Blogging before morning coffee = elementary mistakes

Yglesias looks at a graph that shows that countries that spend more on social services (health, education, welfare) tend to have lower income inequality. He concludes:

[L]et it be said again—you can reduce inequality by having your government spend more on social services[.]

Um, no. Correlation does not imply causation. If I observe that roosters tend to crow at around the same time that the sun rises, I should not conclude that I can make the sun rise earlier by making the rooster crow in the middle of the night. Basic logical fallacy.

The fact is, the Gini coefficient measures pre-tax income. There's no obvious reason why either high taxes or high social spending should affect the distribution of pre-tax income. The point of those policies, in fact, is to equalize after-tax income. So Yglesias' conclusion - that you can raise pre-tax income equality by raising social spending - has no theoretical basis.

A much more likely explanation is that the same thing that causes high social spending also causes low pre-tax inequality. A society that values income equality will both tend to A) pay people more equally, and B) spend more on social services. But this doesn't mean (B) causes (A).

In other words, raising social spending might be a good thing to do (and I think it is, right now at least), but don't expect it to get rid of the "inequality" we read about in the newspapers.

Update: Actually, I discovered that the "inequality" measured by the graph to which Yglesias refers is wealth inequality, not income inequality. Although this doesn't negate my argument - correlation still doesn't imply causation - it's a lot more plausible that increased social spending could increase wealth equality. Better education and health would do more for people's ability to accumulate wealth over time than for rich people's. So Yglesias really isn't on as much of a limb as I thought...

Just one more reason why blogging before morning coffee = elementary mistakes. ;-)