Now, how would you tell if overall productivity went down?
The problem at hand is that we measure "overall productivity" via the proxy of "the number of articles a panel of other scientists considered interesting enough". Everyone who has dipped a toe in academia can tell you how bad that is at even remotely correlating with producing actual value.
But, the actual value of a piece of research often comes to light decades after it has been first published. Rutherford has been at it since 1911, but we only got first value out of his theory in 1940s. Man-millenia have been invested in the meantime. Would this be any better, if all of them had to spend half of their time working on securing funding?
The problem at hand is that we measure "overall productivity" via the proxy of "the number of articles a panel of other scientists considered interesting enough". Everyone who has dipped a toe in academia can tell you how bad that is at even remotely correlating with producing actual value.
But, the actual value of a piece of research often comes to light decades after it has been first published. Rutherford has been at it since 1911, but we only got first value out of his theory in 1940s. Man-millenia have been invested in the meantime. Would this be any better, if all of them had to spend half of their time working on securing funding?