Asian Journal of Economics and Empirical Research ISSN: 2409-2622 Vol. 2, No. 2, 62-73, 2015 http://asianonlinejournals.com/index.php/AJEER 62 Inequality and Unemployment Management Kwangsik Jung 1 1 Instructor, Sisa SLS Foreign Language Academy, South Korea Abstract This work is licensed under a Creative Commons Attribution 3.0 License Asian Online Journal Publishing Group Contents 1. Introduction ............................................................................................................................................................................... 63 2. Inequality and Unemployment ................................................................................................................................................. 63 3. Definition of the Middle Class .................................................................................................................................................. 64 4. Malmquist Index ....................................................................................................................................................................... 66 5. Data ............................................................................................................................................................................................ 67 6. Results ........................................................................................................................................................................................ 68 7. Conclusion .................................................................................................................................................................................. 71 References ...................................................................................................................................................................................... 71 Many studies have claimed that inequality and unemployment should be reduced. They also assert that the middle class should be increased for economic development. These strategies sound like similar outcome strategies. However, what would happen if they were not? In other words, if reducing inequality and unemployment risks aggravating the middle class, how should we manage inequality and unemployment? In this paper, we will examine this. Keywords: Inequality, Unemployment, Middle class, Malmquist index. Productivity index, Data envelopment analysis http://creativecommons.org/licenses/by/3.0/ Asian Journal of Economics and Empirical Research, 2015, 2(2):62-73 63 1. Introduction Inequality and unemployment are key topics for the social stability and economic growth of a country. But what do inequality and unemployment management mean? Which countries manage them well? Most people may answer based on value of the Gini coefficient and the unemployment rate. For example, if a country’s Gini coefficient and unemployment rate are low, people may say that inequality and unemployment are managed well. However, this is only partly correct. The Gini coefficient and the unemployment rate do not consider societal structure. In other words, various social structures can exist under the same inequality and unemployment levels. Therefore, inequality, unemployment, and social structure should be considered comprehensively. For this study, we think of inequality and unemployment as resources. Through this approach, their relationship with social structure is explained, and the definition of ideal inequality and unemployment management is redefined. The paper is organized as follows. In the following section, the concept of this paper is explained, and ideal inequality and unemployment management are theoretically reviewed. In the second half of the paper, the management of inequality and unemployment of 49 countries during, before, and after the financial crisis is analyzed. 2. Inequality and Unemployment In this paper, we handle the inequality and unemployment as the resource, and use the terms “inequality resource” and “unemployment resource”. According to the OECD, “income inequality has a negative impact on subsequent growth”. Federico (2014) From the Okun’s law, we know that there is a negative relationship between GDP growth and unemployment. Therefore, these resources have a negative impact on society. While it is better not to use them, this is impossible. Each country uses these resources to a certain degree. To estimate the consumption of inequality resources and unemployment resources, the Gini coefficient and unemployment rate are used here. Ideal stratification: When inequality and unemployment resources are allocated to society, people’s income level is seen to be diversified, and unemployment occurs. This means that the society is divided into three groups (upper, middle, and lower class). In other words, each country produces stratification by spending inequality and unemployment resources. The outcome of consuming inequality and unemployment resources is stratification. Therefore, we need to consider the structure of stratification because consuming the same amount of resources does not guarantee the same structure of stratification. From the same Gini coefficient and unemployment rates, various Lorenz curves can be generated. Therefore, we pursue ideal stratification. What is ideal stratification? It differs depending on the goal of each country. If each country’s goal is economic development and social stability, then ideal stratification means that there is a relatively large middle class because the middle class is strongly related to economic development and social stability. Kristin Forbes (2000); William (2001) claimed that a “higher share of income for the middle class is associated with higher income and higher growth.” Landes and David (1998) also noted that the “ideal growth and development of society” is related to “a relatively large middle class.” In Africa, strong economic growth has been accompanied by the emergence of a sizeable middle class over the past two decades. Mthuli et al. (2011) Castellani Francesca and Gwenn (2011) and Parent also argue a robust middle class is related to economic and social stability and to better development prospects. In conclusion, the size of the middle class is important. From now on, we will focus on the middle class and use the term “middle class production.” Efficiency: To explore the consumption of inequality and unemployment resources, and middle class production, the concept of efficiency is used. The term efficiency is often used in the input-to-output ratio. Abraham et al. (1978) there are two types of efficiency (absolute or relative efficiency). However, in the case of absolute efficiency, it is almost impossible to know the theoretical levels of efficiency. Therefore, relative efficiency is often used and is used in this paper. Inequality and unemployment resources correspond to the input variable, and the output variable is middle class production. Therefore, the management ability of the inequality and unemployment resources is represented as ( ) ( ) There are two types of efficiency (input-oriented and output-oriented efficiency). William Cooper et al. (2007) Among them, input-oriented efficiency is defined as that the ability to use a minimal amount of input to generate a given level of output. In this paper, the input-oriented efficiency is used to estimate the ability to manage inequality and unemployment resources. Therefore, the goal of ideal inequality and unemployment management is to achieve the lowest inequality and unemployment with a given middle class size. Theoretical review: For simple explanation, we will use only the inequality resource in this section. Let us consider two Lorenz curves described in Figure 1. Middle class refers to households whose income falls between a% and b% of income distribution. At this time, the middle class on Lorenz curve (1) is bigger than that on Lorenz curve (2) 1 . However, the Gini coefficient of Lorenz curve (1) is bigger. Therefore, a little inequality does not necessarily guarantee a larger middle class. In other words, the risk of weakening the middle class exist in polices intended to reduce inequality. 1 D%-A% > C%-B% Asian Journal of Economics and Empirical Research, 2015, 2(2):62-73 64 Figure-1. Paradox example in inequality and middle class Which Lorenz curve is better? We can answer is found in the terms “consumption of inequality resource” and “middle class production.” They can be illustrated as efficiency; the greater the value the better. ( ) ( ) Figure-2. Concept of ideal inequality management In this paper, ideal inequality management means achieving a minimum consumption of inequality resources with a given middle class size. In Figure2 the middle class of the two Lorenz curve are same in a% - b% in terms of income distribution. However, the Gini coefficient of Lorenz curve (2) is smaller than Lorenz curve (1). When Lorenz curve (1) is closer to Lorenz curve (2), the efficiency of Lorenz curve (1) increases. By definition of relative efficiency, the maximum attainable efficiency of Lorenz curve (1) is achieved when it matches Lorenz curve (2). In conclusion, inequality should be managed to achieve Lorenz curve (2). 3. Definition of the Middle Class There is no consensus on a definition of the middle class. Craig Elwell (2014) There are two main approaches (income or consumption based), and each includes a number of sub-definitions. In this paper, we adopt the income- based approach has been used to cope with the Gini coefficient representing income distribution. The income-based definition is also divided into four sub-approaches (PPP-based, distribution-based, median income-based, and poverty-line based). In this study, we will use the distribution-based and median income-based definitions have been used. The former is a relative approach and assumes the relative position with regard to national income distribution. The latter is an absolute approach and considers fixed income ranges. In this study, the two definitions are presented as follows. The distribution-based middle class refers to households whose income falls between the 20th and 80th percentile of income distribution. In addition, a median income-based middle class refers to households whose income ranged within 50 - 200% of the median income. Although a distribution-based middle class can be estimated by a simple calculation, and estimating the median income-based middle class needs additional calculations, as described in the following section. Asian Journal of Economics and Empirical Research, 2015, 2(2):62-73 65 3.1. Estimating the Median Income-Based Middle Class General quadratic Lorenz curve: To estimate the median income-based middle class, we should estimate the Lorenz curve by year for each country. There are two types of Lorenz curves: the general quadratic Lorenz curve (GQ Lorenz curve) and the beta Lorenz curve (Jose and Barry Arnold, 1984; Gaurav, 1998); both are accurate. As the GQ Lorenz curve is more convenient to calculate, we will use it in this study. The equation of the GQ Lorenz curve is below. ( ) [ ( ) ] where, ( ) ( ) ( ) ( )⁄ ( ) ( )⁄ All parameters are estimated from the percentile income distribution data. In the actual calculations, the World Bank’s POVCAL software is used. Head-count index (H): After estimating the GQ Lorenz curve, the headcount index [ ( ⁄ )*( ⁄ ) + ⁄ ] Is used and the middle class size is estimated. It is a method to calculate the ratio of people under a certain income level. Table 1 shows the median income in each year of the three groups according to income level. In addition, Figure 3 shows the median income difference between each group. During 2005-2012, the average median income of high-income countries is about five times larger than that of the middle-income countries. In addition, the average median income of middle-income countries is about two times greater than that the low-income countries. Table-1. Median Incomes (US$ per household, at current prices) High income country Middle income country Low income country 2005 81718.3 10859.9 6600.2 2006 86520.1 12086 6789.7 2007 74736.7 14517.7 7651.5 2008 80019.4 17651.4 8960.1 2009 79804 16504.2 8629 2010 96811 19254.7 9128.2 2011 11562.1 22646.8 10011.3 2012 119550.1 24137.8 10247.5 Average 91847.5875 17207.3125 8502.1875 The middle class estimated range is 50-200% of each median income. In the case of high- and middle-income countries, they are rounded to the nearest hundredth, and in the case of low-income countries, they are rounded to the nearest tenth. The difference between high-income countries and the other two groups is greater than the difference between middle- and low-income countries. The median income difference between high-income countries and the other two groups decreased before and after the financial crisis. However, since 2009, it has sharply increased. On the other hand, between the middle- and low-income countries, it gradually increased, but the difference is not large. In conclusion, the financial crisis created a significant income gap between high-income countries and other countries. Figure-3. Median income difference (US$ per household, at current prices) Asian Journal of Economics and Empirical Research, 2015, 2(2):62-73 66 4. Malmquist Index The Malmquist index shows the change in efficiency over the two periods. The Malmquist index can be decomposed into two components: the catch-up effect and the frontier-shift effect.(Michael James, 1957) Catch-up effect (CU) shows the change in distance from the efficient frontier, namely the technical efficiency change (Rolf et al., 1985); (Rolf et al., 1990); (Rolf et al., 1992); (Rolf et al., 1994). In this paper, efficiency is defined as a country’s ability to manage inequality and unemployment resources under a certain sized middle class. CU refers to changes in this management ability. Therefore, an increase in CU means an improvement in management ability. The country can produce the same middle class size from lower resource consumption. Figure-4. Concept of CU in this paper In Figure 4 the middle class of A and B are same. Over the two periods (t, t+1), B is constant. Although the middle class of A has not changed, inequality is reduced. In all periods, B is more efficient than A. When it is compared to B, although A is inefficient, the efficiency of A improves between period t and t+1. At this time, CU of A is bigger than 1, and we say it is relatively efficient. The frontier-shift effect (FS) shows the shift of the efficient frontier, namely the technical change. This refers to a change in the attainable minimum level of input with a given level of output. FS can be interpreted as the external influence. In this paper, FS mainly refers to the impact of the financial crisis. Figure-5. Concept of FS in this paper In Figure 5 over the two periods, although A is constant, the inequality of B decreased. B is always more efficient than A. A should reduce inequality to improve efficiency. By definition of relative efficiency, A can improve efficiency to the point where it matches B. In period t, the maximum efficiency of A can be attained when it matches the Lorenz curve of B (t). Using the same logic, the attainable maximum efficiency of A is the efficiency of the Lorenz curve of B (t+1) in period t+1. This means A has a greater possibility to reduce inequality and improve efficiency in period t+1 than t. Therefore, period t+1 is more favorable to A than period t. At this time, we say the efficient frontier is shifted upward, and the value of FS is bigger than 1. Productivity considers the CU and FS. The whole period is divided into several sub-periods. In this paper, the whole period (2005-2012) is divided by three-year sub-period, and there are five sub-periods. CU, FS and productivity are calculated for each period. The average productivity of all sub-periods is MI. We explained the concept of CU, FS, and productivity in this paper by using the Lorenz curve. It is necessary to review the Malmquist index more deeply because the concept of unemployment is added and all variables are changed. Figure 6 below illustrates the construction of the Malmquist index, which uses the inputs, x and x+1 in periods, t and t+1 to produce the output y and y+1, respectively. The efficient frontier of the first and second periods is, the CF and BE. is the position of the first period, and is the position of the second period. Asian Journal of Economics and Empirical Research, 2015, 2(2):62-73 67 Figure-6. A Malmquist index The catch-up effect from the first period to the second period is represented below: ⁄ ( ) If CU > 1, it means it is relatively efficient. Oppositely, CU < 1 means it is relatively inefficient. In Figure 6 the reference point of moves from C to B from the first period to the second. Therefore, the frontier-shift effect of is expressed as following. ( ) This can be rewritten as the following equation. ⁄ ( ) Similarly, the frontier-shift effect of is represented as following. ⁄ ( ) According to Rolf et al. (1994) the frontier-shift effect is defined as the geometric mean of and , namely √ ( ) If FS > 1, it means the efficient frontier is shifted upward. As a result, MI is the product of the CU and FS. If MI > 1, it means productivity growth, and MI < 1 means a productivity decline. √ ( ) 5. Data For a time series analysis, data sets from 2005 to 2012 are used. All data are taken from Euromonitor 2 . However, the middle class size of each country is calculated by the present author, using the household income distribution data. In total, 49 countries are analyzed in this study. 3 To increase the accuracy of the analysis, they are divided into three groups depending on the income level. Although we followed the classification of Euromonitor, Taiwan has been regarded as exceptional. According to Euromonitor, Taiwan belongs to the high-income countries. However, when emerging and developed countries are analyzed together, the emerging countries tend to be overestimated. Therefore, Taiwan is analyzed in the group of middle-income countries in this study. 5.1. Input and Output For the analysis, two input variables (inequality and unemployment resources) are used. Table 4 in Appendix shows the statistics of the input variables. On the other hand, the middle class size is the output variable. Statistics of output variable are illustrated in Table 5. 2 World Consumer Income and Expenditure patterns 2014 3 It is ideal to handle all of the world’s major countries. However, in the case of oil-producing countries, due to the peculiarities of their economic structures, the result has been distorted. For this reason, they are excluded from this study. It is desirable to study separately the case of the Middle East. Asian Journal of Economics and Empirical Research, 2015, 2(2):62-73 68 Two types of middle classes (distribution-based and median income-based) are used for the analysis. A distribution-based middle class (DB middle class) is defined as households not including the poorest 20% and the richest 20%. In addition, a median income-based middle class (MB middle class) means households whose income is within 50-200% of the median income. Table 2 shows the correlations of two types of middle classes. There is a high positive correlation between the DB middle class and MB middle class in high-income countries. However, the lower the income levels, the smaller the correlation coefficient. In the case of low-income countries, the correlation coefficient is less than two. This means the results are greatly varied depending on the middle class definition when studying the middle classes of low-income countries. Table-2. Correlation of the middle class sizes High income country Middle income country Low income country 2005 0.891349315 0.558145854 0.081098545 2006 0.852985495 0.570599265 0.125025219 2007 0.81352883 0.607218803 0.225812519 2008 0.864168942 0.687132048 0.162813883 2009 0.845014362 0.68883036 0.061711886 2010 0.845800916 0.583921509 0.251772797 2011 0.757251884 0.578693627 0.300136708 2012 0.715479671 0.531436074 0.293427904 Average 0.823197421 0.600747193 0.187724933 6. Results Three groups are analyzed two times in accordance with the output items. For each analysis, a 2-input (Unemployment rate and Gini coefficient) 1-output (DB middle class or MB middle class) model are used. In the actual calculations, the DEA-Solver-PRO(Professional Version 10.0) software is used. 6.1. The Correlation Coefficient between the Results The correlation coefficient between the productivity results is also large in the case of high- and middle-income countries, as it is greater than 0.9, see Table 3. On the other hand, in the case of low-income countries, it is only 0.49. This can be thought of in connection with Table 2. In the case of low-income countries, the middle class size is different according to the definition. As well, it affects the productivity results, the details of which will be mentioned in section 6.6. Table-3. The correlation coefficient between the results High income country Middle income country Low income country 0.937341 0.921121 0.498141 6.2. Standard Deviation The standard deviation of the analysis results shows a remarkable fact (see Table 6 in Appendix). First of all, the FS in most countries is less than 1. This means the attainable minimum level of resource consumption has decreased. In other words, there is an unfavorable change in the situation. In all analyses, a standard deviation of the FS is very small compared to that of the CU. It means, the MI of each country is determined greatly by the CU. In conclusion, although negative influences existed during the financial crisis period, they had little effect on productivity. Therefore, if a country’s productivity decreased, the main reason was poor management of inequality and unemployment. 6.3. U-Curve Most countries show a U-curve over five sub-periods of analyzing MI. Figures 7 and 8 show the average productivity change of each group. Figure-7. The Productivity change (DB middle class output) Asian Journal of Economics and Empirical Research, 2015, 2(2):62-73 69 Figure-8. The Productivity change (MB middle class output) Overall, except for the low-income countries in the case of the MB middle class output, all are shown in the form of a U-curve. This means productivity was reduced until the financial crisis, but it recovered after the financial crisis. High and middle-income countries have shown the typical U-curve form in the two analysis results. In both results although middle-income countries have recovered productivity at greater than 1, high-income countries have not reached 1. Figure-9. Productivity Change of Greece and Netherlands In particular, although most high- and middle-income countries appear in U-curve form, Greece and the Netherlands have declined their productivity continuously. Figure 9 shows the productivity changes between the two countries. In both analyses, low-income countries are quite remarkable. In most of the sub-periods, the average values of low-income countries are greater than 1. In particular, when the output is the MB middle class, they are greater than 1 in all sub-periods. Furthermore, they sharply increased after the financial crisis. This was possible thanks to remarkable economic growth in China, about which we will mention later. 6.4. High-Income Countries Figure 10 shows that High-income countries show generally low MI. The MI of only three countries among high-income countries is greater than 1 in two analyses at the same time. However, in the case of Austria and Finland, it is hard to say “growth,” because their values are almost 1. Therefore, Germany is the only country we can say exhibited “growth” among high-income countries. Figure-10. MI (High-income countries) Asian Journal of Economics and Empirical Research, 2015, 2(2):62-73 70 On the other hand, five low MI countries demonstrated the relationship between input and MI. Four countries except the USA among the five countries consumed unemployment resource excessively. Figure 11 shows the trend of unemployment rates in four countries. Figure-11. Unemployment rate (Denmark, Greece, Ireland, Spain) On the other hand, the USA consumed inequality resource excessively, rather than unemployment resource. The Gini coefficient of the USA is higher than 0.46. Although the USA is the richest country in the world, it is also the most unequal country among high-income countries. It is the reason of low MI of USA. 6.5. Middle-Income Countries In middle-income countries, it is necessary to look at Argentina and Lithuania. In the case of Argentina, the results are different depending on the middle class definition. When an MB middle class is the output, its MI is high (1.3), see Figure 12. Figure-12. MI (Middle-income countries) However, when the DB middle class is the output, the MI dropped significantly. This means Argentina was not significantly effective in its distribution of wealth. Furthermore, Argentina defaulted on its debt again in 2014, as well as faced a national crisis. On the other hand, even though, Lithuania’s results are the worst in both analyses, since Lithuania joined the euro in 2015, it is worth watching for changes in the future. 6.6. Low-Income Countries The performance of China is prominent among low-income countries. When the MB middle class is the output, the MI is 1.52, see Figure 13. In addition, it has the highest MI among 14 low-income countries. However, when the DB middle class is the output, the MI declines to 1.05, and its ranking dropped to 4th. This difference means economic fluctuations and distributions are moving apart. In other words, even though the economy is growing, distribution is not done well. Countries, such as Ukraine and Indonesia, have the same problem. In the case of low-income countries, the MI of more than half the country is greater than 1 in both analyses. Alternatively, in the case of high-income countries, the MIs of just three countries are greater than 1. However, the important thing is we should not conclude that low-income countries are better than high-income countries from the results. This is because all results are based on the relative value comparison of their group. Thus, even if one country’s MI is very high in its group, it does not mean one country is better than another country of other group. Asian Journal of Economics and Empirical Research, 2015, 2(2):62-73 71 Figure-13. MI (Low-income countries) 7. Conclusion We redefined the definition of ideal inequality and unemployment management, and evaluated the management ability of 49 countries during, before, and after of the financial crisis. This paper has shown that reducing inequality and unemployment is not always same as improving the middle class. Therefore, social structure should be taken into account. Second, during the financial crisis period, in the case of the high-income countries that account for a large share of the world economy, although most countries had the option to manage inequality and unemployment more efficiently, they failed (except for Austria, Finland and Germany). Third, in the case of low income-countries, the reliability of the results is low. The main reason is they are very sensitive to the definition of middle class. Additional problems should be resolved. First, it is necessary to properly define inequality and unemployment management because these concepts are too general. Second, we need to review their relationship with economic fluctuations. Although we saw a productivity decrease during the financial crisis, the specific functional relationship between them should be reviewed. Third, examining the ambiguous definition of the middle class is required. Finally, an analysis method, which can compare different income-level countries, should be developed to make more efficient inequality and unemployment management possible. This will help to achieve sustainable economic development and social stability. References Abraham, C., W. William Cooper and R. Edwardo, 1978. Measuring the efficiency of decision making units. European Journal of Operational Research, 2(6): 429–444. Craig Elwell, K., 2014. The distribution of household income and the middle class. Informing the Legislative Debate Sin 1914, Congressional Research Service. pp: 1-7. Federico, C., 2014. Trends in income inequality and its impact on economic growth. OECD Social, Employment and Migration Working Papers, No. 163. OECD Publishing. Francesca, C. and P. Gwenn, 2011. Being middle-class in Latin America. Working Paper, No. 305. OECD Publishing. Gaurav, D., 1998. Computational tools for poverty measurement and analysis. FCND Discussion Paper, No. 50. Washington, D.C.: International Food Policy Research Institute. Jose, V. and C. Barry Arnold, 1984. The general quadratic Lorenz curve. Mexico: Colegio De Postgraduados. Kristin Forbes, J., 2000. A reassessment of the relationship between inequality and growth. American Economic Review, 90(4): 869–887. Landes and David, 1998. The wealth and poverty of nations. The wealth and poverty of nations: Why some are so rich and some so poor. New York: W.W. Norton. Michael James, F., 1957. The measurement of productive efficiency. Journal of the Royal Statistical Society. Series A General, 120(3): 253– 290. Mthuli, N., L. Charles Leyeka and K.-M. Steve, 2011. The middle of the pyramid: Dynamics of the middle class in Africa. Tunis: African Development Bank. Rolf, F., G. Shawna, L. Bjorn and R. Pontus, 1992. Productivity changes in Swedish pharamacies 1980–1989: A non-parametric Malmquist approach. The Journal of Productivity Analysis, 3(1): 85–101. Rolf, F., G. Shawna and C. Knox Lovell, 1985. The measurement of efficiency of production. Netherlands: Springer. Rolf, F., G. Shawna and C. Knox Lovell, 1994. Production frontiers. New York: Cambridge University Press. Rolf, F., G. Shawna, N. Mary and Z. Zhongyang, 1994. Productivity growth, technical progress, and efficiency change in industrialized countries. American Economic Review, 84(1): 66–83. Rolf, F., G. Shawna, Y. Suthathip, K. Sung Li and W. Zhaoping, 1990. Productivity growth in illinois electric utilities. Resources and Energy, 12(4): 383–398. William Cooper, W., M. Lawrence Seiford and T. Kaoru, 2007. Data envelopment analysis: A comprehensive text with models, applications, references and DEA-solver software. US: Springer. William, E., 2001. The middle class consensus and economic development. Journal of Economic Growth, 6(4): 317–335. Asian Journal of Economics and Empirical Research, 2015, 2(2):62-73 72 Table-4. Statistics of input variables High Income country 2005 2006 2007 2008 2009 2010 2011 2012 Unemployment rate Average 6.77777777 78 6.31055 5556 5.8038 88889 5.91944 4444 7.9833 33333 8.533333 333 8.6088888 89 9.3188888 89 Max 11.28 10.28 8.66 11.33 18.03 20.08 21.64 25.05 Min 4.38 3.91 3.56 3.08 3.73 4.4 4.04 4.19 SD 2.22192432 6 2.08127 0744 1.8063 74786 2.05855 0239 3.2872 10511 3.894533 122 4.7915254 71 6.1793849 77 Gini coefficient Average 0.34577777 8 0.35027 7778 0.3526 11111 0.34983 3333 0.3520 55556 0.352555 556 0.3456111 11 0.3555555 56 Max 0.469 0.47 0.463 0.466 0.468 0.47 0.477 0.478 Min 0.267 0.273 0.28 0.268 0.259 0.266 0.269 0.271 SD 0.04643091 7 0.44471 809 0.4160 195 0.44333 296 0.4484 8073 0.044579 37 0.4511311 1 0.0450231 6 Middle Income country 2005 2006 2007 2008 2009 2010 2011 2012 Unemployment rate Average 8.10647058 8 7.19176 4706 6.4247 05882 6.40117 6471 9.1223 52941 9.936470 588 9.3141176 47 9.2858823 53 Max 17.93 13.97 13.1 12.65 18.23 19.83 16.2 15.85 Min 3.53 3.3 3.23 3.18 3.62 3.22 3.09 3.04 SD & 3.93105733 4 3.28784 1684 2.9800 12831 2.71867 5878 4.1390 05813 4.903756 138 4.1322528 04 4.1744326 84 Gini coefficient Average 0.37588235 3 0.37370 5882 0.3703 52941 0.37252 9412 0.3721 76471 0.372235 294 0.3733529 41 0.3744705 88 Max 0.506 0.5 0.492 0.485 0.48 0.475 0.472 0.471 Min 0.26 0.264 0.261 0.254 0.251 0.252 0.256 0.257 SD 0.06666228 5 0.06387 0733 0.0622 13284 0.06216 1199 0.0610 76013 0.059788 721 0.0591459 44 0.0589068 31 Low Income country 2005 2006 2007 2008 2009 2010 2011 2012 Unemployment rate Average 9.47642857 1 9.05214 2857 8.3728 57143 7.75428 5714 8.3471 42857 8.239285 714 8.1078571 43 7.8657142 86 Max 15.26 12.27 13.79 11.33 12.03 11.79 12 12.68 Min 4.19 4.1 4.02 4.19 4.29 4.1 3.83 3.74 SD 2.68704472 2 2.20906 2379 2.4367 70967 2.08369 3888 1.7835 11785 1.869827 077 2.3421757 75 2.7138610 43 Gini coefficient Average 0.42364285 7 0.42378 5714 0.4292 85714 0.428 0.429 0.429071 429 0.4284285 71 0.4291428 57 Max 0.595 0.594 0.593 0.593 0.592 0.592 2.592 0.591 Min 0.31 0.33 0.343 0.343 0.342 0.333 0.326 0.328 SD 0.08927968 8 0.08891 4115 0.0832 82123 0.08474 2143 0.0830 54199 0.084137 675 0.0830817 15 0.0820992 43 Table-5. Statistics of output variables High Income country 2005 2006 2007 2008 2009 2010 2011 2012 DB middle class Average 52.371666 67 52.087222 22 51.854444 44 52.17833 333 52.15666 667 52.05333 333 51.907222 22 51.867222 22 Max 54.94 54.76 54.58 54.95 55.51 55.37 55.23 55.15 Min 46.41 46.34 46.68 46.73 46.38 46.39 45.79 45.72 SD 2.1383562 96 2.1440226 23 2.2581122 77 2.081691 196 2.132397 16 2.145479 516 2.2444462 32 2.2509496 18 MB middle class Average 67.922222 22 67.705555 56 68.027777 78 67.63333 333 67.36111 111 65.62777 778 65.711111 11 64.238888 89 Max 75.5 74.9 75.5 78 80.4 77.3 78 76.7 Min 52.5 52.5 54.5 53.4 53.7 52.3 52.9 48.3 SD 5.5697244 16 5.3611315 99 5.0855973 74 5.264755 176 5.555086 582 5.668779 883 6.4435000 43 7.4956566 94 Middle Income country 2005 2006 2007 2008 2009 2010 2011 2012 DB middle class Average 50.541764 71 50.825882 35 50.806470 59 50.73 50.67764 706 50.78117 647 50.768235 29 50.708823 53 Max 55.42 55.21 55.22 54.84 54.69 54.82 54.94 54.8 Min 42.04 42.82 43.16 43.47 43.83 44.01 44.19 44.26 SD & 3.5666041 89 3.2785554 34 3.2378085 7 3.217067 764 3.132677 627 3.147447 145 3.1275554 26 3.1027384 73 MB middle class Average 58.552941 18 58.552941 18 59.452941 18 62.51176 471 59.69411 765 61.04705 882 61.682352 94 60.635294 12 Max 77 77.1 77.7 85.4 84.8 85.2 84.2 78.8 Min 34.9 36.7 36.7 41.3 41.2 46.7 45.1 45.3 Continue Asian Journal of Economics and Empirical Research, 2015, 2(2):62-73 73 SD 12.481141 66 10.802726 37 11.871453 03 11.79776 898 11.21413 564 10.81827 838 10.527656 15 9.9303160 31 Low Income country 2005 2006 2007 2008 2009 2010 2011 2012 DB middle class Average 45.664285 71 45.335714 29 44.942857 14 45.4 45.27857 143 45.17857 143 45.107142 86 45.092857 14 Max 52.1 53 51.9 51.9 51.9 51.8 51.8 51.7 Min SD 6.0326811 42 6.2976927 57 6.1967378 87 5.973273 809 5.799360 518 2.774769 868 5.5610547 26 5.5399894 87 MB middle class Average 58.085714 29 48.821428 57 56.957142 86 54.22857 143 52.72142 857 58.02857 143 55.042857 14 56.864285 71 Max 89.5 81.2 82.6 79.8 82 85.5 83.8 82.7 Min 34.6 23 31.3 34.2 27.3 41.8 39.8 37.9 SD 16.237509 15 15.805217 18 13.497162 91 12.78739 936 14.98666 699 12.37804 686 11.958049 38 10.819114 65 Table-6. Standard deviation of CU and FS High Income countries DB middle class output MB middle class output CU FS CU FS 2005→2008 0.095436267 0.037089688 0.108190434 0.035407862 2006→2009 0.12069557 0.020808809 0.129922961 0.08518847 2007→2010 0.116588756 0.026162906 0.133956861 0.033139944 2008→2011 0.102556379 0.015777084 0.140820412 0.010348775 2009 →2012 0.100155306 0.008259692 0.15963529 0.010348775 Average 0.107086456 0.021619636 0.134505192 0.024676021 Middle Income countries DB middle class output MB middle class output CU FS CU FS 2005→2008 0.105347548 0.063119159 0.181023277 0.075170224 2006→2009 0.141520478 0.04469597 0.171126902 0.06139061 2007→2010 0.184350774 0.25327634 0.255543253 0.21240582 2008→2011 0.128440321 0.043864624 0.223649275 0.032986299 2009 →2012 0.117962361 0.051703667 0.180505365 0.067636202 Average 0.135524296 0.04574221 0.202369614 0.051684783 Low Income countries DB middle class output MB middle class output CU FS CU FS 2005→2008 0.119996807 0.024365075 0.197829426 0.057696919 2006→2009 0.09426296 0.022783937 0.219483715 0.025571054 2007→2010 0.126240602 0.031737642 0.1817485 0.036689347 2008→2011 0.150401295 0.029173167 0.319145649 0.067701349 2009 →2012 0.1678201 0.018151432 0.365622042 0.017532905 Average 0.13174436 0.025242251 0.256765866 0.041038315 Views and opinions expressed in this article are the views and opinions of the authors, Asian Journal of Economics and Empirical Research shall not be responsible or answerable for any loss, damage or liability etc. caused in relation to/arising out of the use of the content.