AI summary by TwinMind
Lecture 16: Convergence and Cross-Country Variation
MIT OpenCourseWare
47m 1s
Completing Previous Growth Theory Analysis
- Wrapping up growth theory section of course
- Showing what current models can/cannot explain regarding global income disparities
- Need to identify gaps to explain "great dispersion" in income per capita worldwide
Balanced Growth Path Review
- Complete model: productivity growth + unemployment + population growth
- Balanced growth normalization: effective workers (productivity × population)
- Zero growth for normalized variables in steady state
- Capital per worker grows at rate GA; output per worker grows at rate GA
- Exogenous drivers: GA (technological progress) and GN (population growth)
- In steady state: capital grows at GA + GN, output grows at GA + GN
Cobb-Douglas Production Function Example
- Production function: K^(1-α) × (A×N)^α
- Constant returns to scale (exponents sum to 1)
- Growth accounting: output growth = (1-α) × capital growth + α × effective worker growth
- Balanced growth: GK = GA + GN, therefore GY = GA + GN
- Capital/output per worker: subtract GN to get GA
Measuring Technological Progress - Solow Residual
- Challenge: measuring GA (technological progress rate)
- Bob Solow's approach: competitive model assumption
- Worker payment = marginal productivity ($30,000/year example)
- Labor share contribution: α × GN (wage bill/total revenue × population growth)
- Capital share contribution: (1-α) × GK
- Solow residual: GA = GY - α×GN - (1-α)×GK
- Technological progress = output growth minus factor contributions
China Growth Accounting Example (1978-2017)
- Average growth: 7.2% per year
- Population growth: 1.7% per year
- Capital stock growth: 9.2% per year
- Solow residual (technological progress): 4.2% per year
- Analysis shows non-balanced growth (transitional growth)
- Capital growing faster than output (9.2% vs 7.2%)
- Indicates economy below steady state with rapid capital accumulation
- Balanced growth rate would be: GA + GN = 4.2% + 1.7% = 5.9%
- Transitional growth due to initial low capital stock
Human Capital Extension
- Modified production function: Y = K^(1-α) × (A×H)^α
- H = effective labor = N × e^(φ×S) where S = years of schooling
- φ ≈ 0.1 (one extra year schooling raises human capital ~10%)
- Huge cross-country differences in education levels
- Normalization by population (N) rather than effective workers
Balanced Growth with Human Capital
- Eventually GH = 0 (education cannot increase indefinitely)
- Same balanced growth characteristics as basic model
- Capital per person grows at GA, output per person grows at GA
- Human capital affects steady state income level, not growth rate
Steady State Capital-Labor Ratio
- Capital accumulation equation with human capital
- Steady state: K/H = [s/(δ + n + GA)]^(1/(1-α))
- Positive relationship with saving rate
- Negative relationship with population growth
- Output in balanced growth: function of technology, human capital, and structural parameters
Cross-Country Income Comparisons
- Comparing countries to US baseline
- Relative income: yi = (Hi/Hus) × (si/sus)^((1-α)/α) × other factors
- Assumption: same technology across countries initially
- Higher saving rates → higher income per capita
- Higher population growth → lower income per capita
Technology Differences Across Countries
- Initial assumption of equal technology worldwide = bad assumption
- Measuring relative technology levels across countries
- Technology differences explain 50-67% of cross-country income differences
- Much more important than saving rates, education, population growth
Conditional Convergence
- Models work well for countries with similar technology (A) and human capital (H)
- Example: US, UK, Japan convergence 1870-present
- Negative relationship between initial income and growth rate (for similar countries)
- Countries with similar fundamentals converge to each other
Unconditional vs Conditional Convergence
- Without controlling for A and H: convergence unclear
- With controls for A and H: strong convergence patterns
- Some countries (especially Africa) have persistently low technology levels
- Each country converges to its own steady state, not universal steady state
Technology Disparity Trends
- Growing inequality in technology levels worldwide
- Ratio of 90th to 10th percentile GDP per worker increasing over time
- Different rates of technological progress across countries
- Recent decline possibly due to China's rapid growth and large size
Model Limitations
- Good at explaining convergence to country-specific steady states
- Poor at explaining why countries have different technology levels
- Gap in understanding what drives technology adoption disparities
- Institutions literature provides some explanations but incomplete
