If the best data you have is conjoined, then you must either figure it out, or ignore it.
Do you know what percentage, exactly, are retired vs "giving uo"?
No?
Then when someone says full employment, they're right unless you have specific and precise data to counter that.
We could also discuss people too sick to work. Or on maternity leave. Or wealth vs desire to work. Or 100 other factors.
That's odd. The burden of proof is not exclusively on one side.
Then yes, you have to start the dance. And work at it. And convince the entire planet to change.