Gini coefficient

• 1. Introduction

  ›

  The Gini coefficient is a single number between 0 and 1 that attempts to capture something extraordinarily complex: how equally a society distributes its wealth and income. A score of 0 means every person has exactly the same. A score of 1 means one individual holds everything while everyone else has nothing. In practice, no real society sits at either extreme, yet the distance between nations can be startling. In the late 20th century, Slovakia sat near the bottom of the inequality scale for OECD nations; Mexico sat near the top. South Africa, by some estimates, carried the highest Gini in the world, above 0.63, while the global income Gini in 2005 was estimated somewhere between 0.61 and 0.68. Where did this number come from, and what can it actually tell us? Those are the questions this documentary sets out to answer.

• 2. Corrado Gini's Invention

  ›

  Italian statistician and sociologist Corrado Gini published his foundational paper Variabilità e mutabilità, meaning variability and mutability, in 1912. He was building on the earlier work of American economist Max Lorenz, who had devised a curve plotting how income was distributed across a population. Gini's insight was to measure the gap between Lorenz's curve and a hypothetical straight line representing perfect equality. That gap, expressed as a ratio, became his measure of inequality.

  Two years later, in 1914, Gini pushed the idea further in his work On the measurement of concentration and variability of characters. He applied what he called the simple mean difference to income and wealth data, arriving at what he termed the concentration ratio. That ratio is the direct ancestor of the coefficient used today.

  In 1915, mathematician Gaetano Pietra added a geometrical interpretation, linking Gini's ratio to the visible area of concentration on a chart versus the area of maximum possible concentration. That altered version became the standard form adopted across subsequent decades. According to OECD data, Canada was the first country to use the Gini coefficient officially at a national level, doing so in the 1970s. Canada's income inequality index ranged from 0.303 to 0.284 between 1976 and the end of the 1980s, a modest span that reflects relatively stable distributional conditions during that period.

• 3. Reading the Numbers

  ›

  When the Gini is plotted mathematically, the tool at its core is the Lorenz curve. The curve traces what share of total income is held by the lowest-earning portion of the population. A perfectly equal society would produce a straight diagonal line. A real society produces a curve that sags below that line. The Gini coefficient is the ratio of the area between the diagonal and the sagging curve to the entire triangular area under the diagonal.

  Two simplified examples help illustrate the range. If the wealthiest 20% of a population earns 80% of all income, as in the Pareto principle scenario, the income Gini is at minimum 60%. If 1% of the world's population owns 50% of all wealth, the wealth Gini is at minimum 49%. Neither calculation requires complex data; the logic falls out of the arithmetic directly.

  For OECD countries in the 2008-2009 period, the pre-tax Gini ranged from 0.34 to 0.53, with South Korea the most equal and Italy the most unequal on a pre-tax basis. After taxes and transfers, that range shifted to 0.25 to 0.48, with Denmark the lowest and Mexico the highest. For the United States, the pre-tax Gini was 0.49 and the after-tax Gini was 0.38 during the same period. The OECD average after taxes and transfers was 0.31, and European countries, particularly Nordic and continental welfare states, consistently posted lower figures than the rest of the OECD group.

• 4. A World Getting More Unequal

  ›

  Milanovic's estimates of the world income Gini coefficient stretch back to 1820, and the trajectory is sobering. In 1820 the global Gini stood at 0.43. By 1913 it had risen to 0.61, and by 2002 it reached 0.71, the highest point in the estimated series. That 180-year trend of rising global inequality was driven by industrialization concentrating wealth in particular nations while others stagnated.

  After 2002, something shifted. By 2005 the world Gini had eased back to 0.68, and a separate dataset tracking 1988 through 2013 shows a more pronounced fall, from 0.80 in 1988 down to 0.65 by 2013. Economists attribute this reversal to rapid economic growth in developing nations, especially China and India, whose large populations pulled hundreds of millions of people into higher income brackets. Countries like Brazil improved basic services including health care, education, and sanitation, while Chile and Mexico adopted more progressive tax policies. The pattern suggests that global inequality is not fixed; it responds to policy choices and to the growth trajectories of large, previously poor economies.

  At the regional level, according to UNICEF data, Latin America and the Caribbean posted the highest net income Gini of 48.3 on an unweighted average basis in 2008. Sub-Saharan Africa followed at 44.2, then Asia at 40.4, the Middle East and North Africa at 39.2, Eastern Europe and Central Asia at 35.4, and high-income countries at 30.9.

• 5. Beyond Income: Other Inequalities

  ›

  Three World Bank economists, Vinod Thomas, Yan Wang, and Xibo Fan, applied the Gini framework to education across 85 countries. Their study found that Mali had the highest education Gini index of 0.92 in 1990, meaning educational attainment was extraordinarily concentrated. The United States had the lowest education inequality Gini at 0.14. Between 1960 and 1990, China, India, and South Korea achieved the fastest drops in education inequality.

  Philosopher and economist Amartya Sen proposed that inequality measures should focus on expanding people's choices and capabilities rather than simply tracking income gaps. That idea inspired the Gini opportunity coefficient, which measures how equally a society distributes the chance to succeed based on personal effort and talent rather than on birth circumstances like gender, race, parental income, or place of birth. In 2003, Roemer reported that Italy and Spain showed the largest opportunity inequality Gini index among advanced economies.

  In 1978, Anthony Shorrocks introduced a related measure, now called the Shorrocks index, that uses Gini coefficients to estimate income mobility. The index compares short-term income inequality, such as annual household earnings, against long-term earnings measured over five or ten years. A 2010 study using United States social security data since 1937 concluded that income mobility in the country had a complicated history, largely because of the mass entry of women into the American labor force after World War II. Sastre and Ayala, studying data from 1993 to 1998 across six developed economies, found that France had the least income mobility and Italy the highest during that five-year span.

• 6. What the Gini Misses

  ›

  Bangladesh and the Netherlands both recorded an income Gini coefficient of 0.31 in 2010. Bangladesh had a per capita income of $1,693 that year; the Netherlands had a per capita income of $42,183. The number was identical. The lived reality was not. This is the central tension built into the Gini: it measures relative distribution, not absolute wealth. A developing country can see its Gini rise even while the number of people in absolute poverty falls, because the coefficient tracks proportional gaps, not actual living standards.

  Sweden illustrates a different version of the problem. The country posted a disposable income Gini of 0.31, suggesting a relatively equal society. Its wealth Gini, however, ranged from 0.79 to 0.86, suggesting an extremely unequal distribution of assets. The income measure and the wealth measure told almost opposite stories about the same country.

  Billionaire Thomas Kwok argued that Hong Kong's reported Gini of 0.434 in 2010 was inflated partly by structural demographic shifts. Rising numbers of small households, elderly households, and elderly people living alone meant that combined household income was being divided across more units. The measured inequality rose, but the underlying income distribution may not have changed in the same proportion. United States Census Bureau data in Table C of the source confirms a similar dynamic: between 1979 and 2010, inflation-adjusted incomes rose across all brackets, the share of households in higher brackets grew, and yet the pre-tax Gini climbed from 0.404 to 0.469.

  Informal economies add another layer of difficulty. Schneider and colleagues, in a 2010 study of 162 countries, estimated that roughly 31.2% of world GDP, or about $20 trillion, was informal. In some of the poorest sub-Saharan countries, informal employment accounted for as much as 90% of all employment. Income earned through subsistence farming or barter rarely appears in official statistics, meaning Gini calculations for those economies systematically undercount the resources available to lower-income groups.

• 7. Unexpected Applications

  ›

  In 2004, astronomer Jennifer Lotz applied the Gini coefficient to the distribution of light across galaxy pixels in telescope images, finding it a reliable method for separating ultra-luminous infrared galaxies from normal ones. The coefficient has since been used extensively in galaxy morphological classification.

  Ecologists use the Gini to measure biodiversity, plotting cumulative species proportions against cumulative individual counts. Health researchers have used it to measure inequality in health-related quality of life within populations. In engineering, it has been applied to evaluate the fairness of Internet routers distributing packet transmissions across different data flows. In machine learning, it has served as a metric for measuring similarity across images and text in vector spaces, including in the selection of training samples in settings where data is sparse.

  In credit risk management, a version of the Gini called the discriminatory power measure gauges how well a rating model separates borrowers who will default from those who will not. Kaminskiy and Krivtsov extended the concept further, into reliability theory, proposing a Gini-type coefficient that measures how systems age or rejuvenate over time. Their version ranges from -1 to 1, with negative values indicating decreasing failure rates and positive values indicating increasing ones. A value of zero corresponds to the exponential life distribution, the baseline case of constant failure rate. The Gini coefficient, conceived to describe the difference between rich and poor in early 20th-century Italy, had by the early 21st century become a general-purpose tool for measuring concentration and dispersion across disciplines as far apart as astrophysics and loan underwriting.