Traditional design of wireless networks mainly focuses on system capacity and spectral efficiency. As green networking is an inevitable trend, energy-efficient design for future wireless networks becomes paramount. In this paper, we address energy-efficient resource management in downlink orthogonal frequency division multiple access networks. The focus is targeted toward multi-cell networks, which are composed of multiple base stations (BSs) sharing the available radio resources. Consequently, greater emphasis is given to techniques that take inter-cell interference into account. Resource management in our context refers to the task of allocating the radio resources in order to maximize energy efficiency. We devise resource management techniques that jointly tackle the problems of scheduling and power control. Accordingly, we adopt two different approaches: a centralized approach, where BSs coordinate in order to reach a globally optimal energy-efficient solution, and a distributed approach, where BSs selfishly strive to maximize their own energy efficiency. We portray the centralized approach as a convex optimization problem, whereas we have recourse to non-cooperative game theory to model the distributed approach. In particular, we show that the non-cooperative game converges to a unique Nash equilibrium in low- and high-interference scenarios. We perform thorough numerical simulations to quantify the discrepancy between the centralized and distributed approaches, and identify the conditions where they have precedence over the state of the art. Moreover, the simulation results highlight the fast convergence of our algorithms, which is a precious asset for realistic deployments.