Many of those self-diagnosing features can be defeated by trivial circuits. A 555 timer for a downstream O2 sensor, a 4N4001 diode and a couple resistors for an EGR vacuum sensor hack, and several others.
I have had cars in the past that OBD2 reported as operating correctly, that easily passed smog/inspections, that probably weren't operating as the EPA intended.
But how many people with older cars have the initiative and the the capability to do that. It's a small fraction of a small fraction of a small fraction of vehicle owners.
Chasing a problem like that around is a waste of everyon'e time and money. Environmental regulatory agencies really need to do a better job considering the 90/10 or 80/20 rule (the last 20% of whatever you're optimizing will consume 80% of the resource).
At what expense/frequency, though? If it's just to cover the ocassional (unmodified) car that can't be made to report as "Ready and OK" per OBD2, it's just easier to use the existing repair exemptions process. (Spend a certain amount on repairs and get another cycle of inspection passage.)
If it's to cover modified cars, those are a trickier situation. It's easy for some to game the system so that it "passes" on the magic day and then fails the other 364 or 729 days in the cycle when it's actually being driven.
In either case, it seems like the expense of operating and staffing gas sensing stations, rolling road dynos, etc. Probably better to just turn a blind eye to that case as well.
I have had cars in the past that OBD2 reported as operating correctly, that easily passed smog/inspections, that probably weren't operating as the EPA intended.