This dissertation aims to explore what structural features can generate power-dependence relations in social networks (i.e., A's bargaining power over B is equal to B's dependence on A in the dyadic relation) and how those relations shape individuals' access to social capital. The first study identifies power-dependence relations in structural holes and network closure. Existing network theories and graph-theoretic measures explain how ego within structural holes and network closure becomes a recipient of valuable resources from donor-alters. Yet, their potential bargaining situations are often overlooked. I suggest new propositions and measures of Total Power (i.e., mutual dependence) and Relative Power (i.e., unequal dependence) in bargaining between ego and alters and apply them to examine the relationship between the social capital of actors and directors and their career-survival in the U.S. film industry, 1929-2019. Event history analysis results reveal very different survival mechanisms of actors and directors that previous theories and measures do not capture. The second study identifies power-dependence relations in centrality within context-specific network flows. Centrality measures identify influential nodes that are more involved in certain network flows. Yet, those network flows are purely defined in graph-theoretic terms (e.g., shortest-paths) and do not account for many context-specific patterns potentially embedded in social networks (e.g., paths through racially homogenous nodes). Applying the respondent-driven sampling (RDS) method, I propose a Monte Carlo approach to measuring centrality and power-dependence relations within complex and flexible network flows. I use this approach to explore the well-known networks of Florentine families, 1394-1434, and find interesting snapshots of their power game. The third study identifies power-dependence relations in chain of remittance flows. Remittances are often considered to yield a unidirectional flow of capital—the money from migrant workers' residential country to their home country. However, as social exchanges, remittances could provide certain rewards to migrant workers (e.g., social status in their family; emotional satisfaction), which may lead them to be more committed to their jobs. This hypothesis of the bi-directional flow of capital is tested using the Mexican Migration Project (MMP) datasets and formalized as a new model of power-dependence relations potentially embedded in social networks.