Modeling Strategies to Ensure Food Safety in the US Fresh Produce Supply Chain

Economic models of contamination and risk control often assume producers respond to incentives in a predictable and uniform manner. However, real-world producer behavior is characterized by heterogeneity, adaptive learning, and complex responses to policy interventions. In order to model this, we use an agent-based model (ABM) to simulate the decision-making of 1,000 producers across 10 regions over a 60-month period to analyze the effects of dynamic food safety testing regimes. The simulation with our designated parameters yielded counter-intuitive yet significant results: system-wide contamination rates decreased from 13.1% to below 1%, while the testing rate itself was reduced by 73 percentage points. This outcome was driven by a dramatic increase in producer risk-control effort, which grew 11-fold in response to the perceived effectiveness of the testing system. With our configuration, the investment in capital-intensive technology remained low (approximately 3%), suggesting that behavioral change was the primary driver of improved safety outcomes. The entire system costs fell by 97.3%, from $1.63 million to $43,800 over time as the contamination and testing requirements reduced. These findings suggest that dynamic, feedback-based testing policies can achieve superior food safety outcomes at substantially lower costs than static regulatory approaches.