There is after all a strategic benefit to this; getting your talent for cheaper. Consequently a deal like this b/t direct competitors is harder to get off the ground. Also as an optimization problem, there are many different constraints that come into play at different parts of the curve. When you saturate the employment market with this behavior the market would likely seize up.
There are a lot of people (of varying degrees of foolishness[1]) who want to work at Google or Apple but wouldn't even consider the companies you listed.
[1]Prefer working at Google to Oracle? Fair enough. Rather work at Apple than Microsoft? Ha, have fun with that.
No. Those companies don't have a great reputation among tech people, but have a lot of good people working for them. Therefore they don't push prices up much, and are a great source of good candidates.
If that's the reasoning, why are FaceBook and Twitter not included? They have a good reputation among tech people, a bunch of engineers working for them, and (in FaceBook's case at least) they push salaries up a lot.
The particular companies involved make me think this really is about good partnership relations - until Android, Apple and Google had a really good relationship, Apple and Pixer share a head honcho, they all buy from Intel, Intuit's not really a competitor to any of them, etc. The effect, of course, is to drive down wages for the employees who would otherwise have been poached, and that's why the DoJ got involved. But I doubt that execs at each of those companies were telling themselves "We're going to drive down wages for our key employees" when they made the policy.
I believe that the companies in question would not have liked having a similar agreement with Facebook and Twitter. But it was not in the interest of Facebook or Twitter to agree to it. Therefore no meeting of the minds was possible.
I also believe that collusion was made much easier by the existing good relationships you refer to.
Where we may part ways is on what the execs were saying to themselves. I don't think it was just about maintaining good relations. I believe it was about reducing the cost of having to replace key institutional knowledge. It is worth skipping out on raw talent (that will need training) if it results in keeping key people.
Which brings us back to my point. Microsoft, Oracle and Yahoo do not have good reputations among tech people. So they are not major threats to losing institutional knowledge. And they are a great source of candidates. So it isn't in the interests of Apple, Google, etc to extend the agreement to those companies.
Kind of the same problem that kept it from happening with Facebook, but with the roles reversed.