Determinants of the Optimal Network Configuration and the Implications for Coordination

This paper develops a simulation model to compare the performance of two stylized manufacturing networks: the lead factory network (LFN) and the archetype network (AN). The model identifies the optimal network configuration and its implications for coordination mechanisms. Using an NK simulation model to differentiate between exogenous factors (configuration) and endogenous factors (coordination), we find low complexity of the production process, low transfer costs and high search costs, as well as a larger number of manufacturing plants benefit LFN compared to AN. Optimally coordinating the chosen network configuration of LFN might require to fully transfer knowledge in the short run but to transfer nothing in the long run. Moreover, a late knowledge transfer from the lead factory to the plants increases the pre-transfer performance of LFN but results in a larger performance drop, yielding a lower short-run but a higher long-run performance of LFN.