1 option
Exploring the Sources of Downward Bias in Measuring Inequality of Opportunity / Ibarra, Gabriel Lara.
World Bank Open Knowledge Repository (formerly "World Bank E-Library Publications") Available online
View online- Format:
- Book
- Government document
- Author/Creator:
- Ibarra, Gabriel Lara
- Series:
- Policy research working papers.
- World Bank e-Library.
- Language:
- English
- Subjects (All):
- Economic theory & research.
- Health, nutrition and population.
- Income distribution.
- Inequality.
- Inequality of opportunity.
- Macroeconomics and economic growth.
- Mean log deviation.
- Monte carlo.
- Population policies.
- Poverty impact evaluation.
- Poverty reduction.
- Rural poverty reduction.
- Top incomes.
- Local Subjects:
- Economic theory & research.
- Health, nutrition and population.
- Income distribution.
- Inequality.
- Inequality of opportunity.
- Macroeconomics and economic growth.
- Mean log deviation.
- Monte carlo.
- Population policies.
- Poverty impact evaluation.
- Poverty reduction.
- Rural poverty reduction.
- Top incomes.
- Physical Description:
- 1 online resource (38 pages)
- Place of Publication:
- Washington, D.C. : The World Bank, 2015.
- System Details:
- data file
- Summary:
- This study analyzes the extent of downward bias in the calculation of inequality of opportunity for continuous outcomes such as income. A typically recognized source of bias is the unobserved circumstances as there is a limited set of variables available in household and labor force surveys. Another previously overlooked source is the likely unobservable nature of top incomes. Using Monte Carlo simulations where the underlying inequality of opportunity is predetermined at various levels, the study presents three key findings. First, the omission of a relevant circumstance can bias the inequality of opportunity estimate by as much as 80 percent, depending on how much variation of the outcome such circumstance explains. Second, not observing the top 5 percent of the income distribution can lead to downward biases of anywhere between 12 and 35 percent, and the combination of missing the most favored population and even one relevant circumstance exacerbates the bias of the empirical estimates. The third key result is that the estimated inequality of opportunity is strongly correlated with the amount of variation in the outcome variable explained by the combination of circumstances (measured by the R2). This result suggests that in empirical applications, the inequality of opportunity estimate can be roughly (and quickly) approximated using simple econometric techniques.
The Penn Libraries is committed to describing library materials using current, accurate, and responsible language. If you discover outdated or inaccurate language, please fill out this feedback form to report it and suggest alternative language.