For robotic manipulation tasks in uncertain environments, research typically revolves around developing the best possible software control strategy. However, the passive dynamics of the mechanical system, including inertia, stiffness, damping and torque limits, often impose performance limitations that cannot be overcome with software control. Discussions about the passive dynamics are often imprecise, lacking comprehensive details about the physical limitations. In the first half of this paper, we develop relationships between an actuator's passive dynamics and the resulting performance, to better understanding how to tune the passive dynamics. We characterize constant-contact physical interaction tasks into two different tasks that can be roughly approximated as force control and position control and calculate the required input to produce a desired output. These exact solutions provide a basis for understanding how the parameters of the mechanical system affect the overall system's bandwidth limit without limitations of a specific control algorithm. We then present our experimental results compared to the analytical prediction for each task using a bench top actuator. Our analytical and experimental results show what, until now, has only been intuitively understood: soft systems are better at force control, stiff systems are better at position control, and there is no way to optimize an actuator for both tasks.