Income Inequality and Economic Growth Revisited. A Note

Income Inequality and Economic Growth Revisited. A Note Mª Ángeles Caraballo, University of Seville (Spain), [email protected] Carlos Dabús, Universidad Nacional del Sur (Argentina) and CONICET, [email protected] Fernando Delbianco, Universidad Nacional del Sur (Argentina) and CONICET, [email protected] Abstract The relationship between income inequality and economic growth is estimated using dynamic panel estimation on a sample covering 112 emerging countries for the period 1980-2014. The results show that income inequality has a positive influence on economic growth for richer countries, in line with the classic theory, and a negative effect for poorer countries, as argued by the political economy approach. Keywords: Economic Growth, Income Inequality, Emerging Countries. JEL Codes: O11, O15, O47, C23 1. Introduction The income inequality-economic growth relationship has been widely analysed in the literature. However, no consensus regarding the effect of inequality on growth has yet been reached. The theoretical literature presents two main approaches. The classic theory postulates a positive association between inequality and growth through the saving rate (Stiglitz, 1969). The political economy approach holds that inequality harms economic growth through social instability (Alesina and Perotti, 1996) and credit market imperfections (Galor and Zeira, 1993). Both approaches find empirical support (Delbianco et al., 2014; Lee and Son, 2016; and references therein). An increasing empirical literature shows that the inequality-growth relationship can be either positive or negative, depending on the different scenarios considered or on the properties of the distribution of income (Voitchovsky, 2005; Delbianco et al., 2014; and Lee and Son, 2016). The hypothesis in this paper is that the effects of inequality on growth depend on the stage of development of each economy. Thus, at early stages, the effects are negative because the political economy approach holds. At higher stages they are positive since the classic theory prevails. We assume that, in highly developed countries, the needs of lower-income people are less pressing, and therefore higher inequality does not cause as much social unrest as it could generate in less developed economies, but that it favours greater savings and investment by the rich population, which in turn fosters economic growth. To test our hypothesis, we divide the sample according to the level of income, the degree of income inequality, monetisation as a proxy of credit constraints, political stability, and the level of savings. For each subsample we analyse the inequality-economic growth relationship. Section 2 presents the data and the variables. In Section 3 the model specification and empirical results are shown. 2. Data and variables Our sample covers 112 emerging countries during the period 1980-2014. All data were obtained from the World Bank. To overcome the unbalanced panel, the data is transformed into a five-year average. The explained variable is growth rate (GROWTH) of the Gross Domestic Product (GDP). We use the Gini coefficient (GINI) as an indicator of income inequality. As a proxy of credit constraint, we use the degree of monetisation measured by the percentage of the monetary aggregate M2 over the GDP (M2). For political stability (PS), we use the Political Stability and Absence of Violence Indicator. Savings serve as a proxy for investment. The control variables are those found as robust by Levine and Renelt (1992): the investment/GDP ratio (INV), education (HK) measured as the average number of years of total schooling (15 years or more), openness measured as the sum of exports plus imports to GDP (OPEN) and population growth (POP) (Summary Statistics in Appendix). Table 1. Mean of Initial Income HIGH LOW INITIAL INCOME 5558.99 680.49 Income Equality 3763.84 2878.20 M2 5836.48 4164.67 Political Stability 3580.30 1953.54 INV 3666.009 2410.4 Note: High: those countries with values above the median. Low: those below the median 3. Methodology and results To test our hypothesis, for each subsample in Table 1 we estimate: where is the log of GDP of country i in period t, and Xi,t denotes the set of control variables (current or lagged. Our variable of interest is the Gini coefficient, lagged one period. Finally, is the error term and α, β and δ are the set of parameters to be estimated. The temporal unit is the mean quinquennial observations, starting in 1980-1984. The lagged HK and GINI coefficients are included because their impact on the real economy may take longer than five years to take effect (Voitchovstky, 2005). To estimate the above expression, the Arellano-Bond specification (Arellano and Bond, 1991; Roodman, 2015) is used. Lagged variables and quinquennial dummies were used as instruments. The results appear in Table 2. Table 2. Dynamic Panel results Initial Income Income Equality M2 Political Stability INV Dependent Variable: GROWTH Total HIGH LOW HIGH LOW HIGH LOW HIGH LOW HIGH LOW (1) (2) (3) (4) (5) (6) (7) (8) (9) (10) (11) GROWTH (-1) 0.08 -0.10 -0.06 0.03 -0.33** -0.24* -0.34 -0.06 -0.25 -0.04 -0.12 (0.10) (0.13) (0.16) (0.10) (0.16) (0.14) (0.24) (0.12) (0.19) (0.17) (0.18) GINI(-1) 0.26 0.56** 0.16 0.48** 0.03 -0.08 -0.14 0.70*** -0.25** 0.05 0.17 (0.16) (0.26) (0.13) (0.19) (0.14) (0.11) (0.29) (0.23) (0.12) (0.10) (0.19) INV 0.24** 0.37** 0.03 0.12 0.39*** 0.29*** -0.47 0.25** 0.08 0.10 0.40* (0.10) (0.18) (0.10) (0.10) (0.11) (0.10) (0.35) (0.12) (0.17) (0.08) (0.21) HK(-1) -0.45 0.81** 0.57 0.82 0.41 -0.39 4.87* 0.92 1.10** 0.18 2.18** (0.40) (0.39) (0.73) (0.55) (0.39) (0.32) (2.65) (0.76) (0.52) (0.31) (0.96) POB -1.52 -0.47 0.51 0.03 -0.67 -2.62*** 8.26** -1.90 -1.15 -1.06 5.66* (0.95) (1.21) (1.50) (0.80) (1.99) (0.57) (4.05) (1.67) (1.38) (0.83) (3.31) OPEN 0.06* -0.04 0.06** 0.03 -0.01 0.00 0.17** -0.04 -0.01 -0.02 0.21*** (0.04) (0.03) (0.03) (0.04) (0.04) (0.03) (0.09) (0.04) (0.03) (0.02) (0.08) Observations 185 107 78 80 105 135 50 100 85 105 80 Countries 70 38 32 35 49 59 29 49 34 51 40 Hansen Overid (p-value) 0.253 0.458 0.641 0.543 0.591 0.550 0.896 0.581 0.542 0.560 0.651 Hansen Excluding Group (p-value) 0.299 0.336 0.631 0.366 0.474 0.508 0.852 0.533 0.418 0.693 0.538 Hansen Exogenous (p-value) 0.260 0.778 0.469 0.880 0.755 0.521 0.703 0.542 0.760 0.187 0.724 Notes: Robust standard errors in parentheses. *** p<0.01, ** p<0.05, * p<0.1 The Hansen test does not reject the hypothesis of no over-identification and exogeneity of the instruments, thereby validating the results obtained. Columns (2) and (3) show that income inequality is positive and significant only for richer countries, as stated by the classic theory. These results hold when the subsample is divided in terms of income equality and political stability. For income equality, columns (4) and (5) show that GINI is significant for those countries with higher income equality, which could indicate the existence of a threshold beyond which inequality encourages growth. For PS, columns (8) and (9) show that income inequality is positive and significant for stable (richer) countries, and negative and significant for less stable (poorer) countries, as argued by the political economy approach. When the sample is divided in terms of M2 and INV, income inequality is not significant. These results support our hypothesis of a positive inequality-growth relationship for richer countries, in accordance with the classic theory, and of a negative relationship for countries with lower income (i.e. politically unstable countries), as predicted by the political economy approach. References Alesina, A. and Perotti, R. (1996): Income distribution, political instability, and investment. European Economic Review, 40(6), 1203-1228. Arellano, M. and Bond, S. (1991): Some tests of specification for panel data: Monte Carlo evidence and an application to employment equations. Review of Economic Studies, 58(2), 277-297. Delbianco, F., Dabús, C. and Caraballo, M.A. (2014): Income inequality and economic growth: new evidence from Latin America. Cuadernos de Economía, 33 (63), 381-397. Galor, O. and Zeira, J. (1993): Income distribution and macroeconomics. Review of Economic Studies, 60(1), 35-52. Lee, D.L. and Son, J.C. (2016): Economic Growth and Income Inequality: Evidence from dynamic panel investigation. Global Economic Review, 45 (4), pp. 331-358 Levine, R. and Renelt, D. (1992): A sensitivity analysis of cross-country growth regressions. American Economic Review, 82(4), 942-963. Roodman, D. (2015). xtabond2: Stata module to extend xtabond dynamic panel data estimator. Statistical Software Components. Stiglitz, J. (1969): Distribution of income and wealth among individuals. Econometrica, 37(3), 382-397. Voitchovsky, S. (2005): Does the profile of income inequality matter for economic growth? Distinguishing between the effects of inequality in different parts of the income distribution. Journal of Economic Growth, 10(3), 273-296. APPENDIX Variable Obs. Mean Std. Dev. Min. Max. Economic Growth 710 1.71 4.34 -30.26 33.50 Initial Income 770 3120 4879 143 36069 Gini 425 40.37 9.95 19.40 74.33 M2 673 38.44 24.73 1.23 185.71 PI 449 -0.52 0.88 -3.09 1.40 Investment 679 21.57 8.10 2.53 65.93 HK 616 5.44 3.08 0.05 13.16 POP 787 1.88 1.40 -4.64 14.93 OPEN 693 74.58 38.19 0.67 310.58