It is generally accepted that centralized resource allocation can achieve better performance than distributed resource allocation in terms of system optimization at the cost of signaling overhead. However, it may be not true in D2D broadcast communication, especially when any closed-loop feedback on PHY, MAC, RLC, or PDCP layer is not allowed. In this paper, we provide performance comparison between centralized and distributed resource allocation schemes in an outside network coverage scenario, according to evaluation methodology and simulation environment agreed in current 3GPP RAN1 D2D standardization. Based on our system level simulation results, it is shown that distributed resource allocation using an implicit coordination (e.g., energy sensing) can achieve better performance than centralized resource allocation schemes under consideration, especially when the frequency resource granularity is small. Therefore, distributed resource allocation in D2D broadcast communication can be the baseline approach for Public Safety in outside network coverage.