Presentation is loading. Please wait.

Presentation is loading. Please wait.

Comments on “Human capital and the wealth of nations” by R. Manuelli and A. Seshadri  What the paper is about: Accounting of output per worker Novelty:

Similar presentations


Presentation on theme: "Comments on “Human capital and the wealth of nations” by R. Manuelli and A. Seshadri  What the paper is about: Accounting of output per worker Novelty:"— Presentation transcript:

1 Comments on “Human capital and the wealth of nations” by R. Manuelli and A. Seshadri  What the paper is about: Accounting of output per worker Novelty: quality of education is taken into account, at least conceptually. Main finding: most of cross-country income differences due to factor accumulation, not TFP.

2  Details: y = z k  h 1-  y = output,k = physical capital, h = human capital,z = TFP In the paper:h = h(s, investment) In Hall and Jones :h = e rs Lowest quintile has TFP equal to 73% of the US level Hall and Jones estimate it at 20% (?) Interpretation: ignoring differences in quality of education amplifies differences in TFP.

3  Alternative interpretations: 1.Differences in human capital across countries are exaggerated: –Schooling quantity and quality (and thus h) are estimated from calibration. –Top/bottom quintile quantity : 20% higher in calibration than data –Top/bottom quintile quality: almost 40% higher in calibration 2.Quality is not properly measured: –Proxy is public spending in schooling per pupil/GDP per worker –It ignores private spending –Does a higher ratio really mean better quality? 3.Different PWT databases. Does it matter?

4  Two more caveats: Calibration for the US around 2000 from steady state implications of the model. –Worst year to assume steady state in the US (period of abnormally high growth rates – “new economy”) –Estimates of human capital in the rest of the world also based on steady state assumption. –In general, is a country ever in steady state? Role of h is inflated because of its endogenous response to TFP changes. –In equilibrium h is ultimately determined by TFP (through wages) and life expectancy

5  Further research I: the “Becker Paradox” Convergence in life expectancy but not in years of schooling. What can the model say on this? Life expectancy at age 5 (L5) Years of schooling (s) ∆s /∆L5 1960199019601990 Rich countries72.476.87.810.6.65 Middle & low- income countries (ex SSA) 62.870.33.26.1.39 Sub-Saharan Africa (SSA) 54.459.81.33.1.33

6  Further research II: Macro-Mincer return Better education quality implies a higher return on schooling, ceteris paribus. For each country r = r a +  where r a = average world return;  = deviation from average Standard growth regression:  y = c +  k + r  s + e   y = c +  k + r a  s + , where  =  s +e So omitting schooling quality from growth regressions would bias the estimated r a up. In practice it is 0. Can the model explain this?


Download ppt "Comments on “Human capital and the wealth of nations” by R. Manuelli and A. Seshadri  What the paper is about: Accounting of output per worker Novelty:"

Similar presentations


Ads by Google