Determinants of long-run economic growth including physical capital accumulation, human capital, population growth, and total factor productivity. Built on Solow (1956), Barro (1991), and Mankiw-Romer-Weil (1992).
Constructs
gdp_per_capita
GDP Per Capita
Total economic output divided by population, measured in constant purchasing power parity dollars. The standard measure of average material living standards across countries.
physical_capital
Physical Capital Accumulation
The stock of produced means of production including machinery, equipment, structures, and infrastructure. Accumulated through investment (savings) and depreciated over time.
human_capital
Human Capital
The stock of knowledge, skills, and health embodied in the labor force. Accumulated through education, training, and health investment.
population_growth_rate
Population Growth Rate
Annual rate of change in total population, including natural increase and net migration. In growth models, higher population growth dilutes per-capita capital.
population change rateannual population growthdemographic growth rate
total_factor_productivity
Total Factor Productivity
The portion of output not explained by the quantity of inputs used in production. Reflects technological progress, efficiency, and institutional quality.
convergence_rate
Convergence Rate
The speed at which poorer economies catch up to richer ones in terms of per-capita income, conditional on structural characteristics.
Findings
A 1C increase in temperature reduces GDP per capita growth by ~1.2 percentage points in poor countries. Rich countries show no significant effect.
Direction: negative
Confidence: strong
Method: OLS with country and year fixed effects
Economic productivity peaks at ~13C and declines nonlinearly above that threshold, for both rich and poor countries.
Direction: negative
Confidence: strong
Method: OLS with country and year FE, quadratic
Augmented Solow model including human capital explains ~80% of cross-country income variance vs 59% for baseline. Human capital coefficient positive and significant.
Direction: positive
Confidence: strong
Method: OLS cross-section, N=98
One SD increase in cognitive skills test scores associated with ~2 pp higher annual GDP growth. Robust to IV identification.
Direction: positive
Confidence: strong
Method: OLS and IV cross-country
Years of schooling loses significance once cognitive skills are controlled for — quantity without quality doesn't drive growth.
Direction: null
Confidence: moderate
Method: OLS with joint inclusion
In the neoclassical growth model, long-run GDP per capita is determined by the savings rate and population growth rate given diminishing returns to capital. Higher savings rates lead to higher steady-state income levels, while higher population growth leads to lower steady-state income.
Direction: positive
Confidence: foundational
Effect: Steady-state y* = (s/(n+g+δ))^(α/(1-α)) where s=savings, n=pop growth
Method: Mathematical growth model with diminishing returns
Technological progress (TFP growth) is the only source of sustained long-run growth in per-capita output. Capital accumulation alone cannot sustain growth due to diminishing returns.
Direction: positive
Confidence: foundational
Effect: Long-run per capita growth rate equals rate of technological progress g
Method: Mathematical growth model
Growth rate of real per capita GDP is positively related to initial human capital (school enrollment rates) and negatively related to initial level of real per capita GDP, consistent with conditional convergence. Countries converge to their own steady states at approximately 2% per year.
Direction: positive
Confidence: strong
Effect: β≈-0.0075 on initial GDP (conditional convergence at ~2%/year); school enrollment β≈0.025 (positive)
Method: OLS cross-country regression, N=98 countries, 1960-1985
Investment share of GDP has a positive but diminished effect on growth once human capital is controlled for. Physical capital alone is insufficient to explain cross-country growth differences.
Direction: positive
Confidence: moderate
Effect: Investment/GDP coefficient positive but reduced when schooling included
Method: OLS cross-country regression, N=98
An augmented Solow model including human capital explains approximately 80% of cross-country variation in income per capita, compared to ~60% for the textbook Solow model. The estimated capital share (α≈1/3) is consistent with factor shares when human capital is included.
Direction: positive
Confidence: strong
Effect: R²≈0.78 for augmented model vs R²≈0.59 for basic Solow; α≈0.31, human capital share≈0.28
Method: OLS cross-country regression, N=98 non-oil countries, 1960-1985
Population growth has a strong negative effect on GDP per capita, consistent with the Solow model prediction that higher population growth dilutes per-capita capital.
Direction: negative
Confidence: strong
Effect: Coefficient on ln(n+g+δ) ≈ -1.50 (significant at 1%)
Method: OLS cross-country regression, N=98
Conditional convergence rate is approximately 2% per year across country samples, consistent with the augmented Solow model's predictions.
Direction: positive
Confidence: strong
Effect: Implied λ≈0.02 (2% convergence per year)
Method: Restricted regression with convergence speed estimation, N=98
Institutions have a large causal effect on GDP per capita. IV estimates show that a one-unit increase in protection against expropriation risk increases log GDP per capita by approximately 0.94, much larger than OLS estimates, suggesting OLS understates the institutional effect.
Direction: positive
Confidence: strong
Effect: β≈0.94 (IV), p<0.01; OLS β≈0.52
Method: IV/2SLS, N≈64 former colonies
Once institutional quality is instrumented, geographic variables (latitude, distance from coast, temperature) have no significant direct effect on GDP per capita, challenging geographic determinism theories of development.
Direction: null
Confidence: strong
Effect: Geographic variables insignificant when institutions instrumented
Method: IV/2SLS with geographic controls
Corruption negatively affects investment as a share of GDP. A one-standard-deviation decrease in corruption is associated with an increase in the investment rate of over 4 percentage points.
Direction: negative
Confidence: strong
Effect: β≈-0.3 to -0.9 depending on specification; 1 SD corruption reduction → +4pp investment/GDP
Method: OLS and IV using ethno-linguistic fractionalization as instrument, N≈67 countries
Corruption negatively affects economic growth both directly and indirectly through reduced investment. The relationship is robust to controlling for political instability and other institutional variables.
Direction: negative
Confidence: moderate
Effect: Significant negative effect on growth, robust across specifications
Method: OLS and IV, cross-country regression
Institutions are the fundamental cause of long-run economic performance. They reduce uncertainty in human exchange by providing a structure to everyday life, and the difference between institutional frameworks explains divergent economic trajectories across nations.
Direction: positive
Confidence: foundational
Effect: Foundational theoretical claim
Method: Theoretical framework with historical analysis
Transaction costs are a key mechanism through which institutions affect economic performance. Efficient institutions lower transaction costs, enabling more complex exchange and greater specialization, which drives economic growth.
Direction: positive
Confidence: foundational
Effect: Foundational mechanism linking institutions to growth
Method: Theoretical framework
Long-run GDP per capita is determined by savings rate and population growth given diminishing returns to capital
Direction: positive
Confidence: foundational
Method: Mathematical growth model
TFP growth is the only source of sustained long-run growth in per-capita output
Direction: positive
Confidence: foundational
Augmented Solow model with human capital explains 80% of cross-country income variation
Direction: positive
Confidence: strong
Method: OLS cross-country, N=98
R&D expenditure as share of GDP is positively associated with TFP growth with estimated social returns 2-4 times larger than private returns to R&D investment
Direction: positive
Confidence: strong
Method: ols_regression
Patent applications per capita are positively associated with GDP per capita growth but with heterogeneous effects across technology sectors and country income levels
Direction: conditional
Confidence: moderate
Method: ols_regression
High-technology export share is positively associated with economic complexity and long-run growth in Schumpeterian growth models with empirical validation
Direction: positive
Confidence: moderate
Method: ols_regression
Conditional convergence at approximately 2% per year across country samples
Direction: positive
Confidence: strong
Method: OLS cross-country, N=98