The US government proposes spending $1b on innovation in manufacturing. If one likes government spending on such things, this sounds like a really good thing. Here is a link to the NNMI homepage overview. The preliminary design document for the effort is here. In a nutshell, the plan imagines up to 44 institutes each focused on a technology or theme in manufacturing, spread regionally around the country. A first institute is up and going in additive manufacturing, spread across a multi-state region from Ohio to West Virginia dubbed the “Tech Belt.” Another 15 institutes are being prepped for proposals.
Each institute is conceived as a coalition of leading organizations in a focused area of study:
Institutes will be a partnership between government, industry, and academia, supported with cost-share funding from Federal and non-Federal sources. It is expected that institutes will typically receive $70-120 million in total Federal funds, depending upon the magnitude of the opportunity, maturity, and capital intensity of the technology, and scope of the focus area, over a 5-7 year timeframe. When combined with substantial non-Federal co-investment, for example 1:1, it is envisioned the total capitalization of an institute over this period will be $140 to $240 million.
A key feature of the network and its individual IMIs is a strong focus on building clusters of advanced manufacturing capabilities that join expertise from industry, academia, and government . . . .
Each Institute will have a unique and well-defined focus area, such as a manufacturing process, an enabling technology, manufacturing processes for new advanced materials, or an industry sector . . . .
If one is going to have huge scale institutes led by industry, academic, and government leaders, funded at multi-million dollars a year, requiring an additional multi-million dollars of private money, then the NNMI is set up about as well as one could want. It is, so to speak, “best practices.” In the NNMI one sees the influence of a similar effort for nanotechnology, the NSF Engineering Research Center and Science and Technology Center programs, and the CRADA environment used at federal labs, all wrapped up in a competitive proposal process, with a panel of leading experts to decide what applications are worthy to be provided with funding to create such sustainable institutes.
The design document assumes a linear model of innovation. If we need innovation in manufacturing, so the linear model informs us, then we need basic research in manufacturing, the results of which are passed on to applied research, and as the “technology readiness levels” advance, at some point companies will develop the “last mile” of improvements in the best, most readied technologies into commercially valuable manufacturing processes and products.
They will identify critical manufacturing processes and technologies with potential transformational impact, and through their member companies, they will have the capacity to translate these technologies into market-relevant private-sector manufacturing production.
Now this whole TRL stuff is the work of large organizations determined not to let unready technology slip into their processes or products. The premise is that technologies that have been identified and validated will be the subject of coalition development efforts. Think about this in light of universities patenting up much of what they see. How does one of these institutes come to have access to such technologies if the patent rights are reserved for the exclusive use of speculative investors? Are these investors going to allow their patented inventions to go to an institute for development? Why would anyone accept such stuff into a manufacturing “commons” (as the design document calls it)? Clearly, there’s stuff in reality at odds with the vision of the design document, best practices and all.
The design document asserts that there is a “gap” between basic research and commercial development. This “gap” means that all the wonderful stuff created in basic research is not moving on the linear model conveyor belt to commercial use because there is a lack of government funding in this middle TRLs:
Common weaknesses identified in many of these studies might be labeled as “missed opportunities” or failures to reap the full economic and commercial value from public investments in research.
As summed up the National Science and Technology Council,”A gap exists between the R & D activities and the deployment of technological innovations in the domestic production of goods,” contributing significantly, for example, to the growing trade deficit in high-value-added, advanced technology products.
There are many examples where materials and product technologies were innovated in the United States, but most of the significant commercial market share they enabled was lost to other countries. They include rechargeable lithium-ion batteries, oxide ceramics, semiconductor memory devices, manufacturing equipment such as wafer steppers, flat panel displays, robotics, solar cells, and advanced lighting.
It’s not that these technologies have not got to market, but that somehow the United States, for all its basic research investments, ends up not manufacturing the products. If one wanted to construct a narrative of what is working, the United State spends tens of billions each year on basic research and the world uses the results to develop manufacturing capabilities that answer to the needs of global firms seeking to profit from these results. The system works quite well. There are no gaps, though there are barriers. For instance, when American universities grab up “early-stage” invention patent rights, they take those results out of circulation in all jurisdictions where they can patent–often only the United States. In such cases, the results are not used in the United States, except on the occasion of a monopoly license successfully producing a commercial product–which according to one account I’ve read, is about 0.2% of what’s claimed by universities. That’s 2 in 1000. Vast swaths of the research landscape, and all the nice places to live and work, are being claimed by the university patent licensing franchise.
University patenting is creating a huge disincentive for basic lab results to be used in the United States, all but assuring the continued operation of the pipeline of results to foreign jurisdictions for development. Patents don’t address the gap: they create it. Or at least, they do a great deal to contribute to the gap, and very little to close the gap.
The problem with manufacturing development in the United States is not that there is a lack of innovation in manufacturing. The problem is that it’s difficult to build new manufacturing facilities in the US–delays in permitting, cost of regulations, uncertainty of political will, and the like. We are not talking here about creating huge polluting factories like China has done–we are talking about the efficiency and effectiveness of the processes that approve such work.
To cite one example, Plantronics, the headset manufacturer, has its world headquarters in Santa Cruz, California. There, adjacent to one of its buildings, on its own property, Plantronics wanted to pour a cement slab and build a new design center. Maybe a footprint of 75′ x 75′. In the three years it took to get a permit to pour the cement from the city of Santa Cruz, Plantronics built a new world-class green, fresh, and fun manufacturing facility in China. The problem is not innovation in manufacturing, but a regulatory environment that is antagonistic to manufacturing. Even when folks aim for Cradle to Cradle technology cycles, the place to do it is not the United States. No amount of TLR 4-7 gap funding will alter this situation. The changes have to be in the regulatory environment that does not promote manufacturing–and that means land use, environmental review, construction permits and inspection, financing, and infrastructure support. It means dealing with a court system that can hold up initiatives for years on procedural matters alone.
In all this, we have not even looked at labor issues. Of course, if one can get skilled workers in China, or India, or Thailand, or Indonesia, and one has manufacturing processes that require such workers, then that’s where the investment will be. If a worker in China gets paid $2/hr and a worker in the US gets paid $20/hr, one can see the attraction of shifting labor-intensive work, at least for a period of time, to China. As one commentator has pointed out, the US government, by “exporting” as it were inflation to the rest of the world, helps to drive up foreign costs and makes US manufacturing look a lot more attractive. But wages are not the controlling issue: wages in Germany are higher than in the US, but Germany is thriving on exports, much to the consternation of other parts of the Eurozone.
In places like Italy, laws that were meant to support workers turn out to support one generation of workers, but not the next:
Employers have battled for years with the unions for greater flexibility. The result is a three-tiered labor force, a setup Italians dub “apartheid.” Of 27 million workers, 15 million—most 40-plus—enjoy stable jobs with guaranteed privileges. An additional 8 million, mostly younger, form a growing army of freelancers and employees on continuously rolled-over short-term contracts. They receive none of the benefits that would in theory be granted under the generous labor laws. The remainder, 4 million or so, toil in the unprotected underground economy, according to Italy’s National Institute for Statistics.
Thus, if one is setting out to spark a manufacturing revolution using “innovation” to create “jobs,” then it would be good to stipulate just which kinds of jobs one is intending to create, for whom, and how that stacks up with quality of life for the workers. In a lot of state-financed economic development programs, it would appear that the only ones assured of jobs in the whole process are the administrators of the programs. Yes, some construction workers get temporary employment, too. In Utah, the much-vaunted USTAR program put $93m of state money into innovation research. The result was 4 companies and 13 employees. It’s not clear there were any new products–i.e., innovation–but certainly there was a mighty big pile of aspiration. The state of Washington copied the USTAR program, but the legislature canceled the program before getting a final report, saving itself from finding out that its expenditures performed even worse than Utah’s. In 2000, California embarked on a plan to spend $400m or so on four “Institutes for Science and Innovation.” Despite the expenditure, and requirements for private investment, there’s little one can find reporting any innovation, or jobs, or new stuff to make the citizens of California happy. Of course, administration of the program takes up a lot of the available funds, as does building new research buildings, and paying for research to get done. But innovation? Not really.
What is so strange about the NNMI proposal is that the design of the program follows the same design that has not done much of anything for economic development. The NSF’s Engineering Research Centers and Science & Technology Centers feature the same sort of coalition/cooperation/consortium structures, complete with a thematic focus, requirements for private investment, and a premise of self-sustaining activity after the federal money runs out. There have been about 100 of these efforts, but it’s not clear that more than a handful continue after their ten-year run. For most, a decade is enough, and there’s always more federal money to go after–and getting that is apparently a lot more attractive than trying to keep some old institute going with private money.
Another problem: it appears that “peer” review of proposals narrows the range of possible areas of work (as discussed here). While such review may be “best practice,” there’s no little indication that it is effective practice. In essence, peer review ensures that insiders control research to keep it on track with the leadership positions the reviewers have in their areas of study. To suggest something that ignores such positions, or threatens them–well, that takes a big, two-hearted reviewer to let slip through. Michael Crow and Barry Bozeman argue in Limited By Design that the federal laboratory system is set up for a world that has changed out from under it:
R&D institutions and R&D policies have grown without consideration of system-level issues, coordination issues, life-cycle issues, or even consistent review and evaluation. All of these facts, matched with our data, indicate that policy makers are ill prepared to deal with what is now [in 1998] a 50-year-old enterprise in need of serious renovation. (27)
Crow and Bozeman point out one prevailing problem–ideology without pragmatism:
U.S. science and technology policy and the federal laboratory system revolve to a suprising degree around ideological premises. This is not to say that ideology supersedes politics, but that politics is cloaked in ideology. Among discussions of such big issues as the power of the market, the potential of activist government, and the meaning of globalization, the more mundane issues of “what works?” easily get lost. . . . If one has little evidence of “what works,” then ideology at least provides some decision-making ballast. (33)
The NNMI design document is packed with ideology and has little regard for what works. There are plenty of vision statements, about how clusters will form and create jobs as a result of securing federal grant money, how new products and jobs will arise, and national competitiveness will be restored. It’s as “best practices” a kind of vision as one could hope for. It’s just that there is scant evidence that this sorts of initiatives amount to anything, in terms of innovation, jobs, or stuff citizens need, with special focus on small companies–those are the poster subjects, of course, but the connection between a nifty $150m Institute and the stuff in the posters never gets made–it’s just asserted. It will happen. Or, at least, it’s supposed to happen.
From the linear model, one gets the idea that if the government funds “basic” research, then somehow industry will consume the results, and from this digestion will come new products, jobs, and economic vitality. The problem for “technology transfer” then is to somehow cajole, or force, industry to swallow the IP along with the discovery. That is, the role of the patent is conceived of as a way to make industry pay for the meal, with the idea of “added value” being that industry pays not only for what it may want (discoveries) but also for what it generally does not want (patents). In the university view of things, tech transfer offices are not soup kitchens. Companies should pay, or be made to pay. Sounds ideological, one might think. If tech transfer offices are not signing up a lot of companies to licenses, and receiving a lot of money for doing so, then, well, it may be a failure of the technology transfer office or a failure of the “market” to accept the tech transfer goods on offer. That’s been the slow-boiling fuss as it becomes apparent that the present system is not producing the 30% licensing rates claimed for university-based licensing in the rhetoric that created Bayh-Dole. It’s more like 0.2%–two orders of magnitude lower than claimed.
Into this environment, then, comes the NNMI program, proposing even more money, and placing more requirements on participants, and all but stipulating that the result of this ideological design of research, positioned so rationally in the TRL 4-7 “gap” that seems to be no gap at all in most countries of the world, will be products, jobs, and national competitiveness. It may be well to wish that it could be so. It may be well to aspire to help to make it so. But there’s nothing in the linkage between big, multi-million dollar collaborations sprung out of nothing and innovation, jobs, and national competitiveness besides an assertion of a linkage. The best practices on which the NNMI are based are just those that haven’t “worked” for the purposes claimed.
Crow and Bozeman, discussing federal labs, argue for an institutional design approach, one that gets at what labs are doing, encourages coordination of work, but reduces federal agency management and augments flexibility and autonomy, with more policy formed in the field. The NNMI does not follow this advice. The NNMI design document makes a virtue out of the idea that the government does not expressly dictate the focus for each institute, allowing various competing coalitions to come up with their own proposals. But the government does get to choose who gets funded, and so it comes around to the same thing, cloaked differently. From then on, it is a matter of doing research and development and training under the scrutiny of a federal contract, demonstrating and documenting compliance down to the use of photocopy paper, but not, in the end, accounting for whether anything “works” in terms of new processes, jobs (beyond the NNMI pork), and the location of new manufacturing in the United States.
The problem for the small business that might be doing a bit of research is that the NNMI, despite its protestations to be about small companies, is actually a large flying island that casts a huge shadow. Small companies have a choice, once $150m or so of federal funding (plus, apparently, the same amount or more from the industry leaders) is positioned over their efforts: join up and contribute one’s technology to the larger effort, or get swamped out. Everywhere that the NNMI funds a coalition of “leading” organizations and companies to “innovate” in manufacturing, that’s scorched earth with regard to SME efforts. If the universities involved file patents on everything in sight, all the more reason for the SMEs to not even try. The design of the NNMI is one that aims to bring outlier research under the control of the federal government (the government gets a non-exclusive license in everything), or if not the government then coalitions of companies that are ready to work with the government. The premise is, like it has been for the NNNI in nanotechnology “make the program look successful and you will be rewarded with funding.” That’s the fundamental deal. If it is rawly pragmatic for companies, it becomes sacred text for the universities. If basic research does not appear to contribute materially to “innovation” then why fund universities with $50b or so of federal money each year?
There are good arguments for government funding research, and supporting industry, and taking care of workers. At each point, however, especially as the money scales and the administrative processes descend, the ideology and apparatus of the program come to dominate. “Best practices” is code for “more of the status quo.” While that is fine for the propagation of practice–imitation, in Teece’s taxonomy–it bodes less well for innovation itself–introduced change in the established order. The NNMI design proposes to extend the federal lab program into manufacturing, setting up 15, and then more than 40, “institutes” to dominate areas of study. May good things come from doing so? I certainly hope so! But from what I can see, the NNMI design is based on ideology, not “what works.” For small companies, it means, get swept up, or swamped out by the vision of federal entry into areas of work.
If the federal government wants more manufacturing in the States, then how about it start with regulatory reform of the processes by which manufacturing facilities can be built and operated. Worker pay is not the problem. And the environment does not have to suffer from a blind eye. But the need is for a National Network of Better Governance of Innovation–which means, often, knowing when and how to get out of the way, while still preserving a role as trustee or advocate. Perhaps something like Charter Innovation would do be worth considering.