Individual Submission Summary
Share...

Direct link:

Modeling Input Substitution in Agent-Based Supply Chains for Policy Analysis

Thursday, November 5, 8:30 to 10:00am, Property: Boston Marriott Copley Place, Floor: 3rd Floor, Room: Boston University

Abstract

We extend an agent-based model of supply chains to capture firms’ ability to substitute across inputs when major prices change or disruptions occur, as in cases geopolitical conflict and natural disasters. Most agent-based supply chain models rely on fixed-proportions production functions, which assume no flexibility in sourcing and can overstate the persistence and severity of shocks. By introducing an explicit elasticity of substitution, our framework allows firms to adjust their input mix more realistically, improving the model’s usefulness for public policy analysis. We implement and compare three approaches, balancing model generality with ease of model construction. In the first method, firms can switch among a set of Leontief production functions, subject to an adjustment cost that declines over time. Second, we use a Cobb-Douglas production function, which implies an elasticity of substitution of 1, and interpolate between Cobb-Douglas and fixed-proportion outcomes to represent elasticities between 0 and 1. Third, we incorporate a constant elasticity of substitution (CES) production function, which can represent any nonnegative elasticity, so that we can consider elasticities greater than 1. We demonstrate the different outcomes these production functions yield in an agent-based model of the small arms munitions supply chain constructed from bill-of-lading data. Our analysis focuses on substitution between copper and steel. The results show that substitution materially changes how shocks propagate through the network. Higher substitutability reduces aggregate volatility, shortens the duration of disruptions, and limits cascading bottlenecks. Lower substitutability, by contrast, amplifies bottlenecks and increases network fragility. This contribution is relevant for public policy in three ways. First, it provides a more realistic representation of how firms respond to shortages, price spikes, and trade disruptions. Second, it enables policymakers to test how assumptions about production flexibility affect the performance of interventions such as stockpiling, supplier diversification, industrial subsidies, and export controls. Third, it helps identify the sectors and firms that remain vulnerable even when some substitution is feasible. More broadly, the framework offers a practical way to incorporate production flexibility into agent-based models of strategic supply chains.

Author