I came across the following discussion of the optimal royalty structure for an author.
The agent is really high on The Unicornians. She thinks it's the next Twilight. So she submits it to several editors at once. Editor 1 comes back offering $300,000 for three books. Editor 2 offers $30,000 for three books but with a significantly better hardcover royalty. (Say, 20% instead of 10%.)
Putting aside the (very important) questions of which editor would be a better fit and which publisher is doing a better job with Unicornian-esque books, I would argue that the author of The Unicornians is always better off signing with Editor 2.
Let's say that The Unicornians is not a tremendous success. The first book in the trilogy sells 8,000 copies in hardcover; the second two sell 6,000*. With Editor 1, the author gets her $300,000^^, but The Unicornians comes up $240,000 short^^^ of earning out. With Editor 2, the author only makes $80,000 on the series, but $50,000 of that is royalty, and the publisher has also made a (modest) profit. The publisher will likely ask the author for another series, perhaps something focused in on the werewolf dude...
Okay, so now let's say The Unicornians IS successful. Let's say the first book sells 250,000 copies in hardcover**, because they make a movie, and teens squeal about how hot the unicornian boy's horn looks. The second and third books also sell 250,000.*** With Editor 1's deal, the author earns back her advance and makes $1.2 million, for a total of 1.5 million dollars. With Editor 2's deal, the author earns out and makes $2.7 million in royalties, for a total of $3 million.****
Really Long & Boring Post about Book Advances and Publishing and also see
Book Advances and Marketing and the Cart and the Horse
A few comments. First, if an author has some indication that his book will sell as in the first example, he will make much more money. In fact, it would take five trilogies to earn this much money. Even if that publisher is made at you, the author proved he could sell 14,000 books, and so the next publisher should be interested in giving them another contract, even if it is at the much lower advance level. You also could take the money and give away your next book series on your website. This probably is not an important strategic situation because publishers have more expertise in understanding the commercial viability of the books than their authors do. After all, the authors specialize in writing books while publishers specialize in publishing and marketing them.
Another comment is that economists have discussed optimal contracting. It comes up in labor economics, game theory, and IO at a minimum. The optimal way to split uncertain revenues depends primarily on two factors, the relative risk aversion of the participants and the way that work quality can be observed. A few examples will help.
Let's say that you sew buttons onto suits. Since that can be observed perfectly the optimal contract is simply to pay a flat fee, say a $1 for each one you sew. When you work harder you make more. Slack off and make less. The lesson here is when you can contract over observable quality and productivity it makes sense to have employees bear the risk.
Now say you have a job where no amount of testing will predict how good you are at it, just that experience shows that 50% of people selected will be good at it. Presume also that your effort doesn't matter either. If the employer can employ hold a large number of such employees, then they can diversify all their risk, having little or no risk in aggregate production. In this situation it is optimal to pay employees a flat rate. Employees have to take a single job, so they can't diversify away from the risk of being bad at the job they take. The employer has no aggregate risk. If employees are risk neutral or risk loving, they'd take high pay if they are revealed to be of high quality and low pay if of low quality, but that doesn't describe most people. The lesson here is that when there is uncertainty in the productivity because effort cannot be clearly observed or there are sources of risk outside of the employee's control, those risks are best born by the people who either do not mind risk or can diversify that risk, which is usually the employer.
Now back to authors and publishers. Having a large advance provides a smaller incentive to write a better book than a larger royalty does. Once they give you your advance only your artistic integrity, reputation, and desire to be given more work in the future motivate you to write the best book possible. Your advance size shouldn't motivate you because it is a sunk benefit. That is, it is fixed based on your effort. On the other hand, a larger royalty provides less of an incentive for publishers to promote your book because they get a smaller share of the upside. There is residual risk that is born by the publishers because they front all the money to print and market the books. Giving them less of the upside but leaving the costs the same means that they should be less willing to take a risk on publishing your books.
This doesn't have a clear general solution. You want the most upside to write the best book possible, but you want to leave as much upside as possible to the publisher to ensure that they sell as many of your books a possible. A publishing house is diversified, so they are in a natural position to bear more risk than you are. That suggests that they take more royalties and you take a larger advance. However, we know it cannot be optimal to give the author no upside, because the advance alone would provide limited motivation to write a good book.
Publishers seem to distribute their marketing budgets in a lumpy way. That is, some books with small advances get large marketing budgets. This suggests that publishers can (or believe they can) distinguish much more about commercial viability after the book is written. That alone may be enough to motivate authors if they have a utilitarian value from seeing their book sell well. Alternatively, if they have at least some royalty payments then writing a good book means making more money even though the advance is a fixed payment because it raises the likelihood that publishers will promote your book. That suggests a compensation structure like the one we observe in publishing business, a combination of effort monitoring and risk sharing in the form of advances and royalties. Economics doesn't have anything obvious to say about the relative sizes of these parts without knowing more about risk preferences, the probability of commercial success, and the role of marketing in book promotion. It is however a tractable problem. If we knew the odds of a book being a commercial success dependent on author effort and publisher's marketing budget then we could solve for the optimal contract of profit sharing.
Hat Tip to Boing Boing at Proposal to raise book royalties, lower advances