引用
貧窮持續時間的動態分析:以嘉義縣1990-1998年之低收入戶爲例
The Dynamics of Poverty Spells: An Analysis of the Panel on Households in Poverty in Chia-Yi County, 1990-1998
作者:王仕圖(Shu-Twu Wang)、玉德睦(Te-Mu Wang)、蔡勇美(Yung-Mei Tsai) | 首次發表於 2020-07-06 | 第 26 期 December 2001
DOI:https://dx.doi.org/10.6786/TJS.200112.0211
研究論文(Research Articles)
論文資訊 | Article information
摘要 Abstract
早期許多貧窮研究都存在一種看法,就是影射一群人口長期處於貧窮的狀態。但「所得動態的追蹤研究」(Panel Study of Income Dynamics)的研究卻發現福利使用者雖然有少數是長期的福利依賴者,卻有相當部分的使用者只是短暫的停留。這個論點使得貧窮動態的研究成爲關心的焦點。本研究利用1990-1998年之嘉義縣低收入家戶的追蹤資料,分析貧窮持續時間的動態、不同觀察方式貧窮時間分佈、以及影響脫離貧窮的因素等。以1990年的低收入戶作爲起始年並計算貧窮的持續時間,則貧窮的中位年數爲4.69年。以「貧窮時段」(poverty spell)爲計算單位,貧窮的中位年數爲3.97年半的家戶會在第四年以前脫離貧窮,有四分之一的家戶在追蹤時間結束時尚未脫離。不同觀察方式下的貧窮時問分佈方面,完整的貧窮時段的分配部分,超過30%的家戶會在第一二年以前脫離貧窮。但從某一時間點觀察脫離貧窮者的持續時間分佈,一半以上的福利資源爲持紛續八年以上的家戶所使用。另外從「比例風險模型」的分析,可以瞭解低收入家戶戶長的年齡、教育年齡、健康地位是影響家戶脫離貧窮風險的重要個人特質; 在家戶特質部分,單親家戶和戶內具有工作能力人口則具有顯著的影響。

關鍵字:貧窮、貧窮時段、追蹤調查、存活分析
Most poverty research operates from a similar perspective, looking at the conditions of people during poverty. However, the Panel Study of Income Dynamics revealed that only a small number of welfare recipients receive it for an extended period, and a great number of welfare recipients receive it for only a short time. This has now become a focal point of further poverty research.
Based on panel data from Chia-Yi County during the 1990-1998 period, we analyzed the duration of poverty spells and the distributions of completed and uncompleted spells of poverty. We also report results of our analysis on factors affecting escape from poverty among these households . The results of our analysis indicate that the median duration of poverty was 4.69 years, when cohort-based estimates were used. When the data on the spells of poor households was used, the median spell of poverty was estimated to be 3.97 years long. More than 50 percent of poor households escaped poverty within 4 years and only one-quarter of poor households were in poverty for over eight years. The results of comparing the completed and the uncompleted spells show that among the former the duration tended to be shorter than the latter. Among the completed spells, more than 30 percent of them ended within 2 years. However, at any given point in time, the bulk of poor households were those experiencing ar elatively long-term hou t of poverty. When the proportional hazard model was used, the age of the heads of households, the number of years of schooling, and health status were found to be important factors contributing to the possibility of leaving poverty. Among the household characteristics, we found that the single-parent households and the number of household members with the ability to work directly affect the probability of leaving poverty.

Keywords: Poverty, Poverty Spell, Panel Study, Survival Analysis