本文為了建立房地產預警系統，針對1992年至2007年期間房地產巿場相關指標為警情指標選取之研究對象，利用層級分析法、因素分析法與擴散指數模型，選擇如同房地產景氣指標四個層面之重要代表性（非先行性）為主變數之警情指標，再綜合確定各指標之權重組成之房地產警情指標，不同於以往將選擇房地產景氣指標利用專家意見的定性方法，且其權重皆設定相同之方式。 警界檢查值是根據房地產警情指標預警值級別判斷房地產警情狀況，檢查值的計算運用變動百分比的方式，拔靴法，3σ法及理想目標值法，評估房地產警情指標之適度警界檢查值。 預測房地產巿場未來情況，採用指數平滑法、Box-Jenkins預測方法，以及灰色GM(1,1)Alpha模型，來預測房地產巿場警情指標及其單指標，並比較其預測績效方式選出較佳模型。據此便可對房地產巿場發生不正常現象的可能性和嚴重程度進行預測和監控。 研究發現，在民國95年第1季至民國96年第1季房地產警情狀況，可以看出其中的建築貸款餘額與全體金融機構放款總額之比值(S7)過高燈號為5分，顯示銀行對於房地產業貸款過熱，而住宅使用率(S19)有過低燈號為1分的不正常情況，並產生出過多的空屋的情形。接著預測民國96年第2季，所選擇的重要性指標來看，延續了前一季的狀況，更可看出其中的建築貸款餘額與全體金融機構放款總額之比值(S7) 過高燈號為5分，臺灣地區住宅建照樓地板面積與臺灣地區建照執照樓地板面積之比值(S8) 偏高燈號為4分，反應出供給多，五大銀行新承做購屋貸款金額與全體金融機構放款總額之比值(S16) 過高燈號為5分，需求過多，然而住宅使用率(S19)有過低燈號為1分的不正常情況，過多空屋，部分也可能含有餘屋的情形，此買了不住的情形，使住宅資源浪費，需加以警戒，並應適時提出修正。 關鍵字：房地產預警系統(Real Estate Early-Warning System) ，擴散指數模型(Diffusion Index Model)，層級分析法(Analytical Hierarchy Process)，因素分析法（Factor Analysis），檢查值(Check Point)，3σ法，拔靴法(Bootstrap)，指數平滑法（Exponential Smoothing）、Box-Jenkins預測方法，灰色GM(1,1)Alpha模型(Grey Forecast)。 / In order to establish real estate early-warning system, We select the main indicators from real estate market relative indexes during 1992 to 2007.Selectction are similar as the four layers of real estate cycle indicators that we have the important not proceeding indicators. The important indicators selection we applied the analytic hierarchy process, factor analysis and diffusion index model, forthmore, we integrated to determine the weights of the indicators and composed of main indicators. Unlike the past, the real estate cycle indicators choose the use of qualitative methods of expert advice, and set the same weight manner. Warning check point is based on the real estate early-warning value of indicators to determine the status of real estate market. Refered as the current cycle indicators, we have the check value calculated by using the range percentage, bootstrap method, and also developed the 3σ and ideal goal methods to determine an appropriate warning check point of the real estate early-warning value. To forecast the real estate market situation, we use black-box model of exponential smoothing, Box-Jenkins methods, as well as the gray GM (1,1) Alpha model to predict the real estate market indicators and their single indexs. Comparing and electing the better prediction performance model which can forecast and monitor the real estate market situation. The real estate market situation are estimate based on the empirical analysis during the first quarter of 2006 to the first quarter of 2007. The construction loans (S7) is too high for the signal is 5 points, which show overheated. The residential usage (S19) show too low for the signal is 1 point, which indicated non-normal conditions and produce too many vacant houses. Thereafter, we forecast the second quarter of 2007. We have continuously the situation of the previous quarter which as the construction loans (S7) is too high for the signal is 5 points. The residential building floor area (S8) is too high for the signal is 4 points, which was reacting too much supply. housing loan (S16) signal is too high for the signal is 5 points, which show too much demand. But The residential usage (S19) is too low for signal is 1 point, which show too many empty houses and some may contain redundant house.To buy and not to live, so that waste the resources of residential. We need to be alert and should be the right time to amend. Keywords: Real Estate Early-Warning System, Diffusion Index Model, Analytical Hierarchy Process, Factor Analysis, Check Point, 3σ, Bootstrap, Exponential Smoothing, Box-Jenkins, Grey forecast.

The study attempts to apply the DEFCON Concept as an early alert system to Non-Performing-Loan (NPL) Management in Taiwan’s Banking Industry. The recent financial crises in South East Asia have stimulated a significant body of empirical research on the subject of potential leading indicators for banking crises. Specifically, a number of statistical models have been developed to provide early warning signals of impending risks and also the relationship between the NPL and those leading indicators. The purpose of this study is to 1) Explain the definition of DEFCON and the application of DEFCON in the banking industry, 2) The literature review is on the correlation between Non-Performing Loans and Macroeconomic variables, giving particular importance to regression models. 3) The methodology of DEFCON Planning includes the data used, variables selection via the coefficient analysis, a simple regression model and usage of the selected variables to set the DEFCON triggers 4) Ultimately, to help aid in what Bank’s can undertake under different levels of DEFCON to prevent potential loss. Our empirical results show that 1) Economic Growth Rate 2) the Leading Index 3) Bounced Check Rate, 4) Shinyi Housing Index 5) Unemployment Rate 6) Consumer Price Index 7) Consumer Debt 8) M1B currency supply and 9) Unemployment Rate, are the leading indicators that predict Taiwan’s NPL ratio; however, it is prudent to note that the NPL ratio may be manipulated by banks, and may result to inaccurateness in some indictor’s prediction of the model. It is imperative that constant monitoring be the practice to ensure the effectiveness of the model. The Banks in Taiwan should monitor the overall DEFCON status periodically and use it as early alert system and take proactive actions based on the level of economic deterioration (DEFCON level) to well manage their asset and reduce NPL.

2007年8月美國爆發次貸危機(Subprime Crisis)，如此新型態的金融危機是否可由金融危機預警系統預測？是本文所欲探討的目標。本文採用訊號方法、固定效果下的Panel Logit Model和CART(Classification and Regression Tree)三種計量方法建構危機預警模型。最後利用美國2006年至2008年資料，驗證本文所建構之預警模型是否能夠有效預測次貸危機的發生。 / “Could banking early warning systems help to predict Sub-prime crisis?” That is the main issue that we want to discuss. We combine three kinds of early warning systems models – Signal Approach, fixed effect panel logit model, and CART approach – to create a new banking early warning system(EWS). We will use the US 2006-2008 data to examine whether this new EWS could predict the Sub-prime crisis correctly.

自1980年代各國推行金融自由化後，為穩定金融秩序，建立風險導向金融監理制度更顯重要。一般來說，金融監理工具可分為實地檢查及場外監控兩種，過去以行業別進行之監理，在金融控股公司的發展下，亦發展出對應之監控機制，然而僅止於實地檢查機制，以金融集團為預警對象之場外監控預警系統仍有待建立。 本研究遂在探討如何建立適合我國之以金融集團為預警對象的場外監控預警系統，挑選2003、2004年兩年之本國銀行、票券、證券、壽險、產險公司財務業務比率為樣本，以區別分析法建立預警模型，再以各金融控股公司之子產業公司結果建立各年度金融控股公司之預警模型。 本研究僅嘗試以財務比率建立量化場外監控預警模型，研究結果僅供學術上研究參考，並非運用於真實狀況之評斷，因此，依研究結果提出之結論及建議，僅供參考。此外，(1)模型並未加入質化指標，(2)資料有限的情況下，亦無做樣本外測試，(3)無實際破產金融機構資料，僅能以模擬方法分類，皆是本研究不足之處，仍須修正及改進。

國際研究中雖有針對國家級的銀行脆弱性作分析，卻並未定義或預測台灣系統性危機，本研究在這樣的背景下，決定建構台灣本土的銀行業預警系統，建立銀行危機的領先指標，希望不只順應國際潮流，更能發展適合台灣特殊性的模型。本研究利用貝氏網路模型的特殊性: (1)事後值(2)機率特性，以個體化資料著手，建構一總體性模型。故研究者能確切了解個別銀行財務狀況，對個別銀行發出預警。事後值的特性使研究者能同時考慮多項財務比率。另外，利用機率特性，可幫助研究者了解危機的程度，且能做總體的延伸運用。 本研究發展出兩種方法建構總體模型。第一種為百分比法，以危機銀行佔總銀行個數的比率為基礎；第二種為加權平均法，讓機率值高者有較大權數，機率小者有較小權數去建立一加權平均機率值。 將本研究的推論結果和「台灣金融服務業聯合總會委託計畫-台灣金融危機領先指標之研究」比較，顯示本模型的兩種方法皆與危機之發生有相同趨勢，而考慮危機訊號的設定後，方法二加權平均法顯然具備較佳的預測結果。此外相較總體面衝擊產生的危機，本模型在預測能力上，對來自銀行個體面造成的危機預測明顯較優異。 / International organizations defined and predicted country bank crises events without Taiwan, but they happened in Taiwan in the past twenty years. We construct the early warning system for banking crises in Taiwan and develop the specific model suited to our country. Using Bayesian Model’s specialities: (1) posterior value; (2) probability, we build a systematic model based on microeconomic data. So researcher can understand all financial conditions and predict the financial distresses of individual banks. The concept of posteriority lets researchers can consider a lot of financial ratio at the same time. The characteristic of probability makes researcher to extend the model to macroeconomic. We develop two methods to build systematic model. One is Percentage method which is based on the percentage of financial distress banks to all banks. The other one is weighted average method which used large weight in financial distress bank and small weight in financial sound banks. Comparing our results with the report that Taiwan Financial Services Roundtable issued in 2009, our methods have distress trends which link with crisis directly. But weighted average method has a better predict power than percentage method after considering the signals of distress we specify. Besides, our model has a stronger predictive power in crises from individual effect than crises from macroeconomic shocks.

根據世界銀行的資料顯示，佔全球面積約19%的2500萬平方公里之地球表面，及佔全球一半以上人口的34億人是相對的暴露在一個以上天然災害之威脅下。隨著全球化的來臨，我們居住在一個風險共享的社會中，而在全球環境安全被視為全球公共財的同時，如何在集體行動的邏輯下進行全球危機管理，已成為全球行動者的主要課題。例如如何透過國際合作來對抗SARS，禽流感等危機，皆是當前全球行動者關注的議題。值得注意的是，儘管近二十年來國際社會對於減災所做的承諾與投入的資源日益增加，災害所造成的經濟損失及受到災害影響之人口卻也逐漸上升。面對這些現象，本研究最主要想要探究的研究問題即在於什麼樣的因素影響著全球減災合作。 本研究主要的研究問題，係探求在全球行動者為何要參與減災合作，而此全球減災合作又如何運作的呢？全球減災合作、理性選擇與全球社會資本的分析架構將被運用。從理論上粹取的因素，例如風險意識、能力素養、偏好、制度限制、資訊、可信的承諾與信任等，被用來分析行動者如何決定參與合作，以及此合作如何運作。鑑於2004年印度洋海嘯所造成的重大傷亡以及後續國際社會對救災及減災的承諾，本研究將以印度洋海嘯預警系統的個案為例，並透過在4個國家共計22人次對參與此系統的國際行動者之訪談資料，以及對參與印度洋海嘯預警系統之人員發放共計591份問卷進行調查及分析，回收問卷目前共計61份，然進行論文分析時為59份。換言之，實際上的回收率為10.66%，而本研究用以分析之問卷回收率為10.32%。本研究最主要的發現為風險意識及能力素養的提升，結合理性選擇與社會資本的不同因素作用下，將對全球減災合作的結果有正面的影響。最後，本研究也對未來國際減災合作提出相關之建議。 / Writing on the issue of global environmental security, the World Bank has noted that approximately “25 million square kilometers (about 19 percent of the Earth’s land area) and 3.4 billion people (more than half of the world’s population) are relatively highly exposed to at least one hazard.” With the coming of the globalization era, we .also live in a shared risk society. Since global environmental security is seen as a global public good, how to act for global crisis management under the logic of collective action has become a primary subject for global actors. Coping with the crises of SARS or Bird Flu through international cooperation has become a significant issue for these global actors. One of the main dilemmas of international cooperation for disaster reduction is the reconciliation of different individual actions. Interestingly, in spite of two decades efforts of international cooperation, the amount of damage caused by natural disasters and the total number if people affected have gradually increased since the 1960s. This research focuses on two questions in the present research: why do global actors cooperate in disaster reduction, and how does this cooperation operate? The frameworks of international cooperation in disaster reduction, rational choice and global social capital are employed here, to explore the issue of international cooperation. Several factors, such as awareness of risk, capacity, preferences, institutional constraints, information, credible commitment, and trust, are used to examine how an actor engages in decision-making and how cooperation occurs. Because of the tremendous damage that resulted from the Indian Ocean tsunami of 2004 and the engagement of the global society in disaster recovery and reduction, the above issues will be explored through a case study of the development of the Indian Ocean Tsunami Warning System (IOTWS). Twenty-two interviews were conducted in four countries and these constitute the qualitative data for this analysis. 591 questionnaires also have been sent to the participants in the IOTWS to collect the quantitative data. I analyzed the quantitative data from 59 returned questionnaires (10.32% returning rate) and the qualitative data from 22 interviewees in four countries. These analyses resulted in several suggestions to facilitate international cooperation for disaster reduction.