I don't know if there are that many drugs in the category of "proven to be safe but just not proven to be effective", so I'm not sure how if this proposed change would do anything.

My guess is that the threshold for proving that a drug is "safe" is probably much more stringent than the threshold for proving that a drug is "effective". I wouldn't bother taking a drug if cured fewer than 5% of people who take it, but I would probably want to avoid taking any drug that kills even 1% of people who take it. If these were my effect sizes of interest, it would be much cheaper to run a trial to demonstrate efficacy than to run a trial to demonstrate safety. So, if we retained a requirement for manufacturers to prove that a pharmaceutical is "safe", but we removed a requirement for manufacturers to demonstrate efficacy, it wouldn't reduce trial costs by very much at all. 

I think we can have a strong prior that most vaccines are in the "safe but not effective" category because we have a pretty solid understanding of all of the biological pathways involved, but this doesn't hold for every type of pharmaceutical.