I recently finished The Knockoff Economy: How Imitation Spurs Innovation by Kal Raustiala and Christopher Jon Sprigman, as part of my periodic attempt to educate myself on productivity issues. I decided to tackle to my to-read list from the bottom rather than the top for a change, and since this book had been sitting there for at least two years, it was the winner. It’s very good and clearly written, and a lot of their arguments have become more mainstream since it was published in 2012.

There are many interesting tidbits throughout, but one I found very intriguing came at the very end, in a coda not closely related to the main substance of the book. Here is an excerpt of the relevant section:

Conventional thinking about innovation and IP relies on the concept of a rational innovator. It assumes that innovators calculate, either explicitly or implicitly, the cost of creation versus the size of the return they will likely enjoy. A writer might anticipate a certain advance from her publisher; a musician might estimate the sales of a new song. This expected return shapes how much effort they pour into creation and what kinds of creation they pursue. Abundant research in economics and psychology, however, suggest that their judgments are often likely to be wrong—and systematically so.

As many studies have found, individuals are very bad at assessing their own future prospects. They have a pronounced optimism bias. They think they will succeed where others have not, and they heavily discount the prospect of failure. Nearly all newlyweds, for example, believe they will not get divorced, when in fact a large minority will—and often within a few years. Likewise, students wildly overestimate their likely grades, even in the face of stiff competition. Like the residents of Lake Wobegon, we all want to believe we are above average. … Optimism bias, in short, leads many innovators to think they will gain a greater return from their intellectual creations than they actually do.

Why is this important to understanding the interaction between copying and creativity? Because optimism bias likely acts as a subsidy for innovation. Creators who have an unduly strong belief in their ultimate prospects for success should be willing to invest more in their creativity. And this increased willingness to invest is likely, in turn, to lead to increased creative output as compared with a world in which creators rationally calculated the odds—odds that may include expected losses from copying.

There is another important, and related, factor that skews how innovators assess their expected return on innovation. Many contemporary markets for creative goods are what economists call “winner take all” or “tournament” markets. In these markets, a huge reward goes to a few at the very top—the superstars—while much less goes to those just below them. This dynamic is easy to see in areas like professional sports: just think about Major League Baseball, where the very best players receive enormous salaries, while those who are merely excellent languish on AAA farm teams, earning a tiny fraction of what the true stars do.

Tournament markets amplify small differences in performance into enormous disparities in reward. Given this basic dynamic, we might expect people to shy away from competing in markets like these—the risk of failure is great, competition can be very intense, and the difference between success and failure hard to determine until years of effort have been invested. Yet we see large numbers of individuals competing to become a sports star, a national politician, a CEO, or, most important for our purposes, a musician, writer, or inventor of the next huge Web concept. Many markets for creative goods are tournament-like. A hit song can yield huge sums for the right creative artist. Yet the vast majority of songs go nowhere, commercially speaking. Likewise, books and screenplays can rake in enormous revenues if they are truly successful, but New York and Los Angeles are awash in the tens of thousands of authors who tried and failed. …

Like optimism bias, tournaments induce more investment than is rational. So both optimism bias and tournament markets push innovators toward high levels of innovation. …The important point is that both of these effects exaggerate anticipated benefits. And it follows that exaggerated expectations of benefit will tend to keep innovation buoyant, beyond what a rational calculation of return would predict.

In other words, pursuing innovation to some extent depends on having irrational expectations about the future. To me, this ties very neatly into David Graeber’s bureaucratic theory of technological stagnation: he argues that the more research is driven by corporate bureaucratic practices that require precise estimation of the outcomes and benefits of said research, the less innovation actually happens. This contention directly challenges William Baumol’s idea of the “free-market innovation machine”: that sustained rapid economic growth is possibly precisely because innovation can be turned into a routine organized activity, and does not have to depend on random flashes of insight.

It does seem clear that once innovation becomes a routine corporate activity with a budget and cost-benefit analysis, then that cost-benefit analysis should be more accurate than what individual people do in their heads with all the usual cognitive biases and errors. But if the incentive to innovate depends on systematically over-estimating the potential benefit of innovation, then doing an accurate cost-benefit analysis will not in fact be a good thing. It may lead to less wasted effort at the individual or firm level, but could also mean less innovative activity in aggregate.

To be clear, this is not at all what Raustiala and Sprigman argue–to the contrary, they think that the irrationality of innovation in fact makes it more resilient to bad regulation or intellectual property-rights violation, and therefore is a reason to be more not less optimistic about the future of innovation. And I don’t completely believe the bureaucratic theory of technological stagnation either (Baumol is a genius and more likely to be right than Graeber). It is worth thinking about though.