Education and Growth in a panel of 21 OECD countries

1
Education and Growth in a panel of 21 OECD
countries

• Clotilde LAngevin and Nadine Laïb
• Conference on Medium-Term Economic
• Assessment
• September 2005

2
Introduction

• Education at the center of the debate
  surrounding aggregate productivity in OECD
  countries
• Econometric link between human capital and growth
  in OECD countries?
• Are macroeconomic returns to education
  significant and positive?

3
I) Effects of education on growth a review of
economic theory (1)

• Returns to education
• Microeconomic
• Macroeconomic
• The neo-classical Solow model
• Constant returns to scale
• Two production factors (labor and capital)
• Returns to each factor are decreasing

4
I) Effects of education on growth a review of
economic theory (2)

• Endogenous growth models
• Non-decreasing returns to scale for an
  accumulated factor (Nelson-Phelps 1966, Romer
  1990a)
• But many empirical studies say returns to capital
  are decreasing
• Possible to include human capital in a
  neo-classical Solow model framework
• Extended neo-classical model
• Mankiw, Romer and Weil (1992) human capital is
  an additional accumulated production factor, but
  returns are decreasing

5
I) Effects of education on growth a review of
economic theory (3)

• MRW Aggregate neo-classical Cobb-Douglas
  production function
• Three input factors K, L, H.
• Harrod-neutral technical progress A, exogenous.
• 
  with
• In the LR, income per capita depends on the
  savings rate, the level of human capital, the
  growth rate of population and exogenous
  technological progress.

6
II) The theoretical model (1)

• Medium term relationship the MRW model
• Measuring human capital
• Mincerian (1974) specification
• Measuring physical capital
• Collinearity between S and ln(k) can appear when
  using stock of physical capital
• We replace physical capital by a simple function
  of the investment rate

7
II) The theoretical model (2)

• Measuring physical capital (continued)
• We proxy physical capital by its steady-state
  expression
• The final medium-term equation

8
II) The theoretical model (3)

• The final short-term equation EC Form
• Different transitional dynamics for each country
  of the panel (Bassanini Scarpetta 2001)
• Hypothesis the medium-term production function
  is identical for the 21 countries
• Plausible for OECD countries
• Pooled-Mean Group estimator

9
III) The data (1)

• Panel of 21 OECD countries, 1971-1998
• Sources online OECD databases
• Human capital
• De la Fuente and Domenech (2000) and subsequent
  versions of OECD Education at a glance complete
  annual data
• Physical capital ( )
• 0.10

10
III) The data (2)
11
IV) Empirical results (1)

• Unit root tests and cointegration
• Limited number of observations (27)
• However, the unit root tests tend to indicate
  that
• are
  I(1)
• Johansen method suggests a cointegrating vector
  of rank 1
• We must check for stationarity of the long-term
  relation in the ECM form

12
IV) Empirical results (2)

• Estimation with PMG specification
• Coefficients are stable
• Interpretation
• ?0.26 (not sign. different from theoretical
  value)
• Macroeconomic returns to education 7
• -gt returns to enlarged capital
  around 0.9

13
IV) Empirical results (3)

• Unit root tests on the medium-term levels eq.

14
IV) Empirical results (4)

• Endogeneity of human capital
• Instrument for human capital initial value of S
  linear trend
• Estimation results
• Coefficients are stable

15
V) Example for a single country France (1)

• The data

16
V) Example for a single country France (2)

• Eliminating multicollinearity between S and trend
  -gt trend constrained
• Estimation results
• Residual of the medium-term equation is
  stationary

17
V) Example for a single country France (3)

• Interpretation
• Endogeneity problems results must be interpreted
  with caution
• Elasticity of total output per capita to physical
  capital not statistically different from that of
  the OECD country mean
• France stronger macroeconomic returns to
  education than the average of OECD countries

18
V) Example for a single country France (4)

• Dynamic contributions

19
Conclusion

• We find significant and positive returns to human
  capital in the medium run for OECD
• The medium-term equation is stationary on average
  for OECD countries
• Reliable coefficient estimates require
  estimations for single countries
• For France, macroeconomic returns to education
  are also significant and positive