東亞地區的代理問題與其上市櫃公司集中的股權結構有關。本研究以控制權與現金流量權偏離的幅度衡量公司治理的品質，探討最終控制股東股權治理結構對研發活動的支出、研發活動之攸關性（Relevance）和可靠性（Reliability）之影響。本研究沿用Kothari et al.(2002)的作法，以過去五年盈餘的變異性，作為研發活動可靠性的替代變數。實證結果發現，公司治理較佳的公司會投入較多的研發支出。當股權偏離幅度越小，研發活動會與股價成正相關，且盈餘變異性會較小。 / The agency problems of listed companies in East Asia is closely related to their typical concentrated ownership structures. The objective of this paper is to explore the association of corporate governance quality, measured by the divergence between the ultimate owner’s voting right and cash flow right, with the extent of R&D activities, and with the relevance and reliability of R&D expenditure. Following the Kothari et al.(2002), we use the standard deviation of subsequent earnings volatility as proxy for reliability of R&D outlays. Consistent with our hypotheses, we find that firms with higher quality of corporate governance tend to spend more on R&D activities.We also find that the positive associaton between R&D outlays and stock price is increasing as the control cash flow right divergence decreases. Finally, the positive associaton between R&D outlays and subsequent earnings volatility increases as the control cash flow right divergence increases.

本文以一個簡單賽局理論模型為分析依據,探討企業間的自製或是外包策略對整體社會福利的影響。 傳統經濟理論分析告訴我們自製策略可以解決雙重邊際化的問題。但在本文我們舉出一個反例，亦即當存在研發的可能及上下遊市場的研發外溢效果時，垂直整合可能會造成廠商過度研發的情形，最終產生整體社會福利不效益情形。 / In this paper, we set up a game-theoretic model, in which the whole industry is composed of a downstream monopolist and two upstream firms. This paper discusses the major debate in a firm’s market strategy: to make or to buy the components it needs. Traditionally, economic theories told us that vertical integration strategy (make strategy) can solve the problem of double marginalization and hence increase profits as well as social welfare. Nevertheless, this result does not necessarily hold when R&D possibility is introduced. We show that when there is downstream to upstream R&D spillover, vertical integration may result in an over-investment in R&D from the social standpoint. This result may shed some light on the rationale behind recent antitrust rulings.

本研究主要採用文獻探討以及個案訪談作為主要的研究方法，先藉由文獻探討建立起論文整體之架構以及相關理論之說明定義所需探討之研究變項，之後再透過台灣半導體標竿企業的製程研發專案訪談加以實證。本研究以研究「專案類型」、「技術知識特質」、與「組織架構特質」對「製程開發的創新活動」的關聯關係來探討台灣半導體製程的研發活動過程並說明所觀察現象的具體意義。另外，並透過研究不同性質的專案來觀察與客戶間的互動關係並說明所觀察到現象以及背後的思考邏輯，進而可得到以下初步之研究結論： 一、製程研發專案的技術與知識特質 1. 不同類型專案中，製程合併型程度越高的專案，其技術知識的內隱性較低、多元性較低、標準化程度較高、路徑相依程度較高。而製程技術標準化與相依度會影響專案成員在研發過程所採取的實驗條件設定的方式，進而在實驗開發階段會影響內隱度。 2. 全新元件的開發可分為兩個層次；第一、元件架構的確立。第二、製造流程的確立。 二、製程研發的團隊組成與研發活動 1. 專案組織架構隨專案類型差異而有明顯差異。製程合併型的專案一般為廠級的開發專案，傾向以「輕型團隊」之方式運作﹔而高度創新的製程研發專案傾向以「重型團隊」之方式運作。 2. 技術知識內隱程度愈高，多元程度程度愈高者愈傾向以試製實驗來共同解決問題。 3. 專案的知識內隱程度愈高越需外部知識的來源；製程研發專案的技術知識的內隱性、多元程度愈高，使用者參與程度愈傾向「共同開發」。 三、組織結構與知識的分享平台 1. 為蓄積、分享重要的技術資訊，成立正式組織統籌規劃高壓製程的研發，並於每季邀集各廠提供相關製程經驗分享。研發經驗會藉由團隊的研發過程與分享制度的建立，進行不同型式的知識轉移。 2. 研發專案團隊內的知識分享機制隨著機密程度不同而有所限制。 四、其他發現 1. 不同產品類別、背景客戶的電路設計習慣不一，可藉適當電路設計於開發時期找出元件的弱點。IDM廠對資料要求度較高，設計師習慣所有的文件資料都具備後才開始設計。 2. 使用者的參與製程開發，可讓新製程依使用者的習慣調整。對製程了解越深入的設計公司，其電路設計越游走合法邊緣。 / This thesis adopts reference and case study as the main research approach. It sets up the thesis whole structure by reference and relevant theories to define the factors. Afterward, to demonstrate the thesis structure by interview three projects about IC manufacture process developing of the company, which is the benchmark semiconductor company in Taiwan. This thesis attempts to take an exploratory study of the relationship between characteristics of project type, characteristics of technological knowledge, characteristics of organization structure, and innovation actitity in process developing on that company in Taiwan. In addition, through the case study, to observe the relationship with customer to explain the phenomenon and thinking in different case.There are primary figures found in the thesis： 1. The characteristics of technological knowledge in process developing projector. a. In different kinds of process developing projects, the combinative multi-process project with low degree in tacit knowledge and pluralism, but with higher degree in standardized and route- interdependence. The degree in process standardized and route- interdependence would influence the experiment condition, which would effects the degree of tacit knowledge in experiment period. b. The development of new device can be divided into two levels: First, the establishment of the device structure. Second, the establishment of the process flow. 2. The developing team make-up and developing activity a. The organization structure has obvious differences in different project type. The combinative multi-process project inclined to with the way operation of ' the light-duty group '; the research and develop with high innovation inclined to with the way operation of ' the heavy-duty group '. b. The projector with high degrree in tacit knowledge and pluralism inclined to solving the problem by the trial-producing experiment. c. Degree of technology diversity determines degree of user engagement in development. Projects with high degree of technology diversity tend to engage user in the joint development mode. Projects with low degree of technology diversity tend to engage user in the “Offering Mode”. 3. Organization structure and the sharing platform of knowledge a. It should build an official organization structure to overall planning developing projectors for knowledge accumulataion and sharing. b. The develipong experiment would be transferred in different type by sharing system building and organization set-up. c. The knowledge sharing of research and develop has limitation in different secret degree. 4. Others a. Different product classification, background of customers has different design style. Could make use of proper circuit design to find out the weakness of device. b. IDM factory require high quality documatation support. Designer used to design afeter all documentation ready, c. The new process could be adjusted according to user’s design style if he participates in the developing project. d. More understanding in process, the circuit design might violade the design rule.