Harvard won.  Well, let me back up for a second and give some background: for the past few years there’s been a general movement toward the idea that early admissions is a Bad Idea, because it disadvantages low-income students.  If you’re admitted early under binding early decision, you’re committed to attend mostly regardless of the financial aid package offered — and the college’s definition of need is often rather different than the family’s.  So, to undo this gross injustice early admissions had to be removed.  That, and arguments that the entire process had simply gotten too complex and needed to have decades of accumulated cruft removed have successfully swayed Yale, Princeton, UVA, and Stanford among others away from binding early decision.

I sense astroturfing.   Let’s take a look at some of the background: Harvard, one of the most selective colleges in the world and routinely regarded as among the most prestigious, has the highest yield rate of any of its peers — by nearly 10%!  With a 79% yield, 8 out of 10 students accepted enroll.  When a student is accepted to Harvard as well as other schools, Harvard almost always wins.

As much as students are applying to college hoping for a spot, colleges are hoping for the class they want.  Truly top students are in demand as much as the colleges themselves are — and it’s no surprise that colleges look for tactics to maximize their chances of assembling a star freshman class each year.  How can each college offset some of the lost acceptees to more desirable colleges?  By early decision!  If you commit to a college early, you’re taken out of the pool of eligible students for Harvard, or some other upstream school, to steal.

That the movement away from early admissions started at Harvard should be obvious now: they’re the only college with no fear of, on average, losing students in the cross-admit war of regular decision.  Everyone else has a slightly bigger fish to fear, even if the prestige difference is only a few percentage points.  The “it disadvantages low-income students” argument is especially interesting coming from Cambridge because they used non-binding early action, not binding early decision.  Harvard’s (effective, whether or not it was on purpose) goal was to push the subject into the national consciousness, thus pressuring its peers who did use binding early action out of their positions.  And it worked!

Of course, Harvard might have meant everything they said and been totally sincere in their cause of moving away from early decision.  The observation that the move advantages Harvard’s class makeup is still true regardless.  Whether or not they meant to, they’ve made a deft strategic move.

Now, it’s important to mention that early decision, in general, does have a lower percentage of low-income applicants relative to regular decision and early action, due to the whole legally binding regardless of aid bit.  And, it does strictly disadvantage the students who aren’t in that pool… ipso facto, low income students are disadvantaged by early decision.  But why this is true is a little bit more subtle than is immediately obvious.  It’s up to the colleges what they do with the early pool: the fact that they admit such a disproportionate amount of their class from this more limited set is a separate issue from the makeup of the pool itself.  It comes back to yield — since colleges’  perceived prestige is so dependent on their yield rate, they have a huge incentive to admit lots of students from the pool that’s almost guaranteed to have a 99% or higher overall yield.  In game theory, this is called a “move of assurance” on the student’s part.  This, by the way, makes the fact that Harvard has such a high yield even more striking, since they never used early decision!

So, to be clear: early decision is the conscious trade-off by the university to take a guaranteed student now, or see what they’ll get in regular decision — a game that sounds strikingly similar to the one the students play.  It’s because the universities care so much about their yield and relative prestige that yield becomes an overriding factor and causes such a large percent of the class to disappear early.  Of course, if you don’t accept at a much higher rate early, then the incentive for students to lock themselves up as they do is gone, and the entire thing doesn’t work.  Thus, if early decision is to be used it must provide an advantage, one which necessarily cuts out students who can’t commit to the risk.  Removing this, and making the market more fair — where students will tend to settle at the best school they’re (mostly) well fit for relative to everyone else applying — happens to be against the interests of almost everyone except Harvard.  Interesting games here.