A World of Ginis
This paper addresses the significant inconsistencies in global Gini coefficient estimates by constructing a unified dataset of over 122,000 observations to identify systematic measurement differences—particularly between income and consumption metrics—and provides correction factors to harmonize these estimates for more effective policy analysis.
Original paper licensed under CC BY 4.0 (https://creativecommons.org/licenses/by/4.0/). This is an AI-generated explanation of the paper below. It is not written or endorsed by the authors. For technical accuracy, refer to the original paper. Read full disclaimer
In the world of economics, there is a single number that leaders, journalists, and researchers rely on to understand how wealth is shared within a society. It is a measure of inequality, a way to see if a country's resources are concentrated in the hands of a few or spread more evenly among the many. For decades, this number has been the standard tool for deciding whether a government needs to change its tax laws, whether a region is becoming more stable, or if a policy is actually helping the poor. The idea is simple: if the number goes up, inequality is rising; if it goes down, the gap between rich and poor is shrinking. But for this tool to work, everyone must be measuring the same thing in the same way. If one person measures the height of a building using a tape measure and another uses a laser, their results will differ, and no one will know who is right. This is the problem facing the global community of inequality researchers. They have been collecting this crucial number from hundreds of countries for over a century, yet the numbers they find often disagree wildly, even for the same country and the same year.
A team of researchers from the University of Verona and the University of Bologna set out to solve this confusion. They gathered every available estimate of this inequality number from thirteen different major databases, creating a massive collection of over 122,000 observations spanning 222 countries and territories from 1867 to 2024. Their goal was not just to list these numbers, but to understand why they were so different. They wanted to know if the disagreement was due to bad data, different ways of asking questions, or something deeper about how we define wealth itself. By sorting through this mountain of information, they discovered that the differences were not random errors, but systematic patterns caused by specific choices researchers make when they calculate the number.
The most significant finding was that the definition of "wealth" changes the result more than anything else. In many parts of the world, researchers ask people how much money they earn in a year, while in others, they ask how much people spend on food, housing, and goods. The researchers found that when you measure based on income, the inequality number is consistently higher than when you measure based on spending. On average, the income-based number is nearly five points higher than the spending-based number globally. In some regions, like North America, the gap is even wider, reaching over ten points. This means that a country could appear to have high inequality simply because the researchers chose to look at income, or low inequality because they looked at spending, even if the actual distribution of people's lives had not changed.
The study also revealed that this gap is not static; it has been growing over time. Since the year 2000, the difference between income-based and spending-based measurements has widened significantly. In the past, the gap was smaller, but in recent decades, it has expanded, suggesting that the way we measure wealth is becoming less consistent with how people actually live. Furthermore, the researchers found that other technical choices, such as whether to count taxes paid or whether to adjust for family size, add further layers of difference. For example, counting income before taxes are taken out always produces a higher inequality number than counting income after taxes and government benefits are applied. In countries with strong tax systems, this difference can be substantial, making a nation look much more unequal than it actually is for its citizens after the government has done its work.
One might assume that as more databases were created, the disagreement between them would have grown worse, with different groups pulling the numbers in opposite directions. However, the researchers found a surprising result: the long-standing databases have actually remained quite consistent with each other over time. The increase in disagreement is not because the old sources are drifting apart, but because there are simply more sources now. The proliferation of new databases, each using slightly different methods, has mechanically widened the range of possible answers for any given country. It is not that the old measurements are becoming less reliable, but that the landscape of measurement has become more crowded with different approaches.
To help researchers and policymakers navigate this complexity, the authors developed a set of practical tools. They created adjustment factors that can be used to translate a number calculated in one way into the equivalent of another. For instance, if a researcher has a spending-based number for a country in South Asia and needs to compare it to an income-based number from a different region, they can apply a specific correction to make the two comparable. These adjustments vary by region and by the level of a country's development, acknowledging that the gap between income and spending measurements is larger in poorer nations where cash income is harder to track. The paper concludes that while perfect agreement may never be possible, the field can move forward by being transparent about these differences. Researchers must now state clearly which definition of wealth they are using and apply these corrections when comparing data across borders. The goal is to ensure that when we look at the state of the world, we are comparing apples to apples, not mistaking a different fruit for the same thing.
Drowning in papers in your field?
Get daily digests of the most novel papers matching your research keywords — with technical summaries, in your language.