The Power of Positive Drinking on Health and IncomeGail Heriot
I‘m not much of a drinker. I don’t think I’ve had a drink since June or so, and I seldom have more than one or two in a month. Why? Well, for one thing, I can’t hold my liquor. After one drink, I can start to get affectionate. After two, I’m pinching men’s bottoms. It’s not always a pretty sight.
I’m told there’s a lot of evidence these days that moderate (or even immoderate) drinking is good for your health. Maybe. But I’ve always been skeptical. It’s that old chain of causation thing. How could you tell whether drinking is leading to good health or good health leading to drinking? I was therefore interested to learn of a UC-San Francisco sociologist’s research (as reported in the San Francisco Chronicle) that seems to credit the latter possibility.
Of the 54 studies she and her colleagues reviewed, all but seven grouped former drinkers with abstainers (and not with drinkers). This is probably error. Former drinkers have often given up drinking precisely because they are ill. If you group them with abstainers, you are unfairly weighing down the abstainers with a group that is disproportionately in ill health (when there is no reason to believe that their illness was caused by their decision to stop drinking and, in addition, there is no reason to conclude that it wasn’t caused by their earlier drinking). And by leaving them out of the drinker category, you’re making the drinker category look healthier than it really is.
In the seven studies that grouped former drinkers with drinkers, no health differences were detected between abstainers and drinkers–or at least so the UCSF sociologist reported. Maybe there are flaws with these studies too. But I don’t think I’m ready to change my habits just yet.
Meanwhile, as my fellow Right Coaster Tom Smith reported, last week, a much more surprising hypothesis (at least to my mind) was being advanced–that drinking leads to higher income. According to economists Bethany Peters and Edward Stringham:
[W]e find that self-reported drinkers earn 10-14 percent more than abstainers, which replicates results from other data sets. We then attempt to differentiate between social and nonsocial drinking by comparing the earnings of those who frequent bars at least once per month and those who do not. We find that males who frequent bars at least once per month earn an additional 7 percent on top of the 10 percent drinkers’ premium.
(For some reason, women who frequent bars did not enjoy that additional premium. Go figure.)
Can any of this be so? Of course. But a lot of the reason may be that higher income leads to more drinking rather than the other way around. People with higher earnings can have a drink every day plus pay the rent. Consequently, some do both. Those who are strapped for cash must be more careful.
In their Reason Foundation Policy Brief, Peters & Stringham are interested in a different explanation–“We hypothesize that drinking enhances social capital, which leads to superior market outcomes.” Put differently, drinkers, particularly those who frequent bars, have a wider circle of friends and perhaps better social skills than others. And that could add up to higher income. Peters & Stringham attempt to support their hypothesis by controlling for age, race, religion, education, marital status, parental education, number of siblings, and region of residence. When they did so, they found that both drinking (men and women) and drinking in bars (men only) are associated with higher income. And these two factors are to some degree independent.
I continue to be skeptical about any causal connection. Here are some of my somewhat random thoughts:
1. Again, healthy and robust people are more likely to drink than not-so-healthy people. And they’re more likely to earn more too (even if you’re comparing only full-time employees). This will be so regardless of whether their condition is related to their drinking, so until it can be proven that drinking causes good health, the notion that drinking causes higher earning should be taken with a grain of that stuff found on a margarita glass.
2. High earners are more likely to be encouraged to entertain their clients by buying them drinks than low earners. In some fields, they will have expense accounts precisely for that purpose. But it is the high-end job that is causing the drinking (and the visit to the bar) and not the other way around. For exactly that reason, I used to go to bars more when I practiced law (and earned a higher salary) than I did after I left that job.
3. Once you control for skills learned through education, one of the most important job skills is sociability. It’s crucial in marketing and management, where a gregarious nature is a real asset. But drinking may not cause the kind sociability that increases income (i.e. not bottom pinching) so much as result from it. There are lots of ways to be sociable–active membership in a civic organization, for example. Had Peters & Stringham controlled for some of them, the positive contribution of drinking and frequenting bars might have been reduced or dropped out.
4. There are some delicate issues of religion and social class here too. Peters and Stringham rightly attempt to control for religion. But by dividing the population into only five groups–Catholic, Protestant, Jewish, No Religion (which fared poorly) and Other Religion–they fail to get at the crux of it. If religion/social class is part of the explanation here, that explanation may be obscured by lumping all Protestants together. Some Protestant denominations (e.g. Episcopal) have members who are disproportionately middle- to high-income and encourage (or at least fail to discourage) drinking. Others (e.g. Southern Baptist and Jehovah’s Witnesses) have members that are disproportionately middle- to lower-income and discourage (and sometimes even forbid) drinking. The reason is overwhelmingly unlikely to be that Episcopalians drank their way into affluence (or that Southern Baptists and Jehovah’s Witnesses abstained their way into comparative want. The reasons are historical, cultural and very complex. Insofar as there is a causal link, it is apt to be in the reverse: People who come from higher social strata have a social safety net that to some degree protects them when they drink too much; people from lower social strata may be more at risk. Their differing moral codes reflect this. Lumping them together as “Protestants” obscures this complexity.
Given all these problems and uncertainties, I thought Peters & Stringham were just a wee bit premature in talking about the “policy implications” of their study. When all this is sorted out, there may be none. I found the Reason Foundation’s press release that accompanied the policy brief a bit overly aggressive:
“We’re quick to ban beer at sports stadiums and festivals. The legal blood alcohol level is dropping everywhere, and we’re barraged with overhyped warnings about binge and underage drinking,” Stringham said. “Instead of fear mongering we should step back and acknowledge the proven health and economic benefits that come with the responsible use of alcohol.”
Or not, guys.