One thing I don't see much of in the above posts other than Brian slightly touching on it is the effect of incremental development. What I mean by incremental development is the stuff most people can't see even though it's right in front of them. It is like the bureaucracy you see in the government or a bank or an insurance company -- the process is moving, but it's moving so slow you can barely see progress.
Think about it like this, that 75 to 84 rated hitting coach is likely going to get the job done for the season and keep you at $6 million mark. However, if you doubled that to $12 a season and now you have a pitching coach, hitting coach and bullpen coach in that 88 to 97 range. Is that coaching upgrade going to make a huge difference for one season? My guess is that it likely won't.
However, if you are able to have 88+ coaches more consistently, that's three to four spring training bumps where the better coaches might actually get you those extra 2 to 4 development points, over three to four seasons, and get your players closer to those ceilings. And, since you only have so much control over how well your MiL teams perform in the playoffs, I feel like the extra money spent on coaching behaves as a hedge if you're unable to get those really nice end of season promotion bumps when your MiL teams make it to the MiL WS.
Finally, some folks might think two to four points missing from a player's ceiling is not a big deal, but that two to four points could be multiple ratings and multiple players, so that math starts to add up. Can I guarantee the extra money spent on coaching is making an impact? No. Why? Because I can't take the players I have and re-run them through the development process with mediocre coaches to see if there is a major difference. So, I dont have data to back up my theory, but I also don't think anyone else does or ever will because nobody knows true ceilings and nobody can rerun the same players (as a control) through the development process with different coaches.
My point is that we should not be discounting the affect of incremental development.