| 摘要: | 論文提要 本論文的主要目的，在於建構地區性房屋銷售價格指數，作為地區性房屋市場業者比較之用。經由桃園與新竹地區的房地產公司的二手房屋交易資料，本論文應用特徵價格法(hedonic price approach)，採用基本及課題模型分析，進行房價指數編制，藉以分析自1995年至2005年間桃竹地區的房價變動及發展趨勢。 經由實證的模型分析結果，得到以下的結論： 一、基本模型的實證結果，對房價的形成已具有相當好的解釋能力；加入了更多虛擬變數後的課題模型，則取得了明顯的證據，證實或部分解釋了過去人們對房價漲跌變化的直觀想法多數為真。 二、利用Box-Cox轉換對房價所作的非線性函數關係之研究，確認在大多數情況下，房價模型的非線性關係顯著。 三、課題模型中若同時存在市鎮別(區位)及房屋種類變數時，房價函數可從非線性關係轉換為半對數線性關係，解決了直接採用半對數特徵房價模型之函數不確定性爭議的問題。 四、本研究之特徵房價指數與簡單移動平均法所建立的房價指數進行比較分析，結果顯示前者在年度曲線的完整性及對房價漲跌的解釋均較後者有合理性，故在解釋房價變動趨勢方面，採用特徵價格法應較優於簡單移動平均法。 五、與市場上業者定期公佈的房價指數比較，發現在以2000年為基期時的房價指數圖，兩者顯示幾乎重合的情形，而若將基期移至1995年比較，則可以看出本研究桃竹地區房價指數的各年度跌幅，明顯與業者所公佈較大範圍的房價指數為大。 在研究資源相對不足之下，一般房地產業者或可利用本研究所建立的房價指數模型，對個別小範圍市場進行類似的市場分析，並應用於房屋鑑價方面，對瞭解個別市場的變化應有相當的幫助。 Summary The purpose of this thesis is to construct housing price index used for comparing the local housing market around Taoyuan and Hsinchu area. By the way of using transaction materials from some realty companies in the area, the theory of hedonic price approach is applied to propose a basic model and a topical model, then to compile housing price indices for analyzing house price variation and its development tendency during 1995 to 2005. According to the analysis result of empirical models, conclusions were obtained as follow: First, the empirical results of the basic model show a robust explanation to housing price formation; After added more dummy variables to control the model, then it has obvious evidence confirmed on consumer’s behavior about housing price rise and drop changes. Second, using the Box-Cox transformation method to detect the non-linear relations between the price and characteristics, the results further confirmed that the majority of housing price models has the feature of non-linear relations. Third, if joined together with township (location) and house type variables to the topical model, housing price function may transform from non-linear to linear relations, which found a way for solving the functional uncertainty arguments when directly using semi-logarithm to housing price model. Fourth, housing price index of this thesis carried on the comparison with the simple moving average housing price index, the results indicated that the former is better than the later in rationality, not only in integrity of yearly curve but also in the explanation to house price variation. Then, the hedonic price approach might be superior to moving average method in the explanation of housing price changing tendency. Fifth, comparison with a housing price index in market which announced in regular by a realtor company, discovered that, take 2000 as base period, housing price index chart presented both are almost coincident, and if moved base period to 1995, then the dropped rate of housing price index that Taoyuan and Hsinchu area was obviously more expanding than what they announced of larger area. Under the relative shortages of research resources, any moderate realty companies may use housing price index model that had constructed by this thesis, doing somewhat similar analysis to individual smaller market, and applying to housing price appraisals. It would be very helpful to understand the changes of the local house market. |
