Measuring Human Capital Is Both A Simple And Very Tricky Thing


Human capital is vastly important, for it’s by far the largest portion of total capital. It’s also by far the largest determinant of how rich the future is going to be. However, it does always fall foul of Goodhart’s Law. As soon as you’ve identified some measure then that very measure ceases to be a good guide to policy.

For example, historically more education has boosted growth. The rise of Grievance Studies could well have thrown that into reverse. Or with a little less snark, for a UK male an arts degree reduces his lifetime earnings, not increases. Sure, that’s not the same thing as being societally beneficial or not but still, indicative. Thus we cannot say that more education from where we are now is going to boost growth.

At which point a paper in The Lancet:

Background Human capital is recognised as the level of education and health in a population and is considered an
important determinant of economic growth. The World Bank has called for measurement and annual reporting of
human capital to track and motivate investments in health and education and enhance productivity. We aim to provide
a new comprehensive measure of human capital across countries globally.
Methods We generated a period measure of expected human capital, defined for each birth cohort as the expected
years lived from age 20 to 64 years and adjusted for educational attainment, learning or education quality, and
functional health status using rates specific to each time period, age, and sex for 195 countries from 1990 to 2016. We
estimated educational attainment using 2522 censuses and household surveys; we based learning estimates on
1894 tests among school-aged children; and we based functional health status on the prevalence of seven health
conditions, which were taken from the Global Burden of Diseases, Injuries, and Risk Factors Study 2016 (GBD 2016).
Mortality rates specific to location, age, and sex were also taken from GBD 2016.

So, why doesn’t this work? As Tyler Cowen points out:

You will note the first sentence of the paper’s background claims: human capital refers to “the level of education and health in a population”. The first two sentences of the actual paper immediately contradict this: “Human capital refers to the attributes of a population that, along with physical capital such as buildings, equip­ment, and other tangible assets, contribute to economic productivity. Human capital is characterised as the aggregate levels of education, training, skills, and health in a population, affecting the rate at which technologies can be developed, adopted, and employed to increase productivity.” The paper does an OK job of measuring the former, but absolutely fails on the latter.

We can and should include in human capital such things as general attitudes towards, say, entrepreneurship. That the British ruling classes continually denigrated “trade” could have been a good thing – it allowed people to get on with it – could have been a bad. But it was definitely an important thing.

And something most important to note. If you do use these conventional measures of human capital, getting results like this:

Finland is first, the United States is #27, and China and Russia are #44 and #49 respectively.

Capital is what you use to produce stuff. So, is the US using its human capital more or less efficiently than Finland, given that it’s richer? And do we want to be efficient ourselves or not?