This week, the Large Hadron Collider (LHC) opens back up for business. Important business, to be sure, but expensive – the LHC was originally completed in 2008 at a cost of about $9 billion.
That (relatively) high price-tag is illustrative of an ongoing trend: radical innovation is getting more complex, and thus more costly, over time. In reading the innovation literature and economic history, I can identify 3 likely, primary reasons why:
First, there is the very simple fact that the knowledge frontier is expanding over time, inching forward with each new innovation.
Since innovation, as opposed to imitation, is closer to the knowledge frontier, it requires a higher level of schooling. In the 9th century, discovering a better configuration for the horse harness probably required relatively little education; in 2015, top R&D teams are comprised almost exclusively of PhDs.
We can observe this effect empirically, too: Northwestern economist Benjamin Jones has assembled outstanding data showing that the age of first innovation has been consistently rising over time.
In essence, subject-mastery takes more schooling (and then, often, more professional experience) today than it did yesterday, and that means higher wages for each individual inventor.
Teams of Inventors
The expanding knowledge frontier has another effect – increasingly, innovation occurs in teams. When innovators face an expanding knowledge frontier, a compensating mechanism is to narrow individual specialization, which consequently forces innovators to work increasingly in teams (and with team size growing over time).
So not only are individual innovators being paid more over time, but they are increasingly working in teams (of increasing size), thereby multiplying the total human capital cost of innovation.
Further compounding these increasing human capital costs is the increasingly combinatorial nature of new, radical inventions.
In the mid-19th century, one of the most crucial advances was the invention (and then refinement) of Bessemer steel. An impressive feat to be sure, but one that required only Henry Bessemer himself, and his (extensive) expertise in metallurgy.
Today, the most impressive technological leaps forward require cooperation among masters of many different disciplines. For instance, IBM’s Watson did not merely require an expert in AI, or even a team of experts in AI – the invention required a team of experts across diverse fields, including linguistics, information retrieval, machine learning, and hardware. The story is similar with regards to Google’s development of a driverless car – there is no driverless car expert, the way there was a metallurgy expert in Bessemer. Instead, the technology has come together as a collaboration among a wide range of experts: in software, hardware, digital mapping, laser design, and of course automotive engineering.
Because of this need for cross-disciplinary work, the human capital costs of radical innovation are compounded one last time. Not only are we paying individuals more, and not only have the individuals become teams, but we actually now need teams of teams working across multiple disciplines. (Admittedly, not every radical innovation henceforth will be cross-disciplinary in this way, but many of the highest-reaching ones certainly are).
Fixed Capital Costs?
Finally, I would tentatively posit one additional reason for the increasing costs of innovation: my own casual observation suggests that fixed capital costs are going up over time, and are unlikely to cease doing so. Whereas Galileo was able to build his own telescopes, modern astrophysicists require highly complex and expensive tools like the BICEP telescopes. The taller the shoulders we stand upon, the more expensive the tools are to get there. This line of reasoning makes intuitive sense to me, but I have yet to see an empirical study on the matter (although I would be unsurprised to find one).
In any case, no matter the precise mixture of causes, it is apparent that the costs of innovation are bound to rise over time. This has potential consequences not just in terms of the rate of innovation (which we may cope with by the increased wealth generated by innovation in the first place), but also in terms of who gets to innovate.
The individual inventor was probably always a myth, but certainly is so today. Moving forward, the big question on my mind is this: will the entrepreneurial firm, with resources far above most individuals but far below most large incumbent firms, follow the same fate and lose its place as a primary source of radical innovation?
Multidisciplinary Approach to Economic Success
1 Million Cups City Spotlight: Des Moines