This is a simple communication issue. Really what you're asking is application information.

I had a similar issue when I was working with someone else and they were wondering why I needed a specific part to vary a pulse width. Come to find out, that the answer that they were expecting was not one that I could have thought of on my own. The application engineer needed to get involved just to say that it was meant to specifically effect a sensor and that the sensor worked off of different pulse widths and would be tuned to these pulse widths. Although, I knew this information, it did not occur to me that this was the information that they were after. I thought they wanted to know more about the schematic and how specific blocks of the circuit worked. This came down to a point that the guy asking the question was just wondering if they could simplify the overall scheme by using a finite pulse width.

This scenario is no different. Let's just say that the 12 bits is used to tell the motor it's rotational speed. In this case, that might mean that the rotational speed may not rotate at an acceptable tolerance if the DAC only provided 4 bits. This really is not a problem to be solved, but more like art that needs to be formed because something may functionally be gained (but losing something else) by using a lower bit DAC without even knowing until the final prototype is built.

I do not know the OPs true application information as that was not greatly detailed, but in effect that's what you're asking about. I guess another way to ask would be what functionality is achieved between a 12 bit DAC and an 8 bit DAC? I do not believe that there is ever a "need" to go with one or the other, but whether the desired effects of choosing the part give designer the desired artwork.