Thursday, September 14, 2006

Social Drinking and Social Capital

A new study by Ed Stringham and Bethany Peters shows that men who drink alcohol earn 10% more than those who don't; women who drink earn 14% more than those who don't. Also, men (but oddly, not women) who visit a bar at least once a month earn an additional 7%. Here's Stringham's explanation, which makes sense to me:
"Social drinking builds social capital," said Stringham, an economics professor at San Jose State University. "Social drinkers are networking, building relationships, and adding contacts to their Blackberries that result in bigger paychecks."
A few reactions. First, one implication is that taxes on alcohol could be taxes on the formation of social capital -- which can have deleterious effects on both individual and overall productivity. As the article says, the study "suggests that the growing wave of anti-alcohol legislation at state and local levels will have harmful effects on local economies and individual finances." A possible response to this claim is that the social-drinking boost results not from the alcohol per se, but from the social context in which people drink (and are sometimes expected to do so). So if taxes succeeded in reducing or eliminating alcohol consumption, the social contexts might persist -- or re-form in other locations -- and continue to do their job. That may be partially true, but I suspect alcohol itself performs a crucial "lubricating" function, allowing people to let down their guard, meet more people, and get to know each other better. Alcohol-free social contexts would therefore generate lower levels of social capital.

Second, while the 7% earnings boost associated with visiting a bar at least once a month is good evidence for the social capital hypothesis, it also makes the 10-14% boost from merely drinking somewhat mysterious. The bar-visiting variable is supposed to control for the effect of drinking in social contexts. But perhaps some of the relevant drinking takes place in other locations, such as non-bar restaurants and workplace lounges; the boost from such drinking would be captured by the drinking-at-all variable. Or perhaps a more precise variable measuring the frequency of bar visits (rather than the dummy varible used by the study) would capture more of the social capital formation effect, leaving a smaller effect from merely drinking.

Third, I wonder why the bar-visiting boost applies only to women. The paper doesn't speculate much about this, saying only, "Perhaps, women increase social capital apart from drinking in bars." This might dovetail with my drinking-in-other-locations speculation above.

Fourth, I wonder if differences in drinking behavior might result in different earnings for men and women. The famed "74 cents on the dollar" male-female earnings gap is flawed, because it doesn't take account of all the factors that differ between men and women (such as years of education and labor force loyalty). But even once these explanatory factors are included in regressions, a smaller gender gap remains. If men drink more than women, that could be another factor that helps explain the gap.

1 comment:

Perry E. Metzger said...

I find the whole thing suspicious.

There are serious questions about whether most "moderate drinking brings longevity" studies have properly controlled for people with serious illnesses who can not drink but who are part of the study group. Such people skew the results towards greater mortality for non-drinkers.

Similarly, this study might not be controlling for factors like non-drinking due to illness, non-drinking due to alchoholism (which might have cost a person a few years of capital formation etc.), non-drinking due to membership in ascetic religious sects (which might reduce one's "worldly orientation"), etc., all of which could skew the results towards lower income for non-drinkers.

Generally speaking, such studies suggest immediate conclusions, like "social drinking improves income", but these conclusions can easily be the product of poor study design.