COMMENTS ON THE DESIGN OF THE NATIONAL NETWORK
FOR MANUFACTURING INNOVATION(NNMI)
In response to the first “Blueprint for Action” workshop hosted by the National Institute of Standards and Technology (NIST) to review the design of the NNMI program on January 16, 2013, the program descriptions provided there,[1] and the White House Preliminary NNMI Design released the same day,[2]the Massachusetts Institute of Technology Washington Office submits the following comments to NIST’s Advanced Manufacturing National Program Office (AMNPO) for consideration during the NNMI design process. These comments also consider the Information Technology and Innovation Foundation (ITIF) December 2012 report, “Why America Needs a National Network for Manufacturing Innovation (NNMI)”.[3]
These comments are organized according to the first three of four design topics discussed during the January Blueprint for Action workshop, namely:
- Technologies with Broad Impact
- Institute Structure and Governance
- Strategies for Sustainable Institute Governance
1) Concerning “Technologies with Broad Impact”
- The NNMI role as “testbeds” for technology demonstration is an important one and should be an explicit element in program design.
The majority of U.S. manufacturing is undertaken by some 300,000 small and mid-sized production firms. To compete in international markets, these firms along with larger firms must move to a new level of productivity and efficiency.
The testbed phase for such advanced technologies and processes marks a genuine gap in the manufacturing innovation system, and NNMI is a mechanism to fill it; design of NNMI and its constituent Institutes for Manufacturing Innovation (IMIs) should explicitly emphasize this testbed function. Small and midsize manufacturing firms are generally thinly capitalized, so they must be risk-averse; fewcan implement a new advanced manufacturing technology unless it is fully proven in their production process contexts. In addition, these firms are unlikely to obtain financing to install such new production technologies unless the costs and savings can be clearly shown. According to the White House Preliminary Design, “strategies to encourage the participation of SMEs in Institutes include engaging outreach partners and intermediaries that work closely with SMEs, providing valued information and services tailored to address SME needs, providing a tiered membership fee structure, allowance of all in-kind contributions for new member SMEs, the use of contract-based activities, staged licensing of IP, and similar arrangements.”
By demonstrating technology performance and cost, this testbedrole of the NNMI is one of the strongestarguments for IMI’s and is also critical to obtaining project financing for smaller firms. The testbed role will also provide an opportunity for smaller and midsize firms to train their workforceson the new technologies. Finally, emphasizing the testbed role to potential supporters or partners will allow each IMI to move ideas to products at the target Technology and Manufacturing Readiness Levels, which were outlined in the White House Preliminary NNMI Design.
2) Concerning “Institute Structure and Governance”
- NNMI’s should focus on advanced manufacturing technologies with broad impact and synergies across a series of significant industrial sectors, not on a single industrial sector.
The Advanced Manufacturing Partnership, in its 2012 report to President Obama, explicitly recommended that NNMI’s should foster technologies and processes with broad industrial impact that cut across a series of industries. If cross-cutting technologies are selected that result in manufacturing system-wideimpacts, these will have a larger effect on manufacturing as a whole and will create greater synergy among investments. Thus, the potential target technology areas for NNMI focus should create a web of interconnected, not isolated, industries.
- NNMI launches should initially encourage pilots and design experimentation.
The Administration recommendsa network of 15 Institutes for Manufacturing Innovation (IMIs), with an average annual budget of $40 million, totaling $1 billion per year.If such major initial funding is not available, as the Administration has recognized, the effort can still move ahead using existing authorizing authority available to a series of agencies within their existing budgets.
Rather than locking in the NNMI design ahead of time, allowing a series of government agencies (NIST, DOD, NSF, DOE, and perhaps NASA) to take the lead on one to two pilot NNMI’s each could enable experimentation in design available at each within an overall framework, so that a range of best practices can be evaluated. The particular technology focus for such pilots cannot be entirely “bottom up” and open – there has to be some connection to advanced manufacturing work that would complement the agency mission, otherwise the agencies will not support the pilots over time. However, serving the agency mission must be balanced, of course, by selecting a technology focus with broad industrial impactnot just serving a narrow agency need. With an initial handful of pilots in operation, the overall NNMI design, based on these results, could be practically refined and tweaked before locking into fixed long-term system plans. Manufacturing infrastructure creates a very complex substantive design problem and experimentation should be encouraged within an overall framework.
- It is important to consider the role of federal R&D in the NNMI mission.
A strong portfolio of complementary R&D will be needed if NNMI’s are to achieve their goals. Yet despite the Advanced Manufacturing Partnership report recommendations to this effect, the federal R&D agencies involved in advanced manufacturing research (NIST, NSF, DOD, and DOE) have not yet developed common, collaborative, cross-cutting R&D initiatives that will support the NNMI efforts. Most advanced manufacturing technologies are not yet technologically mature and require ongoing R&D. IMI’s that work at the late stage of development through testbed stages of innovation will need to be supplemented by R&D that feeds into them. Manufacturing innovation inherently must be a connected system that ties together different stages in the innovation process, from research to development to testbeds to implementation, and the NNMI design needs to fully recognize this. If the NNMI design is to succeed the agencies should build a supporting R&D program element to support it, otherwise an IMI may turn into a fish without water.
This manufacturing R&D problem is an aspect of a larger structural problem. The current federal R&D system doesn’t place a major focus on manufacturing R&D; advanced manufacturing is simply not a significant R&D priority (in fact, it is less than one percent of total federal R&D). Thus, it is a secondary focus in university R&D. Federally–funded R&D in advanced manufacturing is also dispersed acrossfederal agencies without a focused priority-setting cross-agency mechanism. Universities historically perform important work in breakthrough research, but more support is needed in the public and private R&D sector dedicated to advanced manufacturing breakthroughs as well as to the traditional incremental/engineering research. Without this connected, cross-cutting foundation, NNMI advances simply will not occur. Underpinning that research should be, as the Advanced Manufacturing Partnership report recommended, a technology strategy developed by industry, agencies and universities.
3) Comments on “Strategies for Sustainable Institute Governance”
- IMI’s need to be designed to foster the translation of regional results to national-scale implementation.
The progress and results emerging from NNMI’s must meet national as well as regional needs, since the program as a whole aims at a broad-based transformation of manufacturing that produces national-scale results. NNMI’s will also need to work on business models and knowledge transfer for the processes and technologies they evolve.
Manufacturing is organized regionally not nationally, so NNMI’s need to be regionally based. In his comments[4] submitted to the Department of Commerce’s Request for Information on Proposed New Program: National Network for Manufacturing Innovation (NNMI), MIT Associate Provost and Professor of Electrical Engineering Martin Schmidt suggest how the regional model could operate:
The business models should be designed to leverage the existing strong connections between universities and strong innovation ecosystems which have developed in many regions of the country, but in which the manufacturing sector is rarely represented well.... Regional investors andstate economic development agencies must be involved in developing andexecuting Institute business models if SME (small and medium enterprises) manufacturers are to be more fully integrated into these innovation ecosystems.
However, a regional ecosystem model also requires a national translation model. For example, achieving advances in additive manufacturing for particular industries in Ohio is useful, but the results really must be disseminated nationwide for a wide spectrum of benefits to flow. How will the lessons from an additive manufacturing NNMI in the Midwest be disseminated to suppliers on the east coast? Schmidt also addressed the need for a national advisory group, which was also recommended by the White House, to assemble broader groups of industry:
Strategic guidance and measurement of progress towards an Institute’s strategic goals would be enhanced by a collaborative advisory board whose membership goes beyond the Institute partners, to obtain a national reach beyond the immediate region. ... The role of an advisory board with industry, university and government technology experts could be modeled on the role a “blue ribbon” committee of visitors plays in evaluating research programs at many agencies or academic departments at most universities. Coordinated national strategies guided by advisory boards providing a big-picture perspective would ensure that, individually and as a collective network, the national reach of the Institutes was maximized.
A collaborative advisory board of national firms and industry groups with interest in the technologies, joined by government agency and university experts, is one translation model for moving regional advances to the national level. National-level workshops, or a form of “DARPA-Tech” or of ARPA-E’s annual innovation conference, hosted by the supporting federal agencies funding the NNMI, on the program’s progress could be another model. Several of these approaches could be utilized, but steps should be taken because regional implementation without national implementation could be a technology implementation barrier.
- NNMI’s are not a single policy “silver bullet;” they should be a key part of a larger and coherent policy framework.
The federal governmenthas a major stake and therefore a rolein driving solutions to America’s manufacturing challenges (see ITIF Report Premise #3).A series of relevant resources, including R&D, tax incentives, education and training support, and procurement, could be brought to light here. It will be the government’s responsibility to use its policy and resource tools for implementation of innovation policy, to prevent market failures, and to carefully consider the NNMI’s design so it is substantively and politically sustainable.
The NNMI alone is not a panacea. Aside from the IMI’s, there is a range of additional innovation policies needed at the macro policy level (taxes, trade, regulatory reform, etc.) and important innovation policies in advanced manufacturing R&D, education, training, and production scale-up support. These policies, however, must be defined and rationalized in a manufacturing context. For example, as noted above, many of the most important advanced manufacturing technologies (advanced materials, robotics, and “smart manufacturing” technologies, for example) are simply not ready for the NNMI testbed step without significantly more R&D input – NNMI’s built around them could fail because of this technology readiness problem. In parallel to NNMI’s there must be a feeder R&D system in advanced manufacturing, with industry/university/government strategies and collaborations.
- NNMI’s address a classic economic “market failure” problem, but further consideration of how this program will be supported over time is required.
The ITIF has characterized the market failure phenomenon in its report:
Markets fail to adequately incentivize manufacturing innovation, particularly process innovation. It is widely acknowledged among economists that successful innovations yield benefits for competitors, suppliers, and consumers as well as the innovating firm. These ‘spillovers’ are a disincentive for investment [by particular firms].
The NNMI aims to fill the gap between research and technology implementation. They should enable new advanced technologies and processes to be implemented more rapidly particularly by the 300,000 small and mid-sized manufacturers in the US, who comprise the majority of the manufacturing sector and are ill-equipped to implement such new processes and technologies on their own. They substitute a collaboration model to resolve the market failure problem faced by individual firms in sponsoring innovative R&D, with the “spillovers” shared by the participants. In theory this should assure collaborative support for NNMI’s, but will it?
In order to meet the White House goal of self-sustained IMI’s within 5-7 years of their launch, further consideration should go into the design of the support system for these institutes. For example, lessons are available from programs such as the Advanced Technology Program (ATP), which was renamed and restructured as the Technology Investment Program (TIP) in 2007, and has now halted. While its purpose was to bridge the “valley of death,” where research ideas fail to develop into commercial technologies, this program faced a number of support design challenges.[5] The Manufacturing Extension Partnership (MEP) at NIST, in contrast, offers a model for successful support design. Its goal to bring the latest manufacturing technologies and processes to small manufacturers around the nation and is strongly supported by both states and by regional manufacturers. How such support, involvement and commitment could be ensured for NNMI’s should be considered in the program design. Regardless of how strong the substantive design is for the NNMI’s, unless the support design both supports a quality program and brings ongoing backing and advocacy for the NNMI program, it will not endure.
- A range of cost sharing should be considered rather than a fixed number.
According to the ITIF report, “manufacturers should generally provide 50 percent of the resources for each IMI, with federal and state (or other regional) co-investment comprising most of the balance.” Cost-sharing dampens the risk that comes with innovative technologies, and is generally advisable.However, some technologies are further along than others – the higher the technical risk the lower the cost share should be. Thus, a range of cost sharing based on technical risk and corresponding industry interest should be considered, not a fixed number.There may be some technologies so critical to advanced manufacturing that a zero or very low cost share is imposed. In addition, there needs to be more work on the term for federal support.Although federal support can be curtailed over time, consideration should be made to phasing this and imposing it gradually, with exceptions made when highly promising technology is still emerging. Not every technology will fit the White House’s five-year support yardstick. If federal support is phased out over time as the White House suggests, work needs to be undertaken on models for some sort of return for industry partners, including recommendations on how IP will be handled among NNMI participants. Again, NIST should keep in mind that different advanced manufacturing technologies will have different implementation pathways, and one size doesn’t fit all when it comes to NNMI models and transition from federal support.
- NNMIs require a concept of “place”, but should be designed to allow a connection with outside researchers.
IMI’s will work in a locality that closely connects them to regional innovators and firms. However, they also need access to the best talent. That talent will not necessarily be based in a single region, so IMI’s need to find a way to access it. For example, DOE’s Energy Innovation Hubsto date have not yet adequately recognized the need for researchers outside of the Hub to connect and collaborate with the Hub, which can be a limiting factor on the technology advances the Hub can pursue. The NNMI should learn from this problem, and the White House’s suggestion for inter-collaboration among IMI’s is one way to address it. The NNMI system as a whole can’t solely fund in-house work, it needs to provide links for critical advances from outside each Institute to enter.
Conclusion
The design of the NNMI program should consider a series of additional design questions beyond those discussed thus far. For example, IMI missions must reach outward as well as inward to the direct participants. They also need to align, for example, with government research and service agencies other than just those putting direct funding into these organizations. Thus, designers of the NNMI need to define to what extent NNMI’s must be both autonomous as well as collaborative and connected. A series of additional program design features should be considered:
- The NNMI role as “testbeds” for technology demonstration is an important one and should be an explicit element in program design.
- NNMI’s should focus on advanced manufacturing technologies with broad impact and synergies across a series of significant industrial sectors, not on a single industrial sector.
- NNMI launches should initially encourage pilots and design experimentation.
- It is important to consider the role of federal R&D in the NNMI mission.
- IMI’s need to be designed to foster the translation of regional results to national-scale implementation.
- NNMI’s are not a single policy “silver bullet;” they should be a key part of a larger and coherent policy framework.
- NNMI’s address a classic economic “market failure” problem, but further consideration of how this program will be supported over time is required.
- A range of cost sharing should be considered rather than a fixed number.
- NNMIs require a concept of “place”, but should be designed to allow a connection with outside researchers.
-- Maggie Lloyd, MIT Washington Office