For example, with current digital cameras, a monochrome sensor like Red Dragon has actual higher resolution than a color Dragon, because no Bayer pattern is needed. I am not sure if the same is true of chemical monochrome film?
This here is likely the reason, as transferring color film will still require the Bayer filter, but the B&W will not. The other answer may be a difference in the quality of B&W vs Color film stocks (especially in the 70's).
Good points. I hadn't considered about the camera being used for the scan. The SOLARIS restoration I saw is billed as a 2K restoration, which may go further to your theory here. If the restoration had been done in 4K, or at least the scans done in 4K, the 30-40% loss of resolution from a Bayer camera would not have been visible in a 2K projection, like I saw at Cannes. But since both the restoration and the projection were done at 2K, that same 30-40% Bayer penalty would in fact be noticeable. I guess we will have to see what resolution the film was actually scanned at for sure until this theory could be confirmed or refuted. I guess it's always possible that the elements were scanned at 4K and the DI and restoration work were done at 2K. The Youtube video about this new MOSFILM restoration should provide some answers.
I spoke to a friend who is very knowledgeable about film stocks. What he told me is that the black-and-white film stocks used today actually date back to the 1960s, so not much has changed with B&W film stock. Color stocks, on the other hand, have improved dramatically. This might help to explain why some of the B&W shots looked like they could have been film recently, while the color portions looked dated, grainy, and low resolution. He said that back in the 50s, 60s, and 70s, B&W was in fact sharper than color stock, for a number of reasons.