Presentation on theme: "Endogenous growth theory"— Presentation transcript:
1 Endogenous growth theory II. The empirics of GDP growth
2 QuestionsWhat are the variables (institutional, cultural, demographic…) which determine GDP per capita and/or LT growthDo we expect poor countries close the gap with rich countries?What are the policies/institutions which allow such convergence to take place
3 An industry developed in the 1990s Take a cross-section of countriesRegress their growth performance over a given period on a set of explanatory variables:InvestmentEducationFinancial developmentCorruptionAgePolitical variables: coups, etc…Then write a World Bank report saying that variable X is “good for growth”
4 The findings:A recent paper by Sala-i-Martin et al. runs a horse-race between a large number of specifications involving more than 67 variablesThey rank variables by robustness using Bayesian techniques
14 The shortcomingsWhether we are really talking about growth depends on the specificationThe economic interpretation of these regressions is not clearMany variables are not robust
15 The initial income problem: If initial income is not included in the regression, we estimate a permanent sustainable growth rateIf it is and has a negative coefficient, we estimate the long-run output levelIt can only grow ifOne of the explanatory variables grow (but most can’t)A growth trend affects all countries
16 The interpretation problem Some variables affect growth because they proxy for the growth in the inputs of the production function: education, investment, etc…Others matter because they affect human behaviour and therefore how the economy accumulates these inputsFinally, whether initial income should enter depends on how the input contributions are specified
20 Convergence in neo-classical models Neo-Classical models: each country converges to its own steady stateAll own steady states grow at the same rateBut the level depend on policies, savings rates, etcSimilar countries converge to same GDP per capita
21 Convergence in endogenous growth models A laggard never closes the gapTherefore, no convergence in income levelsThis because MPK is no higher for the laggardFurthermore, differences in policies affect the long-run growth rate
22 Looking at convergence allows us to Test the relevance of endogenous growth modelsAssess the magnitude of the returns to accumulable factors
23 Two approachesBarro and Sala-i-Martin: take a data set of similar economic units and look at convergence between them in pc GDPMankiw-Romer-Weil: take a cross-country regression of growth rates on initial income controlling for own long-run steady state
24 Barro and Sala-i-Martin They use a data-base of U.S. states over a long-run periodThey estimate the equivalent of our local speed of convergence regression:
32 Findings:The more similar the countries, the more it holds unconditionallyThe less similar the countries, the more likely we find divergenceBut the law is restored if controls are added, controlling for own steady state
33 How to eradicate poverty? 1. Adopt the policies and institutions of advanced countries2. Wait!How long? Suppose I am 10 times poorer than the US. How long does it take to be 2 times poorer?
35 What do we get? With v=0.02, ρ0 = 0.1, ρ1 = 0.5, t = 60 years!With v=0.056, we instead gett = 21 yearsWe want to understand why the speed of convergence is so lowCan policy increase the speed of convergence?
36 Gloom?In principle, the speed of convergence only depends on the deep technological parametersThat it is low tells us that the technology is not what we thought it wasBut it does not tell us we can increase v
37 Mankiw-Romer and WeilNational accounts suggest that the elasticity of capital is 0.3Speed of convergence is more like1-v/(g+δ) = /0.08 = 0.75To reconcile these two facts, they introduce another form of capital: Human capital
43 What have we learned?We have seen that with α = 0.3, it is difficult to explain X-country income differencesBut now what matters is α + β, which acts as αSo with α + β large enough we can explain cross-country differences.A natural question is: can we also expect slow convergence?
45 Empirical strategyInvestment rates and schooling are kept to proxy for own steady stateInitial output is addedCoefficient in initial output related to SOV as in BSMNo other control variable is added in strict interpretation of Solow model
49 SummaryThe Solow model predicts too low income disparities and too quick convergenceThe AK model predicts zero convergence and widening disparitiesThe Augmented Solow model does well to predict both the disparities and the speed of convergence