臺灣博碩士論文加值系統

保險公司是否具備足夠的清償能力，一直都是監理機關、金融業者、投資者、
被保險人所關注的議題，在學術界研究此領域的學者也不少。本研究根據美國保
險監理官協會(NAIC)所提供的財務比率與 A.M. Best 信用評等機構的美國產險公 司失去清償能力且被接管的資料，建構預警模型。
本研究的研究對象是西元 1996-2003 年的美國產險公司，總共 15,508 筆資
料。應變數是產險公司在未來三年內是否失去清償能力並且已經被接管，自變數
總共有十二個。模型採用決策樹(Decision Tree)及羅吉斯迴歸(Logistic Regression) 分別建立產物保險公司失去清償能力的預警模型，透過混淆矩陣(Confusion Matrix)去評估模型的結果，並且找出各模型影響保險公司失去清償能力的變數。
在建構模型時，將資料切割成訓練資料(Training Data)與測試資料(Testing Data)，本研究將西元 1996-2000 年期間，作為訓練資料(Training Data)，西元 20012003 年期間，作為測試資料(Testing Data)，透過實證發現，因資料分類比例過於
懸殊，因此在配適模型中，發現預測結果不如預期。所以透過隨機抽樣的方式，
來解決資料分類不均勻的問題。本研究就將隨機抽樣五十組，每組失去清償能力
且被接管公司與正常公司比例為 1：1，將抽樣樣本五十組分別配適模型，得出五 十組所預測的結果，最終模型採用綜合五十組預測結果得出結論。

Whether an insurance company has enough solvency has always been an issue of concern to the insurance regulators, financial institutions, investors, and consumers. There are many scholars studying this field in academia. This study constructed an early warning model based on the financial ratio provided by NAIC and the bankruptcy data of US property and casualty insurance companies of the A.M. Best ratings.
The study is on the property and casualty insurance company in U.S.A over the period 1996 to 2003, with a total of 15,508 data. The dependent variable is whether the property and casualty insurance companies have insolvency in the next three years. There are twelve independent variables. The insolvency prediction models use Decision Tree and Logistic Regression. The Confusion Matrix is used to evaluate the results of the model and find out which models affect the insurance company.
When constructing the model, the data is separated into Training Data and Testing Data. This study will use the Training Data during the period 1996-2000, and the Testing Data during the period 2001-2003. Because the proportion of data classification is too unbalanced, the prediction results are not as expected. Therefore, the problem of uneven data classification is solved by means of random sampling. the 50 groups are randomly sampled. The ratio of bankrupt companies to normal companies in each group is 1:1. The sample models of the samples are sampled to obtain the predicted results, and the final model uses a comprehensive set of 50 prediction results.

第一章 緒論
第一節 研究動機及背景
第二節 研究目的及範圍
第三節 研究流程與架構
第二章 文獻探討
第一節 清償能力之探討
第二節 財務變數與失去清償能力之探討
第三節 建構模型之探討
第三章 研究方法
第一節 資料來源
第二節 變數選取
第三節 資料分析及處理
第四節 研究設計
第四章 實證結果
第一節 分析模型結果
第二節 分析財務變數及最終模型結果
第五章 結論
第一節 研究結論
第二節 研究限制
第三節 研究建議
參考文獻
附錄一 各變數的盒鬚圖
附錄二 變數型態特性是否影響模型

一、 中文部分
1. 王儷珊(2001)，我國產物保險公司清償能力的探討，國立中山大 學財務管理研究所碩士論文。
2. 王志宏(2014)，資料探勘方法探討保險業失卻清償能力之研究， 東吳大學財務工程與精算數學系碩士論文。
3. 林建智(1997)，「 論保險監理之目標」， 保險專刊，第五十期，頁 180-192。
4. 周品吟(2005)，保險公司清償能力風險管理之研究，朝陽科技大 學保險金融管理系碩士論文。
二、 英文部分
1. Ambrose, J. M., and J. A. Seward (1988), Best's Ratings, Financial Ratios and Prior Probabilities in Insolvency Prediction, Journal of Risk and Insurance, 51, No. 2, pp.229-244.
2. Ambrose, J. M., and A. M. Carroll (1994), Using Best's Ratings in Life Insurer Insolvency Prediction, Journal of Risk and Insurance, 61, No.2, pp.317-327.
3. BarNiv, R., and J. B. McDonald (1992), Identifying Financial Distress in the Insurance Industry: A Synthesis of Methodological and Empirical Issues, Journal of Risk and Insurance,59, No.4, pp.543-574.
4. Beaver, W. H. (1966), Financial ratios as predictors of failure, Journal of Accounting Research, 4, No.3, pp.71-111.
5. Brockett, P. L., W. W. Cooper, L. L. Golden, and U. Pitaktong (1994), A Neural Network Method for Obtaining an Early Warning of Insurer Insolvency, The Journal of Risk and Insurance, 61, No.3, pp.402-424.
6. Butsic, R.P. (1994), Solvency Measurement for Property-Liability Risk-Based Capital Applications, Journal of Risk and Insurance, 61, No.4, pp.656-690.
7. Cummins, J. D., S. E. Harrington, and R. Klein (1995), Insolvency experience, risk-based capital, and prompt corrective action in property-liability insurance, Journal of Banking & Finance, 19, pp.511-527.
8. Cummins, J. D., G. Martin, and R. D. Phillips (1999), Regulatory Solvency Prediction in Property-Liability Insurance: Risk-Based Capital, Audit Ratios, and Cash Flow Simulation, Journal of Risk and Insurance, 66, No. 3, pp.417-458.
9. Cummins, J. D., and D. W. Sommer (1996), Capital and risk in property-liability insurance markets, Journal of Banking & Finance, 20, No.6, pp.1069-1092.
10. Grace, M., S. Harrington, and R. Klein (1998), A Risk-Based Capital and Solvency Screening in Property-Liability Insurance: Hypotheses and Empirical Tests, Journal of Risk and Insurance, 65, No. 2, pp.213-243.
11. Grace, M., R. W. Klein, and R. D. Phillips (2003), Insurance Company Failures: Why Do They Cost So Much? The Wharton Financial Institutions Center, Georgia State University
12. Harrington, S. E., and J. M. Nelson (1986), A Regression-Based Methodology for Solvency Surveillance in the Property-Liability Insurance Industry, The Journal of Risk and Insurance, 53, No.4, pp.583-605.
13. Lee, S. H., and J. L. Urrutia (1996), Analysis and Prediction of Insolvency in the Property-Liability Insurance Industry: A Comparison of Logit and Hazard Models, Journal of Risk and Insurance, 63, No.1, pp.121-130.
14. Pottier, S. W., and D. W. Sommer (2002), The Effectiveness of Public and Private Sector Summary Risk Measures in Predicting Insurer Insolvencies, Journal of Financial Services Research, 21, pp.101116.
15. Pottier, S., and D. Sommer (2011), Empirical Evidence on the Value of Group-level Financial Information in Insurer Solvency Surveillance, Risk Management and Insurance Review, 14, No.1, pp.73-88.
16. Westgaard, S. and Wijst, N. (2001), Default probabilities in a corporate bank portfolio: a logistic model approach, European Journal of Operation Research, 135, No.2, pp.338–349.
17. Chawla, N. V., K. W. Bowyer, L. O. Hall, and W. P. Kegelmeyer (2002), SMOTE: Synthetic minority over-sampling technique, Journal of Artificial Intelligence Research ,16, pp. 321–357.