The modern seismic codes such as EN1998-1 recommend confined boundary elements at the end of the cross-section of shear walls (see for example the paragraph 5.4.3.4.2 of EN1998-1). Generally speaking the confined columns at the end of the cross-section of shear walls improve the ductility of these elements. So, the shear walls with confined columns at the end of their cross-section perform better than the shear walls without confined columns at their ends.
You always have to confine a RC wall. It could be with edge reinforcement or columns/beams. What you opt for depends on a specific situation/demand. You have to use dowels and starter bars (conservative Eurocode request) in columns/beams solution.
I would just like to add to what Konstantinos very comprehensively mentioned, that EN1998-1 also allows the use of (non ductile) lightly reinforced walls if their cross sectional length is large (>4.00m) and under some conditions.
In such cases, it is anyhow probably more difficult to achieve a ductile behavior of the wall by implementing any kind of special detailing on the reinforcement, thus the confinement at the ends of the cross section can be avoided (other kind of mechanisms of dissipating energy are taking place, such as the rigid-body rocking etc).
The shear wall having confining boundary at ends perform well as compared to shear walls without confining boundary. It is provided to increase the deformability of shear wall.
I agree with Dr Kostas Morfidis and other Colleagues that the boundary columns improve the behavior of structural (shear) walls. Also , they increase their torsional stiffness .
I agree with opinion of research colleagues that shear walls with properly confined boundary columns improves the lateral load resisting capacity of shear walls.
What if column, being part of a whole frame system, is overstressed due to the presence of shear wall? (column, as mentioned, is at the end of shear wall, but poured separately).