When I wrote a post about e-book piracy recently, I was surprised to see an argument break out in the comments about the pricing of e-books for higher education. Even putting aside the most extreme remarks (that college texts should be free because the variable costs are zero — which, by the way, is not true) and wishful thinking (no reputable publisher would introduce a revised edition simply to cut off the sale of used books), the comments revealed a lack of understanding of how this market works. There is no other book market like it, and it is its very structure that makes college texts so expensive.
Before I continue, I want to share an anecdote. One of my first publishing jobs was as the college paperback editor for New American Library, which is now part of Penguin. NAL had a long list of Signet Classics, which included classroom versions of public domain classics — Dickens, Thackeray, Austen, et al. These books were solidly profitable, but we noticed that the sales of our Signet Shakespeare series had begun to flatten out in the face of new competition. So we made the decision to revise the books: new introductions and bibliographies, new covers, and in some instances changes to the texts in light of modern scholarship.
And that’s when I learned how college publishing worked. The head of our college marketing group came to me to say she loved the idea behind the new editions, but we had to be careful not to change the page numbers. How’s that again? It seems that many instructors used the same lecture notes year after year, sometimes for decades, and if you changed the page numbers, the notes would have to be revised. This could prompt some professors to switch to competitors’ editions.
The point here is that although college texts are created for students, it is the interests and sometimes the whims of the faculty that have first priority.
With that anecdote serving as background, let’s see why college textbooks are so expensive. To answer this, once again we have to jump into the time machine. We go back 20 years to a strange world populated not by dragons or telepaths or sea monsters but by a now-vanished set of accounting policies for the amortization of capital expenditures in the publishing industry. At that time, publishers of textbooks generally estimated that the useful life of an edition was about five years. There were different ways to get to that number for accounting purposes. Some publishers took the cost of developing a book and spread — amortized — those costs in equal amounts over each of five years. Other publishers estimated the number of copies they expected to sell and then charged their profit-and-loss statement on a pro rata basis each time they shipped a book. There were and are many variations on how to do this, and no two publishers had the exact same rules, and sometimes the rules even within a publishing house would vary discipline by discipline or even title by title. But you get the idea — take the cost of developing a book and spread the cost in some logical manner over the lifetime of the book.
Why spread the costs? Why not simply take the hit all at once at the date of publication? This is how many trade publishers operate — they charge their P&Ls with a book’s plant cost (the one-time cost to create the first copy) on the date of publication. They do this in part because a trade book sells so many copies at the time of publication that the plant cost can easily be absorbed, and in part because you never really know how many copies a trade book will eventually sell. But college publishing is different — few titles sell in the quantities of trade books, the plant cost of a college text is much higher than in the trade, and college texts don’t have the bulk of their sales at the date of publication. So college publishers spread those costs to pair them with the useful life of a book.
We are not talking peanuts here. A widely used industry estimate for the average investment to create the master copy of a college textbook is $750,000. This is much much higher than in the trade, where only superstar authors drive the cost up to that figure and beyond. If you think you can sell 25,000 copies of that book, that gives you a per-unit charge of $30 — and that does not include paper, printing, and binding. This is one of the reasons college texts cost so much.
But why spend so much money developing a textbook? Has calculus changed that much over the years? The high cost of development grows out of the peculiar way instructors select books. Let’s not forget that instructors do not pay for the books themselves; students pick up the tab. So Professor Jones reviews a selection of new books on economics; which one does he choose for his class? He likes the idea of online tutorials (more expense), he likes the integration with a suite of reference tools (more expense), he likes the simulations of the world’s economies (more expense) — and he wants the very best for his students, so spare no expense. You could argue that a publisher could hold the line and decline to add all this expensive baggage, but then Professor Jones will simply adopt the more expensive and fully featured text from the competition. The crucial point is that Jones is invited to make choices (called adoptions) that other people have to pay for. This is another, the principal, reason college texts cost so much.
Our time machine is working its way forward and comes to a time when a group of companies figure out how to ship used college texts around the country relatively inexpensively. They build databases of all the courses taught at every college and the books used in those courses. They then approach college bookstores and say, You can make more money selling students used books than you can selling new books. And it’s true — there is often more margin to the retailer on the used book at $70 than for the new book at $125. So college bookstores begin stocking more and more used books and highlighting them to students, eroding the college textbook market. By some estimates the market share of used books is now about one-third of the total market.
So we have the publisher — that crook, that snake in the grass — who has invested $750,000 against the expectation that they would sell 25,000 copies, but now the sale has dropped to 15,000. The used books, in other words, are the economic equivalent of book piracy — legal, but economically challenging. Now the per-copy charge for development is up to $50. This is another reason college texts cost so much.
Now our time machine has returned to the present day and we find that the old policy of amortizing books over five years has been changed to three years as a means to recoup that $750,000 investment — because the used book market is now a very real competitive threat. So the per-unit cost of development rises again. This is another reason college texts cost so much.
It will be apparent to everyone that the higher and higher the costs and prices go, the greater the opportunity to sell used books, as they can easily undercut new books on price. It’s a vicious circle.
The high cost of college texts would be lowered in a minute if college instructors had any financial responsibility for the cost of textbooks. Of course some instructors are very diligent about watching prices, and some institutions have made it a goal to lower textbook costs. But the average instructor? In the battle between moral imperatives and economic incentives, follow the money.
This is why college textbooks cost so much. It is not the way college texts will be marketed forever (there are chinks in the wall opening up now), but it is an expensive way to present baseline information to undergraduates. And, gosh, ain’t those publishers positively satanic?