標題: 半導體產業價值鏈創新與企業評價關聯性分析:修正式Ohlson模型與多產出資料包絡分析法之應用
Application of Modified Ohlson Model and Multiple Output Dea Model in Exploring Innovation Impact on Firm Value Along the Semiconductor Value Chain
作者: 蔡璧徽
Tsai Bi-Huei
國立交通大學管理科學系(所)
公開日期: 2012
摘要: 過去文獻以發展費用作為研發活動的認列標準(Sougiannis (1994) 與Chan, Lakonishok and Sougiannis (2001)),然而,研究發展費用為公司創新活動的投入,並非 為創新活動的產出。本研究採用專利的數目與引證率指標為企業創新指標,以台灣半導 體為研究對象,在IC產業價值鏈呈現上游IC設計、中游IC製造與下游IC封測垂直分工 (vertical industrial disintegration)的架構下,探討專利產出與專利外溢效果對價值的影響。 本研究預計按三年期進行:(1)第一年計畫先探討專利產出與專利外溢效果對台灣 半導體價值的影響。(2)第二年計畫依據垂直分工理論(vertical disintegration theory),將 IC產業價值鏈區分上游IC設計、中游IC製造與下游IC封測垂直分工的架構,測試各生產 階段專利與產業內專利外溢對市值的邊際效用差異。(3)第三期計畫則是整合獲利與專利 產出作為績效指標,比較台灣積體電路產業各生產階段導入ERP系統前後績效差異。IC 設計階段偏重產品創新,ERP導入能使人力與企業作更有效運用,預期有助於產品創新, 然而,IC製造與封測階段偏重製程創新,ERP導入期間需耗時調整製程,ERP導入期間預 期反而阻礙製程改善。 全球半導體專利訴訟與日遽增凸顯核心技術於高科技產業的重要性,2009年美國法 院判決中芯半導體公司侵害台灣積體電路公司專利權,需賠償台積電公司10億美元,然 而,會計準則能未明確制訂研究發展資本化的準則,無法合理因應高科技管理決策與企 業評價,本研究成果除預期能作為會計準則依據,亦能作為企業評價與決策的基礎。 第一年計畫:修正式Ohlson 模型於半導體專利外溢效果之應用 過去研究Sougiannis (1994)與Chan, Lakonishok and Sougiannis (2001)以Ohlson模型 闡述研究發展活動為公司的無形資產,測試研究發展費用與市場價值的關聯性,然而, 研究發展費用為公司創新活動的投入,並非為創新活動的產出。本研究採用專利指標作 為創新產出的代理變數,運用Ohlson模型架構測試專利指標與台灣半導體公司市場價值 的關係,驗證專利是否提升公司市場價值。 此外,依據經濟學群聚理論,台灣半導體產業群聚於科學工業園區,會促進資訊外 溢,本研究進一步在Ohlson模型的架構下,探討半導體產業同產業專利外溢對於公司價 值的影響,本研究專利指標除了採用量化的公司專利數目外,尚採用專利品質的標準: 公司專利被引證數目、公司專利自我引證數目、公司專利由其他公司引證數目,測試台 灣半導體產業被引證頻繁的專利,對公司價值的影響。實證結果若證實專利與專利外溢 效果會提升公司價值,隱含公司專利能提升企業與產業的經濟效益,會計準則應考量資 本化專利發明的所有支出,研究結果預期能作為會計準則無形資產認列的參考,此為本 研究的預期貢獻。 第二年計畫:垂直分工理論於半導體產業專利外溢效果之應用 依據垂直分工理論,IC產業價值鏈呈現上游IC設計、中游IC製造與下游IC封測垂直 分工的架構,IC設計生產階段偏重的產品創新,其他廠商學習吸收新知困難,反觀,IC 製造與封測階段偏重製程的改善,其他廠商學習吸收新知較為容易,半導體產業價值鏈 下,不同階段的創新型態與有所差異,半導體廠商研發活動所產生的外溢效果亦會不同。 本研究偏重探討不同階段的專利發明與專利外溢效果對市場價值的影響,實證結果 若證實IC設計、製造與封測專利外溢效果均有所差異,隱涵IC設計生產階段偏重的產品 創新,其他廠商難以學習後加以創新,專利外溢效果較低,反觀,製造與封測階段偏重 製程的改善,其他廠商容易應變學習,專利外溢效果較高,研究結果能作為產業趨勢分 析與投資大眾評價的參考,此為本計畫的預期貢獻。 第三年計畫:多產出資料包絡分析法於半導體產業ERP 效率衡量之應用 過去研究僅以會計獲利作為績效指標(Tsai, 2008),然而,對高科技產業,創新能 力是永續經營的關鍵要素,專利為創新產出指標,故本研究以專利及獲利雙產出兩種績 效指標,衡量台灣半導體產業廠商績效。 本計畫依據垂直分工理論,將IC產業價值鏈區分為上游IC設計、中游IC製造與下游 IC封測垂直分工的架構,研究台灣上市上櫃半導體產業公司ERP導入前後研發成長的變 遷,以及影響研發成長與績效改變的因素。研究設計以兩階段的方析方式分析ERP系統 對於公司專利核准與經營效率的影響;第一階段以資料包絡分析法(DEA)計算效率值, 再以Wilcoxon符號等級檢定企業導入ERP系統前後之效率值是否具有顯著差異,第二階 段以Tobit迴歸分析影響企業經營效率的因素。 IC設計生產階段偏重的產品創新,ERP的導入有助於節省時間人力,預期IC設計廠 商ERP導入後專利產出增加,反觀,IC製造與封測階段偏重製程的改善,組織龐大勞力 密集,ERP導入前幾年組織衝突需耗時協調,可能會降低研發績效。研究結果能作為高 科技企業會計資訊系統導入組織協調與創新管理的參考,為本計畫預期貢獻。
Innovation effects on firm value have not been effectively discussed because previous studies took R&D stock as innovation proxy (Lin and Yeh, 2005; Yasuda, 2005; Yang and Huang, 2005). However, R&D expenditure is an input rather than output of technological innovation, and is often allocated in a somewhat arbitrary fashion in firm counts. To avoid this type of biased measurement problem, this study takes patent numbers to represent the quantity of innovations, and citation frequency to quality. Based on the Ohlson model, our study for the first time explains firm value through technological innovation variables (number of granted patents, inventors, total citations, self-citation and other-citations), patent spillover proxies, financial statement variables (book value and abnormal return) and tax shield factors. This study focuses on the Taiwan Integrated Circuit (IC) industry and U.S. fabless IC design houses. The design sector of these two countries took the first and second place of the global market share in 2004, and Taiwan also championed the foundry IC packaging and testing industry. Under the vertical disintegration structure of Taiwan IC industry, each step in IC manufacturing, packaging and testing stages is fixed with standardized machinery and equipments. The innovations focus mainly on “process innovations.” On the contrary, IC design firms are structured differently according to the applications of final electronic products- meaning that their R&D activities are dedicated to “product innovation”. This study applies partial F statistics and likelihood ratio statistics to examine the differences of patent or patent spillover effect on firm value between process and product innovation along Taiwan IC value chain. In addition, this paper employs the multiple outputs, both innovative output (patents), and financial output (net sales), in our Data Envelopment Analysis (DEA) models to compensate the shortcoming of previous studies, which only utilize financial data to examine the firm performance. Furthermore, this study examines the impact of ERP systems on firm innovation. The study will assess whether the firm performs better after the implementation of ERP systems and what might be the reasons for such efficiency elevations. Twenty-five Taiwanese semiconductor firms that implemented ERPs are analyzed based on our two-stage method. First, this study uses DEA method to analyze the firm efficiency. Next, nonparametric Wilcoxon sigh rank test is applied to compare the efficiency between pre-ERP and post-ERP periods. Third, Tobit regression model is employed to analyze the determinants of efficiency improvement under ERP installations. The results are expected as follows. First, patent indicators, including number of granted patents, inventors, total citations, self-citation and other-citations, are predicted to be positively related to firm value. Patent spillover also elevates Taiwan IC firms values because of the clustering environment, which stimulates the information sharing among IC firms. Second, both the innovative impact and patent spillover effect on firm value are significantly different between product and process stages along IC industry value chain. Finally, firm efficiency is likely to upgrade after ERP implementation. The significant improvement is due to the development of patents, the curtailing of firm’s operating cycle days, receivable turnover days and inventory turnover days.
官方說明文件#: NSC99-2410-H009-016-MY3
URI: http://hdl.handle.net/11536/98304
https://www.grb.gov.tw/search/planDetail?id=2391862&docId=380450
顯示於類別:研究計畫