Defense Innovation Reform and Bureaucratic Incumbency: Limits of DARPA Emulation in South Korea

Abstract The U.S. Defense Advanced Research Projects Agency (DARPA) is widely treated as an organizational model for disruptive innovation and has inspired emulation across policy domains and countries. Its performance, however, may depend not only on portable design features but also on institutional conditions specific to the U.S. defense establishment. This article examines that tension through South Korea’s Future Challenge National Defense Technology Development Programme (K-DARPA). With high and stable defense R&D investment and dense policy-learning ties to the United States, Korea was a most-likely case for successful emulation. Yet K-DARPA diverged substantially from its prototype. Drawing on documentary analysis and 25 interviews with defense officials, program managers, scientists, industry practitioners, and policy experts, the article identifies three mechanisms through which incumbent defense bureaucracies filtered the imported model. Organizational internalization placed the programme inside the Agency for Defense Development, whose personnel regime and hierarchy constrained external recruitment, narrowed program-manager discretion, and reintroduced top-down planning. Evaluative carryover applied deployment-oriented selection and evaluation criteria to exploratory research. Accountability-driven risk aversion , rooted in audit and legislative scrutiny, left a statutory “honorable failure” provision unused. Incumbents thus adopted DARPA’s visible features while subordinating them to existing governance routines. Without the administrative precedents and political protections that let DARPA separate exploration from acquisition discipline, imported innovation agencies are domesticated rather than replicated.