Is human capital capital?

2/23/2015 06:22:00 PM
The Harvard Mark II computer Wassily Leontief used for his input-output models.
This is a question that seems to have taken the econosphere by storm.

Of course, human capital is not capital. If it were, we'd just call it capital and lump it in with the rest--the purpose of the term "human capital" is to distinguish it from actual capital.

So, if it's not capital, what is human capital? At most, I'd say it's a modeling insight. The vague notion of "skills" and knowledge that it stands in for are very real, but when it comes to modelling we are really just pretending it's a kind of capital stock. The key insight was in recognizing that skills and knowledge share a few of the same kinds of intertemporal dynamics as capital, and that we can obtain accurate predictions by modeling them as such, even though they differ in lots other ways.

One of the earliest modelling successes of the theory of human capital happened in international trade theory. In 1949, economist Wassily Leontief plugged some imported Soviet input-output models into a computer with data from the BLS and let it chug. When the computer solved the model, the results it spit out were mind-boggling--the US was actually a net exporter in labor-intensive goods, not capital-intensive ones. That was pretty shocking because in 1949, shortly after world war II had destroyed half the world, the US owned essentially all the world's capital, having only a small portion of it's population. Why wasn't the US exporting capital-intensive goods as theory predicted?

The answer, it turns out, is human capital. Leontief's model hadn't incorporated any notion of cumulative job skills or knowledge, and therefore missed the fact that Americans in 1949 actually were exporting the things they had a comparative advantage in--goods that were human-capital intensive--despite the relative dearth of physical capital anywhere else on earth.* Failing to model human capital as a kind of capital stock, it turns out, produces incorrect predictions.

Reading this debate on the econosphere, you'd be forgiven for thinking that this kind of human capital--essentially just job skills--is the only kind that exists. Noah Smith, for example, asserts "creating value from... [human capital] requires the owner to sacrifice some leisure." That's not true of all kinds of human capital! I'm a little disappointed that no one brought up any of the myriad other kinds of human capital that exist (or rather, the myriad other kinds of non-capital which exhibit capital-like intertemporal dynamics), such as health capital.

The concept of health capital originated with Michael Grossman's seminal 1972 paper in which he postulated that an individual's life-cycle of spending on healthcare could be modeled as investments into a capital stock. This, for the first time, captures many of the real-world aspects of health behavior and health spending that static models can't grasp. We knew, empirically, that people's health and longevity depends in part on the cumulative choices they make over the course of their lifetime, such as whether or not to smoke, what kinds of food to eat, how much time to spend exercising, and even how much to spend on healthcare. Each of these choices involves an intertemporal tradeoff--that is, a choice of how much to "invest" into maintaining their stock of health capital. Extensions of Grossman's model can explain all kinds of health behaviors from addictions to yo-yo dieting to trends in national health expenditures.

Do humans literally carry bags full of health capital around with them through life? No. Health capital is not capital. As economists we are merely pretending that the vague notion of healthiness is a kind of capital. But because of the similarity in intertemporal tradeoffs, modeling health choices this way yields many correct predictions, and resolutions to old theoretical puzzles we would not otherwise be equipped to solve.

*This account comes ultimately from Leontief, by word of mouth. He was one of my professor's professor. I do not, however, claim familiarity with the academic literature.