Is a Glass of Wine Harmless? Wrong Question.
The latest alcohol advice ignores the value of pleasure.
This article was featured in One Story to Read Today, a newsletter in which our editors recommend a single must-read from The Atlantic, Monday through Friday. Sign up for it here.
Not so long ago, alcohol was good for you. In a 1991 segment on 60 Minutes, a French researcher claimed that red-wine consumption was responsible for good health in France. This argument proved popular with the wine-consuming public, and prompted academic papers positing an inverse relationship between red-wine consumption and cardiovascular disease. Scientists even put forward a mechanistic theory for why red wine was healthy, involving a compound called resveratrol.
Others soon cast doubt on the possibility that red wine could really have any causal relationship with longevity. The “alcohol is good for you” narrative eroded and, in the past year, seems to have fully collapsed. A number of researchers are now arguing that basically any amount of alcohol is bad for you; a New York Times article from January was titled “Even a Little Alcohol Can Harm Your Health.” Some—including the Canadian government—are suggesting that, as a result, the safest choice is not to drink at all.
Excessive alcohol consumption clearly leads to significant problems, physical and emotional. That is not up for debate. However: Recent rhetoric, veering in the direction of abstinence, goes well beyond the sound advice to avoid heavy drinking and ignores the value of pleasure.
A pleasure-agnostic approach to health advice is now in vogue even outside the domain of alcohol, and is filtering down to the general public with sometimes absurd results. Recently, a reader asked me: Is there any data on health benefits to orgasms? I am not aware of reliable data from randomized experiments suggesting that having more orgasms improves health. That isn’t the point of orgasms, anyway. The point of orgasms is that they are fun. We do not need to prove health benefits to want to have them.
Public-health advice is sometimes based on a “lexicographic” standard—putting the effects on health first, second, and third, and ignoring other considerations, including enjoyment. A lexicographic standard applied to, say, meat consumption would hold that we must always eat burgers well done, because that is the best way to avoid any risk of E. coli, even though well-done burgers are tasteless. More generally, some in public health avoid discussing the negative unintended consequences of absolutism. During the coronavirus pandemic, some officials advocated strongly for long-lasting school closures, arguing that keeping kids at home was the only way to prevent in-school spread among students and teachers. That was, in a technical sense, true, but this recommendation failed to consider the enormous costs to children of those closures, which should have been weighed against any benefits.
Coming back to alcohol, pleasure-agnosticism could make sense if the best available evidence indicated substantial harm from even moderate drinking. It does not. I should also stress that the data are fundamentally flawed because the largest, most commonly cited studies we have are observational, not randomized. And the characteristics of people who consume alcohol in moderation are different from those who do not.
In 2018, The Lancet published a comprehensive study on the link between alcohol consumption and cancer, heart disease, and other illnesses. It is an extraordinary work of scholarship, combining hundreds of previous papers. And the results indicate an upward trend in cancer, in particular, as alcohol consumption increases. But the effects at moderate levels of drinking—say, one to two drinks a day—are very small. For heart disease, we see the familiar decrease in risk at moderate drinking levels, and an increase with higher amounts.
None of these results are convincingly causal. It seems very likely that all associations—positive and negative—are overstated relative to the truth. Generally, when researchers are able to adjust for some demographic differences, the relationship between alcohol and health gets smaller. This, in turns, suggests that if they could adjust for more differences, it would get smaller still. Whether these relationships would be smaller but still positive, or really zero, is something we cannot know from the data we have.
We cannot conclusively prove that moderate alcohol consumption is totally benign, much less beneficial. Based on the data we have, it also seems extremely unlikely that moderate alcohol consumption is fully “bad” for your health.
If you do not enjoy, or actively dislike, alcohol, then the abstinence standard might be the right one for you. But many people do enjoy a drink from time to time: a beer with friends, a cold glass of rosé in the summer, a hot toddy in front of the fire, even just a glass of white wine while cooking at the end of a long day. If we accept that pleasure has value, and that the data are muddy, then the moderation standard makes more sense.
The pendulum on alcohol has swung too far from the 1990s. Alcohol is probably not the key to longevity. But it’s not arsenic, either. In the immortal words of Cookie Monster, it’s a sometime food.