It is widely recognized that the wireless spectrum is a scarce and limited resource and that the present practice of static spectrum allocation and exclusive licensing is inefficient. The proposed approaches generally either focus on maximization of spectrum utilization or profit maximization of individuals (such as the government or users). In this paper, we consider an efficient, or socially optimal, spectrum sharing that consists of three objectives: full (quantitative) utilization, effective (qualitative) utilization, and zero interference. Through a comparative study of these models using suggested objective criteria, we show a hybrid model consisting of a dynamic spectrum market and dynamic spectrum access supported by cognitive radio technologies that can achieve the social optimum. The dynamic spectrum market enabled by a benevolent social coordinator has fundamental differences from existing dynamic market models in that primary licensed user is not involved in the process of allocating underused spectrum. Moreover, the motivation of social coordinator is to reach socially optimal allocation of spectrum resources rather than to maximize profit or revenue of individuals.