I see your point - you're right. For 'perceived' effort, the best metric is your brain, which perceives the effort.
The problem with this is that your brain can perceive efforts quite differently with different situations. You may be redlining in a racing paceline but so caught up with focusing on the pace and rotation that you don't perceive the effort as that hard, which sets you up for a big bonk/dropoff in the later part of the ride. Similarly, you might think you're going hard on an early AM workout, but are really not doing the workout much justice since you're not fully awake.
The PM measures output - you're right it doesn't measure perceived effort one bit. It's a more reliable, more objective number, but for sure you'll have days where for the same power output you may have a perceived effort of "5-6" versus "9-10" depending on your fatigue and training load. They really look at different things, but in most cases, the objective PM numbers will be more reliably reproducible and objective than the much more variable and subjective RPE.
I've found HR to be a pretty decent surrogate for a PM in the right circumstances (steady state training) - it's sort of in between the objectivity and precision of PM vs RPE.
Interestingly, all 3 are useful together as they all measure slightly different things. And also interestingly, you can absolutely get a lot faster with none of them.