本研究選以台灣某大T銀行為例，以該銀行之各分行資料為主，並選取以金融海嘯之後的數個月份月資料進行時間變化下之縱橫資料分析。運用Gonza′lez, Teräsvirta and van Dijk（2004, 2005）之縱橫平滑移轉迴歸模型，自非線性角度出發，深入進行銀行財富管理績效在分行存款業務量差異下之非線性影響探討 研究估計結果，手續費、存款量、員工數、來客數及股價指數資料均為I(0)變數資料。線性關係檢定則發現模型設定的非線性檢定結果顯著性不高，但仍可視存款量對手續費收入存在縱橫平滑移轉效果，而區間個數檢定得知，存款對手續費影響的模型設定為m=1且r=1之縱橫平滑移轉模型。 最後，於縱橫平滑移轉模型之參數估計與檢定中發現，當存款小於門檻值的下區間時，若要提高銀行財富管理的手續費收入，應增加存款額度與來客數，並於股票走勢「熊市」下跌時為之；然而，相反的，在存款量大於門檻值的上區間時，若想要提高銀行財富管理的手續費收入，則應降低員工數，增加來客數，並於股市「牛市」走勢為之。 本研究所得結果，應可作為日後銀行經營財富管理業務之參考依據。 Selecting one of the Taiwan large banks (T-Bank) as an example and utilizing the data from each of its branches, this study analyzes the nonlinear effect of the variation of the branch’s deposit on the performance of bank wealth management for the post-subprime crisis period. The methodology employed is the panel smooth transition regression (PSTR) model by Gonza′lez, Teräsvirta and van Dijk (2004, 2005). The empirical results first find that all the variables considered are shown to be I(0) stationary series from the panel unit root test. Even though the test of linearity in this study can not be significantly rejected, we keep doing our further estimation of the PST effect since there exist a threshold value of the deposit and the model specification is showed having one threshold value with two regimes as m=1 and r=1. Finally, the estimation and testing of our PSTR model shows that each branch of the T-bank should increase the amount of deposit and the number of customer and better to do so during the bearish market in order to increase the fee of the wealth management when the amount of deposit is in the lower regime. On the contrary, each of the branch should decrease the number of employee, increase the number of customer and better to do so during the bullish market in order to increase the fee of the wealth management when the amount of deposit is in the upper regime. The finding of this study can be utilized as a bench mark for the future operations of bank’s wealth management.
