This dataset was created using the Poverty and Inequality Platform (PIP) for Mahler ⓡ Yonzan ⓡ Lakner (2022). It has been updated using the September 2026 vintage of PIP data. The data vintage used in the paper can be found here.

This dataset contains a distribution of global welfare for the years 1990 to 2026. For each year, the dataset contains 1000 welfare (income or consumption) bins per country. The dataset reports the following variables: country name and 3-letter country code, World Bank regions (region_name, region_code), quantile—rank of the bin with 1 being the lowest welfare bin and 1000 the highest welfare bin, welf—average daily household per capita welfare values of each bin in 2021 PPP USD, and pop—population of the bin in millions.

Note that country distributions are based on lining up surveys for each country following the same methodology applied to lineup estimates in PIP. Also following PIP's Fall 2025 release, the distributions for countries without any survey in PIP are imputed using the method outlined in Mahler et al. (2026). Distributions for countries without survey in releases prior to Fall 2025 were based on the geographic region's average. For details, see PIP's Methodology Handbook. 

Recommended citation: 

“Mahler, Daniel Gerszon; Yonzan, Nishant; Lakner, Christoph. 2022. The Impact of COVID-19 on Global Inequality and Poverty. Policy Research Working Papers;10198. © World Bank, Washington, DC. https://openknowledge.worldbank.org/entities/publication/54fae299-8800-585f-9f18-a42514f8d83b License: CC BY 3.0 IGO.” updated with Poverty and Inequality Platform (September 2026).

Disclaimer: 

The data should not be used in lieu of the full survey data, or in lieu of the poverty and inequality statistics that are estimated directly from the survey data. For these purposes, users should use the estimates available in the Poverty and Inequality Platform (PIP). The poverty and inequality statistics published in PIP are estimated directly from the survey data. For example, the binned database would result in a lower level of within-country inequality, since it does not account for the inequality within each bin. The dataset is designed for global and regional or other aggregated analysis. For country-level analysis, users should work directly with the underlying surveys. Survey percentile distributions are available here.