Well, if I were optical engineer/designer, I would say it is a design compromise made somewhere along the process.. Internal iris is obviously not at the optimal place within the system as far as optical distortions are concerned.. Maybe they did not want it to be too far away from camera end, to keep the controls from camera simpler (it is quite complicated mechanical system already..), and/or to keep the smaller size of iris and maintain the lower cost.
http://www.mir.com.my/rb/photography...nses/200mm.htm
Interestingly, my Tair11-A has iris placed immediately after front lens group, but it is 2x larger in diameter, and it is totally manual (that gave me idea to try something similar, by the way), see here:
http://www.zenitcamera.com/archive/lenses/tair-11.html
It is quite obvious that the front group of lenses creates the problem. With external diaphragm, only the central area of lens is used, so CA is much more tolerable.
With internal diaphragm, it seems the peripheral parts of the front lenses are engaged, and hence exaggerated CA.
Whatever was the cause of this hassle (CA), the problem is solved, it seems :-)