Before peer review they were not dealing with the volume of scientific content that is submitted nowadays. One could debate whether citations are the only metric necessary, but there needs to be a filter before that even happens, or else we will be flooded with an avalanche of garbage research. It would be trivial for someone with deep enough pockets to order a network of junk papers citing each other and debunking climate change or evolution and make a real mess. As painful as reviewing is, it's a necessary evil for now.
Things might be ignored, but only after a person has already wasted their time looking at it and determining it is crap.
It was harder for cranks to publish back when journals were purely physical and a crank would have to come up with the printing costs himself. Now that anyone can publish for free on the internet, peer review is even more important for establishing what content out there is worth paying attention to and which is not.
Peer review happens anyway, but faster, and in public, without the insanity of revise and resubmit.
Journals are not where the action is in Economics or Computer Science. It works for them. Why not for everybody?
I neglect those because journals in my own field are both free to publish in and, more often than not, open-access. I do understand that not everyone is so fortunate, however.