This research deals with alternative modeling approaches to multiple agent dynamics. Models of a supply chain system are constructed to make comparisons about the capabilities of aggregated (System Dynamics) and disaggregated (Agent-Based) modeling approaches, based on a query to answer questions such as “Can aggregated, macro-level modeling capture the dynamics of micro-level, agent-based modeling? In what specific cases?” Effects of several factors, including inventory positions, price, shadow orders, loyalty, safety stocks, and ordering policies are analyzed. It is shown that there are factors, effects of which can be captured by System Dynamics at an aggregate level; however it is also observed that System Dynamics may miss the dynamics at more detailed level resulting from the emerging heterogeneity among individual agent behaviors in these cases. There are also cases where System Dynamics cannot capture the dynamics generated by ABM, even at an aggregate level. Regarding the supply chain dynamics, it is shown that when agents try to act ‘rationally’, emergent system behavior may become destructive. Loyalty and reliable safety stocks are proposed as strategies against oscillations in the supply chain.