Damping 'delays' response. When I put a displacement at the end of a cantilever, the maximum load near the anchor occurs after the start of the load. This supports the equation cos(wt+phi), not cos(wt-phi).

Could you provide more information to support your conclusion above?