In this paper, we present a proposed methodology to determine the optimal capacitor locations and sizes for power-loss reduction in a radial distribution system. The objective is to minimize energy loss by considering capacitor cost. Here, both the power loss index (PLI)-based approach and the loss sensitivity coefficient-based approach are comparatively studied to determine the optimal capacitor location. The index-based approach combined with a genetic algorithm is used to determine the capacitor sizes. We also discuss about how customer loads and costs change after reactive power compensation. The proposed method were tested on the IEEE 13-bus and 34-bus test systems, and the results are comparatively analyzed and discussed.