Purity - BSA is a purified protein, while milk contains a complex mix of proteins, lipids, carbohydrates, etc. The additional components in milk can potentially interfere with antibody binding in a Western blot.
Consistency - The composition of milk can vary depending on the source, while BSA from a reputable supplier will have consistent blocking activity.
Reduced background - Milk proteins can sometimes cause high background on Western blots, while BSA tends to give lower background. The complex mixture of proteins in milk likely leads to more nonspecific antibody binding.
Optimized blocking - BSA is optimized for blocking in blotting applications, while the proteins in milk are not designed for this purpose.
GTPase specificity - Milk may contain GTP-binding proteins that could potentially interfere with detection of the GTPase of interest. BSA does not contain GTP-binding motifs.
Binding capacity - BSA has high binding capacity and can effectively coat the membrane to reduce nonspecific binding. The diverse mixture of proteins in milk reduces blocking capacity.
BSA is better for blocking than defatted milk powder, however it is very expensive in comparison 100Euro for 100g vs ~5 Euro for 500g.
You can get non-specific interactions, biotin which is endogenous in milk makes using streptavidin conjugates impossible and it's possible to have poor background.
BSA is not optimised for blocking, it's just available and relatively cheap compared to other proteins and functionally useful for many scientific purposes, e.g. a carrier protein, where you have very little of your valuable protein and need something to "carry" it through processes to avoid loss (like sticking to the side of your tube).
You could use casein equally well.
Binding capacity of defatted milk powder is not a problem either, you just need to have enough.
Supercheap defatted milk powder is unlikely to have been properly defatted and this can cause problems with background. However, don't use NIDO as it has all the fat in it still!!!!