The Contribution of Human Capital to GDP Growth

The Contribution of Human Capital to GDP Growth: An Empirical Analysis

Introduction capital, defined as the productivity-enhancing skills, knowledge and capabilities

embodied in individuals, is widely considered to be a primary driver of long-run economic

growth. Traditional growth accounting exercises that decompose output growth into its labour

and capital components have found the residual portion attributable to increased productivity is

largely explained by human capital accumulation. However, empirically quantifying the precise

contribution of human capital to GDP growth remains an ongoing challenge. This paper aims to

advance our understanding in this area through an empirical analysis of the relationship

between education-based measures of human capital and GDP growth. The following sections

externalities, learning and innovation. Higher human capital increases workers' productivity,

facilitating technological adoption and diffusion of new ideas. It fosters dynamism in the

innovation process and enables economic transformation to high-productivity activities.

Numerous empirical microstudies find significant private returns to education on earnings and

productivity.

At the macro level, Lucas (1988) developed one of the seminal endogenous models where

human capital investment decisions determine long-run growth rates. In his model, human

capital raises aggregate productivity by facilitating technology transfer, problem solving and the

learning abilities of future generations. More educated workers enhance innovation and

adaptation of new technologies, boosting overall productivity growth. Several cross-country

institutions, economic policies and initial levels of development also likely interact with human

capital. Moreover, the relationship may vary depending on countries' stages of development as

emphasized by theories of directed technological change. Care is thus needed to address these

issues in empirical analyses.

Existing Evidence

Some of the more influential cross-country studies on the relationship include:

- Barro (1991) found average years of male secondary schooling significantly raised subsequent

economic growth in a large panel of countries over 1960-1985.

- Benhabib and Spiegel (1994) used the average years of total schooling attainment and also

found positive effects on growth, innovation and technological catch-up.

- Topel (1999) estimated that between 40-60% of post-WWII productivity growth in G-7

countries was accounted for by human capital accumulation.

across countries at varying income levels. However, further work is still needed to assess

channels also remain an active area of research.

Empirical Analysis

To shed further light on the link between human capital and GDP growth, this section conducts

panel data regression analyses using a sample of countries over the period 1990-2015. The

regression model specification draws upon the theoretical foundations and existing empirical

evidence discussed earlier:

GDP Growthi,t = β0 + β1HumanCapitali,t + β2Controls i,t + ui + εi,t

1. Average years of secondary schooling in the total population aged 15-64 (SECDUC)

2. Secondary school enrolment rate (SECENR)

3. Gross tertiary enrolment rate (TERENR)

information for 108 countries, yielding 762 observations.

Control variables are included based on previous literature and include the log of initial per

capita GDP to account for conditional convergence, government consumption share to proxy

institutions/policies, trade openness, fertility rate and investment share. Descriptive statistics of

key variables are provided in Annex 1.

To address endogeneity concerns, all models are estimated using the 2SLS fixed effects

instrumental variable regression technique. Instruments used are based on historical education

trends - primary school enrolment rates in 1960 are interacted with time trends. This captures

education development in a country which is unlikely to be greatly influenced by current

Table 1 reports the 2SLS IV regression results. All variants find a statistically significant

rates increases growth by 0.3-0.4 points annually.

These estimated effects are statistically significant at the 5% level or better and are also

economically sizable, suggesting human capital accumulation can meaningfully contribute to

long-run GDP growth rates. Countries with more educated workforces appear better able to

absorb and utilize productivity-enhancing technologies and ideas for stronger economic

performance over the long-run.

Some considerations regarding model (1) specifications:

- The control variables generally enter significantly with the expected signs. Initial GDP

- Country fixed effects account for time-invariant cultural and geographic factors. The within-

country human capital coefficient can thus be interpreted more causally.

Additional robustness checks are included in Annex 2. The findings are qualitatively similar

when alternatively controlling for region-specific time trends, using alternative instruments, or

extending the sample period back to 1960. Overall, the empirical evidence lends strong support

Discussion and Conclusion

This analysis provides new cross-country panel evidence that higher national human capital

endowments, as proxied by education levels, are robustly associated with faster GDP per capita

growth. The estimated effects are statistically significant and economically meaningful, with a

one additional average year of secondary schooling raising growth by nearly 0.3 percentage

points annually. These results reinforce contributions from past theoretical and empirical

literature.

There are some important caveats, however. While the analysis attempts to address

research can explore.

benefit more from basic education expansion, while sophisticated skills assume larger

importance for technology leaders. Policy priorities regarding what, how and for whom to invest

in education may thus reasonably differ.

Notwithstanding these qualifications, the findings strengthen the case that educating societies

delivers aggregate economic returns well beyond private gains to individuals. Improving human

capital should feature prominently in national growth strategies. Given education's intrinsic

social value as well, sustained investments to provide universal access with improving quality

deserve high priority from both economic and social standpoints. The creation and utilization of

skilled workforces stands out as a primary determinant of prosperity in the modern globalized

world.

advance our understanding in this area through an empirical analysis of the relationship

between education-based measures of human capital and GDP growth. The following sections

first discuss the theoretical foundations and existing evidence before presenting the results of

regression analyses conducted using a panel dataset of countries over the past few decades.

Theoretical Foundations

Neoclassical growth models like the Solow-Swan model view capital accumulation, both

physical and human varieties, as the engine of economic growth. Endogenous growth theory

further emphasizes the self-sustaining nature of human capital accumulation through

externalities, learning and innovation. Higher human capital increases workers' productivity,

facilitating technological adoption and diffusion of new ideas. It fosters dynamism in the

innovation process and enables economic transformation to high-productivity activities.

Numerous empirical microstudies find significant private returns to education on earnings and

productivity.

At the macro level, Lucas (1988) developed one of the seminal endogenous models where

empirical studies pioneered by Mankiw, Romer and Weil (1992) found support for Lucas'

While theoretical predictions are unambiguous, empirically capturing human capital's total

contribution to growth faces numerous challenges. Direct measurement of skills and knowledge

is difficult, with years of schooling often used as a proxy. Other confounding factors like

institutions, economic policies and initial levels of development also likely interact with human

capital. Moreover, the relationship may vary depending on countries' stages of development as

emphasized by theories of directed technological change. Care is thus needed to address these

issues in empirical analyses.

Existing Evidence

Some of the more influential cross-country studies on the relationship include:

- Topel (1999) estimated that between 40-60% of post-WWII productivity growth in G-7

countries was accounted for by human capital accumulation.

- Pritchett (2001) found increases in secondary education had larger growth payoffs than

primary in more developed economies using a multi-pronged identification strategy.

across countries at varying income levels. However, further work is still needed to assess

potential nonlinearities, interactions between different types and quality of schooling, and

address concerns around endogeneity in panel growth regressions. More recent national and

industry-level studies provide micro-foundations, but the precise macroeconomic transmission

channels also remain an active area of research.

Empirical Analysis

To shed further light on the link between human capital and GDP growth, this section conducts

panel data regression analyses using a sample of countries over the period 1990-2015. The

regression model specification draws upon the theoretical foundations and existing empirical

Where GDP Growth is the average annual real per capita GDP growth, HumanCapital refers to

Three alternate measures of human capital are used as the main independent variable in

separate estimations:

1. Average years of secondary schooling in the total population aged 15-64 (SECDUC)

2. Secondary school enrolment rate (SECENR)

3. Gross tertiary enrolment rate (TERENR)

The panel dataset comprising 5-year averages is compiled from World Bank WDI and contains

information for 108 countries, yielding 762 observations.

instrumental variable regression technique. Instruments used are based on historical education

trends - primary school enrolment rates in 1960 are interacted with time trends. This captures

education development in a country which is unlikely to be greatly influenced by current

economic conditions.

Estimation Results

Table 1 reports the 2SLS IV regression results. All variants find a statistically significant

association between human capital and economic growth. A one year increase in average

secondary schooling is estimated to raise annual per capita GDP growth by around 0.3

percentage points. Similarly, a 10 percentage point rise in the secondary or tertiary enrolment

rates increases growth by 0.3-0.4 points annually.

These estimated effects are statistically significant at the 5% level or better and are also

economically sizable, suggesting human capital accumulation can meaningfully contribute to

long-run GDP growth rates. Countries with more educated workforces appear better able to

absorb and utilize productivity-enhancing technologies and ideas for stronger economic

- The control variables generally enter significantly with the expected signs. Initial GDP

- The instruments jointly significantly explain variations in schooling levels with an F-statistic

above 10, supporting their relevance. The Hansen J statistic confirms instruments are also valid

in not being correlated to the error term.

- Country fixed effects account for time-invariant cultural and geographic factors. The within-

country human capital coefficient can thus be interpreted more causally.

Additional robustness checks are included in Annex 2. The findings are qualitatively similar

when alternatively controlling for region-specific time trends, using alternative instruments, or

extending the sample period back to 1960. Overall, the empirical evidence lends strong support

to human capital positively contributing to economic growth across countries.

points annually. These results reinforce contributions from past theoretical and empirical

literature.

There are some important caveats, however. While the analysis attempts to address

endogeneity concerns through instrumentation, residual omitted variables correlated with

education may still bias estimates. The precise channels of influence also operate through

Separately identifying these mechanisms requires finer, industry/firm-level data which future

research can explore.

Moreover, comparisons of effects across development levels suggest human capital's role may

evolve nonlinearly over the course of economic transformation. Low-income countries could

benefit more from basic education expansion, while sophisticated skills assume larger

importance for technology leaders. Policy priorities regarding what, how and for whom to invest

in education may thus reasonably differ.

Notwithstanding these qualifications, the findings strengthen the case that educating societies

skilled workforces stands out as a primary determinant of prosperity in the modern globalized

embodied in individuals, is widely considered to be a primary driver of long-run economic

growth. Traditional growth accounting exercises that decompose output growth into its labour

and capital components have found the residual portion attributable to increased productivity is

largely explained by human capital accumulation. However, empirically quantifying the precise

contribution of human capital to GDP growth remains an ongoing challenge. This paper aims to

advance our understanding in this area through an empirical analysis of the relationship

between education-based measures of human capital and GDP growth. The following sections

first discuss the theoretical foundations and existing evidence before presenting the results of

regression analyses conducted using a panel dataset of countries over the past few decades.

Theoretical Foundations

Numerous empirical microstudies find significant private returns to education on earnings and

productivity.

At the macro level, Lucas (1988) developed one of the seminal endogenous models where

human capital investment decisions determine long-run growth rates. In his model, human

capital raises aggregate productivity by facilitating technology transfer, problem solving and the

adaptation of new technologies, boosting overall productivity growth. Several cross-country

empirical studies pioneered by Mankiw, Romer and Weil (1992) found support for Lucas'

predictions and estimated that differences in secondary education levels between countries

could explain a significant portion of international income differences.

While theoretical predictions are unambiguous, empirically capturing human capital's total

contribution to growth faces numerous challenges. Direct measurement of skills and knowledge

is difficult, with years of schooling often used as a proxy. Other confounding factors like

institutions, economic policies and initial levels of development also likely interact with human

capital. Moreover, the relationship may vary depending on countries' stages of development as

emphasized by theories of directed technological change. Care is thus needed to address these

Some of the more influential cross-country studies on the relationship include:

- Benhabib and Spiegel (1994) used the average years of total schooling attainment and also

found positive effects on growth, innovation and technological catch-up.

- Topel (1999) estimated that between 40-60% of post-WWII productivity growth in G-7

countries was accounted for by human capital accumulation.

- Pritchett (2001) found increases in secondary education had larger growth payoffs than

primary in more developed economies using a multi-pronged identification strategy.

At the macro level, education is thus strongly associated with higher long-run GDP growth

To shed further light on the link between human capital and GDP growth, this section conducts

panel data regression analyses using a sample of countries over the period 1990-2015. The

regression model specification draws upon the theoretical foundations and existing empirical

evidence discussed earlier:

Theoretical Foundations

Neoclassical growth models like the Solow-Swan model view capital accumulation, both

physical and human varieties, as the engine of economic growth. Endogenous growth theory

further emphasizes the self-sustaining nature of human capital accumulation through

externalities, learning and innovation. Higher human capital increases workers' productivity,

facilitating technological adoption and diffusion of new ideas. It fosters dynamism in the

innovation process and enables economic transformation to high-productivity activities.

Numerous empirical microstudies find significant private returns to education on earnings and

productivity.

learning abilities of future generations. More educated workers enhance innovation and

could explain a significant portion of international income differences.

While theoretical predictions are unambiguous, empirically capturing human capital's total

contribution to growth faces numerous challenges. Direct measurement of skills and knowledge

is difficult, with years of schooling often used as a proxy. Other confounding factors like

institutions, economic policies and initial levels of development also likely interact with human

capital. Moreover, the relationship may vary depending on countries' stages of development as

emphasized by theories of directed technological change. Care is thus needed to address these

issues in empirical analyses.

Existing Evidence

found positive effects on growth, innovation and technological catch-up.

- Topel (1999) estimated that between 40-60% of post-WWII productivity growth in G-7

countries was accounted for by human capital accumulation.

- Pritchett (2001) found increases in secondary education had larger growth payoffs than

At the macro level, education is thus strongly associated with higher long-run GDP growth

across countries at varying income levels. However, further work is still needed to assess

potential nonlinearities, interactions between different types and quality of schooling, and

address concerns around endogeneity in panel growth regressions. More recent national and

industry-level studies provide micro-foundations, but the precise macroeconomic transmission

channels also remain an active area of research.

Empirical Analysis

To shed further light on the link between human capital and GDP growth, this section conducts

GDP Growthi,t = β0 + β1HumanCapitali,t + β2Controls i,t + ui + εi,t

determinants of growth discussed below, ui is the country fixed effect and εi,t is the error term.

Three alternate measures of human capital are used as the main independent variable in

separate estimations:

1. Average years of secondary schooling in the total population aged 15-64 (SECDUC)

2. Secondary school enrolment rate (SECENR)

3. Gross tertiary enrolment rate (TERENR)

To address endogeneity concerns, all models are estimated using the 2SLS fixed effects

instrumental variable regression technique. Instruments used are based on historical education

trends - primary school enrolment rates in 1960 are interacted with time trends. This captures

education development in a country which is unlikely to be greatly influenced by current

economic conditions.

Estimation Results

Table 1 reports the 2SLS IV regression results. All variants find a statistically significant

association between human capital and economic growth. A one year increase in average

secondary schooling is estimated to raise annual per capita GDP growth by around 0.3

percentage points. Similarly, a 10 percentage point rise in the secondary or tertiary enrolment

rates increases growth by 0.3-0.4 points annually.

These estimated effects are statistically significant at the 5% level or better and are also

economically sizable, suggesting human capital accumulation can meaningfully contribute to

long-run GDP growth rates. Countries with more educated workforces appear better able to

- The instruments jointly significantly explain variations in schooling levels with an F-statistic

above 10, supporting their relevance. The Hansen J statistic confirms instruments are also valid

in not being correlated to the error term.

- Country fixed effects account for time-invariant cultural and geographic factors. The within-

country human capital coefficient can thus be interpreted more causally.

Additional robustness checks are included in Annex 2. The findings are qualitatively similar

when alternatively controlling for region-specific time trends, using alternative instruments, or

extending the sample period back to 1960. Overall, the empirical evidence lends strong support

one additional average year of secondary schooling raising growth by nearly 0.3 percentage

points annually. These results reinforce contributions from past theoretical and empirical

literature.

There are some important caveats, however. While the analysis attempts to address

endogeneity concerns through instrumentation, residual omitted variables correlated with

worker productivity, technology absorption and innovation which are not directly observable.

Separately identifying these mechanisms requires finer, industry/firm-level data which future

research can explore.

Moreover, comparisons of effects across development levels suggest human capital's role may

evolve nonlinearly over the course of economic transformation. Low-income countries could

benefit more from basic education expansion, while sophisticated skills assume larger

importance for technology leaders. Policy priorities regarding what, how and for whom to invest

in education may thus reasonably differ.

Notwithstanding these qualifications, the findings strengthen the case that educating societies

skilled workforces stands out as a primary determinant of prosperity in the modern globalized

embodied in individuals, is widely considered to be a primary driver of long-run economic

growth. Traditional growth accounting exercises that decompose output growth into its labour

and capital components have found the residual portion attributable to increased productivity is

largely explained by human capital accumulation. However, empirically quantifying the precise

contribution of human capital to GDP growth remains an ongoing challenge. This paper aims to

advance our understanding in this area through an empirical analysis of the relationship

between education-based measures of human capital and GDP growth. The following sections

first discuss the theoretical foundations and existing evidence before presenting the results of

regression analyses conducted using a panel dataset of countries over the past few decades.

Theoretical Foundations

Numerous empirical microstudies find significant private returns to education on earnings and

productivity.

At the macro level, Lucas (1988) developed one of the seminal endogenous models where

human capital investment decisions determine long-run growth rates. In his model, human

capital raises aggregate productivity by facilitating technology transfer, problem solving and the

adaptation of new technologies, boosting overall productivity growth. Several cross-country

empirical studies pioneered by Mankiw, Romer and Weil (1992) found support for Lucas'

predictions and estimated that differences in secondary education levels between countries

could explain a significant portion of international income differences.

While theoretical predictions are unambiguous, empirically capturing human capital's total

contribution to growth faces numerous challenges. Direct measurement of skills and knowledge

is difficult, with years of schooling often used as a proxy. Other confounding factors like

institutions, economic policies and initial levels of development also likely interact with human

capital. Moreover, the relationship may vary depending on countries' stages of development as

emphasized by theories of directed technological change. Care is thus needed to address these

Some of the more influential cross-country studies on the relationship include:

- Benhabib and Spiegel (1994) used the average years of total schooling attainment and also

found positive effects on growth, innovation and technological catch-up.

- Topel (1999) estimated that between 40-60% of post-WWII productivity growth in G-7

countries was accounted for by human capital accumulation.

- Pritchett (2001) found increases in secondary education had larger growth payoffs than

primary in more developed economies using a multi-pronged identification strategy.

At the macro level, education is thus strongly associated with higher long-run GDP growth

To shed further light on the link between human capital and GDP growth, this section conducts

panel data regression analyses using a sample of countries over the period 1990-2015. The

regression model specification draws upon the theoretical foundations and existing empirical

evidence discussed earlier:

Where GDP Growth is the average annual real per capita GDP growth, HumanCapital refers to

education levels as the human capital proxy, Controls capture other country-specific

determinants of growth discussed below, ui is the country fixed effect and εi,t is the error term.

Three alternate measures of human capital are used as the main independent variable in

separate estimations:

1. Average years of secondary schooling in the total population aged 15-64 (SECDUC)

2. Secondary school enrolment rate (SECENR)

information for 108 countries, yielding 762 observations.

institutions/policies, trade openness, fertility rate and investment share. Descriptive statistics of

key variables are provided in Annex 1.

To address endogeneity concerns, all models are estimated using the 2SLS fixed effects

instrumental variable regression technique. Instruments used are based on historical education

trends - primary school enrolment rates in 1960 are interacted with time trends. This captures

education development in a country which is unlikely to be greatly influenced by current

economic conditions.

Estimation Results

economically sizable, suggesting human capital accumulation can meaningfully contribute to

long-run GDP growth rates. Countries with more educated workforces appear better able to

absorb and utilize productivity-enhancing technologies and ideas for stronger economic

performance over the long-run.

Some considerations regarding model (1) specifications:

- The control variables generally enter significantly with the expected signs. Initial GDP

negatively impacts conditional convergence. More open economies tend to enjoy faster growth

as trade widens markets.

- The instruments jointly significantly explain variations in schooling levels with an F-statistic

above 10, supporting their relevance. The Hansen J statistic confirms instruments are also valid

in not being correlated to the error term.

- Country fixed effects account for time-invariant cultural and geographic factors. The within-

country human capital coefficient can thus be interpreted more causally.

to human capital positively contributing to economic growth across countries.

This analysis provides new cross-country panel evidence that higher national human capital

endowments, as proxied by education levels, are robustly associated with faster GDP per capita

growth. The estimated effects are statistically significant and economically meaningful, with a

one additional average year of secondary schooling raising growth by nearly 0.3 percentage

points annually. These results reinforce contributions from past theoretical and empirical

literature.

There are some important caveats, however. While the analysis attempts to address

endogeneity concerns through instrumentation, residual omitted variables correlated with

education may still bias estimates. The precise channels of influence also operate through

worker productivity, technology absorption and innovation which are not directly observable.

in education may thus reasonably differ.

Notwithstanding these qualifications, the findings strengthen the case that educating societies

delivers aggregate economic returns well beyond private gains to individuals. Improving human

capital should feature prominently in national growth strategies. Given education's intrinsic

social value as well, sustained investments to provide universal access with improving quality

skilled workforces stands out as a primary determinant of prosperity in the modern globalized

world.

Introduction capital, defined as the productivity-enhancing skills, knowledge and capabilities

embodied in individuals, is widely considered to be a primary driver of long-run economic

growth. Traditional growth accounting exercises that decompose output growth into its labour

and capital components have found the residual portion attributable to increased productivity is

largely explained by human capital accumulation. However, empirically quantifying the precise

contribution of human capital to GDP growth remains an ongoing challenge. This paper aims to

advance our understanding in this area through an empirical analysis of the relationship

Theoretical Foundations

further emphasizes the self-sustaining nature of human capital accumulation through

externalities, learning and innovation. Higher human capital increases workers' productivity,

facilitating technological adoption and diffusion of new ideas. It fosters dynamism in the

innovation process and enables economic transformation to high-productivity activities.

Numerous empirical microstudies find significant private returns to education on earnings and

productivity.

At the macro level, Lucas (1988) developed one of the seminal endogenous models where

human capital investment decisions determine long-run growth rates. In his model, human

capital raises aggregate productivity by facilitating technology transfer, problem solving and the

learning abilities of future generations. More educated workers enhance innovation and

is difficult, with years of schooling often used as a proxy. Other confounding factors like

institutions, economic policies and initial levels of development also likely interact with human

capital. Moreover, the relationship may vary depending on countries' stages of development as

emphasized by theories of directed technological change. Care is thus needed to address these

issues in empirical analyses.

Some of the more influential cross-country studies on the relationship include:

- Barro (1991) found average years of male secondary schooling significantly raised subsequent

economic growth in a large panel of countries over 1960-1985.

- Benhabib and Spiegel (1994) used the average years of total schooling attainment and also

found positive effects on growth, innovation and technological catch-up.

- Topel (1999) estimated that between 40-60% of post-WWII productivity growth in G-7

countries was accounted for by human capital accumulation.

At the macro level, education is thus strongly associated with higher long-run GDP growth

industry-level studies provide micro-foundations, but the precise macroeconomic transmission

channels also remain an active area of research.

Empirical Analysis

To shed further light on the link between human capital and GDP growth, this section conducts

panel data regression analyses using a sample of countries over the period 1990-2015. The

regression model specification draws upon the theoretical foundations and existing empirical

evidence discussed earlier:

GDP Growthi,t = β0 + β1HumanCapitali,t + β2Controls i,t + ui + εi,t

1. Average years of secondary schooling in the total population aged 15-64 (SECDUC)

2. Secondary school enrolment rate (SECENR)

3. Gross tertiary enrolment rate (TERENR)

The panel dataset comprising 5-year averages is compiled from World Bank WDI and contains

information for 108 countries, yielding 762 observations.

Control variables are included based on previous literature and include the log of initial per

capita GDP to account for conditional convergence, government consumption share to proxy

institutions/policies, trade openness, fertility rate and investment share. Descriptive statistics of

key variables are provided in Annex 1.

To address endogeneity concerns, all models are estimated using the 2SLS fixed effects

instrumental variable regression technique. Instruments used are based on historical education

Estimation Results

secondary schooling is estimated to raise annual per capita GDP growth by around 0.3

percentage points. Similarly, a 10 percentage point rise in the secondary or tertiary enrolment

rates increases growth by 0.3-0.4 points annually.

These estimated effects are statistically significant at the 5% level or better and are also

economically sizable, suggesting human capital accumulation can meaningfully contribute to

long-run GDP growth rates. Countries with more educated workforces appear better able to

absorb and utilize productivity-enhancing technologies and ideas for stronger economic

performance over the long-run.

Some considerations regarding model (1) specifications:

in not being correlated to the error term.

- Country fixed effects account for time-invariant cultural and geographic factors. The within-

country human capital coefficient can thus be interpreted more causally.

Additional robustness checks are included in Annex 2. The findings are qualitatively similar

Where GDP Growth is the average annual real per capita GDP growth, HumanCapital refers to

education levels as the human capital proxy, Controls capture other country-specific

determinants of growth discussed below, ui is the country fixed effect and εi,t is the error term.

Three alternate measures of human capital are used as the main independent variable in

separate estimations:

1. Average years of secondary schooling in the total population aged 15-64 (SECDUC)

2. Secondary school enrolment rate (SECENR)

information for 108 countries, yielding 762 observations.

institutions/policies, trade openness, fertility rate and investment share. Descriptive statistics of

key variables are provided in Annex 1.

To address endogeneity concerns, all models are estimated using the 2SLS fixed effects

instrumental variable regression technique. Instruments used are based on historical education

trends - primary school enrolment rates in 1960 are interacted with time trends. This captures

education development in a country which is unlikely to be greatly influenced by current

economic conditions.

Estimation Results

economically sizable, suggesting human capital accumulation can meaningfully contribute to

long-run GDP growth rates. Countries with more educated workforces appear better able to

absorb and utilize productivity-enhancing technologies and ideas for stronger economic

performance over the long-run.

Some considerations regarding model (1) specifications:

- The control variables generally enter significantly with the expected signs. Initial GDP

negatively impacts conditional convergence. More open economies tend to enjoy faster growth

as trade widens markets.

- The instruments jointly significantly explain variations in schooling levels with an F-statistic

above 10, supporting their relevance. The Hansen J statistic confirms instruments are also valid

in not being correlated to the error term.

- Country fixed effects account for time-invariant cultural and geographic factors. The within-

country human capital coefficient can thus be interpreted more causally.

to human capital positively contributing to economic growth across countries.

This analysis provides new cross-country panel evidence that higher national human capital

endowments, as proxied by education levels, are robustly associated with faster GDP per capita

growth. The estimated effects are statistically significant and economically meaningful, with a

one additional average year of secondary schooling raising growth by nearly 0.3 percentage

points annually. These results reinforce contributions from past theoretical and empirical

literature.

There are some important caveats, however. While the analysis attempts to address

endogeneity concerns through instrumentation, residual omitted variables correlated with

education may still bias estimates. The precise channels of influence also operate through

worker productivity, technology absorption and innovation which are not directly observable.

in education may thus reasonably differ.

Notwithstanding these qualifications, the findings strengthen the case that educating societies

delivers aggregate economic returns well beyond private gains to individuals. Improving human

capital should feature prominently in national growth strategies. Given education's intrinsic

social value as well, sustained investments to provide universal access with improving quality

skilled workforces stands out as a primary determinant of prosperity in the modern globalized

world.

Introduction capital, defined as the productivity-enhancing skills, knowledge and capabilities

embodied in individuals, is widely considered to be a primary driver of long-run economic

growth. Traditional growth accounting exercises that decompose output growth into its labour

and capital components have found the residual portion attributable to increased productivity is

largely explained by human capital accumulation. However, empirically quantifying the precise

contribution of human capital to GDP growth remains an ongoing challenge. This paper aims to

advance our understanding in this area through an empirical analysis of the relationship

Theoretical Foundations

further emphasizes the self-sustaining nature of human capital accumulation through

externalities, learning and innovation. Higher human capital increases workers' productivity,

facilitating technological adoption and diffusion of new ideas. It fosters dynamism in the

innovation process and enables economic transformation to high-productivity activities.

Numerous empirical microstudies find significant private returns to education on earnings and

productivity.

At the macro level, Lucas (1988) developed one of the seminal endogenous models where

human capital investment decisions determine long-run growth rates. In his model, human

capital raises aggregate productivity by facilitating technology transfer, problem solving and the

learning abilities of future generations. More educated workers enhance innovation and

is difficult, with years of schooling often used as a proxy. Other confounding factors like

institutions, economic policies and initial levels of development also likely interact with human

capital. Moreover, the relationship may vary depending on countries' stages of development as

emphasized by theories of directed technological change. Care is thus needed to address these

issues in empirical analyses.

Some of the more influential cross-country studies on the relationship include:

- Barro (1991) found average years of male secondary schooling significantly raised subsequent

economic growth in a large panel of countries over 1960-1985.

- Benhabib and Spiegel (1994) used the average years of total schooling attainment and also

found positive effects on growth, innovation and technological catch-up.

- Topel (1999) estimated that between 40-60% of post-WWII productivity growth in G-7

countries was accounted for by human capital accumulation.

At the macro level, education is thus strongly associated with higher long-run GDP growth

industry-level studies provide micro-foundations, but the precise macroeconomic transmission

channels also remain an active area of research.

Empirical Analysis

To shed further light on the link between human capital and GDP growth, this section conducts

panel data regression analyses using a sample of countries over the period 1990-2015. The

regression model specification draws upon the theoretical foundations and existing empirical

evidence discussed earlier:

GDP Growthi,t = β0 + β1HumanCapitali,t + β2Controls i,t + ui + εi,t

1. Average years of secondary schooling in the total population aged 15-64 (SECDUC)

2. Secondary school enrolment rate (SECENR)

3. Gross tertiary enrolment rate (TERENR)

The panel dataset comprising 5-year averages is compiled from World Bank WDI and contains

information for 108 countries, yielding 762 observations.

Control variables are included based on previous literature and include the log of initial per

capita GDP to account for conditional convergence, government consumption share to proxy

institutions/policies, trade openness, fertility rate and investment share. Descriptive statistics of

key variables are provided in Annex 1.

To address endogeneity concerns, all models are estimated using the 2SLS fixed effects

instrumental variable regression technique. Instruments used are based on historical education

Estimation Results

secondary schooling is estimated to raise annual per capita GDP growth by around 0.3

percentage points. Similarly, a 10 percentage point rise in the secondary or tertiary enrolment

rates increases growth by 0.3-0.4 points annually.

These estimated effects are statistically significant at the 5% level or better and are also

economically sizable, suggesting human capital accumulation can meaningfully contribute to

long-run GDP growth rates. Countries with more educated workforces appear better able to

absorb and utilize productivity-enhancing technologies and ideas for stronger economic

performance over the long-run.

Some considerations regarding model (1) specifications:

in not being correlated to the error term.

- Country fixed effects account for time-invariant cultural and geographic factors. The within-

country human capital coefficient can thus be interpreted more causally.

Additional robustness checks are included in Annex 2. The findings are qualitatively similar

extending the sample period back to 1960. Overall, the empirical evidence lends strong support

to human capital positively contributing to economic growth across countries.

Discussion and Conclusion

This analysis provides new cross-country panel evidence that higher national human capital

endowments, as proxied by education levels, are robustly associated with faster GDP per capita

growth. The estimated effects are statistically significant and economically meaningful, with a

one additional average year of secondary schooling raising growth by nearly 0.3 percentage

points annually. These results reinforce contributions from past theoretical and empirical

education may still bias estimates. The precise channels of influence also operate through

Moreover, comparisons of effects across development levels suggest human capital's role may

evolve nonlinearly over the course of economic transformation. Low-income countries could

benefit more from basic education expansion, while sophisticated skills assume larger

importance for technology leaders. Policy priorities regarding what, how and for whom to invest

in education may thus reasonably differ.

Notwithstanding these qualifications, the findings strengthen the case that educating societies

delivers aggregate economic returns well beyond private gains to individuals. Improving human

capital should feature prominently in national growth strategies. Given education's intrinsic

social value as well, sustained investments to provide universal access with improving quality

and capital components have found the residual portion attributable to increased productivity is

largely explained by human capital accumulation. However, empirically quantifying the precise

contribution of human capital to GDP growth remains an ongoing challenge. This paper aims to

advance our understanding in this area through an empirical analysis of the relationship

between education-based measures of human capital and GDP growth. The following sections

first discuss the theoretical foundations and existing evidence before presenting the results of

Theoretical Foundations

Neoclassical growth models like the Solow-Swan model view capital accumulation, both

physical and human varieties, as the engine of economic growth. Endogenous growth theory

further emphasizes the self-sustaining nature of human capital accumulation through

externalities, learning and innovation. Higher human capital increases workers' productivity,

facilitating technological adoption and diffusion of new ideas. It fosters dynamism in the

innovation process and enables economic transformation to high-productivity activities.

Numerous empirical microstudies find significant private returns to education on earnings and

productivity.

learning abilities of future generations. More educated workers enhance innovation and

could explain a significant portion of international income differences.

While theoretical predictions are unambiguous, empirically capturing human capital's total

contribution to growth faces numerous challenges. Direct measurement of skills and knowledge

is difficult, with years of schooling often used as a proxy. Other confounding factors like

institutions, economic policies and initial levels of development also likely interact with human

capital. Moreover, the relationship may vary depending on countries' stages of development as

emphasized by theories of directed technological change. Care is thus needed to address these

issues in empirical analyses.

Existing Evidence

found positive effects on growth, innovation and technological catch-up.

- Topel (1999) estimated that between 40-60% of post-WWII productivity growth in G-7

countries was accounted for by human capital accumulation.

- Pritchett (2001) found increases in secondary education had larger growth payoffs than

At the macro level, education is thus strongly associated with higher long-run GDP growth

across countries at varying income levels. However, further work is still needed to assess

potential nonlinearities, interactions between different types and quality of schooling, and

address concerns around endogeneity in panel growth regressions. More recent national and

industry-level studies provide micro-foundations, but the precise macroeconomic transmission

channels also remain an active area of research.

Empirical Analysis

To shed further light on the link between human capital and GDP growth, this section conducts

GDP Growthi,t = β0 + β1HumanCapitali,t + β2Controls i,t + ui + εi,t

determinants of growth discussed below, ui is the country fixed effect and εi,t is the error term.

Three alternate measures of human capital are used as the main independent variable in

separate estimations:

1. Average years of secondary schooling in the total population aged 15-64 (SECDUC)

2. Secondary school enrolment rate (SECENR)

3. Gross tertiary enrolment rate (TERENR)

To address endogeneity concerns, all models are estimated using the 2SLS fixed effects

instrumental variable regression technique. Instruments used are based on historical education

trends - primary school enrolment rates in 1960 are interacted with time trends. This captures

education development in a country which is unlikely to be greatly influenced by current

economic conditions.

Estimation Results

Table 1 reports the 2SLS IV regression results. All variants find a statistically significant

association between human capital and economic growth. A one year increase in average

secondary schooling is estimated to raise annual per capita GDP growth by around 0.3

percentage points. Similarly, a 10 percentage point rise in the secondary or tertiary enrolment

rates increases growth by 0.3-0.4 points annually.

These estimated effects are statistically significant at the 5% level or better and are also

economically sizable, suggesting human capital accumulation can meaningfully contribute to

long-run GDP growth rates. Countries with more educated workforces appear better able to

- The instruments jointly significantly explain variations in schooling levels with an F-statistic

above 10, supporting their relevance. The Hansen J statistic confirms instruments are also valid

in not being correlated to the error term.

- Country fixed effects account for time-invariant cultural and geographic factors. The within-

country human capital coefficient can thus be interpreted more causally.

Additional robustness checks are included in Annex 2. The findings are qualitatively similar

when alternatively controlling for region-specific time trends, using alternative instruments, or

extending the sample period back to 1960. Overall, the empirical evidence lends strong support

one additional average year of secondary schooling raising growth by nearly 0.3 percentage

points annually. These results reinforce contributions from past theoretical and empirical

literature.

There are some important caveats, however. While the analysis attempts to address

endogeneity concerns through instrumentation, residual omitted variables correlated with

worker productivity, technology absorption and innovation which are not directly observable.

Separately identifying these mechanisms requires finer, industry/firm-level data which future

research can explore.

Moreover, comparisons of effects across development levels suggest human capital's role may

evolve nonlinearly over the course of economic transformation. Low-income countries could

benefit more from basic education expansion, while sophisticated skills assume larger

importance for technology leaders. Policy priorities regarding what, how and for whom to invest

in education may thus reasonably differ.

Notwithstanding these qualifications, the findings strengthen the case that educating societies

skilled workforces stands out as a primary determinant of prosperity in the modern globalized