臺灣家戶飲食類別之差異分析 —貧窮和一般家庭之比較

Abstract

貧戶在消費上受制於經濟資源的匱乏，導致其消費上與非貧家戶有所差異；政策制定上常忽視此差異，若假設貧窮及非貧家戶消費沒有差異而制定經濟或福利相關政策，政策對貧窮及非貧家戶的影響不見得相同，可能造成政策成效不彰（Shamima and Ahmad，2007）。基於此目的，本研究以官方貧窮、絕對貧窮及相對貧窮設定貧窮門檻，分別區分出雙重貧窮、僅消費低及僅所得低三種經濟弱勢家戶；並以主計處12年（2009-2020）之家庭收支調查，方法上使用階層線性模型（hierarchical linear modeling, HLM）進行分析，從而避免由資料巢狀的結構進而導致係數偏誤的情況，以此探討貧窮家戶在主食、副食、蛋乳製品、水果、糖類、飲料、菸酒及外食消費比例上與非貧家戶之間的差異。據HLM結果顯示，不同貧窮類型家戶在飲食總消費支出或飲食品項消費比例上與非貧家戶相較皆存有顯著差異，且相同類型家戶的消費亦存在時間及地域差異；四類家戶的消費變動在時間趨勢上相同，但非貧家戶的消費變動則相較貧窮家戶小；無論以何種貧窮測量方法，三類貧戶的主食、副食、水果等支出彈性皆高於非貧家戶；而根據不同貧窮測量及均等比之結果，皆顯示貧窮家戶在菸酒消費比例上相較非貧家戶高。基於本實證結果，本研究建議在對飲食或扶貧相關福利政策制定上需考量貧窮與非貧家戶消費上的不同，及不同地域間各類型家戶消費的差異，以避免相關政策成效不彰之問題。

Poor households are constrained in their consumption due to a lack of economic resources, resulting in differences in their consumption patterns compared to non-poor households. The policy-makers often overlook these differences. If we assume that there are no differences in consumption between poor and non-poor households and formulate economic or welfare-related policies based on this assumption, the impact of the policies on the poor and non-poor households may not be the same, potentially leading to ineffective policy outcomes. Under this purpose, this study sets poverty thresholds based on official poverty, absolute poverty, and relative poverty, and differentiate three types of economically disadvantaged households (dual poverty, low consumption only, and low income only) from non-poor by using each poverty thresholds. The data we used in this study was provided by The Survey of Family Income and Expenditure of Taiwan from year 2009 to 2020, considering that data structure is hierarchical, hierarchical linear modeling (HLM) was used to analyze. The HLM results indicate significant differences in total food consumption expenditure and food share of food items between different types of poor households and non-poor households. Consumption patterns within the same type of household also exhibit temporal and regional variations. The consumption trends are similar among the four types of households, but the trend of non-poor is flatter than the other types of households. Regardless of the measurement method for poverty, the expenditure elasticities of staple food, non-staple food, fruits, and other items for the three types of poor households are higher than non-poor households. This study also show that poor households have a higher share of food expenditure on tobacco and alcohol compared to non-poor households. Based on our findings, we suggest that the differences in consumption patterns between poor and non-poor households should be considered before formulating welfare or economic-related policies, as well as the variations in consumption among different types of households across different regions.