They did have to do that, or the observed decay rate of SN1987A would be wrong. This is the same as the frame-rate argument that I got confused about. If you don't speed up decay-rate by the same amount as you speed up c then you won't see the emmissions from SN1987A decaying at normal modern speeds, but you do. I felt that speeding up decay-rate proportionately as c expands is just post-hoc-rationalisation to justify the observations, but MJN says not. I lack the physics to argue the point.
You're right. If they didn't need it in 1981, they have absolutely needed it since 1987 or they would have been falsified in 1987.