Arnold Kling  

Education and Growth

A Footnote to the Economics of... Economic Therapy...

Ed Glaeser writes,

no variable from 1900 better explains success in 2000 than investment in education.

That is, if you had a high percentage of the school-age population enrolled in school in 1900, you would have a high GDP per capita in 2000. Given that 100 year time difference, I think it is safe to rule out pure reverse causality.

Comments and Sharing

COMMENTS (5 to date)
Bob Murphy writes:

Holy cow! Did those people get 38 consecutive PhDs?!

Greg Ransom writes:

The high school dropout rate in LA Unified is well over 50% .....

Stil writes:

You could pick any variable that sets western Europe and America 100 years ago apart from the rest of the world, and that should always be very 'predictive' of current economic outcomes.

If you can squeeze Japan in as well, the regression line should be nice enough to make baby Jesus smile.

But yes, having a functional educational system is probably a Good Thing (TM). What is interesting these days on the other hand is how effective investment in education is on the margin.

ajb writes:

Of course, Glaeser's problem is that Japan is above the regression line -- i.e. its performance is better than its initial standing in terms of human capital from one hundred years ago. In fact, I would guess that most of the countries that have done better than Argentina -- say, Taiwan, So. Korea or Singapore -- were below Argentina in terms of initial human capital and probably income in 1950 or 1955 yet have outperformed since then. Moreoever Italy and Greece were comparable to Argentina and Chile and Portugal were worse in terms of human capital. Yet all have outperformed Argentina from Glaeser's own data.

Moreover, the full paper showed Argentina rapidly converging to the US in terms of basic education, enough that it should have spurred some equivalent in growth.

Arguing from the average is contaminated by the sheer number of failures of performance in Latin America relative to successful nations.

Floccina writes:

Even with a 100 year period I would bet that correlation is not causation in this case. Stil, and ajb make good points. There are many characteristics that would go along with investing in schooling that could explain the outcomes. Also I think that generally the richer people are the higher the percent of their income they spend on education.

It would be interesting to see Cuba added to those graphs.

Comments for this entry have been closed
Return to top