本研究在標準的核心－邊陲模型(core-periphery model)中加入居住成本進行分析討論，模型設定是在Pflüger and Südekum (2008)的基礎上，將quasi-linear效用函數改為Forslid and Ottaviano (2003)所使用的Cobb-Douglas函數，讓所得效果得以呈現在聚集經濟的機制中。本文關於社會最適的分析重點分為兩個部分。第一，探討「缺乏住宅部門考量」的Forslid and Ottaviano (2003)與本研究結果的差異之處；第二，分析本研究與Pflüger and Südekum (2008)雖然同樣加入住宅部門，但在使用不同的效用函數下所形成的結果差異。主要結果顯示：(1)高貿易自由度時，Forslid and Ottaviano (2003)的社會最適為聚集，本研究的社會最適則是分散；(2)不同於Pflüger and Südekum (2008)發現「過度聚集」（市場均衡相對於社會最適）為低貿易自由度時唯一的情況，本研究發現「聚集不足」亦可能發生在低住宅偏好以及低差異化產品替代彈性時，顯示出當模型忽略了所得效果，有可能誤判社會最適的空間配置。

This study discusses a standard core-periphery model by adding the cost of living. The model is based on Pflüger and Südekum (2008), but the quasi-linear utility function changes to the Cobb-Douglas function based upon Forslid and Ottaviano (2003). We present the income effect in a mechanism of an agglomeration economy and focus the analysis on two issues. First, what are the impacts of land rent on the spatial structure under different assumptions? Second, what are the differences between the optimal spatial allocation and the market equilibrium? We then compare our findings with the results of Pflüger and Südekum (2008), with the main conclusions as follows. (1) When trade freeness is high, the social optimum of Forslid and Ottaviano (2003) is agglomeration, but our study exhibits dispersion. (2) Unlike Pflüger and Südekum (2008) in which “over-agglomeration” (market equilibrium versus social optimum) is the only case during low trade freeness, our study finds that “under-agglomeration” may also occur in low housing preference and low elasticity of substitution for differential goods. Our results show that when the model ignores the income effect, it is possible to misjudge the social optimum of spatial configuration.