We published a new topic page on economic inequality

Economic Inequality Public attention to economic inequality has grown a lot over the last twenty years. Alongside, the available data has greatly improved too: estimates now cover many more countries and longer time periods, and new international databases have helped make the data more comparable and much more accessible. Where our view was once limited to individual countries, we can increasingly take a global perspective on inequality. At the same time, the wide range of sources and metrics now available can be difficult to navigate. There are many different approaches to measuring inequality, each with its own strengths and weaknesses, and it can be hard to know how they relate — or where to start. This page is designed to make the range of inequality data easier to understand. Here you’ll find all our data, charts, and writing on economic inequality, organized around three key international databases: the World Bank’s Poverty and Inequality Platform, the Luxembourg Income Study, and the World Inequality Database. Taken together, this data shows that inequality in many countries is very high and, in many cases, has been rising. Globally, the gaps between the poorest and richest are extremely large and are compounded by overlapping inequalities in health, education, and many other dimensions. But it also shows that inequality is not rising everywhere. Global income inequality — measured across the world’s population as a whole — has fallen in recent decades, driven in particular by rapid economic growth in parts of Asia. And within many individual countries, income inequality has fallen in recent decades or remained stable. This variation across countries tells us something important that is often missed: high and rising inequality is not inevitable. Through their institutions and policy choices, individual countries can affect the level of inequality among their citizens. Inequality is not purely determined by global forces beyond our control; it is something that we can change. Inequality around the world: global incomes data from the World Bank The World Bank’s Poverty and Inequality Platform (PIP) publishes data for almost all countries in the world, based on a large collection of national survey data. To achieve such wide coverage, it pools data from two kinds of surveys. For high-income countries, the data measures people’s incomes after taxes and benefits. For most low- and middle-income countries, it instead measures their consumption. These two concepts are closely related but not the same.1 The big strength of the World Bank dataset is the global perspective it provides. It gives us a picture of how much people have to live on all across the world, from the richest to the poorest, and everyone in between.2 When thinking about inequality, what people often have in mind is the level of inequality within a given country. What the World Bank’s global data shows is that this inequality within individual countries is just one part of the story. In addition, there are also large inequalities between countries: the citizens of some countries are much better off than the citizens of others. Below, we take a look at both kinds of inequality. Comparing inequality within countries Measures of inequality try to capture how evenly or unevenly economic resources are spread, or “distributed” across the population. One commonly used measure is the Gini coefficient. Although popular, a downside of the measure is that it is not very easy to interpret. Often, it’s more intuitive to compare incomes at particular points in the distribution. That’s the approach taken in the chart below. For six different countries, it shows the ratio between the 90th percentile and the 10th percentile. These are the income levels that mark the thresholds for the richest and poorest tenth of the population, respectively. The 90th percentile is the income that just puts someone inside the top 10%. The 10th percentile is the income that just puts someone inside the bottom 10%. The ratio of these two numbers — the “P90/P10 ratio” — gives us a measure of inequality that’s easy to interpret. It tells us how many times richer a person just inside the bottom 10% would have to be to just enter the top 10%. The bigger the ratio, the bigger the gap separating the richest and poorest tenth. In the United States, there’s a 7-fold gap between the incomes marking the richest tenth and the poorest tenth. In Germany, the gap is much lower: 4.5-fold. In Brazil, inequality is much higher — more than twice as large as in Germany, with a 10-fold gap. The data for Angola, Vietnam, and Madagascar refers to consumption rather than income.3 Looking within these three countries, we again see inequality varies hugely. The relative gap between the top and bottom tenth in Angola is twice as big as it is in either Madagascar or Vietnam. These large differences demonstrate a basic, but important insight. Inequality varies widely across countries, even among countries with similar levels of economic development, technology adoption, or exposure to global markets. These factors can all play a role, but they do not wholly determine inequality — far from it.4 The variation we see suggests that inequality is not simply dictated by forces beyond a country’s control. Through the institutions they build and the policies they adopt, countries can take different paths that lead to higher or lower levels of inequality. Inequality within and between countries The ratios in the chart above show us the different levels of inequality that exist within these countries. But they miss out on something important: the huge income differences that exist between the countries. Being in the richest or poorest tenth within a country means something very different depending on whether you live in a rich or a poor country. To add this important dimension, we plot the actual levels of income or consumption that mark the richest and poorest tenth in each country. The figures are expressed in international dollars, which makes an adjustment to account for the different cost of living in each country.5 The amounts are per person — each person (including children) is assigned an equal share of their household’s total income or consumption.6 We see what the different levels of inequality shown in the previous chart mean for household incomes. Incomes at the top of the distribution in the United States, for example, are considerably higher than in Germany. But at the bottom of the distribution, that’s not the case. A person just inside the bottom tenth in Germany is actually slightly better off than their counterpart in the United States. Although average incomes are higher in the United States, the different levels of inequality in the two countries mean that this plays out very differently across the distribution. What this way of plotting the data makes particularly clear, though, are the stark differences between countries. In poor countries like Angola and Madagascar, incomes are incredibly low across the entire distribution. They are extremely low among the poorest tenth, far below even the World Bank’s extreme poverty line of $3 a day. But they are also very low, even in the top tenth. When compared to the incomes in high-income countries, even the “rich” in low-income countries are very poor. Why a global perspective on incomes is so important The World Bank data allows us to understand how incomes around the world compare. This is important because it’s something most of us get badly wrong. In particular, people in rich countries tend to greatly overestimate incomes in the rest of the world, relative to their own.7 The World Bank data shows that in 2026 the global median income is $290 per month, or a little under $10 a day. Half the world’s population lives on less than this amount — far below the level of income of even the poorest tenth in high-income countries. This means the majority of the world lives on a very low income that few people in rich countries have experience of. It is no wonder then that people in rich countries tend to underestimate just how unequal the world really is. Although the World Bank data has some limitations, discussed below, it’s a crucial source that helps correct this misperception. It allows us to look beyond our own experience, and that of our fellow citizens, and to think about economic inequality in global terms. By bringing together the large income differences that exist both within and between countries, it shows that the problem of inequality is even bigger than most realize. Explore the World Bank’s data on global incomes The interactive chart below allows you to explore the same World Bank data over time, for countries or regions of your choice. The controls at the top allow you to explore the whole distribution, broken down by decile — each corresponding to a tenth of the population. You can change indicators to see the mean income within each decile, the threshold marking each one, or the share of total income that each decile makes up.8 Incomes have risen across the global distribution since 1990. Globally, the poorest, the middle, and the richest groups are all better off today than one generation ago. In absolute terms, these gains were very unequal. From 1990 to 2026, the income needed to just place someone in the richest 10% rose by about $18, from around $37 to $55 per person per day. The income level marking the poorest 10% also rose, but by only around $1.80, ten times less. In percentage terms, the largest gains were at the bottom half of the distribution. Although the rises at the bottom were small, in relative terms they were large. The incomes of the poorest 10% and the global median — those right in the middle of the distribution — have more than doubled since 1990. This growth in incomes lifted hundreds of millions of people out of extreme poverty. The rise for the richest 10% was much lower in proportional terms, increasing by around a half. What are the limitations of the World Bank’s data? The World Bank data shown above is a comprehensive source of data on global incomes, but it has limitations worth understanding. In order to achieve global coverage, the World Bank pools a wide range of national survey data. Although many steps are taken to harmonize this data, there are still limitations to the comparability of the data points between countries and over time. One important issue is the mix of income and consumption data already mentioned. The two concepts are related but not the same: income equals consumption plus savings. As people earn more, consumption usually becomes lower than income, with richer households tending to save a greater share of income. When people earn less, their consumption can often be higher than their income, as they can spend more than they earn by borrowing or drawing down savings. That’s often the case, for example, for older, retired people. As a consequence, inequality estimates tend to be somewhat higher when based on income rather than consumption data. One paper by World Bank researchers found that, on average, income-based Gini coefficients are about 13% higher than consumption-based ones for the same country.9 There are also comparability issues within countries over time, when survey methodologies change. The World Bank publishes a comparability indicator to help identify these breaks, and you can choose to reveal them in the chart above. But there are also important ways in which this data is made comparable. Two crucial points are worth mentioning. First, in many poorer countries, a large share of people produce food for their own use or exchange goods without money, and have little or no monetary income. To make a fair comparison of living standards, the data accounts for this: researchers estimate what these goods would have cost if the household had bought them, and add that value to their consumption. Second, to make income and consumption more comparable across countries, all figures are expressed in international dollars — a hypothetical currency that adjusts for differences in the cost of living between countries and for inflation over time. This is important because prices are generally much lower in poorer countries — without this adjustment, incomes in those countries would appear even lower relative to richer ones than they actually are in terms of what people can buy. This is a complex task, and the adjustments are inevitably not perfect. You can read more about some of the limitations in our article: What are international dollars? Beyond comparability, there are other limitations related to the surveys themselves. Household surveys struggle to capture the incomes of the very richest, who are few in number, less likely to participate in such surveys, and whose incomes tend to be undercounted when they do — partly because income from sources like investments and businesses is easily missed or misreported, and partly because some prefer not to disclose it at all. We discuss how another data source attempts to address this in the section below. The very poorest are also often missed because they are living without a fixed address, in institutions (such as shelters, care facilities, or prisons), or in areas that surveyors cannot reach (for example, due to conflict). Finally, not all countries conduct household surveys regularly. For some countries and years, no survey data is available, and this can give an incomplete picture. For example, as of 2026, around 30 countries had no survey data from the last decade. In our charts of World Bank data, we only show country data points for years with a survey in that country. However, when the World Bank produces the aggregates for entire world regions or the world as a whole, it must extrapolate the data to arrive at annual estimates, and this adds uncertainty to these broader estimates. Harmonized cross-country data from the Luxembourg Income Study How does income inequality compare across countries? It’s a harder question than it sounds. That’s because international comparisons of incomes — and the international datasets we can use for this — have to rely on national data that is often not directly comparable by default. All good international datasets tackle this comparability problem in some way — including the sources on this page and on Our World in Data more broadly. But there is a trade-off: wider global coverage typically comes at the cost of less consistent definitions across countries. The World Bank data presented above prioritizes broad global coverage, accepting that there are some differences across surveys. The Luxembourg Income Study (LIS) is a cross-national database, located in Luxembourg, which instead prioritizes comparability. To achieve this, LIS only includes surveys that are compatible with its harmonization framework. It takes the original national survey data (the responses from individual households) and harmonizes it by mapping the data onto a common set of definitions.10 One important adjustment LIS makes is for household size, known as “equivalization”. LIS adjusts household income to account for the fact that people in a household can share some costs, like heating and rent, so they need less income per person to achieve the same standard of living.11 This is also how statistical agencies in many rich countries typically measure inequality. LIS also captures incomes both before and after taxes and benefits, which makes it possible to study government redistribution, which plays a particularly large role in richer countries. The cost is that LIS covers fewer countries than the World Bank and has especially sparse coverage of low-income countries.12 The benefit is greater confidence in the comparisons it does make. The chart below shows the Gini coefficient — a widely used measure of income inequality that ranges from 0 (everyone has the same income) to 1 (all income goes to one person). You can explore how it varies across countries and over time, and how it changes when we look at income before and after taxes and benefits. In most countries, the Gini is lower after taxes and benefits — but the size of this gap varies widely. We discuss what this comparison tells us and what it doesn’t in another article: Income inequality before and after taxes: how much do countries redistribute income? For the United States and the United Kingdom, the data reaches back to the 1960s — long enough to track how inequality changed across very different political and economic periods. Through the 1970s, both countries had experienced broadly stable levels of income inequality. What followed in the 1980s was very different: the UK saw a large, steep rise in inequality in the late 1980s, followed by a more steady increase across the 1990s, while the US saw a more continuous, gradual increase throughout this period. Economist Tony Atkinson drew attention to what this comparison reveals in his book Inequality: What Can Be Done?13 He argued that comparing the speed and timing of these different paths across countries can help us understand their causes. The fact that they differ so much — despite both countries being exposed to similar global pressures — points to the role of national institutions, politics, and policy choices. It suggests that high and rising inequality is not inevitable. Tracking the incomes of the richest: data from the World Inequality Database All the inequality data we have seen so far comes from household surveys. Surveys are the most common source of data for inequality estimates, and they tell us a lot about incomes across most of the population. But survey data has a known blind spot at the very top of the income distribution. The extremely rich are few in number and so less likely to be asked to join a survey. When asked, they are less likely to participate. And even when they do participate, surveys tend to undercount their income — particularly income received from investments and business ownership. This is sometimes called the “missing rich” problem, and it means that survey-based measures of inequality likely understate the true level of income concentration at the top.14 To address this, researchers have developed a different approach drawing on other sources — in particular, tax records and national accounts.15 Tax records are often a better source for tracking top incomes because, unlike surveys, they draw on administrative reporting rather than voluntary participation, and can capture income from investments and business ownership in more detail. Tax data also has the important benefit of providing a longer-run perspective on inequality, thanks to the historical records available for some countries. National accounts data — countries’ official statistics that measure the whole economy — is used to ensure that all the income generated in the economy is accounted for. Together, these sources are used to construct distributional national accounts (DINA): estimates of how a country’s total national income is distributed across the population. This is the approach used by the team behind the World Inequality Database (WID). The database provides an important window into how the incomes of the very richest compare with the rest of society. However, this strength regarding top incomes does not mean the WID dataset is straightforwardly “better” than traditional survey-based data in general. First, the difference here is not just a question of accuracy; the WID data is measuring something different. As we explain more in the section below, it uses a much broader concept of “income” than found in survey data. This has some important advantages — especially for tracking top incomes — but it includes components people wouldn’t normally think of as part of their income. Second, the WID data also has some measurement weaknesses. In particular, its methodology is limited by the fact that, for many countries and periods, the underlying data it requires is not actually available.16 The WID project is very ambitious. It includes estimates of the entire income distribution, both before and after taxes and benefits, for almost every country in the world — in some cases dating back over a hundred years. This broad coverage is, in theory, a strength. But because the underlying data is missing in many cases, bold assumptions are often needed, making it hard to know how accurate the resulting estimates are. The WID team describes the issue like this:17 We should emphasize however that due to lack of proper data access in a large number of countries, these series should be viewed as imperfect and provisional. They are based in some cases on regional and country imputations based on regions and countries with similar characteristics… The specific assumptions made vary widely by country and period, and are documented by the WID team across many detailed papers and technical notes. In many cases, income estimates are made for countries where no income data exists at all.18 But even in data-rich countries like the United States, some important assumptions were needed to bring together the different underlying data sources. Researchers can, and do, disagree about these assumptions in ways that can meaningfully affect the resulting inequality estimates.19 Overall, the WID dataset offers some crucial contributions to our understanding of inequality around the world. It gives us a sense of the scale to which survey data may underestimate top incomes, and of the path inequality has taken over the long run in some countries. When looking at the data for any particular set of countries, though, the ambitious nature of the DINA approach and the limitations of the data need to be borne in mind. You can read more about the methodology below and on the WID website. The chart below lets you explore the WID data for the share of income received by different income groups, both before and after taxes and benefits. In some countries, the richest 1% receive more than a quarter of all national income For a relatively low inequality country like the Netherlands, WID estimates that the richest 1% received around 7% of all income before taxes and benefits in 2022. In a high inequality country like Brazil, they estimate that over a quarter of all income went to the richest 1% in the same year. Even among the richest 1%, income is highly concentrated. In the US, for example, the richest 0.1% — that is, one in a thousand people — received about 10% of the national income in 2024. That’s almost the same as the rest of the richest 1% combined. Inequality of what? Why WID and survey-based data measure different things A key innovation in the WID approach is that it combines multiple sources of data: tax records, national accounts data, and the more common survey data. The main motivation is to provide better estimates of top incomes than survey data alone can. But this approach also has implications for what kind of income is being measured. The concept at the heart of the WID definitions is net national income. Like the more familiar GDP, this is taken from the national accounts.20 WID uses tax and survey data to estimate the spread of different types of income across the population — wages, capital income, property income, and so on. These distributions are then scaled up so that the totals match the figures from the official national accounts. National income is a broader concept of income than that measured in survey-based datasets — including the LIS and World Bank data mentioned before. Let’s look at an example to see why this matters. Take the United States in 2024. Within the WID data, the average net national income per adult was $95,280.21 WID estimates that the top 1% of adults received 20.7% of pre-tax national income in that year — so the average income of the top 1% was around $2 million per year.22 If that seems high, part of the reason is that it includes elements you might not think of as income — items you would not find on a paycheck, tax return, or bank account, and crucially, money that an individual cannot directly spend. This is because WID aims to distribute all national income, including components that no individual directly receives. One important component is what is known as “undistributed corporate profits”: the income that companies retain after paying costs, taxes, and dividends to shareholders. No one receives this directly when profits are retained, but it still benefits the owners because it increases the value of the business. This benefit can be substantial at the top of the income distribution — and given WID’s focus and overall approach, it makes sense that it is counted in this way. However, it isn’t income in the usual sense — people cannot spend undistributed profits.23 Another aspect that is counted as income in WID is the imputed rent of owner-occupied housing. Homeowners who live in their own homes, of course, don’t receive rent as cash income, but they do receive a real economic benefit from the property they own: the income they would receive if they rented it out, or the foregone cost they would have to pay to rent a similar house. To account for this, national accounts treat homeowners as if they were paying rent to themselves. Like the retained profits, these imputed rents are not considered income in the usual sense, and the value of the benefit can be difficult to estimate.24 The income concept differs in other ways, particularly in how WID treats taxes and benefits. LIS and PIP measure income either before or after all taxes and benefits — a relatively straightforward distinction.25 But WID’s definitions are different: - Pre-tax income — also described by WID as “pre-tax, post-replacement income” — is measured before taxes are paid and most benefits are received, but after the receipt of social insurance benefits, such as public and private pensions (even though public pensions would normally be treated as government benefits). This unusual choice around pensions is made because countries organize pensions very differently — in public systems, private ones, or a mix of both — and treating them in the same way avoids the comparison being driven by those differences. - Post-tax national income captures income after taxes have been paid and government benefits have been received. This includes not only social benefits in cash, but also the value of public services like schools and hospitals, and other government spending like roads and defense. Not every part of national income can be straightforwardly attributed to individuals. In some cases, this is a difficult task. For example, how should government spending on public goods like education, defense, or infrastructure be attributed to individuals? Should it be attributed to everyone equally? Proportionally to income? To those who use each service the most? Researchers can, and do, disagree about this question in ways that can meaningfully affect the resulting inequality estimates. In its benchmark series, WID allocates most government spending proportionally to income, and public healthcare spending roughly equally across all adults.26 This difference in income concept reflects a trade-off at the heart of WID’s method. On the one hand, the broader, more standardized approach makes the data more comparable across countries and over time, and better suited to tracking what happens at the very top of the distribution, which is precisely where standard survey data falls short. But it comes at a cost: the definitions are more technical and further removed from what people would recognize as their income. Producing the estimates requires many assumptions that, in some cases, are inevitably somewhat arbitrary. Featured Data on Economic Inequality Research & Writing July 6, 2023 How has income inequality within countries evolved over the past century? While the steep rise of inequality in the United States is well-known, long-run data on the incomes of the richest shows countries have followed a variety of trajectories. August 25, 2025 Global inequality is huge — but so is the opportunity for people in high-income countries to support poor people People in high-income countries could dramatically improve lives worldwide with minimal financial commitment, yet few do. July 3, 2023 Income inequality before and after taxes: how much do countries redistribute income? The redistribution of income achieved by governments through taxes and benefits varies hugely. December 9, 2021 Global economic inequality: what matters most for your living conditions is not who you are, but where you are How much does it matter to be born into a productive, industrialized economy? April 17, 2017 The history of global economic inequality The inequality in people’s living conditions across the world is extremely large. How did the world become so unequal, and what can we expect for the future? June 30, 2023 Measuring inequality: what is the Gini coefficient? The Gini coefficient is the most common way of measuring inequality. But what does it actually measure? And how does it differ from other measures of inequality? August 28, 2019 Global Inequality of Opportunity Today’s global inequality of opportunity means that the good or bad luck of where you were born matters most for your living conditions. We look at how this chance factor is the strongest determinant of your standard of living, whether in life expectancy, income, or education. Data Insights on Economic Inequality Endnotes Income equals consumption plus savings. For households that save, income is higher than consumption by definition. On average, richer households tend to save a bigger share of their income, meaning that the gap between income and consumption rises with income. But at the very bottom of the distribution, the relationship often flips: by borrowing or spending down savings (i.e., negative savings), people may consume more than they earn. A common example is retired people drawing down their savings: they may have a very low income but still have a high level of consumption. A known weakness of the survey data on which the World Bank data relies is that the very richest and poorest are often not captured very well. The very rich are few in number and often don’t respond, or underreport their incomes when they do. And the very poorest can be hard to reach. We discuss this and other limitations in the section below. Because richer households tend to save a larger share of their income, inequality tends to be somewhat lower when measured in terms of consumption. So the ratios would likely be higher if we were measuring income for these countries. See Haddad et al. (2024), The World Bank’s New Inequality Indicator, World Bank Policy Research Working Paper 10796, Annex A.3. The large variation in inequality levels across countries — even accounting for plausible economic, technological, or demographic determinants — is a consistent finding in inequality research. Branko Milanovic expressed it like this: “[O]nce the ‘given’ elements are accounted for — there is still sizable discretion regarding income inequality. Income distribution is viewed also as the product of social choices mediated through elections, lobbying of various social groups, societal preferences or historical development.” - Milanovic, B. (1994). Determinants of Cross-Country Income Inequality: An “Augmented” Kuznets Hypothesis. World Bank Working Paper, 1246. The updated 2000 version is available at the Stone Center. Anthony Atkinson, in his 2015 book “Inequality: What Can Be Done?” and paper from the same year, “Can we reduce income inequality in OECD countries?” emphasized the many policy options that are open to countries, as evidenced by past trends and cross-country comparisons. - Atkinson, A.B. (2015). Can we reduce income inequality in OECD countries?. Empirica, 42(2), 211–223. - Atkinson, A.B. (2015). Inequality: What Can Be Done? Harvard University Press. For a more global analysis based on an earlier version of the World Bank dataset we are using here, see also Ravallion, M. (2018). Inequality and Globalization: A Review Essay. Journal of Economic Literature. Economic data is recorded in local currencies — rupees, dollars, pounds — and without adjusting for inflation. Before incomes (or consumption) can be compared across countries or over time, they need to be converted into a common unit. International dollars are a hypothetical currency used for this. Using international dollars adjusts for two things: inflation within each country (so values across years are comparable) and differences in living costs between countries, using purchasing power parity rates (PPPs), which reflect how much local currency is needed to buy what one US dollar would buy in the United States. One 2021 international dollar is defined as the value of goods and services that one US dollar would buy in the US in 2021. We explain this in more detail in our article What are international dollars? The surveys on which the World Bank estimates are based are conducted at the household level. Accounting for household size matters when considering what a household’s income or consumption means in terms of living standards. The World Bank takes a simple approach and divides the total household income or consumption by all members of the household, including children. Other datasets handle this in different ways. For example, the Luxembourg Income Study discussed below adjusts incomes to account for the fact that people in a household can also share some costs. A study by political economist Gautam Nair showed that Americans overestimate the global median income by a factor of ten. Nair, G. (2018) Misperceptions of relative affluence and support for international redistribution. The Journal of Politics, 80(3), 815–830. Giving What We Can, an organization that advocates for people to donate some of their income to help effective causes, made a video based on the same World Bank data we show here, which captures well how skewed our picture of the global income distribution tends to be. The data is originally expressed as daily figures. To convert to monthly or annual values, we simply multiply by 30 or 365, respectively. This is a mechanical conversion, and it doesn’t account for the fact that income and consumption can fluctuate over the course of a year. Household surveys ask about a specific reference period — often the past week or month — which is known as the survey “reference period”. In many lower-income countries, for example, earnings vary greatly with the agricultural seasons, and a person may be in poverty or out of it depending on when the survey was done. The World Bank data doesn’t capture this kind of variation within the year. You can read more about this in this World Bank blog. This is based on a sample of 84 country-year observations where both income and consumption were available, from Haddad et al. (2024), The World Bank’s New Inequality Indicator, World Bank Policy Research Working Paper 10796, Annex A.3. The World Bank PIP we presented earlier does this too — in fact, for many high-income countries, it relies on data from LIS. The difference is that LIS applies stricter comparability criteria, while the World Bank aims for broader global coverage and combines both income- and consumption-based surveys — a mix of concepts that can affect inequality comparisons. The LIS harmonization doesn’t resolve all the problems — for example, differences in survey design and response rates across countries remain — but it does reduce some of the methodological differences that would otherwise distort cross-country comparisons. LIS publishes the Compare.it tool to communicate which series have been harmonized for each country, as well as document any issues remaining. For example, a single person and a family of four cannot live equally well on the same income, but by living together, they can share some costs within the household: a household of four does not need four times the income to have the same standard of living. LIS uses the square root equivalence scale: it divides the total household income by the square root of the number of people in the household. This is one of many possible scales researchers use, and other datasets handle this in different ways. For example, the World Bank PIP discussed earlier provides figures per person, that is, dividing the total household income by the number of people in the household, including children. The World Inequality Database main series (discussed below) uses an “equal split” and divides it by the number of adults (aged 20 and over) in the household. The LIS coverage remains thinner than the World Bank’s PIP. At the time of writing in 2026, LIS covers 52 countries and PIP 172. But it is continuously growing to include more countries and periods. Atkinson, A. B. (2015). Inequality: What Can Be Done? Harvard University Press. Nora Lustig (2020) provides a good overview of the problems involved in capturing top incomes in household surveys accurately, and how researchers and statistical offices attempt to correct this. Lustig, N., 2020. The “missing rich” in Household Surveys: Causes and Correction Approaches (Vol. 520). ECINEQ, Society for the Study of Economic Inequality. Anthony Atkinson, Thomas Piketty, and Emmanuel Saez (2011) provide a good overview of initial research efforts in this field. Atkinson, A.B., Piketty, T., Saez, E (2011). Top Incomes in the Long Run of History. Journal of Economic Literature 49(1), 3–71. There are other limitations too. Tax data only captures income visible to the tax system. In countries where informal employment is very common, for example, tax data alone can miss income from a large share of the population. In addition, it can miss income that is unreported, evaded, or held offshore. For more on the limitations of tax data, see Atkinson, Piketty, and Saez (2011). Although national accounts data — to which the WID estimates are anchored — is often thought of as being more complete or otherwise more accurate than survey data, it is not straightforwardly the case that one is better than the other for the purposes of measuring inequality around the world. National accounts data has measurement problems of its own. For more on the differences between surveys and national accounts estimates, see Ravallion, M. (2003). Measuring aggregate welfare in developing countries: How well do national accounts and surveys agree?. Review of Economics and Statistics, 85(3), 645–652. And Deaton, A. (2005). Measuring poverty in a growing world (or measuring growth in a poor world). Review of Economics and Statistics, 87(1), 1–19. Taken from the WID website. We quote the WID team directly because they explicitly discuss the limitations and publish detailed technical notes documenting the assumptions and data availability for each country. For most African countries, for example, the WID’s income inequality estimates are based on surveys of household consumption. This is then adjusted based on the relationships between consumption and income data observed in a small number of countries. For most Middle East countries, some income survey data is available, but it is adjusted based on the relationship between tax and survey data observed just in one country: Lebanon. Moreover, capital income — which is counted in national accounts data, but often poorly captured in surveys — is added on to the adjusted survey data for these countries, assuming a pattern observed in the United States and France. See the technical note for Africa and a paper by WID researchers on income measurement in Africa: Chancel et al. (2023). Income inequality in Africa, 1990–2019: Measurement, patterns, determinants. World Development, Volume 163. For the Middle East, see the technical note and the earlier working paper. The debate between, on the one hand, Piketty, Saez, and Zucman — whose work the WID data for the US relies on — and Auten and Splinter on the other is a good example of this. Both teams use the same underlying US data and aim to measure the same thing: how national income is distributed across individuals. But they disagree about whether (post-tax) inequality has risen in recent decades, and by how much. One important reason for their different results is that they make different choices about how to allocate the large share (around 40% in recent years) of national income that doesn’t appear on individual tax data. This includes, for example, decisions on how to handle payroll taxes, social security benefits, and how to estimate tax evasion. You can read more in this post or, in more detail, in the 2025 paper by Clarke and Kopczuk. Gross domestic product (GDP) measures the total value of final goods and services produced within a country. To get from GDP to national income, you subtract the “consumption of fixed capital” (depreciation) — the costs to replace wear and tear in buildings, machinery, and other assets used in production — and add “net foreign income”: the income earned abroad by a country’s residents, minus the income earned domestically by non-residents. This figure comes from WID (accessed in July 2026) and is measured in constant 2025 US dollars. The net national income total is drawn from official national accounts data, and WID collects and standardizes these figures across countries to enable comparisons. This figure is per adult (aged 20 and over), so it is higher than what you would see in per capita measures. Calculated as net national income per adult ($95,280) multiplied by the top 1% income share (20.7%), divided by the population share of the top 1% (1%): $95,280 × (0.207 ÷ 0.01) = $1,972,296. This increase in the value of the business is also referred to as a latent capital gain — it exists on paper, but it hasn’t been realized. Realizing these gains means converting them into cash, for example, by selling shares. Whether a company pays out or retains profits often depends on tax incentives. The richest can choose when to realize profits and often have an incentive to take income in this form to defer paying taxes. By including undistributed profits as income, WID aims to make the data more comparable across countries and over time, and independent of specific tax regimes. See the DINA guidelines for how WID estimates imputed rents: depending on the country and year, it draws on estimates already included in household surveys, or imputes them itself. For a discussion of the different estimation methods, see Balcázar, C. F., Ceriani, L., Olivieri, S., & Ranzani, M. (2017). Rent‐Imputation for Welfare Measurement: A Review of Methodologies and Empirical Findings. Review of Income and Wealth, 63(4), 881–898. The before-tax measure in LIS excludes public pensions but includes private ones, meaning a retired person’s before-tax income depends on how their country organizes pensions. PIP focuses on after-tax measures. For more details, see the DINA guidelines. Cite this work Our articles and data visualizations rely on work from many different people and organizations. When citing this topic page, please also cite the underlying data sources. This topic page can be cited as: Joe Hasell, Bertha Rohenkohl, Pablo Arriagada, Esteban Ortiz-Ospina, and Max Roser (2023) - “Economic Inequality” Published online at OurWorldinData.org. Retrieved from: 'https://ourworldindata.org/economic-inequality' [Online Resource] BibTeX citation @article{owid-economic-inequality, author = {Joe Hasell and Bertha Rohenkohl and Pablo Arriagada and Esteban Ortiz-Ospina and Max Roser}, title = {Economic Inequality}, journal = {Our World in Data}, year = {2023}, note = {https://ourworldindata.org/economic-inequality} } Reuse this work freely All visualizations, data, and articles produced by Our World in Data are completely open access under the Creative Commons BY license. You have the permission to use, distribute, and reproduce these in any medium, provided the source and authors are credited. The data produced by third parties and made available by Our World in Data is subject to the license terms from the original third-party authors. We will always indicate the original source of the data in our documentation, so you should always check the license of any such third-party data before use and redistribution. All of our charts can be embedded in any site.