Why GRID?
Several cross-country databases on income inequality already exist. So why GRID?
The reason is that cross-sectional inequality measures offer a series of static snapshots of an economy. The key limitation is that they do not contain information on the dynamics of individual incomes or on the degree of mobility across the income distribution.
This information requires tracking individuals over time. And GRID does exactly that. This longitudinal dimension is necessary for evaluating the welfare implications of changes in the income distribution and thinking about the design of social insurance programs.
Four Pillars of GRID
The four pillars of GRID are: (1) longitudinal, (2) administrative, (3) granular, (4) harmonized.
Longitudinal: GRID statistics are computed from longitudinal micro data, which allows documenting properties of individual income fluctuations, understand the nature of income risk (e.g., the size, asymmetry, and persistence of income shocks), and estimate mobility across the income distribution. Longitudinal data also allows comparisons of inequality trends in current income to trends in measures of long-run income status (“permanent income”).
Administrative Records: The record-based nature of the underlying data reduces concerns about measurement error. It therefore allows to produce a wide range of well-measured non-parametric (quantile-based) statistics that are informative about the shape of the distribution of income changes (whether they are left- or right-skewed, fat-tailed, leptokurtic, etc.) and analyze the nature of tail shocks (e.g. very large income declines that disrupt labor market trajectories of individuals workers).
Granular: The granular nature of the underlying micro data (typically, millions of observations per year) enables the computation of a wide range of well-measured statistics for finely defined sub-populations, which allows researchers to assess, for example, whether trends in inequality and economic uncertainty have evolved differently for different groups. These are key inputs into all public policy decisions that take redistribution into consideration.
Harmonized: GRID was designed from ground up with a focus on comparability across countries. The focus on harmonization guided our choice of which administrative dataset to use for each country, sample cleaning and selection, and most importantly, our decision to write a single master code to produce all the statistics for every country in GRID. The resulting harmonized statistics can allow researchers to study the importance of institutions vs. the role of market forces, especially during recessions and crises, and the effects of structural reforms or policy changes on different parts of the society.
Citing GRID
If you use GRID statistics in your research, please cite it as
”Guvenen, Fatih, Luigi Pistaferri, and Giovanni L Violante, 2022. “The Global Repository of Income Dynamics,” https://www.grid-database.org. Accessed DD.MM.YYYY.
For methodology, data description, or variable construction in GRID, please cite
”Guvenen, Fatih, Luigi Pistaferri, and Giovanni L Violante, “ 2022. “Global Trends in Income Inequality and Income Dynamics: New Insights from GRID, Quantitative Economics, Vol 13, pp. 1321-1360.
Download BibTex Citation Record
Additional Documents
Below, we provide several documents. The first one details the goals of the project and our expansion plans. The others are technical documents that describe our sample selection guidelines, the exact definitions for all our statistics, and some country-specific information on data construction.