That's for the smart guys, the engineers, to answer. I know that #12 used to have a #14 wire for an egc and they felt that it did not always clear the fault.
I do not see why you are confused. I have issue with this rule and obviously knew very little about they way others interpet it. To simply state increase the breaker to 40 amps means you did not fully read my rant.
No doubt in my mind you , me and others will run into an issue where the breaker may need to be downsized because of unforseen circumstances and using cables greater in size than #10. No longer can the contractor throw in a larger wire because he has it on the truck , other consequences will prevail with this interpetation.
I will empisize the other forum posts where it is claimed that Voltage drop upsizing is not required. So better to use #8 NM breaker at 40 amps when you only need 20amps on a long run .
If you folks want to interpet it this way then I think there should be an added caviot regarding distance , impedence or something because this as written is non-sense. This rule has a place but not as broadly defined as many of you have opined.
The problem is the rule is written to cover all installations, yet not all installations have the same conditions.
I think most of us would agree if you had a circuit only 100 feet long and you increased your conductor sizes by just one size or maybe two because of voltage drop that leaving the EGC the same size would likely result in fairly reliable enough low impedance for ground fault protection. There will be increased impedance and slower response time from the OCD but you are still fairly well protected. Take that same size circuit and increase it to 1500, 2500, even 3200 feet and now you not only have the voltage drop issue requiring larger circuit conductors, but you also have a significant amount of impedance in the EGC if you don't also increase it in size. But there is likely many more 100 foot circuits installed than there is 3200 foot circuits, yet the rule is written so it applies to all.
I see this in some of the irrigation work I get involved in. Many long circuits involved in some of this work. I have seen many line to line faults that occur at the far end of a circuit that end up burning open a termination, or switch contacts or something else before opening overcurrent devices. Impedance over the distance of the lines limiting the amount of current that flows is the main reason. Less impedance means more current flows which means overcurrent device responds to the condition quicker.
We have the opposite problem at the source end when determining if the gear has high enough interrupt rating, and if not high enough, sometimes just making conductors longer adds enough impedance to lower the available fault current to a level below the interrupt rating of the gear.
What works for you in one place can work against you in another.