I would not necessary agree with you on that when you have a line to ground fault. One reason why is the there tied together at both end forming a single conductor path. Not saying single conductor, Form a Single Conductor Path. So the line to ground fault would be feed to the fault from both ends.
Thus the wire type EGC in that raceway must carry this total amount of fault until the OCD opens. What is key is open time and Conductor Withstand rating.
Another way to look at maybe is the impact on the faulted conductors in that raceway. We already know that set is no good and has to be replaced. what about this effect on the other three sets. To small of an EGC could damage the other three sets. I say could because we do know the open time and the withstand rating of the EGC. I have seen this happen and the wire MFG would not cover cost of replacement of the second set due to the EGC was undersized and damaged the conductors in that raceway. Long story however relevant. Long story short the OCD opened due to an install Error leading to a ground fault in the first set, It did its job. what happened next was the EC said Um and turned up the break to max blaming the trip on setting. It was not a load issue it was a ground fault on that 480 volt feeder. The EGC was sized as you mentioned ignoring the engineering. He thought he knew better than the professional that designed it and went with sizing the EGC from the OCD for each run. It was up sized for voltage drop from the MDP underground to a remote panel in the same building, pretty typ. They simply nicked the line conductor and warped it up with a scotch pad/ tape and finished pulling, It was dug up at the fault. The withstand of the EGC was not is spec and melted to the other conductors and damaged them. We know because we dug them up and cut section of raceways out per engineers request. This was not our install. I was called in because the local EC needed help. By the time it all went bad the EC that did the job was out of warrenty and said not his problem so the building owner called out local EC. I recommended before we proceed call first EC and then up the chain of command after noteing conductor size did not match plans and the breaker had been re set and adjusted after the Gear provider had done final settings.
Based on that I can see why it does have to larger than the underground conductors. The example you listed if on a 200 amp OCD would be 4 to 1 and than the not larger rule applies. I say bad design. This would be in case of a underground PVC run. Above ground EMT, IMC, RGC, I would be fine with it. I would also think about the available let through fault current and the open time of the OCD for a EGC with stand rating and make sure it will open with out damage. If I see a red flag I RFI engineering for a ruling.
SO what happens when some one works around the rule. I am sure you asking where is he going by know. What if we design a 2 to 1 for the same install. 400 amp OCD for the same load. Granted if a panel and it was rated for 225 it would now be 400 amp. I have seen this as well and recently. The engineer designed around the rule for a long run with a 180 amp load at the 2 to 1. The EGC works out right under 1/0, so instead of 3/0 you can use 1/0 EGC and be done per NEC. The time open works on the 1/0 for the withstand rating. Better yet I seen it designed with GRC underground and no wire type and EMT above ground with no wire type. I would be careful when you choose to go against what written unless you understand the outcome in a what if. Sometimes putting that green wire in can be a bad thing if not installed correctly.
Just my take on it anyway. May be all wet and need more schooling. An engineer may help with this question
Quick withstand rating on #6 and 1/0 @ 1/2 cycle
#6 =15,213 amps
1/0=61,223
Which one you want in that pipe with a let though of 42K