Abstract A large amount of research is done on the effect of income inequality on national health. Still a final conclusion on the effect between those two variables has not been made. Previous papers suggest a negative effect between the two variables, but some researchers are critical against this finding. In order to find new information on this topic a panel data regression was done, with the independent variable being the Gini coefficient, and with the dependent variable being life expectancy and later infant mortality. No relation was found between the Gini coefficient and life expectancy, and no relation was found between the Gini coefficient and infant mortality. Name: Svend Olaf Henriksen Student-number: 10580026 Programme: Economics & Business Track: Economics Supervisor: dhr. R.E.F. van Maurik Date: February 17 2017
Table of content 1. Introduction...p.3 2. Literature review...p.4 3. Data...p.6 4. Methodology and research method...p.7 5. Findings and analysis...p.11 6. Discussion...p17 7. Bibliography...p.19 8. Tables...p.20 9. Figures...p.23 This document is written by Student Svend Olaf Henriksen, who declares to take full responsibility for the contents of this document. I declare that the text and the work presented in this document is original and that no sources other than those mentioned in the text and its references have been used in creating it. The Faculty of Economics and Business is responsible solely for the supervision of completion of the work, not for the contents.
Introduction This thesis is about the relation between inequality and the health of the citizens in a country. Hence the research question of this thesis is as follows: What is the effect of income inequality on health in a country? With this question the goal is to find out if there is a positive or negative relation between those two variables, or maybe even no relation at all. Interest on this topic came from other research on inequality and how it is growing (OECD, 2011). Also there are various politicians who are trying to lower income inequality with their policies. As health is important for the individual, but also for having an economic strong country as a whole, it is important to know how we could possibly better the health of citizens in a country.
Some previous research has been done on this topic but never in the way it is done in this paper. The variables are different and more recent numbers are used to get relevant results from the data. Also more data has been used to ensure the numbers are significant and reliable. One reason to use other variables and different countries is to see if the results from the regression are different, compared to those of others. By using more variables and more data, this paper adds new knowledge in the field of this research and could find, or could help finding a definitive conclusion on what the effect is of income inequality on health.
In this paper the relation between life expectancy and the Gini coefficient was analysed using a panel data regression. Secondly the relation between infant mortality and the Gini coefficient was studied. Afterwards results were compared with previous done research.
This paper is divided in a few sections. After the introduction there is a literature review where previous research on this topic is discussed. Following there will be a data section, where all the data used in the research is going to be explained. Next there will be a methodology and research method section where the research and the way of doing it is going to be explained. After the research has been explained, the findings and analysis section will give and analyse the results of regression and explain what can be seen in this result. Then there is a short discussion section including a conclusion.
Literature review
The effect of income inequality on health has been widely investigated by different researchers and for different countries. Although the researches done are all on the same topic, variables and samples vary broadly. Most of the research is done on developed countries, but some of the researches also focus on less developed countries like China (Pei & Rodriquez, 2006) or India (Rajan, Kennedy, & King, 2013).
One big research using developed countries was written by Lynch et al. (1998). They wrote a paper on the correlation between income inequality and mortality in the United States and Canada. For their regressions they used different income inequality measures, and they used excess mortality as a health measure. They concluded that income inequality has a positive effect on mortality, so when income inequality rises, excess mortality will also rise. Because the effect of income inequality on health could be different in development countries than it is in developed countries, it is relevant to do research on both of these. An example of a paper written on a then less
developed country is the paper by Pei and Rodriguez (2006). In their research they use the Gini coefficients from Chinese provinces as the independent variable and as dependent variable they used the self-rated health taken from a survey including 9594 households. Their regression is done with data from 1991, 1993 and 1997. In the research was found that the chance of reporting fair or poor health was greater when there was more inequality in that province. Additional relevant research was done by De Vogli, Mistry, Gnesotto and Cornia (2004). They did two regressions with the first one focusing on Italian regions and the second one using 21 developed countries. For their regressions they also used the Gini coefficient, but as their dependent variable they used life expectancy at birth. In their research they found a significant effect of income inequality on life expectancy in the Italian regions. Also they found a strong negative relation between life expectancy and inequality in the 21 developed countries. Other research also found a negative effect of income inequality on life expectancy (Wilkinson, Income distribution and life expectancy, 1992). A very broad research was done by Wilkinson and Pickett (2009), in their paper they do different regressions using both the Gini coefficient and the 20/20 ratio as variables for inequality. For health however they use a much broader variable then seen before, namely the Z score. This Z score is a score based on health and social problems and is composed for every country and U.S state they use in their regressions. In their research they found a negative effect of inequality on the Z score in 50 U.S states and in a sample of 20 rich countries.
A lot of papers find that lower income inequality would lead to better health, Judge (1995) however criticizes that conclusion and states that "it seems extraordinary that a predominantly mono-causal explanation of international
variations in life expectancy should ever have been regarded as plausible". Judge however does not prove that there is no relation between these two, he just asserts that better regressions are needed to prove a relation between life expectancy and inequality. There are also other sceptical views on how researchers interpreted the regressions (Lynch, Smith, Kaplan, & House, 2000). In their paper they write, "We do not deny negative psychosocial consequences of income inequality, but we argue that interpretation of links between income inequality and health must begin with the structural causes of inequalities". In their paper they also criticize the finding that gross domestic product per person has no effect on health (Wilkinson, Income distribution and life expectancy, 1992) and say that this depends on what sample of countries is used.
As stated before the research question in this paper is: what is the effect of income inequality on health in a country? Research done for this paper is on many points pretty similar to previous done research or has the same idea behind it. What will distinct this research from others is that different variables will be used than in some of the other papers, although there also are other papers that use the same variables. Also the dataset will be larger than some of the other datasets and this paper will use data from over a longer time period. Because this paper uses a panel data regression it is possible to both use different countries in the sample but also use eight different years.
Data
The Gini coefficient data is the disposable income Gini and is taken from the website of the Organization for Economic Co-operation and Development, or the OECD. As only OECD countries are used all data was available from that website.
The Gini coefficient data was taken from the year 2006 until 2013 because these years were available for most of the countries and these numbers were reliable. The life expectancy numbers also came from the OECD website and were also taken from the years 2006 until 2013. The life expectancy is measured in years and is defined as how long, on average, a newborn can expect to live if the current death rates do not change. Infant mortality data was also found on the OECD website and is defined as the number of deaths of children under the age of 1 per 1000 live births. The numbers for infant mortality were also taken from the year 2006 until 2013. The first Control variable, national income per capita, was also found on the OECD website. The national income per capita is defined as a countries gross domestic product deviated by the amount of inhabitants in that country. It is measured in US dollars per capita in millions of dollars and at current prices. The second control variable, which had to be a measure for female schooling, was chosen to be adult education level for woman. This measures the percentage of woman in a country with a tertiary degree. The variable is measured in percent and is measured for woman between 25 and 64 years of age. The adult education level for woman was also taken from the year 2006 until 2013 and could also be found on the OECD website.
Methodology and research method
In order to find the effect of income inequality on health there was chosen to use panel data regressions. Both fixed effect and random effect panel regressions were performed. Four regressions were done in this paper; the first one uses life expectancy at birth as the health variable and the second one used infant
mortality as a health variable. The third and fourth regressions are almost the same as the first two regressions, but now with an extra control variable added. The regressions were done on samples of 34 and 31 OECD countries over 8 years, data from the year 2006 until and including the year 2013.
The variable for income inequality in the regression is the GINI coefficient. The GINI coefficient was used because it is widely available for many countries and it includes a lot of data compared to the 20:20 ratio for instance. As stated above the variables for health were both the life expectancy at birth and the infant mortality rate. Life expectancy at birth is described as the expected age a person would reach born in that year. The life expectancy was chosen as a health measure because it probably is closely related to the health of a person because if one would live less healthy, one is expected to die at a younger age. Infant mortality was taken because it reflects the circumstances a child is born in, in the year of measurement. If fewer children survive in the circumstances it is born in, one could also expect the average health in that country to be lower.
The control variables used in the regressions were gross national income per capita and adult education level for woman. The first control variable was used because there was expected that national income per capita would have a positive effect on life expectancy, as people with more money can afford a healthier lifestyle and better medical care. The adult education level for woman variable was used because research showed that infant mortality is lower when woman have higher education (Arntzen, Moum, Magnus, & Bakketeig, 1996). Also there is an expected effect between the adult education level of woman and
national income per capita, so leaving the adult education level of woman out could lead to omitted variable bias.
The research was performed using a panel data analysis with the two dimensions being countries and years. Both fixed and random effects were used to find the results.
The first regression was done using random effects. The reason to use random effects instead of fixed effects in the first regression is that the Hausman test between the two regressions gave an outcome for the Chi-square of 0.25, which means it is more efficient to use the random effects. Also for this regression robust standard errors were used because the errors were found to be heteroskedastic. The Regression equation used is referred to as equation (1).
𝑙𝑖𝑓𝑒𝑒𝑥𝑝𝐵!" = 𝛼 + 𝐺𝑖𝑛𝑖!"∗ 𝛽!+ 𝐺𝑑𝑝𝐶𝑎𝑝𝐷𝑜𝑙𝑙𝑎𝑟!" ∗ 𝛽!+ 𝑢!" + 𝜀!", (1)
In equation (1); 𝑙𝑖𝑓𝑒𝑒𝑥𝑝𝐵!" is the life expectancy at birth in country i at year t, 𝐺𝑖𝑛𝑖!" is the gini coefficient in country i at year t, 𝐺𝑑𝑝𝐶𝑎𝑝𝐷𝑜𝑙𝑙𝑎𝑟!" is national income per capita in country i at year t, 𝛼 is the constant, 𝑢!" is the between entity error and 𝜀!" is the within entity error.
The second regression looks a bit different than the first regression because in the second regression fixed effects were used. After doing the Hausman test for the second regression there was concluded a fixed effect regression should be used because the Chi-square was found to be 6.89. In the second regression life expectancy was changed with infant mortality. Also for the second regression the errors were found to be heteroskedastic so robust errors were used. The second regression equation is referred to as equation (2).
𝐼𝑛𝑓𝑎𝑛𝑡𝑚𝑜𝑟!" = 𝛼!+ 𝐺𝑖𝑛𝑖!"∗ 𝛽!+ 𝐺𝑑𝑝𝐶𝑎𝑝𝐷𝑜𝑙𝑙𝑎𝑟!" ∗ 𝛽!+ 𝑢!". (2)
In equation (2); 𝐼𝑛𝑓𝑎𝑛𝑡𝑚𝑜𝑟!" is the infant mortality in country i at year t, 𝐺𝑖𝑛𝑖!" is the gini coefficient in country i at year t, 𝐺𝑑𝑝𝐶𝑎𝑝𝐷𝑜𝑙𝑙𝑎𝑟!" is national income per capita in country i at year t, 𝛼! is the unknown intercept for each country, and 𝑢!" is the error term. Regression three and four were almost the same as regressions one and two, but now an extra control variable were added. Again the Hausman test was performed and the test for heteroskedasticity was done. These tests found that in both regressions fixed effects needed to be used, and the errors again were heteroskedastic, so the robust option needed to be used. Regressions three and four are referred to as equation (3) and equation (4). 𝑙𝑖𝑓𝑒𝑒𝑥𝑝𝐵!" = 𝛼! + 𝐺𝑖𝑛𝑖!"∗ 𝛽!+ 𝐺𝑑𝑝𝐶𝑎𝑝𝐷𝑜𝑙𝑙𝑎𝑟!" ∗ 𝛽! + 𝐴𝐸!"∗ 𝛽!+ 𝑢!" (3) 𝐼𝑛𝑓𝑎𝑛𝑡𝑚𝑜𝑟!" = 𝛼! + 𝐺𝑖𝑛𝑖!"∗ 𝛽!+ 𝐺𝑑𝑝𝐶𝑎𝑝𝐷𝑜𝑙𝑙𝑎𝑟!" ∗ 𝛽!+ 𝐴𝐸!" ∗ 𝛽!+ 𝑢!" (4) In equation (3) the dependent variable was 𝑙𝑖𝑓𝑒𝑒𝑥𝑝𝐵!" which is life expectancy at birth in country i at year t. In equation (4) the dependent variable was 𝐼𝑛𝑓𝑎𝑛𝑡𝑚𝑜𝑟!" which is infant mortality in country i at year t. The other variables were 𝐺𝑖𝑛𝑖!" which is the gini coefficient in country i at year t, 𝐺𝑑𝑝𝐶𝑎𝑝𝐷𝑜𝑙𝑙𝑎𝑟!" which is the national income per capita in country i at year t and 𝐴𝐸!" which is the adult education level of woman of country i in year t.
The main reason to use panel data was that it made it possible to use more and more reliable data. To run the regression and get reliable results developed countries had to be used, also the data on these countries had to be trustworthy. Because of these criteria only OECD countries were used because for these countries lots of data is available, all the data is available in one place, and this data is reliably measured. The reason for only using 34 of the 35 OECD countries in the first two regressions is that Canadian data for life expectancy and infant mortality was not available for all the years, thus was Canada omitted from the sample. For the last two regressions also Chile, Israel and New Zealand were taken out of the sample because the Adult education level data for those countries was not available or not available for all the years.
Findings and analysis
For the first regression the life expectancy was used, which ranged from the lowest life expectancy of 70.8 years in Latvia in year 2006 to the highest life expectancy, which was the Japanese one in 2013 namely 83.4 years. The change in life expectancy varied widely per country during the years, some countries saw a rise of more than four year in life expectancy, while others only saw a rise of 1 year. The change in the Gini coefficient was not only positive nor only negative. Some countries saw a rise in income inequality during the years, which means a higher Gini coefficient. Other countries saw a decrease in inequality or saw it rising first and falling again later. As shown in the table, the change in income inequality was really different for every country.
In the first regression no effect of the Gini coefficient on life expectancy was found. In the random effects regression with robust standard errors, 𝛽! was
found to be 7.147 but was not found to be significant on a 1, 5 or 10 percent level. Different variants of this regression were done, where for instance logarithmic values of de dependent or independent variable where used, or where the differences between the years were used. Whatever form was used for the regression formula, 𝛽! would always be positive but never significant. The 𝛽! in a random effect model gives the average effect of X over Y when X changes across time and between countries by one unit. In the case of this paper it would be the effect of the Gini coefficient on the life expectancy. The result from this regression suggests that there is no effect of a change in income inequality on the life expectancy in OECD countries. There was however found an effect of national income per capita on life expectancy. In the first regression, the regression with random effects, robust errors and non-logarithmic variables, the 𝛽! variable was found to be 0.0001639 and it was found to be significant on a 1% level. All the results from the first regression can be seen in table 1. Table 1.
The finding that the Gini coefficient has no effect on the life expectancy was different than results in other researches. Previous research suggested that the Gini coefficient would have a negative effect on the life expectancy (De Vogli, Mistry, Gnesotto, Cornia, 2005), which would mean that less income inequality would give a higher life expectancy. The negative effect of the Gini coefficient on
Independent
variable Coefficient Robust Std. Err. z-value
Gini 7.146995 7.261907 0.98
GdpCapDollar 0.0001639*** 0.0000211 7.76
the life expectancy was both found to hold for Italy and for 21 wealthy countries. The negative effect of income inequality on life expectancy was also found in a research by Wilkinson (1992), Wilkinson however did not use the Gini coefficient as variable for inequality. Instead of the Gini coefficient a proportion of income going to the least well off 70% off the population was used in Wilkinson’s research. The differences in outcomes can be due to different things; firstly it could be due to the difference in samples. The regression done in this paper used 34 different countries, and for those countries data from 8 different years were used. In the regressions done by De Vogli et al. (2005) the amount of countries used was 21 and in a regression done with Italian regions 20 different regions were used. Also in the research done in this paper a time variable was used, which was not the case in the two researched mentioned before. In the second regression almost all numbers were the same as in the first regression but the numbers for the dependent variable. The infant mortality numbers ranged from 24.5 in Turkey till the lowest number of 0.9 in Iceland. Most of the countries saw a decrease in infant mortality, where only Luxembourg and Iceland saw a minor elevation. In the regressions done with infant mortality as dependent variable there was found a positive coefficient for 𝛽!, but this number was highly insignificant. In the all non-logarithmic regression formula with fixed errors and robust standard errors, the 𝛽! was found to be 2.1199, with a p-value of 0.744. This very insignificant number means that there is not found an effect of the Gini coefficient on the infant mortality rate. Again different forms of this regression formula were used, with for instance the differences between years or logarithmic values. Also a regression with a lagging infant mortality variable was done. In non of the regressions a significant 𝛽! coefficient was
found. There was however found an effect between national income per capita and infant mortality. In the regression there was found a negative effect of national income per capita on infant mortality. The 𝛽! was found to be
-0.0000976; this number was significant on a 5% level. The results from the second regression can be seen in table 2.
Table 2.
The result found in the second regression, the Gini coefficient has no effect on the infant mortality, would mean that more or less income inequality would not affect the infant mortality in a country. This is a plausible result but it was not in line with results from a previous done research, which found a positive effect of the Gini coefficient on infant mortality (Wilkinson & Pickett, Income Inequality and Social Dysfunction., 2009). The result was also not the same as the result in a paper written by Babones (2008), who used a different variable for income inequality and used a sample of less developed countries.
The third and fourth regressions were the same as the first two regressions but now the adult education level for woman was added, which measures the percentage of woman in a country aged between 25 and 64 years having a tertiary degree. The adult education level ranged from 9 percent up to 51 percent. After testing for heteroskedasticity and doing the Hausman test for both formula(3) and formula(4), there was found that robust standard errors and fixed effects needed to be used.
Independent
variable Coefficient Robust Std. Err. t-value
Gini 2.110772 8.659623 0.24
GdpCapDollar -0.0000976** 0.0000415 -2.35
Both in the third and fourth regression there was not found any effect of the Gini coefficient on de dependent variable. In both regressions a positive 𝛽! was found, but in both regressions it was insignificant. There was however found a significant effect for the adult education level for woman. In regression 3 the 𝛽! was found to be 0.1942 and was significant at the 1 percent level. In regression 4 the 𝛽! was found to be -0.0598, which was significant at the 5 percent level. Beside an effect of the adult education level for woman there was again found an effect of the national income per capita on the dependent variable in both regressions, those effect were only significant however when using logarithmic values for the national income per capita variable. The outcomes of regression 3 and 4 with logarithmic values for national income per capita can be found in Table 3 and Table 4, the regressions without logarithmic values can be found in table 6 and 7 in the tables section. Table 3. Independent
variable Coefficient Robust Std. Err. t-value
Gini 8.716769 5.4346 1.60 LogGdpCapDollar 2.142562*** 0.6385365 3.36 AE 0.15938*** 0.0199344 8.00 Constant 49.54636*** 6.546126 7.57
Table 4. If we define both life expectancy and infant mortality as health, a change in income inequality would not have an effect on the health in a country, but a higher national income per capita would raise the health in a country following the outcomes of these regressions. A lot of previous research suggests that more income inequality is linked with worse health in a country. Surprisingly is then that the findings in this research are different from what the literature suggests. An explanation for finding different results in the first and third regressions than in the literature could be that life expectancy is not a good variable for health. What makes life expectancy somewhat less reliable as a variable is that the life expectancy at birth is based upon historical life expectancies, so there would have to be a lag in the regression to find a more reliable outcome. Another problem with using life expectancy is that it is not changed directly by the change in Gini coefficient. If the Gini coefficient changes now, the life expectancy tomorrow would not change.
What was found in the regressions using infant mortality could also very well be true and one could find that infant mortality is a better variable to estimate health in a country. Infant mortality is a number that does not have a lag because its not based on historical numbers. A cause then for finding
Independent
variable Coefficient Robust Std. Err. t-value
Gini 0.2437892 8.339569 0.03 LogGdpCapDollar -5.835797* 3.08021 -1.89 AE 0.0297701 0.0573605 0.52 Constant 64.10607** 29.41954 2.18
different numbers could be difference in samples. Because one could expect developed countries to have a better national health system then less developed countries, which could explain lower infant mortality and no effect of the Gini coefficient on infant mortality. Although Wilkinson's (2009) paper only used rich countries, the sample was still different.
Discussion
Other than in this paper a negative effect of income inequality on health has been found in different big countries like Japan (Kondo, Kawachi, Subramanian, Takeda, & Yamagata, 2008), China (Pei & Rodriquez, 2006) and The United States (Lynch & Kaplan, 1997 ). Different things have said to be the reason for the found negative effect. Wilkinson (1992) argued that income inequality affects health trough psychological factors; people have the feeling that they are lower in the social hierarchy because of lower income, and those negative feelings give people poorer health inside the body and translate into habits like smoking. Other research suggests that because places with more inequality seem to have more crime, more people without health insurance, less investment in education, and poorer educational outcomes, which again translates into more difficult life for the people living in those areas (Kaplan, Pamuk, Lynch, Cohen, & Balfour, 1996). Besides the effect of the income inequality coefficients, some of the regressions in this paper found a positive effect of national income per capita on life expectancy. This finding was different than what Wilkinson (1992) found, but it is in line with the finding of Lynch et al. (2000).
In the analysis was found that the Gini coefficient has no effect on the life expectancy, and that the Gini coefficient has no effect on infant mortality in 34
OECD countries. Furthermore some regressions showed that national income per capita has an effect on the life expectancy and also on the infant mortality. When comparing the results from the first regression with the second regression there can be seen that if both life expectancy and infant mortality would be used as a variable for health in a country, the results of the regressions conclude the same thing. Both regressions suggest that there is no effect between the income inequality and health. Both regressions could be right and both effects could be right, but the numbers found can also be due to wrong variables or less good variables.
Although this paper added new information on this topic by using a panel regression with a large sample and a time variable, the regression was still too narrow to get one clear final conclusion. In this paper two different variables for health were used, but it is not known how well these variables represent the total health in a country. Future research should include a wider health variable that includes mental health, infant mortality and all other sorts of health indicators. Not only should the health variable be wider and based on more indicators, but also it should again be used in combination with a time variable and a country variable for a large sample. Besides a different health variable, there should also be regressed upon different sorts of income inequality indicators instead of only using one indicator in a research.
Bibliography Arntzen, A., Moum, T., Magnus, P., & Bakketeig, L. S. (1996). The Association between Maternal Education and Postneonatal Mortality. Trends in Norway, 1968–1991. Int J Epidemiol , 25 (3), 578-584. Babones, S. (2008). Income inequality and population health: correlation and causality . Social Science & Medicine , 66, 1614-1626. De vogli, R., Mistry, R., Gnesotto, R., & Cornia, G. (2005 ). Has the relation between income inequality and life expectancy disappeared? Evidence from Italy and top industrialised countries . Journal of Epidemiology and Community Health , 59 (2), 158-162. Judge, K. (1995). Income distribution and lifeexpectancy: a critical appraisal . BMJ: British Medical Journal , 311, 1282-1285. Kaplan, G., Pamuk, E., Lynch, J., Cohen, R., & Balfour, J. (1996). Inequality in income and mortality in the United States: analysis of mortality and potential pathways. BJM: British Medical Journal , 312, 999-1003. Kondo, N., Kawachi, I., Subramanian, S., Takeda, Y., & Yamagata, Z. (2008). Do social comparisons explain the association between income inequality and health?: Relative deprivation and perceived health among male and female Japanese individuals. Social Science & Medicine , 67, 982-987. Lynch, J. W., Kaplan, G. A., Pamuk, E. R., Cohen, R. D., Heck, K. E., Balfour, J. L., et al. (1998). Income inequality and mortality in metropolitan areas of the United states. . Amercian Journal of Public Health , 88 (7), 1074-1080. Lynch, J., & Kaplan, G. (1997 ). Understanding how inequality in the distribution of income affects health. J Health Psychol , 2, 297-314 . Lynch, J., Smith, G., Kaplan, G., & House, J. (2000). Income inequality and mortality: importance to health of individual income, psychosocial environment, or material conditions. BJM: British Journal of Medicine , 320, 1200-1204. OECD, 2011. Divided We Stand: Why Inequality Keeps Rising. OECD Publishing. http://dx.doi.org/10.1787/9789264119536-en. Pei, X., & Rodriquez, E. (2006). Provincial income inequality and self-reported health status in China 1991-7. J Epidemiol Community Health , 60 , 1065- 1069 . Rajan, K., Kennedy, J., & King, L. (2013). Is wealthier always healthier in poor countries? The health implications of income, inequality, poverty, and literacy in India. Social Science & Medicine , 88, 98-107. Wilkinson, R. G. (1992). Income distribution and life expectancy. BMJ: British Medical Journal , 304 (6820), 165. Wilkinson, R. G., & Pickett, K. E. (2009). Income Inequality and Social Dysfunction. Annual Review of Sociology , 35, 493-511.
Tables Hausman test Prob>Chi2 Regression 1: 0.6166 Regression 2: 0.0086 Regression 3: 0.0001 Regression 4: 0.0008 Table 5.
Variable Obs Mean Std. Dev. Min Max
Gini 272 0.3126944 0.0541313 0.2291205 0.484 LifeexpB 272 79.40331 2.776514 70.6 83.4 Infantmor 272 4.477941 2.950394 0.9 24.5 GdpCapDoll ar 272 34766.39 14298.01 12976.34 95035.34 AE 248 30.34343 9.946388 9.013416 47.10034 Table 6. Summary statistics
Table 7. Note: The dependent variable for table 6 was LifeexpB Table 8. Note: The dependent variable for table 7 was Infantmor Independent
variable Coefficient Robust Std. Err. t-value
Gini 3.127464 9.207883 0.34 GdpCapDollar -0.0000509 0.0000423 -1.20 AE -0.0597604** 0.0279717 -2.14 Constant 7.052481*** 2.146337 3.29 Independent
variable Coefficient Robust Std. Err. t-value
Gini 7.671198 5.926099 1.29 GdpCapDollar 0.000016 0.0000175 0.91 AE 0.1942159*** 0.0215214 9.02 Constant 70.52369*** 1.952362 36.12
Figures Figure 1: The Gini coefficient change in time for 34 OECD countries. 0,2 0,21 0,22 0,23 0,24 0,25 0,26 0,27 0,28 0,29 0,3 0,31 0,32 0,33 0,34 0,35 0,36 0,37 0,38 0,39 0,4 0,41 0,42 0,43 0,44 0,45 0,46 0,47 0,48 0,49 0,5 2006 2007 2008 2009 2010 2011 2012 2013 AUS AUT BEL USA CHE CHL CZE DEU DNK ESP EST FIN FRA GBR GRC HUN IRL ISL ISR ITA JPN KOR LUX LVA MEX NLD NOR NZL POL PRT SVK SVN SWE TUR
Figure 2: The change in the average Gini coefficient of 34 OECD countries. 0,309 0,31 0,311 0,312 0,313 0,314 0,315 0,316 2006 2007 2008 2009 2010 2011 2012 2013
Average Gini CoefWicient
Average Gini CoefoicientFigure 3: The life expectancy change in time for 34 OECD countries. 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 2006 2007 2008 2009 2010 2011 2012 2013 AUS AUT BEL USA CHE CHL CZE DEU DNK ESP EST FIN FRA GBR GRC HUN IRL ISL ISR ITA JPN KOR LUX LVA MEX NLD NOR NZL POL PRT SVK SVN SWE TUR
Figure 4: The change in the average life expectancy of 34 OECD countries. 77,5 78 78,5 79 79,5 80 80,5 2006 2007 2008 2009 2010 2011 2012 2013
Average Life expectancy
Average Life expectancyFigure 5: The infant mortality change in time for 34 OECD countries. 0 5 10 15 20 25 2006 2007 2008 2009 2010 2011 2012 2013 AUS AUT BEL USA CHE CHL CZE DEU DNK ESP EST FIN FRA GBR GRC HUN IRL ISL ISR ITA JPN KOR LUX LVA MEX NLD NOR NZL POL PRT SVK SVN SWE TUR
Figure 6: The change in the average infant mortality of 34 OECD countries. 3 3,3 3,6 3,9 4,2 4,5 4,8 5,1 5,4 2006 2007 2008 2009 2010 2011 2012 2013
Average infant mortality
Average infant mortalityFigure 7: The national income per capita change in time for 34 OECD countries. 0 10000 20000 30000 40000 50000 60000 70000 80000 90000 100000 2006 2007 2008 2009 2010 2011 2012 2013 AUS AUT BEL USA CHE CHL CZE DEU DNK ESP EST FIN FRA GBR GRC HUN IRL ISL ISR ITA JPN KOR LUX LVA MEX NLD NOR NZL POL PRT SVK SVN SWE TUR
Figure 8: The change in the average national income per capita of 34 OECD countries. 30000 31000 32000 33000 34000 35000 36000 37000 38000 39000 2006 2007 2008 2009 2010 2011 2012 2013
Average GDP/Capita
Average GDP/CapitaFigure 9: The adult education level for woman change in time for 31 OECD countries. 8 10 12 14 16 18 20 22 24 26 28 30 32 34 36 38 40 42 44 46 48 2006 2007 2008 2009 2010 2011 2012 2013 AUS AUT BEL USA CHE CZE DEU DNK ESP EST FIN FRA GBR GRC HUN IRL ISL ITA JPN KOR LUX LVA MEX NLD NOR POL PRT SVK SVN SWE TUR
Figure 10: The change in the average adult education level for woman of 31 OECD countries. 26 27 28 29 30 31 32 33 34 35 2006 2007 2008 2009 2010 2011 2012 2013
