If there's one thing that economists agree on, it's that the United States needs more "human capital." That's jargon for "people with skills." Just as buying physical capital—like computers, robots or factory equipment—lets companies produce more goods and services with less labor, hiring people with skills or specialized knowledge that makes them more productive helps companies produce more without increasing the man-hours spent producing it.

There's a lot of evidence to back this up. Results from a randomized study showed that a better kindergarten education leads to higher lifetime earnings, largely because those who received a good education early on are more likely to have the mental, social and other skills necessary to be productive workers. Similarly, there's a huge wage premium for college, which is likely attributable at least in part to skills learned there. And there's evidence that letting in more high-skilled immigrants reduces poverty for native-born workers (and, obviously, improves life for those immigrants considerably).

Now, economist William Yu at UCLA provides another data point. In a study for First 5 LA's Early Childhood Education conference set to be released in a few days, he constructed a "city human capital index" or CHCI that measures the average amount of schooling attained in a given city, county or metropolitan area, as recorded by the Census Bureau. For children and young adults under 25, Yu measures how much education they are likely to receive, based on elementary and secondary school enrollment rates in the area and on college enrollment, for those over 18. Each year of schooling gets 10 points.

Of major cities, D.C. comes out way ahead:

It's the same thing for metropolitan areas: the D.C. metro area is number one, with a score of 140.5 (or 14 school years, two years of college), followed by Boston and San Francisco, with the New York and Los Angeles metro areas far behind. The D.C. suburbs in particular fare very well. Falls Church, Va. is the best-educated independent city or county in the country, with a score of 154.7. Arlington County is third and Howard County, Md. is fourth, with Los Alamos in New Mexico at second.

There are some surprises in the data. Both as cities and metro areas, St. Louis and Kansas City are both more educated than New York and Chicago, and Cincinnati, Detroit, Cleveland and Tampa all outperform Los Angeles. The overall pattern is that the coasts, the plains and the Midwest all perform quite well, with the South notably less educated than the rest of the country:

Yu finds that the CHCI correlates very well with per-capita income at the county and city levels.

The correlation with unemployment rates is rougher, but still statistically significant. These figures alone don't show causation. It could be that more educated people just flock to richer places and don't themselves make those places rich. But given the evidence showing that higher earnings for educated workers is due to demand for their skills, it's not too big a leap to say that more education leads to richer cities or counties.

It's always good to take new metrics like this with a grain of salt. For one thing, CHCI falters as a measure of human capital because it fails to incorporate migration patterns. Regions like Silicon Valley with a high proportion of skilled immigrants likely have more human capital than this metric would pick up. But the simplicity of the metric is also a virtue. It's basically a collation of solid Census data on educational attainment, and is likely to be reliable on account of that source.

More to the point, it confirms what we already knew. More educated places tend to be richer ones. And D.C. is both very educated and quite rich compared to the rest of the country.