CEPAL FISCAL POLICY SEMINAR Blanca Moreno Dodson World Bank

Transcription

1 CEPAL FISCAL POLICY SEMINAR 2010 Blanca Moreno Dodson World Bank

2 Structure of the Presentation Introduction: Motivation Literature Review Methodology Function Specification and Methods Empirical Results Conclusions

3 Motivation Many studies show that public spending is significant for economic growth but the results are still mixed Since many governments are increasing and reallocating public spending to provide economic stimulus in the midst of the global economic crisis, the link between public spending and growth becomes more relevant than ever Alternative public spending allocations involve dynamic tradeoffs in their short and medium term impact on growth

4 Motivation Most studies mix developed and developing countries, and countries with different growth patterns Country sample selection may explain conflicting results Is the link between public spending and growth stronger for countries experiencing higher and sustained growth rates?

5 Literature Review Within an endogenous growth framework Barro introduced public services in the production function (1990): Government expenditure is positively linked to economic growth when the share of government expenditure (and consequently the tax rate) is low, but then turns negative due to increasing inefficiencies as the share of expenditure increases

6 Literature Review Barro and Sala i Martin (1992) introduce the notion that if social returns on public investment are larger than private returns, public investment can increase economic growth The rest of the literature provides conflicting results partly explained by the country and data sample, function specification, and econometric issues

7 Literature Review Grossman (1990): Both positive (higher productivity) and negative (inefficient provision and distortionary effects of taxation): net positive impact Levine and Renelt (1992): Public consumption negative and public investment positive impact Easterly and Revelo (1993): Public investment in communication and transportation, as well as general government investment: positive impact Glomm and Ravikumar (1997): Public spending in health, education and infrastructure: positive impact Bleaney, Gemmell, and Kneller, (1999); and Bleaney, Gemmell, and Kneller (2001) (developed countries): Productive expenditure good for growth but distortionary taxes lower its impact

8 More Recent Empirical Results Folster and Henrekson (2001): The more the econometric problems are addressed, the more robust the link between government size and economic growth gets Gupta, Clements, Baldacci, and Mulas Granados (2002): Public spending (in developing countries), especially capital, has positive impact only if the budget deficit is low Agell, Ohlsson, and Thoursie (2006): There is no robust relationship between growth and the share of government expenditure Park (2006): Comparing productive public investment and lower taxes versus consumption and higher taxes does not lead to any robust results Schaltegger and Torgler (2006): Large public expenditure has a negative impact (high income ) Wahab (2004) and Colombier (2009) (for OECD countries) and Ang (2009) for Malaysia: Only capital expenditure is significant for growth

9 Methodology While most empirical studies in the literature use a heterogeneous sample of countries, Moreno Dodson (2008) includes only fast growing developing countries: the link between total public spending and growth is overall positive with some productive components of public spending being particularly significant Second paper (2009) presented today with N. Bayraktar attempts to validate these results by introducing a comparison country group

10 Methodology The empirical specification introduced in the paper includes the government budget constraint to avoid biases associated with incomplete specification ignoring financing options of governments and budget balance Some public spending categories are left out to avoid perfect collinearity Test of non linearities in the empirical specification (with the government budget constraint)

11 Country Sample A sample including 7 fast growing developing countries and 7 other developing countries whose growth rate patterns have been somehow less stable and consistent during the period of analysis ( ) First set: Korea, Singapore, Malaysia, Thailand, Indonesia, Botswana, and Mauritius: among the top performers in the world in terms of GDP per capita growth (3% and above) Second set: Chile, Costa Rica, Mexico, Philippines, Turkey, Uruguay, and Venezuela

12 Function Specification y ˆ b y + b p + b HC + b FR + b PE + b FS + it = ˆ 1 it 1 2 it 3 i 4 it 5 it 6 it b7 i is the country index t is the year index is the rate of growth of GDP per capita p is the ratio of private investment to GDP (or openness) HC is the initial human capital index CPIINF

13 Function Specification FR is the ratio of total fiscal revenues to GDP PE is the ratio of total public expenditures to GDP FS is the ratio of the fiscal balance (deficit or surplus) to GDP CPIINF is the inflation rate b 1, b 2, b 3, b 4, b 5, b 6, and b 7 are the coefficients assigned to the independent variables.

14 Methodology Unlike other studies testing only the impact of public investment on growth separating it from current spending, this analysis includes total public spending. Evidence that some categories of current spending items are indeed critical to ensure the profitability of investments Operations and maintenance expenditures, are critical to ensure the profitability of infrastructure investments since they can facilitate access and prevent accidents, permitting citizens to arrive safely to markets, schools, hospitals or any other destinations Salaries of teachers, usually classified under the current spending rubric, are closely connected to the quality of education provided In addition, it would not be realistic to try and isolate public investments completely since in many countries capital budgets include de facto, explicitly or implicitly, salaries and current spending items

15 Public Spending Classifications Economic Social Others: GFS standard classification Productive Unproductive Others: a priori definition introduced by Bleaney, Kneller, and Glemmell (2001) Alternative definition excluding defense spending from the productive group

16 Productive a priori Bleaney, Kneller, and Glemmell (2001): public spending items likely to be included in the private sector production function: Productive expenditure: General public services expenditure, Defense expenditure, Educational expenditure, Health expenditure, Housing expenditure, Transportation and communication expenditure Unproductive expenditure: Social security and welfare expenditure, Recreation expenditure, Fuel and energy expenditure, Agriculture, forestry, fishing and hunting expenditure, Mining mineral resources, manufacturing, and construction expenditure, Other economic affairs and services expenditure Specific definition to each country: subject for debate

17 Some Comparative Facts about the Two Groups Similar budget size in both groups (around 36% of GDP) On average, the fast growing group has a much higher share of productive expenditure in total, while unproductive components of public spending are larger for the comparison group Better governance in the first group

18 Econometric Methodology Three alternative econometric methods for panel data are used (OLS, SURE, and GMM) and their results are then compared First OLS and SURE are used, and then the dynamic panel technique (GMM) is applied; the results are compared with those obtained with the static panel regressions

19 Results The influence of public spending on economic growth is clearly different in the two groups, even though the size of government measured as total public expenditures as percent of GDP is similar In the fast growing countries, the productive components of public spending consistently have a positive and statistically significant impact on economic growth, after controlling for macroeconomic and private sector variables, initial conditions of countries, and other budget components

20 Results Private investment and openness, tend to have higher significance in explaining growth for the fast growing group: The expected complementarity effect between public and private sector in the growth process may only be manifested in the first group There are differences in investment climate between the two groups Inflation is negatively correlated with growth only for the fastgrowing group, indicating that reducing inflation leads to faster growth and therefore growth is responsive to improvements in macroeconomic stability in those countries

21 Results When we combine the two groups of countries in the regression analysis, the link between government spending and growth again gets weaker or disappears In the regressions with combined data, when we control the group effects by interactive dummy variables, it can be seen that only the fast growing panel produces a positive and statistically significant link between the productive component of public spending and growth

22 Conclusions Our overall conclusion therefore is that public spending can be a significant determinant of growth, if countries are able to devote a significant fraction of these expenditures to productive uses, in connection with private sector investment Perhaps the use of heterogeneous country samples in previous analyses can explain why many results are inconclusive

23 Future Areas for Research Study fiscal reaction to global crisis and how changes in the composition of public spending has affected growth (data including 2009) Study impact on poverty as same rates of GDP per capita growth can lead to different poverty reduction outcomes and fiscal policy is key to determine this effect

24 Muchas gracias por su atención Agradezco sus comentarios y preguntas