NIST’s “substantially fueled” premise for unleashing innovation

Here is a claim from the opening of a recent NIST report–1234–on “Unleashing American Innovation,” a “draft green paper”:

The U.S. innovation system is substantially fueled by the discoveries and inventions arising from federally funded R&D at the Nation’s universities, research institutes, and Federal Laboratories.

What supports this claim? On the face of it, it cannot be true. “Substantially” is wrong. “fueled” is utterly ambiguous. And what constitutes a “U.S. innovation system”?

Consider: about 60% of university patents in the Bayh-Dole era don’t recite federal funding–that’s about 70,000 patents that don’t to the 50,000 that do. And while university patent counts are a direct measure of the extent to which research innovation has been blocked and delayed, I don’t expect that NIST is making the subtle argument that because federally funded inventions are blocked by university patents less extensively than are other inventions, that somehow they contribute more to the “U.S. innovation system.”

And university research itself is hardly the source of “substantially” all “innovation” in the United States. We import innovation—technology, research, products—from the rest of the world; our industrial research produces new technology; and innovation—technological and otherwise—also arises in non-research (and non-institutional) environments. Put it this way, in terms of patents–in 2017, the USPTO issued over 300,000 utility patents. Of these about 7,400 recited federal funding. Do some math. That’s about 2% of the total with federal funding. NIST starts its report with the silly claim that this 2% of inventive work somehow is the substantial “fueler” of the “U.S. innovation system.” It doesn’t add up. 

Perhaps “fueled” refers to “funding.” In that case, let’s look at R&D funding. There is a fine chart at Wikipedia based on World Bank, UNESCO, and OECD figures. Let’s be coarse. The U.S. spends $500 billion a year on R&D. According to the NSF, U.S. universities spent $75 billion on research in 2017, with $31 billion (about 40%) coming from federal sources. On the measure of university funding for research, then, the federal component cannot be said to be “substantially” fueling much of anything nationally. Sure, $31 billion is a “substantial” sum–but that’s not the meaning of “substantially fueled”–there, “substantially” means “mostly, for the most part, in the greatest way.” Or, of course, “substantially fueled” is just political bluster that only a fool or a partisan would take at face value. Put it this way–the pharmaceutical industry spends more in advertising each year than the NIH spends on research. Or, consider federal expenditures generally–about $150 billion in 2017, of which $80 billion was military spending. At best, then, the federal government supplies 30% of the funding for the “U.S. innovation system.” Here’s a helpful graph from a 2004 CBO publication that will save another 1,000 words:

Or, consider the situation by who does the research rather than merely who provides the funding:

If anything, industry “substantially fuels” research and development in the U.S. But there’s more.

Back to that table of global R&D. While the U.S. provides $500 billion a year, China spends $450 billion and the EU another $390 billion, another $500 billion for the rest of the world. There’s somewhat under $2 trillion per year in play in R&D, much of that looking to do commercial business with the U.S. market. The U.S. share of global R&D would appear to be about 25%.

How much R&D “innovation” does the U.S. import from the rest of the world? I don’t know, but given foreign patent activity, it would appear that the U.S. imports a great deal of “innovation,” which then enters the “U.S. innovation system.” In 2017, of the 320,000 U.S. utility patents issued, nearly 113,000 went to assignees from China, South Korea, Japan, German, Great Britain, and France. We might then estimate that a third of the U.S. “innovation system” deals with non-U.S. technology. Of course, we might think that foreign organizations take out U.S. patents to disrupt innovation in the U.S.–and in that case, what are we to make of the U.S. patent system and its role in promoting the progress of the useful arts? Otherwise, we are looking, in big round numbers, of drawing in something on the order of another $250 billion of foreign R&D into the “U.S. innovation system.” That is, in a truly gross estimate, foreign organizations provide more funding to the “U.S. innovation system” than the federal government does.

Now all this is fine streetlamp statistics. But where do we get the idea that research money and patents are good proxies for innovation? Innovation, after all, involves some form of introduced change in an established order (Godin) or something considered new to an adopting group (Rogers) or a new commercial practice or product (OECD). Money spent doesn’t equate readily to what is introduced as new. An “innovation” may come at great research expense, or at not hardly any research expense at all, as when a society for putting things on top of other things simply stops doing that–innovation! but without much foo-fah of research first.

We might propose then that it is an open question whether research must always precede and cause innovation, or whether there are other pre-conditions for innovation. On the face of it–brief survey of history of innovation, say–it would seem clear that research is not a prerequisite for innovation, that much innovation has taken place without research by people recognizing something to their advantage, as while working at a craft or servicing a locomotive or hammering flint into a spear point. Thus, while research funding is an interesting thing, there’s good reason to doubt that money is the only, or even a significant, driver of innovation. And just to look closely into this face, we might wonder if money offered by government bureaucrats would be the best, most effective money in choosing research to support, in the odd chance that research might cause innovation.

We can wonder, as well, about patents as a proxy for innovation. Yes, patents tend to track things that are new, useful, and non-obvious. But how much of the new, useful, and non-obvious enters the patent system as a patent rather than as published “art” that limits the scope of what can be claimed as an exclusive right. That is–how much do issued patents track the total of “innovation” that arises each year in the U.S.? Estimates out there suggest that less than 5% of inventions with issued US patents ever see commercial play. There does appear to be a desire by some that it would be a great loss if anything were innovative without a patent–that is, the agenda is to increase the scope of the patent system so that anything that would be “innovative” would first have to pass through the patent system, would have to be owned, would have to enjoy two decades of exclusive control before anyone else would be permitted to adopt the innovation without purchasing it or instances of it or things made by using it. But that sort of agenda–expand patenting relative to total innovation–is a very different thing from “unleashing innovation” or “stimulating innovation.”

Even “unleashing” or “stimulating” innovation as a matter of policy suggests a deep dissatisfaction with present practices, products, and living conditions. Yet NIST’s report does not come out and argue that we are just too stupid, we are doing things all wrong, and that without changing established habits, practices, and even bureaucratic ideologies, we are stuck in the mud of crappy lives, crappy products, and crappy ways of doing business, and crappy ways of stimulating innovation through research money and patenting. But no, we see nothing of this sort, as if innovation exists as an abstract policy goal, to make what’s great already even greater, to gild the lilies of paradise. It’s difficult to imagine a screwier starting point for a policy discussion about how to make the results of federally supported research more broadly available and useful.

If “U.S. innovation system” is a secret code for “whatever NIST wants it to mean” then perhaps NIST’s claim is true, that federal funding is the “substantial fueler” of that “U.S. innovation system,” the one that NIST means, the one that federal funding does “substantially fuel.” And if we distinguish a “system” from other ways by which innovation might arise, then perhaps all NIST is claiming is that the bureaucratic part of the effort to innovate is “fueled” mostly by federal money. Or, tautologically, the federal effort to innovate is funded primarily by federal funds. But other efforts to innovate may not be systems, may not be U.S.-based, and may not be driven primarily by federal funds or federal bureaucrats or pundits paid by federal bureaucrats.

As a policy discussion, NIST’s report starts with a set of assumptions that precludes almost all discussion of policy that might lead to changes in practices at federal laboratories that would make research more relevant to the American public, or to industry, or to entrepreneurs. NIST assumes that federal research is more important than anything else, that research drives innovation, and that all this has produced wonderful results–but that there is even more wonderfulness per year possible. Nothing about this suggests any deep urgency to change much of anything. Not whether we have too much federal research money about, sucking talent away from really interesting and useful work; not whether we have poorly chosen and sited research activities; not whether there’s too much emphasis on the use of the patent system; not whether public purposes really might be worth staying with all the way through and not allowing private speculators to intervene and fragment a collective effort into gobbets of exclusive rights.

In a sense, we have a Strictly Ballroom problem–the NIST pundits cannot contemplate rule changes for the federal research dance competition because if they changed the rules much, then they would not be the pundits anymore. Others, who know how to dance under the changed rules, would have the opportunity. They cannot contemplate change that doesn’t include them retaining their present positions of punditry. Thus, the “innovation ecosystem” that the NIST pundits refer to must be the one that they control, not just how it comes about that new things come into being and change established orders, including NIST punditry.

Thus, we might posit that federal changes in research and discovery administration may take place only at the speed it takes federal pundits to restate or walk back from their public positions, which in practice doesn’t much happen for punditry that takes the form of prophecy (“federally funded research produces innovation”; “federally funded research substantially fuels the U.S. innovation system”). We wait a generation and hope that the current pundits fail to train their successors or are rescued from this world of woe by UFOs from the planet Clarion.


This entry was posted in Innovation, Metrics, Policy and tagged , , , . Bookmark the permalink.