Excerpt
List of contents
1. Introduction
2. Theoretical Background
3. Data and Methods
4. Empirical Results
a. Descriptive Analysis
b. Clustering of the European Union
5. Policy Implications
a. National policies
b. Supranational policies
6. Conclusion
7. References
a. Bibliography
b. List of abbreviations
c. Data sources
d. List of figures
e. List of tables
f. Appendix:
Abstract
Why are health and social problems in the EU related to income inequality within countries, rather than per capita income?
With regard to Wilkinson and Pickett’s studies in “The Spirit Level” (2010), I am demonstrating the relation of health and social problems with income inequality for EU countries and compare the results with the European Social Policy Models described by Boeri (2002) and Sapir (2005).
“Income inequality has soared to the highest levels since the Great Depression, and the recession has done little to reverse the trend, with the top 1 percent of earners taking 93 percent of the income gains in the first full year of the recovery” (New York Times, 2012)
1. Introduction
At least since the “Occupy Wall Street” and the “We are the 99 percent” movements started to dominate newspaper headlines, the problem of unequally allocated disposable income has gained more attention by policy makers around the globe. In reconstruction times following WWII, gains in income have been shared almost equally between income quintile groups until the late 1970s - when the Great Convergence ended - as pointed out in Figure 1.
illustration not visible in this excerpt
Figure 1: Change in Real Family Income by Quintile and Top 5%, 1947-2009 USA
Source: Analysis of U.S. Census Bureau data in Economic Policy Institute, The State of Working America 1994-95 (M.E. Sharpe: 1994) p. 37
Politicians and economists have therefore increased their interest in finding other indicators for economic performance rather than only casting an eye on GDP growth (Bonesmo, 2012). I will thus investigate why health and social problems are far more related to income inequality rather than GDP growth or per capita income. To understand the situation especially in the European Union (EU) I will subsequently explain the underlying circumstances.
Through the establishment of the EU as a single market and with deeper integration, similar problems like in the USA have arisen and the removal of social discrepancies and regional disparities has become a goal declared in the EU’s Europe 2020 strategy to confront its cohesion problem, which is drastically influencing welfare in most European countries (Artis and Nixson, 2007). In its strategy the EU targets 20 million Europeans to be lifted out of poverty at a total amount of 115 million in 2010 (Eurostat). Complementary, the EU promotes inclusive growth to ensure social and territorial cohesion in line with the goals of the Euro Plus Pact and the Structural and Cohesion funds. To achieve cohesion, the EU invests more than a third of its budget (OECD, 2007) - that means a total of € 347 billion during 2007 until 2013 - into its regional policy to support job creation, competitiveness, economic growth, improved quality of life and sustainable development (European Union, 2012). Furthermore it focuses on stimulating growth in areas where incomes are relatively low.
There are not only income inequalities within countries, but also between members and its regions due to different levels of development. Convergence Objective regions are funded by the ERDF and ESF if their GDP per capita is below 75% of the EU average, which has as a whole declined since the latest enlargements. It was originally also intended for regions having a high number of unemployment like former GDR, Spanish and UK regions facing serious unemployment problems of up to 18% (Artis & Nixson, 2007). Economists (De Grauwe, 2008; Kuznets, 1934; et al.) have risen their worries about the GDP per capita indicator to be insufficient to describe the socioeconomic situation in regions since welfare cannot be measured only by national income which is calculated only by items which can be priced excluding important welfare indicators such as poverty, human development et cetera (Raghav et al, 2012). The lacking capability of the GDP per capita indicator has motivated me to consider for other indices such as health and social problems, used by Wilkinson and Pickett (2010) in their studies on how income inequality is related to social gradients within OECD countries.
A further driver of income inequality is social immobility, which means that wealthy families are more likely to bring out wealthy offspring, because of resources promoting them into higher income positions. That next generation is also more probable to start a family with partners of a similar income group because of community backgrounds, which means that this process, taking place over generations, is encouraging income inequality (Cooper, Durlauf et al. 1994; Cooper 1998; Durlauf 1995). On the opposite the same process is happening for families with a poor background and no resources to promote their descendants into better earning careers. This very basic, almost unalterable process is a main driver of the income gap becoming wider due to lacking social mobility, as empirical analysis have shown (Cooper, 1998).
When looking at the EU it is obvious that income disparities do not only occur in poor countries but in highly developed ones as well. The “blue banana”, consisting of regions from Milan over the Ruhr up to Manchester, is the highest developed area in the EU. In Germany the comparison of former GDR and FRG displays an extreme example of internal disparities. Former Eastern Germany, lacking of decades of progress, consisted of some of the least developed parts in the EU after the reunification in 1991, when its GDP per capita was around 40% and unemployment being at a critical rate (10%).
Assuming that income inequality is a problem most of the EU countries have to face, Boeri (2002) mentioned that the reduction of poverty along with inequalities is depending on the efficiency and equity of a suiting Social Policy Model. By plotting the percentage change of reduction of the Gini coefficient due to taxes and transfers against social expenditures as a percentage of the countries’ GDP, Boeri (2002) has shown that countries with higher social expenditures tend to reduce their Gini coefficient more significantly. Especially smaller countries seem to have a more effective redistribution rather than larger nations (except the UK). In my own elaborations, found in the Empirical Results section, it is not obvious anymore that higher social expenditures will lead to lower income inequality. Instead I compared income inequality with a slightly modified index of health and social problems in contrast with Wilkinson and Pickett (2010), who have observed a high relation between health and social problems and income inequality for OECD countries and US states.
2. Theoretical Background
Bonesmo (2012) has shown that differences in income inequality across the EU may be initiated by differences in labour market outcomes, concentration of capital income and especially by the progressivity of tax and transfer systems. With the Nordics having low dispersion in labour earnings while cash transfers still are universal and taxes are not very progressive they also tend to have one of the lowest income inequalities of all EU countries.
Wilkinson and Pickett (2010) mentioned that income inequality within countries is still a problem that governments are yet to be made aware of. Developing countries may perform better regarding the removal of inequality due to economic growth in comparison to industrialised countries which have reached saturation limit in terms of reducing income inequality via economic growth. The comparison of countries with a different socioeconomic classification and varying stages of development is consequently difficult and could effectively not be expedient. Wilkinson and Pickett (2010) have therefore only discussed developed OECD countries and their issues affecting society and economy. In Figure 2, I present a graph to demonstrate why growth in per capita income is not enhancing life expectancy anymore at a certain stage. Countries with a certain amount of GNI per capita tend to be saturated in terms of enhancing life expectancy through growth of the GNI per capita in comparison to economies which are still growing more rapidly.
If we just look at the blue observations (higher income per capita countries) in Figure 2, I can say that an increase in GNI per capita barely has any effect on life expectancy anymore and a guess on the trend for the blue observations without the regression line could reasonably not be made. I chose this graph to consider health and social problems to be something worth thinking about since life expectancy is considered to be representing the quality of life or well-being and plays an important role in the UN Human Development Index (HDI). To come back to my research question that health and social problems are related to income inequality within countries rather than per capita income, I am referring to Wilkinson and Pickett (2010) who have shown that the problem lies within societies. When looking at Figure 3, a huge gap between income groups, concerning their life expectancy, becomes obvious.
illustration not visible in this excerpt
Figure 2: Life Expectancy vs. Gross National Income per capita
Source: Own elaborations, data source: World Bank (GNI/capita = PPP in international $)
Wilkinson and Pickett (2010) have shown that there are extraordinary social gradients in health running right across society – England and Wales in this case. Even at the top there are differences between the richest and the very rich. Wilkinson (2011) argues that within societies we are looking at relative income or social status and the size of the gaps between income groups and that is what makes income inequalities within countries a driver of social gradients.
illustration not visible in this excerpt
Figure 3: Life Expectancy vs. Income Groups
Source: Richard Wilkinson & Kate Pickett, (2010): “The Spirit Level”
Parallel to Wilkinson and Pickett’s (2010) investigations in “The Spirit Level” - of how socioeconomic indicators within a country may be affected by income inequalities or vice versa - I am investigating if there is a similar outcome for EU countries. Posterior, I will present the answer to my research question if the index of health and social problems for EU countries bears a closer relation to income inequality than per capita income. Additionally I will compare my results with the classification of European countries by Boeri (2002) and Sapir (2005) regarding their Social Policy Models. As proposed the Mediterraneans and Anglo-Saxons are more unequal than the Continentals and Nordics which implies that those countries suffer from higher income inequality. The Anglo Saxons for example are struggling with increasing wage dispersion and a high incidence of low-pay employment (Boeri, 2002), while the Mediterraneans concentrate their social spending on old age pensions whereas youth unemployment has reached new dimensions in a negative way with Greece and Spain having rates beyond 50 % and Portugal and Italy following with around 35 % and counting (Eurostat, 2012Q2). Additionally, their social welfare systems typically draw on employment protection and early retirement provisions. On the other hand the Nordic countries are still performing best with the highest level of social protection expenditures and universal welfare provision. They also apply extensive fiscal intervention in labour markets based on active policy instruments. The Continental countries are quite similar to the Nordics with a higher spending on pensions and insurance based non-employment benefits (Sapir, 2005). In the empirical section I will additionally have access to data of the latest enlargement countries which might affect the results compared with Boeri (2002) and Sapir (2005).
To classify the European Social Policy Models concerning their efficiency, Boeri (2002) has evaluated the countries regarding their degree of redistribution and reduction of inequality, with the Nordics and Anglo-Saxons having the most efficient ones. As earlier mentioned I came to different results, being discussed in the Empirical Results section.
3. Data and Methods
The data being used for the analysis has been obtained from reliable sources namely Eurostat, WHO, UNICEF, OECD and the US Census Bureau. Data for all 27 EU countries was accessible for calculating the index of health and social problems. No data was available for non-OECD countries concerning their pre-tax and -transfer Gini coefficient. For a full list of the variables and its sources please see the appendix.
illustration not visible in this excerpt
Figure 4: Gini Coefficients in Europe; Source: Eurostat
To start analysing, I first obtained the Gini coefficient, being the most prominent and widespread index of income inequality. Leading providers of statistics and databases like the OECD, Eurostat et cetera, are using the Gini coefficient to describe income inequality in countries worldwide. It is a simple tool to relatively define how unequal income is allocated among population.
The coefficient can be a figure between 0 and 1, where 0 describes perfect equality (every individual receives the same amount of income) and 1 indicates that all income is allocated to one person. However there are problems arising when interpreting the coefficient and its resulting Lorenz curve[1] (Wilkinson and Pickett, 2010).
Because of its “relative” meaning, two countries with completely different income structures can possess the same Gini coefficient. For example Belgium had a coefficient of 28 and Albania had quite the same one (26.7) in 2005, while Belgium had a per capita income of around $ 38,000 and Albania only $ 7,800 (CIA, 2012), hence Albania may have a more equal distribution of income but has a huge disadvantage in terms of welfare as also described in the Human Development Report (2010). Therefore the Gini coefficient does not take absolute numbers into account what might cause problems when comparing countries with differences in absolute income (Bellù, & Liberati, 2006). A second issue occurs when computing a Gini index. As seen in Figure 5, income distribution in country X differs from the one in country Y, still they have the same Gini coefficient. To find the differences between those countries, one has to look at the Lorenz Curves as seen in Figure 5, hence the Lorenz Curve describes not only how much inequality is found graphically but also in which part of the income distribution. For example: Country X has a more equal distribution in the lowest quintile, therefore the closer the Lorenz Curves gets to the 45° line the lower the Gini coefficient becomes. Yet it is very important to note that equal must not inevitable mean fair because if every individual gains the exact same income, quality of competition is most likely to lose value which is in complete contrast with free market economy (Chuen, 2010).
illustration not visible in this excerpt
Figure 5: Lorenz Curves of different income distributions with same Gini index
Source: Lorenzo Giovanni Bellù and Paolo Liberati (2006): "Inequality Analysis - The Gini Index".
Further approaches on measuring income inequality are income share ratio metrics, which are used to get a more detailed look at the income distribution in comparison to the Gini coefficient. It is used to compare two income groups, with the S80/S20 income quintile ratio share being the most prominent one. It compares the top 20 % earners with the bottom 20 %. A problem occurring again is the relative nature of the ratios describing differences only between two income groups. In Figure 6, I compared the growth of income for various income groups with the national median income. For example the bottom quintile compared with national median income (1:M) has seen no significant growth, whereby it even has to register losses especially in the Eurozone. Compared with the Gini coefficient of the Eurozone, the income growth of the fourth quintile and 95th percentile led to higher income inequality and hence an increase of the Gini coefficient. In contrast to that the allocation of income has become fairer in the “new” EU countries[2] of the latest enlargements because of a significant decrease of income especially for top earners. When looking at the data of the whole EU (EU27) we observe a salient point at 2008, which might have been triggered by the economic crisis. The Gini coefficient slightly decreased afterwards but could not prevent the top-earners to be the only ones profiting from income growth as can be seen in the real income growth ratios diagram in Figure 6, with increases of income only to be found in the fourth quintile and the 95th percentile.
[...]
[1] An approach on the calculation of the Gini coefficient can be found in the appendix.
[2] As the new, or the latest enlargement countries I am referring henceforth to the Czech Republic, Estonia, Hungary, Latvia, Lithuania, Poland, Slovakia, Slovenia, Malta, Cyprus, Romania and Bulgaria.