Statistical Inference and Decomposable Poverty Measures.

The paper develops simple statistical test procedures for generalized decomposable poverty measures. It is shown that the estimates of the decomposable poverty indices, including the overall and various sub-group poverty indices, have (asymptotic) jointly normal distributions. The full variance-covariance structure is derived and can be estimated consistently without prior specification of the population density underlying the sample data. The (overall/decomposed) poverty indices can thus be used as tools for statistical inference instead of simply as descriptive statistics. We illustrate the test procedure by examining the contributions of whites and non-whites to total U.S. poverty in 1975 and 1985. Copyright 1995 by Blackwell Publishing Ltd and the Board of Trustees of the Bulletin of Economic Research