Under most normal circumstances, using reverb on the mix bus is no different to using a send on every track, with every send set to the same level. Usually this it not a good idea – it’s better touse sends to apply reverb in different levels to different tracks. Some sounds can ‘take’ more reverb than others. Some sounds need more reverb then others to emphasise the depth in the mix. A send level of 0dB (unity – meaning the reverb is the same level as the dry sound) might still be not enough for sustained sounds like pads and organs. On the other hand, a send level of -21dB might sound extremely wet for staccato sounds or hand percussion.
Having said that, there is a place for mix-bus reverb. While it’s not as refined or tailored as using individual sends, it is much faster. I’ve done it myself on occasion when I’ve had a project that’s up against a hard deadline. Mix-bus reverb also sounds different to individual sends when it’s placed after other mix-bus processing, such as compression or other dynamic effects (for example, NOT eq). Whether this sound is useful for you and worth the greatly-reduced flexibility is up to you.
Reverb in mastering is a slightly different matter. In this situation it’s too late to adjust the reverb in the mix, so it can only be applied to the stereo mix. Reverb may also serve a slightly different purpose when used in mastering – to make all the songs in a release have a similar ambience. This might be particularly important on compilation albums or albums with a wide variety of sonic approaches.