Quantile means and quantile share standard errors and a toolbox of distributional statistics
推导了分位数均值和分位数份额的渐近方差与协方差公式，并基于此构建了分位数分解的不平等测度工具箱，应用于加拿大2000和2005年收入数据，揭示了高收入增长、中等收入份额下降及性别差异等显著统计特征。
This article derives the (asymptotic) variances and covariances – and hence standard errors – of quantile means and quantile shares in terms of explicit formulas that are distribution-free and easily computable. The article then develops a toolbox of quantile-based disaggregative inequality measures, based on the means and shares, which allow for detailed inferential analysis of income distributions in a straightforward unified framework. The analytical formulas are applied to Canadian Census public-use microdata files on workers’ earnings for 2000 and 2005. The results highlight the statistical significance of how upper-earnings levels have advanced beyond middle earnings, how much the share of mid-range earnings has eroded over even a five-year period, and how decile mean growth rates for women were everywhere higher than for men – except at the top decile, where the opposite phenomenon was highly significant.