Amazon.com Inc. is at it again. To the consternation of much of the book industry, the online giant is again offering digital titles for less than major publishers think books are worth. And this time, the price is zero.

If you own an Amazon Kindle, as opposed to just using the Kindle app on another device, and you also belong to the company’s $79-a-year Amazon Prime service, you can now “borrow” one digital book a month from the new Amazon Lending Library for free. You can keep the book as long as you want, but you can have only one at a time.

The new service worries Wall Street, too, because it increases Amazon’s out-of-pocket costs. The company is paying wholesale prices for some of the books in the lending library. For others, such as the titles from Lonely Planet travel guides, it is paying a flat fee for a group of books over a period of time. (It will report sales figures on individual titles back to those publishers.)

Beyond short-term earnings, however, the lending library is just the latest innovation to raise big questions about the whole publishing ecosystem. In an environment where books are increasingly digital, what’s the most effective way to create value for readers, for authors and for intermediaries? And -- the biggest question -- which intermediaries will survive the transition?

Big Six Balk

The lending library doesn’t include any books from the Big Six U.S. publishers -- Random House, Simon & Schuster, HarperCollins, Macmillan Publishers Ltd., Penguin Books Ltd. and Hachette -- because Amazon can’t control what it charges for their digital books. They are undoubtedly relieved to be excluded. But the pricing control they value so highly reflects rigid arrangements they may come to regret.

Amazon used to pay publishers a wholesale price for e-books, just as it does for physical copies. It set whatever price it thought best for its overall business, even if that meant losing money on an individual title in order to boost traffic or sell more Kindles. It could adjust prices up or down to reflect new information or offer special promotions. Its standard price was $9.99, which was often less than it paid for each copy. Major publishers thought that was too low, but most couldn’t do anything about it.

Then came the iPad and the accompanying iBooks store. Apple Inc. struck a different deal with publishers, known in the business as the “agency model.” Publishers set the retail prices, with Apple taking a percentage for its services. The Big Six liked that deal and wanted it to be the industry standard.

Amazon resisted, going so far as to remove all the physical books from Macmillan off its site in hopes of forcing the company to continue the wholesale arrangement. But that sales strike alienated Amazon customers, who were angry when they went to the site and couldn’t buy the books they wanted. Amazon blinked.

As a result, most of the big-publisher titles in the Kindle store now sell for $12.99 to $14.99 each -- a range Amazon called “needlessly high” when it capitulated.

I should say at this point that I am not an entirely disinterested observer. I’m an author, with two books available in digital form. And I agree with Amazon that, at $14.99, my 1998 book “The Future and Its Enemies” was priced needlessly high when its Kindle edition was released last spring. You have to either love me or your Kindle a lot to pay that much for a 13-year-old book you can get in paperback for $6. But, like Amazon, I have no say over how my e-book is priced.

Publishers, for the most part, don’t believe customers care much about the difference between Amazon’s old price and their new, higher ones. They’re skeptical that consumers respond to small price differences. A former publishing executive recently told me he simply didn’t believe that “if I really want a book for $9.95 I don’t also want it for $10.95 or $12.95.”

Look at Research

People in publishing say things like that all the time. While they admit that charging $100 for the typical hardback would be foolish, they don’t believe that changing the price of a book by a dollar or two will significantly change the number of copies sold.

The economic research suggests the opposite. In a 2009 paper that looked at consumers using computer price-comparison systems, or shopbots, to buy physical books online, economists Erik Brynjolfsson, Astrid Andrea Dick and Michael D. Smith found that a 1 percent drop in price -- a mere 25 cents on a $25 book -- increased the number of units sold by 7 percent to 10 percent. Shopbot users tend to be more price-sensitive than most consumers, but that’s a huge difference.

Publishers resist such evidence. The standard response is that it’s hard to know anything about pricing because “every book is different.” Every title is a unique good, and every customer values each book a little differently. So you might as well trust your gut.

Every book is indeed different, but that’s no excuse for charging more than the market will bear. And, at least for digital copies, there’s a way around the “every book is different” problem: bundling a lot of books together, charging a flat fee, and letting customers use whichever ones they like best. Bundling eliminates some of the statistical variation and unpredictability in consumer behavior. The differences from person to person and book to book cancel each other out.

“It is easier for a seller to predict how a consumer will value a collection of goods than it is to value any good individually,” wrote Brynjolfsson, who directs the Massachusetts Institute of Technology’s Center for Digital Business, and Yannis Bakos of New York University’s Stern School of Business in a 2000 Marketing Science article. They noted that, “at the optimal price, more consumers will find the bundle worth buying than would have bought the same goods sold separately. Because of the predictive value of bundling, large aggregators will often be more profitable than small aggregators, including sellers of single goods.”

Buying in Bundles

That’s why cable companies and video rental services such as Netflix charge subscription fees rather than trying to price each program or movie separately. Amazon’s lending library, which comes along with free shipping, streaming video and whatever else the company next decides to throw into an Amazon Prime subscription, is a move toward bundling digital books.

So is the flat-fee arrangement the company made with a few publishers and tried unsuccessfully to make with many more. Of the three pricing models, Brynjolfsson suggests, “the least popular with the publishers is the one that makes the pie the biggest,” maximizing total sales.

The pie gets bigger not only because the bundler can more accurately choose prices but because the bigger the bundle, the more valuable it is to consumers. Variety itself is worth something. For digital goods, where the cost of selling one more copy is nearly zero, Brynjolfsson suggests, the way to maximize both profit and revenue is to bundle as many titles as possible. (There may be a few exceptions, like the boxing matches sold on pay-per-view, that are extremely valuable to a small market segment and worth almost nothing to everyone else.)

Unfortunately, traditional publishing contracts don’t easily accommodate such bundles. Contracts are based on specifying in detail what cut of a list price per copy goes to the author. A bundle isn’t priced that way. Its price reflects its value as a package.

In theory, a book could make the whole bundle more valuable without itself selling a single copy. So should authors be compensated by how many copies they sell? Or should they get some kind of flat fee just for joining up? “When we start including publishers who have classic book deals that involve net royalties for e-book sales for authors, how will they figure out what to pay them?” asks agent Simon Lipskar in a blog post. It’s a tricky question.

Writing more specifics into contracts will only help until the next innovation comes along. “Things are evolving very rapidly,” says Brynjolfsson. “It’s a mess to try to pre-specify all the different structures and prices.” That’s the traditional approach and it has, he argues, made publishing arrangements “very brittle.”

Publishing Themselves

So brittle, in fact, that he and his co-author Andrew McAfee took their analysis to its logical conclusion with their most recent book, “Race Against the Machine.” Although a prominent publisher offered them an advance, they decided instead to self-publish through Amazon. The deal gave them a nimbler publication schedule and more control over pricing -- the ability, for instance, to offer a two-day sale when the authors are speaking at a conference. Now they’d like to join the lending library.

Why go with an intermediary when you can sell directly to the world’s biggest bundler?

Brynjolfsson cautions that authors won’t necessarily be better off if Amazon’s bundles replace more traditional book marketing. “The pie would be bigger (more revenues total),” he wrote in an email, “but they might not get as big a slice.” It’s hard to know in advance.

One thing is certain, however. Publishers are in trouble. They think their problem is that they are losing their retailers. But the real danger is that, over time, they are going to lose their authors as well. No wonder they are afraid of Amazon.

(Virginia Postrel is a Bloomberg View columnist. She is the author of “The Future and Its Enemies” and “The Substance of Style,” and is writing a book on glamour. The opinions expressed are her own.)

To contact the writer of this article: Virginia Postrel at vp@dynamist.com.

To contact the editor responsible for this article: Tobin Harshaw at tharshaw@bloomberg.net.