ECON2915 Economic Growth Lecture 4 : Human capital. Productivity - PowerPoint PPT Presentation

ECON2915 Economic Growth Lecture 4 : Human capital. Productivity measurement. Andreas Moxnes University of Oslo Fall 2016 1 / 40

Human capital and income So far: Workers assumed to be identical over time and across countries. How can differences in human capital explain cross-country income differences? Human capital: Factors that influence the productivity of the worker, e.g. education & health. Production function with human capital: Y = F ( K , hL ) , where h is effort/quality per worker. 2 / 40

Human capital characteristics 1 A productive input : Increased HC − → increased output. 2 Can be produced (as opposed to e.g. natural resources) 3 Its value depreciates over time. 4 Yields a return, e.g. investment in education increases the wage. 5 Cannot be rented, as opposed to physical capital. 3 / 40

Human captial: Health We’ll focus on two determinants of human capital: Health & education. Better health: Improves productivity - increase output by working more or improving quality. Brings more people into the workforce. 4 / 40

Nutrition vs GDP/capita UK: Better nutrition estimated to raise output by 95% over 200 years (0.3% per year). Output growth over the same period was 1.15%. 5 / 40

Life expectancy vs GDP/capita 6 / 40

Health and income Two-way causality. 7 / 40

An exogenous shift in income A − → B − → C 8 / 40

HC : Education Education & skills also human capital. ◮ Boost productivity & wages ◮ Intrinsic value, higher utility. 9 / 40

Changes in the level of education 1975-2010 10 / 40

Norway, 1970-2014 100% ¡ 10% ¡ 20% ¡ 30% ¡ 40% ¡ 50% ¡ 60% ¡ 70% ¡ 80% ¡ 90% ¡ 0% ¡ 1970 ¡ 1980 ¡ 1981 ¡ 1982 ¡ 1983 ¡ 1984 ¡ 1985 ¡ 1986 ¡ 1987 ¡ 1988 ¡ 1989 ¡ 1990 ¡ 1991 ¡ 1992 ¡ 1993 ¡ 1994 ¡ 1995 ¡ 1996 ¡ 1997 ¡ 1998 ¡ 1999 ¡ 2000 ¡ 2001 ¡ 2002 ¡ 2003 ¡ 2004 ¡ 2005 ¡ 2006 ¡ 2007 ¡ 2008 ¡ 2009 ¡ 2010 ¡ 2011 ¡ 2012 ¡ 2013 ¡ 2014 ¡ Primary ¡ High ¡school ¡ University ¡<= ¡4 ¡years ¡ University ¡> ¡4 ¡years ¡ 11 / 40

Education : Costs and benefits Costs: Tuition. 6.2% of GDP in the U.S. The opportunity cost. Benefits: The return to education. ◮ e.g. % increase in wage for an additional year of schooling. 12 / 40

Returns to schooling Hall and Jones (1999). 13.4% per year for years 1-4, 10.1% for next 4 years, 6.8% for 8+ year. Both developing and developed countries. 13 / 40

Decomposing wages We know that capital’s share of income is around 1/3 (the α ). For the remaining 2/3, how much is due to human capital and how much is “raw labor”. 14 / 40

Share of HC in wages For complete higher, raw labor is 1/4 of wages. For the economy as a whole, the HC share is larger in advanced countries. ◮ Higher wages for more education. ◮ Larger share of population with more education. ◮ HC share 59% in developing, 68% in advanced. 15 / 40

Education and income 16 / 40

Education and income Higher education − → higher income. Higher income − → spend more on education. 3rd factors that affect both income and education. 17 / 40

The Solow model with HC Let’s introduce HC in the Solow model: Y = AK α ( hL ) 1 − α = h 1 − α AK α L 1 − α , where h is effort/quality per worker. 18 / 40

Intensive form Y = h 1 − α AK α L 1 − α or y = h 1 − α AK α L 1 − α = h 1 − α Ak α L Change in capital stock: KL − K ˙ ˙ k = ∂ ( K / L ) L ˙ = ∂ t L 2 ˙ ˙ K L − K L = L L = γ Y − δ K − kn L = γ f ( k ) − ( δ + n ) k (as before). 19 / 40

Steady state Steady state defined by ˙ k = 0: γ h 1 − α Ak α = ( δ + n ) k � A γ h 1 − α � 1 / ( 1 − α ) k ss = δ + n � A γ � 1 / ( 1 − α ) = h δ + n And output y ss = h 1 − α Ak α � � 1 / ( 1 − α ) � α � A γ = h 1 − α A h δ + n � � α / ( 1 − α ) γ = hA 1 / ( 1 − α ) δ + n Compared to Solow model without h ? 20 / 40

Predictions Two countries i and j , with h i and h j , all else equal. Then y ss = h i i y ss h j j Income per capita proportional to HC. E.g. twice as high h in i yields twice as high y . 21 / 40

A numeric example 12 years of schooling in i and 2 in j . What is human capital h in the two countries? ◮ Recall wage increase per year of additional schooling (13.4% for grades 1-4, etc). ◮ Assume human capital h proportional to wages. Then h i = 1 . 134 4 × 1 . 101 4 × 1 . 068 4 × h 0 = 3 . 16 × h 0 h j = 1 . 134 2 × h 0 = 1 . 29 × h 0 And y ss = 3 . 16 h 0 i = 2 . 47 . y ss 1 . 29 h 0 j 22 / 40

Predicted vs actual GDP per worker Calculate h for every country and use the formula to predict y : Positive correlation but too few poor countries in model. Why? 23 / 40

The quality of education It’s not only years of schooling that matters. E.g. student-teacher ratio, teacher quality, access to teaching materials. Test scores capture both quantity and quality. 24 / 40

Externalities 25 / 40

Externalities Externalities: An incidental effect of economic activity for which no compensation is provided. The Solow model could not generate sufficient income inequality coming from human capital. ◮ One reason could be externalities. Education: Additional schooling for individual x − → private return to x but also returns for y . ◮ E.g. x adopts new technologies that y also will use. 25 / 40

Next topic: Productivity measurement Productivity is how efficient we use the factors of production. The Solow model: A . ◮ Knowledge ◮ Organizing production ◮ Effort Over the long run, productivity growth is considered the major factor behind income growth. 26 / 40

Recall 27 / 40

Identification What’s productivity and/or factor accumulation? 28 / 40

Development accounting Can we identify the 3rd case? Consider 2 countries with production function i ( h i L i ) 1 − α Y i = A i K α ⇐ ⇒ y i = A i k α i h 1 − α i Then i h 1 − α k α y i = A i i ⇐ ⇒ k α j h 1 − α y j A j j j h 1 − α k α A i = y i j k α i h 1 − α A j y j i Yes. Intuition: Given same α , we adjust GDP/capita ratio with factor accumulation ratio to infer A ratio. 29 / 40

An example Given α = 1 / 3, we get 2 h 1 − α k α 1 1 / 3 1 2 / 3 A 1 = y 1 = 24 2 27 1 / 3 8 2 / 3 = 2 . 1 h 1 − α k α A 2 y 2 1 1 30 / 40

Cross-country evidence Enormous differences in the A ’s. Where do A differences come from? Measurement issues? 31 / 40

Productivity or factor accumulation? 32 / 40

Productivity or factor accumulation? Both - surprisingly similar plots. 33 / 40

Measurement Productivity A is the residual, after having removed h and k . Maybe we are capturing differences in ◮ the quality of capital. ◮ unobserved worker differences. ◮ other factors of production. 34 / 40

Growth accounting The previous decomposition was in levels. Here, ask the question: is growth coming from factor accumulation or productivity? 35 / 40

From levels to growth Define x = ˙ x ˆ x We know ∂ ln y = ˙ y y = ˆ y ∂ t Then we have xy = ∂ ln ( xy ) = ∂ ln x + ∂ ln y � = ˆ x + ˆ y ∂ t ∂ t ∂ t � x / y = .. = ˆ x − ˆ y x a = .. = a ˆ � x E.g. if x and y grow by 2% and 4%, then xy grows by 6%. 36 / 40

Growth accounting Production function again y = Ak α h 1 − α Growth rates: y = ˆ A + α ˆ k +( 1 − α )ˆ ˆ h ⇐ ⇒ ˆ y − α ˆ k − ( 1 − α )ˆ A = ˆ h Productivity growth = output growth - growth in inputs. 37 / 40

Example Set α = 1 / 3. Then ˆ y − α ˆ k − ( 1 − α )ˆ A = ˆ h = 0 . 04 − 1 30 . 02 − 2 30 . 02 = 0 . 02 38 / 40

Historical data Following countries over time, does factor accumlation (FA) or productivity explain growth? 39 / 40

Historical data Appears that ˆ A is more important - more variation in ˆ A than ˆ FA . 40 / 40