Just to be geeky for a second: although corporate governance is a vast topic, it boils down to something called "principal-agent theory" in academic circles. The best way to understand this is to frame it in terms of delegation. For instance, the owners of an organisation (shareholders or taxpayers) are the "principal". They delegate tasks and responsibilities to the managers of an organisation (the Board, for example). The managers are the "agent", who speak and act on behalf of the principal. Agency theory deals with the risks and problems that can arise as a consequence of such a delegation. Governance is the framework and the processes that ensure that the agents remain aligned to the principals.
The aspect of corporate governance that PMOs are concerned with has to do with the investment on change made by the Board . The Board (on behalf of the owners) delegate authority and responsibility for delivering change (in the shape of projects) to a cascading hierarchy of managers and overseeing bodies that eventually apportion chunks of that authority/responsibility all the way down to the frontline agents: the project team. The specifics of this arrangement are typically shown through RACI matrices in most projects.
Complexity arises in large organisations because not only are there multiple principals and agents even in the simplest case: shareholders and board of directors, but there are many levels within the organisation below the board, each of them delegating to the next level down. Alignment must persist all the way down, or the principals could well be let down by agents acting on their own interest. This is one of the reasons why Portfolio Offices concentrate on benefits, not deliverables. Benefits is principal language, while deliverables is agent language. PMO (such as a Portfolio Office) must be the reliable translators between these languages, which must remain different to enable people to do their work.
In practice, the many PMOs at the various levels will deal with their own section of the principal-agent cascade. With the advent of P3O (more on that later) there is now wide acceptance that all the PMOs must be linked in a framework of governance, leading from the particular focus of projects, through programmes, up to the largely external outlook of a Portfolio Office.
There are two broad areas where PMOs can provide value in terms of corporate governance. One is working out and implementing the detailed cascade of delegation & decision-making (downward), and delivery & issue escalation (upward) that is necessary to maintain the whole project environment aligned. The other is in devising the least intrusive methods to detect and correct mis-alignment, while still ensuring that obstacles to delivery are removed from the way of project teams.
Typically, most of the effort of PMOs is expended on the latter area, and too much of the focus goes on trying to resolve the apparent conflict between support and assurance in the mix of services that a PMO provides. If PMOs took a metaphorical step back, they would see that the best way of adjusting the balance between support for the projects (agent) and assurance for the business (principal) can only be achieved if they have a good understanding and deployment of the appropriate governance framework for the organisation. It is because of the need to manage this framework in the context of corporate project management that the most valuable PMO discipline is risk management, not planning.
The challenge for PMO governance is that it needs to perform the basic function of governance — a consolidated view of the risks of change, and the diagnostic capability to provide insight that allows re-alignment — in the context of a fragmented project environment. PMOs need to take into account what the ultimate purpose of governance is, so that a given PMO can play its part in the overall governance of the organisation. A PMO also needs to be aware, in great detail, of the strengths and weaknesses that exist in the project environment of the organisation to which they belong.