This paper links two applications of Extreme Value Theory (EVT) to analyze Taiwanese financial markets: 1. computation of Value at Risk (VaR) and Expected Shortfall (ES) 2. estimates of cross-market dependence under extreme events. Daily data from the Taiwan Stock Exchange Capitalization Weight Stock Index (TAIEX) and the foreign exchange rate, USD/NTD, are employed to analyze the behavior of each return and the dependence structure between the foreign exchange market and the equity market. In the univariate case, when computing risk measures, EVT provides us a more accurate way to estimate VaR. In bivariate case, when measuring extremal dependence, the results of whole period data show the extremal dependence between two markets is asymptotically independent, and the analyses of subperiods illustrate that the relation is slightly dependent in specific periods. Therefore, there is no significant evidence that extreme events appeared in one market (the equity market or the foreign exchange market) will affect another in Taiwan.

總體政策的有效性與國外干擾的隔絕，乃是國際金融理論的重要研究課題。以往的研 究，大多為固定匯率與浮動匯率的比較，本文則引進Marion的國民對國外債券的預期 報酬率，探討在資本完全移動與金融資產完全替代之下，雙元匯率的國外干擾隔絕與 總體政策的相對有效性。 第一章為緒論，說明研究動機與研究方法；第二章則建立本文的準均衡模型，並探討 總體政策的有效性及國外干擾；第三章建立本文的充分均衡模型，並探討其總體政策 有效性及國外干擾；第四章為結論，將本文的重要結論做一個摘要性的說明。

本文旨在探討隨機模型（stochastic model）下，最適干預政策的選擇問題。此一問題最旱由Poole在一有七Ｏ年的文章中 提出，但他所考慮的僅是封閉經濟下，最適貨幣政策工具的選擇。其後學者們乃陸續 以其文為藍本，君入供給面，考慮理性預期與資本移動等因素，並將模型擴充為開放 經濟模型。 因此本文擬分三大部分，第一部份是文獻的探討，將介紹隨機模型下有關理論的演進 。 第二部分， 較具代表性的模型，比較在不同的干預政策下，其結困的差異。 第三部分，分析不同目標函數下，最適干預政策間的衝突性，並兼論單一目標與多目 標函數對模型結論的影響。 最後一章則是結論。

本文從貝氏統計理論以及時間數列分析法探討動態線性理性預期模型。並用之以國內 的消費、所得以及其他經濟變動數的理性預期因素。 第一章：緒言：研究動機、研究目的、研究方法及資料來源。 第二章：理性預期之假設的含意及其應用。 第三章：數種理性預期模型之建立。 第四章：理性預期模型的統計特性：認定、估定及假設檢定。 第五章：理性預期與移動平均的預測比較。 第六章：理性預期之實證研究─台灣之經濟模型。 第七章：結論與建議。

本研究目的在整合依附理論、負向情緒調適預期和壓力因應三種理論觀點，建構面臨生涯抉擇的大四學生之情緒經驗歷程模式。研究採問卷調查法，以台灣八所大學783位大四學生為對象。研究工具包括依附風格量表、負向情緒調適預期量表、生涯抉擇壓力因應量表、快樂感受量表和憂鬱量表。資料分析方法為因素分析、信度分析、皮爾遜積差相關分析、t考驗、單因子變異數分析及結構方程模式。 初步模式驗證發現，初始模式與觀察資料間無法適配，故根據修正指標，刪除不適當路徑、整併觀察變項、將部分觀察變項的殘差連結後，產生三個理論模式。進一步進行模式適配度考驗，顯示三個模式與觀察資料間適配度良好，並能有效解釋變項間的關係。模式一：依附會影響負向情緒調適預期；依附會透過正向壓力因應影響情緒經驗。模式二：依附會影響負向情緒調適預期；依附會透過負向壓力因應影響情緒經驗。模式三：依附會透過情緒焦點因應影響情緒經驗；依附會透過負向情緒調適預期和情緒焦點因應影響情緒經驗。 三個模式的共通處為：依附會影響個人的壓力因應；個人的壓力因應會影響其情緒經驗；依附並不會直接影響個人的情緒經驗，但依附會透過壓力因應而對情緒經驗產生影響；依附會影響個人的負向情緒調適預期。 研究建議為：宜積極發展生涯輔導方案與生涯探索課程；協助個案瞭解依附、負向情緒調適預期和壓力因應對其情緒經驗的影響；針對不安全依附者提供預防介入；以不同性別和不同依附型態者為對象，進行模式的建構與驗證；採焦點團體法或縱貫研究法進行研究；將其他重要變項納入模式中，或探究變項間的關係；依附和情緒經驗等測量工具的改進。 / The purpose of the study was to integrate attachment theory, coping theory, and the perspective of expectancies for negative mood regulation (NMR) to develop the process model of emotional experiences for senior college students encountering the stress of career choice. The participants of the study were 783 senior college students in Taiwan. The participants were evaluated by Attachment Style Scale, NMR Scale, Coping Scale, Happiness Scale, and Depression Scale. The data were analyzed by factor analysis, Cronbach α analysis, Pearson product-moment correlation, t-test, one-way ANOVA, and SEM. In the primitive model testing, the original model couldn’t fit with the observed data. Adjustment is thus made, in accordance with the modification index, to delete the unsuitable paths, combine the observed variances and line the error variances of some observed variances. Three models are generated. As shown in the result of model testing, the three models coordinate with the observed data, explaining the relationships among main variables. Model one: Attachment can affect NMR; attachment can affect emotional experiences through positive coping. Model two: Attachment can affect NMR; attachment can affect emotional experiences through negative coping. Model three: Attachment can affect emotional experiences through emotion-focused coping; attachment can affect emotional experiences through NMR and emotion-focused coping. The similarities in the three models were as follows: Firstly, attachment can affect coping. Secondly, coping can affect emotion. Thirdly, attachment can’t affect emotional experiences directly, but attachment can affect emotional experiences through coping. Fourthly, attachment can affect NMR. The suggestions were as follows: 1.The programs of career guidance and the curriculum of career exploration should be developed. 2. The clients should be assisted to understand the influence of attachment, NMR, and coping on their emotional experiences. 3. The preventive intervention should be provided to students of insecure attachment. 4. The models should be developed and tested according to the participants of different gender and attachment styles. 5. The focus group interview and the longitudinal method should be adopted. 6. Other important variables should be added to the model; alternatively, the relationships of these variables be explored. 7. The instructions of measurement in attachment and emotional experiences should be improved.

本論文主要是應用過去探討盈餘反應係數（ERC）的研究模型，驗證公司與分析師預測值間正負預測誤差的方向性與幅度對於市場反應之影響；並且在公司持續一致存在正/負向誤差的情況之下，探討市場對於此種趨勢的反應，並進一步以實證結果分析影響經理人員進行盈餘管理的可能原因。主要的研究問題有三： (1)當公司超出分析師預測時，市場獎勵公司的程度，是否更甚於當公司未達分析師預測時，市場懲罰公司的程度? (2)在控制未預期盈餘（unexpected earnings）的幅度之下，當公司超出分析師預測時，市場獎勵公司的程度，是否更甚於當公司未達分析師預測時，市場懲罰公司的程度? (3)當公司存在持續一致報導正/負向預測誤差趨勢的情況下，市場對於公司的此種趨勢給與的獎勵或懲罰的性質及程度為何? 本研究係以在台灣證券市場上市公司為研究對象，研究期間為民國77年至民國90年共計14年。研究方法分別使用Pooled regression與Annual regression兩種迴歸模式分別對樣本觀察值進行複迴歸。共計三個迴歸模型，六個迴歸結果。 本研究結果如下： 一、在是否超過分析師預測方面，當公司超過分析師預測時所得到的市場獎勵程度，並不一定大於當公司未能達到分析師預測時所得到市場懲罰的程度。因此就此部分之結論而言，公司為了得到市場獎勵而從事盈餘管理與公司為了逃避市場懲罰因素而從事盈餘管理，兩種行為之動機程度並無明顯差別。 二、倘若在控制未預期盈餘的幅度的情況之下，當公司超過分析師預測時所得到的市場獎勵程度，則是顯著的大於當公司未能達到分析師預測時所得到市場懲罰的程度。因此就本研究此部分之結論而言，公司為了得到市場獎勵的盈餘管理動機程度甚於避免受到市場懲罰之盈餘管理動機程度。 三、在是否存在持續一致報導正/負向預測誤差趨勢方面，可區分為兩部分說明： （1）對於有持續一致未達分析師預測趨勢的公司而言： 持續一致未達的未預期盈餘部分（系統因子），市場會對該系統化部分之行為失去信心而給予其懲罰。 然而，非一致性未達的未預期盈餘部分（非系統因子），相較於其他沒有持續一致未達分析師預測趨勢之公司，市場則不一定會對該部分之行為失去信心而給予其超額懲罰。意即因非系統未達之部分所得到的市場懲罰，並不一定大於沒有持續一致未達分析師預測趨勢之公司。 （2）對於有持續一致超出分析師預測之趨勢的公司而言： 持續一致超出的未預期盈餘部分（系統因子），市場不一定會將該系統化部分之行為視為經理人員從事盈餘管理之結果而給予懲罰。 然而，非一致性超出的未預期盈餘部分（非系統因子），相較於其他未能持續一致超出分析師預測之公司，市場則亦不一定予以實質肯定並給予其超額獎勵。意即因非系統超出之部分所得到的市場獎勵，並不一定大於未能持續一致超出分析師預測的公司。 / This study investigates whether the market rewards（penalizes）firms for beating（missing）analysts’ earnings forecasts. Specifically, we examine the market response to positive and negative forecast errors. In addition, we examine whether the sensitivity of stock prices to positive or negative forecast errors is affected by the firms’ history of consistently beating or missing analysts’ forecasts. In this paper, we provide empirical evidence on three issues: (1) is there a differential market response to the level of unexpected earnings for firms that beat analysts’ forecasts versus those that do not? (2) does the market respond to whether firms beat or miss analysts’ forecasts independent of the magnitude of the forecast error? (3) whether the market gives rewards or penalties for firms with a historical tendency to report positive / negative forecast errors? The results indicate that the earnings multiple applied to positive unexpected earnings is not significantly greater than for negative unexpected earnings. However, we find that after controlling for the magnitude of the forecast error, the market penalty for missing forecasts is significantly greater in absolute terms than the response to beating forecasts. Finally we find that for the firms that consistently beat analysts’ forecasts, the market doesn’t give penalties to the systematic component of the forecast error, and doesn’t give excess rewards to the unsystematic component. But for the firms that consistently miss analysts’ forecasts, the market gives penalties to the systematic component of the forecast error, and doesn’t give excess penalties to the unsystematic component.

本研究探討費率自由化政策與台灣產物保險市場經營效益，利用實證資料評估產物保險公司之整體經營績效及核保利潤。本研究自保險年鑑、台灣經濟新報資料庫及保險市場重要指標，選取1998年至2006年間財務及業務資訊來進行實證分析。 依Klein (1999)分類，台灣產物保險市場屬於類似獨占性競爭市場，產物保險公司高度競爭。依實證結果顯示，實施費率自由化政策，公司自留費用率持續緩慢增加，自留綜合成本率亦呈現上升趨勢。多數產物保險公司之實際核保利潤仍大於依Fairley (1979)計算之均衡核保利潤。 實證結果摘要如下： (1) 類似獨占性競爭市場：赫芬多指數皆小於0.1，市場呈現競爭狀態。1996年後，火險及車險之簽單保費成長率似乎每四、五年會呈現負成長。另外，1998-2003年費用率與公司規模呈現顯著負相關(p=0.01)。2004-2006年，此負相關並不顯著(p=0.1)，顯示大公司漸不具有成本優勢，即產險市場之進入障礙有減少之趨勢。 (2) 成本費用支出增加：除2001年外，1998至2006年之產物保險公司自留綜合成本率介於91.09% 和 93.49%。2006年之自留綜合成本率為93.49%。2006年之產物保險公司費用率上升至40.51% 且損失率下降至52.97%。 (3) 核保利潤呈現正值：依Fairley(1979)計算之2006年預期均衡核保利潤。20家產物保險公司，17家實際核保利潤大於預期均衡核保利潤。 / This paper provides an empirical study of rate deregulation plan and profit performance in Taiwan property-liability insurance market. The data used in this study are from Insurance Year Book, Taiwan Economic Journal Data Bank, and Important Indexes of Insurance Industry (Taiwan) during the period from 1998 to 2006. Based on the classification by Klein (1999), the market structure of Taiwan property-liability insurance industry is similar to monopolistic competitive market and the property-liability insurers are engaged in intense competition. The results of this study show that the expense ratios of insurers rise slightly from year to year and the combined ratios also follow a trend of increase. The actual underwriting profits of most property-liability insurers are larger than the expected numbers estimated by using the methodology in Fairley (1979). The empirical results are as follows: (1) The market structure is similar to monopolistic competitive market for that the Herfindahl indexes are all below 0.1. The growth ratios of written premium on fire insurance and automobile insurance seem to become negative every four or five years after 1996. Besides, the negative correlation between expense ratio and the scale of economics in insurance market from 1998 to 2003 was significant at the p = 0.01 level. However, this correlation from 2004 to 2006 was not significant at the p = 0.1 level. It seems that larger insurers do not have significant cost advantages over smaller insurers in the recent years, namely the entry barriers decline. (2) The expenses and costs keep increasing. Except 2001, the combined ratios ranged from 91.09% in 1998 to 93.49% in 2006. The expense ratio increased to 40.51% while the net loss ratio decreased to 52.97% in 2006. (3) The underwriting profits of most insurers are positive. The expected underwriting profits by using the methodology in Fairley (1979) is less than the actual ones in 2006. Among the 20 property-liability insurers, the actual underwriting profits of 17 insurers were larger than their expected underwriting profits.

非語言敏感度（nonverbal sensitivity）指正確察覺與解釋他人非語言訊息的能力。過去研究發現女性的非語言敏感度優於男性，可以權力取向或社會角色理論解釋。權力取向認為女性在社會上缺乏權力，故需展現良好的非語言敏感度保障自己權益。社會角色理論認為女性擔任，或被期待擔任照顧者，因此培養良好的非語言敏感度。本研究同時探討權力取向與社會角色理論，嘗試比較兩個理論對非語言敏感度性別差異的解釋。本實驗採用兩人互動，透過權力差異的操弄與不同的議題（經濟提供者議題或照顧者議題）討論，探討互動時的非語言敏感度。每位受試者還觀看四部影片，測量無動機影響的非語言敏感度。結果顯示權力差異可解釋兩人互動的非語言敏感度，且資料型態符合權力差異者的動機解釋。而觀看影片時，相較於探討經濟提供者議題受試者，探討照顧者議題受試者有較佳的非語言敏感度。此外我們亦發現性別、權力與社會角色三因子交互作用，非語言敏感度會受到實驗分派的社會角色是否符合生理性別的傳統作業影響，符合預期狀態理論的觀點。 / Nonverbal sensitivity refers to the ability to accurately detect and decode nonverbal cues communicated by others. Past research showed that women enjoy superior nonverbal sensitivity over men. Power perspective and social role theory each provides explanations for women's superior nonverbal sensitivity. Power perspective suggests that due to the lack of power in the society, women count on their nonverbal sensitivity to get what they want and for protection. Social role theory suggests that women are more likely to be in or are usually expected to be in caretaker roles, which require their attending to others' verbal or nonverbal cues. As a result, women develop superior nonverbal sensitivity compared to men. We designed an experimental study to test explanations offered by power perspective and social role theory. In the experiment, two individuals were told to collaborate on a task either required them to be in the provider role mode (focusing on earning money) or caretaker role mode (focusing on taking care of others). Power difference was manipulated by assigning one of the two individuals to be a leader and the other follower in this task. Interactions were recorded upon participants’ consent. After interaction, the participants were asked to interpret nonverbal cues in their interaction. They were also told to view four video clips, and then report and interpret nonverbal cues in each video clip. The results suggest that nonverbal sensitivity in interaction can be best accounted for by explanations consistent with the power perspective. The powerful was better at “B sees Self” nonverbal sensitivity, and the powerless was better at “B sees Me” nonverbal sensitivity, consistent with motivational explanations. In addition, we found a three-way interaction of participant sex, power, and social role, suggesting that women had the best “B sees Me” sensitivity when they had no power and were in the provider role. And men showed the best “B sees Self” sensitivity when they had power and were in the provider role. The three-way interaction finding is consistent with predictions made by expectation states theory. However, when viewing four video clips, participants in the caretaker role had better nonverbal sensitivity than those in the provider role. Motivation was offered to account for the differential effects of power and social roles. By analyzing clips of the dyadic interactions, we found that the powerful showed more leader behavior such as deciding discussion direction, and the powerless showed more follower behavior such as stammering. However, the results of the clips did not account for the aforementioned findings of nonverbal sensitivity.

本文以資料包絡分析法衡量2004年第一季至2008年第三季台灣商業銀行之效率與生產力，並以其代表銀行之經營績效。不同於過去文獻之處在於本文以前瞻之觀點（沈中華，2005）求出放款預期損失以衡量放款風險，作為放款過程中的非意欲產出。本文並利用Ray and Desli（1997）架構拆解Malmquist總要素生產力指數，探討銀行績效變動之來源，此外，亦將該拆解結果與Färe et al.（1994）拆解方式比較，發現在技術變動部分有明顯不同。本文並探討影響銀行效率之因素以及檢定不同類型銀行之生產力表現是否有顯著差異。 / The purpose of this paper is to examine the efficiency and productivity change of Taiwan’s banks over 2004Q1-2008Q3 using data envelopment analysis. Unlike the literature of the past, this paper uses the forward-looking viewpoint to derive expected losses from loans to measure the credit risk. In addition, the loan expected loss is thought of as an undesirable output in the estimation of efficiency and productivity. On the other hand, this paper uses a VRS frontier benchmark (Ray and Desli, 1997) to analyze the sources of productivity change. This paper also compares this result with that of Färe et al.（1994）decomposition, and finds that the estimated technical changes of two approaches are significantly different. Further, this paper analyzes the factors influencing banks’ efficiency, and investigates the productivity differences between different ownerships.

國外研究證實，由於基金績效具有持續性，則理性的投資人會以過去績效最為投資參考依據，將資金投入過去表現佳的基金，而此一投資決策應能持續創造超額報酬或風險溢酬，因此市場資金應會流向未來績效佳的基金（smart money effect)，此即為現金流量報酬預期效果且由於基金的現金流量變動代表投資人的投資決策變動，故現金流量報酬預期效果亦即為投資人對於股票型基金報酬的預測能力。 為瞭解台灣基金投資是否具有報酬預測能力（選對好基金，將資金由壞基金中抽離的決策），而此能力是否會因基金基金規模產生差異，且市場投資人可否根據此一公開資訊（上上期的現金流量）、累積資訊（累積前三期的淨現金流量）作為投資參考，並賺取超額報酬。本研究根據建構八組投資組合，包括三組不同基礎的現金流入（出）交易策略，比較各投資組合的報酬預期效果。 結果發現，台灣股票型基金投資人並不具備報酬預期能力，且常做出錯誤的決策，通常由好基金中籌離資金，喪失獲取較佳報酬的機會。而市場投資人無法藉由遞延一期的現金流量資訊獲取較佳超額報酬機會，但可藉由過去累積三期的現金流量資訊，將資金由淨現金流入金額大的基金中抽離，並投資於淨現金流出金額較大的基金，可因而獲取較佳的績效。另外，投資人對小型基金的報酬預期能力優於大型基金。