Outlier Detection for Welfare Analysis

Extreme values are common in survey data and represent a recurring threat to the reliability ofboth poverty and inequality estimates. The adoption of a consistent criterion for outlier detection is useful in manypractical applications, particularly when international and intertemporal comparisons are involved. This paper discussesa simple, univariate detection procedure to flag outliers in the distribution of any variable of interest. It presentsoutdetect, a Stata command that implements the procedure and provides useful diagnostic tools. The output of outdetectcompares statistics—with focus on inequality and poverty measures—obtained before and after the exclusion ofoutliers. Finally, the paper carries out an extensive sensitivity exercise, where the same outlier detectionmethod is applied consistently to per capita expenditure across more than 30 household budget surveys. The resultsare clear-cut and provide a sense of the influence of extreme values on poverty and inequality estimates.