One reason anyone would be resistant to your theory is that there are many more reasonable possible explanations for these data points.
Perhaps in the year of demarcation, more students started taking both SAT and ACT. Some states mandate the ACT so it’s possible that was a year in which more states introduced the requirement, bumping up the percentage in a way that has nothing to do with a change in Bowdoin’s reporting practices.
To the OP: a 33 is highly unlikely to move the dial one way or another, but Jeff Selingo’s Friday article on MIT’s reintroduction of testing requirements has some useful insights on the submitting vs test optional decision.