To President Barack Obama’s auto industry task force, as well as to most commentators on the industry, the main problems that General Motors and Chrysler faced this year were problems of cost. If the companies could shed mountains of debt, lower their labor and retirement costs, and cut dealers and brands, the standard story went, then they could once again become competitive. Summarizing his conditions for assisting GM and Chrysler, Obama said, “Now, what we’re asking for is difficult. It will require hard choices by companies. It will require unions and workers who have already made extraordinarily painful concessions to do more. It’ll require creditors to recognize that they can’t hold out for the prospect of endless government bailouts…it will require efforts from a whole host of other stakeholders, including dealers and suppliers.”

But the biggest obstacle to the prosperity of the Detroit-based automakers is not high costs but insufficient innovation. For too long, U.S. automakers failed to adopt world-class production methods that would bring quality up to the standards of their best foreign competitors. They introduced hybrids and other alternative-powered cars too late. And they did not cultivate the innovative capacities of their suppliers that make most of the components of their cars. The results have devastated not only the American auto industry, but also the Great Lakes regions that depend on it.

Federal policymakers’ failure to see the domestic automakers’ innovation problem cannot be blamed on the shortsightedness of individuals, nor was it unique to the auto industry. Instead, the problem was structural: Almost uniquely among economically advanced nations, the United States has no government agency or program responsible for innovation as a whole, either for a particular industry or for the entire U.S. economy. Innovation is the key to economic growth and rising standards of living. Yet there is perhaps no other issue of comparable national importance on which federal policy is so lacking. It is high time for the United States to establish a National Innovation Foundation—new, nimble, lean, and collaborative agency devoted entirely to helping businesses innovate.

For about four decades after World War II, there was no need for the federal government to worry about how innovation occurred. The U.S. economy was so dominant that there was little external, competitive pressure to cut costs—and thus R&D spending—and little chance of major innovation occurring outside the United States. Innovation was industry-driven, with leading manufacturing firms, such as the Detroit Three automakers, doing most of the work. For the most part, they required only limited and indirect government support to make their innovation systems work. These firms would do much of their research in-house and use it to develop new products. With the United States the clear leader in technology-based industries, there was no question that both the R&D for and the manufacturing of these products would happen domestically. Leading firms would make most of the components for their own products and, when using (mostly domestic) suppliers, would design the components they wanted suppliers to make. With a product mix that changed only slowly (in part because of the market power of leading firms), managers and engineers would make marginal improvements to production processes and production workers would implement them, leading to steady growth in productivity and living standards.

The federal role in innovation at that time was limited to funding basic research, subsidizing the education of scientists and engineers (key inputs into the innovation process), and maintaining a system of intellectual property law (defining property rights so as to promote innovation). To be sure, commercial innovations have resulted from federal efforts (with the Internet, originally an outgrowth of a Defense Department program, as the leading recent example). But when innovation results from these efforts, it is usually their by-product rather than their purpose.

In the last three decades, however, the U.S. innovation system has changed radically. Leading firms now face more product market competition from abroad and more frequent changes in consumer demand at home; consequently, they are less willing to make long-term, risky investments in basic and applied research. With many countries now offering competitive conditions for production and increasingly for R&D, innovation by leading U.S. firms often means that production and R&D occur abroad. Meanwhile, large manufacturers, including automakers, now rely heavily on smaller suppliers to design and produce much of what goes into their products—in cars, that means everything from radios to seats. It is now more important for smaller firms to innovate; yet many are not well equipped to do so. As a result of all these changes, firms are left with fewer incentives to innovate and little external support to help them do so.

The federal role, however, is largely what it was in the old system: funding inputs to and defining property rights for innovation. Although a handful of federal programs work directly with businesses to promote innovation, they are small and, despite welcome budget increases in the Obama Administration’s first year, they remain underfunded. To be sure, virtually all state governments and many local governments have established technology-based economic development programs to promote innovation. But states and localities do too little (because the benefits of innovation don’t stop at state lines) and often focus on the same high tech industries (such as information technology and biotechnology), regardless of whether their regional economies are well suited to those industries.

America’s relative decline is due not only to other nations catching up, but also to gaps in our ad hoc innovation system. Although businesses increasingly do their research in partnership with universities, the research priorities of universities and business firms often differ, hindering the coordination necessary for those partnerships to work. And even when they do work, corporate-academic partnerships generally involve only one firm, limiting the diffusion of their benefits throughout the economy. Large manufacturers’ growing reliance on smaller suppliers pushes problems of innovation and productivity improvement onto smaller firms that often lack design capabilities and lag in adopting the best technologies and methods of organizing work. Although it is now essential for firms to involve production workers in continuously improving production systems, many firms do not do so because they are unwilling to give up managerial control or because their managers do not know how to change the way they manage.

Perhaps most important, the American innovation system at present does not have a means of coordinating different kinds of innovation. In the old system, individual large firms made most of the decisions that resulted in innovation. In the new system, responsibility for innovation is dispersed among many large and small firms, universities, and state and local governments, and there is nothing to ensure that the sum of all their efforts is in the interest of the nation as a whole. And the market won’t do the job because none of these actors is able to capture the full benefit of its innovative activity.

To fill the gaps in America’s de facto innovation system, the federal government should create a National Innovation Foundation (NIF), akin to the National Science Foundation. An NIF would support commercial innovation directly, not just by funding inputs to innovation. It would work with firms and other economic actors, such as universities, unions, and state and local governments. It would help firms do more basic and applied research by offering matching grants to a consortium of companies, to do university-based research that would chart out and overcome the barriers to technology commercialization that all the firms in the consortium shared. It would offer states matching grants to do more and better innovation-focused economic development. It would offer similar grants to self-organized groups of businesses, educational institutions, local governments, and other organizations to support industry clusters within metropolitan areas and smaller economic areas. To help manufacturing and service firms of all sizes adopt best practices in technology, work organization, and business relationships it would, in partnership with state governments, operate technology diffusion centers in each state (this effort would build on but go beyond the federal government’s current Manufacturing Extension Partnership program). All NIF programs would require that funding recipients direct their innovations toward substantial R&D and production in the United States, so that it did not contribute to the outsourcing of either.

An NIF would be the federal government’s central agency for formulating innovation policy. It would measure, research, and advocate for innovation. The knowledge that it would gain in this way would enable it to be more than a combination of innovation-promotion programs. It would not try to pick particular industrial or technological “winners” but instead would solicit proposals from firms and governments and evaluate each proposal on its own merits. It could even fund competing research programs to see which technology works better.

What would be the effects of such a broad-based approach? The auto industry provides a perfect example of why innovation-promotion programs and innovation research and policy need to be brought together in a single agency that has the knowledge and ability to coordinate them. State governments and regional groups could use NIF grants to figure out what kinds of non-auto work their auto suppliers could realistically do to replace lost demand from automakers. They could then use the grants to assist those suppliers in bidding on and performing that work. Auto-dependent Great Lakes states are already trying to do these things on their own, but despite the similarity of the problems they face, they are pursuing these efforts independently, and sometimes at cross purposes, For example, some states are as interested in poaching auto suppliers from their neighbors as they are in retaining and diversifying their own suppliers. NIF funding would make such counterproductive state efforts less likely.

An NIF would have recognized that the U.S. automakers have long faced problems of both product quality and more far-reaching product innovation. It could have been an early warning system, publicizing these problems and their consequences for national and regional economies well before GM and Chrysler’s recent financial troubles began, and suggesting solutions. It could have helped those companies work with their unions and suppliers to adopt the waste-reducing “lean production” methods that use workers’ knowledge and skill effectively and underlie the high quality of such world-class companies as Toyota and Honda.

An NIF would also have understood how the stages of the innovation cycle, from R&D through production, were connected, and it could have used this knowledge to improve the way automakers and suppliers behave at each stage. For example, it could have funded a proposal to research the technological, economic, and organizational obstacles to the mass production of all-electric cars and another proposal to do the same for hydrogen-powered cars. Understanding those obstacles and the prospects for overcoming them could help NIF work with the producers of alternative-powered cars to anticipate and solve problems that arose in the production process.

An NIF would have been able to assess and respond to the impacts that alternative automotive technologies, such as electric or hydrogen power, could have on the innovative capacity of auto-dependent regions. Would existing auto suppliers in those regions be able to adapt to such technologies, perhaps with NIF assistance? Would new suppliers be likely to locate in those regions or elsewhere? Is the entire supply chain likely to be reconfigured to make auto parts production more modular (like computer hardware production), potentially reducing or eliminating the economic advantage of existing auto-dependent regions to the auto industry?

Finally, an NIF would have viewed the Detroit Three, their suppliers, and the Great Lakes regions in which their facilities were located as components of a single innovation system, albeit one greatly in need of improvement. It would have been able to assess the prospects for suppliers and auto-dependent states and regions to replace part of their auto work with other kinds of manufacturing. It would have been able to offer assistance to help them do so. To the extent that alternative kinds of manufacturing would not fully replace the shortfall in auto-related work, an NIF would have been able to evaluate the consequences for innovation in the Great Lakes regions and in U.S. manufacturing as a whole. It then would have been able to recommend a federal policy toward the Detroit Three that was informed by an assessment not only of the policy’s impact on those firms but also of its impact on innovation more broadly.

If there had been an NIF, would the President’s auto task force have taken a different approach to the problems of GM and Chrysler, perhaps conditioning federal aid to those companies less on short-term cost reductions and more on quality improvement and product innovation? Or, given the Obama Administration’s decision to take a more hands-off approach to the latter, would the new managements of GM and Chrysler have used NIF assistance to help their companies succeed? Perhaps. At the very least, though, an NIF would have given innovation a more prominent—if not decisive—place in the policy debate about the auto industry.

The benefits of an NIF would extend well beyond the auto industry and manufacturing, though. Low productivity growth, a consequence of a lack of innovation, plagues many industries, especially service industries. Because the innovation challenges are often both technological and organizational, an NIF is ideally suited to address them. Consider health care. Electronic medical records have the potential to improve the quality of health care and keep costs down, but implementing them has been a challenge even within individual hospitals. Or take residential construction, an industry plagued by an outdated form of business organization and a reliance on low-skilled workers. If public policy is going to encourage green building materials for both retrofitting and new construction, then the industry may not be well equipped to respond. In both industries, an NIF could combine research, technology diffusion, and industry cluster grants to identify the obstacles to innovation and help firms overcome them.

Without the direct federal boost to innovation that an NIF would provide, productivity and American workers’ wages will not rise as rapidly. U.S. companies will introduce fewer new products and services, and those products and services will be less likely to be developed and produced in this country. Other economically advanced nations have established effective agencies to promote innovation. It is time for the United States to do so as well.