地方派系的過度介入、經營規模過小以及專業人才與制度的缺乏，使得基層金融機構弊端叢生，但由於過去的金融體系封閉穩定，並未有全面性的擠兌或恐慌(panics)現象產生。直至八十四年七月彰化四信爆發違法貸款事件，並進而引發存款人爭相前往擠兌，基層金融危機四伏的冰山一角才露了出來。 本文即運用存活模型(Survival Model)中的加速生命模型(accelerated life model,Cox and Oakes,1984)及相關文獻的回顧來探討基層金融機構擠兌現象發生的原因。本文之主要結論為：1.由於地方派系的介入及長期以往的經營不善，時間的累積不僅無法使基層金融機構得到學習曲線的效果，反而成為基層金融機構成長的阻力；2.以放款餘額表示的規模大小對信用合作社與農會信用部之存活期間有顯著的正向影響，以資產總額表示的規模大小僅對農會信用部之存活期間有顯著正向影響；3.逾放比率對基層金融機構之存活期間具有正向影響，其中對信用合作社存活期間之影響並滿足10%之顯著水準；4.淨值比率愈高，基層金融機構之存活期間愈長；5.流動性準備比率對信用合作社之存活期間具有顯著正向影響；6.存放比率對基層金融機構之存活期間並無顯著影響；7.存款戶的預期心理及擠兌現象的擴散蔓延(Park,1990)是造成基層金融機構發生擠兌的非數量化因素。

台灣的證券市場中，散戶投資人佔高達八成的比例，其在市場的參與程度影響證券市場的發展；本研究針對自民國七十九年市場崩盤後，散戶投資人繼續參與市場或離開市場的情況進行長達15年的追蹤分析。 研究發現民國79年2月12日的崩盤跌幅雖然曾高達83%，但是離開市場的比例並不大，高達九成的人在後續的年度仍然繼續進出市場，且其交易次數與平均交易金額都逐漸增強，顯示持續積極參與市場外，也是市場主要的穩定投資人；在分析中可以發現，不論在交易金額或交易次數的分析上，樣本群在市場崩盤後的各波段交易上，反應程度都比整體大盤投資人大。 對存活者、非存活者在崩盤前的交易次數、交易金額分析中，發現存活者為平均交易次數、金額較高者，但是變異數很大，且崩盤後後續交易記錄也顯示存活者群中，存在高交易次數、低單次交易金額與低交易次數、高單次金額兩類，與台灣證券市場散戶實證研究中，周轉率與報酬率呈U型分配，交易次數特別高及特別低的兩群樣本投資人，可能報酬較佳，所以得以存活於市場的情況相呼應。 / 80% investors in Taiwan Stock Market are individual investors. In that, their participations have great impacts to the stock market prosperity. This research focuses on the market crash which has happened since 1990 and tracks the individual investors’ trading activities of the following fifteen years. In research we found that, although the market index has decreased 83% since Feb.12, 1990, there are 90% individual investors remained trading in the market. Besides, we found the individual investors who keep trading (the survivor) increase both in their trading frequency and in trading amount which shows the survivors play an important role in the stock market during the following years. In further analysis, we also found the survivors had greater responses to the rises and falls of the stock market than total market individual investors’ average. When we analyze the survivors’ pre-crash characters, which showed survivors were those who had greater average trading frequency and average trading volume. Besides, there also existed big variances in the survivor group. Both evidences showed there existed survivors who were high trading frequency with low trading volume and those who were low trading frequency with high trading volume. These proved the prior research evidence of individual investors’ character that the return and trading frequency distribute as U shape in Taiwan stock market. Investors in highest trading frequency group and lowest trading frequency group could get better return to survive.

在論文中我們以不可重複發生的事件為討論的對象，介紹離散時間存活模式的架構，並利用離散時間存活模式與連續時間存活模式、離散時間存活模式與多期二元反應變數模式的關係，找出在某些條件之下，模式之間相對應的情形。另外，針對資料的異質性或相關性的問題，我們在模式中加入隨機效應，以捕捉個人的特質，並彌補解釋變數的不足。同時，我們將說明如何運用最大邊際概似估計法估計模式中的參數。 / This thesis is mainly discussing non-repeatable events, describing the structure of Discrete-Time Survival model, and at the same time discovering the corresponding phenomenon among models under certain conditions by taking the knowledge from the relationship between Discrete-Time Survival model and Continuous-Time Survival model, as well as the relationship between Discrete-Time Survival model and Multiple-Period Dichotomous Response model. Moreover, random effects are added to the models in order to capture the individual characteristics and make up for the shortage of explanation variables for the assessment of heterogeneity and correlation. This thesis also explains how to estimate the parameters in the models by taking the approach of Maximum Marginal Likelihood Estimation.

本文研究資料取自於中央銀行經濟研究處調查之全體民營企業資金流量統 計資料，樣本期間為民國 75-79年。研究方法主要採用描述性統計來報導 台灣雙元金融體系之特質以及廠商之基本營運特性。在廠商存活函數及成 長函數設定方面，則採用Evans (1987)之研究架構，且以 Heckman (1979)二階段估計方法來進行估計，以修正樣本選擇性偏誤，並對函數作 對數化展開二階，以解決函數非線性之問題，最後則以Bresch-Pagan之方 法檢定是否存在變異數異質性，若有異質變異數之現象產生，則以White (1980)之方法加以修正。實證結果發現，在存活函數之估計方面，規模越 大、沈沒成本佔總資產比率越高及資產使用效率越高之廠商越不容易退出 市場，借款比率則無明確之影響方向。在成長函數估計方面，規模越大之 廠商成長率越低，而且制度外借貸市場借款比率越高者成長率亦越低。

自全民健保開辦後，洗腎病人平均每年成長16％到20％，截至民國95年3月國內請領洗腎重大傷病者已達52,000名，發生率高居世界第一，盛行率為世界第二，可想而知洗腎費用相當驚人，目前透析治療一年花費高達250億元，占全民健保整體支出的6％以上，若再加上尿毒症相關併發症醫療支出則超過400億元，是健保單一治療最大耗費項目。 洗腎人口的增加 使得洗腎費用急遽躍升，基於財務規劃的考量，健保局屢次希望調降洗腎的支付點值，然而這個想法卻引起醫界強烈的反彈，強調相較於美國和日本的洗腎費用，台灣的透析給付已無調降空間，認為若強行調降血液透析給付恐會影響洗腎品質；相反的，健保局依據新加坡的資料及調查門診洗腎耗用資源相對值的結果，卻呈現臺灣洗腎費用過高的現象，醫界與健保局各有依據，雙方始終無法達成共識。 整個問題的核心在於了解影響透析治療的相關因素，究竟臺灣的洗腎品質如何？是否會因洗腎支付調整而有變動？不同醫療院所之間的洗腎品質是否存在差異性呢？由於臺灣過去的透析資料均掌握在醫界團體手中，社會大眾對於洗腎議題縱有諸多討論，因受資料的限制總是不得其門而入，然則自健保開辦後，所有的洗腎病患皆列入重大傷病紀錄，透過國衛院的健保資料庫理當能了解洗腎病患的就醫費用和存活變項，可惜的是透析治療資料繁雜，外界不容易進行分析，因此本文將健保資料庫加以彙整出完備的透析病患資訊。 由於洗腎病人多為慢性腎臟患者，需要長期治療，故本研究以洗腎三個月以上的血液透析病患為主要研究對象，同時為了排除不同洗腎院所的影響，本文進一步以未轉院的新進病人為分析基礎，利用危險率模型（Cox Proportional Hazard Model）探討病人特性（年齡別、性別、疾病別），以及醫院特性（醫院權屬別、層級別、地區別與洗腎治療床規模）對長期血液透析病患的存活率與死亡相對危險性之影響。 研究結果發現病人的特性指標扮演較重要的角色，年齡愈大、男性、僅罹患糖尿病，或同時合併糖尿病以及高血壓的多重診斷病人，其死亡的相對危險性較高；僅罹患高血壓的透析病人短期死亡風險不顯著，長期則有顯著差異。醫院特性方面，洗腎治療床規模愈大，病患存活率愈高，公立醫療院所的病患存活率最低，基層診所的病人存活率最高。不區分長短期效果時，私立的死亡風險高於公立，基層診所的相對危險性低於醫學中心；但長期而言，公立的相對危險性就高於私立，基層診所的死亡風險仍顯著低於醫學中心，和存活率分析結果一致。 探究公立醫療院所存活率較低的原因，本研究認為可能是源於公立醫療機構缺乏追求高績效的動機，致使經營績效疲弱，進而影響到血液透析病人的洗腎品質；本文從三方面探討醫學中心存活率較基層診所為低之因，一可能源於醫學中心的病人滿意度較低，二為醫學中心的糖尿病透析病患比例較基層診所為高，三則可能由於醫學中心的院內感染機率較大之故。 本文的研究貢獻主要有四點，一是國內對於洗腎品質的研究少有以存活率作為品質之指標者，同時相關研究多受限於樣本數不足的困境，而本文利用全民健保資料庫可進行大規模且全面性的研究；另者，本文改進過去對於洗腎病患存活情形的分析方法謬誤，以新進病人為分析基礎；三為連結醫院特性與存活率的關係，針對所有透析醫療院所探討供給者特性對於洗腎病人存活率之影響；最後，本文發現醫學中心和其他醫院權屬別相較並未有較高的存活率，是以可藉此教育民眾正確就醫觀念，以落實轉診制度，使醫學中心的功能得以真正發揮，醫療資源得以善加分配。

在過去文獻中多有討論遷徙行為者，但少有對家戶的居住時間進行分析，本研究首先檢視自有者與租賃者之居住時間，其次為瞭解家戶對實質住宅的經濟需求較大，還是對已建立的社會關係網絡需求較大，提出居住時間的時間相依性問題，即隨著居住時間的增加，家戶會因為對住宅需求改變而愈容易遷徙？還是在累積惰性下愈不易遷徙？最後則是探討家戶居住時間的影響因素，藉由不同的住宅與家戶特徵來瞭解家戶可能的居住時間，以作為住宅遷徙、流動性及其他住宅政策分析之參考。 本研究在實證方法上採取存活分析法，並利用華人家庭動態資料庫的橫斷面跨時資料進行研究。實證結果顯示，自有者的存活函數明顯大於租賃者，且自有家戶之居住時間為租賃家戶的3.11倍。另外，家戶遷徙機率具有正時間相依性，反應出家戶隨著居住時間的增加，遷徙機率會增加，然而自有家戶模型的遷徙機率卻呈現負時間相依性。居住時間影響因素實證結果，每人坪數較大，居住時間會較短，而年齡愈大與居住人數愈多者，居住時間會較長。 / Although there are lots of studies exploring migration, few of them focus on analysis household duration of residence. Duration of residence is an important source of information for a variety of private and pubic decisions and for the understanding of various social and economic phenomena. This study makes a residence duration comparison between owners and tenants. Due to cumulative stress and inertia, moving probability will be increased or decreased. We attempt to describe the duration of residence is positive or negative time dependent. This study also provides the evidence that house and household characteristics determine the duration of residence. This empirical study employs the survival analysis and uses the data from Panel Study of Family Dynamics. The result indicates that the survival function of owners is larger than of tenant. Duration of owner residence is 3.11 times longer than that of tenants. Generally, household moving probability is positive while the owner’s is negative. Research finding further reveals that ping per person is negatively related to residence duration. Age and household size are positively related to residence duration.

土地開發決策行為具備不可回復性及可延遲性等特色，於未來情況不確定市場下，土地開發決策包含了選擇權價值。根據Titman（1985）等人將實質選擇權理論應用於土地開發方面之研究指出，土地開發決策中所內含之實質選擇權，將影響土地所有人的開發決策。而根據實質選擇權理論，未來情況的不確定性乃是實質選擇權價值來源，隨著不確定性提高，將延後土地開發時機並增加未開發土地價值。過去二十餘年來，將實質選擇權應用於土地開發之相關理論研究已獲得豐碩成果，然而於臺灣地區，相關實證研究相當稀少，甚至若干分析結果呈現實質選擇權適用程度不顯著的情況。唯過去分析方式多採靜態之迴歸模型進行，且未將不同市場情況區域分開探討，因而導致估計結果可能的不準確，而曲解實質選擇權理論於臺灣地區之應用。針對上述情況，本研究採動態之存活分析方式針對同樣亦屬動態之土地開發決策行為進行分析；並利用迴歸分析方式觀察未來房價不確定性如何影響未開發土地價值。藉由上述分析，探究實質選擇權理論於臺灣地區土地開發市場之適用程度。 本研究利用臺灣地區大規模開發地區土地開發及未開發土地價格資料進行實證分析。結果顯示，於不確定性程度高且變化程度大之地區，房價不確定性的提高將延後土地開發時機並增加未開發土地價值，符合實質選擇權理論所預測。然於不確定性程度低且變化幅度小之地區，由於選擇權價值的淡化，導致其他因素影響了房價不確定性與土地開發時機及未開發土地價值間之關係，使得分析結果未如實質選擇權理論所預期。此外，本研究於土地開發時機方面採存活分析方式進行，並將不同市場情況之地區分開探討，而得出實質選擇權理論於臺灣地區適用之顯著成果，與過去之相關實證研究指出之不顯著情況不符。表示在進行實證研究時，必須選擇最適當之分析方式，並儘可能就不同市場情況加以探討，方能獲得最符合實際情況之分析結果。 / Recent theoretical research suggests that land development exercises a real option. Option pricing theory has been successfully applied in the valuation of real investments in the last two decades. According to Option pricing theory, the real option in undeveloped land will effect the development decision of land developer. Because the uncertainty of future condition is the source of real option, this research use the data set of land development and undeveloped land price in Taiwan area to test two predictions of real options with respect to land markets: greater price uncertainty should delay the timing of development and raise vacant land prices. Different from other empirical studies, this research uses the survival analysis method to test the prediction of real option theory. The result of our research is consistent with the prediction of the real option theory. It shows that greater price uncertainty lowers the likelihood of development and raises vacant land prices in the area of higher price uncertainty. This finding suggests that in the area of high uncertainty, we would expect land developer in the property market to be more prudent and decreasing the current investment activities compared with the area of a relatively stable market environment.

Who can survive longer and what factors could prolong the trading life of individual investors in the market? This is the questions we ask in the dissertation. Based on our knowledge, there is not any research about the issue of survival analysis on analyzing individual investors in stock market. The paper classifies three possibilities could affect the trading life of investors: personal characteristics, trading behavior, and market condition. In the dissertation, we use tick-by-tick transaction data from the Taiwan Stock Exchange to profile survivors versus non-survivors, to investigate how the traders’ characteristics (such as, gender), trading behaviors (such as the degree of diversification, trading amount and trading frequency) and market condition affect the trading life of investors. We borrow the proportional hazard models proposed by Cox (1972) who used in bio-statistics to analyze the survival rate. Using the Kaplan-Meier curves for male and female, we find that survival functions and hazard rates of female investors have better survival prognosis than the male investors. Different timing of entering results in distinct patterns of survival curves and hazard rates. Investors entering that market in the bull and bear market have a larger survival rate than those who enter the market in normal time during the trading life from 1 to 7 years. Moreover, as the trading life increases larger 7 year, the three curves of bull, bear and normal market conditions, respectively, appear to get closer, suggesting that if trading life is shorter than 7 years, the investors entering in the bull and bear markets seemly have lower hazard ratio than that in the normal market to leave the market. Finally, the results of Cox’s proportional hazard model show that female investors stay in the market 74 days longer than the male. Trading cycle increasing by one day will prolong the traders in the market by 4.8 days. Average volume per trade measured in ten thousands does not have economic effect on the trading duration, although its estimate is statistically significant. A one percentage increase of portfolio return will reduce about 151 days of the trading life. One more stock in the portfolio will prolong about 133 days in the trading life. The effect on the trading duration of trading performance of those who enter in the bull market is positive. / Who can survive longer and what factors could prolong the trading life of individual investors in the market? This is the questions we ask in the dissertation. Based on our knowledge, there is not any research about the issue of survival analysis on analyzing individual investors in stock market. The paper classifies three possibilities could affect the trading life of investors: personal characteristics, trading behavior, and market condition. In the dissertation, we use tick-by-tick transaction data from the Taiwan Stock Exchange to profile survivors versus non-survivors, to investigate how the traders’ characteristics (such as, gender), trading behaviors (such as the degree of diversification, trading amount and trading frequency) and market condition affect the trading life of investors. We borrow the proportional hazard models proposed by Cox (1972) who used in bio-statistics to analyze the survival rate. Using the Kaplan-Meier curves for male and female, we find that survival functions and hazard rates of female investors have better survival prognosis than the male investors. Different timing of entering results in distinct patterns of survival curves and hazard rates. Investors entering that market in the bull and bear market have a larger survival rate than those who enter the market in normal time during the trading life from 1 to 7 years. Moreover, as the trading life increases larger 7 year, the three curves of bull, bear and normal market conditions, respectively, appear to get closer, suggesting that if trading life is shorter than 7 years, the investors entering in the bull and bear markets seemly have lower hazard ratio than that in the normal market to leave the market. Finally, the results of Cox’s proportional hazard model show that female investors stay in the market 74 days longer than the male. Trading cycle increasing by one day will prolong the traders in the market by 4.8 days. Average volume per trade measured in ten thousands does not have economic effect on the trading duration, although its estimate is statistically significant. A one percentage increase of portfolio return will reduce about 151 days of the trading life. One more stock in the portfolio will prolong about 133 days in the trading life. The effect on the trading duration of trading performance of those who enter in the bull market is positive.

新竹科學園區對我國的社經發展有著卓越的貢獻，就2006年的統計數據來看，年營業額佔全國工業生產毛額的比率約35%，佔台灣GDP的比率則將近10%，佔全國出口貿易總值的8%，就業人數則約佔全國就業者的1.2%，其實質重要性可見一般，所以竹科廠商的存活是與全國、地方的社經發展緊密交織無可劃分，竹科廠商的退出可能會造成經濟成長停滯、工作機會短少、政策資源錯置等問題。 既有研究多半強調竹科成功的要素，但往往會使我們將目光只放在成功廠商，而忽視廠商退出竹科的現象，又或者未正視高科技產業的特質之一：高風險，因此本研究將試圖運用事件史分析來捕捉在竹科特殊的產業生態中，高科技產業其風險的真正面貌，並運用資源基礎、資源依賴以及組織生態學的觀點來探討竹科的不同廠商所涉風險之影響因素。 / The Hsinchu science park has made remarkable contributions to Taiwan’s development, so far as statistics in 2006, the annual sales volume takes up 35% of the national industrial production volume, nearly 10% of Taiwan’s GDP, it also accounts for 8% of the national export trade. The employment of the Hsinchu science park takes up 1.2% of the total employment of Taiwan. Its significant importance is easy to be understood fully at a glance, so the survival of firms in the Hsinchu science park interweaved with Taiwan’s society and development. Firm exit not only makes the working opportunity deficient, but also affect the usage of policy resources and the socio-economic development of Taiwan. Academic researches about the Hsinchu science park only emphasized the successful element, placed sight on the successful manufacturer, and ignored the phenomenon that some firms have exited from the Hsinchu science park. Actually, one characteristic of the Hi-Tech industry was high risk. So this research will attempt to use the event history analysis to realize firm survival in the Hsinchu science park, catch the real risk of Hi-Tech industry and understand the important factor that effect firm exit from the Hsinchu science park through resource-based theory, resource dependency theory, and organizational ecology perspective.

本研究利用DEA的效率衡量工具，分析球隊可能的解散因素，推論台灣職業棒球之市場機制。根據效率估計的結果約略可將台灣職業棒球二十年的發展分成三個時點，1990年至1997年—第一次簽賭爆發前：效率值呈現遞減的現象；1998年至2002年—兩聯盟共存期：多數的球隊呈現經營績效不佳；2003年至2009年—兩聯盟合併後時期：球隊普遍的經營績效均拉升。 Tobit迴歸分析結果顯示兩聯盟共存對經營績效有負向影響，二軍制度則對球隊的經營績效有正向影響，但令人訝異的是觀眾數對經營績效有負向的影響。存活分析結果推論觀眾數、球隊經營績效為影響球隊存續最重要的兩個因素。資料比對的結果符合上述Tobit迴歸及存活分析之結論，強化在DEA、Tobit、存活分析的推論結果。總而言之，台灣職業棒球市場經營機制尚未健全，本研究將其稱為「半人為市場機制」。台灣職棒必須進行改革，避免非市場因素介入市場運作，台灣棒球產業的未來才有發展性。 關鍵字：職業棒球、DEA效率分析、績效分析、Tobit迴歸分析、存活分析 / In this study, we use DEA efficiency measurement as a tool to analyze the survival of teams in the Taiwanese professional baseball market. According to the efficiency measurement results, one can divide the development of the Taiwanese professional baseball teams in the last twenty years into the following three periods: first, from 1990 to 1997—the outbreak of the first gambling scandal, overall efficiency performance was falling; second, from 1998 to 2002—when two professional leagues coexisted, overall efficiency performance was at the lowest; third, from 2003 to 2009—after two leagues were merged, significant improvement in terms of efficiency performance was observed. Tobit regression results suggest that the coexistence of two professional leagues has negative impact to the efficiency performance of teams while the introduction of the minor league system has positive impact to the efficiency performance. Surprisingly, the number of audience who attended baseball games has negative effect on the efficiency performance. The results from survival analysis identified the number of audience who attended baseball games and efficiency performance are the two main factors for the survival of a baseball team. A detailed data analysis confirms the Tobit regression and survival study results. In general, the market mechanism is yet well-developed. Non-market factors often affect the development of baseball teams and their survival. Drastic reforms and changes are required for the future development of the professional baseball market. Keywords: Professional Baseball, DEA Efficiency Analysis, Performance Analysis, Tobit Regression Analysis, Survival Analysis