Distributed generators (DGs), coupled with suitable control and communication infrastructures, are expected to play a key role in improving the efficiency of electricity grids. In this paper, we focus on low-voltage and single-phase microgrids exploring the interplay of distributed power loss reduction and communication. We select representative power-loss reduction algorithms from the state of the art and provide design rules for the required networking strategies in the presence of lossy communication links, assessing the impact of communication as well as electrical grid features. Toward this end, we devise a novel statistical cosimulation (electricity grid, communication, and control) framework that faithfully mimics the characteristics of real-world microgrids in terms of communication and grid topologies, power demand, and distributed generation from solar sources. Our numerical results highlight the role of communication procedures and the differences among the selected optimization techniques for power loss reduction, assessing their convergence rate and quantifying the impact of communication failures, line impedance estimation error, communication and electricity grid topologies, network size, and number of DGs.