The contribution by Sornpaisarn, Shield & Rehm [1] provides a description and analysis of the interesting alcohol-taxation scheme instituted by the Thai government. The tax rate for a given type of alcoholic beverage is assessed as the greater of two possibilities: a specific excise tax, which is proportional to alcohol content, and an ad valorem excise tax, which is proportional to wholesale price. One effect of this system (compared to a simpler system, where all beverages are subject to the specific tax) is to impose a higher tax on the more dilute and heavily advertised beverage types favored by youths. The authors observe that the intended result of the higher tax rate (and hence price) is to promote abstention by this group, which is desirable as abstention is normative in Thai society.

The authors note that moderation rather than abstention tends to be the goal of alcohol-control measures in high-income western nations, and that a specific excise tax, while effective in promoting moderation, is not effective in promoting abstention. However, in the United States, abstention for youths under age 21 is the official goal, and is currently written into the law in every state. Also, of direct relevance to the Sornpaisarn et al. paper, there is good evidence that even relatively low excise taxes encourage abstention among youths.

In support of that conclusion, I would like to recap research from the National Longitudinal Survey of Youth (NLSY) [2]. The NLSY commenced with a national probability sample of young Americans aged 14–21 years as of 1 January 1979, with detailed questions about drinking collected in 1982–5, 1988 and 1989. Using these 6 years of longitudinal data, we analyzed the effect of state alcohol-control polices on whether the respondent had had a drink in the previous 30 days, and if so whether the respondent had been binge drinking (defined in our analysis as four occasions of six or more drinks). Our analysis thus provides an assessment of the effects of both specific excise taxes and the minimum drinking age for an age group ranging from 17 to 31 years.

At that time (the 1980s) the state excise tax on beer averaged just 3.9 cents per 12-ounce drink (adjusted to 2011 dollars). (Note that these excise taxes are proportional to overall volume of liquid, rather than to volume of alcohol.) Beer was the source of most of the alcohol consumed by youths in the United States at that time, as it is now. The minimum drinking age (set by state law in each of the 50 states) ranged from 18 to 21 years during the period that we analyzed. For our sample, the 30-day prevalence of drinking was 70% and the 30-day prevalence of binge drinking was 14% (p. 388).

Findings regarding the minimum drinking age were consistent across a variety of specifications: youths who were too young to drink legally were in fact about 6 percentage points more likely to abstain, other things being equal, indicating that the law had a modest effect on abstention rates. The minimum drinking-age law also promoted moderation among those who did drink: those bound by the minimum drinking age law were about 2.5 percentage points less likely to binge, other things being equal.

The same analysis documented significant effects for the beer excise tax on the prevalence of drinking by these youths. A typical result (of several specifications) is that an increase of 6.5 cents (adjusted to 2011 prices) in the tax on a 12-ounce bottle would reduce the 30-day prevalence of drinking by about 2 percentage points. Estimates of the effect of the excise tax on the prevalence of binge drinking were sensitive to the specification, but suggestive of a moderating effect.

We also found evidence from the same sample that is relevant to Sornpaisarn et al.'s concern about abstaining and moderation in later life. In particular, we found that the drinking environment during early adolescence affected the likelihood of binge drinking, but not of abstention. Youths who at age 14 lived in a state that had a relatively permissive minimum-age law (legal to drink beginning at age 18) were more likely to binge drink in their late teens and early 20s. That result emerges from an analysis that controlled for contemporaneous alcohol-control measures as well as individual characteristics. It supports a conclusion that alcohol controls that are in place affecting adolescents at about the time they may be initiating alcohol consumption can have long-term effects.

Sornpaisarn et al. observe that the Thai system of taxation, with its combination of specific and ad valorem tax rates, ‘may prevent drinking initiation in addition to discouraging harmful patterns of alcohol consumption’. In response, this comment makes a simple point—that the evidence from one high-income country, the United States, suggests that a specific tax on beer (the beverage of choice for American youths) has a direct effect on abstention prevalence for youths. The higher the tax, the higher the rate of abstention. The Thai system is one approach to engineering a higher tax rate on beer, but another would be to simply legislate a higher specific tax rate for that type of beverage. (It would be prudent, in line with Sornpaisarn et al.'s discussion, to increase the specific excise on spirits at the same time to avoid substitution to that source of ethanol.) Age-based prohibition is also effective at encouraging abstention and moderation for the affected age group, and for encouraging moderation over the longer term.

Of course, the United States differs from Thailand in myriad respects, including income level and drinking patterns, but these national contexts are perhaps not so different that they cannot learn from each other.

Declarations of interest