How advanced technologies are reshaping manufacturing

LIVE

How advanced technologies are reshaping manufacturing
Sections

Research

How to boost long-run growth after COVID-19

shutterstock_1798980148
Editor's note:

This brief is part of the Brookings Blueprints for American Renewal & Prosperity project.

While the economic shocks of COVID-19 have occupied the full attention of policymakers, once the pandemic is under control the problem of slow long-term growth will resurface. This report outlines the challenges of long-term growth that America faces and policy proposals to help overcome them.

  • An initiative on worker training and re-training is essential to a more productive but inclusive economy. Such an initiative is a gamble because of past failures, but policymakers must try again and learn lessons from past failures and successes.
  • To sustain employment and support productivity going forward, there should be renewed funding available for loans and grants to struggling companies, both large and small. Taxpayer losses are inevitable as some companies are rescued only to fail later. Given the impact of the pandemic, that cost is worthwhile.
  • Our current immigration policy is counterproductive for long-term growth. Immigrants with specialized skills, both technical skills and entrepreneurial talents, are vital to America’s role as a technological leader, and our immigration policy must welcome them here.
  • Public research and development (R&D) funding has declined, which is a mistake. A small amount of additional federal funding can overcome that problem. Support for R&D is particularly important given the weakness in manufacturing productivity described in this report.

Back to top


Challenge

As of the fall of 2020, the priority for economic policymakers should be to sustain employment, family incomes, and business viability. Even when the economy is on its way to recovery, however, slow long-run economic growth will remain a concern. The Gross Domestic Product (GDP) has grown at about two percent a year since the early 2000s compared to over three percent a year in the past. One reason for slower growth is that the population, and hence the labor force, are growing much more slowly. There is not a lot that can be done about that. The other reason is slow productivity growth, the focus here.

What can be done to boost productivity growth? This is hard to do, but even modest improvements to productivity growth constitute success because they compound over time into more substantial increases in living standards.

Why has productivity growth been so slow?

A standard measure of productivity is labor productivity, defined as real (inflation-adjusted) output per hour of work in the nonfarm business sector of the economy. Table 1 below shows how its growth rate has varied since 1948.

Growth in labor productivity table

From the end of World War II until the early 1970s there was strong productivity growth that powered the booming American economy and an era of rising living standards. Since then, productivity has grown at only half the rate, with the exception of the years from 1995–2004. In a 2016 study I examined the reasons for slower growth, and the brief surge starting in the mid-1990s. The conclusions can be summarized as follows:

  • The kind of innovation that traditionally powered productivity growth has slowed. These innovations included the automation of factories and economies of scale, such as building large electric power plants, and replacing small stores with supermarkets.
  • The service sector, where productivity increases are often slow or unmeasured, has become a larger part of the economy.
  • Regulations to protect worker safety, the environment, and consumer safety may have also slowed productivity and diverted investment into different uses. Though there is not a consensus on the issue, on balance it does not seem that regulation was a primary determinant of slower growth.
  • The surge of productivity growth in the 1990s was fueled by a wave of computer investment, rapid growth in big-box retailing, and in parts of the finance industry. These sources of growth ran their course in about a decade.
  • The level of education of the U.S. workforce is increasing but more slowly than in the past.
  • There is a widening gap between the productivity of the best firms and that of the average firm. Many new technologies have emerged, but have proven difficult to implement widely; they require skills that are not widespread, especially among employees of small firms.
  • In some cases, the productivity benefits of major innovations may lie in the future. For example, the shift to online shopping promises a boost to productivity but has not yet done so.

How has the economy changed, and will there be a post-pandemic productivity boost?

There were important economic trends emerging in the economy before the onset of the pandemic. Three of the most important are, first, the service sector was expanding its share of employment and GDP. Second, workers without a college degree saw their relative earnings decline as a result of automation and trade. And third, the rise of online retail transformed both big-box stores and shopping centers.

The pandemic accelerated the shift to online sales, which have increased to 16.1 percent of retail sales in the second quarter of 2020, up substantially from 10.8 percent the previous year. Automation has also been accelerated as companies have reduced person-to-person exposure to the virus. Not all trends have been accelerated by the pandemic. In the manufacturing sector, supply chains are now regionalizing in response to pandemic disruptions. This is a reversal of prior trends, where companies were instituting global supply chains based on the lowest-cost location for production (a trend that had already slowed in response to President Trump’s trade policy). The turn towards regionalizing production will not induce a significant reshoring of manufacturing jobs to the United States. Though less supply will come from China, any production that moves back to the United States will be highly automated, while the more labor-intensive activities will move to Mexico or remain in Asia.

Whether the pandemic accelerated, slowed, or even reversed a past trend, in each case it provided a huge shock to the economy, driving many companies out of business while opening opportunities for others. Jan Hatzius and his colleagues at Goldman Sachs see the pandemic triggering a wave of creative destruction, the term made famous by Schumpeter in the 1940s. Hatzius et al. forecast that there will be a productivity boost from the pandemic from three sources. First, they predict a shift away from low productivity industries like restaurants, hotels, recreation and traditional retail. Second, they predict the closure of unprofitable firms. And third, they predict that there will be pressure on surviving firms to cut costs and raise productivity. They estimate productivity growth of 2.7 percent annually between 2020-22, compared to 1.3 percent in the previous three years.

This positive post-pandemic productivity scenario is open to question. We know the pandemic’s effect on workers has been very uneven. Minorities, women, and those without higher education will find it harder to find good jobs in the recovery. Any predicted boost to productivity will be lost if full employment is only restored by consigning more and more workers to low-wage, low-productivity jobs. Productivity soared after the Great Recession, rising at a 3.5 percent annual rate in 2009 and 10. However, productivity growth averaged just under 0.6 percent a year over the subsequent four years as employment gradually returned.

Recent data from the Bureau of Labor Statistics reinforces the concern over U.S. productivity growth. Historically, the manufacturing sector has been a major contributor to productivity improvement for the whole economy, despite the sector’s modest size. In the eight years through 2019, however, manufacturing labor productivity growth has actually declined at about 0.4 percent a year, with a large decline of 1.1 percent in 2019. Chemical products, motor vehicles, printing, paper products, plastics and machinery were among 17 out of 19 manufacturing industries that saw productivity declines in 2019.

While a temporary productivity boost is possible, it will be a major challenge to achieve and sustain productivity gains in future years, while simultaneously restoring employment.

Back to top


Limits of historic and existing policies

The two main historical approaches to improving long-term growth have been to encourage investment in business capital and human capital.

Tax incentives for corporate investment

The Trump Administration cut corporate taxes in 2017, arguing this would spur greater investment. Following the Great Recession, investment hit a peak at the end of 2014 of just under 14 percent of GDP and has fluctuated within a narrow range since then. The predicted surge in investment after the tax cuts did not happen.

It is important to make sure that taxes do not stifle investment and that U.S. corporate tax policy is in line with competitor countries, but the tax cuts in 2017 were costly in terms of tax revenue lost and did not have the desired impact on investment.

Support for higher education

Both federal and state governments provide financial support to colleges and universities. The concern about current policy is that it leaves out a large part of the population that does not attend college or attends and then drops out. Support for higher education has been important to U.S. economic growth and has given the United States a large share of the world’s best research universities. It is important, though, to acknowledge the limitations of this policy. More needs to be done to provide skills and training to those who do not receive a degree from a four-year college.

Back to top


Policy recommendations

 

Improve workforce skills in the post-COVID-19 economy

Members of the workforce who have not completed a college degree and receive very little on-the-job training are shortchanged by our current educational system. These workers have seen their earnings stagnate, and they are undercutting their lifetime incomes by leaving the workforce early. To improve the post-pandemic labor market and to help companies be more productive, policymakers should invest in an expanded skills training program. This is not a short-term policy, it is a long-run project. Achieving stronger productivity and more inclusive growth depends on creating a more skilled workforce.

In designing a training program it is vital to avoid the problems that have plagued such efforts in the past. Anthony Carnevale, a training expert now at Georgetown, has laid out the key problems. Many previous programs were designed to train workers for manufacturing jobs even though the availability of manufacturing jobs is limited. Most past programs were run in collaboration with unions and employers have adversarial relationships with them. Some past programs have been costly to employers and workers they have trained quit for a better jobs elsewhere. Some parents are opposed to training programs, seeing them as pushing their children into second-class careers.Carnevale notes the federal training budget is only $8 billion, so we are not really trying to train the workforce. This proposal by Brookings expert Annelies Goger includes elements I just described. It recommends scaling up “earn and learn” programs in states and regions, in which employer and community college partnerships provide workers a compensation as they receive on-the-job training.

There are also lessons to be drawn from successful training programs. Harry Holzer, also of Georgetown and a Brookings nonresident senior fellow, has written extensively on training programs, finding that those with the right design have worked well. Programs do better when they provide training through community colleges that is linked to the needs of local employers. An assessment by Mathematica of registered apprenticeship programs in 10 states concluded that program participants had substantially higher earnings than nonparticipants. Other countries’ successes with training programs are also worth considering. Germany has a very extensive apprenticeship program that trains young people not just for manufacturing jobs but also for service sector jobs. Denmark has a program called Flexicurity which provides retraining for workers who have lost their jobs. If a company lays off workers, those workers are given generous financial support, but this is linked to their participation in a training program. If workers fail to participate in training or if they turn down a suitable job, they lose the financial support.

The potential for improvement is great. A recent study by Peter Q. Blair and colleagues found that as many as 30 million workers that do not have Bachelor’s degrees have the potential to obtain better jobs and earn more than they are paid the positions they currently hold. They have intrinsic skills that are not being using in their current jobs.

In designing a new program, I propose that states and cities have the flexibility to organize their efforts in accordance with local labor market conditions. Programs should be run in collaboration with employers that can specify the skills they need, provide instructors, and promise jobs to those that pass qualifying exams. In return, the employers get trained workers at low cost. Community colleges provide a natural setting for this training to take place, but instructors must have the necessary experience and skills to provide hands-on training. It is futile to train more people for manufacturing jobs than there are jobs available, but as the older generation retires there will be job openings in this sector. Further, there are many blue-collar jobs in the service sector that pay well and remain in demand, such as auto mechanics, heating and air conditioning servicers, and plumbers. Participants in training should be paid a stipend, conditional on attendance and receive a bonus for passing the program, which may help overcome resistance to vocational education.

Experimentation by local governments should be encouraged as long as data is made available to allow independent post-program evaluation of successes and failures. The federal government should underwrite the cost of the programs, recognizing that a percentage of the funds will be wasted but that a more skilled workforce benefits the whole economy. There should be extensive marketing and messaging around the training programs and investments in outreach. Providing a stipend plus a good outreach program will help overcome the resistance of younger people to a non-academic program and the resistance of older workers to re-training. Funds should be included to set up an easy interface to inform people about the programs and to help people enroll online. As part of the training, people should be coached on how to apply for jobs. Technology can help in the learning process as long as there are good teachers and mentors available.

Preserve productive companies

Thousands of companies are being driven out of business by the pandemic. While the CARES Act provided a lot of money to help businesses avoid failure, that money has mostly run out. There is still tremendous pressure on businesses where revenues have declined sharply.

While business failure is commonplace in normal times (400,000 businesses fail each year) the pandemic is highly atypical, and the failure of many or most of the companies going under is not related to their efficiency or productivity, but rather to how hard-hit their line of business has been. Even among companies that survive the pandemic, many will suffer long term damage. Retained earnings that would have gone to investments and to the development or adoption of new technologies will be lost and the companies may be forced to postpone actions that would have increased their efficiency and productivity. The loss of businesses that are basically sound will make it much harder for the economy to return to full employment.

Of all firms that close in normal times, over 90 percent do so without bankruptcy. Even among larger companies, most that enter bankruptcy end up being liquidated rather than reorganized. Bankruptcy filings have increased among large firms that were already struggling before the pandemic hit, but were pushed over the edge by the economic crisis.  Among small businesses, the hardest hit industries have been arts, entertainment and recreation, accommodation and food services, and educational services. Small retailers have been hit but not especially hard. It is likely that business distress, failures, and bankruptcies are getting worse as the pandemic drags on.

How should policymakers respond? The most hard-nosed approach is to let companies fail, but most economists and policymakers prefer to provide business subsidies. CARES provided such funding and any further rounds of support should also direct money to businesses. Economists highlight the potential market failures, or frictions, that can justify providing policy support. And another line of argument is that the bankruptcy courts will become congested and lead to more company failures than would be optimal. Adding extra judges and other legal reforms are part of the proposed solution.

I propose that policymakers provide generous, subsidized loans to businesses to help them survive, even though a program like this will inevitably sustain losses. It should look particularly for minority-owned businesses that were viable before the pandemic and could be so again. Large companies that fail should go through the bankruptcy process but policymakers should create a fund of debtor-in-possession financing because the private market will not be able to support all the companies that are viable in the long run. To qualify for access to this funding, private sector lending must be secured as well and these lenders must have skin in the game. There should be rigorous accountability and clawback provisions so that profitable companies are required to pay back taxpayers once they recover.

Encourage talented people to come to the United States

Foreign-born individuals have made major contributions both to U.S. entrepreneurship and to science and technology. A 2016 study looked at U.S. startups that were, at the time, valued at $1 billion or higher and found that more than half, 44 out of 87, were started by immigrants, and immigrants were key members of management or development teams in over 70 percent of these companies. The New York Times in 2017 argued that Silicon Valley simply would not work without immigrants.

Under President Trump, immigration has been discouraged. Of course, the thrust of his policy has been to build a border wall and exclude low-income immigrants from Latin America. I do not support the draconian measures that have been used to exclude this group and a more lenient immigration policy would bolster the low growth rate of the U.S. workforce. Further, immigrants with specialized skills, especially those with science, math, engineering and computer skills add substantially to productivity and create jobs for Americans when they start new companies.

Support public R&D

Federally funded R&D plays an important role in the advancement of knowledge and basic science. It is the major source of support for research in academic institutions. The largest part of federal research funding comes from the military budget (44 percent of the total), while health and retirement account for over half of the remainder. Federal support for R&D rose strongly (in current dollars) from 2000 to 2009, followed by a period of much slower growth and then outright decline. There was a small uptick in funding in 2017, the last year available.The weakness in federal R&D spending has been more than offset by business spending which has risen strongly since 2000 (with modest cyclical downturns) reaching $381 billion in 2017, when it was over three times larger than federal support for R&D. The growth of business R&D helps U.S. long-run growth, but federal R&D support is important because it provides the main financing for basic research and scientific advancements. Private businesses will fund some basic R&D, but they are looking for profits and depend upon universities and research labs to focus on advances that do not yield returns in the short run. Also, the education of scientists and engineers is supported by federal R&D grants. Federal R&D and business R&D serve different goals.

One reason for federal research support is to keep the United States globally competitive. There are four countries or regions that carry out the bulk of global R&D: the United States, China, the EU and Japan. As of 2017, the United States was the largest spender on R&D (business and public combined), but China’s spending has grown much more rapidly and may have overtaken the United States by 2020. The Chinese have thrown money at technology and probably have wasted a portion of that money. Still, the Chinese learn from their mistakes and are becoming a formidable competitor in technology industries. To a degree, advancing scientific and engineering knowledge is a global enterprise, but the United States should remain one of the global leaders.

To support productivity growth and maintain U.S. competitiveness, I propose that the federal government increase its spending on R&D by seven percent a year, the annual growth rate achieved from 2000 to 2009.

Back to top


Conclusion

It appears that effective vaccines are now being produced and, as long as Americans are willing to be vaccinated, I am cautiously optimistic that there will be a solid recovery of growth in 2021, followed by above-average growth for about three years before full-employment is restored. A job training initiative would help in restoring full-employment in the short-run because so many workers have lost jobs and will not get the same jobs back again. Such an initiative can also improve the pace of growth in the long-run. There are many exciting technologies currently being introduced, or that are on the horizon—robots, artificial intelligence and the internet-of-things, for example. A well-trained workforce is required to take advantage of the productivity benefits of these breakthroughs.

Policies of the recent past have restricted the inflow of immigrants with specialized skills and cut back on the funding of public R&D. These policies have been a serious mistake that must be reversed to foster stronger productivity growth in the future.

Back to top

  • Footnotes
    1. For a proposal to restore GDP to its pre-pandemic path by mid-2021 see Wendy Edelberg and Louise Sheiner, What Could Additional fiscal policy do for the economy in the next three years? Brookings, October 9, 2020.
    2. The data in Table 1 are not very sensitive to the exact years chosen. The first slowdown started in the early 1970s, either 1973 or 74. John Fernald has explored the starting date for the second slowdown and he uses statistical tests to determine that 2004 is the best date to use. See John G. Fernald, Productivity and Potential Output before, during, and after the Great Recession, NBER Macroeconomics Annual 29, 1-51.
    3. One suggestive piece of evidence is that there was extensive deregulation in the 1970s and 80s and yet productivity growth remained very sluggish.
    4. Online retail at the end of 2019 was 11.3 percent of the total according to Statista. It had jumped to 16.1 percent by the second quarter of 2020, but of course other factors are impacting productivity in the pandemic. Online retail by itself is very productive, but there are delivery costs involved in delivering each package separately.
    5. Joseph A. Schumpeter, Capitalism, Socialism and Democracy, London, Routledge, 1942.
    6. US Economics Analyst, A Productivity Boost from Creative Destruction and Cost Savings (Hill), Goldman Sachs 9/28/2020 (pay wall)
    7. Schumpeter is not a good guide to the pandemic economy because it is not innovation that has created the destruction of many firms. Almost no one predicted what the pandemic would do, and no firms were prepared for how it would affect them. Firms that were troubled before the pandemic are more likely to fail because they lack a profit cushion, but there is randomness to the impact of the pandemic that is uncorrelated to the productivity and efficiency of firms.
    8. Bureau of Labor Statistics, Multifactor Productivity Trends in Manufacturing – 2019, US Department of Labor, USDL-20-2138.
    9. The data release showed declines in multifactor productivity growth by industry. This concept adjusts labor productivity for the impact of capital input.
    10. Gross fixed nonresidential investment as a share of GDP from the Bureau of Economic Analysis, reported on FRED, BEA code A008RE.
    11. Male labor force participation was 86 percent in 1950 and 69 percent at the end of 2019 – US Bureau of Labor Statistics, accessed through FRED at https://fred.stlouisfed.org/series/LNS11300001
    12. Minority parents are right to be concerned if teachers push bright students towards vocational programs when they should be on an academic track.
    13. For a discussion of labor market programs in Europe see Martin Neil Baily and Jacob Kirkegaard, Transforming the European Economy, Peterson Institute, 2004.
    14. This and the next two paragraphs draw on Greenwood et al. (2020). Their results use Compustat data for public companies.  For small businesses data are taken from Census Bureau Data (The Pulse Survey). Robin Greenwood, Ben Iverson, and David Thesmar, Sizing Up Corporate Restructuring in the COVID Crisis, paper presented at the Brookings Papers on Economic Activity conference, September 24, 2020.
    15. Samuel G. Hanson, Jeremy C. Stein, Adi Sunderam and Eric Zwick, Business Continuity Insurance: Keeping America’s Lights on During the Pandemic, Working Paper, April 8, 2020, https://www.igmchicago.org/wp-content/uploads/2020/04/Business-Continuity-Insurance-20200408-FINAL.pdf; Brunnermeier and Krishnamurthy (2020), Greenwood, Iverson, and Thesmar (2020).
    16. National Science Board, The State of U.S. Science and Engineering 2020, Figure 17.