In my experience most RCEs are necessitated by the Examiner applying new prior art in a second Office action and making it final. The applicant not having seen the new art then needs to further amend the claims, but the Examiner will not permit the Applicant to do so after final.
When I was an Examiner, I was taught to read and understand the specification before scoping out the field of search, and to search the disclosed as well the claimed invention. IF Examiners understand and search for the disclosed invention at the outset, then the best prior art will be found at the time of first Office action. If this is done, the Examiner can have a high level of confidence that no matter how the Applicant amends the claims, the closest prior art is already in the record. Then, no matter how the Applicant amend the claims, a new rejection based on new prior art should never be necessary. Claim amendments not supported by the spec can be rejected on 112 1st paragraph grounds. There should never be a need to do a different search in response to an amendment.
In my view if we can eliminate or avoid new prior art rejections in response to an amendment, the need for RCEs will be reduced dramatically if not totally.
The best and 100% effective way to accomplish this is for Examiners to read and fully understand the specification, and search for the DISCLOSED invention. If prior art is found that teaches the DISCLOSED invention, you will reach closing of prosecution immediately.
In this regard I propose that Examiners be given credit for writing a summary of their understanding of the disclosed invention and search reflecting that understanding. This will be beneficial for both Examiner and Applicant as no Applicant wants to waste time and money on an unpatentable invention.