Also a keen observation. This is probably the root of COH2's balance problems. How could it/should it be resolved?
I don't think it can, at least not without major changes in CoH2. They have taken steps already to try and mitigate it through patching, which is part of why the current meta feels quite unrelated to the original design statements for USF/OKW. USF lategame got more consistent with changes to Jackson (although it's still not great), OKW late game got a bit nerfed, although it's still very good. OKW isn't as weak as most people think in the early game. Sov / OH have fluctuated wildly over patches about whom was stronger at which stages of the game. Further patching could slightly tweak this but it's a bit late to shut the gate once the horse has bolted.
Resource inflation in team games is a natural consequence of cache sharing and map's being flooded with more and more units to make sure everything is capped. People have proposed solutions to this before. One would be to disable cache sharing. Another is to make caches only "partially share". A third is to make it so caches don't share, but you can build one cache per player on each point, so that OKW wouldn't ever benefit from the resources, and additionally the OH / Sov / USF players would all need to build their own caches and thus all pay the manpower price. Mechanically that would require some srs patching from Relic, methinks, but it's not a bad idea conceptually.
Games going longer in team games is a regression to the mean type thing, one player failing early in a 1v1 means gg, 1 player failing early in a 2v2 means its "probably" gg but could be recoverable, in a 4v4 its bad but not uncommon to recover gracefully. Other players can cover for build weaknesses (you almost never get an "oh shit nobody built an AT gun" moment although it can still be badly positioned), other players cover for you getting driven off the field early, etc. The effects of individual failures get averaged out and because there are so many bodies per square feet on a lot of maps, it becomes really difficult to exploit small breakthroughs unless you've REALLY romped them. Even games where one side is clearly winning sometimes drag out for quite a while in big team games.
The lessons are more valuable for designing the sequel, imo. In CoH3 the need to seriously think about the differences between game modes both economically and tactically when designing factions, and attempt to differentiate them in ways that do not create asymmetry in the problem areas I've pointed out.
Obviously the other main area they need to have a long hard thing is the urban / enclosed vs open maps, and how they can try to minimise the impact of bad maps on balance without just making every faction a carbon copy. It's definitely a design challenge!