We took a long term look — years 1-2034 — at global income equality using the C-GIDD dataset. In line with other research, we found that inequality was more or less flat before the industrial revolution, rose sharply from around 1800 till 1980 and then started to decline, and is projected to decline further.
We measure income inequality with the often used Gini coefficient. If Gini=0 there is perfect equality, if Gini=100 there is perfect inequality. Most countries have a Gini between 30 and 50.
The graph below shows Gini from year 1 till 2034. The sharp rise in Gini from 1800 is mainly explained by increased income differences between countries, while the within-country effect is moderate. Even if all countries had perfect equality (Gini=0), the global Gini will be high since countries have vastly different income levels. We calculated this scenario for 2015 and global Gini is 48 with perfect equality within each country (Gini = 0), to compare with actual global Gini of 62.
- Gini coefficients 1990-2019 are based on C-GIDD’s more than 2,500 geographic units (i.e., subnational level). Therefore, accuracy is high.
- 1970-1990 and 2019-2034 coefficients are calculated using C-GIDD’s long-term national dataset (212 countries). Accuracy is slightly lower than above
- The period 1-1970 uses Angus Maddison’s long term GDP and population data, further broken down by us to the country level. GDP is used as a proxy for income, scaled to the income level in 1970. Gini coefficients are based on researchers like Popov and Piketty. The accuracy is moderate 1820-1970.
- In all years, the Gini coefficient is calculated by first estimating income as a function of cumulative population (a non-trivial mathematical problem). The coefficient is then calculated by summing the number of people across countries within a small income bracket, creating a global Lorenz curve, and converting it into the Gini coefficient.