The reliability of the data

Since the 1950s, there has been a data revolution, resulting in a significant stock of information on incomes based on comparative tax data and household surveys.1 These data have enabled researchers to provide more precise statistics on global poverty and inequality. However, Atkinson and Brandolini (2009) are critical of the way in which researchers have dealt with the data. They argue that the Gini coefficient is generally indeterminate, and call instead for an investigation into the development of key variables that determine inequality country by country. This approach forces researchers to assess and demonstrate how the Gini value fluctuates depending on the parameters used, for example net or gross taxes or expenditures, and gross or disposable income.2

Another problem affecting the measurement of inequality is that of ‘undersurvey’, i.e., data that is not captured by surveys. Milanovic points out that rich people often withhold their actual income, meaning that national accounts present a more encompassing picture than household surveys in this respect.3 Scholars have therefore increasingly been making use of fiscal data (the reported pretax income of the rich) to estimate income shares at the top, but whether this has indeed improved the data sets remains questionable (in the US, for example, it did not make a significant difference).

Footnotes