Battelle is out with a new study forecasting “Global R&D Funding” for 2014. The report covers countries and various areas of research such as biotech and energy. The findings are rather bland–the US will spend $465 billion next year on research, with industry spending 71% of that, and universities (mostly with federal money) spending 16%. The US spends about one third of the world’s research budget, with Asia spending two fifths and Europe spending about one fifth. The US will spend about 2.8% of its GDP on R&D, about like previous years. The report goes out of its way to suggest that China will outspend the US on R&D by 2022. I get the impression that this is supposed to be a problem.
In another way, however, the report is noteworthy–for what is left out and unremarked. Research and development appears to have little by way of outputs. Impact is reduced to the effect of expenditures on R&D, and the beneficial effect of that spending on the economy. That’s how the number of jobs involved is calculated. The impression the report gives is that the activity of research is way, way, way more important than the findings of research. The justification for policy-makers is in the first-order spending and the second-order effects of that spending, not in findings, discoveries, inventions, new products, new industries, new ways of understanding the world, markets, company organization, and the like. The report does mention “impact metrics” in the context of the results of a survey in which an unknown number of respondents mentioned “publications” as the most important metric of the impact of R&D, followed by patents. While a publication or patent might mark a kind of research event, neither in itself has the prospect of much economic impact of a “positive” form. Yes, a publication might help swamp the literature with another announcement of incremental “progress” that plumps the feathers of a research organization, and yes, a patent might make an area of inquiry less attractive for other investigators, and so disrupt collaboration and competition alike, and delay overall progress in favor of establishing fragmentary fiefdoms to what ever organization has the money to pay for patenting. But these are not substantial economic benefits, certainly not worth the $465 billion price tag for next year’s purchase.
The report, on page 21 for those of you following along, has a full-page figure that sets out in spectacular form a version of the linear model of innovation, in which research flows out from basic to applied to development to new products in a virtuous circle that requires care at each step. This figure is presented in the context of an “innovation ecosystem” in which, somehow, new things will get created by the global R&D engine. What is odd is that it is an “ecosystem” that Walter Valdivia in his recent Brookings Institute report sets in opposition to the linear model, which he argues is discredited. Apparently not at Battelle. Take a look:
The primary green arrows point south, from research to money-making, which then fuels more research. The figure is reasonably clear on this point: Money from research products should fund more research as “reinvestment.” The money-making itself comes after new products in the form of “clusters” and startups–things that governments and regional economic development folks might consider important, but which the Battelle report does not take up as an impact metric. Again, the apparent benefit of R&D is a kind of economic clustering, which cannot happen just anywhere but rather must create polarizations–big winners and lots of losers, as companies, talent, and money move to join the dominant “clusters”–Silicon Valley or Boston or Shanghai or wherever. One would not see this in the figure, however, which makes it appear that cluster formation is a general property arising from the flow of research activity towards economic activity based on products and outcomes. A casual government policy-maker might think every region will benefit from such cluster outputs, but clearly this is impossible.
More to the point, if this figure reflects a “comprehensive R&D ecosystem,” why isn’t “reinvestment” limited to money made from the outputs of research? Why should government money from general sources have anything to do with it? But of course, money comes into research from all sorts of sources, not simply “reinvestment” from profits from research-supported products. Thus, the figure presents a puzzle: is the figure advocating that research funding be restricted to profits from research-supported products? or is the figure something of a fiction, failing to indicate that research is supported broadly by a range of sources, including from tax revenue, and that those sources may not have anything at all to do with research? or in an even stranger twist, does the figure claim that all sources of economic activity ultimately derive from basic research? It is a nice figure, graphically speaking. I have little confidence that it is in any sense “true.” One might call it an “emblem” or a “hope” that a comprehensive R&D ecosystem might be represented by a tidy diagram, even if little on the diagram actually corresponds to anything in an actual “ecosystem.” One might go so far as to say I’m being a bit harsh on the whole thing, as anyone can see that the figure is not meant to represent even a general R&D ecosystem, but only to underscore the point that more money is needed for R&D, and some of it should come from companies making money in areas of technology in which research also operates. No doubt the diagram has been created to make a rhetorical point, and not much else.
The figure’s graphical rhetoric is disturbing in other ways. The apparent start of the process depicted here is “basic research.” There is a way of describing the development of any idea that might indeed start with a finding of “basic research,” but basic research is not the sole source of new ideas about the world. Practice-based findings may be as important, or more important, to innovation than basic research. The same may be true with regard to other sources of innovation–the salesman or the science fiction writer may have an insight that leads to the things to do, the places to look, for new ideas about the world, about technology, about products. There are good arguments that basic research often comes after an important innovation, to wrap theory and sweep out the obscure corners of cause and effect, which might be happy news for efficiency gains and the like but utterly uneventful in terms of the relationship of the cows to the barn.
Innovation may owe much more to long-established technology and practice than to the current results of basic research. This may be the case even more so now, with universities patenting every “early-stage” finding that they can, ensuring twenty more years of drought before a result can be productively used in business. Microsoft and Facebook, one might argue, were not founded on the near-term results of basic research, but rather on a body of practice that had developed and was generally available. Some of that, of course, may have benefited from basic research–but not stuff covered by patents at universities, not stuff that their own basic research scientists (they had none in starting) had created to give them monopoly advantages. There are, of course, instances of companies starting that were involved in basic research, such as Genentech and perhaps Google, with the page rank algorithm, and both companies also relied on a host of things that were readily available for use, without regard to the origins in basic research of those things.
Thus, one might ask whether the mix of available stuff (from all sources) combined with the introduction of some new idea or finding (from any source) has more to do with the opportunities for new products and economic activity than does the total amount of money spent on basic research in a country or industry. If it’s just total money, then the argument would have to be that whatever places are dominant will continue to be dominant, so long as they spend more money than everyone else. I would be surprised if total dollar expenditure is the key component of either discovery or of innovation. Expenditure would appear to be most important to those on top, seeking to stay on top. As a market expands, and competition comes from multiple directions, one has to improve at the margins to maintain control of product definitions, services, and income streams. That’s important stuff, and it may even create clusters of service-providing companies that soak up that expenditure, but such a value-chain economy is not necessarily an innovation economy–one might think of it more as a hegemony-chain economy, built by the dominant for the dominant, rather than an innovation economy that introduces change into status quo activities.
When some new thing has happened, and been taken up in to a successful product, then it is possible to construct a story that picks out features of the new product that might be ascribed to recent “basic research” (often marked by a patent owned by a university) and describe how that feature came to be designed and validated for commercial use. Such a story may well leave the impression that the product itself came from basic research, through just the sort of “technology readiness” levels sketched generally by Battelle’s vision of the linear model. The problem is, this is a story written in reverse, often leaving out a great deal, to be told forward, and then generalized as the expected manner in which new stuff happens. If one is in basic research, perhaps it is a happy story to tell, one of centrality and aspiration. If one is in policy, however, it is one story among many possible, and to fixate on it is to attempt to live in an rhetorical illusion. The real world is mostly different, chaotic, mysterious, multiplicitous, and changing. The illusion of economic order imposed by a figure such as Battelle’s is hardly reassuring.
As a general economy, the Battelle figure fails in yet another way. The steps in the progress from basic research to cluster-formation do not identify either a finding or an organization that is put in play. The boxes are simply categories of research, and the arrows are simply flow of something–what? Inventions? IP? “Knowledge”? People? If in the boxes are lots of different organizations at each step, then how does whatever is flowing from one box to the next actually get there? Is everything in the previous box in the public domain? Is it licensed to everyone in the next box as a standard? It would appear that the boxes, until after the “technology transfer” step, are all within the same organization. That would suggest a model of innovation in which each organization seeks to develop its own work, itself. “Publication” would not matter, except as opportunities for self-promotion and warnings to other would-be developers to “keep off.”
The Battelle figure is not one of collaboration, or of general access to a vast range of long-existing technology, or of the movement of talent to areas of greatest innovation activity. Instead, this is a figure of dominant organizations maintaining their dominance through internal control of creative processes. If universities want to be “like Mike” with their research findings and “Mike” happens to be large companies seeking to maintain monopoly or hegemonic dominance of their industries, then this Battelle figure would appear highly attractive. It is a production model, not an exploration model. It is a model on which to build pyramids, not one to discover a new sort of geopolymer to make pyramid building so easy that anyone could do it.
Here’s the thing about basic research. We don’t know what to look for. We don’t know where to look for it. We don’t know what it looks like. We don’t know why we want it. These are the attributes of the unknown future. This is the home of black swans. This is the unexpected. The Battelle figure does not reflect an engine that might get us to the unexpected; the Battelle figure is an engine of more of the same, as if the future is the result of a controlled process of development which, in coming to make the most money for their possessors, necessarily attract everyone else to cluster around seeking business by offering services and technology as hopeful offerings.
I am all for development under managed control. That’s a useful form of development. It doesn’t bother me that companies operating at huge scales are involved in development. They have the financial and labor and political resources to make some things happen that otherwise might not. Managed development is how one gets from a design concept to a working product, just like the figure indicates. How much one spends in getting there, however, is not a sign of “better.” If one adopted a managed development model that had less waste, less uncertainty, less failure, and more useful products in areas of unmet need, perhaps a country or industry could spend *less* on research and have much greater impact, profit, and pizzazz. My impression, for working in university IP management and talking with folks working at major industry and federal lab research centers, is that only a tiny, tiny, let’s repeat for effect, tiny bit of research output makes it from lab to marketing. Even in the big corporate labs, the ones spending north of a billion dollars a year, much of what is discovered or developed does not make it into that company’s products. The experience of Xerox and its Palo Alto Research Center is well known, with a succession of discoveries that became valuable, but not for Xerox, not within Xerox. The reality is that it is very difficult to invent what you want, or need.
Even generally insightful folks like Chunka Mui get misled on this point. Mui wrote a year ago about how Apple and Microsoft benefited from the research at Xerox, and argued that it is essential that Apple launch its own internal research program so it will have a source of innovation–as if it cannot also draw on research done elsewhere, as it has in the past. Oddly, where Apple has done some nice design and development work, it appears to have been raided by Apple’s own supplier, Samsung. Apparently you can have too many cows in a barn. And what of the billions Microsoft spends each year on research? How much of that makes it across the brain-blood barrier to the seat of marketing?
Almost all basic research seems, in practice, to be on behalf of someone else, in the future, or far away, randomly selected, that one only in the luckiest of circumstances has the chance to meet. Nothing an organization does to try to own such results increases the likelihood that a match between inventors and users will be made, no matter how rosy the scenario of technology transfer, that with patents an institution taking ownership can somehow find that perfect match, do something good in the world and make scads of money doing it. Maybe in the Barbie and Ken world, administrators playing doll house. Not for the most of us. The patent serves little good. Institutional ownership serves little good. Government demands to follow some set protocol serve little good. Policy statements inducing government spending to make a show of people following protocols serve little good. More spending on research, for the sake of more spending, to outspend China, serves little good.
If we care about innovation for the sake of progress, or to redistribute opportunity, or to remedy the ills to which our lives are exposed, then we might start by accepting that more research funding is not the answer, especially if that funding is of such amounts that it is immediately “captured” by the dominant organizations of the day, who then (according to the figure) seek to monetize their positions (via IP, products, research funding) with the expectation that they will “re-invest” in more research, so long as the results come back primarily to those in power. It is an understandable vision. It makes a lot of sense. It’s difficult to argue with it, on its own premises, just as it would be difficult to suggest to a very, very, very rich, phenomenally wealthy merchant banker that he might give up some portion of his wealth just for the good of it. That would appear to run against his nature, his telos or purpose, and therefore against virtue itself. It would be unethical.
What then is the answer? Start with diversity: there will be multiple answers. Consider where innovation does arise, and how, as Steven Johnson has documented. Accept we know way less than we would like to believe, and that our pattern recognition systems are terribly unreliable, as Nassim Taleb and Daniel Kahneman in rather different ways have made clear. Pay attention to what Vannevar Bush wrote about establishments, special projects, and gadgeteers. Rig for the unexpected. If we are to go hunting the Black Swan, and get back alive, and have something good to show for it, then it will take some imagination–and luck–to know what our rigging should look like. Otherwise, we mount many, expensive Franklin expeditions to the unknown that never return, ill-equipped and ill-prepared, without ever learning what we ought to know. We don’t “do research” any better than we “do good.” Research is not something “done” for its own sake, any more than good is. The situation does not improve to make “research” an industry, any more than philanthropy as an industry. As Thoreau would have it, to be the attempted object of such philanthropy should start one running for the hills:
If I knew for a certainty that a man was coming to my house with the conscious design of doing me good, I should run for my life, as from that dry and parching wind of the African deserts called the simoom, which fills the mouth and nose and ears and eyes with dust till you are suffocated, for fear that I should get some of his good done to me — some of its virus mingled with my blood. No — in this case I would rather suffer evil the natural way.
If we looked to Thoreau for research policy advice–and I may be the only one in recent memory to do so–we (meaning me, I suppose) might latch onto this simple statement:
What good I do, in the common sense of that word, must be aside from my main path, and for the most part wholly unintended.
It may be that research is also of this nature. Rather than being a directed, professional activity, full of credentials, competitive grant proposals, and sufficient funding and standing to be invited to talk at conferences full of others envious of one’s resources and publications, productive research might be something “aside from the main path” and its consequences “for the most part wholly unintended.” Absurd? Is it? Or is it absurd to think we can do a thing that is a description of the consequence of some other thing? Does it really help to call something “research” before you do it? Or is it “research” because afterwards, it was clear that this is what you did? I might add that Vannevar Bush made a case for much the same thing as Thoreau points out: what a team working on a problem in the interest of an area of endeavor but outside the control of its establishment might draw on will come from unexpected places, and be created by people with unexpected credentials–it may be academics, or industry scientists, or gadgeteers who messed with radios or automobiles or college mug books or 3d printers. How does one go about accounting for such associations, or teams, or efforts?
The Battelle report suggests that spending on R&D is important, that more is better than less, and that all aspects of R&D must be spent upon to enable a linear model to create an ecosystem, and if that happens, clusters will break out everywhere, money will flow like honey, and some portion of that can be invested back in yet more research. I believe it, I really do, but only for a moment.