本研究基於資本效益之概念，探討金融控股公司合理之資本配置，並以2002年第4季至2005年第2季間富邦金融控股公司之財務資訊進行實證分析。首先，本研究應用Margrabe (1978)互換選擇權評價模式，計算金融控股公司經營所隱含違約風險之賣權價值。其次，利用Myers and Read (2001)邊際違約價值相等之原則，將資本分配至旗下各子公司。實證結果歸納如下： 1.各子公司所需配置之邊際資本大小依序為台北富邦銀行、富邦人壽、富邦證券、富邦產險。即邊際資本越大，子公司每承擔1單位負債需要配置更多之資本。 2.富邦產險、富邦人壽與富邦證券實際資本均遠高於理論配置之資本，台北富邦銀行實際資本則低於理論資本。在給定固定門檻利率下，各子公司資金成本大小依序為台北富邦銀行、富邦人壽、富邦證券、富邦產險。 本研究結果顯示，富邦產物保險公司以減資方式，將過剩之資本退還富邦金融控股母公司，與Myers and Read (2001)邊際違約價值相等之資本配置原則相符。

本研究引用Myers and Read(2001)邊際違約價值相等原則進行資本配置。以金融控股公司為對象，考慮市場資產及負債，將金融控股公司各子公司資產負債之加總。利用Margrabe(1978)交換選擇權評價模型計算違約價值，推導各子公司邊際違約價值公式，利用邊際違約價值相等計算資本配置。由Cummins and Phillips(2005)估計產險公司各個險種之資本成本，利用CAPM及Fama-French三因子模型估算beta值，估計充分資訊下各子公司beta值與資本成本，最終利用風險調整報酬(RAROC)比較配置資本報酬與資本成本，進行績效評估。

鑒於金融機構跨業經營乃係現況發展趨勢，而風險管理已成為金融機構業務管理之首務，本研究將探討國際金融機構風險管理的新規範－新版巴賽爾資本協定（Basel II），並蒐集民國八十三年六月至民國九十二年三月我國上市公司公開發行之財務資訊，分別以商業銀行、證券公司、人壽保險公司、產物保險公司及票券公司為代表，模擬為金融控股公司下之各個子公司，將結合營運性、風險性、及法令為考量之資本配置模型；進一步探討金融控股公司之風險管理與資本配置。 歸納模擬結果如下： 以營運性為考量並配合法令要求之資本配置，此配置模型係在設定盈餘目標下，追求風險極小值，或設定風險容忍水準，以追求盈餘之極大值。經過目標值的變動下，此最適化模型可得一效率前緣曲線。由此效率前緣圖可知，在盈餘維持在平均盈餘水準下，此模型可使風險值由原來的8,860佰萬元降至7,045佰萬元，其降幅為20%，RAROC由原來的0.77升至0.97，升幅為25%；若固定風險值在原來未分配前的平均水準，則盈餘由先前的7,681佰萬元提升至8,821佰萬元，其提升幅度為14%，RAROC由原來的0.77提升至0.90，提升幅度為16%，亦相當可觀。若將資本維持在歷史平均水準，則可使其盈餘達到7,305佰萬元，而風險值為6,446佰萬元，RAROC為1.00，升幅為29%。據此推論，依此配置模型分配結果，將可改善整體金融控股公司之經營績效。 綜合營運性、風險性並考量法令要求之資本分配模型，納入了風險限制條件，主要係考量高風險之業務，須有高資本以因應其非預期性損失，但同時為兼顧經營績效，必須在營運性與風險性間求得平衡點；實證結果發現，受到風險性限制條件的影響，使得此條效率前緣曲線均落在以營運性為考量資本配置模型之下方，這代表的是在此模型下，無法達到較高之盈餘，原因來自於高盈餘伴隨著高風險，但高風險在此配置模型中是不允許的。惟此模型依然有提升整體經營績效的功能。例如，將盈餘目標設為原來的7,681佰萬元，則風險值降為7,601佰萬元，降幅約14%，RAROC也提升至0.90，升幅為16%；若資本維持在平均水準177,185佰萬元，則盈餘可達到5,802佰萬元，風險值則為4,528佰萬元，RAROC為1.09，較原先高出41%。 / As cross business managing is the modern development trend and risk management has been the first task for the financial institutions, this study attempts to analyze the new standard of the international financial institutions’ risk management – new Basel II. The data concerning business operations, risks, regulations from June 1992 to March 2003 are collected for each group of commercial banks, security houses, life insurance companies, non-life insurance companies, bill finance companies to build a capital allocation model. The financial holding companies’ risk management and capital allocation is further discussed. The results of this study are summarized as follows: The capital allocation model considers business operations and regulations. This allocation model sets up profit target, seeks minimum risk or sets up level of risk tolerance to seek maximum profit. After the variable target, the suitable model can get a efficient frontier curve. From this curve we find out that the profit maintains under the average profit level. The model can make Value-at-Risk reduce from 8,860 million to 7,045 million, down 20%, RAROC rise from 0.77 to 0.97, up 25%; if the fixed Value-at-Risk is before distributing, the profit will rise from 7,681 mission to 8,821 million, up 14%, RAROC will rise from 0.77 to 0.90, up 16%, still outstanding. If the capital remains at historical average level, then the profit can reach 7,305 million, and the Value-at-Risk is 6,446 million, RAROC is 1.00, up 29%. According to the inference, the distributing result of the allocation model can improve the operation performance of the financial holding company. The capital allocation model synthesize operation, risk and consider legal requirement, bring into the restriction of risk is to consider high risky business should have high capital to deal with unexpected loss, but also to consider operation performance, need to seek balance between operation and risk; From the result of this study finds that under risk restriction, the efficient frontier curve is within the capital allocation model which considers operation, this means under the model, higher profit is hard to achieve, the reason is high profit accompanies high risk, and high risk is prohibited from the model. But this model still has the function to approve whole operation performance. For example, if the profit target is 7,681 million as original, the Value-at-Risk will reduce to 7,601 million, down 14%, RAROC will rise to 0.90, up 16%; if the capital remains at the average level’s 177,185 million, the profit can reach 5,802 million, the Value-at-Risk is 4,528 million and RAROC is 1.09, up 41%.