Decisions should be made at the lowest level where all the relevant information is available. In this case, the executive understands the pressing reasons for making the change, so is properly able to weigh the incremental risk (with the technical advice of Shirley et al) vs. the business priority, whereas Shirley can only see half of the picture.
Imagine if Shirley had gone ahead and broken policy, then something went badly wrong. She likely would've been blamed for the whole fiasco.
This is what I was looking for in this entire thread. It's one thing to push through a small change in a small internal program for something that isn't critical like a batch job that clears old data from a server. That's not what this is.
It changes the potential output of the entire company. Yes, it's a single LOC totaling out to a single byte. Yes, it passed all of the tests. However, tests are written by people, and tests are only as good as the people who write them. This is partially why there are human eyes in the QA/IT/UAT phases. Shirley would have been viewed as the single point of failure had something go wrong, because she made the decision to break policy.
Talk about policy and it's politics and escalation and such. Talk about principles and it becomes a moral issue.
Alternatively, think of it as breaking a specific law (such as the speed limit) vs being immoral in a more general sense of the word.