<p>Basically, HYS get the best of both worlds with SCEA. It looks friendlier and nonbinding, so more people apply and they have a better idea of who has their university as a first choice. In addition, they suffer only minimal consequences from the nonbinding nature; those schools have no real competition anywhere because the student elected to apply there in lieu of its main rivals. For example, if you really want to go to Yale, there's no strategic advantage in applying early to Harvard or Stanford instead. It's not any easier to get into than the other two, and no matter what decision Harvard and Stanford gives you, you're probably going to pick Yale in May (provided you get in).</p>
<p>Contrast this with a school like Brown. Don't get me wrong, I'm not a Brown-hater, but I would acknowledge that when offered a choice between Brown and a HYPSMC (and friends), a fair percentage of students would go for the latter. Brown has less to gain from SCEA because if a student is able to gain admission into both Brown and some HYPSMC, it is much more difficult for the admissions officers to decide where the student's priorities lie. In other words, it lacks the same luxury of being the absolute #1 best. ED makes it much easier for students to do so, and this explains the other Ivies resistance to (SC)EA. Early action programs outside these 6 universities open the university up to more loss, so in the end it becomes a game for the individual school to decide which program will get them the most applicants AND the most stable yield rate.</p>
<p>Contrast this again with MIT's and CalTech's early applications. The admissions officers who read the application can't really tell where they are on the student's priority lists (ergo CalTech's need for a "Why us?" essay). Granted, they can tell with limited certainty (see next paragraph) that HYS aren't #1 on their lists; otherwise they would most likely already be bound by the respective SCEA agreement of that school and not an early applicant. </p>
<p>Finally, HYS have the advantage of a relatively small group of top schools and the fairly clear-cut preference for them over America's other universities (e.g. most students would pick Harvard over Georgetown). If you're scared of your #1 school's early program's restrictiveness, you might opt for a different one, so while someone who wants to go to Stanford but feels the need to apply EA to a few schools may do so, the other schools (or the student), not Stanford, lose out in the long run because the student will either get in regular and say no to the other schools or get rejected regular (regular admissions can be more selective because those admissions letters "cost" more in terms of yield). As all the ED-using schools but Princeton tend to fall lower in the hierarchy, there's not much of a probability of a student applying (SC)EA to their #2 or #3.</p>
<p>Side note, why won't Princeton, MIT, or CalTech use SCEA? I don't really know. They could just be stubborn or cautious, and I expect this is the case, at least for MIT and CalTech. Princeton, however, has been criticized in the past for using strategic admissions techniques considered unethical in order to increase yield (i.e. during regular, rejecting many students considered likely to be admitted to another elite school). Whether or not this is true, if Princeton WERE afraid of being 2nd choice to other universities, 99%-yield ED is a better option than the comparatively unknown yield of SCEA, though I don't really understand why these fears would exist at a place like Princeton. Whatever.</p>