在房地產發展之相關研究中,交易價格一直是重要的研究議題,本文以台南市安南區民國90年至97年之住宅價格為實證資料,從空間層面去探討台南市安南區住宅交易價格之空間分布以及聚集地區,並且從時間層面來分析安南區住宅交易價格空間分布之變化,以了解安南區房地產市場歷年發展的重心與變化情形,以及與地區之發展概況。此外,藉由觀察新建住宅之空間分布與變化以了解房地產市場的發展重心。 本研究經實證後發現,台南市安南區民國90年至97年之住宅價格的空間自我相關程度會隨著設定界限範圍距離半徑之增加而遞減,因此愈近的住宅樣本間將存在愈高之相依性,而地區型空間自我相關測試結果及Moran’s I值顯示歷年安南區歷年住宅價格之間有較明顯的正相關趨勢,且其空間分佈上是呈現以生活機能較佳或交通便利之生活圈為主要聚集區域。 經比較特徵價格模型與空間自相關模型發現,空間自相關模型對特徵價格模型解釋能力之改善程度明顯提高,尤其又以空間誤差模型更適合用來推估安南區之住宅交易價格。 在安南區新建住宅價格方面,以交易價格在301~600萬區間,且所佔的土地面積以30坪較為購屋者所接受,且交易價格愈高之新建住宅較明顯集中於重劃區內,因此,重劃開發恰好提供一個很好的發展誘因,吸引新建住宅聚集,而藉由觀察新建住宅之空間分佈,亦可了解區域發展之脈絡。
The housing transaction price usually is one of the most important topics in housing research field. This study analyzes the housing transaction price in Annan district of Tainan City during 2001 to 2008 by using spatial techniques. The focus of the study is on the changes in spatial distribution of housing transaction prices as well as their spatial concentration. In addition, this study discusses the housing market development by exploring the spatial distribution and concentration of new housing transactions in Annan district. The empirical results show that the degree of spatial autocorrelation of housing transaction prices decreases with an increase in the bounded distance, which indicates that the closer of the distribution of housing transaction prices, the higher of their spatial dependences during these eight years. The tests of Moran’s Index show positive correlations among housing transaction prices, and the results of LISA also present that high-high housing prices are concentrated on areas with better living and traffic conditions. With a comparison of the hedonic price model and the spatial autoregression models, the results show that spatial autoregression models have better goodness-of-fit and also can significantly improve bias and inefficient estimations. The spatial error regression model has the highest goodness-of-fit in all these eight years. This indicates that the spatial error regression model is more suitable for estimation housing transaction prices, especially in new development areas like Annan district. Regarding new housing transaction prices, the results show that housing purchasers are more likely to accept the houses with price ranges from 3.01 to 6 million NTD and with 30 Ping of site areas. The new housing transaction prices were significantly concentrated on land readjustment areas where these areas offer sufficient public space and facilities and thus not only attracting housing developers but also housing purchasers.