The difference, from an RF perspective, of an encrypted signal and a non encrypted signal is non-existent as the encryption takes place before the modulator. Therefore encrypted signals won't impact the noise flow for everyone in the band. This would only occur if the operator were using frequency hopping.
>you're still monopolizing part of a very small band with a private conversation preventing others from using the shared resource
Depending on the technology used in the receiver, this also isn't true. There are numerous techniques to allow for multiple carriers on the same band. CDMA for example would allow for this. It would take some coordination to make sure each carrier is using a different code but should be feasible for amateur radio operators. It all depends on how sophisticated their receiver is.
>Error correction necessitates a higher duty cycle
This is true and is why encrypted signals have lower data rates. So instead of having a longer transmission you lower the data rate to get the same effect.