Most articles about drinking or not drinking alcohol assume two things about consumption patterns. They assume that the population forms a rough bell curve with a small group of teetotalers at one end, a small group of “alcoholics” at the other end, and a vast bulge of “moderate drinkers” in the middle. And they assume that most of the alcohol sold is consumed by those moderate drinkers. These assumptions shape everything from how drinking appears on TV to how alcohol is taxed and regulated. As Prohibition vanishes in our rear-view mirror, US media, commerce, and policy decisions invoke a vast population that enjoys a drink or two every evening, maybe a little more on weekends or special occasions. We’re collectively sorry if our policies harm underage drinkers or alcohol dependents, but we can’t be expected to give special consideration to these small minorities with the needs of so many moderate drinkers to consider. It wouldn’t be fair.

There’s just one problem with this assumption: it’s wrong. Alcohol consumption patterns don’t form anything like a bell curve; they more closely approximate the 80/20 Pareto Principle, in which most alcohol is consumed by a small number of drinkers.  Look at this chart, from the Washington Post Wonkblog, which is based on a rigorous book-length study by Duke University Professor Philip J. Cook.

That vast, moderate-drinking middle doesn’t exist. What we actually have in the US is a 70 percent majority of non-drinkers and extremely light drinkers. Then we have some moderate drinkers, less than 20 percent of the population because the top of the ninth decile verges on heavy drinking. Finally, we have a decile of really heavy drinkers, people who average some ten drinks a day. They consume 74 percent of all alcohol sold and do considerable damage to themselves and to other people.

We’ve been willing to overlook that damage in order to accommodate . . . twenty percent of the population? Seems bizarre when you think about it. In fact, our collective actions make sense only if we don’t realize how small the population of “moderate drinkers” actually is. Here’s one way to think about it: the fraction of the US adult population that drinks moderately is the same size as the fraction of the US adult population that believes alcohol should be illegal: just under one-fifth. I don’t think either of these minority interests should dominate our alcohol policies, yet moderate drinkers do–or seem to. The reality is that the interests of moderate drinkers are a cover for the real driver of our alcohol policy: the money to be made from the tenth decile.

In 2017, the most recent year for which we have statistics, the alcohol industry generated a whopping $1.2 trillion in worldwide revenue. Just to give you a sense of scale, that’s a couple hundred billion more than the revenues of OPEC or the global pharmaceutical industry. The US’s share of that $1.2 trillion is $234.4 billion, dwarfing revenues from tobacco, soft drinks, pharmaceuticals, illegal drugs, and the entire defense industry. Of Big Tobacco, Big Pharma, and Big Oil, only Big Oil is as big as Big Alcohol.

Where do those earnings come from? As I mentioned, Professor Cook reports that ten percent of Americans consume 74 percent of all alcohol sold, downing an average of ten drinks per person per day. Seventy-four percent of $234.4 billion is $173.5 billion. In other words, in 2017, the US alcohol industry earned almost as much from the abuse of its product as Amazon earned from its entire commercial and technical empire ($177.9 billion). By contrast, light and moderate drinking generated $61 billion in revenue–a hell of a lot of money, to be sure, but dwarfed by the $173.5 billion from the tenth decile. The fact that alcohol abuse generates such vast amounts of money is the industry’s best-kept secret.

The secret is easy to keep because the media rarely investigate–or report on– the business or policy side of alcohol abuse. Or much of anything having to do with the alcohol industry, for that matter. That’s why you’ve never heard the phrase “Big Alcohol” and may even have squirmed a little when I used it a moment ago.

On the rare occasions that the media do confront alcohol abuse as a public policy issue, they always shift their focus to moderate drinking. Always. Last August saw a classic example of this technique when leading medical journal The Lancet published a worldwide study concluding that alcohol had killed 2.8 million people in 2016 and represented a global public health crisis. The news outlets that covered the study–and many didn’t–zeroed in on a single, relatively minor point made by the Lancet researchers: that the very safest level of alcohol consumption is zero. The study never said people shouldn’t drink moderately; it said that, if governments decide to issue recommendations about alcohol consumption, those recommendations should reflect what the science shows. But the news coverage, especially the shorter articles, treated the whole complex issue as a debate over whether it’s healthier to drink a little or not at all. Forget the 2.8 million dead people; what really matters is whether Mike and Mary Moderate can continue to believe that their daily glass of merlot will help them live a little longer.

Here’s an analogy. In the bad old days, when gays and lesbians wanted to date publicly, they often employed a friend of the opposite sex as a “beard,” someone to pose as their date while the actual date posed as a tag-along friend. Now, the alcohol industry uses moderate drinkers as a “beard.” It can’t admit its passionate devotion to the consumers that furnish most of its profits, heavy drinkers, so it pretends to court moderate drinkers. For example, when Costco and company pushed my state’s voters to let large chain stores sell liquor, they did not say, “Privatizing liquor sales will help people with alcohol use disorders buy much more liquor with much less effort and embarrassment.” They certainly didn’t add, “And it will make liquor much easier for underage drinkers to acquire.” Instead, they painted a picture of Mike and Mary Moderate longing for a wider range of choices and the convenience of picking up those choices along with their groceries, prescriptions, or furniture. And the strategy worked because voters assumed Mike and Mary represented the majority of the population and the bulk of liquor sales.

Because anyone who challenges the dominant narrative about alcohol tends to be called a “neo-prohibitionist,” let me say clearly that I think alcohol prohibition is a terrible idea, as bad an idea as drug prohibition. Harvard historian Lisa McGirr has spelled out the lasting damage it caused: not just contributing to a steep rise in organized crime, but strengthening the Ku Klux Klan and expanding the right of federal law enforcement to surveil, terrorize, incarcerate, and kill citizens. I’m not a prohibitionist or a neo-prohibitionist; I’m an anti-prohibitionist to the core.

What I want is for our public conversation about alcohol, alcohol policy, and alcohol use disorders to take account of how people actually drink, rather than continuing to assume that most people are moderate drinkers. I don’t think that’s unreasonable.

---

Appendix: Rebuttal of the Trevor Butterworth Critique of Philip Cook and the Washington Post Wonkblog.

This is an essay defending the chart I cited above. Warning: it’s written for people interested in statistics, alcohol policy, and argumentation; others may find it dry. I append it here because it’s relevant to the main post and because Butterworth is occasionally cited as someone who “demolished” Cook’s analysis when he did nothing of the kind.

Half a decade ago, Forbes columnist Trevor Butterworth attacked the Washington Post Wonkblog for uncritically publishing a chart based on statistics from Philip J. Cook’s Paying the Tab.  The chart shows alcohol consumption patterns for US adults, patterns based on survey data from the National Epidemiologic Survey on Alcohol and Related Conditions. Butterworth’s principal objection is Cook’s use of an across-the-board multiplier of 1.97 to account for the difference between the amount of alcohol people claimed to consume and the amount of alcohol sold, which was nearly twice as high as the self-reports predicted. Though every alcohol researcher knows that people wildly under-report their alcohol intake and sales figures are a logical way to adjust self-reports, Butterworth believes a multiplier of 1.97 is too high. So he attacks the figure, not with evidence, but with his own intuition that it just can’t be right.

Do people drink nearly twice as much as they claim? Sure–and then some, if a couple of recent films are any indication. In the film, The Truth about Alcohol, now on Netflix, host Javid Abdelmoneim, a physician specializing in emergency medicine dons a BAC monitor for a week to track his drinking, and he estimates his consumption at less than half what it turns out to be. In the HBO film, Risky Drinking, a young woman claims to consume five drinks per night on the town but actually downs fourteen, her score kept by a digital tracker in the corner of the screen. In other words, there’s nothing outlandish about a multiplier of 1.97, so the onus is on Butterworth to show that Cook’s calculations are wrong.

He doesn’t. He just calls it “a stretch” to imagine that light drinkers could under-report by nearly 50 percent. Then he speculates that many who claim not to drink at all are likely fibbing, even though we know from other sources that the US has lots of teetotalers. For example, almost 45 percent of Latinxs avoid alcohol, as do a majority of Evangelicals of all ethnicities, and those groups are 17 and 25 percent of the US population respectively. And there are smaller communities that avoid alcohol, as well as individuals who’ve given it up for one reason or another. In challenging the honesty of abstainers, Butterworth projects the assumptions of his own class: that everyone drinks and that not drinking is so unnatural and so difficult that teetotalers are bound to “cheat.” In other words, though Butterworth slams Cook’s chart for everything it “requires us to believe,” his own critique “requires us to believe” that Cook’s data are wrong without any evidence except his own incredulity.

That said, Butterworth may have a couple of legitimate, though extremely minor, objections. First, he complains that underage drinkers are missing from Cook’s chart, which is based on an 18-and-up data set. If Cook didn’t take younger drinkers into account when calculating his multiplier, then, yes, that’s a problem because they consume enough alcohol to reduce that multiplier by 11 percent. But Butterworth’s phrasing is so coy that it’s impossible to tell whether he’s objecting to a multiplier error or just pointing out that the chart doesn’t include underage drinkers, which is not a shortcoming. Because so much underage drinking is episodic binge drinking, counting a teen’s drinks per week is not as useful as counting an adult’s. It makes sense to talk about legal and underage drinking patterns separately, which is what Cook does, quite exhaustively, in his book.

Butterworth’s second legitimate-but-minor objection is that Cook’s calculations do not reflect alcohol wasted after sale: e.g. spilled or unfinished drinks. Pointing to the 30-40 percent of food-calories wasted, he invites readers to imagine a comparable percentage of alcohol waste despite the fact that most wasted food is food that spoils. Most alcoholic beverages do not spoil until they’re decanted, so the percentage wasted is far smaller. I doubt there’s any way to estimate that percentage, but I’m willing to give Butterworth the benefit of the doubt and say that Cook’s overall totals are probably very slightly high, except where they are zero.

Butterworth may not deserve the benefit of the doubt, however. To point out how “inflated” Cook’s totals are, he compares them to other studies that show lower consumption rates at the hard-drinking end of the spectrum. But these studies are all . . . wait for it . . . self-reports not adjusted in some way (e.g. alcohol sales data) to compensate for under-reporting. So of course the numbers are smaller! Butterworth scolds Wonkblog for not taking these other studies into account, yet he deserves a scolding himself for comparing adjusted data to self-reported data, which ignores the undisputed fact that Americans dramatically under-report their alcohol consumption.

Then he misrepresents one of Cook’s most startling conclusions.

Cook, it turns out, is . . . an advocate for high taxes on alcohol, and the Wonkblog piece ends with his claim that a policy curbing the avidity of those in the tenth decile such that they’d consume no more than the ninth would cause alcohol sales to drop by 60 percent.

First, Cook does advocate higher alcohol taxes, so that part is true. The misrepresentation is in suggesting that Cook advocates a policy to make the tenth decile drink like the ninth. Cook doesn’t do anything of the sort; in fact, he jokes that only “rigorously enforced rationing” could produce such a change. His point is merely to emphasize the huge jump in consumption from the ninth to the tenth decile, a point the Wonkblog piece underscores. In suggesting otherwise, Butterworth hints that Cook’s chart points toward extreme and unreasonable policy proposals and so should be dismissed.

Despite his “gotcha” rhetoric, the only criticism Butterworth lands is very minor: the suggestion  that Cook’s multiplier may have been a trifle high, though not nearly high enough to invalidate or even significantly modify his conclusions. Nonetheless, Butterworth’s attack earned him lots of attention, sending Cook back to the drawing board, where he revised his figures to say that the consumption of the top ten percent of drinkers may be closer to nine drinks a day than ten. Of course, all his other totals had to come down, too, so his essential point: that the top ten percent of the population drinks nearly three-fourths of the alcohol, remains unchanged.

So why the “gotcha” rhetoric? Why the specious reasoning? Why depend so heavily on innuendo, rather than data? Some answers to those questions lie in Butterworth’s other work.

Butterworth heads an organization called Sense About Science US, which claims to cut through conflicting data about health and the environment but actually promotes a pro-industry, anti-regulatory agenda funded by donors such as the Koch brothers. A 2016 article in The Intercept reveals that, while Sense About Science does occasional work in the public interest to justify its claim of scientific objectivity, the organization spends most of its time and energy attacking the scientific consensus on issues such as the dangers of asbestos or e-cigarettes. The group, which has roots in the tobacco PR “wars,” works on behalf of industries to undermine independent research that could alarm consumers about their products and damage their bottom line.

Butterworth himself has defended BPA, fracking, poisons in toiletries, sugary sodas, Oxycontin, and other products about which scientists have sounded alarms. Well-connected and prolific, he’s adept at pitting weak, industry-sponsored research against strong evidence to create the kind of doubt that fuels climate change denial, all the while concealing his industry affiliations and insisting he’s simply a disinterested watchdog. As we know from the Forbes article, his work appears “above the line” in major periodicals. According to a report in the Milwaukee Journal Sentinel, it also appears “below the line,” as he scours the internet for articles and blog posts on which he can comment.[1]

I have said before that the alcohol industry desperately wants to conceal basic facts about the way people drink. They don’t want it generally known how much of their product is consumed by people who drink too much, and Cook’s work, especially as interpreted by the Wonkblog chart, makes those consumption patterns crystal clear. It was probably inevitable that some hack would attack the work; I’m just sorry people didn’t read and analyze Butterworth’s piece more carefully, because it’s not even a very good hatchet job. It landed well only because it supports familiar, comforting myths about drinking.

---

[1] Information taken from “Trevor Butterworth, Sense About Science, and STATS Spin Science for Industry” on the U.S. Right to Know web site.