You said "Quick questions, on what do you base that assumption?" I can't directly answer your question because YOUR QUESTION is wrong. IT'S NOT AN ASSUMPTION BUT AN OBSERVATION. The observation over many years of doing it both ways and any difference (I assume there must be some overhead that would make it theoretically slower) are so small as to be undetectable with tens of thousands of inserts per day. Both methods yield an instantaneous result as perceived by the human users brain.

You said "Do you have a realistic test data and harness to proof the point?" Real world usage by thousands of users

You said "Is there a problem calculating the hash on insert?" Yes. There are numerous (unknown number and unknown type) ways users can input the data so we have no control over what is insert so the only way to get a hash is with a calculated field or running a process after an insert.

This is the last supersize in futility I am going to offer. Anyone else please have the common decency to directly answer my question instead of offering alternatives that make stupid assumptions.