Sunday, September 3, 2023

No Connection Between ABC Laws and Alcohol-Related Outcomes

States in red have ABC laws. (Source: wiki)

American states have varying levels of state control over alcohol sales. When the topic is discussed at all, it's typically framed as an issue of controlling a potential vice - ostensibly, those laws are there to protect the states' citizens from the bad effects of alcohol. So it's obviously worth asking: do these laws actually help?

One of the nice things about a federal system is we can give some states the slack to conduct their own individual experiments, and see if they work out, so in that sense it's a good thing that some states have ABC laws and some don't. I initially wanted data either for QALY lost due to alcohol per capita by state, or alcohol use disorder by state. I was unable to find that data (if you have it, contact me and I will add it to the analysis.) But there were three outcomes where I could find data that are on their face relevant to why we would care about this: alcohol-related death rate, cost of alcohol related morbidity and mortality per capita, and volume of alcohol consumed per capita.

Using two-tailed t-tests and data as cited below, there is no significant difference between ABC states and non-ABC states in death rate, alcohol-related health costs, or consumption.

If you don't trust fancy statistics and just want to know the averages, then I have to tell you, the (again, not significant) averages were WORSE for death and consumption in ABC states, and better for costs in ABC states. There's just no way to look at these numbers and conclude that ABC laws improve alcohol outcomes.

Are there outliers? Yes, unsurprisingly, Utah. Removal of Utah from the data makes the consumption difference significant the other way. That is to say, if we take out Utah, there actually IS a significant difference in consumption between ABC and non-ABC states - and ABC states drink MORE.

Of course this says nothing about causality, with or without the difference in consumption that becomes significant when we remove Utah. In that case, you could understand the statistics as meaning
  • ABC laws actually make people drink more, or
  • People in ABC states already drank more, and their governments passed ABC laws to reign them in, or
  • It's something else, and ABC laws have no effect.
Obviously none of these are palatable as publicly debated arguments for ABC laws ("The Governor announced today we're keeping the ABC laws because you people are impulsive animals who can't be trusted with liquor") and the states who are collecting revenue from such laws are not going to do anything to damage the fragile justification for continuing the system as-is.

Of course, the next question is "Then why do states keep ABC laws in place?" The answer is likely revenues, veto blocs (the employees at the state stores who want to keep their jobs more than they want the state government to make sense for citizens) and some degree of virtue signaling to voters. If you're ever fortunate enough to hear public officials and state store employees and unions defend the status quo on this, you'll be treated to some vaguely moralizing, gyratory, evidence-free statements; gambling produces similar sorts of 2 + 2 = 5 rhetoric, which is why the arrangement is rarely if ever directly addressed. Consequently, in an environment where this kind of absurdity is accepted with a shrug, adjacent nonsensical laws, e.g. the Johnstown Flood Tax still being collected today from alcohol sales in Pennsylvania, can persist and metastasize easily.*

Remember this next time you're at your local state store, talking about this with a friend, or most importantly, voting.


REFERENCES
Death rates: drugabusestatistics.org
Alcohol health-related costs: CDC
Alcohol consumption per capita: wisevoter.com

*There are those who claim Johnstown has been rebuilt, but reasonable people can disagree on this. That said, guys, it's been well over a century. Either they've rebuilt Johnstown by now, or they should give up.