本研究透過2010年人口及住宅普查資料探索臺灣358鄉鎮之空屋率現象,首先探討臺灣空屋產生的原因,並透過空間分析來了解空屋率的地區差異與其空間分布。迴歸分析中納入人口消長因素以及都市化程度等經社指標進行分析,並透過空間分析來探索空屋率是否存在鄰近效應;本研究另一目的為探討空屋與價格之間是否因地區差異存在空間異質,因此設置虛擬變數與交互作用進行探討。 研究結果發現除了生育率與農業人口比兩個變數與空屋率不顯著外,淨遷徙率、死亡率、都市化程度與老化指數皆與空屋率有正向顯著;而空間變數結果顯示空屋率確實受到鄰近空屋率影響而存在鄰近效應,但調整後R平方仍只有0.183。因此本研究加入設置之虛擬變數:投資炒作區與交互作用:價格乘以投資炒作區進行探討,結果實證出除了農業人口比不顯著外,淨遷徙率、死亡率、都市化程度與老化指數皆與空屋率為正向顯著;生育率與空屋率為負向顯著,符合預期假設;而地價與空屋之間在非投資炒作區為負向顯著,但若為投資炒作區則為正向顯著,其代表空屋與價格之間存在空間異質。
The purpose of this research focuses on Taiwanese 358 counties’ vacancy rate by spatial analysis. The research starts from the reasons which caused the vacancy rate. Furthermore, we puts population and urbanization factors into the regression and to examine whether it has neighbor-hood effect. Another purpose of this research is to explore the vacancy rate’s spatial heterogeneity by setting up dummy variable and interaction effect. The results have proven that there is positive effect in migration, mortality, density, old rate factors and vacancy rate have neighborhood effect. However, the adjusted R-squared is only 0.183. Therefore, the research puts dummy variable and interaction effect into the regression. The results show that there is positive effect in migration, mortality, density and old rate factors; negative effect in fertility rate. In addition, vacancy rate and land price vary with different areas. Thus, vacancy rate has spatial heterogeneity in this research.