Bayesian Approach to Lorenz Curve Using Time Series Grouped Data
Bayesian Approach to Lorenz Curve Using Time Series Grouped Data
复制标题
使用时间序列分组数据的洛伦兹曲线贝叶斯方法
DOI：
10.1080/07350015.2021.1883438
复制
发表时间：
2021
影响因子：
3
通讯作者：
Sugasawa Shonosuke
中科院分区：
文献类型：
--
作者：
Kobayashi Genya;Yamauchi Yuta;Kakamu Kazuhiko;Kawakubo Yuki;Sugasawa Shonosuke
本研究关注的是从时间序列分组的收入比例数据中估计与潜在的假设收入分配相关的不平等措施。我们采用Dirichlet似然方法，将Dirichlet似然参数设置为连续收入类别假设收入分布的Lorenz曲线之差，并通过时间序列结构将Lorenz曲线的变换参数结合起来，提出了一种状态空间模型。本文还研究了通过考虑Dirichlet分布的广义版本扩展似然模型的可能性，其中平均值是基于具有附加层次结构的Lorenz曲线建模的。日本月收入调查的模拟数据和真实数据证实，所提出的方法比现有的没有时间序列结构的独立估计模型的方法对不平等措施产生更有效的估计。
This study is concerned with estimating the inequality measures associated with the underlying hypothetical income distribution from the times series grouped data on the income proportions. We adopt the Dirichlet likelihood approach where the parameters of the Dirichlet likelihood are set to the differences between the Lorenz curve of the hypothetical income distribution for the consecutive income classes and propose a state-space model which combines the transformed parameters of the Lorenz curve through a time series structure. The present article also studies the possibility of extending the likelihood model by considering a generalized version of the Dirichlet distribution where the mean is modeled based on the Lorenz curve with an additional hierarchical structure. The simulated data and real data on the Japanese monthly income survey confirmed that the proposed approach produces more efficient estimates on the inequality measures than the existing method that estimates the model independently without time series structures.