Archives: Publications

These are AIESG Publications

Rebuilding the Post-Pandemic Economy

After suffering the worst economic shock since the Great Depression, the American economy is recovering in fits and starts. While many businesses are reopening their doors and thriving, continued uncertainty about the course of the virus, the inflation outlook, labor shortages, and many other factors are hampering a full return to normal activity. The COVID-19 pandemic reinforced and exacerbated many of the biggest structural economic challenges in our society. It precipitated the largest economic relief and stimulus spending in US history and transformed the way that millions of Americans live and work, with automation, e-commerce, and telework all playing a bigger role.

The policy volume Rebuilding the Post Pandemic Economy examines important questions about how the post-pandemic economy will take shape. What are some initial lessons we can take away from the novel government programs that were deployed to provide economic relief and stimulus? How can we implement new infrastructure investments to maximize efficiency and equity, and best respond to the climate crisis? After a year of widespread school closures, what have we learned about the role of K-12 education in perpetuating or reducing social and economic inequities? And how should American trade policies evolve to promote economic recovery and strengthen America’s role in the global economy?

DOWNLOAD BOOK

Brookings Institution: Six reasons why an expanded Child Tax Credit or child allowance should be part of the US safety net

The Child Tax Credit has been part of the federal income tax code since 1997. It has been expanded many times, most recently as part of the American Rescue Plan. Under this plan, for the year 2021, the maximum Child Tax Credit amount is increased from $2,000 per child to $3,600 for children below the age of 6 and to $3,000 for children under age 18.  This credit amount is phased out at high levels of income in two steps. The increased credit is phased down to the previous credit amount starting at income of $112,500 for single parents and $150,000 for married parents; that reduced amount is then phased out completely, beginning at income of $200,000 for single parents and $400,000 for married parents. In addition to the increased credit amount and expanded range of qualifying income up the income distribution, for 2021, the credit is fully refundable. This means that parents receive the full credit amount, regardless of the amount of taxes they owe. This has the effect of extending the credit to families with no taxable income. These changes result in a child benefit, administered through the tax code, that is nearly universal and unconditional on parental work status.

House Democrats have proposed extending this expanded Child Tax Credit through 2025. Other policy makers have expressed reservations about this proposed extension, citing concerns about having a credit payment that is unconditional on work and on extending a payment to high income families. A consideration of the pros and cons of a nearly universal, unconditional child tax credit or benefit amount raises a host of issues. It is useful to consider these issues within the framework of the optimal design of a social safety net system. Within this broad framework, there are six specific points about the desirability of such a payment.

1. We should have social insurance against child poverty.

In this country, we provide social insurance to people who find themselves in what economists call “the bad state of the world.” We have disability insurance for people who find themselves unable to work and earn a substantial amount of money in the labor market. We have unemployment insurance for people who lose their jobs. We have social security old age payments for people who live beyond their working years and are either no longer able to work or have not saved enough. This is social insurance against elderly poverty.

However, we have no social insurance for kids who—through no fault of their own—find themselves living in a home with material deprivation. I think if most of us were going to design a system of social insurance from behind a Rawlsian veil of ignorance—meaning, we were going to design a world that we’d like to be born into before knowing our place in that world—providing social insurance against child poverty would be the first type of social insurance we’d provide.

From the perspective of social insurance and a reasonable social welfare function, there is a strong case for a basic guarantee of income for kids and in particular social insurance against child poverty. (This is something I wrote about in an article for Brookings last October.)

2. From the perspective of government spending as an investment, this would yield a positive social return.

Here I am implicitly equating a tax credit with a spending program. Conceptually they are the same. In practical administrative terms, they are not the same thing. There are practical challenges with running social policy through the tax code, but for the moment, I want to set those considerations aside and focus on the social return to government “spending”—whether that “spending” comes in the form of foregone tax revenue, tax credits, or an explicit spending program.

Spending on children, especially children from low-income homes, has a large social return. We have a lot of evidence showing that increasing the income and material resources of low-income families with children leads to better school performance, better child and maternal health outcomes, and better long run outcomes for children. We are vastly underinvesting in our nation’s children currently and a child allowance or expanded child tax credit would go some way toward rectifying that. (This is something Diane Schanzenbach and Hilary Hoynes and I highlighted in a piece we wrote for the Conversation earlier this year.)

The investment aspect for the Child Tax Credit is harder to make the further up the income distribution you go. The current expansion of the Child Tax Credit up the income distribution means many high-income families are receiving this additional income. We have very little evidence that supplementing the income of higher-income families has a positive social return, so you’d need to lean on other arguments to justify expansions up the income distribution.

I think there is a very compelling social investment case to be made for the anti-poverty aspects of the Child Tax Credit coming from the expansions in credit amounts to lower-income families. That case falls apart the farther up the income distribution we go.

3. Providing income assistance to children regardless of their parents’ work status would fill a hole in our current safety net.

To my mind, the key benefit of introducing an unconditional child allowance or child tax credit is because there is a gaping hole in our safety net. We don’t have any meaningful source of income support for kids whose parents don’t work and who don’t qualify for categorical income assistance, say, through a medical eligibility that qualifies them for Supplemental Security Income (SSI). 

An unconditional child allowance or refundable child tax credit differs from a policy that conditions benefit or credit receipt on work, like the Earned Income Tax Credit (EITC) does. The EITC does two things—it supplements the wages of low-income workers and it also transfers money to families with children. These are two laudable goals, but they don’t have to be accomplished with the same program.

I like the idea of divorcing the supplementation of wages of low-income workers and transferring money to children. These policy goals can be separated so that kids whose parents have very low or no earnings are not left without government assistance. An unconditional child allowance or tax credit rightly fills in this gap in our safety net.

4. A nearly universal child tax credit or allowance limits work disincentives.

On the one hand, as I’ve mentioned, there is little investment rationale for extending this type of benefit payment up the income distribution. On the other hand, a near universal design limits work disincentives. The usual work disincentives of transfer programs come from both substitution effects, which come from the fact that benefits are clawed back as people’s earnings income, which provides a disincentive to work, and income effects, by which people might choose to work less because they have more money.

A nearly universal child tax credit means that the credit is not clawed back as people earn more money (except at very high levels of earnings), so there is only an income effect. Those who worry about the work disincentives of transfer programs should find this design appealing. On net, we should expect to see less work reduction from this design than we would from a more typical transfer program design, like in the old Aid to Families with Dependent Children (AFDC) program.

5. There is a case to be made for additional cash benefits to supplement existing in-kind program benefits.

There is a strong case to be made for providing more cash assistance to low-income families with children, supplementing our existing panoply of programs that provide health insurance, food and nutrition assistance, and limited housing assistance. (I would argue we need more housing assistance as well.) This would help families meet the different needs that they face, which arrive at unpredictable times, and are hard to meet with the various siloed programs that currently exist.

The criticism of this approach is that some worry that some parents will not spend money in ways that benefit kids. But we have good evidence from a variety of income shocks—things like EITC expansions, for example—that low-income children generally benefit from the additional income coming into the family. So even if there are some parents who might not spend the money in the ways that would most benefit their children all the time, in general, we can expect that the money will be spent in ways that improve children’s outcomes.

6. This is just not that expensive relative to the likely social benefits.

The final point to make is about the fiscal costs of an expanded child tax credit. This is just not that expensive, relative to the benefits we will get from that government spending. Estimates suggest that the Child Tax Credit costs about $118 billion a year; the temporary Child Tax Credit expansions under the America Recovery Act are projected to cost about $105 billion a year.

Another option would be a child benefit or allowance (as opposed to a tax credit), something like what I proposed in a simple thought experiment I wrote up in a Brookings post last October. I noted that if we gave each child living in poverty the average Social Security benefit received by a Social Security recipient age 65 and over (about $17,000 per year), the rate of childhood poverty in this country would fall to less than one percent. If we gave each child living in poverty half the average Social Security benefit ($8,556 annually), the rate of childhood poverty in this country would fall to about 3 percent. The cost of this dramatic reduction in child poverty would cost $179 billion for the full benefit award and $90 billion a year for the half benefit award.

To put that money in perspective, let’s compare this to a Universal Basic Income (UBI), a policy idea that has gotten a lot of attention over the past couple of years. A UBI would guarantee a meaningful level of income to every adult in the United States. A payment of $10,000 per year to every US adult would cost $2.5 trillion.  Why are we even having that conversation when we could essentially eradicate child poverty in this country for $180 billion? Even if we reduced the UBI award to be $10,000 a year to adults who make less than $20,000 and then phased out at 30 percent, it would still cost $1.5 trillion per year. A targeted child allowance would cost a fraction of that, and constitute an investment in the next generation.

To conclude, for these main six reasons, an unconditional child tax credit or a child allowance fits into a well-designed social safety net in our country.

This article was originally published by the Brookings Institution.

AESG Member Statement: A Call for US Leadership on Global Vaccination Efforts

The United States government should take up a position of world leadership on ending the global COVID-19 pandemic through vaccine outreach to the world. Such an effort would serve a clear humanitarian purpose. It would represent forward defense of our security interests by slowing the virus’s rate of mutation. No other action would so clearly signify a US commitment to enlightened international leadership at a time when our strength and outward looking vision is increasingly being challenged and no other initiative would do more to stabilize the global economy.

Wise policy making requires balancing rigorous attention to detail with bold vision. If national commitment awaited detailed technocratic plans, the U.S. would never have launched the Marshall Plan, sent a man to the moon, or launched the PEPFAR policy that has done so much to combat AIDS. With the Delta variant of COVID becoming pervasive and more mutations likely, now is the time for bold action. The necessary global approach will involve financial commitments, as well as active policies to encourage production capacity at home and abroad, to strengthen local health system capacity in partnership with developing countries, and to promote private sector initiatives.  We must reject the false dichotomy between domestic and global efforts. Given mutation risk, safety anywhere depends on safety everywhere.

The International Monetary Fund estimates that roughly 60 percent of the world could be vaccinated by mid-2022 for $50 billion. This is less than one percent of what the US has already committed in response to COVID. Other nations would join a US initiative. Because US spending could be leveraged by the international financial institutions and the private sector, each dollar of US contribution could be leveraged 5 or 10 to 1. In all likelihood, US investment in global vaccination would pay for itself several times over by strengthening the US and global economy, reducing further virus spread, and building international good will towards the United States.

 

SIGNATORIES*

Mary Barra
General Motors

Ben Bernanke
Brookings Institution

Joshua Bolten
Business Roundtable; Former White House Chief of Staff

Erskine Bowles
Aspen Economic Strategy Group; Former White House Chief of Staff

Chad Bown
Peterson Institute for International Economics

Ken Chenault
General Catalyst

Susan Collins
University of Michigan

David M. Cote
Vertiv Holdings

James S. Crown
Henry Crown & Company

Roger W. Ferguson, Jr.
TIAA (Emeritus)

Michael Froman
Mastercard; Former US Trade Representative

Jason Furman
Harvard University

Timothy Geithner
Warburg Pincus; Former US Treasury Secretary

Austan Goolsbee
University of Chicago

Glenn Hubbard
Columbia University

Doug Holtz-Eakin
American Action Forum

Neel Kashkari
Federal Reserve Bank of Minneapolis

Melissa S. Kearney
The University of Maryland; Aspen Economic Strategy Group

Jacob Lew
Lindsay Goldberg; Former US Treasury Secretary

Maya MacGuineas
Committee For a Responsible Federal Budget

David McCormick
Bridgewater 

Marc Morial
National Urban League

Janet Murguia
Unidos US

Michael A. Nutter
Former Mayor of Philadelphia

Jim Owens
Caterpillar (Emeritus)

Henry M. Paulson, Jr.
The Paulson Institute; Aspen Economic Strategy Group; Former US Treasury Secretary

John Podesta
Center for American Progress; Former White House Chief of Staff

Ruth Porat
Alphabet and Google

James Poterba
Massachusetts Institute of Technology

Penny Pritzker
PSP Partners; Former US Secretary of Commerce

Robert Rubin
Centerview Partners; Former US Treasury Secretary

Paul Ryan
American Idea Foundation; Former US Speaker of the House

Matthew Slaughter
Tuck School of Business at Dartmouth

Margaret Spellings
Texas 2036

Michael R. Strain
American Enterprise Institute

Lawrence H. Summers
Harvard University; Former US Treasury Secretary

Mark Weinberger
EY (Emeritus)

Robert Zoellick
Brunswick Group; Former President of the World Bank

 

*Affiliations are listed for identification purposes only and do not necessarily reflect the position of members’ institutions. 

Internet Access and its Implications for Productivity, Inequality, and Resilience

The past year has brought an unprecedented change in the way Americans work, with millions of workers working from home and connecting to colleagues and clients virtually. In an AESG report titled “Internet Access and its Implications for Productivity, Inequality, and Resilience,” economists Jose Maria Barrero (Instituto Tecnológico Autónomo de México), Nicholas Bloom (Stanford University), and Steven J. Davis (University of Chicago) examine the role that home internet quality plays in driving worker productivity under this new work paradigm.

The authors draw timely insights and lessons from the Survey of Working Arrangements and Attitudes (SWAA), an original cross-sectional survey they have collected since May 2020 on over 40,000 working age Americans. They estimate that universal internet access – defined as high quality, fully reliable home internet service for all Americans – would raise earnings-weighted productivity in the post-pandemic economy by 1.1%, which implies flow GDP gains of $160 billion per year, or a present value gain of $4 trillion at a 4% discount rate. They estimate that universal access would raise the extent of work from home (WFH) in the post-pandemic economy by about seven-tenths of a percentage point.

Universal internet access promotes economic and social resilience by facilitating commerce and socialization at a distance. Internet technologies enabled large sectors of the economy to function well during the pandemic. In addition, existing evidence suggests that home internet access may help to mitigate the negative health effects of loneliness and social isolation in a time of pervasive social distancing.

THE SURVEY OF WORKING ARRANGEMENTS AND ATTITUDES
Barrero, Bloom, and Davis have fielded the Survey of Working Arrangements and Attitudes since May 2020, collecting 2,500 to 5,000 responses per month. Survey respondents report higher productivity when working from home during the pandemic as compared to when working on employer premises before the pandemic.

In previous research, the authors combined this survey data with information about employer plans regarding post-pandemic work arrangements to predict what a re-optimization of work arrangements post-pandemic would look like. They estimated that one-fifth of paid workdays will be supplied from home in the post-pandemic economy and more than a quarter of workdays on an earnings-weighted basis. They also estimate that re-optimization could be expected to boost productivity by close to 5%, largely through saved commuting time.

PROJECTING THE PRODUCTIVITY EFFECTS OF UNIVERSAL ACCESS
The authors combine individual-level data on the planned extent of WFH in the post-pandemic economy with individual-level estimates for the productivity impact of universal access to project the effects of a hypothetical move from the current state of internet access to one with high quality, fully reliable internet access in all households. The authors find large productivity consequences as a result of imperfect internet: a 3% aggregate labor productivity shortfall during the pandemic and a 1.1% gain from universal access in the post-pandemic economy. To the extent that better internet access improves WFH efficiency, the authors conclude that the overall estimated impact on the extent of WFH—an increase of 0.7 percentage points—is quite modest both during and after COVID. Importantly, the impact also varies little across demographic groups.

Barrero, Bloom, and Davis also derive implications for aggregate output, finding that the flow output loss during the pandemic is nearly three times as large as the projected flow benefits from universal access in the post-COVID economy. They note that this comparison underscores the economic resilience value of universal access: the output payoff is much larger in pandemic-like disaster states when output is unusually low and the marginal value of output is unusually high.

Barrero, Bloom, and Davis additionally analyze the earnings gains to workers from the associated productivity enhancements and estimate that such gains would be nearly uniform across income and demographic groups, meaning that productivity improvements would not come at the expense of widening inequality.

The authors recognize that there is uncertainty around their estimates of the impact of universal access on productivity and output. For instance, insofar as pandemic-related stressors (kids at home, lack of technological familiarity) pulled down WFH productivity during the pandemic, their estimates might understate the positive effect of internet access on WFH productivity after the pandemic. In addition, they lack survey data on the relative efficiency of WFH for respondents with no WFH experience during the survey period, accounting for 43.3% of respondents on an equal-weighted basis and an estimated 34.2% on an earnings-weighted basis. The lack of data for these respondents may lead them to overstate the effects of universal access.

INTERNET ACCESS AND SUBJECTIVE WELL-BEING DURING THE PANDEMIC
The authors examine the social effects of fully reliable internet access, particularly during crises such as the COVID-19 pandemic. They cite evidence that suggests that social distancing during the pandemic and pandemic-related stresses had negative health effects for many Americans. Such evidence motivates the hypothesis that better internet access during the pandemic alleviated the harmful psychological and other health effects of social distancing and pandemic- related stresses.

Using responses from the SWAA, the authors conclude that universal access would materially improve well-being during pandemic-like disasters for persons who currently lack home internet service while smaller improvements in well-being would accrue to persons who currently have subpar access.

UNIVERSAL ACCESS AS A SOURCE OF RESILIENCE
Barrero, Bloom, and Davis emphasize that by raising output in the face of infectious disease outbreaks, biological attacks, and other disaster states that involve physical distancing, universal access to high-quality home internet service would strengthen U.S. economic resilience. For society as a whole and for individual firms and workers, the capacity to quickly switch between production modes of roughly equal productivity is a valuable option that pays off especially in bad states of the world. The authors’ analysis suggests that the output payoff to universal access during pandemic-like disasters is nearly three times as large as the payoff during normal periods and that universal access promotes resilience by providing a ready means of engagement and socializing when circumstances compel physical distancing.

The authors also cite other important benefits of universal access that they do not quantify, such as the ability of households to turn to online shopping and home delivery services during a pandemic-like disaster, increased compliance with stay-at-home orders, and ameliorated gaps in remote learning opportunities. The authors underscore the importance of such enhanced economic and social resilience in the face of recurring outbreaks of COVID-19 and other infectious diseases.

Suggested Citation: Barrero, Jose Maria, Nicholas Bloom, and Steven J. Davis. December 10, 2021. “Internet Access and Its Implications for Productivity, Inequality, and Resilience”. In Rebuilding the Post-Pandemic Economy, edited by Melissa S. Kearney and Amy Ganz. Washington, DC: Aspen Institute, 2021. https://doi.org/10.5281/zenodo.14057577.

Chapter

Executive Summary

Aspen Economic Strategy Group Releases New Policy Analyses Examining the US Infrastructure Agenda

Three papers examine how infrastructure investments can promote economic growth and broader prosperity.

Washington, DC, July 14, 2021 – The Aspen Economic Strategy Group (AESG) today released a set of three papers on infrastructure and technological innovation and their impact on the post-pandemic economic recovery.  These papers will be included in the group’s annual policy volume, “Rebuilding the Post-Pandemic Economy,” which will be released in December 2021, but it was critical to release these papers now so they can help inform debate as the US infrastructure agenda evolves.  The authors of these papers, all renowned experts on the topics covered, provide important economic insights relevant to ongoing private and public sector discussions about the size and scope of infrastructure spending. 

The AESG, a diverse, bi-partisan group of distinguished economic leaders and thinkers, will convene in late July to discuss important questions about why and how to invest in American infrastructure.  How much infrastructure investment do we need and what type of infrastructure should be prioritized (roads and bridges vs. digital infrastructure)?  How do we enhance the efficiency of the current infrastructure stock? How should we think about the equity dimensions of new infrastructure investments? 

As the Senate reconvenes this week and resumes work on a bipartisan infrastructure plan, the papers provide economic analyses on which types of new infrastructure investments can be a springboard for innovation, energy efficiency, sustained competitiveness, and broader prosperity.

 

Economic Perspectives on Infrastructure Investment
Edward Glaeser, Fred and Eleanor Glimp Professor of Economics in the Faculty of Arts and Sciences at Harvard University
James Poterba, Mitsui Professor of Economics at MIT and the President of the National Bureau of Economic Research

As the country embarks on a major infrastructure initiative, the authors highlight relevant policy lessons from economic studies of infrastructure projects.  Key themes include: 

  • Projects vary widely in their benefits and costs, even within categories such as roads and bridges. Overall, interstate highways today are smoother, and fewer bridges are structurally deficient, than several decades ago. Careful cost-benefit analysis, perhaps carried out by a nonpartisan federal agency created for the purpose, can help identify which projects should be undertaken.
  • While many potential infrastructure projects have substantial benefits, infrastructure costs in the United States are very high by international standards.  Controlling costs by improving procurement practices and project management can raise the benefits per dollar of infrastructure spending. 
  • Maintaining existing infrastructure, rather than building new projects, is one of the most cost-effective ways to deploy new infrastructure dollars.

 

Challenges of a Clean Energy Transition and Implications for Energy Infrastructure Policy
Severin Borenstein, E.T. Grether Professor of Business Administration and Public Policy at the Berkeley Haas School of Business
Ryan Kellogg, Professor and Deputy Dean for Academic Programs at the University of Chicago Harris School of Public Policy

A central challenge in the transition to clean energy is the need to simultaneously control costs and ensure a reliable energy supply.  The authors present key issues the country faces in transitioning to a low-carbon energy system, and the infrastructure that will be needed to support this transition:  

  • Broad incentives – such as carbon pricing, clean energy standards, or clean energy subsidies – that do not discriminate across zero-emissions resources will be essential for directing capital towards cost-effective investments in clean energy infrastructure.
  • The climate challenge is a global challenge, and an essential way to encourage other nations to reduce their own emissions will be to invest in development of zero-emissions technologies and then export those technologies around the globe.
  • Reforms to wholesale and retail electricity markets, and to the regulatory process for long-distance transmission investment, are needed to enable a low cost energy transition that distributes its costs and benefits in an equitable way.
  • The clean energy transition must provide for decommissioning of legacy fossil fuel infrastructure, which also offers an opportunity to employ displaced oil and gas workers.

 

Science and Innovation: The Under-Fueled Engine of Prosperity
Benjamin Jones, Gordon and Llura Gund Family Professor of Entrepreneurship, and Professor of Strategy at the Kellogg School of Management at Northwestern University

The paper explores the central role of science and innovation in the national interest and what role the government needs to have in enacting policies that strengthen public investment.  Key findings include:  

  • The United States massively underinvests in science and innovation, with implications for our future standards of living, health, national competitiveness, and capacity to respond to crises.
  • Public R&D investment is near its lowest level in the last 60 years and doubling current funding would more than pay for itself. 
  • The social returns to R&D investment are huge: for every $1 that is invested, society reaps an average return of $5.

 

“As policymakers continue to shape the US infrastructure agenda, these papers provide valuable insights and recommendations about how to promote efficient and equitable investments,” said Melissa S. Kearney, AESG Director and the Neil Moskowitz Professor of Economics at University of Maryland. “Smart investments in infrastructure will advance the goals of a more robust, inclusive, and resilient nation.”  

The Aspen Economic Strategy Group (AESG), a program of the Aspen Institute, is composed of a diverse, bipartisan group of distinguished leaders and thinkers with the goal of promoting evidence-based solutions to significant US economic challenges. Co-chaired by Henry M. Paulson, Jr. and Erskine Bowles, the AESG fosters the exchange of economic policy ideas and seeks to clarify the lines of debate on emerging economic issues while promoting bipartisan relationship-building among current and future generations of policy leaders in Washington. More information can be found at https://economicstrategygroup.org/.

The Aspen Institute is a global nonprofit organization committed to realizing a free, just, and equitable society. Founded in 1949, the Institute drives change through dialogue, leadership, and action to help solve the most important challenges facing the United States and the world. Headquartered in Washington, DC, the Institute has a campus in Aspen, Colorado, and an international network of partners. For more information, visit www.aspeninstitute.org.

 

Contact:
Suzanne Pinto
Suzanne.Pinto@argos-communications.com

Challenges of a Clean Energy Transition and Implications for Energy Infrastructure Policy

Economists Severin Borenstein (Berkeley Haas School of Business) and Ryan Kellogg (University of Chicago Harris School of Public Policy) discuss the major barriers that need to be overcome in order to successfully execute a transition to a reliable low-carbon energy system at reasonable cost. The authors observe that the country must rapidly reduce greenhouse gas (GHG) emissions while maintaining energy affordability, reliability, and resilience, and address the historical socioeconomic and racial disparities in both access to energy consumption and harm from energy production.

Their report offers a number of concrete suggestions for an energy infrastructure policy agenda, in addition to broad clean energy production incentives:

  • Invest in research, development, and early-stage deployment of novel technologies.
  • Improve the design and price transparency of wholesale power markets.
  • Enhance federal authority over long-distance transmission siting.
  • Reform retail electricity rates to more accurately reflect society’s full marginal cost.
  • Address local pollution and involve local communities.
  • Ensure funding for infrastructure decommissioning.

THE CORE CHALLENGE
Unlike fossil fuels, which provide cheap, reliable sources of energy that can be stored and transported, wind and solar generation are intermittent, costly to store, and are not dispatchable. Providing reliable, zero-carbon emission supply will require combining wind and solar resources with investments in dispatchable zero-emission sources (such as nuclear or hydroelectric sources), long-distance transmission, demand flexibility, and storage technologies. Doing so will require large, policy-driven investments in new energy infrastructure for generation, transmission, storage, and distribution of zero-emission power. A central policy challenge is how to catalyze the necessary investments without unduly increasing costs to end-users.

LOW CARBON TECHNOLOGIES
Reducing costs of nuclear power and developing and scaling carbon capture and sequestration technology would make significant strides toward achieving a zero-carbon economy. In the near-term, the primary pathway for providing clean energy is electrification. Thus, over at least the next decade, widespread use of intermittent renewables—wind and solar power—will be necessary. However, wind and solar alone cannot support a stable electricity system because their production pattern is determined by nature.

The authors argue that a combination of four different approaches could maintain grid stability in a system with very high levels of intermittent generation: long-distance power transmission, storage, demand flexibility, and dispatchable generation. However, they acknowledge that each of these four approaches has its own challenges:

  • Transmission: Electricity transmission is critical for taking electricity from its point of generation from sun and wind and moving it to population centers. Transmission also connects energy markets, which enhances competition among electricity generators. More market options mean less need to curtail renewable production if the output of a given facility exceeds the needs of nearby customers, which then implies higher capacity utilization and lower cost per kilowatt-hour. The cost of transmission is substantial, but according to the authors, the larger barrier is the complex web of authorities that must sign off for an investor to build multi-state transmission lines. This problem could be overcome by allowing the Federal Energy Regulatory Commission to have authority over rights-of-way and use of eminent domain, as it does for natural gas transmission.
  • Storage: Large-scale battery storage can partially substitute for long-distance transmission connections by filling in when intermittent generation fades or demand surges. The cost of battery storage has dramatically decreased over the last decade and it likely has a significant role to play in balancing short-term fluctuations in supply and demand. Current battery technologies, however, are not cost competitive for long- duration storage and not likely to become so in the next decade.
  • Demand Flexibility: Electricity demand has become more volatile over time due to climate change, amplifying the challenge of meeting demand with intermittent generation sources. Dynamic demand adjustment is one tool for managing fluctuations in electricity consumption. For example, electric meters can measure consumption and communicate data directly and frequently to the utility. Such technologies can also be used with heating and cooling, refrigeration, and vehicles. Dynamic electricity rates can also provide incentives for customers to change consumption patterns as supply availability changes.
  • Dispatchable generation: The ability to store and transport energy to match supply to changes in demand is also a challenge. Current battery storage technology is not well suited for long-duration storage. Nuclear and hydroelectric energy generation can provide low-carbon, long-duration grid balancing in the near-term as wind and solar ramp up. Longer term, improvements in nascent technologies such as geothermal energy production or carbon capture and sequestration would be valuable for enabling greater amounts of dispatchable, low-cost clean power generation.

One specific area where policy can also promote and accelerate low-carbon technology is supporting the development of widespread electric vehicle (EV) charging infrastructure, including the electrical distribution system upgrades needed to support the service. Overcoming the network economics challenge of rapidly growing both EVs and EV charging infrastructure is important, but there is still great uncertainty about the best technologies, locations, and business models for EV charging. Hence, flexible approaches to government support are needed, with a goal of supporting experimentation on charger siting strategies and business models.

RESEARCH AND DEVELOPMENT INVESTMENT
Rapid innovation is needed on multiple fronts to address the challenges and opportunities outlined above. Given uncertainties about technological progress, the authors argue that broad incentives that do not discriminate across zero-emissions resources—such as carbon pricing, clean energy standards, or clean energy subsidies—will be essential for directing capital toward cost-effective investments in clean energy infrastructure.

In addition, direct government investment in research and development in the energy sector is necessary to promote uninhibited knowledge sharing across industries and countries, and particularly with the developing world. Most of the GHGs in the atmosphere today were put there by what are now wealthier nations. But wealthy countries also have a self- serving incentive to share low-carbon technologies in order to avoid having nations now coming out of poverty ramp up their growth through intensified use of fossil fuels, because emissions from that pathway will create spillover damages to the wealthier countries.

WHOLESALE MARKET DESIGN
A core challenge in electricity markets is how to reward suppliers whose value does not fit the standard per-kilowatt-hour compensation paradigm. These include sources that can change output rapidly on demand, such as battery storage and hydroelectric generation, as well as sources that can be stored for very long periods of time and then reliably operate when needed.

Wholesale electricity markets have also been hindered by policies that subsidize and/or mandate certain sources of electricity generation rather than taxing greenhouse gas emissions. Instead of increasing wholesale prices, as a carbon price would do, these approaches have reduced wholesale prices, which has squeezed the profits of nuclear, hydro, and other generation sources that are near carbon free but have not been included in states’ subsidies and mandates. This problem has been especially severe for existing nuclear power plants, many of which are not earning revenue sufficient to cover their ongoing operating costs. Because wholesale prices are depressed by renewable generation policies, the standard market test doesn’t convey appropriate information about whether these plants should continue to operate. In addition, the authors contend that transitioning from bilateral power trades to highly organized and centralized wholesale auctions could help reduce costs and promote dispatch of renewable generation.

SUSTAINABLE RETAIL PRICING AND DISTRIBUTED ENERGY RESOURCES
Many states that have aggressively pursued decarbonized energy have covered costs by increasing volumetric electricity rates, oftentimes above social marginal cost. These high prices discourage electrification, create perverse incentives for behind-the-meter (BTM) generation and storage (such as rooftop solar installation), and disproportionately burden low-income customers, making it one of the least-desirable funding options. Alternative funding sources, such as fixed monthly charges indexed to income, or allocating state and federal funds from other sectors would be less regressive than raising electricity prices.

ENERGY INFRASTRUCTURE AND LOCAL COMMUNITY IMPACTS
Energy infrastructure, while integral to US economic growth, has historically produced negative local environmental impacts and local pollutants that have disproportionately harmed disadvantaged communities. There are many policy choices that could have inequitable impacts in the transition to zero-emissions sources. For example, broad, zero- carbon incentive policies could fail to reduce emissions from sources of local pollution that are concentrated in “hotspots” near disadvantaged communities. Similarly, if carbon capture technologies become a significant mechanism for achieving zero emissions, then plants that continue to burn fossil fuels might still emit local pollutants, even if the carbon is captured. Thus, policies that incentivize zero-emission energy infrastructure must also be accompanied by policies that guard against the possibility of local pollution “hotspots” during the transition.

STRANDED FOSSIL FUEL ASSETS
The environmental hazards presented by the current stock of orphaned wells must be addressed by direct public investment in their decommissioning. The decommissioning of all of these wells – including restoration of their surroundings – is likely to cost more than $10 billion by one estimate. A silver lining to this necessary expenditure is that it provides an opportunity to employ oil and gas workers who would otherwise be displaced by the transition to zero-emission energy infrastructure. One estimate suggests that decommissioning the full inventory of orphaned wells would require 100,000 person-years of labor. The natural source of such labor would be displaced oil and gas workers, whose skills would naturally transfer to well decommissioning work. Well decommissioning would employ a valuable stock of human expertise that would otherwise depreciate away were these workers to move to another sector or drop out of the labor force. These job opportunities may also soften political opposition to the zero-emission energy transition. Although well decommissioning is not a long-run solution to displacement of oil and gas workers, it could serve as a temporary bridge to future employment opportunities in geothermal energy, offshore wind, carbon sequestration, or other clean energy technologies.

CONCLUDING OBSERVATIONS
The authors conclude their report by noting that the defining energy challenge of the 21st century is to transition the provision of energy services to zero-emissions sources, while simultaneously controlling costs and ensuring the reliability of energy supply. This transition will require historic investments in zero-emissions energy generation, transmission, storage, and distribution infrastructure. Federal policy choices will play a leading role in determining whether, where, and when these investments will occur, how costly they will be, and who will bear those costs.

Suggested Citation: Borenstein, Severin, and Ryan Kellogg. December 10, 2021. “Challenges of a Clean Energy Transition and Implications for Energy Infrastructure Policy”. In Rebuilding the Post-Pandemic Economy, edited by Melissa S. Kearney and Amy Ganz. Washington, DC: Aspen Institute, 2021. https://doi.org/10.5281/zenodo.14057590. 

Chapter

Executive Summary

Science and Innovation: The Under-Fueled Engine of Prosperity

Benjamin Jones (Kellogg School of Management at Northwestern University) argues that the United States is vastly underinvesting in science and innovation, hindering productivity growth. He presents evidence that increased public spending on research and development would lead to improvements in standards of living and health, enhance our economic competitiveness, and advance our nation’s capacity to respond to crises.

Jones reviews the empirical evidence of the benefits of basic science investments and, drawing on the evidence, calls for three spheres of policy action:

  1. Scaling funding resources: Evidence suggests that doubling the total U.S. public investment in R&D would easily pay for itself through social benefits in excess of costs. Jones notes that a sustained doubling of all forms of R&D expenditure could raise U.S. productivity and real per- capita income growth rates by an additional 0.5 percentage points per year over a long time horizon.
  2. Scaling the people pipeline into science and innovation careers: Jones emphasizes that achieving the benefits of the science and innovation system will rely on having a robust science and innovation workforce—a process which requires long-term investment to cultivate individuals with relevant training and talent. In particular, he highlights the potential large returns from increasing domestic investment in STEM workers and importing talent through reforms to high-skill immigration policies.
  3. Making diverse investments: Risk is inherent to the innovative process. Instead of trying to pick “winners,” policymakers and the public must approach high-risk investments as making a wide portfolio of bets. This approach can produce more efficient search, lower collective risk, and increase returns in the science and innovation system. Public funding should push for diverse pathways instead of crowding into particular areas, and focus on increasing the scale of funding and human capital, and the diversity of approaches that are taken.

Finally, Jones cautions that policymakers should not become hamstrung over which policy levers to employ, such as R&D tax credits, funding the NIH, investments through DARPA. There is a need for more research about which channels feature the highest returns, but, Jones argues, the biggest failure would be to continue to underinvest, given the enormous social returns that could be gained.

SCIENCE AND INNOVATION AS A PUBLIC GOOD
Left to its own devices, the private market will underinvest in R&D. When one party engages in costly and risky work to discover a new idea, any discovery can be imitated and built upon by others, creating broader social value to the original work. The original innovator is unlikely to capture this broader social value, and the private value to the original innovator can fall far short of the social value it creates. As a result, private incentives to invest in creating a new idea may be well below the social interest in making that investment. This is the basic market failure and incentive problem that surrounds the advance of ideas.

Public policy plays an important role in addressing this market failure by providing additional funding or incentives for R&D through a variety of mechanisms: government-sponsored research funding, intellectual property systems, research and development (R&D) tax credits, prizes, public research contracts, and advanced purchase commitments. Public funding for R&D investment as a share of US GDP is now at a 60-year low, as shown in Figure 1. This lack of investment is striking in light of recent and evolving challenges. The United States has faced a slowdown in productivity growth and rising concern about the international competitiveness of the U.S. workforce over many years, real wages for the median household have struggled to rise and have failed to keep pace with the gains in prior generations. Yet, as productivity has lagged, U.S. R&D intensity has slipped compared to other countries. In the mid-1990s, the United States was in the top five of countries globally in both total R&D expenditure as a share of GDP and public R&D expenditure as a share of GDP. Today, the United States ranks 10th and 14th in these metrics.

FIGURE 1: U.S. R&D SPENDING OVER TIME

THE EVIDENCE ON SCIENCE AND INNOVATION
Jones summarizes a large number of studies that estimate the social return to investment. He documents evidence on the very high social rates of return—in excess of 20% —found across agriculture, manufacturing, biomedical research, and in studies of the economy as a whole. In his own research conducted with Lawrence H. Summers of Harvard University, he finds the social rate of return to R&D expenditure in the U.S. economy appears to exceed 50%. Their analysis indicates that $1 invested in innovation produces, conservatively, at least $5 in social benefits on average—and quite possibly $10 or even $20 in social benefits per $1 spent.

COUNTERING THE SKEPTICS
Despite very high expected returns, the economic uncertainty and political risk inherent to federal investment in basic research have fueled skepticism about allocating scarce public funding to R&D. Innovative endeavors have a high failure rate. Since R&D is explicitly a search into unknown territory, judging success and failure ahead of time is difficult. And even when an idea is in hand, there is uncertainty about its future prospects, with eventual success often preceded by apparent failure. For every CRISPR breakthrough, there is also a Solyndra anecdote. Despite the risks involved, the evidence is abundantly clear: society reaps extremely large, long-run returns from public R&D investment.

Skeptics also question whether government officials have the capacity to identify and invest in good opportunities, or they question the value of experts themselves, who are often depicted as disengaged from “real-world” problems. Countering the skeptics of direct government funding, Jones documents the direct link between public funding and the public use of scientific research. Publicly-funded research has rich interfaces with public use, including for marketplace invention and government policy.

THE PEOPLE WHO DRIVE INNOVATION
A crucial driver of innovation are policies that advance human capital. Children exposed to innovators and entrepreneurs are more likely to become one themselves. In particular, girls who move to regions with higher shares of female inventors are more likely to become inventors. Importing talent through immigration is also critical. US immigrants patent more often than native-born Americans and make up a disproportionate share of the science and engineering workforce. Immigrants also account for a disproportionate share of entrepreneurs and are more likely to start companies of all sizes, including high-growth start-ups.

A VARIETY OF R&D POLICY LEVERS
There are myriad channels through which the federal government invests in R&D. For instance, DARPA invests in technology for national security, the tax code features R&D tax credits for private sector innovators, the NIH, Department of Energy, and the National Science Foundation fund research through a large network of national laboratories, research universities, and private-sector research contracts.

Large social returns to R&D are observed across all channels. It is less clear from the research which channels yield the highest returns. Jones cautions against allowing this uncertainty to lead to inaction altogether. While it is possible that such information would produce even higher returns, the true failure is not to put more resources toward R&D investment overall, which would lead to enormous positive returns.

Suggested Citation: Jones, Benjamin F. December 10, 2021. “Science and Innovation: The Under-fueled Engine of Prosperity”. In Rebuilding the Post-Pandemic Economy, edited by Melissa S. Kearney and Amy Ganz. Washington, DC: Aspen Institute, 2021. https://doi.org/10.5281/zenodo.14057609.

Chapter

Executive Summary

Economic Perspectives on Infrastructure Investment

THE ECONOMICS OF FEDERAL INFRASTRUCTURE INVESTMENT
Bipartisan support for new infrastructure spending reflects a consensus view that well- chosen infrastructure investments would enhance American economic competitiveness and increase the economy’s productive capacity. Sound investments also have the potential to accelerate the US economy’s transition to sustainable energy sources and to address some of the sources of income inequality, such as access to transportation services.

In Economic Perspectives on Infrastructure Investment, James Poterba of MIT and Edward Glaeser of Harvard highlight policy relevant lessons from the voluminous research literature on the economics of infrastructure projects. They draw four main conclusions that inform the design of an infrastructure-spending program:

  1. Project-based consideration of costs and benefits is an essential input in determining the optimal level of infrastructure investment.
  2. Cost control should be an integral part of infrastructure project planning, starting with procurement procedures and planning for the possibility of project delays and unexpected costs.
  3. User fees can, in some settings, be an attractive source of revenue to cover capital costs and ongoing maintenance needs. When set to reflect the marginal cost of using infrastructure, these fees encourage efficient utilization.
  4. Though public-private partnerships and the privatization of infrastructure assets can sometimes enhance operational efficiency and improve both procurement and management, private provision can shift risks to the public sector.

Glaeser and Poterba describe in detail the research that leads to these policy-relevant conclusions. Their paper describes various approaches to assessing the appropriate level of infrastructure investment and to undertaking cost-benefit analyses. They also discuss why the costs associated with U.S. infrastructure projects appear high by international standards, and why there is such variability in costs across similar projects. They discuss the challenges of financing of infrastructure projects and describe some of the political economy challenges that they face.

OPTIMAL INFRASTRUCTURE INVESTMENT
The authors stipulate that the optimal level of infrastructure investment should be determined by a project-by-project consideration of the costs of acquiring infrastructure capital and the benefits of using it. This approach contrasts with an “engineering” approach —such as that of the American Society of Civil Engineers (ASCE) and some historical approaches to estimating infrastructure demand—which define infrastructure needs without reference to the costs of meeting them, and often without recognizing alternatives to infrastructure spending as a way of meeting these needs.

Cost estimates include inputs such as concrete and steel, construction labor, borrowing costs, potential disamenities such as pollution, and distortions associated with tax finance or other funding mechanisms. Benefit calculations should should include the project’s impact on the users of the infrastructure, as well as society as a whole through channels such as macroeconomic stimulus. With a fixed budget for infrastructure investment, projects should be undertaken in order of their benefit-to-cost ratio. Because borrowing costs are one of the costs of infrastructure projects, a fall in real interest rates translates into an increase in the optimal level of infrastructure investment would rise, all else equal.

The paper presents some evidence that calls into question the prevailing narrative of America’s “crumbling infrastructure.” Interstate highways today are smoother, fewer bridges are structurally deficient, and dam collapses are less frequent than in the past. The cost-benefit approach to infrastructure analysis recognizes that in some cases, when estimated benefits of a given project are low, it may be more cost-effective to a structurally compromised bridge or road than rebuild or repair it.

PROJECT SELECTION USING COST-BENEFIT ANALYSIS
The application of project-specific cost-benefit analysis to individual infrastructure projects, and a comparison of cost-benefit ratios across projects, is quite different from a policy process that establishes an aggregate spending number and then allocates that amount to different jurisdictions based on population or other considerations. Even with a top-down allocation, however, cost-benefit analysis can play an important role in project selection at the state or local level when a fixed pool of resources must be allocated across alternatives.

Glaeser and Poterba acknowledge that cost-benefit analysis involves considerable uncertainty, which means the evaluator has opportunities to exercise discretion. It is thus important to develop institutions that can perform non-partisan, rigorous cost-benefit calculations and use a systematic approach. They highlight six features of cost-benefit analysis that are similar across many forms of physical infrastructure: estimating future benefits, measuring systemic impact, assessing ancillary benefits and costs, estimating macroeconomic effects (such as anti-recessionary stimulus and agglomeration economies), measuring impact on GDP and productivity, and considering distributional impacts.

The authors acknowledge the practical difficulty of performing project specific cost-benefit analyses before legislation is passed. Doing so would involve a lengthy process that doesn’t often fit with the political reality of legislative debate and passage. Given this reality, they consider three ways in which cost-benefit analysis could be inserted more directly into the process of allocating U.S. infrastructure spending. One is to estimate the benefits of infrastructure spending as a whole, rather than the benefits of particular projects, and to compare this with the social costs of raising funds for infrastructure projects. There are important limitations of such an approach, notably the difficulty of establishing a credible

estimate of the social benefit of national spending. The second is the creation of an infrastructure bank that would receive some fraction of federal infrastructure spending and allocate the funds to projects that appear to have particularly high benefits relative to costs. The authors caution that while in principle, an infrastructure bank could be a nimble agency that chooses high return projects, it runs the risk of becoming a political tool. A third option, which the authors view as the most likely to succeed, would involve federal mandates for states to make more use of cost-benefit analysis when they spend federal dollars, perhaps with input from a federal agency that develops and applies cost-benefit methods.

AMERICA’S UNIQUELY EXPENSIVE INFRASTRUCTURE AND OPTIONS FOR COST- CONTAINMENT
American infrastructure costs are very high by international standards, which could imply that the optimal amount of infrastructure investment in the U.S. is lower than in other countries. Some recent projects, such as the East Side Access project in New York City which cost $4 billion per kilometer of tunnel, are the most expensive infrastructure expansions anywhere. Costs also appear to have risen over time. Highway costs, according to the DOT National Highway Construction Cost Index, increased by 32% in real terms between 2003 and 2020.Though it is difficult to draw general conclusions about cost drivers, expensive designs, procurement costs, and labor costs all seem to contribute.

The authors suggest several strategies to control costs. Maintaining existing infrastructure is often more cost-effective than building new projects, although the political system sometimes exhibits “ribbon-cutting bias” in favor of new projects. Other strategies include:

  • Applying cost-benefit analysis
  • Purchasing from low-cost suppliers, if necessary by relaxing “Buy American” provisions
  • Streamlining environmental reviews
  • Harmonizing implementation of prevailing wage requirements Strengthening local procurement offices.
  • Improving procurement practices and project management.

FINANCING INFRASTRUCTURE EFFICIENCY AND EQUITY CONSIDERATIONS
User fees and congestion charges are efficient ways to pay for infrastructure construction and maintenance, although distributional considerations often restrict their use. Requiring users to pay for their infrastructure limits overuse and generates revenues. Although in some cases low-income users of infrastructure would be disproportionately burdened by user fees, this is not the case for all types of infrastructure. Given different patterns of infrastructure use at different income levels, some user fees, such as airport user fees, would be charged primarily to higher income individuals. Furthermore, the impact of user fees on low-income users could be offset with targeted rebates or vouchers.

Another approach to funding infrastructure spending is the use of public private partnerships (PPPs). Such partnerships are a means of bringing in financial support beyond that available from the public sector. However, public-private partnerships and the privatization of infrastructure assets can also pose risks to the taxpayer. Glaeser and Poterba note that private providers sometimes negotiate advantageous terms in bargaining with state and local governments, and they point out in some cases, the cost of capital for the public entities may be lower than that for private investors.The private sector may bring important operational and procurement advantages to infrastructure projects.Those considerations, as well as capital access advantages in some cases, can justify public- private ventures or privately-provided infrastructure.

POLITICAL CHALLENGES AND PITFALLS
There are numerous political challenges that confront infrastructure policy in the U.S. These include limited interest in maintaining existing assets and special interests blocking valuable projects.States have the greatest opportunity to address many of these problems, given their ability to govern local politics, pass state laws relating to the environment, or change the procurement process for state agencies. Reducing costs may require the federal government to interact with the states and to make reform a precondition for funding. The authors suggest that a federal authority could negotiate with states on a project-by-project basis in order to lower costs and increase value.

Suggested Citation: Poterba, James, and Edward Glaeser. December 10, 2021. “Economic Perspectives on Infrastructure Investment”. In Rebuilding the Post-Pandemic Economy, edited by Melissa S. Kearney and Amy Ganz. Washington, DC: Aspen Institute, 2021. https://doi.org/10.5281/zenodo.14057633.

Chapter

Executive Summary

14 Facts about US Investments in Infrastructure and R&D

DOWNLOAD

Bipartisan support in Congress is emerging for new strategic investments in science and technology, in addition to a trillion-dollar infrastructure deal. These proposals reflect an emerging consensus that large-scale government investments are necessary to support the US economy’s transition to sustainable energy sources, address underlying sources of domestic inequality, and promote American economic competitiveness. 

In order to inform the debate over the size and scope of new investments, we present this set of 14 facts with the goal of advancing an accurate understanding of the current state of public investment in the United States and, in particular, investment in infrastructure and research and development (R&D). As demonstrated in this memo, however, this is impossible to characterize the current situation in simple soundbites. On the one hand, US public investment in R&D is at a 60-year low. On the other hand, private R&D investment in the United States is reaching historic highs. America’s highway infrastructure is aging and the cost of building new roads has risen dramatically, but road quality also continues to improve and highways are becoming smoother. The United States invests less in transportation infrastructure as a share of GDP, on average, than the OECD or China. But the United States still leads the world in total (combined public and private sector) R&D investment.    

The charts below provide a nuanced picture of federal investment and specifically nondefense investment, which falls into three categories: education and training, physical capital, and R&D. We focus on the latter two categories, in particular.1 Section I describe long-run trends in R&D and physical capital investments in the United States. Section II describes trends in overall infrastructure investment, by sector and across different categories of investment, highlighting the rise of digital infrastructure investment. Section III looks at unique features of the US highway infrastructure. Finally, Section IV places trends in US public investment into context by comparing them with investment trends in China and other advanced economies.

I. Overall Trends in Government Investment

Government investment has the potential to boost long-run productivity, wages, standards of living, and international competitiveness and to help redress long-standing sources of inequality.  Government investment also provides public goods and services that the private sector tends to underproduce. For example, public funding for research and development has large, positive social externalities that private firms do not take into account when making decisions about how much to invest. Unlike public consumption, which often focuses on smoothing business cycles, government investment is made with an eye toward long-term economic goals.

In 2018, the US government invested less as a share of the economy relative to any other year since the early 1960s. This fact is especially striking since government debt as a share of the economy reached 104% in 2018, nearly surpassing the previous peak during World War II. With the exception of a temporary increase in funding for infrastructure projects and investment in response to the Great Recession, the recent run-up in federal debt has largely financed public consumption rather than investment.

1. Total federal investment as a share of the economy is declining.

Figure 1: US federal investment as a share of GDP, 1962–2018

Source: CBO.

Figure 1 shows the breakdown of federal investment as a share of GDP across defense and nondefense categories. At 2.4% of GDP, total investment is far below the 1968 peak of 6.2%. Although both types of investment have declined, defense investment as a share of the total is shrinking at a faster rate. In 1962, defense investment made up 72% of the total. By 2018, that share had fallen to under 40%.

2. US federal investment in R&D and physical capital is at historic lows, while federal investment in education and training remains on par with the historical average.

Figure 2: US federal nondefense investment as a share of GDP, 1962–2018: physical capital, education and training, and R&D

Source: CBO and author’s calculations.

Figure 2 breaks down nondefense investment into its three primary categories: physical capital, education and training, and R&D. As a share of the economy, the US federal government invested just under 1.5% of GDP across all three categories in 2018. This is well below the historical average of 2% of GDP per year and is the second lowest level between 1962 and 2018. 

US investments in physical capital peaked in the 1960s and 1970s at 1.1% of GDP, a period which included the construction of the interstate highway system. Physical capital investment increased temporarily to 0.9% in 2009 and 2010 due to investments associated with the passage of the American Recovery and Reinvestment Act (ARRA). 

Federal investment in education and training rose from less than 0.5% in the 1960s to nearly 1% of GDP by 1975 due to a large increase in spending on elementary, secondary, and higher education and on education benefits for Vietnam War veterans (CBO 2019). By 1985, it had shrunk back down to just under 0.5%. Education investment as a share of the economy rebounded to 1% of GDP again in 2010 under ARRA, which temporarily increased funding for Pell grants for higher education, along with spending for K-12 education and secondary and vocational training. Today, at 0.6% of GDP, education and training investment is on par with the average over the previous six decades. 

Federal nondefense investment in R&D has steadily declined for several decades. It has consistently been below 0.5% of GDP since the early 1980s, and now, at just 0.3% of GDP, has reached the lowest share on record.

Recently, there has been heightened political interest in increasing federal R&D spending to increase American competitiveness with China. The Innovation and Competition Act (2021), passed by the US Senate in June 2021 with bipartisan support, authorizes over $140 billion over five years to fund R&D activities through the National Science Foundation, the Department of Commerce, the Department of Energy, and NASA. If implemented, the Act would increase federal R&D spending to more than $90 billion per year over the next five years—almost double the historical average of $46.4 billion per year (in 2018 dollars).

3. Although federal R&D investment is near a historic low, business R&D has been rising for decades.

Figure 3: US R&D spending by source, 1953–2019

Source: National Center for Science and Engineering Statistics (NCSES) National Science Foundation, Table 1.

Alongside declines in federal R&D investment, R&D investment by the business sector has been on an upward trend for over half a century. In 1980, US private sector R&D investment surpassed that of the federal government (Figure 3) and the gap continues to widen. From 2010 to 2019, total inflation-adjusted R&D investment grew at a pace of 3.3% per year, which was driven entirely by private investment and outpaced the average real GDP growth of 2% over the same period (NCSES 2021). One might suspect from looking at these trends that public and private investment are substitutes, but existing research suggests the opposite is the case. For example, OECD (2000) finds an additional dollar of government R&D investment increases business-financed R&D by 1.7 dollars, on average.  

Business R&D reached a high of 2.2% of GDP, or roughly $463 billion in 2019. Meanwhile, total federal R&D spending was just over 0.6% of GDP in 2018, among the lowest levels on record. “Other” R&D spending remains a small but growing portion of total R&D spending and includes investment by state and local governments, higher-education, and non-profit institutions.

Although direct federal spending on R&D is declining as a share of the economy, the federal government also promotes private sector R&D through the federal R&D tax credit. In 2021, the Joint Tax Committee (2018) estimates federal expenditures through the R&D tax credit to be worth $11.6 billion. These tax credit benefits are not included in the direct federal spending amounts depicted in Figure 3. However, the share of R&D tax incentives in total government support for R&D has increased over time, from approximately 29% in 2000 to 38% in 2016 (OECD 2020).

II. US Infrastructure Assets and Investments

This section explores trends in US infrastructure investment and the stock of infrastructure assets. As recent public debates illustrate, defining what constitutes infrastructure is a difficult undertaking. We rely on definitions provided by Bennett et al. (2019), who analyze the national economic accounts to define and measure three broad categories of infrastructure investment: basic, social, and digital. The authors define “basic” infrastructure as traditional assets such as transportation, water, sewer, and power. “Social infrastructure” refers to public and private hospitals, schools, and public safety facilities. Finally, “digital infrastructure” includes communications structures and equipment, software, and computers.

Public capital (infrastructure) and technological progress (driven by R&D) can play an important role in stimulating both long-run economic growth and labor productivity (and thereby long-run wage growth). Increasing the amount of capital available per worker hour, known as capital deepening, and technological progress are important drivers of labor productivity, along with human capital investments. Public investments that raise the total amount of capital available per worker and accelerate technological innovation through investments in R&D fuel productivity growth and therefore long-run wage and economic growth. Ramey (2020) summarizes existing estimates of the elasticity of output to public capital, which range from .08 to .39.

4. Basic infrastructure still accounts for the largest share of infrastructure spending across public and private sectors, but the share of digital infrastructure investment is catching up.

Figure 4: Infrastructure investment by type, 1947–2017

Source: Bennett et al. (2019), BEA, author’s calculations.

Total infrastructure investment across all categories (basic, social, and digital) and across public and private sectors has remained relatively steady since the early 1980s, but the composition of spending has changed significantly. In 2017, digital infrastructure investment was 1.4% of GDP compared to 1.4% for social infrastructure and 1.7% for basic. This is in contrast with the year 1980, when digital infrastructure was just 0.33% of GDP, social was 1.2%, and basic was more than 3.1%. Since the year 2000, constant dollar digital infrastructure investment has increased by an average of 6.7% per year, while social infrastructure has increased by 1.9% per year, and basic infrastructure has grown at an average pace of just 0.5% per year.

5. Total spending on basic infrastructure across the public and private sectors is at a historic low.

Figure 5. Basic infrastructure investment by sector, 1947–2017

Source: Bennett et al. (2019), US Census Bureau, author’s calculations.

Figure 5 focuses only on basic infrastructure, including water, sewer, power, and conservation and development across public and private sectors. Total investment in this category peaked in the late 1950s, driven by public investments in the federal highway system. Both total basic investment and public sector basic investment as a share of GDP reached a minimum in 2017 at 1.68% and 0.87%, respectively. When measured on a per-capita basis, public and private basic infrastructure investment have converged over time. In 2017, per-capita private infrastructure spending was $455 per person per year, compared to a historical average of $340 per person. The same year, public spending was $486 per person per year, compared to a historical average of $566 per person.

6. The growth rate of digital infrastructure assets per capita reached an all-time high in 2017, while growth in the stock of other infrastructure assets is slightly above population growth.

Figure 6. Growth rate of the real net stock of basic, social, and digital infrastructure assets, per capita, 1990–2017

Source: Bennett et al. (2019), US Census Bureau, author’s calculations.

Figure 6 shows the growth rates of the stock of infrastructure per capita, in contrast to the flow measures of infrastructure investment  shown in Figures 4 and 5.  Measuring the stock of infrastructure assets helps to inform how well infrastructure investment is keeping up with the depreciation of assets and is thus an indicator of infrastructure’s productive capacity. We focus on per capita measures of infrastructure assets rather than stocks as a share of the economy because the amount of capital per worker (including public capital) drives labor productivity. 

Figure 6 illustrates that basic infrastructure investment has just kept pace with population growth, growing at an average rate of 0.5% per year since the late 1990s. The stock of social infrastructure assets per capita has grown at an average pace of 1.6% per year since 1990, peaking in the early 2000s at 2.5% before the financial crisis of 2008. For the period 2010–2017, social infrastructure investment has grown at just 1% per year, on average. The story of digital infrastructure investment stands in sharp contrast to the other categories. The growth of the stock of digital infrastructure per capita has accelerated since 1990, growing at an average pace of 3.5% over the period 1990–2017 and peaking in 2017, the latest year for which data was available, at 7.2%.

7. The United States has made significant progress in expanding broadband internet access over the past decade.

Figure 7: Deployment of fixed terrestrial 25/3 Mbps services, by area type, 2012–2019

Source: FCC.

The availability of broadband internet is an issue of longstanding interest among policymakers. Moreover, the COVID-19 pandemic further underscored its increasing importance for economic opportunity and education. Figure 7 shows that broadband internet coverage has seen consistent and significant gains in recent years. The Federal Communications Commission (FCC) defines broadband internet by a benchmark download speed of 25 Mbps and an upload speed of 3 Mbps for fixed services. By 2019, 95.6% of Americans had access to fixed terrestrial broadband service, up from 81.2% in 2012. The largest gains over the period took place in rural and tribal areas, although they continue to lag behind urban residents. In 2019, an estimated 14.5 million Americans (4.4%) lacked access to fixed terrestrial broadband, with 11.3 million (17.3%) in rural areas, 846 thousand (20.9%) in tribal lands, and 3.2 million (1.2%) in urban areas (FCC 2016 – 2021).

III. Transportation Infrastructure

The idea that America’s transportation infrastructure is aging and therefore decaying is a common motivation for new physical infrastructure investments. This section highlights relevant facts about the US highway infrastructure that challenge this notion. We also highlight the large differences in the amount of spending on highways across states and the fact that the cost of building new highways has risen over time. For the sake of length, we focus on highways, but refer readers to the work of Duranton, Nagpal, and Turner (2020) for a discussion of the quality and condition of other types of transportation infrastructure, including bridges, bus transit, and subways.

8. The US interstate system is aging, but road conditions are improving.

Figure 8: Average age of US highways and streets vs. share of US interstate miles with the lowest International Roughness Index ratings, 1995–2019

Source: Federal Highway Administration, BEA, author’s calculations. Note: IRI data not available for 2010.

The majority of the US interstate highway system was built between the 1960s and early 1990s. As a result, the average mile of road on an interstate highway is nearly 40 years old. However, this aging does not necessarily reflect a crumbling infrastructure system; highway quality today is as good as ever. Figure 9 illustrates the relationship between the average age of US highways and the percent of US interstates that rate the best in the International Roughness Index (IRI). The IRI was devised by the World Bank in 1986 to create a standardized roughness measurement of roads and highways (TRID 1995).2 IRI ratings can range from less than 60 inches per mile to more than 220 inches per mile, though any rating less than 95 inches per mile is generally considered “good” (MDOT). As shown in Figure 9, despite the aging of the US interstate system, the percentage of US interstate miles with a “good” IRI rating has increased more than 30 percentage points since 1995. Continued federal funding for maintenance and operations (shown in Figure 11) has resulted in smoother highways over time.

9. There are large differences in the level of highway spending per capita across states.

Figure 9: Per capita public highway infrastructure investment by state, 2017

Source: Bennett et al (2019), author’s calculations.

Figure 9 illustrates the variation in per-capita spending by state on highway infrastructure. The range of funding extends from $129 per person in Michigan to ten times that amount ($1,377) in North Dakota. Several of the least populous states in the Midwest and North Central states (Iowa, Minnesota, Nebraska, North Dakota, South Dakota, and Wyoming) are in the  highest quintile of states in terms of highway spending per capita. In contrast, states in the western United States, including Arizona, California, Colorado, Oregon, and Utah, are in the bottom quintile of highway spending per capita. Many economists have documented that expected returns to additional highway infrastructure vary significantly across states, with greater returns in more dense areas (Allen and Arkolakis 2019; Poterba and Glaeser 2021). For example, Allen and Arkolakis (2019) find large welfare benefits from additional highway construction in California and in the greater New York City area. However, per-capita spending appears to be more closely related to geographic location rather than to states’ population density.

10. States spend most of their infrastructure funds on operations and maintenance.

Figure 10: Spending on transportation and water infrastructure, by category of spending, 2017

Source: CBO.

Figure 10 shows the breakdown of US transportation and water infrastructure spending across the federal vs. state and local government, after federal grants and transfers to states. The figure also illustrates the different uses of infrastructure funding at different levels of government. The vast majority of state and local funding goes to operations and maintenance rather than new physical infrastructure. Given the maturity of US transportation and water infrastructure, maintenance and operations can be expected to account for a large share of total spending.  In fact, total state and local infrastructure spending has increased over time, driven by operations and maintenance. In contrast, the majority of federal funding for transportation and water infrastructure goes to capital investments rather than operations and maintenance.

11. The cost of US highway infrastructure has dramatically increased since the mid-1970s.

Figure 11: Spending per mile of new highway construction, 1955–1993

Source: Reproduced from Brooks and Liscow (2020).

An important consideration for policymakers when determining the desired amount of infrastructure is cost. As Poterba and Glaeser (2021) highlight, building infrastructure in the United States is far more expensive than in the rest of the world. Figure 11 focuses specifically on highway construction, illustrating the dramatic increase in cost per new highway mile that occurred between the mid-1970s and the 1990s, when the US interstate system was constructed. Over the period from 1973 to 1991, highway spending per mile increased 193%. Brooks and Liscow (2020) document that increased spending does not correlate with increases in construction wages or the cost of materials, such as concrete, construction machinery, construction sand, paving mixtures, and steel mill products. Over the period, materials prices rose by only 7% and construction hourly wages decreased in real terms by 39%. The authors find increased highway construction costs are positively correlated with increases in incomes and housing values over the same period, as well as an increase in local demands for physical structures to be built along with Interstates that alleviate concerns about noise or other negative externalities.

IV. International Comparisons

A primary motivation among policymakers for dramatically increasing US infrastructure investment is a desire to remain economically competitive with the rest of the world. This section highlights how the United States compares to other countries by some of the metrics discussed above. While it is true that the United States trails other advanced economies in transportation spending, the opposite is true for R&D investment when private sector investments are taken into account. We also highlight the dramatic increase in public investments by China in recent decades.

12. US gross capital formation is in the middle of the pack among advanced OECD economies.

Figure 12a: Government gross fixed capital formation as a share of GDP in advanced OECD countries, 2019

Source: OECD Statistics Datasets 1 & 11 and authors’ calculations. Note: Data for Japan not available for the year 2019.

Figure 12b: Government nondefense gross fixed capital formation in advanced OECD countries, 2019

Source: OECD Statistics Datasets 1 & 11 and authors’ calculations.

Figures 12a and 12b compare levels of US fixed capital formation—or investment in productive assets—relative to other advanced OECD countries. In 2019, the United States ranked 11th of 24 in total capital formation among advanced OECD economies when defense investments are included. As Figure 12b shows, the United States falls to 17th among advanced OECD economies when comparing only nondefense investments. The United States ranking among this group has not improved in recent decades: in 2000, the United States ranked 11th in total fixed capital formation and 14th in nondefense fixed capital formation.

13. The United States exceeds China and the OECD average in total R&D investment as a share of GDP.

Figure 13: Total R&D investment (public and private) as a share of GDP in selected countries, 1981–2019

Source: Figure 3 in Van Reenen (2019), OECD, author’s calculations. Note: The OECD defines gross domestic spending on R&D as “the total expenditure (current and capital) on R&D carried out by all resident companies, research institutes, university and government laboratories, etc., in a country.” It includes R&D funded from abroad, but excludes domestic funds for R&D performed outside the domestic economy.

Figure 13 illustrates overall R&D investment in the United States, China, and the OECD average, excluding China and the United States. Although federal funding for R&D in the United States has been declining, the private sector’s contribution places the United States above the OECD average as a share of GDP. Despite being above average, the United States invests less in R&D than some advanced economies as a percentage of GDP. South Korea leads the OECD overall, investing nearly 5% of its economy into R&D in 2019. Germany and Japan also exceed the United States in R&D investment as a share of their economies. 

China has experienced rapid growth in R&D investment since the mid-1990s, outpacing its overall economic growth and rising to 2.2% of its economy by 2019. The country’s total R&D investment when measured in constant dollars increased by an average of 14% per year between 1991 and 2019. 

Among the four largest western economies—the United States, Germany, the United Kingdom, and France—only Germany has made a meaningful increase in R&D investment as a share of its economy since 2000, increasing investment by 2.7% per year.

14. Transportation infrastructure spending as a share of GDP is far higher in China than in the United States or the OECD average.

Figure 14: Transportation infrastructure across select economies, 1995–2019

Source: OECD.

Figure 14 shows investment in transportation infrastructure, including both new structures and maintenance, as a share of GDP across advanced economies. Although the initial stock of infrastructure varies by country, China far outpaces other advanced economies in both the share and level of transportation investment. The United States also trails most other advanced OECD countries in transportation infrastructure investment. At just 0.6% of GDP, the United States lags behind France, Germany, Japan, Korea, and the United Kingdom in transportation infrastructure investment as a share of their economies.

 

The data underlying these figures are available for download here.

1 See Expanding Economic Opportunities for More Americans

2 According to the FHWA, a road’s IRI rating is “calculated from a single longitudinal profile measured with a road profiler in both the inside and outside wheelpaths of the pavement. The average of these two IRI statistics is reported as the roughness of the pavement section.” (FWHA).