Cyclical budgetary policy and its effects on growth: Cyclical budgetary policy and its effects on growth Philippe Aghion, Robert Barro and Ioana Marinescu


Does macroeconomic policy (budget deficit, interest rates, taxation,...) matter for (long-run) growth?: Does macroeconomic policy (budget deficit, interest rates, taxation,...) matter for (long-run) growth? Debate on ECB policy and the Stability and Growth Pact. Does it matter for growth that Eurozone shows less countercyclical deficit than US/UK? How does the degree of development, and in particular financial development, affect the efficiency of macro policy?


Recessions, investment and growth : Recessions, investment and growth Financial development Government intervention Liquidity constraints of private firms Growth-enhancing investments by private firms Recession Schumpeter Hypothesis: Countercyclical fiscal and budgetary policy should be more growth enhancing when a country is less financially developed.


Previous literature: Previous literature Calderon et al. (2004): institutions (ICRG). Alesina-Tabellini(2005): corrupt democracies. Lane (2003): growth volatility, trade openness and political divisions.


Preview of results: Preview of results We use OECD panel data. Public deficit in the OECD gets more countercylical over time, but less so in the EMU. Lower financial development is associated with a less countercyclical fiscal and budgetary policy. More countercyclical public deficit, investment and consumption increase growth; but this effect is lessened when financial development is higher.


Outline: Outline First stage: the cyclicality of public debt and spending and its determinants. Second stage: the effect of the cyclicality of public debt and spending on growth.


Data used: Data used OECD Economic Outlook. Ross Levine’s dataset on financial development: private credit/GDP. Penn World Tables.


First stage: theory: First stage: theory The variation of public debt (or spending) is determined by (Barro 1979 tax smoothing theory): the size of government and the variation in government spending the stock of debt at the previous period GDP gap (tax revenues)


Econometric specification: Econometric specification Correlation won’t do (panel). Problem: how do we estimate a time-varying coefficient on the GDP gap? First method: Coefficient in the linear regression assumed to follow an AR(1) process for each country i at time t: Second method (check): OLS 10-years rolling window:


Procyclicality of government debt (AR(1)): Procyclicality of government debt (AR(1))


Determinants of the procyclicality of fiscal and budgetary policy: Determinants of the procyclicality of fiscal and budgetary policy Explained variable: procyclicality as estimated by the AR(1) method. All regressions also control for EMU country status, government share of GDP, relative GDP per capita. Year f.e. Year & Country f.e. Year f.e. Year & Country f.e. Year f.e. Year & Country f.e. Private credit/GDP 0.175 -0.638 0.010 -0.029 0.053 0.065 (0.186) (0.193)*** (0.010) (0.008)*** (0.014)*** (0.016)*** Standard error -1.979 1.019 0.785 of GDP growth (1.873) (0.104)*** (0.186)*** Openness 0.016 0.028 0.000 0.002 0.000 0.000 (0.002)*** (0.007)*** (0.000) (0.000)*** (0.000) (0.000) Observations 486 486 453 453 453 453 R-squared 0.26 0.70 0.41 0.87 0.29 0.86 Robust standard errors in parentheses * significant at 10%; ** significant at 5%; *** significant at 1% Public Debt Public Investment Public Consumption


Second stage: theory: Second stage: theory Aghion, Angeletos, Banerjee, Manova, 2005. Productivity a(t) Firms borrow to invest in short-term investment s(t) and long-term investment (R&D) l(t) Production: a(t)F(s(t)) Liquidity shock c(t) is realized. Firms borrow to pay c(t). Productivity a(t+1) Long-term investment yields a(t+1)q(l(t)) if liquidity shock covered Government intervention increases net cash flow through lower taxes or demand stimulation Depends on credit constraints/ financial development


GDP growth and government investment and consumption procyclicality: The explained variable is the growth of GDP per capita. All regressions include the following controls: relative GDP per capita, average years of schooling for the population over 25 years old, trade openness, inflation, population growth, government share of GDP (in %), investment/GDP (in%), terms of trade shock, price shock. GDP growth and government investment and consumption procyclicality Country f.e. Year f.e. Country year f.e. AR(1) 10YRW AR(1) AR(1) (1) (2) (3) (4) -0.239 -0.043 -0.064 -0.180 (0.069)*** (0.022)** (0.034)* (0.065)*** -0.058 -0.038 -0.014 -0.056 (0.032)* (0.020)* (0.019) (0.030)* lag(Private credit/GDP) -0.017 -0.022 -0.007 0.003 (0.010)* (0.009)** (0.004)* (0.010) 0.156 0.029 0.081 0.164 (0.043)*** (0.017)* (0.034)** (0.043)*** Observations 370 304 370 370 R-squared 0.30 0.26 0.43 0.53 Robust standard errors in parentheses * significant at 10%; ** significant at 5%; *** significant at 1% lag(Procyclicality of government consumption) lag(Procyclicality of government investment*Private credit/GDP) lag(Procyclicality of government investment)


Implied growth effects of the procyclicality of government investment: Implied growth effects of the procyclicality of government investment Table calculates the implied growth effect of the government investment becoming acyclical. Region Estimated coef. on lag (Procyclicality of government investment) -Average (lag (Procyclicality of government investment)) Estimated coef. lag(Procyclicality of government investment*privat e credit/GDP) -Average (lag (Procyclicality of government investment)) Average( lag (private credit/ GDP)) Implied effect on growth Panel A: AR(1) EMU -0.1799 * -0.0748 + 0.1642 * -0.0748 * 0.7951 = 0.0037 US -0.1799 * -0.0841 + 0.1642 * -0.0841 * 1.2094 = -0.0016 EMU with US private credit -0.1799 * -0.0748 + 0.1642 * -0.0748 * 1.2094 = -0.0014


Conclusion: Conclusion Macro policy over the cycle matters for growth. Procyclicality of government investment, and to a lesser extent consumption, is harmful to growth of GDP per capita. Less financially developed countries could increase growth substantially by reducing procyclicality of government investment. Effect is particularly strong for EMU.


THE END: THE END


Econometric specification: Econometric specification Problem: how do we estimate a time-varying coefficient on the GDP gap?


GDP growth and government investment and consumption procyclicality: The explained variable is the growth of GDP per capita. All regressions include the following controls: relative GDP per capita, average years of schooling for the population over 25 years old, trade openness, inflation, population growth, government share of GDP (in %), investment/GDP (in%), terms of trade shock, price shock. GDP growth and government investment and consumption procyclicality Country f.e. Year f.e. Country year f.e. AR(1) 10YRW AR(1) 10YRW AR(1) 10YRW (1) (2) (3) (4) (5) (6) -0.239 -0.043 -0.064 -0.062 -0.180 -0.057 (0.069)*** (0.022)** (0.034)* (0.023)*** (0.065)*** (0.023)** -0.058 -0.038 -0.014 0.004 -0.056 -0.036 (0.032)* (0.020)* (0.019) (0.018) (0.030)* (0.018)** lag(Private credit/GDP) -0.017 -0.022 -0.007 -0.003 0.003 -0.002 (0.010)* (0.009)** (0.004)* (0.003) (0.010) (0.008) 0.156 0.029 0.081 0.052 0.164 0.043 (0.043)*** (0.017)* (0.034)** (0.018)*** (0.043)*** (0.018)** Observations 370 304 370 304 370 304 R-squared 0.30 0.26 0.43 0.41 0.53 0.52 Robust standard errors in parentheses * significant at 10%; ** significant at 5%; *** significant at 1% lag(Procyclicality of government consumption) lag(Procyclicality of government investment*Private credit/GDP) lag(Procyclicality of government investment)