Selecting suitable entities to implement blockchain in green agricultural supply chains.

Blockchain implementation in green agricultural supply chains can be undertaken independently by manufacturers or retailers or jointly by both in industrial practices. However, most existing studies predesignate implementation entities artificially and largely ignore firms' endogenous economic willingness, which restricts the practical explanatory capability. Addressing this research gap, this study systematically explores firms' implementation willingness and the optimal selection of blockchain implementers in green agricultural supply chains, thereby offering targeted decision-making insights for blockchain policymakers and industrial practitioner. Four game-theoretic models are constructed involving entities: none, retailer-only, manufacturer-only, and both. By comparing the equilibrium outcomes of different game models, this study employs the dual thresholds of deployment costs and operational costs to quantitatively evaluate the economic willingness of supply chain entities to implement blockchain. Different from prior studies that treat blockchain implementers as exogenous and only consider a single cost dimension, this paper endogenizes firms' blockchain implementation decisions based on economic intentions and establishes a dual-cost threshold framework to categorize and identify the optimal implementer under different scenarios, and illustrate the results with a case study. The key findings are concluded as follows. First, economic willingness varies within thresholds. Manufacturers and retailers may take the initiative to implement blockchain regardless of their partners' behavior, or act reactively by following their partners; beyond the thresholds, they refuse to implement it. Second, entity selection varies across scenarios. Under moderately high dual costs, independent retailer implementation is optimal; under balanced costs, manufacturers may implement alone. Extremely high operational costs lead to no willing implementer, while extremely high deployment costs require joint implementation. Third, the likelihood of implementation modes differs substantially within thresholds. The most likely scenario is retailers' independent implementation, the least likely is manufacturers' independent implementation, with joint implementation in between. Beyond these thresholds, no entity implements it at all, a situation accounting for over 50% in the case study. Methodologically, this paper applies global optimization rather than the local optimization strategy with fixed designated implementer settings in previous studies, effectively capturing the internal economic motivation driving voluntary blockchain implementation in green agricultural supply chains.