Yes, they could "possibly" do that, but they aren't, so why the hell is this relevant?
Harvard admitted 2,200 undergrads in 1988. In 2018 it admitted 2,000. It took me 30 seconds of googling to find this.
It is absolutely true that if you do something for group X, you could have done more for group Y instead with the same resources. But saying that doing something for X automatically takes away from group Y is arguing the counterfactual, "Here is what you would have done if you had not done that." But you don't actually know what they would have done. Maybe they would have reduced admissions even more than they actually did. Can you prove otherwise?
That may seem like a rhetorical question, but it is not. Discovery gives access internal communications which could make that case one way or the other. But it is a case that has to actually be made, and isn't simply a foregone conclusion.
It is generally possible to expand enrollment by paying money to construct larger dorms, open classrooms, and hire staff. Virtually no university in the country is anywhere near its theoretical physical capacity. If I give an institution a large grant earmarked for doing this to create slots for a particular minority, that institution can create those slots without changing what is available to any other minority.
Now I provided a hypothetical case of a large grant from a single individual. That's an admittedly unlikely scenario. But Harvard's affirmative action programs are one of the wonderful things that can be talked about to potential donors. How many were pushed over the edge into donating because of it? Particularly given the fact that a former minority student who went to Harvard was in the press a lot over the last decade. Being the first black President will do that.
The point remains. The fact that one group got something doesn't mean that another group had to get less.
For another area where the same type of fallacy is rampant, consider immigration. People object to immigrants because "they are taking our jobs". When in fact immigrants make a net contribution to our economy, and make a net contribution to how many jobs are available for existing workers.
The way that they did the quarter system reduced breaks and filled the campus year round. The result was an increased number of hours of classroom time. Yes, this was more work for the faculty. But faculty is cheaper than buildings, and the primary mission of the school is teaching.
They did not double the hours. But in the early years, they also had a pretty lopsided male/female ratio. They kept their promise.
Of course, this solves nothing. There are still a finite number of slots, the number has just changed. The question of how to allocate the slots remains, and there are still more applicants than slots.
For a concrete example, consider how Dartmouth College went about admitting women back in 1972. The campus was already full. Yet they committed to not reducing how many men were admitted, with no new construction, and without increasing class size. In short they said that the new women would come at no cost to men trying to apply, and little cost to the university itself (certainly less than the benefit of the tuition.)
They succeeded. How? They moved from a semester to a quarter system. Then arranged to have required second year courses that were only available during the summer to get rid of any stigma about "summer school". These changes let them go from the usual part time academic calendar that most schools follow to operating at capacity year round.
The result? They met their apparently impossible commitment. And the new slots opened for women indeed came at no cost to the existing population of men.