An International Comparison of the Incomes of the Vast Majority Anwar Shaikh and Amr Ragab April 12, 2007
Introduction
Income levels and income inequality are two major dimensions of national and international well-being. The last two decades have witnessed a growing concern about both issues, from policy makers, social scientists, and the media. But the two dimensions are generally treated separately, with GDP per capita (GDPpc) as the paramount measure of national income and the Gini coefficient (G) as the central measure of inequality. Sometimes these are implicitly combined, as in the case of poverty measures which count the number of people in each nation who live on less than one dollar a day (World Bank, 2000/2001, p. 3)
There are well-known problems with these traditional approaches. First of all, average income per capita also tells us nothing about income variations within a population. For instance, if four people with an average income of $50,000 are joined by one more with an income of $300,000, the per capita income of the group doubles1. Knowing that the Gini coefficient of the group is "high" alerts us to the fact that the average is unrepresentative, but does little to help us understand its real magnitude.
A simple alternative would be to measure the income per capita of the vast majority of this group, say the first 80 percent in the income ranking. Such a measure would combine the average level of income and its distribution into an intuitively useful statistic. Moreover, it would have obvious political resonance in any modern political system. There is evidence, for instance, that state-wide voting preferences in the US are correlated with changes in average local incomes (Altman, 2006) and it would be interesting to see if this relationship is stronger with vast majority incomes.
This paper is part of an ongoing project to analyze international inequality. International comparisons tend to focus on either GDP per capita or the incomes of the very poor (e.g. those living on less that $2 per day). The VMI adds a new dimension, because it combines information on income levels and their distribution into a single measure which is the average income of the vast majority of the population. We believe that this broadens the discussion of international inequality, and will ultimately shed new light on several important issues in the development literature such as the relationships between development and inequality, growth and inequality, trade liberalization and living standards, and political instability and inequality. In this paper we focus on the first issue by exploring the links between development and the incomes of the vast majority of the world's population.
In this paper we develop the preceding measure on an international scale. We begin by calculating the ratio of the disposable income per capita of the first 80 percent (the vast majority) of the population to the average income per capita, in any given nation. We call this the Vast Majority Income Ratio (VMIR), and show how it can be derived from a Lorenz curve. Multiplying this ratio by an appropriate average real per capita income measure (Net National Income per capita) then gives us the real income per capita of the vast majority of the population across nations, regions and/or time. It is shown that the VMIR varies considerably across countries. This implies that average per capita income measures such as GDP or Net National Income are not reliable proxies for the per capita incomes of international vast majorities. Indeed, we show that ranking countries by their vast majority incomes (VMIs) gives different results than ranking them by average per capita income. In next section of the paper we demonstrate that the VMIR, which is itself an equality measure, bears a constant ratio to (1- G) across countries and across time. This unexpected internal relation, which we call "The 1.1 Rule", also leads us to a new interpretation of the Gini Coefficient as the relative per capita income of the first seventy percent of the population in any given country. The theoretical foundations of these two empirical rules are addressed in a section on distribution theory and "econophysics". Policy implications are outlined next, as are connections to an earlier literature in which (1-G) was proposed as a discount factor for various measures of well-being. The paper ends with a summary and a discussion of potential areas of future research. The Data Appendix explains our sources and methods.
[...]