住戶的遷移行為，其結果一方面調適住戶的住宅服務水準；另外一方面也藉以改善住宅福祉。拌隨著住戶的遷移所引發之換屋過程「空屋鍵」將住屋單元及住戶相互聯結，當此聯結規模愈大即表示空屋鍵愈長，住宅福祉的改善較多；反之則否。空屋鍵模式即建構一套計算新建住宅所誘發換屋的最大可能性模式，本文的目的即在建立空屋鍵結計量解析模式，進一步計算一新建住宅社區誘發的空屋鍵長，並且利用程序性選擇模式建立空屋鍵與相關屬性之間的因果關係，藉以衡量社區住宅福祉受相關屬性的衝擊邊際效果。本文利用空屋鍵模式，討論新逼住宅誘發之乘數效果，並且討論住戶換屋所誘發的住宅過濾過程的顯著程度，其次再以空屋鍵乘積甫算住戶平均遷移次數，作為住宅福祉計量指標，藉以解析住宅福祉於此過程中之分配結構，最後據此提出改善與促進住宅福祉的策略。在經驗實證結果方面顯示住宅過濾比預期較不明顯。過濾的住宅也多半是屋齡老舊而且鄰里環境與相關住宅條件也都不佳，住宅福祉的提升相當有限。利用過濾來改善中低收入住戶的住宅品質的策略，也確實缺乏效率及公平性。其他諸如租賃住宅的供給、中低收入戶住宅補貼等，對弱勢族群住宅品質的改善，的確有其存在之必要性。另一方面，實證結果也顯示高價位住宅仍比中低價位住宅有較顯著的過濾效果。在程序性選擇模式操作結果顯示，戶長年齡、住戶規模等屬性較為顯著。最後可以歸結出一個結論，增加住宅過濾效果之顯著性以提升住宅福祉必須以住宅產品特質為主要考量，突顯產品區隔性，並且適度增加中高收入住戶換屋誘因，以達到過濾的目的。

One of the most important reasons for the household- relocation is to adjust the level of housing service to match the utility changes on the one way as well as to improve the housing welfare on the other way. Vacancy chains incurring by the household relocation is the linkage between dwelling units and households. Theoretically, the larger the length of the vacancy chain is, the higher the level of the housing welfare improved. The meaning of the length of vacancy chain is the maximal opportunity of the household relocation incurring from a new construction in the market. The purpose of this paper is to establish the interpretation model to calculate the lengths of vacancy chain as a quantitative index of improved level of chain length and some selected attributes to estimate the marginal effects by Ordered Logit and Probit models. This paper uses a vacancy chain model to calculate the multiplier effects incurring from the new constructed residential community and to interpret the significance level of housing filtering. To define the product of the lengths of vacancy chain and the frequencies of household movement as the index of housing welfare improved. Finally, this paper interprets the structure of housing welfare distributed within the area and then submits the strategy. The result shows that the housing filtering process is not significant even if it does happen that the filtered housing will be either obsolescent or suffering from a lack of maintenance and a protracted decline in neighborhood or environmental condition. It is therefore inefficient to improve the housing quality of middle-low income households by accelerating housing filtering process. Other policies as adding rental housing and subsidies to low income- households are necessary to raise the level of housing welfare to weak- groups. The result also indicates that the filtering effect from the higher price- housing is more significant than the lower one. Based upon the ordered discrete choice model individual household’s attributes such as age of household-headship, household size, are quite significant. The conclusion is that for the purpose of accelerating filtering process, the key-factor of improving housing welfare will be setting on housing distinctness to increase the incentives for household relocation.