The Conditional Gini: Estimation and Application to the Relationship between Wealth, Financial Use, and Income Inequality

The Gini measure of inequality can be written as the ratio of dyadic means. We define the conditional Gini as the ratio of conditional dyadic means. The conditional Gini shows how inequality varies with observables, both levels and differences. We illustrate the approach in a simple analytical example, and discuss how it compares to other techniques for attributing inequality to covariates. We show how to estimate the conditional Gini using dyadic regressions, and how to conduct inference using a system dyadic clustering formula. The conditional Gini is estimated using a large, nationally representative household survey from Thailand. The motivation is a development literature implying that financial development can weaken or strengthen the links between income inequality and wealth differences. We find that wealth differences matter significantly less for income inequality among households that use the financial sector than among those that do not, consistent with the idea that financial access relaxes self-financing constraints and broadens economic opportunity. ∗Preliminary Draft: Do not cite or distribute without author’s permission. We thank Todd Elder, Marcel Fafchamps, Steven Haider, Joe Kaboski, Krislert Samphantharak, Gary Solon, Jeff Wooldridge, and seminar/conference participants at CIRPEE/Laval, Notre Dame, and BREAD-Trento-Verona. All errors are ours.