Of course the examples in this article are rather oversimplified, but the concept is important. An organization's PfM function should be a strategically driven role that remains focused on ensuring that there is the highest possible likelihood of achieving the business goals and objectives that have been set during the planning process. Unless PfM remains focused on those goals and objectives it is simply another level of project management.
The resulting role may still add some value, but also has the potential to create confusion and (more importantly) leaves no one focused on the overall objectives that the organization is trying to achieve. If PfM does not play the strategic, objective focused role, then it will be left to the business sponsors. And as we know from past experiences that will inevitably result in partisan decision-making and internal politics none of which benefits the organization at large.
The implications of this are clearly that PfM has to be an executive function within the organization. PfM must have the authority to make decisions that may negatively impact individual business units and potentially damage the ability of those departments to achieve their own goals. Obviously, PfM must always act collaboratively with department heads. Nevertheless, there needs to be a clear understanding and acceptance that the focus is on maximizing the chances of the overall portfolio achieving the collective goals and objectives of the whole organization. To repeat, PfM is not a role that can simply be appointed. It is something that requires the portfolio manager to develop strong relationships with all department heads and create levels of trust that allow for portfolio management decisions to be recognized as driven by the overall organizational needs. And that must be true even when individual groups are negatively impacted by those decisions.
I sometimes compare a portfolio manager to a general managing the entire battle, and that's a good image to think of here. They are managing the resources available to them in order to achieve the strategic objectives that the organization has deemed important. In the "war" to achieve the complete set of objectives that have been set, it is acceptable to lose a few "battles". That means sacrificing one or two projects and leaving others only partially complete as long as they have been sacrificed in order to further the interests of the overall portfolio.
That can be a tough thing for organizations to accept, but it's what elevates portfolio management to a true strategic function, and improves the ability for the organization to achieve its destination.