There is only one explanation for the cost of textbooks: greed! How this plays out is interesting, however.
When I took calculus in 1955, the text cost around $5 new and I bought a used copy for half that. It was around 300 pages in a 6" x 9" format, at a guess. The last time I taught calculus, the cost was pushing $100, the monstrosity had close to 1000 pages and it was in a giant format. I got a bad back just lifting it.
One observation is that the people who choose these texts get their copies free and invariably have no idea what they cost. Once, choosing a text, I wrote to publishers of 4 possible texts, enquiring about the price. One replied that that information was proprietary (!), one told me the price and the other two ignored my request. I chose the second.
How are these texts chosen? Well, for single section courses, the instructor usually chooses them. In ignorance of the price, as I said. In large multi-section courses, they are usually chosen by a textbook committee. Each person on the committee has his own pet topic that he insists of being in the textbook chosen. One will want some problem in mechanics, another on financial mathematics and so on (here I am picturing a calculus book). The textbook publishers understand this all too well and their solution is to include any topic that anyone might conceivably want. That is part of the reason for the bloat. My personal feeling is that none of this belongs in a calculus book; if an instructor wants to include a particular topic, write some notes on the subject and distribute them. The plain fact is that the calculus courses are generally too crammed to include any of it anyway.
The current crop of books have a couple of four color illustrations (enough to be able to claim “four color” on the cover) that are utterly useless, loads of graphs (some of which actually are helpful) and giant margins that the students are expected to make notes in (no Fermats here) and, incidentally, render the book unresellable. Next year, there will be a “new” edition that will correct some of the errors, to be sure (result of careless editing), but are otherwise changed only enough to render it awkward to use the older one. The main technique, since calculus has not changed in the past 100 years, is to rearrange the exercises so that the instructor cannot assign problems by number unless everyone has the same edition.
The only solution I see would be to bypass the publishers entirely. Assign one of the professors the task of writing a text in return for some course relief (it costs no more than $5 or $6 K to hire a part timer for one course) and then make that the official text for the next five (or ten or twenty) years. Such a book can now be printed-on-demand and bound at the local copy shop for under $20 and the department will readily recoup its investment in a year or two. Who knows, other schools might even adopt it. Why hasn’t it happened? Damfino.
Prices will beome unlimited when the person who prescribes is not the person who pays. Can you think of another situation like this? Now physicians cannot be expected to formulate their own drugs, but professors can and certainly should take back the publishing of textbooks from the publishers.