INCOME INEQUALITY AND ECONOMIC GROWTH Eva Kotlánová 1 1 Silesian University, School of Business Administration, Univerzitninam. 1934/3,73340 Karvina, Czech Republic Email:kotlanova@opf.slu.cz Abstract Every human society which is formed in a framework of the specific economic structure is characterised by different poverty rate, social cohesion and income inequality. The aim of this paper is to determine the effect, which has income inequality on economic growth in developed countries. The selected representative sample consists of 34 OECD member countries in the period 2000 2012. The empirical analysis is based on the neoclassical growth model. As the independent variable approximating the effect of income inequality is used the Gini coefficient, one of the best known and used measures of income inequality, prepared in accordance with OECD methodology 1. Significant results of the dynamic panel model confirm the hypotesis that the impact of income inequality on economic growth is negative. Keywords: Income Inequality, Economic Growth, OECD, dynamic panel model Jel Classification: C23, D33, E64, O15 1. Introduction A number of developed countries were forced to cope with the consequences of the economic crisis (2008 2010) for much longer than originally expected. Among other things, the recession led to a growth in unemployment, the decline in household income, the loss of wealth and rapid growth of poverty. All of these factors led to a further growth of income inequality in society. Income inequality creates a space for social and economic discrimination. However, this problem was only accentuated by the economic recession. The statistics of organizations involved in measuring and analysing income inequality at the international level (OECD, World Bank) point out to the fact that the gap between the richest and poorest classes has been increasing for around forty years and is at the highest level in the past 30 years. This is further accentuated by the fact that while the income of the richest grow very rapidly in times of prosperity, incomes of the poorest grow, but slowly and in times of economic stagnation or decline they decrease more rapidly. According to a number of authors, income inequality and its changes affect, inter alia, economic growth. The relationship between these variables was studied by e.g. Kuznets (1955). Although it might seem that income inequality will have negative effect on economic growth (e.g. Alesina and Rodrik, 1994; Owen and Weil 1998; Keefer and Knack, 2002), there are studies that claim the opposite (Kaldor, 1956; Lazear and Rosen, 1981). 2. Effect of income inequality on economic growth The effect of income inequality on economic growth has been addressed by many authors. Some studies have analysed its direct impact on growth variables, others point to the fact that the inefficient distribution of income can also act indirectly, e.g. through institutional factors, political instability etc. The conclusions of the studies are often different, but the majority of them support the hypothesis of a negative effect on economic growth. Growth investments represent one of the variables. In their studies, Benabou (1996), Aghion et al. (1999) argue that if the less wealthy do not want to borrow because of distrust towards the creditors, it will lead to restrictions on their ability to invest. 1 A detailed description is available in Income Distribution Database (http://oe.cd/idd). -288-
Another growth-promoting variable is the human capital. Investments in human capital were dealt with in studies by Galor and Zeira (1993). They emphasize that lower income groups cannot afford costly higher-level education. If this low-income group is comprised of a larger part of the population, low level of education will be reflected in economic performance. Other factors which may lead to income inequality negatively affecting economic growth include political instability and pressure of voters towards inadequate fiscal policy. Alesina and Rodrik (1994) or Persson and Tabellini (1994) postulate that if income inequality in society is high, a low-income voter tends to put pressure on politicians to eliminate this inequality. This situation will lead to tax increases, which can lead to limiting growth. In their studies, Alesina and Perotti (1996), Keefer and Knack (2002) even indicate that, in extreme cases, growing income inequality may lead to political instability and social unrest, which is in turn adversely reflected on economic development. Growth may also be affected by income inequality through non-productive expenditure. The government is forced to deal with high income inequality in society by higher expenditure on social policy (non-productive expenditure) rather than investing in the education and health of the population (productive expenditure). The issue of the effect of productive and non-productive expenditure on economic growth is dealt with e.g. by Machová (2013) or Drobiszová and Machová (2014). On the other hand, some studies show a positive effect of income inequality on economic growth. According to Galenson and Leibenstein (1955) or Kaldor (1961), assuming higher interest rates, a certain concentration of income may lead to a greater accumulation of capital. Assuming higher interest rates, marginal propensity to save is higher in wealthier people than in the poorer ones. Lazear and Rosen (1981) argue that high income inequality may encourage the poorer part of the population to higher labour productivity and increase their willingness to take investment risk assuming a higher rate of return on investment. 3. Trends in income inequality in OECD countries The OECD monitors the evolution of income inequality in its member countries since its inception. Unfortunately, the data base for the early period is poor. According to OECD statistics, the average household income has grown over the last 25 years before the global economic crisis by 1.6% per annum on average. However, most countries also report a rise in income inequality. This is due to the fact that the incomes of the richest 10% of the population grew much faster than the incomes of the poorest 10%. Ways to measure income inequality are numerous; in this case, income inequality is expressed using the Gini coefficient, which is probably the best-known index used for this purpose. The Gini coefficient is derived from the Lorenz curve and essentially compares the cumulative percentage of households (population) with a cumulative share of income that these households receive. It has values from 0 (the country has an even distribution of income in relation to households) to 1 (all income in a society is concentrated with one household or individual). Consequently, this means that the higher the coefficient, the greater the income inequality in a given society (country). -289-
0,50 0,45 0,40 0,35 0,30 0,25 0,20 0,15 Fig. 1. Changes in the Gini coefficient in OECD countries (1985 2013) 1985 2013 Increase Little change Decrease Source: OECD. Figure 1 shows a change in the Gini coefficient in 22 OECD countries, 1985 2013 On average, the Gini coefficient increased over the period by three points (from 0.29 in 1985 to 0.32 in 2013). Seventeen countries showed an increase in income inequality, with Finland, Sweden, Luxembourg, Israel, the United States and New Zealand of more than 5 points. The decline, very slight, occurred only in Turkey (however, it should be noted that in 1985 Turkey had one of the highest income inequalities). In four countries, this quantity remained almost unchanged on average during this period. If we look at the evolution of income inequality across OECD countries in more detail, we find that most of them experienced a decline during the 1990s and a rapid growth in 2000. The subsequent decline was prevented by the economic crisis, which started in 2008 and deteriorated the trends in income inequality. The OECD publishes several types of Gini coefficient. For the purposes of this article, the authors used Gini coefficient before taxes and transfers (GINIb) and the Gini coefficient after the inclusion of taxes and transfers (GINIa). The table in the Annex shows the average values of these quantities in 2000 2012. As the table shows, taxes and transfers have a significant influence on the value of the Gini coefficient. One of the fiscal policy functions is the redistribution of wealth in society, which is to ensure the balancing of income differences. If we exclude these factors, the lowest income inequality is in Korea; however, after factoring in taxes and transfers, the lowest income inequality is in Denmark. How and to what extent fiscal policy tools will be used (or if they will be used at all) will be decided by the government of each country individually. For example, the Czech Republic improved thanks to the effect of fiscal policy, placing fourth (originally eleventh), while the results of Chile indicate that Chile does not use fiscal policy tools in the area of wealth redistribution, or does, but ineffectively. In connection with the selection and use of fiscal policy tools such as taxes and transfers, it should be mentioned that while serving to eliminate income inequality, they can themselves affect economic growth. If, on the one hand, the government increases taxes and, on the other hand, it increases transfers, it can lead to restrictions on activities of economic entities. This issue will be addressed in further research. -290-
4. Methods and data used Empirical analysis of the impact of income inequalities on economic growth is based on panel data, which map 34 OECD member countries 2. These represent a relatively homogeneous sample of available data, which fulfils the condition for the use of regression analysis, as noted e.g. by Barro and Sala-i-Martin (2004). The reference period is 2000 2012. This period was chosen with regard to the possibility of obtaining reliable data for all variables that are included in the model. Since this is a relatively short time series, the authors used panel data regression analysis method and the software E-Views, version 7. This program allows performing all econometric tests, as noted e.g. by Greene (2008). All data was drawn from the OECD database, which is available online (OECD ilibrary). Given that most empirical analyses that deal with the effect on economic growth are based on the neoclassical growth model, or its extended version including human capital (see Mankiw, Romer and Weil, 1992). Both models are alternatively estimated. The equation can be written as: RGDP t = α 0 + α 1 RGDP t-1 + α 2 RINV t + α 3 GINI t + µ t (1) and after the inclusion of human capital, RGDP t = α 0 + α 1 RGDP t-1 + α 2 RINV t + α 3 HC + α 4 GINI t + µ t (2) The dependent variable is the real gross domestic product per capita (RGDP) in USD (PPP). Independent variables then include, based on standard growth theory, the accumulation of physical capital (RINV) approximated by the share of real investment to GDP in PPP per capita, and in the extended version also the accumulation of human capital (HC). When testing the stationarity of time series of all variables using the unit root test according to Levin, Lin and Chu (2002), the authors detected non-stationarity in variables approximating human capital at conventional levels of significance. It was removed by introducing the first difference. Furthermore, with regard to the interpretation of the results, it was necessary to calculate logarithms of the variables. The results presented can be seen as elasticities, which reflect the percentage change in the dependent variable in response to the percentage change in the dependent Since the authors performed the analysis using the dynamic panel model, they estimated it using the generalized method of moments (GMM), and used lagged values of the dependent variable as instruments. The suitability of instruments was tested using the J-test. They also used the Arellano-Bond estimator (Arellano and Bond, 1991), which eliminates the risk of endogeneity of the dependent variable, which is also the independent variable. To adjust the standard deviations for autocorrelation and heteroscedasticity, the "White Period" method was used. 5. Results of the empirical analysis The empirical analysis estimated four dynamic panel models. Due to the insignificance of model results with the inclusion of human capital, the estimate results are presented without its influence. 2 If the analysis covered developed countries as well as the developing ones, averaging may occur, making the results misleading and biased. Such conclusions were made by Barro (2000) when he dealt with the determinants of growth. Therefore, he divided countries to the developed and developing ones and described the different effect of income inequality on economic growth in these two groups. -291-
The results of both tested models confirm the inverse relationship between economic growth and income inequality, which is statistically significant. If income inequality grows, there is a drop in economic growth. In other words, if we manage to reduce inefficient distribution of income in society, it should have a positive effect on economic growth. Also, the influence of the standard variable (investments in our case) is in accordance with the theories of growth in both models. Table 1 shows the results of the model, which used the Gini coefficient before taxes and transfers (GINIb) as the independent variable. Table 1 The effect of income inequality on economic growth in OECD countries 2000 2012 (Gini coefficient before taxes and transfers) Dependent variable ln(rhdp) Variable Coefficient St. Error t-statistic ln(rgdp-1)) 0.790*** 0.007 113.653 ln(rinv) 0.167*** 0.003 49.679 ln(ginib) -0.268*** 0.016-16.421 J-statistic 33.748 Instrument rank 34 NB: *, **, ***; they represent the levels of significance of 10%, 5% and 1%. Source: Authors' own calculations. The results of the model with Gini coefficient after the inclusion of taxes and transfers as the independent variable are summarized in Table 2. Table 2 The effect of income inequality on economic growth in OECD countries 2000 2012 (Gini coefficient after taxes and transfers) Dependent variable ln(rhdp) Variable Coefficient St. Error t-statistic ln(rgdp-1)) 0.787*** 0.009 85.658 ln(rinv) 0.186*** 0.005 40.421 ln(ginia) -0.053*** 0.015-3.576 J-statistic 32.699 Instrument rank 34 NB: *, **, ***; they represent the levels of significance of 10%, 5% and 1%. Source: Authors' own calculations. In both models, the increase in income inequality affects economic growth in the expected direction, i.e. negatively. However, if we compare the results in terms of the Gini coefficient before and after the inclusion of taxes and transfers, it is evident that income inequalities which are not corrected by fiscal policy tools have far greater impact on economic growth. It can therefore be concluded that the correction of income inequality through taxes and transfers makes sense with regard to its impact on economic growth. The question is whether the fiscal policy tools used directly affect economic growth, as noted above. 6. Conclusion -292-
Statistical data (OECD and other) show that in recent decades, income inequality has increased in all types of economies. The aim of this paper was to determine the effect income inequality has on economic growth. Some earlier studies did not allow a clear conclusion to be made, their results being inconclusive or insignificant. It was often due to a large sample of inhomogeneous countries included in the model, or to the inconsistency of the data. Therefore, the authors selected a relatively homogeneous sample of OECD member countries (34) in the timeframe of 2000 2012. In addition to standard growth variables, the authors used a control variable approximating income inequalities, this variable being the Gini coefficient. Significant results of the dynamic panel model confirms the conclusions of those authors who believe that the impact of income inequality on economic growth is negative. Following the implementation of the Gini coefficient before and after taxes and transfers into the model, it became evident that the government's effort to correct income inequalities in the society using fiscal policy tools may have a positive effect on economic growth. This conclusion is derived from the results, which show that the effect of the Gini coefficient before taxes and transfers on economic growth is much stronger than the one after taxes and transfers. On the other hand, it is also necessary to take into account that the use of standard fiscal policy tools such as taxes and transfers can have a direct negative impact on the economic performance of the country. 7. Acknowledgement This paper was supported by the project SGS/13/2015 "Influence of Selected Macroeconomic and Microeconomic Determinants on the Competitiveness of Regions and Firms in Countries of the Visegrad Group Plus". References [1] Aghion, P., Caroli, E. and C. García-Penalosa, 1999. Inequality and economic growth: The perspectives of the new growth theroies. Journal of Economic Literature, vol. 37, issue 4, pp. 1615-1660. [2] Alesina, A. and R. Perotti, 1996. Income distribution, political instability and investment, European Economic Review, vol. 40, issue 6, pp. 1203-1228. [3] Alesina, A. and D. Rodrik, 1994. Distributive politics and economic growth. Quarterly Journal of Economics, vol. 109, issue 2, pp. 465-490. [4] Arellano, M. and S. Bond, 1991. Some tests of specification for panel data: Monte Carlo evidence and an application to employment equations. The Review of Economic Studies, vol. 58, issue 2, pp. 277-297. [5] Barro, R. J. 2000. Inequality and growth in panel of countries. Journal of Economic Growth, vol. 5, issue 1, pp.5-32. [6] Barro, R. J. and X. Sala-i-Martin, 2004. Economic Growth. Cambridge and London: MIT Press. [7] Benabou, R. 1997. Inequality and Growth. NBER Working Paper No. 5658. [online] [cit. 2015-05-27]. Available from Internet: http://www.nber.org/papers/w5658.pdf. [8] Drobiszová, A. and Z. Machová, 2014. The appropriate mix of government spending and taxation: How to promote economic growth? In Soliman, K. S. (ed.). Vision 2020: Innovation, Development Sustainability, and Economic Growth, vols. 1-5, pp. 1882-1890. IBIMA. -293-
[9] Galenson, W. and H. Leibenstein, 1955. Investment criteria, productivity and economic development. Quarterly Journal of Economics, vol. 69, issue 3, pp. 343-370. [10] Galor, O. and J. Zeira, 1993. Income distribution and macroeconomics. Review of Economic Studies, vol. 60, issue 1, pp. 35-52. [11] Greene, W. 2008. Econometric Analysis. New Jersey: Prentice-Hall. [12] Kaldor, N. 1956. Alternative theories of distribution. The Review of Economic Studies, vol. 23, issue 2, pp. 83-100. [13] Kaldor, N. 1961. Capital accumulation and economic growth. In Lutz, F. A. and. D. C. Hague (eds.). The Theory of Capital. New York: St. Martin's Press. [14] Keefer, P. and S. Knack, 2002. Polarization, politics and property rights: Links between inequality and growth. Public Choice, vol. 111, pp. 127-154. [15] Kuznets, S. 1955. Economic Growth and Income Inequality. The American Economic Review, vol. 45, issue 1, pp. 1-28. [16] Lazear, E. P. and S. Rosen, 1981. Rank-Order Tournaments as Optimum Labor Contracts. Journal of Political Economy, vol. 89, No. 5, pp. 841-864. [17] Levin, A., Lin, C. F. and C. Chu, 2002. Unit root tests in panel data: Asymptotic and finite-sample properties. Journal of Econometrics, vol. 108, issue 1, pp. 1-24. [18] Machová, Z. 2013. Pro-growth effects of (Un)productive government spending in the OECD. In Machová, Z. (ed). Proceedings of the 3rd International Scientific Conference Taxes in the World, pp. 103-112. VSB-Technical University of Ostrava. [19] Mankiw, N. G., Romer, D. and D. N. Weil, 1992. A contribution to the empirics of economic growth. The Quarterly Journal of Economics, vol. 107, issue 2, pp. 407-437. [20] Owen, A. and D. N. Weil, 1998. Intergenerational earnings mobility, inequality and growth. Journal of Monetary Economics, vol. 49, issue 1, pp. 71-104. [21] Persson, T. and G. Tabellini, 1994. Is inequality harmful for growth. American Economic Review, vol. 84, issue 3, pp. 600-621. -294-
Appendix Ranking of OECD countries according to the average Gini coefficient values (2000 2012). OECD Country Average value 2000 2012 GINI after tax and GINI before tax and transfers transfers OECD Country Korea 0.337 0.237 Denmark Switzerland 0.371 0.242 Slovenia Iceland 0.386 0.251 Sweden Denmark 0.419 0.260 Czech Republic Netherlands 0.421 0.260 Slovakia Norway 0.428 0.261 Norway Sweden 0.433 0.264 Finland Slovakia 0.437 0.269 Iceland Canada 0.440 0.272 Austria Slovenia 0.452 0.274 Luxembourg Czech Republic 0.462 0.274 Belgium New Zealand 0.462 0.279 Netherlands Japan 0.464 0.286 Germany Australia 0.466 0.287 Hungary Turkey 0.468 0.288 Switzerland Luxembourg 0.470 0.291 France Mexico 0.472 0.309 Korea Spain 0.474 0.310 Ireland Finland 0.480 0.317 Canada Austria 0.483 0.322 Australia Estonia 0.485 0.326 Japan Hungary 0.485 0.328 Estonia Belgium 0.488 0.329 Italy Italy 0.491 0.329 Spain USA 0.493 0.332 New Zealand France 0.494 0.336 Poland Germany 0.498 0.338 Greece Israel 0.501 0.339 United Kingdom Greece 0.507 0.362 Portugal United Kingdom 0.507 0.375 Israel Poland 0.509 0.376 USA Portugal 0.522 0.420 Mexico Chile 0.535 0.421 Turkey -295-
Ireland 0.538 0.509 Chile Source: OECD, authors' own calculations. -296-