First, I must point out that I know very little about plastics and molding and I'm very much aware of that myself, so I'm not at all trying to play an expert here. However, the last 7 years, I've worked for an automotive supplier who primarily manufactures plastic parts. So, with my interest in RC- and static plastic models, I've taken the opportunity to speak with some of the specialists at work.
Choice of type of plastic to mold parts in is not just about the cost of the plastic itself. Tough and reinforced plastics cause a lot more wear on the molds, both because they have to be injected with higher pressure and because of their abrasive effect. That isn't so relevant for (relatively) small volume manufacturers like Associated and Schumacher, but it certainly is for a high volume manufacturer like Tamiya. As pointed out here already, Tamiya use high strength plastics in their TRF-models, but not just because they are more expensive and sold to more demading customers, but also because mold wear isn't so relevant for these low volume models. For high volume models however, mold wear is of course an issue, so if Tamiya had used for instance fibre reinforced nylon throughout in the DT-02, DT-03, TT-02 and similar high volume chassises, the cost of repairing, maintaining and replacing molds would have been significantly higher and eventually handed down to the customers.
Also, takt time is a factor. Generally, high grade plastics demand a longer takt time. The longer it takes to mold (heat, inject, cool, eject) a part/sprue, the longer the machine is occupied for a production run and the higher the labour cost, energy cost etc. Furthermore, shrinkage, warpage and general tolerances are critical issues that Tamiya need to have under control and some plastic materials are much more demanding than others on these issues, resulting in higher cost for monitoring the processes and for corrections whenever the tolerances start to drift.
There surely are many other factors too, but just the above is in my humble opinion enough to feel confident that Tamiya are very much aware of their choices and that they leave very little to chance. That said, I don't excuse Tamiya for poor designs though. Meanwhile, Tamiya have decades of experience making RC-models of relatively fragile plastics, so they should know pretty well how to design parts to make them suitable for their specific tasks, like for instance required stiffness and tight tolerances for a gearbox housing, balance of stiffness and toughness for suspension arms, gloss, homogeneous colours and crispness of decorative parts and not the least the required wall thicknesses and dimensions. And I dare say that the design flaws are not so common anymore, and Tamiya know pretty well how to cleverly design (plastic) parts to compensate for their use of "brittle" or soft plastics.
Also, outside Japan many of us may feel that Tamiya models are (sometimes) not so good value for money, but in Japan, the prices for Tamiya models and spare parts are pretty low and parts availability is probably not even remotely as good in any other part of the world. So Japanese customers possibly think they get good value for money and when a plastic part breaks or is worn out, replacement is cheap and easy to get. And after all, Japan is still Tamiya's most important market by far.
And think of it, has there been any new Tamiya chassis the last 20 years with as severe design flaws (causing breakage of plastic parts) as those commonly found back in the eighties and into the nineties? I can’t think of any. The Juggernaut was an extremely poor effort for so many reasons, some of which were also related to poorly designed plastic parts, and some later chassis designs have issues, but if I’m not mistaken, none of which are directly caused by poorly designed plastic parts? (Not counting issues that affect handling).