Kernel Density Estimation Based on Grouped Data: The Case of Poverty Assessment
IMF Working Papers, July 1, 2008
Source details
- Canonical URL
- Kernel Density Estimation Based on Grouped Data: The Case of Poverty Assessment
Other formats
Bibliographic details
- Authors: Camelia Minoiu, Sanjay Reddy
- Published: July 1, 2008
- Series: IMF Working Papers
- DOI: https://doi.org/10.5089/9781451870411.001
Summary and main conclusion
- Study analyzes the performance of kernel density methods applied to grouped data to estimate poverty (as applied in Sala-i-Martin, 2006, QJE).
- Using Monte Carlo simulations and household surveys, the authors find that the technique gives rise to biases in poverty estimates.
- The sign and magnitude of biases vary with the bandwidth, the kernel, the number of datapoints, and across poverty lines.
- The authors conclude: "Our findings challenge the validity and robustness of poverty estimates derived through kernel density estimation on grouped data."
Key quantitative findings
- Depending on the chosen bandwidth, the $1/day poverty rate in 2000 varies by a factor of 1.8.
- The $2/day headcount in 2000 varies by 287 million people, depending on the chosen bandwidth.
Methodology
- Approaches used: Monte Carlo simulations and analysis of household surveys.
- Focus: Application of kernel density estimation methods to grouped income/consumption data for poverty measurement.
- Factors examined: bandwidth choice, kernel choice, number of datapoints, and different poverty lines.
Scope and metadata
- Authors: Camelia Minoiu, Sanjay Reddy
- Date: July 1, 2008
- Series: Working Paper No. 2008/183
- Issue: 183
- Pages: 34
- ISBN: 9781451870411
- ISSN: 1018-5941
- DOI: https://doi.org/10.5089/9781451870411.001
IMF Working Paper — "Kernel Density Estimation Based on Grouped Data: The Case of Poverty Assessment", Camelia Minoiu and Sanjay Reddy, July 1, 2008.
Content in this bundle
- _wp08183 — References