在經歷一連串的金融風暴與投資市場開放下,相對的共同基金的發展也愈來愈成熟,投資人不僅可透過各種管道瞭解共同基金的資訊,亦可自由選擇基金手續費最優惠的銀行或基金公司來進行投資。 根據中華民國信託商業同業公會所公佈的資料中顯示,共同基金投資市場每年都有擴大的趨勢,在銀行財富管理中,共同基金已成為必要且重要的金融商品之一。因此,在如此大的競爭壓力下,銀行需要更瞭解客戶的需求,可透過分析結論,瞭解客戶特性,定期檢討改善行銷方向,達到銀行銷售收入增加之目標。 本研究利用分量迴歸方式來探討影響客戶貢獻度之變數,分量迴歸的優點不僅是不需假設母體分配,亦不受離群值的干擾,在實際應用方面,對於兩尾端之估計,分量迴歸模型較最小平方法更為準確。 同時,透過條件分量數方式,對其應變數「基金手續費」進行排列,不僅可深入瞭解有關「高、中、低等客戶貢獻度」其變數的變化,亦可藉由變數顯著性來制定未來行銷方向。
After experienced quite a long period of financial crisis and follow by investment market opening, the development of mutual fund is getting more and more mature. Not only that investors could capture mutual fund related information via various channels, but also selecting banks or fund houses with the best subscription fee discount for investment. From the statistic data published by Trust Association of R.O.C, mutual fund keeps increasing in the investment market. Moreover, mutual fund has become parts of significant and important investment product in the wealth management, for banks. Therefore, bank has to understand customers’ needs in such competitive investment market. It’s believed the sales volume and revenue will growth from the analysis of customers’ needs, realizing their characteristics and regularly review business campaign. This research is using Quantile Regression as method to investigate customers’ contribution for a bank based on variables. Population distribution is not required, which is one of advantage of Quantile Regression. Moreover, the method won’t be impacted interfered by outliers. Besides, from actual practice point of view, using Quantile Regression is much precious and accurate for two-tailed estimation, comparing to method of Ordinary Latest Square. Meanwhile, it can be further investigated the changes of variables from high, medium to low contribution clients using conditional quantile methodology sorting by mutual fund subscription fee respective dependent variables. Finally, it’s believed that mutual fund sales campaign and future marketing can be designed accordingly based on these variables with significances.