In static equilibrium problems, strictly, there is no rotation actually happening and so no actual axis of rotation. However, there is often a potential axis of rotation-- rotation would occur in a certain direction due to torque from a given force if there were no balancing torques. (You can think of a torque in balance as ``trying'' to rotate an object in a certain direction.) The best place to pick a pivot point for determining torques is often where the axis of rotation would be (such as a hinge or joint).