Revenue-Sharing in China's Civil-Military Integration: Multi-Level Incentive Coordination and Optimization

Civil-Military Integration (CMI) has become a strategic imperative as states pursue technological innovation, economic efficiency, and resource optimization, yet implementation is hindered by moral hazard and adverse selection within multi-level principal-agent chains. This study develops a dynamic, multi-task principal-agent model in the Holmström-Milgrom (H-M) tradition tailored to China's CMI systems. The model accommodates intermediaries that act simultaneously as both principals and agents, and employs a revenue-sharing contract to align incentives across layers, thereby promoting the high-quality and cost-efficient delivery of dual-use technologies. The model further integrates risk aversion, interdependent efforts, and incomplete contracts. Analytical results and comparative statics show that optimal revenue-sharing and fixed-transfer parameters are highly sensitive to agents' risk preferences and marginal cost structures. Effort spillover among agents has a material impact on aggregate output and shifts the optimal sharing ratio; we identify parameter thresholds under varying institutional and economic constraints. Numerical simulations are complemented by robustness checks, including Monte Carlo experiments, bootstrap resampling, and extreme scenario analyses. Validation based on representative Chinese CMI projects (e.g., the Beidou-3 navigation system, J-20 fighter aircraft, and Y-20 transport aircraft) further confirms that the proposed mechanism remains stable across heterogeneous environments and is consistent with observed implementation outcomes. Policy recommendations include real-time performance monitoring, stratified incentives matched to agent characteristics and risk attitudes, and flexible, renegotiable clauses to accommodate external uncertainty. The study provides a formal, behaviorally grounded contract design that improves incentive alignment and operational efficiency in complex CMI systems.