Followership is generally defined as a strategy that evolved to solve social coordination problems, and particularly those involved in group movement. Followership behaviour is particularly interesting in the context of road-crossing behaviour because it involves other principles such as risk-taking and evaluating the value of social information. This study sought to identify the cognitive mechanisms underlying decision-making by pedestrians who follow another person across the road at the green or at the red light in two different countries (France and Japan). We used agent-based modelling to simulate the road-crossing behaviours of pedestrians. This study showed that modelling is a reliable means to test different hypotheses and find the exact processes underlying decision-making when crossing the road. We found that two processes suffice to simulate pedestrian behaviours. Importantly, the study revealed differences between the two nationalities and between sexes in the decision to follow and cross at the green and at the red light. Japanese pedestrians are particularly attentive to the number of already departed pedestrians and the number of waiting pedestrians at the red light, whilst their French counterparts only consider the number of pedestrians that have already stepped off the kerb, thus showing the strong conformism of Japanese people. Finally, the simulations are revealed to be similar to observations, not only for the departure latencies but also for the number of crossing pedestrians and the rates of illegal crossings. The conclusion suggests new solutions for safety in transportation research.