My "argument" is simply to state that U3 is a measure defined so that it can rely on population (not sampled) data. This is done precisely to reduce the amount

My "argument" is simply to state that U3 is a measure defined so that it can rely on population (not sampled) data. This is done precisely to reduce the amount of processes error and bias. The definition of U3 leaves little room for adjustment, and that is intentional. If you were worried about "manipulation," then U4/5/6 are much more open to manipulation than U3 is.