of which was tax-free.
If we assume that the desire to drink, especially among the poor, is an important motivation in peoples’ lives, you would expect alcohol markets to shed light on political conditions across states.
Jack Daniels is the world’s most popular brand of whiskey, and is widely distributed in every state in the US in a standard 750ml container; it is produced fairly close to the mean center of US population, so it should therefore work as a good lens on alcohol politics. Left to a free market, one would reasonably expect that the cost of a retailed bottle would vary with transportation costs, and somewhat with labor rates, or alternatively, that lower-income consumers would spend about the same fraction of their income on Jack Daniels across the USA, or in other words, that the working time per bottle would be constant.
To test this, I researched the cost of a standard bottle of Jack Daniels in each of the states at a high-volume liquor store in the largest city in each state, and compared it to the average wage at the 25th percentile:
Figure 1: Labor Cost of Jack Daniels (Image source: Craig Roche)
The results were not what I expected. It turns out that the constant-working-time-per-bottle hypothesis is not even close.