Americans generally recognize inventiveness as one of their nation’s competitive strengths. They understand that invention is a powerful engine of economic growth. Yet it gets amazingly little direct attention or funding from product makers, universities, or the government.
—Nathan Myhrvold, cofounder of Intellectual Ventures, one of the top five owners of US patents1
THE HISTORY OF SCIENTIFIC RESEARCH, TECHNOLOGY DEVELOPMENT, AND ECONOMIC GROWTH IN the United States gives us three important lessons. First, publicly funded programs—such as the Manhattan Project, the Atomic Energy Commission, the National Institutes of Health, the National Science Foundation, DARPA, and the Apollo missions—took on tasks that the private sector would not or could not tackle.
Second, these programs created technology that boosted growth and generated good jobs for millions of Americans. The United States became the undisputed innovation leader of the world in large part due to these efforts. Almost all our major technological breakthroughs of the past half century or more—including by private-sector companies in the computer, health care, or transportation space—have their roots in those public investments.
Third, these dramatic and high-impact surges in public R&D were, for the most part, not sustained. The economic basis of these programs was compelling, but political support quickly dwindled. Funding for science has repeatedly been seen as excessive and as primarily favoring relatively few highly educated people.
The United States now faces problems that are quite different from 1941 or 1957. German work on the atomic bomb and the development of Soviet missile capability posed very real existential threats. In both instances, American fears may have been exaggerated, but given the available information—and the strong scientific infrastructure of America’s enemies—policy makers had good reason to be afraid. Germany developed long-range rockets capable of delivering deadly explosives, and the Soviet Union launched the first-ever satellite. Investing heavily in American technology, along with a commensurate expansion in educational opportunities, made a great deal of sense as a national security response. The strongly positive benefits for the broader economy were, ironically, a side effect.
Today, we again face serious external competition. Europe, Japan, and especially China are investing increasing amounts in science and technology. As a result, the rest of the world is threatening to erase our seventy-year lead in technology creation—and in some cases has already done so. Our more serious problems, however, are internal. Growth over the past several decades has been slow, and the gains that the economy has delivered have been concentrated in the highest-income groups and a few superstar cities. We need to boost our rate of economic growth in a way that creates good jobs, with decent wages, for more people and across the entire country, not just along the booming coasts.
In spring 1940, Vannevar Bush walked into the Oval Office with a half-page proposal that said, in effect: trust the scientists, and we will work out the details. Such an approach would not work today. The relationship between scientists and policy makers has changed irrevocably, and we live in an era of constrained budgets at least when it comes to this kind of endeavor. Politicians—and the public—quite reasonably want greater assurances that they would get value for their money and a broad sharing of benefits.
Rebuilding the American growth-and-jobs machine in an economically sensible and politically sustainable manner requires weaving together three major elements: more support for basic science and related commercial development; emphasis on a national strategy of developing new tech hubs as a cost-effective way to take advantage of local research spillovers and agglomeration; and a funding mechanism that results in direct and transparent returns for all Americans.
For evidence that such a strategy can work, start with Orlando, Florida.
Orlando brings lots of things to mind: amusement rides, fairy castles, and giant mice. Fairly low on the list would be computer simulation. Yet Orlando is not only a destination for family fun, it has also emerged as the world’s center of the $5 billion modeling, simulation, and training (MS&T) industry. This industrial cluster is anchored by the University of Central Florida, one of the nation’s largest universities, and the Central Florida Research Park (CFRP), which hosts 130 private companies and ten thousand employees.2
It was not always this way. The East Orange County area in which both the university and the research park reside was once a sleepy subdivision of Orange County, dominated by Disney.3 In 1980, East Orange had fewer than 40,000 residents and 17,000 workers. Thirty years later, it had grown sixfold, with more than 220,000 residents and 107,000 workers. This growth had little to do with Disney; the rest of Orange County grew, but at a much slower rate. As a result, East Orange tripled the percentage of jobs it held in the overall county, from only 8 percent to almost 25 percent.
What happened is a modern version of what once made America the most prosperous nation on earth. The federal government, a local university, and the private sector came together to create a dynamic jobs engine.
The story begins with the closing of the Orlando air force base that had been established during World War II. Whenever a military base is scheduled to be closed, there is substantial lobbying to keep it open, which usually fails. The story ended differently for Orlando—thanks to an influential resident with an important connection.
Martin Andersen, the publisher of the Orlando Sentinel, got to know a young politician named Lyndon Johnson through a common mentor. Andersen’s paper endorsed Johnson during his 1956 and 1960 campaigns for the Democratic Party presidential nomination. During Johnson’s 1964 reelection campaign, Andersen organized a motorcade for Johnson during a visit to Orlando. When Johnson offered to reciprocate with a testimonial, Andersen reportedly responded, “What I’d really like, Mr. President, is a military base.” After his visit, Johnson called Andersen and reportedly said simply, “I’m sending you a naval base.”4
Shortly thereafter, Orlando received a terrific gift. The departing air force base was not closed but rather replaced by a new naval base. Part of the new naval base was the Naval Training Device Center (NTDC), whose mission was the development of combat simulation devices.5
Coincidentally, the University of Central Florida opened in 1968 as Florida Technological University, with a mission to provide personnel to support the growing space program at the Kennedy Space Center and Cape Canaveral Air Force Station. The university grew and was renamed the University of Central Florida in 1978. A major step was taken in 1980, when University president Trevor Colbourn decided to create a research park connected to the university. At the time, with sky-high interest rates, it was difficult to sell the land near the university for residential purposes. So instead, the CFRP was established and bought slightly more than one thousand acres of land south of its campus for $2,500 per acre. The CFRP was set up as its own authority, financing its growth by issuing special purpose bonds of the type used by airports.6
The next step for CFRP was to find an anchor tenant around which the park could build. And they found the perfect anchor with the naval simulation center, which in the early 1980s needed a new building.7 Colbourn and other business leaders convinced the simulation center to move the few miles from the base to the new CFRP and to construct its new building there instead.
The navy’s move turned out to be a motivating force for other companies to locate at the CFRP. An early participant was Perceptronics, a training-simulation company, which arrived in 1987. Today, the park has ten thousand employees, with employment in the park growing at about five hundred employees per year.8
The park and the university have a symbiotic relationship as new ideas flow from the university to the park, which functions as a “real-world lab” for testing these ideas. This relationship has paid off for the university as well as the local economy. From 1979 to 2015, the federal grant dollars going to the university rose from $2.5 million to $82 million. And commitments from the state, the private sector, and other university sources followed in step; total research-and-development funding at UCF rose from $3.4 million in 1979 to $188 million in 2016. In 1978, the UCF had an enrollment of ten thousand students; today it is one of the largest universities in the United States, with more than sixty thousand students.9 In 1982, UCF had no highly ranked science departments; by 2005, its electrical engineering program was in the nation’s top twenty.
Meanwhile, entities based at this research park receive more than $1.4 billion in federal financing, mostly from the military but also from the US Geological Survey, US Army Corps of Engineers, and others.10 But despite the substantial government catalyst, the park is to a large degree a privately funded venture and a centerpiece of the country’s computer simulation industry; moreover, about 60 percent of the park workforce is involved in other activities in which the university excels, ranging from lasers and optics to medical devices to computer technology.11 And not only has the CFRP created jobs, it has created land value as well; the land in CFRP is now worth about $350,000 per acre.12 The federal government, a budding university, and an entrepreneurial private sector working together have created thousands of jobs and billions of economic value in a place known by many for only citrus and Disney.
The UCF/CFRP relationship is an economic success story in terms of developing, changing, and growing the local economy, but there are three important limits. First, the genesis of this hub was a political favor, which is not the most economically efficient way to allocate federal funds. To create a national initiative, we would recommend a more objective way of allocating research-and-development dollars to the places where they can be productively employed.
Second, the increase in research funding has not led to the development of large homegrown technology companies, due in part to the lack of financing in this area for the transition across the valley of death. A report from the National Venture Capital Association found that Orlando was the forty-sixth-ranked city in the nation in the number of companies receiving deals for financial backing.13
The companies in the CFRP have struggled to grow their size and customer base. As a result, the success of the CFRP is still overly tied to the military budget. In September 2013, for example, San Diego–based Cubic Corporation cited US spending reductions as the reason it dismissed an undisclosed number of workers from its 350-employee workforce in Orlando. Two months later, it landed a large new contract worth as much as $112 million to provide simulation training systems for an advanced navy warship, which led to an increase in hiring in Orlando.14
Finally, the Orlando area suffers from a lack of the skilled workers required to grow the technology sector. Primary and secondary education in Florida remains quite weak. US News ranks Florida fortieth for pre-K through 12 education. Florida ranked forty-third for high school graduation rate, forty-second for math scores, and thirty-second for reading scores.15 CFRP manager Joe Wallace has been told by companies that for every PhD they hire, they need ten technicians—but that such a pool of talent is not available in the area.16
The Orlando example is not an isolated one. The classic example of a place transformed by research enterprise is North Carolina’s Research Triangle Park, founded in the 1950s and now the nation’s largest research park. More recent examples include the Georgia Research Alliance (GRA). This program devotes state and private resources to recruiting top research talent to Georgia universities and to leverage those researchers for research funding and economic development. The GRA’s “eminent scholars” program has recruited dozens of the nation’s top scientists to Georgia universities, resulting in $4 billion in outside federal and private investment in Georgia. GRA efforts have led to a portfolio of 180 spin-off companies with more than $660 million in revenue and employment of more than 1,300 professionals.17
At the national level, we should build on the example of Orlando—and other places that have become strong hubs—while keeping in mind these shortcomings. The goal is to create a lasting set of successful technology hubs spread more widely around the country.
The heart of our proposal for jump-starting America is a substantial federal investment in R&D. If we devote an additional half of one percentage point of GDP to research funding—roughly $100 billion per year—we would return public funding to its level in the 1980s. Based on history and available evidence, this investment would lead to a significant growth boost, arising from more invention and faster productivity growth.18
The evidence reviewed in Chapter 5 shows that past expansions of public R&D have been a cost-effective way to increase employment. On the high end, our own regression estimates of the impact of more university R&D funding on jobs suggest that university research spending raises employment at a cost of $28,000 per job, while the study of the New Zealand public R&D program increasing employment implies a cost per job of $29,000.19 On the lower end, the study of the Finnish Tekes program implies a cost per job of $8,100, while the study of military R&D in Europe suggests a cost per job created of only $2,100. This entire range is quite low, relative to what it costs to create jobs in other contexts. For example, existing estimates of the jobs created by stimulus spending in the Great Recession suggested a cost per job created of about $50,000.20
If we conservatively assume that expansions of research and development create one additional job per $25,000 in spending, then an investment of $100 billion per year would create four million good new jobs. This would be a major step toward addressing the shortfall in quality jobs in the United States.
Alternative calculations produce similar numbers. Orlando receives $1.4 billion in government funding for its computer simulation industry. Since 1980, if the East Orange County subdivision had grown at the same rate as the rest of the county, it would have had thirty-eight thousand new jobs. Instead, it added ninety-one thousand new jobs. If we divide the government funding by the number of extra jobs created in East Orange County relative to the rest of the county, we get an estimate of $26,770 per job.
This is just the tip of the iceberg. The really big potential gains come from being at the forefront of the next wave of global superstar technologies. No one knows exactly which technology will be the next blockbuster, but there are plenty of candidates. Being the first to develop radar, jet engines, and the internet was worth a huge amount to the United States in terms of good jobs created and stronger national security.
What we need is a portfolio of high-risk and capital-intensive research and development that can lead to broad-based economic growth by creating high-paying middle-class jobs. We should focus on funding science that looks likely to have a strong return in terms of creating sustainable growth into the future. And the possibilities are indeed endless. In Chapter 8, we highlight promising potential developments in sectors such as synthetic biology, alternative energy, and ocean exploration, but keep in mind these are just a few examples of what may be possible in the near future.
As we noted with the Human Genome Project, a relatively small ($3 billion) federal investment gave birth to an industry that created good jobs for hundreds of thousands of people. But, as we will discuss in the next chapter, other countries are already moving to take advantage of these opportunities. If we want these new jobs to be in the United States, we need to take the lead in developing the technologies.
There is a successful history of organizations such as the military and NIH using government grants to finance research ranging from radar to the human genome. Our initiative for jump-starting America, however, is focused not just on creating knowledge but also on increasing economic growth and the number of good jobs. These additional steps require translating the discoveries made by researchers into products that are valued in the economy.
We reviewed the problem of transforming research into products (here), especially the valley of death faced by new ideas in getting from the laboratory to the prototype. We talked about barriers faced by innovative new firms, from economies of scale in manufacturing to a venture capital sector that does not prioritize large-scale and long-term investments.
During World War II, America solved this problem by just having the federal government pay for both the research and product development. Initiatives such as the Rad Lab and the Manhattan Project were not just about the basic science but about taking the ideas all the way to products that the military could implement quickly. Such an approach does not make sense today, because the goal is not the development of weapons per se.
The focus today must be on partnering with, not displacing, the private sector. Recall our discussion of the limitations of the venture capital sector in the United States: limited funds lead them to avoid the valley of death. But the potential problem when the government tries to fill in these missing areas is that it can end up competing with the private sector—and that is a competition the government is bound to lose, in terms of finding the best investments to finance. Private-sector investors have the expertise and the incentives to pick the best investments—so the projects the private sector passes on are likely to be less productive unless government officials start to strong-arm people or hand out bigger firm-specific subsidies.
Entrepreneurship expert Josh Lerner confirms the valuable role of some public initiatives to promote R&D, but he also discusses some compelling examples of failed government initiatives: Malaysia invested in a massive bioscience complex that is now known as the Valley of the Bio-Ghosts, and Norway “squandered much of its oil wealth in the 1970s and 1980s propping up failed ventures and funding ill-conceived new businesses by relatives of parliamentarians and bureaucrats.”21
Lerner suggests that a better solution is to address head-on the problem of limited capital available to invest in start-up companies, by partnering with private investors to find the best opportunities to fill the valley of death in a way that is likely to yield long-run returns. Public partnerships can increase the compressed time frame of existing venture capitalists that leads them to be unwilling to invest in long-run bets. Our review of the VC sector showed that when capital is more readily available, VCs are willing to fund longer-run, riskier projects—resulting in more innovation. The government can operate strategically in partnership with VCs to loosen those capital constraints.
Experienced venture capitalists could also play an important leadership role in this initiative. The US venture capital sector is the envy of the world—we should be harnessing its strengths.
Crossing the valley of death most effectively means coordinating between scientists, manufacturers, and financiers. And despite the advances in communications technologies of the past thirty years, such collaboration still happens best face-to-face. Research and development yields its highest return, for example, in areas where venture capital is most readily available. Venture investors, however, naturally prefer to locate where there are already a good number of start-ups and potential deals for them to look at—leading to the excessive concentration of VC dollars in just a few cities today. That is why we need a geographic focus for our new investments—through the creation of innovation hubs.
A long-standing feature of both private tech development and public R&D spending is the focus on a few outstanding locations. Cities like Boston and San Francisco have all the preconditions for science-based success, from world-leading universities to vibrant venture capital communities. Naturally, these places have become hubs for US innovation and the focus of federal government funding as well. This pattern shows no sign of slowing, with the top metro areas pulling further and further away from the rest of the country.
There is a strong economic argument for this research concentration. Public research dollars will be most effective when provided to the scientists with the best ideas. This suggests that research should be allocated nationally through a competition in which the best scientists win. Should those scientists be concentrated in a few cities, so be it.
On the other hand, there may be many high-quality research ideas that are being ignored in other parts of the country because there is less research funding available in those areas. There are certainly restrictions on taking existing ideas to scale because of the limited reach of private venture funding in these places. When it comes to establishing new hotbeds of innovative job growth, we face a chicken-and-egg problem: venture capital investors are reluctant to focus on places without a large existing tech presence, but it is hard to build a tech presence in today’s economy without VCs willing to fund the new enterprises.
At the same time, conducting research in the existing small set of coastal locations is unambiguously a lot more expensive than doing the same in lower-cost locations elsewhere in the country. The constraints that local real estate regulations place on property development have costs for research and development, as projects become increasingly expensive to carry out, and it is hard to find affordable places to live for the workers—of all education levels—who are an essential part of the research infrastructure.
A centerpiece of our proposal is to spread the availability of public research funding more broadly across America. Based on existing evidence regarding the complementarity between private and public research, if we expand publicly funded research in new places, private research and development dollars will follow. Places like Orlando—in terms of tech potential—exist all over the United States and are well positioned to become hosts to the next breakthrough technology. Great universities, talented residents, and productive business environments exist alongside reasonable housing costs and a high quality of living all around the country. In Chapter 6, we reviewed the wide variety of places in the United States that meet these conditions.
While there may or may not be some economic costs to redirecting public R&D toward new locations, there are unambiguous political gains. More government funding directed to a small subset of already successful places will be significantly less popular than ensuring that the whole country is represented in any new research initiative. If additional public R&D spending simply follows existing patterns, then most of the country will be left with little to show for a sizeable expansion.
Senator Harley Kilgore had a version of this insight in the late 1940s, but Vannevar Bush pushed back—most likely contributing to the creation of a smaller federal research enterprise than might otherwise have been possible. It would be unwise to ignore the regional pressure for scientific and economic opportunity today. We do too little research and development in America. However, if we want to do more, there needs to be political buy-in. A hypothetical reduction in the efficiency of R&D spending is a small price to pay for a large expansion in the pipeline of new ideas.
However, this won’t work if the dollars are spread too thinly either. The agglomeration effects that we discussed earlier are not going away. To capture the benefits of agglomeration, places must create a compelling case for skilled workers, researchers, and investors to locate there. Incremental investments in the research infrastructure of an area are unlikely to have this effect.
What is needed is a big push: a major leap forward that announces to the world that these new locations are ready and qualified to become technology hubs. This means picking winners and not simply succumbing to political pressure to give money to any qualified city. To do so, we need to follow the lesson of Amazon and other successful US companies.
Political advantages notwithstanding, there is a huge risk to having Congress and the White House simply pick the next research hubs. The Orlando technology and research center got its beginnings from a political favor by then president Lyndon Johnson, and while it turned out well for this area, relying more generally on political whim and favor would be unwise. While it is desirable to introduce new places in the United States into the forefront of technology research and development, it only makes sense if those places are ready. Otherwise, this simply becomes a federal transfer program to politically favored areas and does not maximize the potential for economic growth for our nation as a whole.
We recommend spreading research-related dollars wider than the existing superstar cities to which they flow today. At the same time, we want dollars to go to places where they can be employed productively, not just where powerful politicians call home. And we want sufficient investment in areas so that they can get over the hump of agglomeration economies and become a desirable destination for the technologies of the future. We propose to resolve this tension by proceeding exactly as private companies do, through a competitive process.
In a catalyst competition, areas would apply to become one of the new hubs. Evaluation criteria should be based on measurable dimensions consistent with the idea that an area could become a new center of discovery and job growth. This is discussed at more length in Chapter 6, and our THIS index provides one illustration of how this might be done, although many variants of this approach are possible.
In addition, places would have to demonstrate substantial local buy-in along several dimensions. One is having pro-growth zoning regulations. If the technology hubs are successful, they will become larger and denser urban areas that are attractive both for businesses and individuals. This raises the risk that we re-create in these new hubs the model of restrictive regulations in the superstar cities that have led to high house prices and slower-than-expected job growth. Areas must have a long-run plan to promote sensible growth that allows sufficient affordable real estate within reasonable commuting distances of the new research centers.
The recent string of natural disasters, such as the experience with Hurricane Harvey in Houston, have highlighted, of course, the perils of zoning regulations that are too loose. Clearly, sensible restrictions are needed. There is a large middle ground between the overly tight zoning restrictions that have made the Bay Area and Boston unaffordable places for many people to live and the overly loose zoning rules that led to environmental disaster in Houston. An advantage of the competition for hubs is that we would encourage areas to look for that sweet spot in their urban planning—or they pass up the chance for this new opportunity for local growth.
Another is having a successful infrastructure plan for building a hub that both promotes basic research and its development into commercial products. This means, for example, having strong transportation networks within the area to facilitate interaction and from the area to key markets where commercialized products might be sold.
The third is a plan for building and sustaining an educational base that can support the growth of a new technology hub. Areas need to show that they can not only support growing demand for skilled jobs but also ensure sufficient supply of workers to take those jobs. In part, this will be university-based; there are many excellent universities around the nation that could be drawn on for this endeavor. But this will also require higher-quality high school, vocational, and local college education to train the workers that will support research scientists in creating the products of tomorrow. We discuss such an approach later (here). In competing to be a technology hub, areas will have to show commitment along many dimensions to raising the supply of skills in their area.
Competition between areas to attract new business is hardly novel, and large companies play localities against each other on a regular basis. Amazon used a version of exactly this process in determining where to locate its HQ2. The problem is that the existing competition across states is a zero-sum game that simply transfers to shareholders of major companies much of the gain they bring to a city. States and cities already spend almost $50 billion per year in tax breaks that may or may not make sense for their location, but which leaves the nation as a whole less well off. In contrast, our proposal creates a positive-sum game within which areas benefit from better jobs and higher productivity, while the country as a whole wins from increased innovation.
The federal government already uses a version of this approach, though on a smaller scale. For example, the US Army’s recent choice of Austin for its high-tech Futures Command was the result of a process that started with a list of 150 cities and then whittled down to 5—very much along the lines of the Amazon-type corporate location process. Key criteria included distance to people with math and science expertise, as well as private sector (including academic) history of research and development. Naturally, the army also considered quality of life, including the cost of living, and what kind of support was available from local government. The University of Texas reportedly provided space in a downtown office building, making it easier for the army’s Futures Command to potentially interact with local technology companies.22
A major issue that will be raised with such a new large commitment is governance. How do we make sure this does not turn into a congressional (or executive branch) boondoggle, with funds handed out as political favors rather than rewarding the most deserving research projects?
We agree that this is a serious issue. It is critical that this new program be administered through a new independent entity that is not part of either the administration or Congress. It should follow the structure of the Base Realignment and Closure (BRAC) Commission, which has, over recent decades, successfully reduced the number of military bases.
The BRAC is an underappreciated mechanism for making the kind of hard decisions that seem to constantly bedevil our paralyzed and polarized Congress. Following the end of the Cold War, the United States realized it had substantial excess military real estate but that it would be hard to shutter this capacity since elected politicians have a strong parochial incentive to keep local bases active. The BRAC process was introduced to address this problem in 1988, with rounds in four subsequent years, the most recent being 2005.
The BRAC Commission consists of nine members appointed by the president. The Department of Defense presents this commission with a list of bases to be closed, and the commission reviews and modifies this list before presenting it to the president. Along the way, there is substantial scope for public input to the process, with the 2005 committee hearing from hundreds of public officials and receiving around two hundred thousand pieces of mail from private citizens. All nonconfidential information and proceedings were provided to the public, and each site on the list had to be visited by at least two members of the commission. If the president approves, the full set of changes goes into effect—unless the entire list is rejected by Congress.
In a world where it is hard to take on entrenched interests, the BRAC Commission was an unqualified success. These rounds resulted in the closure of more than 350 installations that were no longer necessary.
BRAC is our model for the creation of an Innovation Commission (IC) that would make recommendations to Congress for areas that would receive funding to establish the new hubs. The commission would make a recommendation on which Congress could vote up or down, but with a presumption of acceptance. This commission would be fully transparent in all its work. Commissioners would be appointed for a fixed period of time and would be charged with generating a financial return for society as a whole—in the form of an innovation dividend, discussed later. This financial return will provide a clear measure of whether the commission does its job properly.
To remain both politically viable and economically productive, the Innovation Commission has to address head-on the issue of failure. While, as we have shown, there are high returns to science-based investments, there are also high risks. The real benefit comes from a small number of investments yielding extraordinary returns. Recall (from Chapter 4) that even for venture capital investors, who currently get to pick the best bets, a small minority (8 percent) of their investment dollars result in the vast majority of their overall returns, while three-fifths of their projects do not even make back the cost of investment.
If we really aim to improve our existing system for converting good ideas into a more productive economy, we must also recognize that many projects will fail—and the ones that succeed may take a while to do so. Unfortunately, as the example of Solyndra (here) shows, our political system provides too many incentives to jump on failures without acknowledging the big-picture path toward success.
A proper means of achieving the balance between long-run risk taking and accountability is to incorporate effective evaluation of the initiative. While we would be investing in projects whose success is only revealed over the long run, constant evaluation and readjustment are possible. The IC must use careful and objective evaluation of performance to adjust how public resources are allocated.
The technology hubs would ultimately propose the most effective mix of spending to maximize the returns in their areas, but their spending would be broadly focused in several categories.
The first is basic research. The United States has a high-quality peer-review mechanism for evaluating and supporting scientific research through organizations such as the NIH and DARPA, and the stories and evidence throughout this book—ranging from lifesaving drugs to the internet—testify to their success. Ongoing peer review would be a central component of allocating new dollars.
The second is development. This importantly includes providing the manufacturing infrastructure to take ideas from the lab to the market. No longer should the United States lose the ability to develop innovative and job-producing technologies to other nations because of a lack of capacity to develop those products. Scaled-up research hubs provide a perfect mechanism for coordinating the manufacturing needs of related projects to overcome problems of economies of scale, as well as for internalizing the spillovers of developing innovative new generalized R&D assets that benefit all producers in the area. One example could be publicly funded manufacturing resources for developing new and innovative drugs, addressing the shortfall that we highlighted with cell and gene therapies. Other examples are highlighted in the next chapter.
The third is financing. As shown in Chapter 5, programs such as the Small Business Innovation Research program have been successful on a fairly modest scale in helping companies cross the valley of death. The government could dramatically increase the size of SBIR-like programs, while partnering with the private sector to ensure that we supplement, and do not displace, private-sector support for innovation.
Fourth, at the early stages of developing technology hubs, spending would be focused on infrastructure. New research hubs require the proper infrastructure to both do basic research and to convert that research to new products. In the competitive proposal stage, hubs need to propose what they would undertake in terms of building research infrastructure, schools, business parks, and other amenities to promote their areas. In addition to building strong infrastructure, this stage has the advantage of employing some workers for whom retraining in technological skills is not cost effective.
Finally, a major target of spending would be on improving technical and scientific education at the educational institutions associated with research hubs.
As discussed in Chapter 2, the twenty years after World War II was a “Goldilocks” period where both a rise in demand and supply of skills led to a rise in pay throughout the income distribution. A major push on publicly funded research and associated development drove up the demand for skills. At the same time, the expansion of science education in our nation’s primary and secondary schools, as well as the availability of low-cost higher education through the GI Bill, raised the supply of skilled workers. Put these developments together and you get the creation of a high-wage middle class built on technological advancement. In the last few decades, however, the rapidly rising demand for skills has not been met with increased supply, leading to rising inequality.
Our plan so far would create the demand for skills, but without an adequate increase in supply, the plan will boost the wages for already-skilled people only, leading to rising inequality. To create genuinely shared prosperity, we need to raise the supply of skills as well. This is an area that has received much more attention than the failure of public research and development, with proposals from groups as disparate as the Brookings Institution, McKinsey & Company, and the Trump administration.23 Several key ideas stick out as highly practical, politically feasible, and economically important.
Increasing the supply of skills starts with investments before college. The 1958 National Defense Education Act led to a large increase in science education across high schools in the United States. We need to reinvest in the type of skills training for high school students that allows them to succeed through a variety of channels, from technical schools to community colleges and four-year colleges. As noted earlier, as part of the innovation hub competition, areas would have to demonstrate a commitment to skills training starting in high school.
A second step in increasing skills supply is making college education affordable for the middle class. Faced with extreme fiscal pressures, states are raising tuition at the state universities that are main avenues of higher education in the United States. While elite private universities are removing financial barriers through more generous financial aid for some people, their admissions barriers rise exponentially as students from around the nation and the world seek attendance at these institutions.
For many students, college means taking on significant student loan debt to complete their education. The average student graduating a four-year college in the United States today leaves with a debt burden of around $30,000.24 The student debt load in the United States is more than $1.4 trillion, and the total cost of college each year is over $500 billion.25 Many individuals who take on these loans go to schools that do not provide productive career opportunities, as witnessed by recent scandals at a set of for-profit universities; many others take large loans to pay for college but never complete the degree that can provide them access to the skilled labor market.26
One option to increase skills supply could be financing for students to study in the universities associated with new hubs. This would include expanded access to student loans, as well as targeted grants to students studying at those universities in the types of science fields that provide training for future work in the high-skill economy.27 Importantly, this should be combined with strong ongoing support for students once they enroll in college, to promote not only college attendance but also completion.
It is imperative to incorporate vocational training as well. It is unrealistic to think that every job that will be created through an initiative such as this will be a high-value-added research job. A larger structure is required to make the research endeavor a successful one, and that larger structure will require a variety of jobs. They will range from semiskilled jobs such as lab technicians to less-skilled jobs such as maintenance staff. The educational plan for the area should incorporate the ability to provide the training that is needed for these jobs as well.
Finally, it is critical that businesses work together to provide the type of general skills training that will benefit entire industries, not just specific businesses. Business investment in training suffers from the same type of spillover problem we discussed with R&D: when businesses train workers, they are providing skills that may be valuable not only to their businesses but to others. As a result, businesses may underinvest in training to avoid paying for skills that workers can just take to other jobs.
Again, cell and gene therapy manufacturing highlights these barriers. For the types of positions that are likely to open up in greater numbers in the near future, college graduates are (according to one expert) “overqualified, but underskilled,” suggesting that firms are likely to turn to apprenticeship training. However, demand from a single firm alone would not warrant establishing a large training center. Instead, “it will be necessary to aggregate the demand for apprenticeship training across cell therapy employers so that the number of trainees exceeds the minimum required to make it worthwhile for a provider to offer training, and also to reduce the risk faced by potential training providers.”28
Public-sector involvement can help fix this underinvestment problem also. One form of such involvement is financing institutions that provide midcareer skills development for workers; as a recent report from the Council of Economic Advisers emphasizes, public education spending peaks at age fifteen and is largely nonexistent after age thirty. There are a host of private and public partnerships for worker skills development that appear successful and could be greatly expanded.29 Areas bidding to be innovation hubs could propose coordination strategies across businesses and such institutions to provide the type of ongoing training needed for successful skills upgrading.
Remember that this is not just education for the future discovery of new technologies. The higher education sector has been an engine of job growth for decades in the United States, and we should build on that success.30
A realistic plan must include provisions not only for how the funds get spent but for how they get raised. In particular, an initiative such as this one requires not only politically independent allocation of funds but independent financing as well. If the financing of this initiative is subject to annual political debate, it will cause difficulties.
The danger is that the financing could become a political bargaining chip, which could interfere with the independence of the Innovation Commission. For example, politicians could condition appropriations on selection of particular sites for the innovation, which would interfere with the most productive set of sites being chosen for the project.
In addition, this project represents a long-run investment in select areas around the United States. In return for this financing, these hubs are committing to major structural changes, ranging from zoning law changes to infrastructure development. Cities will be unwilling to make such a commitment if the proposed funding is not fully guaranteed for many years.
As a result, successful implementation of this program would require a onetime, multiyear authorization that would provide independence and financing certainty to the initiative. Congressional approval is needed for each round of hubs, just as it is needed for a round of base closings recommended by the BRAC. But once a round of hubs is authorized, there is no second-guessing or congressional fiddling around the margins.
Moving out of the annual appropriations process will be challenging, but precedents do exist. For example, multiyear appropriations provide that obligated funds are available until some future date; examples include appropriations for military construction, although these typically last about five years. There are also examples of no-year appropriations, which are available “until expended,” including appropriations for the Federal Aviation Administration to purchase an aircraft, or appropriations to the AIP (Airport Improvement Program) that provides grants to public agencies or private entities for public-use airport projects.31 We are proposing a scale that is beyond these examples, but the structural precedents are in place.
The American people have benefited enormously from past public investments in research and development—but nowhere nearly as much as they could have. The federal government has spent billions of dollars on R&D that has directly led to commercialization of goods from prescription drugs to cell phone apps that use GPS. The companies producing these goods have hired millions of workers and paid billions in taxes, making America richer.
But continuing to rely on such indirect returns to public investment is problematic for two reasons. The first is that the returns are increasingly concentrated in a smaller and smaller set of wealthy entrepreneurs. From the end of World War II through 1970, the share of GDP going to workers as compensation rose from 54 percent to over 58 percent, but it has steadily declined since, and it is now back at its pre–World War II levels. This means that an ever-larger share of the returns to innovation is going to a smaller share of capitalists. If the “labor share” had remained at its 1970 level, compensation today would be $800 billion higher, or $5,000 per person in the US labor force today.
The second, and related, issue is that the capital owners deriving a larger and larger share of our national income are paying less and less tax on that income. The past several decades have seen a large decline in the effective tax rate paid by corporations on their profits in the United States.32 And the recently passed (2017) Trump tax cuts will further continue this trend by significantly reducing the taxation of corporate profits and, at least temporarily, high-income individuals.33
A key element to the success of our proposal is that US citizens directly see a return from their investment into science. US taxpayers are investors in this new initiative, and they should see regular dividends from that investment. So we propose that the returns to this large new public investment accrue more directly to the citizens of the United States through an innovation dividend.
The innovation hub model we propose will lead to exciting new superstar cities around the country. Individuals and firms will want to move to these cities and to live near the new research hub. And as a result, the price of land around these hubs will rise. As discussed earlier, one feature of a successful proposal to be an innovation hub will be zoning rules that allow for affordable development. But even with such rules in place, there will be a rising value both for companies to be located at the research hub and for individuals to live nearby. We propose that the government own some of the land on which the hub is based—and that the rising rents on this land finance the innovation dividend.
Kendall Square near MIT is a perfect example of the value of technology hub real estate. Boston is a dense urban environment with a large variety of locations in which companies could locate. Yet companies will consistently pay rents that are many multiples of nearby areas to be in Kendall Square; Kendall Square recently passed Midtown Manhattan as the most expensive commercial real estate market in the United States.34 This is not surprising given the evidence reviewed earlier about the local nature of research spillovers.
But as we discussed in Chapter 6, this real estate was not always so valuable. The rapid rise in rents in this area have benefited real estate developers such as Joel Marcus, chairman and cofounder of Alexandria Real Estate Equities. From the thirty-five buildings it owns in Cambridge, Alexandria brought in $318 million in rental income in 2017—more than triple what it made one decade earlier.35
In other words, a lot of the benefit from publicly and privately financed research has accrued to the owners of the local land. Our proposal is that in these new research hubs, those benefits are captured and shared with the taxpayer—through public ownership of the land on which the research hubs sit.
Public ownership of land under the research hub can be accomplished in several ways. The first is by relying on existing, and often underused, government-owned real estate. The US federal government is the largest real estate manager in the nation, with more than 3 billion square feet of buildings owned or leased, as well as 34 million acres of land. Moreover, there remains significant unused capacity in federal buildings—only 79 percent of federal buildings are used at 75–100 percent of capacity.36 To pick an example that scores highly on our THIS index, for example, Pittsburgh has 8.2 million square feet of federal property, of which 665,000 is underused or unused, as well as more than 1,300 acres of federal land.
We are not proposing to use undeveloped public lands in the western part of the United States—wilderness areas or national forests would not make sense as tech hubs! Our suggestion is about making better use of real estate that has already been developed and that is indisputably available for commercial purposes—for example, as offices or labs. Many of the potential locations for these research hubs have substantial underused federal real estate.
Of course, existing federal real estate holdings may not be sufficiently concentrated to help create dynamic areas that will attract workers and businesses. Local governments who want to apply for hubs may need to use their own real estate or purchase land in the open market through standard transactions. Land swaps between the federal government and local government or even universities should be considered. A prominent example was the recent land swap between the US Department of Transportation and MIT of a valuable fourteen-acre parcel near Kendall Square.37 MIT paid the federal government $750 million and agreed to create a “vibrant mixed-use site that will benefit MIT’s mission and the Cambridge community,” including a new federal facility.
Once federal funds start flowing into an area, however, the very conditions that make the area attractive will come under pressure from natural market reactions. The owners of the land targeted for research infrastructure and business development will realize that while that land may not be worth much today, it may be quite valuable in the future. Rational investors will therefore incorporate those expectations into prices. Depending on the timing of announcements, there is a risk that the government will end up paying higher prices for any land on which research hubs would be located, and consequently, there will be lower gains to be distributed to taxpayers. In the worst-case scenario, government funds will have served more to enrich local real estate owners rather than all Americans. Indeed, real estate investors bought up land and buildings in cities considered likely winners in the Amazon HQ2 contest, while others were reported to be raising funds so that they could purchase property as soon as the winner was announced.38
For this reason, it is important that communities acquire land before the announcement of a hub investment. This is what Amazon did in 2018, buying the real estate it needed before revealing its HQ2 choices. More broadly, when applying to the competition, areas should be encouraged to demonstrate the political and legal feasibility of providing publicly owned land for research, development, and commercialization.
The government should also explore innovative lease structures that are performance-related. One disadvantage of trying to capture returns through a fixed payment mechanism such as leases, compared to holding equity, is that it is not nearly as flexible and responsive to company performance. A five-year lease to a company that becomes incredibly successful will still pay at the initial lease rate, whereas the value of equity holdings in the company would rise rapidly.
The government could instead offer initially lower lease rates that are linked to company performance; the government then shares the risk—and the returns—of the enterprise. This allows the government to share in some of the upside created by the hubs, while providing some insurance for firms that might not able to cover high rents in the early stages of their ventures.
This is not a particularly new or radical idea—in fact, such a lease structure has long been a feature of retail leases.39 What would be new would be extending this beyond leases that are a function of retail sales to leases that are a function of firm growth—that is, the rental rate could be tied to the measures of success such as sales, employment, profitability, or market value.
Moreover, by tying the government’s return to the success of local companies, it provides a further mechanism to ensure that the government doesn’t engage in cronyism by giving prime spots to favorite companies that are not productive or synergistic with the existing firms in the area. Doing so could lower the earnings in the area broadly, leading to a noticeable reduction in the revenues flowing in to finance the innovation dividend. Over time, the dividend promise provides a natural check on government malfeasance.
The returns on government holdings of land would go into a national endowment fund. The resources in this fund would be immediately distributed to all Americans through an innovation dividend, a flat per capita check each year—with everyone receiving an equal dollar amount.40 This innovation dividend idea draws directly on one of the most successful redistribution programs in the United States, the Alaska Permanent Fund.
In Alaska, leases and royalties for oil exploration and the creation of the massive Trans-Alaska Pipeline System added up to almost $1 billion in the early 1970s—and was just as quickly spent by the state legislature.41 Alaskans voted in 1976 by a margin of 2–1 to amend the constitution to put at least 25 percent of oil revenues into a dedicated fund called the Alaska Permanent Fund.42
The Alaska Permanent Fund Corporation (APFC) is managed by a board of trustees and currently manages about $60 billion in assets. One of the key factors in its success has been an annual dividend that is paid to every man, woman, and child who is resident in Alaska (residents must reapply every year to establish their residency). There is remarkably little fraud, with only 0.03 percent of applications viewed as ineligible. The dividend each year amounts to roughly 10 percent of the net income earned by the APFC during the year.43 Since 1982, the fund has paid out $40,000 per resident of the state. The dividend in 2016 was $1,022; it peaked in 2015 at $2,072.44
This payment provides a way to ensure that all Alaskans benefit equally from oil revenues and has the advantage of lifting state residents out of poverty; one 2016 study found that the dividend annually lifts 15,000–25,000 Alaskans out of poverty.45 Indeed, the dividend may be one reason that Alaska is the most equal state in the nation in terms of income distribution.46
Alaska is one of the most strongly Republican states in the country. Their state congressional delegation is consistently highly conservative and votes against many government spending initiatives, yet this government payment is highly popular. Moreover, this model has spread to one of the most liberal states in the United States: California. California’s Global Warming Solutions Act, also known as Assembly Bill (AB) 32, requires all power plants, natural gas distributers, and other large industries that emit greenhouse gases to pay a fee based on the amount they pollute. The fees are then redistributed to individuals in the state as a “credit” on utility bills. Anyone who is an electricity or natural gas bill customer can receive this “credit,” which is essentially a reduction in the utility bill. The distribution varies slightly by electricity provider, but it is typically a similar amount for every electricity user.47
The success of programs like this in states as disparate as Alaska and California suggests that a similar structure could draw bipartisan support in distributing the returns from the government investment in scientific research.
The enormous rise in public financing of research and development during and after World War II transformed our nation, creating the new products on which our modern economy was built and generating economic opportunities for all. The subsequent falloff in public funding contributed to slower productivity growth and reduced opportunities for most Americans. This should not be surprising given the economics of R&D. The private sector is unlikely through its own profit-seeking behavior to finance enough research to capture the broader social benefits of new ideas, especially those expensive and risky ventures that create industries and jobs.
To jump-start America, we need to return to the model of public-sector leadership in research and development that marked the postwar period. We have provided here the outline of a plan for doing so. Obviously, turning this outline into actual legislation would raise a wide variety of more detailed logistical hurdles, but we view this as a roadmap toward a feasible plan to return the government to its leadership role in promoting technology-led growth in the United States.
The Central Florida Research Park illustrates some of the principles that we have in mind with our proposal—but also some of the limitations. There is a lack of investment capital for firms in the CFRP, resulting in fewer large homegrown employers. And there is a skills deficit that keeps companies from finding the skilled labor they would need to grow. This is why we need to build on and improve models such as the Orlando model—combining innovative R&D infrastructure with sufficient funding and with the supply of skilled employees that are needed to fill the jobs of the future. And we need to do this soon—because other countries are already there.