It’s all about headroom. As you know, 0 DBfs is as loud as you can be: anything above is distortion. In the analog world, you didn’t have a hard ceiling, just increasing distortion. You wanted to get your level high enough so it didn’t sound noisy but low enough so it didn’t contain too much distortion. Many analog audio recorders set peak levels at 3% distortion which gave them around 60 db noise floor. They set their “0” to a certain magnetic signal strength (like 186 nanowebers per meter) to fit audio between distortion and noise.
Since a Digital Betacam VTR samples 20 bits in audio, your S/N ratio is very high. Subtracting 20 db won’t hurt at all.
While many broadcasters will accept tapes with incorrect reference levels, you should use the proper reference level of the machine you are dubbing to. It is a mark of professionalism to set proper audio and video reference levels. -20 db can be used for all broadcast formats.