Social or human infrastructure? Re-examining essential determinants of output per worker among countries

Abstract

Hall and Jones (1999) presented Social Infrastructure as an essential determinant of differences in Output per Worker among countries. We show that other types of infrastructure may influence Production per Worker even more. We call this type of infrastructure Human Infrastructure.

Notes

¹ Infrastructure determines factor endowments and then income differences. as in HJ (1999).

² We have tested the same relationship, using GDP per worker as an average from 1985 to 1989 – from the PWT 6.1, and conclusions do not change.

³ Survival Rates and Life Expectancy are comprehensive measures for health ‘output’, widely used when assessing the relationship between health, human capital, growth and income, while physicians are a proxy for health ‘input’, less used in the earlier literature.

⁴ A specification that use ln(H_j ) instead of H_j would not change results. They are available upon request.

⁵ The source of these instruments are worldclimate.com and un.org.

⁶ The only difference is that we did not bootstrap the SEs. This does not affect the quantitative or the qualitative significance of the HJ (1999) results.

⁷ It is worth noting that the instruments are not particularly good in regression (3).

Table 2. Social infrastructure vs. Human infrastructure enlarged set of instruments: Kumar (2003) instruments

⁸Results in the table use coefficients from regressions with the HJ set of instruments. The use of the enlarged set of instruments described earlier does not change results, that are available upon request.

⁹ The use of estimates in Table 2 to perform this accounting exercise does not change results. In fact, this would increase the share of Human Infrastructure and decrease the Social Infrastructure in explaining output per capita differences.