本文考察了自20世紀70年代末台灣的高新技術產業化的過程中，由工業技術研究院（工研院）經歷了轉變。本文提出了一個框架，與經濟和政治變化的環境中所發生的高科技產業整合工研院的互動。使用機構學習模型，本文標識的組織學習能力。在本質上這些學習能力突出I工研院的相互作用調整的變化的機制。歷史分析表明三個階段，其中的研究院通過主要組織變革發生的背景。階段之間的過渡是不同的，嵌入在特定的經濟和政治背景。 / This thesis examines the transformations undergone by the Industrial Technology Research Institute (ITRI) during the course of Taiwan’s high-tech industrialization since the late 1970s. This study proposes a framework that integrates ITRI’s approaches in interacting with high-tech industries on the one hand, with the broad economic and political changes happening in its environment on the other. Using an institutional learning model, the thesis identifies the organizational learning capabilities, which in essence underlines the mechanisms of ITRI’s interaction adjustment to the changes. The historical analysis suggests three stages, in the contexts of which the institute undergoes through major organizational transformations. The transitions between the stages are different and embedded in the particular economic and political contexts.

研究目的-中國從1989年就開始實行火炬計畫並投資了大量的資金在高新技術產業，在過去的三十年間，中國的經濟發展也快速地成長，本研究旨在透過中國高新技術產業的統計資料來估計其創新對企業生產績效之影響。 研究方法-本研究使用來自中國國家統計局的全國企業專利資料與全國企業績效資料並結合兩者為一資料庫，運用Olley and Pakes 生產力模型來估計生產力，最後利用追蹤資料的迴歸模型來進行創新對企業生產績效的實證研究。 發現-本研究把企業不分類別的專利總數當成依變數與企業績效如產出、生產力、出口等自變數進行追蹤資料的迴歸分析後，其呈現正向且顯著的關係。進一步把專利總數分成發明專利數、實用新型專利數與設計專利數三類依變數並與企業績效如產出、生產力、出口等自變數進行追蹤資料的迴歸分析後，其均呈現正向且顯著的關係。此外，本研究亦發現企業進出市場行為會反向的影響專利數與企業績效之間的關係。 價值-過去有關於創新與企業績效的研究通常都是以整體製造業的規模來進行分析並提供一個總體的估計，因此本研究針對中國製造業中的高新技術產業進行分析以提供對於中國的火炬計畫成效更詳盡的評估，並對未來有關高新技術產業的研究提供初步的研究方向。 / Purpose – China has started a Torch program and invested huge amount of money in high-tech industries since 1989. During last three decades, China’s economy also rapidly grows. According to above conditions, this study aims to create a reliable estimation for confirming the relationship between innovation and firm performance from the evidence of China’s high-tech industries. Design/Methodology/Approach – By building a sample composed of the patent stock and firm-level performance data from China’s National Bureau of Statistics (NBS), we apply Olley and Pakes method to estimate the productivity and use the regression model for panel data to do the empirical study. Findings – Both the total patent stock and the patent stock of three different categories maintain a positive and significant relationship toward firm performance such as output, productivity and exports. On the other hand, the finding also implies a negative effect of firm dynamics on the relationship between patent stock and firm performance. Value – The researches related to innovation and firm performance in the past are usually conducted with the sample of whole manufacturing industries’ data and report an overall estimation. However, this study focuses on high-tech industries to provide a more detailed evaluation for China’s innovative efforts of Torch program and carves out a direction for future research on high-tech industries.