貧窮本身具有多樣性特質,無法以單一指標或元素來衡量,近年來國際上已有越來越多研究致力於多面向貧窮的測量。然而,我國貧窮測量研究雖然相當豐沛,但大多使用和經濟資源有關的單面向貧窮做為衡量基礎。有鑑於此,本研究應用Alkire-Foster計數法,計算多面向貧窮指數(由教育、醫療、環境、健康以及生活水準等五個剝奪構面所組成),分析我國多面向貧窮的樣貌。結果發現:一、同時考慮多面向貧窮發生率和貧窮強度,可避免多面向貧窮指數被高估。二、我國縣市之間的多面向貧窮指數分布差異大,貧窮縣市主要集中在中南部。三、剝奪構面在各縣市的貢獻度皆呈現地區差異。四、貧窮組成要素之分解對辨識各類群體在多面向貧窮的嚴重程度有幫助,且剝奪構面在不同組成要素之分解,亦有助於找出對減輕多面向貧窮較有影響力的剝奪構面。
Poverty cannot be accurately measured using a single indicator because of its multidimensional nature. In recent years, considerable research has been conducted internationally on multidimensional poverty measurement. However, most research on poverty in Taiwan has focused primarily on unidimensional poverty measurement related to economic indicators. Therefore, this study explored multidimensional poverty in Taiwan by applying the Alkire Foster method. We also used the multidimensional poverty index (MPI), a composite measure of education, medical resources, environment, health, and standard of living. We found that the MPI was not overestimated when the incidence and intensity of multidimensional poverty were accounted for. The MPI substantially varied across different cities and counties and was high in South-Central Taiwan. Regarding decomposition, the contributions of deprivation in each dimension to overall poverty varied by region. The subgroup decompositions allowed us to identify each subgroup’s multidimensional poverty and compute the contribution of each subgroup to overall poverty.