We propose a parallel and distributed algorithm for the problem of beamformer design and power allocation in multi-cell massive multiple-input multiple-output (MIMO) networks. Inter-cell coordination and overlapping clusters are considered to enhance the system performance. To overcome the tight coupling induced by inter-cell coordination and overlapping clusters, we formulate the system into a two-layered framework. In this framework, the problem is decomposed into independent subproblems which are naturally solvable in a parallel and distributed manner. For each subproblem, an efficient parallel algorithm is proposed to speed up the convergence rate. The simulation results show that our inter-cell coordination strategy can approach the performance achieved by a high-quality limited inter-cell coordination strategy at low signal-to-noise ratio (SNR), and outperform it at high SNR. Moreover, the proposed parallel algorithm enjoys great parallelization speedup with moderate transmit power consumption.