貧富差異是否能夠解釋空屋率？ 動態追蹤資料模型的實證分析Can Income Inequality Explain Housing Vacancy Rates? An Empirical Study using a Dynamic Panel Data Model

過去已有不少文獻研究臺灣高空屋率的成因，本研究推論貧富階層差異亦可能是造成高空屋率的原因之一。本研究使用「人口及住宅普查」之空閒住宅率、以及內政部不動產資訊平台的低度使用(用電)住宅比例來衡量各縣市之空屋率，建立縣市追蹤資料庫進行實證估計，估計方法為動態追蹤資料模型。估計結果顯示貧富差異的確是影響空屋率的因素之一，5等分位與10等分位所得差距倍數對於空屋率都有正向的顯著影響。其他變數，包括前期空屋率、住宅存量、房價成長率、所得成長率、自有住宅比率、失業率、農林漁牧人口比例對於空屋率都有顯著的影響。
關鍵詞：空屋率、貧富差異、動態追蹤資料模型、臺灣

This paper examines the determinants of housing vacancy rates in Taiwan for the past two decades, paying special attention to the role of income inequality. The data are drawn from the Population and Housing Census, the Real Estate Information Platform, the Statistical Books of Assessed Individual Income Tax Filing Data, the Survey of Family Income and Expenditure, and the Indicators of Municipalities and Counties (Cities) in Taiwan. The empirical methodology is a Dynamic Panel Data Model. The estimation results show that income inequality was indeed associated with higher housing vacancy rates, other factors held constant. Other variables, including the vacancy rates from the previous year, housing stock per capita, the growth rate of housing price, the growth rate of average annual disposable income, unemployment rate, the rate of home ownership, the percentage of employed persons in the agriculture, forestry, fishing, and animal husbandry sector were also significantly associated with the housing vacancy rates.
Key words: vacancy rate, income inequality, dynamic panel data model, Taiwan