隨著社群網路不斷快發展而使得上社群網站成為使用者每日生活的一部分，而社群商務也因此蓬勃發展。社群商務是社群網站使用者因社交活動而彼此交換分享產品或服務資訊的行為。根據Forrester Research在2011年的報告，預測至2015年時社群商務的產值將達300億美金。因此社群商務將隨社群網站不斷的發展，將越來越重要。 因此，本研究文獻回顧及焦點訪談藉由社群網站行為-參與及瀏覽二種行為、社會資本-結合型社會資本(親密朋友)及橋接型社會資本(普通朋友)來預測社群網站使用者採用社群商務-給予及接受二種意願。 本研究以社群網站臉書的使用者為研究對象，研究以紙本及線上二種方式來搜集使用者問卷。紙本問卷主要以北部某大學生為主要搜集對象，線上問卷則以在電腦教室的大學生、社群網站臉書的線上使用者及BBS-批踢踢實業坊的使用者同時也是社群網站臉書上的使用者為主。有效問卷970份。經由信度分析、效度分析及共同方法變異分析，確認本研究之信效度，並採用偏最小平方法(Partial Least Square,PLS)，來進行結構方程模式(Structural Equation Modeling, SEM)研究模型分析。 總體研究一開始先進行分析社群網站行為、線上社會資本及採用社群商務三者主構面皆為二階構面之間的關係。結果發現社群網站行為對線上社會資本及線上社會資本對採用社群商務三個主構面皆是顯著正向的關係。接著本研究模型整體分析社群網站行為-參與及瀏覽、線上社會資本-結合型(親密朋友)及橋接型(普通朋友)對使用者採用社群商務的意願-給予及接受皆有正向顯著的影響。因此本研究並繼續在就個別研究子構面進行探討。 參與行為和瀏覽行為皆對結合型(親密朋友)社會資本有顯著的正向影響，但瀏覽行為對於結合型(親密朋友)社會資本的正向影響程度大於參與行為對於結合型(親密朋友)社會資本的正向影響程度。參與行為和瀏覽行為皆對橋接型(普通朋友)社會資本有顯著的正向影響，但瀏覽行為對於橋接型(普通朋友)社會資本的正向影響程度大於參與行為對於橋接型(普通朋友)社會資本的正向影響程度。 參與行為對採用社群商務意願（給予）有顯著的正向影響，但瀏覽行為對採用社群商務意願（給予）沒有影響。參與行為和瀏覽行為皆對採用社群商務意願（接受）皆有顯著的正向影響，但參與行為與瀏覽行為二種行為對於採用社群商務意願（接受）沒有顯著的不同。 結合型(親密朋友)及橋接型(普通朋友)社會資本對採用社群商務意願（給予）皆有顯著的正向影響，且橋接型(普通朋友)社會資本對於採用社群商務意願（給予）的正向影響程度大於結合型(親密朋友)社會資本對於採用社群商務意願（給予）的正向影響程度。結合型(親密朋友)及橋接型(普通朋友)社會資本對採用社群商務意願（接受）皆有顯著的正向影響，但結合型(親密朋友)及橋接型(普通朋友)二種社會資本對於採用社群商務意願（接受）沒有顯著的不同。 綜上所述，總體而言，瀏覽行為對於社會資本優於參與行為對於社會資本的影響。參與行為對於採用社群商務意願優於瀏覽行為對於採用社群商務意願。橋接型(普通朋友)社會資本對於採用社群商務意願優於結合型(親密朋友)社會資本對於採用社群商務意願。 此外，在經由量化分析研究，本研究獲得全部研究假設分析的結果，接著本研究並採用事後質性分析(Post Hoc Qualitative Analysis)，利用半結構式深度訪談，對象包括大學部1-4年級的學生及上班族來做事後質性分析。事後質化分析的結果是將量化分析的所支持的假設加以驗証及探究更深一層的說明，並提供解釋量化分析所不支持假設的可能原因。 本研究理論貢獻在於採用社群網站行為-參與及瀏覽及線上社會資本-結合型(親密朋友)及橋接型(普通朋友)來預測採用社群商務-給予及接受之意願。較少研究採用社群網站行為及社會資料理論來解釋社群商務。管理意涵在於社群網站的營運者，應思考如何藉由社群網站來增加對使用者的〞黏性〞及提昇使用者瀏覽網站的流量，並依據使用者之前瀏覽的記錄，來推測並提供使用者所喜好的内容。更進一步來說，為增加使用者的參與行為可依照使用者彼此間的互動來分析使用者社群網站上的行為以推薦使用者加入其有興趣的粉絲團及社團，以增加親密朋友及普通朋友之使用者彼此的互動，同時增加線上結合型及橋接型社會資及使用者採用社群商務的意願。此外，本研究建議社群網站經營者應努力發展線上朋友介紹功能，並藉由大數據的資料分析擴大使用者的交友範圍。對於使用者線下的朋友，希望將其成為本身結合型(親密朋友)或橋接型(普通朋友)的社會資本。對於使用者線上不認識的朋友，希望能其成為橋接型(普通朋友)的社會資本，因為他們可能加入同一粉絲團或社團，以促進社群網站的使用者採用社群商務的意願。 / Following the fast growing of social network sites (SNS) such as Twitter, LinkedIn and Facebook in the cyber world recently, the social commerce has become an important emerging issue in these SNS. According to the Forrester Research (2011) predicted the total value of output for social commerce will reach US$30 billion in 2015. Because the development of social commerce will follow the continueous growing of SNS, the social commerce will play a pivotal role in the e-commerce. According to the literature review and focus group discussions, the study explored and applied SNS behavior including participating and browsing, social capital theory including bonding (close friends) and bridging (ordinary friends) to predict SNS users to adopt giving and receiving social commerce intention (SCI). The study applied an empirical research on SNS users i.e. FB users to be research target samples. The data collection were applied by paper_based and online_based as two way to collect sample data. With regard to sample data from paper_based survey, the study adopted undergraduate students of a university located in northen Taiwan was the research main collection samples for paper_based survey. As to online_based survey, the study utilized undergraduate students, who are the same as paper_ based survey in the same university with computer lab classes, FB users of researchers’ friends and Taiwan largest Bulletin Board System-PTT who are also FB users to collect online users’ data. The total valid samples are 970 from online_based and paper_based survey. After reliability analysis, validity analysis and common method variance testing, the study confirmed that reliability and validity of study samples met the requirement of each statistical testing and were qualified for testing the study whole research model. Then, the study adopted partial least square (PLS) to proceed with structural equation modeling (SEM) research model testing. As to whole model research testing, in the beginning the study was testing the relationship among SNS behavior, online social capital and SCI. The three constructs are second order constructs. The testing results show the effects of SNS behavior on online social capital and the effects of online social capital on SCI are significantly positive. Hence, the research continues to proceed with the research model testing. The research findings of whole research model testing have shown SNS behavior as a second order construct of participating and browsing, online social capital-bonding and bridging are both significantly positive influence to adopt social commerce intention as a second order construct as SCI (giving and receiving). Hence, the study will continue to explore the relationships between each sub-construct in the research. Participating and browsing behavior are both significantly positive on bonding social capital. Moreover, browsing behavior is more significantly positive associated with bonding social capital than participating behavior is associated with bonding social capital. Participating and browsing behavior are both significantly positive on bridging social capital. Browsing behavior is more significantly positive associated with bridging social capital than participating behavior is associated with bridging social capital. Participating behavior is significantly positive on adopting SCI (giving). However, browsing behioavr is no significant influence on adopting SCI (giving). Participaiting behavior and browsing behavior both have significantly positive influences on SCI (receiving). However, there are no significant differences between participating behavior and browsing behavior on adopting SCI (receiving). Bonding and bridging social capital are both significantly positive on SCI (giving). Moreover, bridging social capital is more significantly positive associated with SCI (giving) than bonding social capital is associated with SCI (giving). Bonding and bridging social capital are both significantly positive on SCI (receiving). However, there are no significant differences between bonding and bridging social capital on adopting SCI (receiving). To Sum up, generally speaking, browsing behavior is better on social capital than participating behavior is on social capital. Participating behavior is more suitable on adopting SCI than browsing behavior is on social capital. Bridging social capital is more appropriate on adopting SCI than bonding social capital is on adopting SCI. Besides, after quantitative data analysis, the study proceeds with the post hoc qualitative analysis. Utilizing semi-structured in-depth interviews with undergraduate students and working persons to continue post hoc qualitative analysis. The aim of post hoc qualitative analysis is to validitate and provide insight for the hypotheses which are supported by quantitative analysis. Moreover, it provides the explainations for the hypotheses which ae not supported by the quantitative analysis. The theory contributions of the study are applying SNS behavior including participating and browsing, social capital theory including bonding and bridging to predict SNS users to adopt SCI (giving and receiving). Seldom researches SNS behavior and social capital theory to explain social commerce individually. With regard to managerial implications are for SNS operators to think how to utilize SNS to increase users stickiness and raise users’ browsing quantities - following the users’ accumulated browsing behaivors history to predict and provide the contents which users may have interest. Moreover, regarding the interacting behavior for SNS users of participating behavior is to analyze the behaviors of SNS users and to recommend uers to join the fan pages or clubs that they may be interested in. Doing so is to increase the interactions with close friends and ordinary friends to enhance online bonding and bridging social capital and the users’ intentions to adopt SCI simotenously. In addition, the study recommends SNS operators to adopt friends introducing functions thoroughly and apply big data statistical analysis to enlarge the users’ friends rang to make more SNS friends. For the SNS offline friends, the study wish to let SNS users’ offline friends to become their online bonding or bridging social capital. For the SNS online unknown friends, the study wish to let SNS users’ online unkown friends to become their online bridging social capital because they may join the same fan pages or clubs to enlarge and accelerate SNS users to adopt SCI.

科学研究費補助金 研究種目:一般研究(B) 課題番号:06452428 研究代表者:松田 仁樹 研究期間:1994-1995年度

蒙地卡羅模擬是在組合信用風險的管理上相當實用的計算工具。衡量組合信用風險時，必須以適當的模型描述資產間的相依性。常態關聯結構是目前最廣為使用的模型，但實證研究認為 t 關聯結構更適合用於配適金融市場的資料。在本文中，我們採用 Bassamboo et al. (2008) 提出的極值相依模型建立 t 關聯結構用以捕捉資產之間的相關性。同時，為增進蒙地卡羅法之收斂速度，我們以 Chiang et al. (2007) 的重要性取樣法為基礎，將其拓展到極值相依模型下，並提出兩階段的重要性取樣技巧確保使用此方法估計一籃子信用違約時，所有模擬路徑均會發生信用事件。數值結果顯示，所提出的演算法皆達變異數縮減。而在模型自由度較低或是資產池較大的情況下，兩階段的重要性取樣法將會有更佳的估計效率。我們也以同樣的思路，提出用以估計投資組合損失機率的演算法。雖然所提出的演算法經過重要性取樣的技巧後仍無法使得欲估計的事件在所有模擬路徑下都會發生，但數值結果仍顯示所提出的方法估計效率遠遠優於傳統蒙地卡羅法。 / Monte Carlo simulation is a useful tool on portfolio credit risk management. When measuring portfolio credit risk, one should choose an appropriate model to characterize the dependence among all assets. Normal copula is the most widely used mechanism to capture this dependence structure, however, some emperical studies suggest that $t$-copula provides a better fit to market data than normal copula does. In this article, we use extremal depence model proposed by Bassamboo et al. (2008) to construct $t$-copula. We also extend the importance sampling (IS) procedure proposed by Chiang et al. (2007) to evaluate basket credit default swaps (BDS) with extremal dependence and introduce a two-step IS algorithm which ensures credit events always take place for every simulation path. Numerical results show that the proposed methods achieve variance reduction. If the model has lower degree of freedom, or the portfolio size is larger, the two-step IS method is more efficient. Following the same idea, we also propose algorithms to estimate the probability of portfolio losses. Althought the desired events may not occur for some simulations, even if the IS technique is applied, numerical results still show that the proposed method is much better than crude Monte Carlo.

次級房貸風暴使得包裝複雜的衍生性金融商品紛紛遭受波及後，目前結構型金融商品的條款設計將朝簡單化和透明化的趨勢發展，有助於全球金融市場的效率性、完整性與穩定性。本文從市場上選擇具代表性的利率結構型商品，應用模型來推導商品的價格，並深入分析商品的報酬與風險型態。 本文分析的個案商品為全球知名的匯豐銀行所發行之十年期「固定期限交換利率利差連動債券」，在評價上將採用LIBOR市場模型，利用市場上既有的資料求算出期初遠期利率，並校準模型所需的參數化波動度函數與相關係數函數，建立與市場一致的利率期間結構與利率波動度期間結構。模擬路徑時應用最小平方法蒙地卡羅來求得該商品發行之期初價格，此外，亦採用反向變異法加速收斂效果，並針對商品的條款設計作拆解與分析。最後，本文探討了發行機構發行商品之風險與避險策略，並且從投資人之報酬及風險層面作詳盡地剖析。

本研究探討歐洲已開發市場之主權信用違約交換與主權債券和無風險利率之債券信用價差之間的動態關係以及價格發現現象。此外亦分析可能影響歐洲已開發市場主權信用違約交換與債券信用價差變動之因子。 實證結果發現信用違約交換有較明顯之價格發現功能，且信用違約交換與債券信用價差間之基準差與信用風險呈現正向關係。而歐洲主權債券因其性質特殊，其使用德國政府公債作無風險利率反而較歐元交換利率為佳。此外我們發現利率變化與股市皆為影響歐洲主權信用價差之因子，而波動率之影響不明顯，原因也可能是歐洲主權債券過去低風險而成為資金避險標的之特殊性質。 / The thesis examines the dynamic relation between CDS and bond spread on developed European sovereign bonds. We also investigate which variables will affect the changes of CDS and bond spreads. We found that price discovery occurs on CDS more often, and the basis between CDS and bond spread has a positive relationship with credit risk. Due to the special characteristic of developed European sovereign bonds, the German sovereign bond yield is a better benchmark for risk-free rate than the Euro swap rate. Also we found that the change of rates and the return on stock market affect the European sovereign credit spread, but the effect of volatility on credit spread is limited. The reason should be the low-risk nature of these bonds in the past, which made them “safe” products for capitals to park.

可轉換公司債是給予持有者於債券存續期間內行使轉換為股票之複合式證券，除了債券性質外，內嵌的股票選擇權便屬於美式選擇權。而在本文中，針對內含美式選擇權的公司債評價是使用最小平方蒙地卡羅的數值分析，主要原因在於可轉債本身的條款彈性高，加上可轉債可能涉及之標的資產為兩個以上或狀態變數也可能具有多個維度(dimension)。此外，針對可轉債發行公司本身的信用問題，本文則採用縮減式(reduced-form)模型來處理其違約風險問題。依據A. Takahashi, T. Kobayashi, and N. Nakagawa認為採用結構式(structured-form)的缺點為參數難以校準，並列出下面兩論點認為使用縮減式的優點在於： 1. 違約事件將可能造成股價跳躍(jump)現象。 2. 在Duffie and Singleton方法下，資產隨機過程不必設定jump term，仍可設定為擴散過程(diffusion process)。 至於在利率期間結構方面，雖然Brennan and Schwartz(1980)認為實務上，考量利率的隨機性除了降低評價的效率性之外，與利率設定為常數相比，其差異不大。但針對為何差異不大的原因，本文認為利率對於純粹債券之價值影響為負向關係，而對於股票買權則是正向關係，故使得最後可轉債的影響則不明顯。然而，在目前「可轉債資產交換」等可轉債相關衍生性商品相繼推陳出新之下，使得可轉債的純粹債券與選擇權的個別要素評價也是相當重要。所以本文在利率風險的建構上將使用BGM模型來描述利率的隨機過程。

近年來由於金融自由化的發展，台灣已陸續開放新金融商品，除了股權相關的新金融商品之外，也陸續開放利率相關的新金融商品，如新台幣利率交換、新台幣利率選擇權、債券遠期交易、債券選擇權等。在信用衍生性商品市場方面，我國銀行從2002年底開放承做信用衍生性商品，目前正準備開放證券商承做。隨著金融國際化及自由化，未來將會從國外引進更新穎的金融商品，使金融市場更為完備。 本文以Hull – White利率模型及LIBOR市場模型為架構，藉由數值方法評價分析兩個衍生性商品──信用連結票券及利率交換。首先在信用連結票券方面，運用Li（1998）建立信用價差曲線（Credit Curve），將之應用至Hull – White三元樹，評價信用連結票券之價值，並作敏感度分析與避險參數分析。其次在利率交換方面，由於投資人端連結「雪球型」的支付型態，為路徑相依商品，故使用LIBOR市場模型以蒙地卡羅模擬法（Monte Carlo）進行評價與分析，再進行發行者損益兩平分析及情境分析。最後針對兩個商品的評價結果作結論，分析發行者及投資人的利潤及避險，並給予後續研究者模型改進之建議與方向。

在可預見之未來無線點對點的世界裡，資訊交易就如同現實世界人與人間的交易模式，需考量到交易資訊的成本、價值與人際關係衡量之因素；傳統廣告行銷的效率問題也待創新的資訊交易平台來解決。本研究提出的資訊交易模型，以人性考量為基礎，在各種不同情境下，動態衡量資訊成本、價值與使用者間關係，透過助理軟體，協助加入資訊交易平台之個體，以更便利方式進行資訊交易，預期讓交易結果更貼近使用者的需求；而企業方面也能透過點與點之間快速傳遞資訊的特性，預期以更低成本、更高效率，完成商務行銷目的，達到破壞性行銷之目標。 / This paper presents a novel ambient e-service aiming at distributed marketing through sensible bartering in foreseeable wireless Peer-to-Peer (WP2P) environments. A variety of influential factors (e.g., cost, value, relationship) are proposed and formalized for empowering the bartering mechanism, unfolding a rich arena of ambient distributed trading and a disruptive paradigm of e-marketing.

本研究之目的在於探討線上社群協作與其前置因素之關係，並且檢驗社群投入度之中介效果。首先，本研究應用角色內與角色外行為分類，提出角色內與角色外線上社群協作之類型。其次，依據社會認知理論與社會交換理論，本研究提出線上社群協作之前置因素，包括對個人之結果預期、對社群之結果預期、知覺社群信任、知覺社群規範、知覺社群支持與知覺社群認同。最後，本研究基於投入度觀點，檢驗社群投入度對於線上社群協作及其前置因素之中介效果。 本研究之研究情境為玩家公會社群，本研究自一知名大型多人線上遊戲中收集340份有效問卷進行資料分析與假設檢驗，研究結果顯示：(1) 除社群禮貌外，社群投入度對於社群合作行為、助人行為與運動家精神皆有正向影響；(2) 對個人之結果預期與線上社群協作行為之關係，會受到社群投入度所中介；(3) 對社群之結果預期與線上社群協作行為之關係，會受到社群投入度所中介；(4) 知覺社群信任與線上社群協作行為之關係，會受到社群投入度所中介；(5) 知覺社群規範與線上社群協作行為之關係，會受到社群投入度所中介，但社群規範對於運動家精神亦有直接負向效果；(6) 知覺社群認同與線上社群協作行為之關係，會受到社群投入度所中介，但社群認同亦會直接正向影響社群合作與助人行為；(7) 然而，知覺社群支持與線上社群協作行為之關係，並不會受到社群投入度所中介。針對上述之研究結果，本研究進一步闡述其學術研究意涵、實務管理意涵，以及研究限制與未來研究方向。 / The purpose of this study is to investigate the relationships between online community collaboration and its antecedents by examining the mediating effect of community engagement. First, this study proposes a classification of online community collaboration by following the typology of in-role and extra-role behaviors. Accordingly, this study will further examine some of online community collaboration behaviors, including community cooperation, helping behavior, community courtesy, and sportsmanship. Second, drawing upon social cognitive theory and social exchange theory, this study identifies several antecedents of online community collaboration, including person-relevant outcome expectancy, community-relevant outcome expectancy, perceived community trust, perceived community norms, perceived community support, and perceived community identification. Finally, from the perspective of engagement, this study will examine the mediating effect of community engagement on the proposed model. The research setting of this study is online gaming communities. After collecting 340 valid responses from a famous Massive Multiplayer Online Game (MMOG), our results show that (1) community engagement affects community cooperation, helping behavior and sportsmanship positively, but does not have a significant effect on community courtesy; (2) the relationships between person-relevant outcome expectancy and three online community collaboration behaviors are mediated by community engagement; (3) the relationships between community-relevant outcome expectancy and three online community collaboration behaviors are mediated by community engagement; (4) the relationships between perceived community trust and three online community collaboration behaviors are mediated by community engagement; (5) the relationships between perceived community norms and three online community collaboration behaviors are mediated by community engagement, but perceived community norms also has a direct and negative impact on sportsmanship; (6) the relationships between perceived community identification and three online community collaboration behaviors are mediated by community engagement, but perceived community identification also has direct and positive impacts on community cooperation and helping behavior; (7) the relationships between perceived community support and online community collaboration behaviors are not mediated by community engagement. According to these findings, this study concludes with research implications, managerial implications, research limitations and future research directions.

本篇論文將厚尾分配（Heavy-Tailed Distribution）應用在財務及保險精算上。本研究主要有三個部分：第一部份是用厚尾分配來重新建構Lee-Carter模型（1992），發現改良後的Lee-Carter模型其配適與預測效果都較準確。第二部分是將厚尾分配建構於具有世代因子（Cohort Factor）的Renshaw and Haberman模型（2006）中，其配適及預測效果皆有顯著改善，此外，針對英格蘭及威爾斯（England and Wales）訂價長壽交換（Longevity Swaps），結果顯示此模型可以支付較少的長壽交換之保費以及避免低估損失準備金。第三部分是財務上的應用，利用Schmidt等人（2006）提出的多元仿射廣義雙曲線分配（Multivariate Affine Generalized Hyperbolic Distributions; MAGH）於Boyle等人（2003）提出的低偏差網狀法（Low Discrepancy Mesh; LDM）來定價多維度的百慕達選擇權。理論上，LDM法的數值會高於Longstaff and Schwartz（2001）提出的最小平方法（Least Square Method; LSM）的數值，而數值分析結果皆一致顯示此性質，藉由此特性，我們可知道多維度之百慕達選擇權的真值落於此範圍之間。 / The thesis focus on the application of heavy-tailed distributions in finance and actuarial science. We provide three applications in this thesis. The first application is that we refine the Lee-Carter model (1992) with heavy-tailed distributions. The results show that the Lee-Carter model with heavy-tailed distributions provide better fitting and prediction. The second application is that we also model the error term of Renshaw and Haberman model (2006) using heavy-tailed distributions and provide an iterative fitting algorithm to generate maximum likelihood estimates under the Cox regression model. Using the RH model with non-Gaussian innovations can pay lower premiums of longevity swaps and avoid the underestimation of loss reserves for England and Wales. The third application is that we use multivariate affine generalized hyperbolic (MAGH) distributions introduced by Schmidt et al. (2006) and low discrepancy mesh (LDM) method introduced by Boyle et al. (2003), to show how to price multidimensional Bermudan derivatives. In addition, the LDM estimates are higher than the corresponding estimates from the Least Square Method (LSM) of Longstaff and Schwartz (2001). This is consistent with the property that the LDM estimate is high bias while the LSM estimate is low bias. This property also ensures that the true option value will lie between these two bounds.