Key Takeaways
Treat publicly funded science as infrastructure for everyone, not a corporate subsidy
The central bargain. Jonathan Gruber and Simon Johnson argue that government should fund research private firms are unlikely to undertake, then work with universities and businesses to turn discoveries into useful products and jobs. The reason is simple: new knowledge often benefits many firms and the public, while the original investor captures only part of the return. That mismatch leaves valuable research undone. The authors propose a renewed federal commitment, including about $100 billion more per year for research and development, directed toward economic growth and shared prosperity.
A shrinking engine. Federal research spending reached nearly 1.86 percent of GDP in 1964, then fell to about 0.67 percent in 2015. The book links this retreat to weaker productivity growth, stagnant incomes for many families, and a widening geographic concentration of good jobs.
The infrastructure analogy is useful because it shifts the debate from whether government should pick winning companies to whether society should maintain the conditions that make innovation possible. Basic research, skilled workers, and shared technical facilities can benefit firms that never paid for their creation. The hard question is how to preserve broad access without turning public support into permanent protection for favored industries. A sound policy therefore needs open competition, clear public goals, and evidence that support complements private investment rather than merely replacing it.
A wartime research surge helped build America's postwar middle class
A crisis reorganized science. In 1940, engineer and administrator Vannevar Bush persuaded President Roosevelt to create the National Defense Research Committee, connecting government, university scientists, and private industry. The committee mobilized researchers to solve urgent military problems. One example was radar: British scientists supplied a key component, and US researchers at MIT's Radiation Laboratory developed it into systems used to detect aircraft and submarines. Allied U-boat losses fell sharply in 1943 as radar and related tools reached the battlefield.
The peace dividend. Wartime research built expertise and technologies that later supported commercial aviation, telecommunications, computing, and medicine. Federal research spending rose roughly twentyfold from 1940 to 1964. The postwar expansion of college access, including through the GI Bill, helped supply workers for new industries. Median family income roughly doubled from 1947 to 1970, spreading gains beyond a small group of scientists or investors.
The wartime story shows that institutions and coordination matter as much as isolated flashes of genius. Government absorbed early risk, universities assembled expertise, and companies supplied manufacturing capacity. This resembles a mission-oriented innovation system: a defined public problem can align actors who otherwise would not collaborate. Yet wartime urgency is an unusually powerful organizing force, and military research can carry ethical costs. The broader lesson is not that every economic challenge needs a Manhattan Project, but that long-term public goals can coordinate research when commercial incentives alone are too narrow.
Private firms cannot capture enough of the discoveries they help create
Three market gaps. The authors explain why private research and development often falls short. First, spillovers let other firms build on discoveries without paying the original research costs. Second, companies keep failed experiments and safety findings secret, so rivals repeat work or abandon promising paths. Third, long development timelines make investment unattractive when returns arrive after patents lose much of their value.
A costly illustration. Xerox's PARC lab developed ideas behind the graphical computer interface, but Apple and Microsoft captured much of the commercial value. In pharmaceuticals, Merck paused work on lovastatin after hearing an incomplete warning about a rival's related drug. The concern proved unfounded, but the information was proprietary, delaying development. Venture capital adds another constraint: its funds often seek results within a few years, making large manufacturing investments a poor fit.
These examples distinguish invention from value capture. A firm may create knowledge that makes the whole economy more productive, yet rationally choose not to invest if competitors can appropriate the gains. This is a classic case for public support, but not a case for assuming every spillover deserves a subsidy. Public agencies still need to identify additional research that would not otherwise happen, and to avoid paying firms for work they already planned to do. Sharing research results, especially negative findings, can also improve efficiency without requiring government to own every project.
Match technological change with broad access to education and technical training
Skills shape who gains. New technology can raise demand for skilled workers while reducing demand for some routine jobs. The authors call this skill-biased technological change. After World War II, however, the supply of educated workers also expanded. The GI Bill helped millions of veterans attend college, while the National Defense Education Act, passed after Sputnik, strengthened science and math education. More people could fill the technical jobs created by new industries, limiting pressure on the wage gap between skilled and less-skilled workers.
Growth reached families. From 1947 to 1973, real family incomes roughly doubled. Since 1973, median family income had risen only about 20 percent by 2016, even as average output per person grew much faster. The authors argue that research spending should therefore be paired with affordable higher education, vocational training, apprenticeships, and opportunities to retrain workers.
This is a reminder that technology's labor-market effects depend on institutions, not just machines. Education can help workers adapt, while training systems and employer partnerships can connect learning to actual jobs. But schooling alone cannot guarantee that gains are shared: bargaining power, discrimination, regional access, and job quality also shape wages. A policy that expands demand for technical labor without building accessible training risks subsidizing a narrow talent pool. The strongest version of the proposal treats skills as part of the innovation system, not as an afterthought.
Evaluate public research as a portfolio, not by its loudest failure
Breakthroughs are uneven. Public research can create large economic and social returns, but no agency can guarantee that every project succeeds. The Human Genome Project, a publicly funded effort to map human DNA, cost about $3 billion. A study cited by the authors estimated that genomics-related activity generated nearly $965 billion in economic impact from 1988 to 2012, with 280,000 jobs associated with it in 2012. NIH research has also supported advances in vaccines, heart disease prevention, and treatments.
Failure is part of discovery. The Solyndra solar company received a $535 million federal loan guarantee and went bankrupt. The authors note that it represented only a small share of a much larger clean-energy initiative, whose overall loan defaults were low. They argue for transparent evaluation across many projects, rather than treating one failure as proof that public investment cannot work.
Portfolio thinking is a better standard for uncertain research than demanding a perfect record. Venture investors expect many failures because a few successes can repay a portfolio. Public programs deserve the same recognition of uncertainty, alongside stricter accountability for conflicts of interest and poor execution. Economic-impact estimates, including those for genomics, also depend on assumptions about what would have happened without public funding. Good evaluation should track scientific outputs, private investment, jobs, health benefits, and distribution, while using credible comparisons to test whether public spending caused the gains.
Build innovation hubs where talent can grow, not only where it already clusters
Clusters compound advantages. In a knowledge economy, firms and skilled workers benefit from being near universities, investors, suppliers, and one another. This concentration, called agglomeration, can make established technology centers increasingly productive. But high housing costs and restrictive zoning in successful cities keep many workers out. Venture capital is concentrated too: about two-thirds of US funding went to five major metropolitan areas, according to the book.
Opportunity is wider than the coasts. The authors' Technology Hub Index System, or THIS, identified 102 possible hubs across 36 states, based on population, education and research strength, patents, housing costs, crime, and commuting. These places included more than 80 million residents. The index is a starting point, not a guarantee that any selected city will succeed. Public investment could help promising places assemble the research and business networks that private investors tend to avoid until a cluster already exists.
The case for geographic diversification is economic as well as political. Local research spillovers mean that placing a lab in a new region can create benefits that do not travel freely across the country. At the same time, dispersing funding too thinly may prevent any site from reaching the density needed for collaboration. Housing policy is central: a research hub that creates jobs but blocks workers from living nearby can reproduce the exclusion it was meant to address. Any location strategy should therefore pair research investment with housing supply, transport, and workforce development.
Replace corporate tax-bid wars with a national competition for science hubs
Compete on capacity. When Amazon sought a second headquarters, 238 cities and regions bid, often offering tax breaks and other inducements. The company chose Northern Virginia and New York City, two already successful regions. Such deals may shift jobs from one American location to another while reducing public revenue. The authors cite more than $45 billion in annual state and local business incentives as evidence of the scale of this competition.
Make the contest serve the country. Their alternative is a federal competition in which places demonstrate research capacity, university and business partnerships, skilled-worker plans, infrastructure, affordable housing, and zoning that allows sustainable growth. An independent Innovation Commission, modeled on the Base Realignment and Closure process, would evaluate applications transparently and recommend selected sites. Funding should be substantial enough to help new hubs achieve the scale that attracts firms and workers.
A competitive process can harness local initiative while directing resources toward national goals, rather than rewarding whichever government offers the largest subsidy. Its credibility depends on public criteria, independent review, and meaningful local commitments. However, measurable criteria can be gamed, and current research strength may favor familiar institutions over places with overlooked potential. A strong process would combine baseline standards with room for different regional strategies, then reassess results over time. Competition is not automatically fair; its rules determine whether it reveals capacity or simply reproduces existing advantage.
Give taxpayers a visible share of the value public research creates
Capture some of the upside. Public research often creates valuable companies and raises nearby land prices, while the public that funded the research receives only indirect returns. The authors propose that governments acquire land for research hubs before announcing them, then lease it to firms and use rising rents or land values to fund a national endowment. They also suggest performance-linked leases, so public returns can grow when local companies succeed.
Make the return tangible. The endowment would pay an equal annual innovation dividend to Americans, borrowing from Alaska's Permanent Fund, which distributes part of its investment income to residents. To protect the program from political bargaining, the authors call for independent oversight and multiyear funding. A visible dividend could help citizens see public research as an investment rather than an expense whose gains accrue only to private owners.
The dividend addresses a real legitimacy problem: when public risk produces private fortunes, people may reasonably question who benefits. Land value capture offers a practical route to sharing gains without taking ownership of every company. Yet it depends on careful land acquisition, transparent leases, and enough revenue to make dividends meaningful. Equal cash payments also cannot substitute for public services or correct unequal access to good jobs. The proposal is most persuasive as one layer of a broader social contract: public investment, accountable stewardship, and a clear share of returns for the people who finance it.
Lead future industries by building the path from discovery to responsible deployment
The frontier is industrial. The authors point to synthetic biology, lower-carbon energy, and deep-sea exploration as areas where research may create new products and jobs. They also show how countries use research parks to connect laboratories, universities, manufacturers, and investors. Taiwan's Hsinchu Science Park, for example, became a major semiconductor center by linking research with production. The US challenge is not only to invent, but to manufacture and commercialize at home.
Leadership sets the rules. New technologies raise environmental, safety, and ethical questions. Deep-sea mining might supply minerals for high-tech products but could damage poorly understood ecosystems. Gene editing can produce medical benefits while raising risks of unintended effects. The book argues that active participation in research gives the United States a stronger role in setting standards, while emphasizing that innovation must be governed openly and responsibly.
The book's strongest closing insight is that technological leadership includes governance, not just patents or market share. Countries that build expertise can shape standards and safety practices, but speed alone is not a public good. Research hubs need independent scrutiny, public engagement, and rules that keep environmental and health costs from being displaced onto communities with little power. The industrial case also deserves realism: a research lead does not guarantee domestic manufacturing or durable employment. Policy must connect discovery to production while remaining willing to stop projects whose risks exceed their benefits.
Analysis
The book's argument is most compelling when understood as a design problem, not a plea for government spending in the abstract. Research generates knowledge that is difficult for any single firm to own, while commercial development requires capital, facilities, and patient investment. Government can address those mismatches, but the American record also shows that research institutions are political institutions: public trust, geographic distribution, and accountability shape whether funding survives long enough to matter.
The historical account usefully complicates the lone-inventor story. Radar, computing, biotechnology, and other fields grew through networks of public agencies, universities, and firms. The broader lesson is that innovation depends on an ecosystem, including technical education and production capacity. The postwar experience suggests that new technology need not automatically widen inequality when education expands alongside demand for skills. It does not prove that research spending alone can reproduce that period's wage growth, since trade, labor institutions, housing, and social policy also influenced who benefited.
The proposal's central tension is between excellence and inclusion. Research funding tends to reward established capability, yet concentrating it in existing hubs can leave other regions without the infrastructure needed to compete. A place-based competition could help break this cycle, but indices and application criteria risk privileging what can be measured over what is emerging. The authors' emphasis on independent review and local commitments is therefore essential, though it cannot eliminate political pressure or uncertainty.
The innovation dividend adds a distinctive legitimacy mechanism by linking public risk to public return. Its success would depend on durable governance, credible land valuation, and revenues large enough to matter. More fundamentally, cash returns should not obscure broader public value, including health, environmental quality, and access to opportunity. The strongest reading of the book is a call to build institutions that fund uncertain discovery, diffuse its capabilities, and openly negotiate its costs. That is more demanding than simply increasing a budget, but also more likely to sustain public support.
Glossary
Spillovers
Benefits beyond the investorSpillovers are benefits from research or innovation that reach people or firms beyond the organization that paid for the work. Other companies may use the resulting knowledge to create products, and society may benefit through improved health or productivity. Because the original investor cannot capture all these gains, private firms may invest less in research than is beneficial overall.
Valley of death
The risky commercialization gapThe valley of death is the costly, uncertain stretch between a promising research result and a product that can be manufactured and sold. The work may require substantial investment before commercial success is clear. Conventional venture capital often avoids this gap when projects are capital-intensive or take too long to demonstrate a market, leaving technologies stranded between the laboratory and production.
Agglomeration
Advantages of geographic concentrationAgglomeration describes the economic advantages created when related firms, skilled workers, universities, and investors locate near one another. Proximity can improve hiring, access to shared resources, and the exchange of knowledge. These advantages can help a cluster grow, but they can also concentrate jobs and investment in existing hubs and contribute to high housing costs.
Technology Hub Index System
A screen for potential hubsThe Technology Hub Index System, abbreviated THIS, is the authors' method for comparing US metropolitan areas as possible technology hubs. It combines measures of population, education and innovation, and quality of life, including housing costs, crime, and commuting. The authors use it to identify 102 candidate hubs, while stressing that the ranking is one possible starting point, not a definitive verdict.
Innovation dividend
A public return from innovationThe innovation dividend is the authors' proposal to distribute some returns from publicly supported research directly to Americans as equal cash payments. Publicly owned land and related lease income around new research hubs would help finance an endowment, which would pay the dividends. The proposal draws on Alaska's Permanent Fund and aims to make the public's role as an investor visible.
Small Business Innovation Research program
Federal grants for small firmsThe Small Business Innovation Research program, or SBIR, is a federal program that awards competitive research grants to small businesses. Its staged grants help firms test ideas and develop prototypes, potentially making it easier to attract private investment. The book presents SBIR as evidence that public funding can support early innovation while complementing, rather than replacing, private capital.
Human Genome Project
Public effort to map human DNAThe Human Genome Project was an international, publicly supported research effort to determine the sequence of human DNA. Launched in 1990, it combined government agencies, universities, and private firms, and its work helped improve sequencing technology. The resulting knowledge supported research across medicine, biotechnology, agriculture, and other fields, illustrating how a public research investment can create broad downstream uses.
Download PDF
Download EPUB
.epub digital book format is ideal for reading ebooks on phones, tablets, and e-readers.