本文探討景氣因素、研發外溢效果及技術知識特質對IC設計業研發支出遞延效應之影響。本研究主要採用Lev and Sougiannis (1996)所發展之研發支出遞延效應實證模型，以國內IC設計業上市櫃公司為對象，透過分析產業特性，區分為多應用性IC設計群組及單應用性IC之資訊、通訊及消費性群組，並探討其研發支出遞延效應。研究發現如下： 一、在研發支出遞延效應中，IC設計業受景氣之影響僅限於當期之研發支出；就研發支出效益受影響程度上，IC設計業亦低於其他IC產業。此結果顯示，IC設計業由於具備產品多元化和應用多元等利基市場特性，且在國內是屬於成長型產業，故景氣影響程度不如其他下游產業來的大。 二、多應用性IC設計公司之研發支出所創造之未來效益，比單應用性IC設計公司高。多應用性IC設計公司因為受到研發外溢效果之影響，其當期及遞延一期之研發支出，會比單應用性IC設計公司創造更高之效益。此結果顯示，多應用性IC設計公司投入於不同領域產品之研發支出，存在類似產業內外溢效果，因此亦加強研發支出對未來效益貢獻之程度。 三、資訊類單應用性IC設計公司之研發支出，對未來所創造之效益金額及持續年限，均未優於通訊類及消費類單應用性IC公司。本研究發現，技術路徑相依度及技術變動程度兩種技術知識特質，並不能完全解釋單應用性不同群組之研發支出所創造未來效益的程度，而必須同時考量下游應用市場之目前狀況及未來潛力，才能對其研發支出遞延效應做出更正確之推論。 / This thesis analyzes three R&D performance issues of the IC design firms in Taiwan. First, it addresses the effect of the fluctuating economic cycle in the semiconductor sector on the R&D performance of design and non-design firms in the IC industry. Secondly, this study examines the R&D spillover effect on the R&D performance of the multifunctional and single-functional groups of IC design firms. Finally, this study discusses how technological knowledge (path independence and complexity) influences the R&D performance of the three subgroups (computer, communication, and consumer) of single-functional IC design firms. Three major findings of the study are as follows: 1.The fluctuating economic cycle in the semiconductor sector has less influence on the R&D performance of the IC design firms than that of the IC non-design firms. The fluctuation affects the R&D expenditure of IC design firms only in the current year, but that effect on the IC non-design firms exist in the current year and also the following year. The R&D performance of IC design firms is also less influenced. 2.The multifunctional IC design firms generate more benefit from R&D expenditure than single-functional ones, suggesting that the former group has a stronger R&D spillover effect. 3.Although the computer subgroup of IC design firms possesses high technological path dependence and low technological complexity, its R&D performance is not better than the other subgroups. This finding suggests that technological path dependence and complexity do not fully explain the difference in R&D performance among the three subgroups of single-functional IC design firms.