Syndicate content

Top Incomes and the Measurement of Inequality in Egypt

Paolo Verme's picture

The January 2011 revolution in Egypt that overthrew 30 years of autocratic rule was in part due to a sense of deteriorating economic situation, injustice and growing inequality in the society. This was voiced by protesters during the revolution but also by intellectuals and the press after the fall of the old regime in an effort to explain the revolution. The World Values Surveys administered in Egypt in 2000 and 2008 confirm that the subjective aversion to inequality has intensified between the two years and for all social groups.  

It is thus surprising that economic inequality in Egypt as measured by household surveys is low and has been declining during the past decade. In 2009, the most recent year when a large household income and expenditure survey was administered, the Gini coefficients for inequality in incomes and expenditures were 32.9 and 30.5 respectively , far lower than in surrounding countries, southern Europe or the United States. This finding led observers to dismiss these figures as “unreliable” and incapable of capturing the true economic inequality in Egypt.

The Gini coefficient is indeed known to be susceptible to various statistical problems, including households under reporting or not reporting incomes or the sensitivity of the Gini to extreme values of incomes. If rich households tend to respond to surveys less than poor households, the Gini coefficient may be underestimated. On the other hand, if incomes of the rich are reported inaccurately these inaccuracies may bias the Gini coefficient significantly. Cowell and Victoria-Feser (1996)  have shown that a few observations in a survey can account for several percentage points of inequality while Korinek, Mistiaen and Ravallion (2006) have shown how households non-responses can seriously affect the estimation of inequality.

A recent paper by Hlasny and Verme (2013)  put the Egyptian data to the test by using recently developed statistical techniques designed to correct the Gini for various types of estimation problems. The findings of this paper show that the Egyptian data are of good quality and confirm inequality to be low by international standards. Survey nonresponse appears to be systematic in Egypt, since geographic areas with higher distributions of income tend to have significantly higher nonresponse rates. Rich households are indeed less likely to participate and are represented less satisfactorily in the survey than poorer households. In particular, the top income households have a much lower likelihood of participation than the rest of the population as initially expected. However, correcting the data for the underrepresentation of top incomes leads to a higher Gini of only 1.3 percentage points for both income and expenditure resulting in figures for inequality that are still very low in regional and global terms.

The paper also tested the validity of extreme observations in Egypt by substituting these observations with a theoretical distribution of incomes which is known to represent top incomes well across countries. Over a hundred years ago, the Italian economist Vilfredo Pareto noticed that top incomes typically follow a particular distribution. As income increases, the number of observations under this distribution declines following a certain regular pattern and this seems true across countries. The literature on top incomes has used this principle to study tax data while Cowell and Victoria-Feser (2007) used the Pareto distribution to correct inequality measures for such bias. The paper followed a similar approach and found that the distribution of top incomes in the Egyptian surveys follows very closely the Pareto distribution and is not atypical when compared to 418 household surveys worldwide. Replacing the actual top incomes with their estimates from the Pareto distribution does not affect our estimate of inequality in Egypt. This is a further confirmation that the Egyptian data are not atypical and that inequality in Egypt is low.

These findings suggest that the perception of inequality is a phenomenon which is much more complex than the measurement of economic inequality alone and that perceptions about high inequality and social injustice can co-exist with low levels of economic inequality as measured by household surveys. It is therefore important for researchers to better understand how perceptions of inequality are shaped and how economic inequality relates to feelings of social injustice. While the economics profession has made enormous progress in the measurement of objective economic inequality, it is still very much behind in understanding perceptions of inequality and how these perceptions relate to the perceptions of social injustice. This is an important area of research if economics wishes to better explain the socio-political evolutions of countries.


Submitted by William Butterfield on

Having lived in Egypt for two years, my impression is that yes indeed, Egypt is certainly "more equal" than many other developing countries and indeed the US. But Egypt's inequality takes a different form than many other developing country models. What you have in Egypt is a fairly large group of people who are "connected" to the incumbent elite ruling class, and a very large group of people who are not. The "connected" group does about 2-5 times better in terms of income than the "unconnected" group. Given that the connected group is much larger than in most other countries (because the military and protected enterprises are large and diverse, and spread benefits fairly widely) we see a relatively lower Gini than other countries. Compare this to, for example, Latin America or, say, the Philippines, where you tend to have a much smaller elite class that does about 50-100 times better on average than the unconnected classes. When you have a smaller group commanding much larger shares of rent per capita, that is what gives you the big Gini coefficient.

So despite the relatively small Gini, Egypt still has a glaring "insider vs. outsider" problem that is driving the perceptions that you refer to.

Add new comment