The equation typically normalizes stress values to the unstressed section so that comparisons can be made. Obviously when the part necks down, as a malleable material might, the local stress will be far higher than reported but, since the design decision is based on the unstressed section, there isn't typically much use for the local stress calculation.
stress is defined as F/A, without limit to elasticity or plasticity; the issue is "what area to use?".
UTS is an "engineering" stress, F/Ao (Ao is the initial undeformed area)
as distinct to "true" stress where A is the deformed area (eg the necked area of a tension coupon).
"true" stress is only used in labs and schools.
another day in paradise, or is paradise one day closer ?
It depends.
In some cases, you have some code that specifies an allowable stress and you calculate the actual stress and compare it to the allowable.
In some cases, you have some code that specifies nominal strength- you calculate actual load, multiply by some factor, and then compare that to the nominal strength of the item.
In some cases, there isn't any kind of code or standard involved. In that case, you can use either of the two approaches above and assume the appropriate factors.
Or you can look into theories of failure and evaluate that way.