The existing studies assume or treat an authoritarian party organization as a static and uniform structure, in which national party leaders dominate the party on the ground. Moreover, the extant explanations of authoritarianism focus on the effect of macro-level factors (e.g. the changes in the nature of democratic competition, political culture and institutional structure) over the internal strategies of the party leadership. Thus, little attention is paid to the role that local party actors play in authoritarian party structures. This study attempts to enhance our understanding of dynamics and factors behind party authoritarianism by raising the following questions: What does constitute party authoritarianism? Is it really a uniform or a static phenomenon as assumed? If not, how can we explain the variance in party authoritarianism? What might be the theoretical and policy implications of such an analysis for democratic development and party governance? By conducting a comparative case study of four political parties (AKP, CHP, MHP, DTP) in four geographically and politically distinct urban districts (Karşıyaka, Ümraniye, DiyarbakırMerkez and Tarsus) within Turkish political system, this study identifies four types of authoritarian party structures: benign, clandestine, challenged and coercive. In order to explain this variance, this study utilizes principal-agent approach, which is modified in two ways. First, as opposed to internally democratic parties, it is the national party leaders (principals) that delegate authority to local party actors (agents) in authoritarian parties. Second, the interest configurations between the principals and agents are based on not only material but also social interests. Material interests are those associated with power-seeking aims such as a desire for a position in public office. Social interests refer to the shared ideas and values such as ideological attachment, policy interests or loyalty to the leader. It is argued that interest configurations, which constitute the power structures between the national party leaders (principals) and local party actors (agents), vary across space and time. Second, the endogenous and exogenous triggers such as the outcomes of candidate selection processes or electoral defeats have the potential to cause a change in the power equilibrium between principals and agents, which might generate a new type of party authoritarianism or an exit to democratic party governance. Empirical analyses indicate that the agents motivated primarily by material interests are subordinate to party authoritarianism due to the material benefits received from the principals (benign authoritarianism). The agents motivated by social or ideational interests accept the subordination because of their loyalty to the party leader or the party ideology (clandestine authoritarianism). That been said, the agents whose interests conflict with the principals as a result of exogenous or endogenous triggers might attempt to shirk from the authority of the principals and object the authoritarian party structure (challenged authoritarianism). The authoritarian-leaning principals, in response, may exert coercion over the challenging agents (coercive authoritarianism). The success of the challenging agents over the principals depends on their power resources, such as information, social and economic status, legitimacy and networking with other agents. This work, thus, shows that party authoritarianism should be understood as a dynamic and heterogeneous phenomenon, which shows significant degree of variance across space and time. To have a better sense of this dynamic phenomenon, we need to focus on the role of micro-level factors (i.e. interest configuration and power relationships among principals and agents). With respect to broader implications, the principal-agent (PA) relationship must be understood in a different way in authoritarian party organizations where the major responsibility of the local party actors is to fulfill the tasks set by the national party leaders. Therefore, in studying the power structure of authoritarian party organizations, contrary to the conventional understanding, it is useful to assign the role of the principal to the national party leaders and the role of the agent to the local party actors. Another implication of this study is that exit from party authoritarianism is always a possibility not only because the national party leaders choose to do so, but also because the local party actors have the potential to cultivate new power resources and create power networks against authoritarian party structures. Yet, this possibility arises only when there is a conflict of interests between the agents and principals. Therefore, what causes the rise of such intra-party conflicts (e.g. electoral defeats, outcomes of candidate selection processes) and what prevents them from arising (e.g. material benefits, or social interests such as leadership loyalty, ideological attachment) must be given further recognition in studying internal dynamics of party authoritarianism.