New WIID Companion to improve the study of inequality
There is a growing need to understand income inequality trends and how they interplay with other social, economic, and political outcomes, both at the country level and worldwide. Despite significant progress during the last decades in improving the availability of data to study inequalities, substantial constraints still limit our ability to conduct consistent cross-country comparisons and analyses of long-term trends of income distributions.
UNU-WIDER has contributed to this important task by hosting and updating the World Income Inequality Database (WIID), which includes all of the main data on income distributions which is available for most countries. With the new version of the database (March 2021), UNU-WIDER has also released two new WIID companion datasets, one with information about the income distribution at the country level for 200 countries and historical entities, the other with estimates of the global income distribution.
Integrated series standardize the data across the WIID
These new datasets complement the WIID first by simplifying the information into selected, single time series that best describe the income distribution trend in each country. These integrated series cover the longest possible period with the highest possible consistency and are primarily based on household surveys from the main international sources, such as the Luxembourg Income Study (LIS), PovcalNet, Eurostat, and national statistical authorities.
The country-level dataset is constructed with the minimum necessary adjustments to the original survey data which are needed to integrate and standardize the information so that it becomes more comparable across countries and over time. Importantly, these adjustments achieve improved comparability while maintaining the main patterns found in the original source data. Additionally, the distributional data has been standardized so that it always refers to the same welfare concept — i.e., household net income per capita at the country level.
The entire global income distribution over time
Finally, these integrated series for country-level income distributions over time are aggregated to produce the global income distribution reported in the global WIID companion, where inequality is measured among the entire world population. Having both aggregated and disaggregated components allows for the study of between- and within-country components separately, or for examining income distribution by region or country income group. The step-by-step construction of these companion datasets is fully documented in a series of technical notes, alongside the necessary Stata codes to reproduce them.
The new datasets enhance, at the country level and globally, the information that is available in the WIID providing the entire distribution of income at the percentile level alongside a variety of inequality measures, whether relative or absolute. These data will facilitate more comprehensive and integral distributive analysis within and across countries and worldwide. The first analysis using these data has recently been published in the WIDER Working Paper Series and provides one example of their potential.
A multiplicity of approaches to studying and understanding inequality
These data should allow researchers to identify the complex distributional changes that have taken place and come to some consensus on how best to qualify these changes. This requires admitting that there are different, legitimate, distributive sensitivities apparent from the data. Rather than imposing a specific approach, we have tried to give users the flexibility to choose their own approach with the implicit, or explicit, value judgements that come with it.
The approach we followed in constructing the global database — producing an annual series — facilitates regular updates and revisions of the global distribution as new information becomes available. It also allows for projections based on different scenarios. The new database will thus contribute to better monitoring of progress in reducing inequalities, in the spirit of Sustainable Development Goal 10.
Overall, this is a good representation of inequality trends, as far as they can be judged given the paucity of information. Obviously, accuracy improves in the most recent decades, particularly after 1980 or 1990, depending on the country. There is no way to overcome the fact that in earlier years, reliable information on within-country inequality is scarce and often of lower quality. Therefore, the global dataset needs to be highly imputed (e.g., extra/interpolated), implicitly or explicitly, to capture earlier years.
We report the global inequality levels for early years, despite the risks associated with data quality/availability issues. The best existing data provides, at least, a reference and some context to better interpret what happened afterwards. Once distributive information becomes denser and more accurate values estimates become quite good for most countries.
The WIID database provides the type of distributive information that is usually represented in household surveys. The standard limitations of using survey-based information apply, particularly regarding the potential of misestimation at the extremes of the distribution (incomes of the very poor and the very rich). There has been no attempt in this release to make corrections for this, which can be explored for future versions of the data or included in our research analyses.
Finally, people’s wellbeing is not only determined by individual monetary incomes. An overall assessment of inequality, whether at the country level or globally, must consider other inequalities that affect the freedom of people in developing their capabilities along other various dimensions, especially health or education, which are not always perfectly correlated (and therefore well captured) by monetary income alone. However, the dynamics of income inequalities contributes to understanding on more general social inequalities.
The views expressed in this piece are those of the author(s), and do not necessarily reflect the views of the Institute or the United Nations University, nor the programme/project donors.
Carlos Gradín is a Research Fellow at UNU-WIDER. He specializes in the study of inequalities and his work work regularly enhances the empirical evidence and methodological tools used to measure and understand inequalities. He is the project leader for the World Income Inequality Database project and led on the creation of the WIID Companion datasets.