Well the thing is, PESA doesn't cost that much more over MSA. It is just that the West has built a large installed base with MSA, and there was no point in moving that to PESA.
With the Soviet Union, they prefer to go cheap, since MSA requires a delicate and precision mechanical servos, which is not something they can easily mass manufacture. So they went to the Twist Cassegrain, which combines the classic Cassegrain design (same principle as a reflector telescope) but uses a flat focal plane design on the end which uses polarizers to "twist" emissions towards different directions for a scanning motion. Thus the Twist Cassegrain combines the characteristics of Cassegrain with that of a focal plane array. Mainly for the latter, if you are receiving signals against a flat array it will be subjected to less distortion as opposed to focusing it to a recieving point. The antenna used in the Su-27, J-11s and even the MKKs is of this principle.
Since they bypassed the MSA stage, the Russians is just skipping from Twist Cassegrain to PESA. At least with NIIP Tikhomirov. With NIIR Phazotron, they decided to go with MSA, seeing it cost effective with MiG-29 and MiG-21/23 upgrades. and with that launched the Zhuk line.
For China, due to the extent they were aping the West, they moved from parabolic like in the early J-7 radars, to Cassegrain, like the Type 204/208 used in the J-8II, then to MSA. China actually had the opportunity to skip MSA entirely since they don't have the legacy hardware like the West does. And you can see China has managed to develop PESA quite fast, given the KJ-2000/200, various ground AEW radars and fire control radars for the HQ-9 and KS-1A. Yet they stuck MSA on their fighters as it seemed like a safe and "traditional" choice to do.