資本適足率（capital adequacy ratio，即實業界所稱之BIS比率）為金融界評估商銀風險之重要指標，在反映資本結構以至於倒閉風險的意義上，相較於財務分析常用的權益值對總資產比率，BIS比率應是一個更精確的指標。本研究先藉資本市場銀行股長天期窗口超額報酬率反映投資人所要求報酬中之風險貼水，探討投資人是否可以引用資本適足率衡量國內商銀的倒閉風險與流動性風險。此外，本研究亦針對壞帳費用與票券買賣損益兩項富裁量空間之科目，分析國內商業銀行策略性操縱帳面盈餘與資本問題。最後，就現行我國資本適足率規定之缺失，作進一步之檢討，並檢測調整部份風險性資產之風險權數後，對資本適足率解釋投資人所要求必要報酬間關係之影響。

實證結果發現，不論是商銀呈報金融主管機關之資本適足率，或是就銀行所發布資料，儘可能比照公訂資本適足率核算辦法所自行設算、不含資產負債表外風險性資產所計算之比值，甚至自行設算、僅考慮自有資本中之第一類資本（Tier 1 Capital）估算值，均與商銀股市超額報酬有顯著之負血關係，顯示資本適足率對投資人而言，屬攸關資訊，能幫助評估銀行倒閉風險，進而決定其所要求之必要報酬率。又國內商銀中，民營銀行股超額報酬對資本適足率之迴歸係數，較公營銀行更具負向關係，而民國八十一年後新成立之銀行對資本適足率之迴歸係數，亦較八十一年前成立之舊銀行更具負向關係，而景氣較蕭條時，資本適足率與報酬間之關係並未較繁榮期敏感。

在盈餘與資本調控部份，或因使用不同調控工具之成本差異，致使商業銀行在帳面資本不足時，傾向于增加提列壞帳費用；另一方面，銀行似乎為了損益平穩化之目的，而以多實現或少實現票券買賣損益作為調控當期盈餘之工具，此兩項潛在之盈餘調控工具，彼此間有著相互替代代，惟因實現票券買賣損益之成本因時而異，國內商業銀行引用此兩項工具相互替補的程度實隨資本市場榮枯而改變。在估算國內商銀壞帳費用不可裁量部份時，本研究發現以上期壞帳、本期逾催收款、應收匯兌承兌款及無擔保放款餘額估計壞帳，比過動國外文獻所採變數組更恰當。

This thesis empirically examines the explanatory power of capital adequacy ratio (BIS ratio) to Taiwan's commercial bank long-windowed returns minus risk-free rates (hereafter excess return), investigating whether the ratio serves to measure the level of risk of these banks equity securities. Findings indicate the followings: (1) ceteris paribus, long-windowed bank returns negatively correlate with each and every measure of BIS ratio in this study. These results are consistent with the notion that capital adequacy ratio conveys relevant information regarding the bank shareholders risk; (2) required rate of security returns appears to be more (less) sensitive to the BIS ratio for banks founded after (prior to) 1992 and for non-state-owned (state-owned) commercial banks; (3) there is not corroborative evidence that macro-economic variables have incremental explanatory power to the regression coefficient for the BIS ratio.

Further, by identifying and examining the potential discretionary components of Taiwan's commercial bank loan loss provisions (LLPs) and securities gains and losses (RSGs), this study aims at exploring these banks' accruals management practices. Robust against various sensitivity tests, empirical findings support the notion that commercial banks strategically increase their LLPs to avoid unfavorable capital adequacy ratios. On the other hand, this study finds these banks smooth reported earnings via RSGs. Moreover, our evidence is consistent with the hypothesis that LLPs and RSGs serve as substitutes for each other in commercial bank accruals management. However, the extent these banks exercise discretion via either measure varies with domestic capital market performance. For tests in this study, the specification of simultaneous equations outperforms the competing ordinary least square regression models.

This study also provides an innovative design for estimating bank loan loss provisions. As compared with competing designs, our model, which relates commercial bank LLPs to non-performing assets, unsecured loans, accrued acceptances and prior-period loan loss provisions, produce a more efficient predictor for Taiwan's commercial bank LLPs.