This thesis contributes to the experimental economics literature on principal agent relations. Chapter one and Chapter two present lab experiments on an endogenous trust game in which one player (the principal) may decide to leave the investment choice to the agent or to take the investment decision himself/herself. The main result is that voluntary leadership increases efficiency if the players are symmetrically endowed while this is not the case if the principal is better endowed. Chapter three addresses the question whether organizing work in teams may be beneficial for the principal. We investigate a game where agents with different productivities choose whether to work in a team or individually. While standard neoclassical theory denies that team incentives have strategic value in this environment, we show that the principal can influence self-selection and increase performance with team incentives. Chapter four addresses a framework to estimate parameters of utility functions which are assumed to generate the observed data of experiments, namely: the logit equilibrium model introduced by McKelvey and Palfrey (1995, 1998). A three stage two player ultimatum game with advance production that includes sequential and simultaneous stages is analyzed. Parameters of a social utility function proposed by Fehr and Schmidt (1999) are estimated with maximum likelihood methods and compared to a standard utility function that assumes egoistic payoff maximizing preferences. The results suggest that the Fehr and Schmidt (1999) social utility model explains the data better than the standard model.