The economics, statistics, and data behind the percentile
When you enter your annual income and country into our calculator, we answer a deceptively complex question: what percentage of the world's 8 billion people earn less than you?
This is not as simple as looking up your income on a single global list. There is no master database of every person's earnings. Instead, economists model income distributions within each country using statistical distributions, then aggregate across all countries weighted by population.
Our calculator follows the same methodology used by the World Bank's Poverty and Inequality Platform and academic researchers like Branko Milanovic, whose pioneering work on global income distribution has been widely published. The process works in three steps:
A software engineer in San Francisco earning $120,000 and a software engineer in Bangalore earning ₹2,000,000 (roughly $24,000 at market exchange rates) might actually have similar living standards. That's because a dollar buys far more in India than in the United States — more food, more housing, more services. Market exchange rates don't capture this difference.
Purchasing Power Parity is an economic concept that adjusts for these price differences. The idea originates from the "law of one price" — identical goods should cost the same everywhere when expressed in a common currency. In practice, the World Bank's International Comparison Program (ICP) sends researchers to price a basket of hundreds of goods and services in every participating country, then computes conversion factors.
Conversion: Income in PPP dollars = Local income × PPP conversion factor
Example (South Korea):
Annual salary: ₩50,000,000
PPP factor: 1 ÷ 855.4 ≈ 0.001169
PPP income: ₩50,000,000 × 0.001169 ≈ $58,450 PPP
PPP conversion factors are updated periodically by the World Bank. The most recent comprehensive round (ICP 2021, published 2024) involved 176 participating economies. Our calculator uses 2023 estimated factors.
Why not use market exchange rates? Exchange rates are heavily influenced by capital flows, interest rate differentials, speculation, and trade imbalances. They can fluctuate 20–30% in a single year without any change in actual purchasing power. PPP provides a much more stable and meaningful basis for comparing living standards across borders.
How is income distributed within a country? This has been studied extensively by economists since Vilfredo Pareto first observed in 1896 that wealth in Italy followed a power law. Since then, researchers have found that the log-normal distribution provides an excellent fit for the vast majority of the income spectrum (roughly the bottom 99%).
A variable is log-normally distributed if its natural logarithm follows a normal (Gaussian, or "bell curve") distribution. Income is a textbook example: most people cluster around the median, with a long right tail of high earners. Unlike a normal distribution (which is symmetric), the log-normal is right-skewed — it naturally captures the fact that there are many more people earning slightly below the median than there are people earning 10× the median.
The log-normal distribution is defined by two parameters:
Parameters:
μ (mu) = ln(median income) — the center of the distribution on the log scale
σ (sigma) = √2 × erfinv(Gini coefficient) — the spread, derived from inequality
The connection between σ and the Gini coefficient is an elegant mathematical result. For a log-normal distribution, the Gini coefficient G equals erf(σ / √2), where erf is the Gaussian error function. Inverting this relationship lets us derive σ directly from the Gini coefficient.
Once we have μ and σ for a country, we can calculate the probability that a randomly selected person in that country earns less than any given income level. This is the CDF of the log-normal distribution:
P(Income ≤ x) = ½ × [1 + erf((ln(x) − μ) / (σ × √2))]
Where erf is the Gaussian error function, ln is the natural logarithm, and √2 ≈ 1.4142
The error function (erf) is a well-studied mathematical function that appears throughout probability and statistics. It computes the probability that a normally distributed variable falls within a given range. Our implementation uses the Abramowitz and Stegun approximation (Formula 7.1.26 from the Handbook of Mathematical Functions, 1964), which provides accuracy to within 1.5 × 10⁻⁷ — more than sufficient for percentile calculations.
For the inverse error function (erfinv), needed to convert Gini coefficients to σ values, we use Winitzki's approximation (2003), a compact closed-form formula that provides good accuracy across the full domain.
The Gini coefficient, developed by Italian statistician Corrado Gini in 1912, is the most widely used measure of income inequality. It ranges from 0 (perfect equality — everyone earns the same) to 1 (perfect inequality — one person earns everything).
Geometrically, the Gini coefficient is defined as twice the area between the Lorenz curve and the line of perfect equality. The Lorenz curve plots the cumulative share of income (y-axis) against the cumulative share of the population ranked by income (x-axis). In a perfectly equal society, the Lorenz curve would be a straight diagonal line; in practice, it bows below the diagonal, and the further it bows, the higher the Gini.
| Country | Gini | Interpretation |
|---|---|---|
| South Korea | 0.312 | Relatively equal (OECD member) |
| Japan | 0.329 | Moderate inequality |
| United Kingdom | 0.351 | Moderate inequality |
| United States | 0.398 | High for a developed nation |
| China | 0.382 | Moderate-high |
| Brazil | 0.489 | High inequality |
| South Africa | 0.630 | Among the highest in the world |
Scandinavian countries (Denmark, Norway, Sweden) typically have Gini coefficients around 0.27–0.29, while sub-Saharan African and Latin American countries often exceed 0.45. The global Gini coefficient (treating all humans as one population) is approximately 0.65–0.70, reflecting the enormous gap between rich and poor nations.
The global percentile is where all the pieces come together. For each country in our database, we compute the fraction of that country's population earning less than your PPP-adjusted income, then weight by population:
Global Percentile Calculation:
For each country i:
1. Compute σᵢ = √2 × erfinv(Giniᵢ)
2. Compute μᵢ = ln(median incomeᵢ)
3. Compute fractionᵢ = LogNormalCDF(your PPP income, μᵢ, σᵢ)
4. Compute weightedᵢ = fractionᵢ × populationᵢ
Global percentile = Σ(weightedᵢ) ÷ Σ(populationᵢ)
This is equivalent to imagining that you randomly select a person from the entire world population and asking: what is the probability that they earn less than you? Each country contributes to this probability in proportion to its population — China and India, with populations exceeding 1 billion each, dominate the calculation. This is why even a modest middle-class income in a developed country often places you in the top 5–10% globally.
Example: An American earning $55,000/year (close to the U.S. median) translates to roughly $55,000 in PPP terms (the U.S. dollar is the PPP reference currency). This income is higher than the median income in virtually every other country on Earth. After summing across all countries weighted by population, this person typically lands around the 92nd–95th global percentile.
The log-normal distribution is an excellent model for the bottom 99% of the income distribution, but it systematically underestimates the thickness of the upper tail. In plain terms: there are more extremely wealthy people than a log-normal model would predict.
This observation goes back to Vilfredo Pareto himself, who noticed that the upper tail of wealth distributions follows a power law — the famous "Pareto distribution." The probability of earning more than a given amount x falls off as x−α, where α (alpha) is typically between 1.5 and 2.5 for income distributions across different countries.
Our calculator applies a Pareto correction for incomes above a threshold (the approximate entry point to the global top 1%):
Pareto correction (for top 1% incomes):
α (Pareto exponent) = 1.8
Survival function: S(x) = (x_min / x)^α
Corrected percentile = 1 − (0.01 × S(x))
Where x_min is the income threshold marking the top 1%
The Pareto exponent of 1.8 is empirically derived from wealth distribution studies, including the World Inequality Database (WID) and research by Thomas Piketty, Emmanuel Saez, and Gabriel Zucman. A lower alpha means a "fatter" tail (more extreme wealth), while a higher alpha means wealth drops off more quickly.
Without this correction, someone earning $500,000/year might calculate as the 99.99th percentile under the log-normal model, when the Pareto-corrected estimate is closer to the 99.5th — because the real distribution has far more people at $500K+ than the log-normal predicts.
Our calculator also shows your percentile within your own country. This uses the same log-normal model but applied to a single country rather than aggregated globally:
National percentile = LogNormalCDF(your PPP income, μ_country, σ_country)
The difference between national and global percentile can be dramatic. Someone at the 50th percentile in the United States (median U.S. income) is simultaneously at roughly the 93rd percentile globally. Conversely, someone at the 50th percentile in India is at about the 25th percentile globally.
This gap illustrates one of the central findings of global inequality research: where you are born matters more than almost any other factor in determining your income. Milanovic (2016) estimated that about two-thirds of global income inequality is explained by country of birth alone — the "citizenship premium."
Our calculator draws on authoritative economic data from multiple sources. For each country in our database, we maintain six key data points:
| Data Point | Primary Source |
|---|---|
| Median income (PPP) | World Bank Poverty & Inequality Platform |
| Gini coefficient | World Bank / OECD Income Distribution Database |
| PPP conversion factor | World Bank International Comparison Program (ICP 2021) |
| Nominal exchange rate | IMF International Financial Statistics |
| Population | UN World Population Prospects (2023 revision) |
| Wage benchmarks | National statistical offices (minimum wage, median, top 10%) |
Our database currently covers countries representing over 85% of the world's population and GDP. For countries not individually modeled, the global distribution calculation still provides meaningful results because the major population centers (China, India, the US, Indonesia, Brazil, etc.) are all included.
To make the calculator more useful, we include three reference points for each country sourced from national statistics bureaus:
These benchmarks are provided in local currency to give users an intuitive sense of where their income falls relative to their country's norms, before seeing the global comparison.
Ready to find out where you stand globally?
Calculate Your Wealth Rank