本研究之目的在探討我國上市公司其現金增資是否具有繳款行情之現象，再者，更進一步探討該繳款行情是否與公司財務風險特質有關。本研究以140家上市公司為研究對象，選擇民國八十七年及八十八年為研究期間，首先對現金增資公司在繳款截止日前後的股價效果進行實証，再者，將各公司之累積異常報酬與該公司之財務風險變數進行迴歸分析，彙總以下結論： 1、款截止日(t=0)之前的事件期(t=-30∼-12)內，其異常報酬多為正向的異常報酬，此實証結果大致與本研究預期結果相同，卻未達統計上之顯著水準。 2、在繳款截止日(t=0)之前幾天(t=-11∼0)，依本研究實証的結果發現其異常報酬(AR)多為負，且累積異常報酬(CAR)亦逐漸降低，此點與本研究之預期不同。 3、在繳款截止日(t=0)之後的事件期內，本研究之實証結果發現，在繳款截止日後的事件期內，其異常報酬多為負向。 1、公司現增規模：達到5%之顯著水準，亦即公司現金增資規模與累積異常報酬呈顯著正向關係。 2、負債比率：未達顯著水準，但負債比率與累積異常報酬呈反向之關係。 3、業外收支佔總稅前淨利之比：未達顯著水準，但業外收支佔總稅前淨利之比與累積異常報酬呈反向之關係。 4、現金增資差價比例：達1%顯著水準，表示現金增資差價比例與累積異常報酬呈顯著反向之關係。 5、市場多空頭：達5%顯著水準，表示市場在多頭時，其事件期的累積異常報酬會顯著高於市場在空頭時的累積異常報酬。

近年來有許多研究將機器學習應用於財務方面的股價走勢與風險預 測。透過分析股票價格、財報的文字資訊、財經新聞或者更即時的推 特推文,都有不同的應用方式可以做出一定程度的投資風險評估與股 價走勢預測。在這篇論文中,我們著重在財務報表中的文字資訊,並 利用文字資訊於財務風險評估的問題上。我們以財報中的文字資訊預 測上市公司的風險程度,在此論文中我們選用股價波動度作為衡量財 務風險的評量方法。在文字的處理上,我們首先利用財金領域的情緒 字典改善原有的文字模型,情緒分析的研究指出情緒字能更有效率地 反應文章中的意見或是對於事件的看法,因而能有效地降低文字資訊 的雜訊並且提升財報文字資訊預測時的準確率。其次,我們嘗試以權 重的方式將股價與投資報酬率等數值資訊帶入機器學習模型中,在學 習模型時我們根據公司財報中的數值資訊,給予不同公司財報中的文 字資訊權重,並且透過不同權重設定的支持向量機將財報中的文字資 訊結合。根據我們的實驗結果顯示,財務情緒字典能有效地代表財報 中的文字資訊,同時,財務情緒字與公司的風險高度相關。在財務情 緒字以權重的方式將股價與投資報酬率結合的實驗結果中,數值資訊 顯著地提升了風險預測的準確率。 / In recent years, there have been some studies using machine learning techniques to predict stock tendency and investment risks in finance. There have also been some applications that analyze the textual information in fi- nancial reports, financial news, or even twitters on social network to provide useful information for stock investors. In this paper, we focus on the problem that uses the textual information in financial reports and numerical informa- tion of companies to predict the financial risk. We use the textual information in financial report of companies to predict the financial risk in the following year. We utilize stock volatility to measure financial risk. In the first part of the thesis, we use a finance-specific sentiment lexicon to improve the pre- diction models that are trained only textual information of financial reports. Then we also provide a sentiment analysis to the results. In the second part of the thesis, we attempt to combine the textual information and the numeri- cal information, such as stock returns to further improve the performance of the prediction models. In specific, in the proposed approach each company instance associated with its financial textual information will be weighted by its stock returns by using the cost-sensitive learning techniques. Our experi- mental results show that, finance-specific sentiment lexicon models conduct comparable performance to those on the original texts, which confirms the importance of financial sentiment words on risk prediction. More impor- tantly, the learned models suggest strong correlations between financial sen- timent words and risk of companies. In addition, our cost-sensitive results significantly improve the cost-insensitive results. As a result, these findings identify the impact of sentiment words in financial reports, and the numerical information can be utilized as the cost weights of learning techniques.

本文主要探討股票選擇權的隱含波動率是否能夠有效反應公司的財務風險與違約風險，並使用Merton (1974)與Black and Scholes (1973)的選擇權評價模型推導出每日的負債權益比率，作為公司財務風險的代理變數；違約風險的代理變數則是使用Bandyopadhyay (2007)的風險中立違約機率與真實世界違約機率。首先，本文觀察到隱含波動率和股票報酬率之間的確存在負向關係，除此之外，也發現非系統隱含波動率與股票報酬率之間也有負向關係。進一步研究非系統隱含波動率是否能夠反應公司風險，結果顯示當公司的財務風險與違約風險增加時，非系統隱含波動率會上升。最後，本文比較非系統隱含波動率與GARCH模型的波動率對公司財務風險與違約風險的資訊內涵，並執行包圍檢定、工具變數兩階段迴歸分析與非包覆模型的檢定，發現非系統隱含波動率的資訊內涵無法包圍GARCH模型的波動率，但兩者的資訊內涵互相交集。

本論文的目標是利用文本情緒分析技巧，針對美國上市公司的財務報表進行以句子為單位的風險評估。過去的財報文本分析研究裡，大多關注於詞彙層面的風險偵測。然而財務文本中大多數的財務詞彙與前後文具有高度的語意相關性，僅靠閱讀單一詞彙可能無法完全理解其隱含的財務訊息。本文將研究層次由詞彙拉升至句子，根據基於嵌入概念的~fastText~與~Siamese CBOW~兩種句子向量表示法學習模型，利用基於嵌入概念模型中，使用目標詞與前後詞彙關聯性表示目標詞語意的特性，萃取出財報句子裡更深層的財務意涵，並學習出更適合用於財務文本分析的句向量表示法。實驗驗證部分，我們利用~10-K~財報資料與本文提出的財務標記資料集進行財務風險分類器學習，並以傳統詞袋模型（Bag-of-Word）作為基準，利用精確度（Accuracy）與準確度（Precision）等評估標準進行比較。結果證實基於嵌入概念模型的表示法在財務風險評估上比傳統詞袋模型有著更準確的預測表現。由於近年大數據時代的來臨，網路中的資訊量大幅成長，依賴少量人力在短期間內分析海量的財務資訊變得更加困難。因此如何協助專業人員進行有效率的財務判斷與決策，已成為一項重要的議題。為此，本文同時提出一個以句子為分析單位的財報風險語句偵測系統~RiskFinder~，依照~fastText~與~Siamese CBOW~兩種模型，經由~10-K~財務報表與人工標記資料集學習出適當的風險語句分類器後，對~1996~至~2013~年的美國上市公司財務報表進行財報句子的自動風險預測，讓財務專業人士能透過系統的協助，有效率地由大量財務文本中獲得有意義的財務資訊。此外，系統會依照公司的財報發布日期動態呈現股票交易資訊與後設資料，以利使用者依股價的時間走勢比較財務文字型與數值型資料的關係。 / The main purpose of this paper is to evaluate the risk of financial report of listed companies in sentence-level. Most of past sentiment analysis studies focused on word-level risk detection. However, most financial keywords are highly context-sensitive, which may likely yield biased results. Therefore, to advance the understanding of financial textual information, this thesis broadens the analysis from word-level to sentence level. We use two sentence-level models, fastText and Siamese-CBOW, to learn sentence embedding and attempt to facilitate the financial risk detection. In our experiment, we use the 10-K corpus and a financial sentiment dataset which were labeled by financial professionals to train our financial risk classifier. Moreover, we adopt the Bag-of-Word model as a baseline and use accuracy, precision, recall and F1-score to evaluate the performance of financial risk prediction. The experimental results show that the embedding models could lead better performance than the Bag-of-word model. In addition, this paper proposes a web-based financial risk detection system which is constructed based on fastText and Siamese CBOW model called RiskFinder. There are total 40,708 financial reports inside the system and each risk-related sentence is highlighted based on different sentence embedding models. Besides, our system also provides metadata and a visualization of financial time-series data for the corresponding company according to release day of financial report. This system considerably facilitates case studies in the field of finance and can be of great help in capturing valuable insight within large amounts of textual information.

壽險公司積極開發新商品以因應大量退休人口的需求，讓退休屋主得以所居住之房屋為抵押物，向金融機構貸款以獲得退休後之資金來源的反向房屋抵押貸款商品也應運而生。但這類的退休商品，除了讓壽險公司因人類平均壽命延長的現象而曝露在長壽風險的威脅下之外，其中所牽涉到之多樣的財務風險，也讓壽險公司在經營上面臨另外一個挑戰，但是反向房屋抵押貸款商品因其商品特性，似乎也可以提供壽險公司不同的風險分散的效果，有助於提升整體商品組合的避險效果。 本研究所提出之多因子免疫模型，可供壽險公司依照其所銷售之商品及所欲規避之風險，選擇一個最適的商品銷售數量，讓整個商品組合獲得最佳之避險效果。本研究透過多因子免疫模型進行數值分析，發現商品中加入反向房屋抵押貸款商品時，其避險效果明顯的優於未包含反向房屋抵押貸款之商品組合，顯見壽險公司發行反向房屋抵押貸款商品將有助於達到風險分散的效果，獲得更佳的避險成效。 關鍵字：長壽風險、財務風險、反向房屋抵押貸款、多因子免疫模型。 / Life insurance company try to meet the demand of the elder who has been retired by designing new products. The mortgage instruments to enable elderly homeowners to borrow by using the equity in their home as collateral, called “reverse mortgage”. With the launch this kind of product, life insurance company exposures in the threat of longevity and involves in others financial risks. However, the features of reverse mortgage may create the different effects of diversification for life insurance company to catch the better effects of hedging. We propose the Multi-Factors Immunization Approach to calculate the optimal product portfolio which attain the best hedging effects for life insurer by adjusting the number of units sold and recognizing the risks they want to hedge. We discover that the product portfolios which include reverse mortgage have the better hedging effects than these don’t include by numerical analysis. It is obviously that life insurer can acquire the effect of diversification and better hedging effects. Key words: Longevity risk, Financial risk, Reverse mortgage, Multi-factors immunization approach.

本研究以確定給付型退休基金為對象，建構廣義隨機評價模型，以衡量不確定情況下退休基金之財務風險。希望藉著模型建構的過程，適切地描述基金評價過程中所應考量的各項要素。 為了強調基金評價時同時考量內外部精算假設的重要性，本研究將模型分為存活函數、經濟函數和給付函數三部份討論；存活函數利用離散時間非同質性半馬可夫過程（Discrete Time Non-Homogeneous semi-Markov Process）描述成員狀態轉移的機率，把成員工作年資、年齡和及狀態納入評價過程，有別於傳統僅以年齡為假設基礎之精算方法；經濟函數則以隨機過程表達外部環境的不確定性，結合上述假設資訊預估未來給付後，成為半馬可夫隨機精算評價模型，此一般性的模型能推展至基金評價時所需的各項流程。因此，本研究將模型應用於我國公務人員退撫基金，針對公務人員退撫基金的給付特性發展財務評價公式，完整地描述基金精算成本計算、未來人力與現金流量結構模擬以及敏感度分析等過程。 最後，本研究撰寫公務人員精算評價資訊系統，具體化半馬可夫隨機精算評價模型，實證公務人員退撫基金財務評價公式。實證結果也顯示，不論基金的性質或外部經濟環境，都將影響退休基金財務評價結果，為基金評價時不可忽略的精算假設。 / This study focuses on constructing a generalized valuation model for the defined benefit pension schemes. Financial soundness and funding stability are critical issues in pension fund management. In this study, a realistic stochastic model is built to monitor the uncertainty factors in affecting the financial risk and cash flow dynamics along the decision process. In order to evaluate the importance of the interior and exterior actuarial assumptions in pension valuation. Detailed models in describing the turnover patterns, economic uncertainties and benefit structures are explored. Semi-Markov process proposed by Dominicis, Manca and Granata (1991) and Janssen and Manca (1997) is extended in structuring the transition pattern of the plan’s population and the economic based factors are generated through stochastic processes. Modifications according to classification and movements of the plan member and the plan’s turnover pattern are employed to improve its practical usefulness. Then the actuarial valuations, cash flow analyses and workforce projection are performed and investigated. We has explicitly formulated the plan’s realistic phenomenon and implemented the proposed mechanism into a risk management framework for pension finance. By using this realistic approach, the cost factors could be monitored throughout the valuation. Typically these analyses involve substantial assumptions. This article has outlined the procedure of building the proposed model. Finally, Taiwan Public Employees Retirement System is simplified to illustrate the proposed methodology in pension valuation. The results from this study show that the structure of the pension schemes and the assumed economic factors are the significant factors in pension valuation. It also indicated that the fund manager can evaluate these impacts through the proposed model.