由政府利益角度探索美國國防專利表現與技術分布

Abstract

國防科技被視為國家安全與高階產業能力的重要支柱，各國政府長期透過公共研發投資推動關鍵技術與國防相關創新，然而政府資助是否必然能轉化為具高度技術影響力的創新成果，仍缺乏系統性且以實證為基礎的檢驗，特別是在高度任務導向與系統整合的國防與工程技術領域中，公共研發成果的創新表現未必能充分反映於傳統專利影響指標。本研究以美國專利商標局（USPTO）專利資料為基礎，聚焦揭露政府資助資訊之政府利益專利（Government-Interest, GI patents），系統比較企業公司、學研機構與政府機關三類組織中GI與非GI專利的技術表現差異，並透過短、長期專利引用影響力、專利集中度及技術影響力，分析政府資助在不同組織類型與技術結構下的創新成效。實證結果顯示，GI專利在授權初期具有一定程度的可見度，但其長期引用影響力普遍低於非GI專利，且此現象在技術更新快速、競爭激烈的國防研發領域尤為明顯；進一步分析亦發現，即使在相近的政府涉入程度下，國防研發組織仍呈現高度異質的創新表現，顯示政府資助本身並非決定技術影響力的充分條件。相較之下，專利組合結構、技術聚焦程度，以及組織是否能將公共資源有效內化為可持續的技術能力，才是影響國防創新成效的關鍵因素。整體而言，本研究指出，評估國防公共研發成效時，不宜僅以政府資助規模或專利數量作為判斷依據，而應從技術結構與長期影響力的角度，重新理解GI專利在國防研發中的角色，並為國防科技政策設計與研發管理提供實證參考。

Defense technology is widely regarded as a critical pillar of national security and advanced industrial capability, and governments have long relied on public research and development (R&D) investment to promote key technologies and defense-related innovation. However, whether government funding can necessarily be translated into highly influential technological outcomes remains insufficiently examined through systematic empirical evidence, particularly in mission-oriented and highly integrated defense and engineering domains where innovation value may not be fully captured by conventional patent impact indicators. Using patent data from the United States Patent and Trademark Office (USPTO), this study focuses on Government-Interest (GI) patents—those explicitly disclosing government funding involvement—and systematically compares the technological performance of GI and non-GI patents across business enterprises, academic and research institutions, and government agencies. Multiple indicators, including short- and long-term patent citation impact, patent concentration measured by the Herfindahl–Hirschman Index (HHI), and technological impact measures, are employed to examine how government funding translates into innovation outcomes under different organizational and technological structures. The empirical results show that while GI patents exhibit a certain degree of visibility shortly after grant, their long-term citation impact is generally lower than that of non-GI patents, with this pattern being particularly pronounced in fast-evolving and highly competitive defense R&D domains. Further analysis reveals substantial heterogeneity in innovation performance among defense R&D organizations operating under similar levels of government involvement, indicating that public funding itself is not a sufficient condition for achieving high technological influence. Instead, patent portfolio structure, technological focus, and an organization’s ability to internalize public resources into sustainable technological capabilities play a decisive role in shaping defense innovation outcomes. Overall, the study suggests that the evaluation of public defense R&D performance should move beyond funding scale or patent counts and instead adopt a structural and long-term technological impact perspective, thereby reinterpreting the role of GI patents in defense R&D systems and offering empirical insights for defense technology policy design and R&D management.