as a managing editor, I can tell you that they do not incur substantial expenses, and that academics provide the important parts of the service, essentially for free in the cases of most journals. It's not like putting out a magazine; we didn't even have copy or layout editors for our journal, the most inexpensive components of editorial labor. It paid the university department that hosted the journal a mere thousands (single digits) per year. There were two "paid" staffers—myself and one other person, The rest of the "editorial board" consisted of faculty of our and another several universities doing the work for free, under the auspices of the "professional duties" of the academics involved (not as paid by Springer, as paid by their respective institutions). Peer reviewers—free. Editorial labor (copy, layout to production files according to specs, submissions queue, even rough line editing, style work)—graduate students looking for a title to add to their emerging CVs.
Essentially Springer's total cost for putting out the journal amounted to the several thousand (again, single digit thousands, split between myself and one other individual) that they (usually belatedly) paid our department annually for the entire journal in its substance, plus printing/distribution (a pittance given the circulation size of academic journals and the cost per print subscription—not to mention the increasing number of electronic-only subscriptions). They had one liason that handled our entire "account," and the level of labor involved allowed this person to be "over" several _dozen_ journals as just a single person. That's as much a labor footprint, in its entirety, as our journal actually had inside the "publisher."
And for this, they held onto the reprint/reuse rights with an iron fist, requiring even authors and PIs to pay $$$ to post significant excerpts on their own blogs.
Seeing the direction the wind has been blowing over the last half-decade, the department decided (and rightfully so) that it's basically a scam, that academic publishing as we know it need not exist any longer, and wound down both the print journal and the relationship with Springer several years ago, instead self-publishing the journal (which is easy these days) to much higher revenue for the department, and the ability to sensibly manage rights in the interest of academic production and values, rather than in the interest of Springer's oinking at the trough on the backs of academics.
Oh, and many university libraries (particularly in urban areas) do not admit just anyone off the street; you must generally hold an ID that grants access to the library (often student or faculty, plus a paid option for the general public, either monthly or annually, that can vary from somewhat affordable to somewhat expensive). Not to mention that for many people, yes, it is a significant professional hardship to lose a day or two of work to be trekking into foreign territory and sitting amongst the stacks—and that this hardship is made much more irritable by the fact that the very same articles are sitting there online, in 2013, yet can't be accessed at reasonable cost.
As an academic, I have the same frustration. We bemoan the state of science in this society, yet under the existing publishing model we essentially insure that only a rarefied few scientists and the very wealthy elite have access to science at all. $30-$60 is not a small amount for the average person—and that is the cost to read _one_ article, usually very narrowly focused, and of unclear utility until they've already paid the money, that is borderline unreadable for the layperson (or for the magazine author hoping to make sense of science _for_ the layperson) anyway. Why, exactly, would we expect anyone to know any science at all beyond university walls, under this arrangement?