Clarifying Boards’ Relationship to CEOs’ Interpretations

By John Carver and Miriam CarverDecember 22, 2011 | Print

(First published in the November-December 2011 issue of Board Leadership, available to subscribers electronically at publication.)

“Our board worries that the CEO might change her interpretations of board policies without letting us know, so we expect her to inform us of any change right away when it is made. We can’t tolerate waiting a year or quarter or month to learn of interpretations at regular monitoring time.”

John and Miriam Carver understand boards’ anxiety, but find that this solution reveals an important misconception about a board’s relationship to its CEO’s policy interpretations.

Can a Policy Governance board wait a week, a month, or longer to monitor its organization’s performance? Indeed, can it wait to know the CEO’s interpretations that lead to that performance? Given that boards have no choice but to wait some amount of time in each case (because instantaneous notification of performance or the interpretations that lead to performance is virtually unattainable), for a given board, how much delay is tolerable?

With regard to performance data, this question is addressed in Policy Governance by a monitoring schedule that requires periodic demonstration of the degree to which the organization is performing as measured against a reasonable interpretation of Ends and Executive Limitations policies. But the issue raised here concerns not performance data but the CEO’s interpretations that ostensibly aim the organization at performance.1

Policy Governance provides no specific method for a board to check interpretations alone. It does, of course, allow the board to know anything it wishes under the rubric of “incidental information.” The worry expressed in the quotation at the opening of this article, however, seems to concern more than mere knowledge, but opportunity for the board to take action when it acquires the knowledge. Since incidental information is not actionable in Policy Governance, what is a board to do if it is worried about interpretation considerably prior to a regular monitoring point—whether due to a new interpretation or environmental or other circumstances that render an existing interpretation obsolete?

The answer lies in the monitoring frequency itself. Policy Governance enables a board to set whatever periodicity it wishes, for whatever reason it wishes. If the board wants to see CEO interpretations more often or earlier, it can simply demand more frequent monitoring. If the board’s concern is only about a limited time—for example, early in implementation when comfort with the system’s reliability has not been established—then it might demand more frequent monitoring only during the first year or so. In any event, the board can accelerate what it would have considered sufficient monitoring had there been no concern about pre-performance interpretations.

In other words, the board factors in its concerns when setting or adjusting monitoring frequency. If the resulting monitoring frequency causes a report that comes too early for the desired performance to have occurred (as would be the case in the opening quotation), the board may receive a monitoring report that has interpretations but no performance data at all. The CEO’s interpretations would include what he or she contends is reasonable at this time, given his or her interpretation for complete fulfillment. (Although being done for a different reason, there is a similarity here to monitoring expectations that take a long time to achieve. See “How Best to Govern Long-Term Ends Accomplishment,” in this issue of Board Leadership.)

Monitoring in Policy Governance is about performing reasonable interpretations rather than about making them. But for practical purposes, one has to go through the interpretation, so to speak, to make a judgment about performance. In the case considered here, it is important for the board to remember that zero performance can be a reasonable interpretation, but, of course, it must be justified as such. So a pre-performance monitoring report would likely not have, or be expected to have, performance data.

This approach addresses a board’s pre-performance interpretation concerns by using the standard Policy Governance monitoring format. It is therefore subject to the monitoring rules and the discipline that is built in. When a board receives a report, it can rule that the entire report demonstrates neither performance of a reasonable interpretation nor even a reasonable interpretation itself.

However, having recognized that boards using Policy Governance have access to pre-performance interpretations, we are compelled to note the possible downsides boards should avoid.

One is that it’s very easy for the board’s pre-performance focus on CEO interpretations to morph into an approval of those interpretations, so that interpretations not tacitly “approved” then become implicitly disapproved. If this happens, the CEO’s latitude to perform any reasonable interpretation can deteriorate into extending board policy into a lower (more detailed) level, in that the CEO becomes accountable for a certain interpretation of the applicable policies. The board does so not by amending policy as it has the right to do, but by allowing the integrity of de jure expectations to be spoiled by de facto expectations thus created.

An associated trap is for the board to come to rely on steering organizational performance by operating directly on the CEO’s interpretations. It may then not be as compelled to be sufficiently careful with policy language when adopting it. In other words, sloppiness in policy creation—not getting the wording right to begin with—becomes more likely since the board can maneuver the CEO’s interpretations of that language. This is a road back to the kind of ad hoc governance the board had intended to put behind it when adopting Policy Governance.

Furthermore, the board should be mindful that monitoring imposes a cost. Monitoring reports are not free; they displace other important uses of staff time and focus. There is also a cost to governance, for the board must divert effort from other parts of its own job. These are not reasons to forgo needed monitoring, but a caution to be circumspect about what monitoring is necessary for the board to be prudent and lawful.

Although Policy Governance does permit a board to address its pre-performance anxiety, we believe two practices will reduce the worry. One is for the board to openly and explicitly test the range of policy language. We mean that board members and executives together brainstorm the various interpretations that might reasonably be made of policy language under consideration. This is not a decision-making exercise, so mixing the functions of governance and management is not a problem. Neither board nor staff is committed to anything, but the board gets valuable information that it can use to keep or adjust policy language. Moreover, beside that immediate effect, there will likely be a larger payoff in the board’s comfort with the controlled permissiveness of the “any reasonable interpretation” rule, as well as the CEO’s competence in using the range accurately.

The other practice—one that should be in place whether or not the issue is interpretation—is that the board remain obsessively consistent in its application of model principles. Policy Governance, because it is a system, is not an approach that tolerates patchy implementation. Even slight inconsistencies can cause safeguards built into the model to weaken. For example, if the board allows CEO accountability for organizational performance to be watered down or the authoritativeness of board policies to deteriorate, every feature of the model will be diluted, including the board’s confidence in powerful but limited CEO authority.

Finally, we recommend periodic board review and, if appropriate, adjustment not only of policy language but of monitoring frequencies. In the case of a monitoring frequency previously increased to disclose preperformance interpretations, the board should occasionally check to be sure the solution has not outlived the original problem.

* * *


1. We assume for these remarks that (1) there is a CEO (although our points hold good whether or not there is one) and (2) that the board has created Ends and Executive Limitations policies to the depth or degree of specificity that enables it to promise the CEO that accomplishment of any reasonable interpretation will be considered successful performance. We may, for brevity, simply refer to “board policies” when we mean only these two categories.



Rupert Murdoch, in answer to an MP’s question, “Do you accept that ultimately you are responsible for this whole fiasco?” Hearing by the Culture, Media, and Sports Committee, House of Commons, London, July 19, 2011.

Why Wait?

Get the current newsletter and
Sign up to receive exclusive content and special offers in the areas that interest you.
Copyright © 2000-2014 by John Wiley & Sons, Inc. or related companies. All rights reserved.