Optimal monetary policy using an unrestricted VAR