本研究針對台灣2000年至2004年之全國家計單位調查資料，利用分量迴歸模型分析人際網絡密度對家計單位金融服務支出的影響。實證結果發現，即使在控制了家計單位的所得、平均年齡、風險態度、居住地、戶長性別、教育程度、職業別等變數下，人際網絡密度對家計單位的金融服務支出仍呈現顯著正向影響，表示人際網絡越密集的家計單位，金融服務支出越高，金融服務活動越積極。而且，分量迴歸分析的結果顯示，金融服務支出較高的家計單位，受人際網絡密度的影響比中、低家計單位更明顯，呈現出M型化社會的特質。

Following the rise of behavioral finance, interpersonal networks have gradually earned attentions in Finance. Using the dataset of 2000~2004 National Survey of Households' Income and Expenditures in Taiwan, we construct quantile regression models to examine the impacts of interpersonal networks on households' financial service expenditures. According to our empirical analysis results, the more intensive is the interpersonal network, the more is expended in financial services. Moreover, the quantile regression models estimated demonstrate that the households of high quantile are more impacted by interpersonal networks than the median and low quantiles, indicating the characteristics of M-type society.