Do Alcohol Taxes Still Work to Save Lives?

Avid readers of our blog would note that Twenty-One Debunked supports raising the tax on alcoholic beverages almost as wholeheartedly as we support lowering the drinking age to 18.  And there are reams and reams of research evidence over many decades--locally, nationally, and internationally--that find that higher alcohol prices (and thus taxes) save lives both on and off the highways as well as reduce crime, violence, and other alcohol-related problems.   In contrast, the ageist abomination that is the 21 drinking age has not consistently demonstrated similar effectiveness in that regard, no matter what sort of pseudo-consensus exists in the minds of its most ardent supporters, and the best evidence thus far has exposed the specious claim of saving lives as little more than a mere statistical mirage all along.

That said, for the specific endpoint of alcohol-related traffic fatalities, among the reams of evidence there have been a few outlier studies that seem to cast doubt on the lifesaving effect of alcohol prices/taxes as well.  The most recent one in 2017 by McClelland and Iselin of the Urban-Brookings Tax Policy Center studied the effects of the Illinois alcohol tax hikes in both 1999 and 2009, and found no long-term lifesaving effect from either one in terms of drunk driving deaths.  In contrast, a previous 2015 study by Wagenaar et al. had found a fairly large drop in alcohol-related traffic deaths following the 2009 Illinois tax hike, even after controlling for the effects of the Great Recession.  The biggest difference between the two studies was that McClelland and Iselin used the Synthetic Control Method (SCM) while Wagenaar et al. did not, and while a good method, like all methods it too can have its own share of pitfalls.

As for the other outlier studies, most of those are dissected and discussed in a 2015 replication review by David Roodman, which still concludes that a true lifesaving effect is likely.  One such outlier study is by Dee (1999), whose control for state-specific time trends apparently removed too much useful variation in state-level beer taxes.  And while Roodman did not discuss our all-time favorite study by Miron and Tetelbaum (2009), we should note that this study in fact began by replicating Dee (1999) using more years of data and including Alaska, Hawaii, and DC, and in contrast to Dee they did apparently find a fairly strong inverse correlation between beer taxes and 18-20 year old traffic deaths even after adjusting for state-specific time trends.

Granted, it is true that for the specific endpoint of DUI deaths, the price of alcohol may not be quite as important as it once was.  Drunk driving is far less common and far less socially acceptable than it was a generation ago, and legal sanctions against it are much stiffer now as well.  And with alcohol prices currently at a record low in relative terms, and alcohol taxes generally being a small portion of the overall price, the link between the two may not be as salient or noticeable as it once was due to being swamped or masked by other factors.  But that does not mean that it is ineffective, given the fact that several more recent studies continue to find such effects, and the numerous studies that continue find fairly large benefits in terms of reducing non-traffic deaths and harms as well (cirrhosis, unintentional injuries, cancer, crime, violence, STDs, etc.).

Thus, the overwhelming weight of the evidence still continues to support the idea that raising alcohol taxes/prices is an effective (and especially cost-effective) public health policy in terms of saving lives both on and off the highways as well as reducing alcohol-related problems in general.  And if it is high enough, it is also justified on Pigouvian grounds as well.  So what are we waiting for?