Can inequality really be measured?
On one level the notion seems absurd, like trying to quantify fairness, honesty or corruption.
Dictionaries define inequality as a "lack of equality" in opportunity, treatment, or status. Though these concepts are themselves abstract, people tend to use something very concrete--personal income--as evidence that some individuals are more equal than others, to quote the popular adage. In effect, most societies view the distribution of income as a de facto indicator of inequality.
Economists have turned the human preoccupation with pay into a more scientific measure of inequality. They are helped by the fact that income, unlike opportunity or status, can actually be measured with some precision. For the IDB's new report on economic and social development, Facing Up to Inequality in Latin America, Bank researchers used answers to questions about income in household surveys conducted between 1994 and 1996 in 14 countries representing more than 80 percent of the region's population. These answers can be plotted as an ascending curve on a graph that shows the percentage of all income earned by segments of a population (see graph "Utopia vs. Reality" at left).
In a perfectly equal society, this line would be a straight diagonal (because, for example, any given 20 percent of the population would earn exactly 20 percent of all income). In Latin America and the Caribbean, the curve sags downward markedly. Look at points A and B on the graph, and you'll see that the poorest 20 percent earn only 4 percent of all income while the richest 10 percent earn 40 percent of national income.
For most people, however, a curve is much less descriptive than a score. To make themselves understood, economists have popularized a measure known as the "Gini coefficient." Named after Corrado Gini (1884 1965), an Italian statistician and demographer who pioneered studies of the measurable characteristics of populations, the coefficient represents the gap between the perfect distribution diagonal and a country's actual distribution curve.
A Gini coefficient of 0 indicates perfect income equality, while a 1 would imply that all wealth is concentrated in a single person. For the 95 out of 100 countries for which comparable data are available, Gini coefficients range between 0.26 and 0.60. In Latin America and the Caribbean, they range from a high of 0.59 in Brazil to a low of 0.43 in Uruguay (see graph "Two indicators, same result" at left ).
Skeptics might dismiss the Gini coefficient as yet another statistical toy cooked up by researchers obsessed with rankings. Indeed, there are several other ways of parsing income numbers in order to assess income inequality. One of the most common consists of simply dividing the income of people in the top 10 or 20 percent of the curve by that of the bottom decile or quintile. But as the larger graph on this page demonstrates, Gini coefficients tend to correlate very closely with these so-called "income gap" ratios.
For all their appeal, however, the Gini and other income indicators only illuminate a narrow slice of reality. "The real goal is to measure standard of living," says Miguel Székely, an IDB economist who processed many of the statistics in the Bank's report, "and income is only one dimension of standard of living." Other dimensions, such as health, education and consumption, are supposed to be reflected by income, but sometimes they aren't. Countries with comparable Gini scores can have radically different health indicators, for example.
As national averages, income indicators also obscure the huge variations in inequality within individual countries. If the household surveys from which an index is derived include only urban areas (as is the case with Argentina, Bolivia and Uruguay in the graph "Two indicators, same result" at left), the distortion is even greater. Székely says the inequality score for these three countries would have been considerably worse if rural areas also had been included.