Where more women work, people drink more

In countries where more women work for pay, people drink more alcohol. The link is not about money: it holds among countries with the same income. It is a neat example of something statisticians worry about all the time, a third factor that goes with both measures. Here, that factor explains most of the link.

The pattern

Each dot is a country. Across the page is female labour participation, the share of women aged 15 and over who are working or looking for work. Up the page is alcohol consumption, litres of pure alcohol per person aged 15 and over in a year. Across 172 countries the link is +0.41, and +0.45 comparing countries with similar income.

Every figure here is r: how closely two measures rise and fall together, from −1 (one goes up as the other goes down) to +1 (they rise together). 0 means no link.

The third factor

Religion is the obvious candidate. Islam forbids drinking, and among countries of similar income, a larger Muslim share goes with less alcohol (−0.56) and with fewer women in paid work (−0.55). A third factor like this, one that goes with both measures, can make two things look linked when neither has anything to do with the other.

So we allow for the share of people who are Muslim. The link falls from +0.45 to +0.20. Only about 45% of it is left. Comparing countries only within their own region as well leaves +0.09, close to no link at all.

Inside each region, the answer is mixed

Look inside each region, comparing countries of similar income, and the link is positive everywhere. Then allow for religion too, region by region. In some regions the link mostly goes: in Europe & Central Asia (+0.20 becomes +0.09) and Middle East, North Africa, Afghanistan & Pakistan (+0.43 becomes −0.20). In others a link of about +0.29 remains, as in Sub-Saharan Africa, Latin America & Caribbean and East Asia & Pacific. So religion explains most of the worldwide link, but not all of it everywhere.

Region by region, comparing countries of similar income:

The same, after also allowing for the share of people who are Muslim:

The lesson

A link that survives the obvious check, here wealth, is not proven. The question is always what else differs between the places being compared. Within countries over the years there is no link either: as women’s work rose or fell, drinking did not move with it (−0.03 across 175 countries).

What this cannot tell you

• It describes countries, not people. It says nothing about whether working women drink more.
• Alcohol figures include an estimate of drinking that is never recorded, which is rougher in some countries than others. The series ends in 2020.
• A link is not proof that one thing changes the other.

See the full figures for female labour participation vs alcohol consumption.

More articles

• Countries where people have a say collect more tax
• The puzzle of women managers and homicide
• In tight-knit counties, fewer people move in or out
• Friendships across income lines, and how far poor children climb
• US Murder (Homicide) Rate: What Drives It?
• Where the rich move

These articles describe places, not people, and show links, not causes. Every article on PlaceStats.