The biggest problem with "what we know" is what we don't know.
SAE net rules require the ENGINE to be tested with the accessories its delivered with. But we don't know what actual configuration ANY engine is tested with, or if it was ever tested at all. as most engines were made available in multiple models or multiple trim levels, there's often some combination that doesn't have PS. but the issues don't stop there.
the ITCS includes cars from 4 decades and at least 7 countries on 3 continents. those care were often tested to local standards and the numbers converted to US hp using a "standard" equation. but there's little verification available and the test standards were all out of sync. there's a lot of disagreement in the source materials for the older cars. from 1971, US market, domestic manufacturer and many import cars are tested to SAE net - so theoretically WITH the offending PS pump. BUT there was no requirement to post the hp figure accurately (many cars were actually UNDERRATED, largely for insurance or to protect the image of a halo car, among other reasons) and there were loopholes. 2005 on, US market cars should be tested to SAE certified hp and that clears up a LOT of confusion, and puts the engine in the configuration as delivered in a specific make, model, and trim. remember when a bunch of toyotas and hondas were suddenly derated in or around 2006? yeah. they were testing engines with non production ECU calibrations, lighter weight fluids, etc... i.e. "loopholes". GM went from calculated hp (they didn't even test to SAE net, apparently) and FOUND hp when they tested. in some cases, they found a bunch of it.
so there are 8 model years of cars where the HP numbers are no suspect, though many OEMs only went to SAE certified for models released DURING that time (nissan comes to mind). IT rules require the car to be 5 model years old, reducing that window to 3 years, and reducing the new model count substantially. I don't have a count of the number of cars that use SAE certified HP in the ITCS, but it's not a big number.
Yes, we KNOW that the info we work from is suspect to a degree. so we consider things like "known to be..." and if we have a good reason to worry, can suggest a higher or lower multiplier than the standard 25% over stock assumption.
the system is imperfect, but works pretty well regardless. the assumptions about a hp number being tested in accordance with the same standard across the board and in the configuration we are interested in are founded on real standards although they can, at times, be a bit optimistic. Maybe luckily, they do seem to have worked well. like so many other things, the active cars in IT follow the pareto principle, so the front running models are pretty well understood and we know we have their numbers very close to ACTUAL power output in "maxed" IT trim. but most cars in the ITCS are not very well known, still at their initial process classification, and could be grossly over or under their ideal weight once someone puts the effort in to finding out. throwing in another variable is not desired.
as for allowing de-powered steering on those spec lines where a manual rack is also allowed, I understand where you are coming from and agree that it seems like a simple solution, but there's always the possibility that it IS NOT, and unintended consequences have already created more allowances than we might actually want.