The RAW file itself has no understanding of a colour space. It is truly agnostic in this respect. Any colour setting for a RAW *in camera* is purely informative not decisive.
However, once we move the captured data into anything treating the data as an image, the manner in which the data should be interpreted in terms of colour becomes important. The camera is capable of recording colour information outside of any existing colour space. In the processing stages, depending on what use the photo is intended for, we might want to avoid clipping and losses of extreme colour(s) while *editing*. Some loss is inevitable though and more later when the image is printed or exhibited. If we keep the wide gamut space as long as possible, transforming into other spaces more suitable for printing etc. can occur under full user control and utilising the best possible mapping into the final oiutput.
