Coevolution of opinion and consumption behavior under a two-layer network framework

Individuals’ optimistic or pessimistic perceptions of the economic outlook typically spread rapidly through online networks, while consumption intentions and behaviors play a critical feedback role in this process with active consumption boosting market confidence and cautious consumption reinforcing pessimistic expectations. To elucidate this complex coupling mechanism, we construct and analyze a dynamic model – within a two-layer network framework – that integrates opinion propagation with a consumption activation–deactivation process (i.e., transitioning between active and inactive consumption). Specifically, we employ a threshold model on the first layer to simulate the evolution of “positive/negative economic-outlook” opinions, and an Susceptible–Infected–Susceptible(SIS)-like model on the second layer to capture individuals’ consumption behaviors. These two layers are then tightly coupled through an interaction formula. In experiments on a synthetically generated ER–$k$-regular two-layer network the system flips abruptly, jumping from a pessimistic-low-consumption state of roughly (0.16, 0.45) to an optimistic-high-consumption state of roughly (1.00, 0.95); for the real two-layer network derived from the Wiki-Vote dataset the deterministic and Monte-Carlo steady-state values differ by no more than 5 percent, and the phase-transition point shifts by only about 0.03, confirming a robust, quantifiable phase change. This sharp sensitivity defines a narrow policy-sensitive region where modest interventions can steer collective expectations and revive consumption.