National Academies Press: OpenBook

Innovation, Global Value Chains, and Globalization Measurement: Proceedings of a Workshop (2022)

Chapter: 13 Colocation of Production and Innovation: Evidence from the United States

« Previous: 12 Capital Flows in Global Value Chains
Suggested Citation:"13 Colocation of Production and Innovation: Evidence from the United States." National Academies of Sciences, Engineering, and Medicine. 2022. Innovation, Global Value Chains, and Globalization Measurement: Proceedings of a Workshop. Washington, DC: The National Academies Press. doi: 10.17226/26477.
×

13

Colocation of Production and Innovation: Evidence from the United States

Paper Authors: Teresa C. Fort (Dartmouth College), Wolfgang Keller (University of Colorado Boulder, National Bureau of Economic Research [NBER], and Centre for Economic Policy and Research [CEPR]), Peter K. Schott (Yale School of Management, CEPR, and NBER), Stephen Yeaple (The Pennsylvania State University and NBER), and Nikolas Zolas (U.S. Census Bureau)

Presenter: Teresa C. Fort (Dartmouth College)
Moderator: Justin Pierce (Federal Reserve Board)

The workshop paper by Teresa Fort, Wolfgang Keller, Peter Schott, Stephen Yeaple, and Nikolas Zolas studies how firms colocate production and innovation in the context of the United States.

Teresa C. Fort, associate professor of business administration at the Tuck School of Business at Dartmouth College, introduced the workshop paper by highlighting policy questions about the potential impact of the decline in manufacturing employment in the United States on U.S. innovation. There is concern that the manufacturing loss in the United States will reduce U.S. innovation, either because of increased competition from imports and the resulting decline in domestic research and development (R&D), or because of complementarities in the colocation of innovation and production. Bloom and colleagues (2020) show that innovative efficiency is falling and that maintaining similar levels of innovation now requires more workers than in the past.

Fort described the three main questions the paper is intended to address:

  • How have the characteristics of innovating firms evolved with time?
  • Is R&D effective only if it is colocated with production?
  • What are the mechanisms behind the relationship between production and innovation?

Manufacturing declines in the United States may affect domestic innovation in one of two ways:

Suggested Citation:"13 Colocation of Production and Innovation: Evidence from the United States." National Academies of Sciences, Engineering, and Medicine. 2022. Innovation, Global Value Chains, and Globalization Measurement: Proceedings of a Workshop. Washington, DC: The National Academies Press. doi: 10.17226/26477.
×
  • Complementarities from the colocation of production and R&D may exist that allow innovation to occur more efficiently when located near the production site, meaning declines in domestic manufacturing decrease innovative activity in the United States.
  • Alternatively, innovative activity may benefit from reallocation and specialization due to comparative advantage effects, which would benefit the United States. The problem of colocation has two dimensions: geographic and firm borders, and their interaction.

CONTRIBUTION

Fort explained that the paper’s first contribution is a set of stylized facts on the trends of innovating firms in the United States from 1977 to 2016. Patenting activity moves from manufacturing firms (MFs) to nonmanufacturing firms (NMFs). An MF is defined as a firm that has at least one manufacturing establishment in a given year in the United States. MFs account for a vast majority of patenting activity early in the sample, decreasing to about half at the end of the sample. The authors focused on a subset of NMFs, those that were once manufacturers but are no longer, which they called former manufacturing firms (FMFs). FMFs that transitioned from MF to NMF later in the sample continue to innovate, which suggests that colocation is not as important for this subset.

As Fort described, firms that contain both manufacturing (M) and innovation (P) establishments patent more relative to other firms. The distance between M and P establishments within a firm tends to increase over time, but some remain close. The firms that locate M and P establishments within 5 miles of each other are found to patent about 12 percent more.

Fort explained that she and her coauthors plan to analyze where innovative activity occurs within firms and understand the margins that alter the colocation decisions of firms. In addition, they plan to causally identify and understand the relationship of innovation and distance across firms.

DATA

Fort described the U.S. Census Bureau datasets used for this study:

  • The Longitudinal Business Database (LBD), which contains all private, nonfarm establishments from 1977 to 2016. Establishment-level North American Industry Classification System [NAICS] classifications and establishment geocodes from the Business Register (from Fort and Klimet, 2018) were added for all establishments in the dataset.
  • The economic censuses of manufacturing wholesale, retail, and services firms provided establishment-level sales and inputs data at 5-year increments.
Suggested Citation:"13 Colocation of Production and Innovation: Evidence from the United States." National Academies of Sciences, Engineering, and Medicine. 2022. Innovation, Global Value Chains, and Globalization Measurement: Proceedings of a Workshop. Washington, DC: The National Academies Press. doi: 10.17226/26477.
×
  • The Longitudinal Foreign Trade Transactions Database provided firm-level import and export transactions from 1992 to 2016.
  • Two R&D surveys provided information on innovative activity of firms from 1977 to 2016.
    • Survey on Industrial Research and Development (1977–2007)
    • Business R&D and Innovation Survey (2008–2016)
  • The U.S. Patent and Trademark Office (USPTO) database provided the names and addresses of firms matched to the LBD and identified manufacturing and processing patents from 1973 to 2018, an extension of the prior matched database.

PORTRAIT OF U.S. INNOVATION

Fort showed that innovation activity in the United States, measured by patents granted and R&D expenditures, has grown in the 40-year span of 1977 to 2017 (see Figure 13-1). She and her coauthors found a decline in activity after 2015, which is an artifact of the application-to-grant lag.

Fort then explained that she and her coauthors categorized firms into three mutually exclusive types for each year: MFs, at least one manufacturing plant in year t; NMFs, no manufacturing plants in year t; and FMFs, a firm that was an MF prior to year t. NMFs dominated the distribution of firms both in the number of firms and number of employees. The number of MFs was also flat across the sample. Employment in NMFs grew from 1977 to 2017, while employment in MFs, including both manufacturing and nonmanufacturing employees, was relatively flat. The share of employment in MFs classified as manufacturing decreased over the sample. FMFs grew in number and employment over the sample.

The share of innovation by MFs declined from 91 percent in 1977 to 54 percent in 2016, but the MFs were still innovating at nearly double the rate in

Image
FIGURE 13-1 U.S. innovation growth over the last 40 years.
NOTES: Granted patents are assigned to application year. R&D = research and development.
SOURCE: Presentation by Teresa Fort.
Suggested Citation:"13 Colocation of Production and Innovation: Evidence from the United States." National Academies of Sciences, Engineering, and Medicine. 2022. Innovation, Global Value Chains, and Globalization Measurement: Proceedings of a Workshop. Washington, DC: The National Academies Press. doi: 10.17226/26477.
×

2016 as they did in 1977 (see Figure 13-2). In 2016, NMFs accounted for 28 percent of patents and FMFs accounted for 18 percent of patents. This implies that a sizable number of firms are capable of innovating without operating a manufacturing establishment. R&D expenditures show similar trends.

Fort said that permanent FMFs, or firms that exit manufacturing permanently, were split into five cohorts: 1978–1996, 1997–2001, 2002–2006, 2007–2011, and 2012–2016 (see Figure 13-3). Firms in the 2007–2011 cohort showed the strongest growth in terms of innovative activity, whereas the 2012–2016 cohort displayed the second-highest innovative activity, although this growth was muted relative to the 2007–2011 cohort. The 2007–2011 cohort also dramatically increased imports from China over the sample, in absolute terms and relative to the other cohorts. This suggests that these firms, predominantly in computer and electronics manufacturing, likely moved production offshore while continuing to innovate in the United States.

The main takeaways, according to Fort, are that NMFs accounted for an increasing share of patents and R&D expenditures over the sample; FMFs continue to patent, especially firms that recently exited manufacturing; and finally, FMFs that exited manufacturing more recently show a large increase in imports from China, indicating offshoring of production.

MEASUREMENT OF COLOCATION

Fort explained that she and her coauthors developed new measures for colocation of U.S. production and innovation. They identified establishments within a firm that focus on innovation (P establishments) from the following

Image
FIGURE 13-2 Manufacturing firms dominate U.S. innovation.
NOTES: MF patent shares declined from 91 percent to 54 percent in 2016. NMFs accounted for 28 percent of patents in 2016. FMFs accounted for 18 percent of patents in 2016. FMF = former manufacturing firm; MF = manufacturing firm; NMF = nonmanufacturing firm; R&D = research and development.
SOURCE: Presentation by Teresa Fort.
Suggested Citation:"13 Colocation of Production and Innovation: Evidence from the United States." National Academies of Sciences, Engineering, and Medicine. 2022. Innovation, Global Value Chains, and Globalization Measurement: Proceedings of a Workshop. Washington, DC: The National Academies Press. doi: 10.17226/26477.
×
Image
FIGURE 13-3 Patents by permanent FMFs differ by cohort.
NOTES: Firms that exited manufacturing from 2007 to 2011 exhibited the strongest growth. The 2007–2011 cohort also exhibited dramatic growth in Chinese imports. Employment dynamics were similar across cohorts.
SOURCE: Presentation by Teresa Fort.

NAICS codes: establishments that focus on scientific R&D services (NAICS 5417); professional scientific and technical services (NAICS 5413–5416); corporate, subsidiary, and regional managing offices (NAICS 551114); and information and telecommunications (NAICS 5112, 517, 518).

Fort and her coauthors found that firms with both M and P establishments patent 65 percentage points more than firms that lack M or P establishments. When including firm fixed effects—therefore looking at within-firm effects—firms with both M and P establishments patent 15 percentage points more than firms that lack M or P establishments. The authors removed variation from time-varying firm size, age, and stocks of patents.

To measure the distance between manufacturing and innovating establishments, Fort and her coauthors focused on firms with both M and P establishments (MP firms). The average and minimum distances between M and P plants was calculated. According to Fort, the median of the average distance between M and P firms was hundreds of miles across the sample, and this distance grew over time from 301 miles in 1977 to 416 miles in 2012. This may suggest that there is no colocation in the United States; however, the median firm has at least one pair of M and P establishments that are colocated within 6 miles of each other.

RESULTS

Fort and her coauthors estimated a relationship in which the dependent variable is the inverse hyperbolic sine transformation of the sum of subsequently granted patents (or citations) applied for by firms in year t to t+4. This specification allows the results to have an elasticity interpretation. The explanatory variables are the minimum distance, split into two subsets: 0–5 miles and 5–60 miles. The depreciated and 1-year lagged patent stock and a set of controls that account for time-varying differences in firm size and age, along with

Suggested Citation:"13 Colocation of Production and Innovation: Evidence from the United States." National Academies of Sciences, Engineering, and Medicine. 2022. Innovation, Global Value Chains, and Globalization Measurement: Proceedings of a Workshop. Washington, DC: The National Academies Press. doi: 10.17226/26477.
×

year, county, and firm fixed effects are used to remove unwanted variation in the empirical results.

Across all outcome variables of interest (patents, citations, manufacturing patents, and process patents), firms with at least one set of M and P plants within 5 miles of each other innovated more. This relationship also held for firms with a set of M and P plants between 5 and 60 miles of each other. However, the coefficient on patents, which is the measure of innovation, declined from 11.6 percentage points to 7.6 percentage points when the colocation distance increased. According to Fort, this suggests that colocation does matter. At the same time, she explained that the mechanisms affecting innovation are not clearly understood when there is colocation of production and innovation, and there may be something special about the management of firms that have both manufacturing and innovation establishments.

FUTURE WORK

Fort then discussed future extensions of her and her coauthors’ work. They would like to see whether patent creation occurs in a firm’s colocated region; in other words, does innovation at the P plant influence behavior at the nearby M plant? Or is innovation influenced by behavior at another M plant that is not nearby? They will be looking at patenting behavior in a region to understand the relationship between innovation and manufacturing, and to identify any spillover effects from other firms’ M and P plants in a region.

According to Fort, she and her coauthors would also like to understand policy-induced changes to the cost of R&D, such as how a measure of R&D tax credits interacts with manufacturing and innovation establishments.

SUMMARY

Fort concluded by summarizing the new stylized facts from her and her coauthors’ work:

  • The share of aggregate patents by NMFs increased from 9 percent in 1977 to 46 percent in 2016. This suggests that manufacturing is not essential for continued innovation, even in manufacturing-related activities.
  • Firms with both manufacturing and innovation establishments innovate more. The spatial distribution of firms is broad across the establishments, but firms’ patents are higher when the establishments are colocated. In future work, they will explore the mechanisms that drive this pattern, particularly around the influence of the innovation establishment on the behavior of local manufacturing and the productivity of innovation in colocated regions.
Suggested Citation:"13 Colocation of Production and Innovation: Evidence from the United States." National Academies of Sciences, Engineering, and Medicine. 2022. Innovation, Global Value Chains, and Globalization Measurement: Proceedings of a Workshop. Washington, DC: The National Academies Press. doi: 10.17226/26477.
×

DISCUSSION

Discussant: Nick Bloom (Stanford University)

Nick Bloom, William D. Eberle professor of economics at Stanford University, began by placing the paper into two important trends in the U.S. economy: productivity growth is slowing and manufacturing share of employment is falling. Productivity growth is down across all countries in the Organisation for Economic Co-operation and Development—essentially across all developed countries. While manufacturing output is flat, the manufacturing employment share of total employment has been consistently decreasing since 1945. Historically, innovation was concentrated in the manufacturing sectors, so one may think that these trends are connected; the authors sought to understand if this is the case. These questions are relevant not only to the academic audience, but also to policy makers, financial markets, and the general public.

Bloom offered a summary of the work, starting with a brief discussion of the data. The authors combined several U.S. Census datasets with the novel contribution of patent and geocode data. Bloom emphasized the amount of work that went into this process and stated that this work is relevant to policy questions currently being asked about innovation, such as those related to global value chains for solar panels, with respect to China and the United States. Additionally, he noted that the Biden administration is pushing to shift manufacturing back to the United States because it believes that it is important for innovation.

The paper offers four key findings:

  • First, the stylized fact that NMFs account for 46 percent of total U.S. patents in 2016, an increase from 9 percent in 1977. The FMFs are a part of this shift of innovation away from manufacturing, which is interesting as these firms continue to locate their headquarters in the United States, while shifting production to foreign countries. This is consistent with the “coastal” story of such firms as Apple, Levi, Hasbro, or Dell—innovating in the United States and producing abroad while achieving high profit rates. The most innovative U.S. firms are primarily technology firms, which suggests that manufacturing is not the only path to innovation.
  • Second, FMFs that shifted production to other countries continue to innovate in the United States.
  • Third, the firms that have both manufacturing and innovation establishments nearby, or colocated, appear to patent much more than firms that do not. Firms that have only manufacturing plants patent 4 percentage points more, firms with only innovative plants patent 2 percentage points more, and firms with both patent 66 percentage points more, suggesting strong colocation effects, the opposite of the coastal story.
Suggested Citation:"13 Colocation of Production and Innovation: Evidence from the United States." National Academies of Sciences, Engineering, and Medicine. 2022. Innovation, Global Value Chains, and Globalization Measurement: Proceedings of a Workshop. Washington, DC: The National Academies Press. doi: 10.17226/26477.
×
  • Finally, firms with manufacturing and innovation plants in close proximity are more likely to patent. If the plants are within 5 miles of one another, then patenting increases by about 11 percentage points. For firms between 5 and 60 miles of each other, the increase is about 7 percentage points, though this measurement is sensitive to the specification.

Bloom suggested that the key question remaining is about the underlying mechanisms at play. Do these results suggest colocation effects (colocation causing more innovation), or are they a consequence of other firm characteristics, such as high-quality management? Identifying the correct mechanism has policy implications. If there are serious colocation effects, then domestic policy makers may have an incentive to introduce tariffs and subsidize manufacturing. If innovation is independent of manufacturing, then policy makers have an incentive to invest in domestic R&D and education. The fact that the Biden administration is currently discussing these policy questions indicates the importance of this paper in nonacademic circles.

Fort’s response focused on two points raised in the discussion:

  • The results regarding FMFs and colocation effects appear to contradict each other. The FMFs continue to patent, but firms with both manufacturing and innovation plants patent much more. One fact that may help reconcile this is that for firms that colocate within 5 miles, manufacturing employment for these collocating firms has fallen precipitously over the sample period.
  • Second, Fort disagreed with the Bloom’s policy suggestion to use tariffs and subsidies to incentivize domestic manufacturing, even if colocation effects are determined to be strongest. Firms that are shedding manufacturing are not targeting colocated manufacturing plants, but rather their noncolocated plants—likely the larger, labor-intensive plants. The plants that benefit from colocation remain. She sees across-firm spillovers as the more important policy dimension.

One audience member asked if the FMFs are altering the type of innovation in which they participate. Fort responded that the authors have looked at this and the firms appear to continue to patent in the same innovation type. Another audience member wondered if firms outsource manufacturing domestically. Fort responded that this intuition is correct. In the 2007 data from the Bureau of Economic Analysis (BEA), there are no firms that produce outside of the United States that do not also produce within the United States. Another audience member wondered if the authors have looked at the variation of geographic region in the data. Are the manufacturing and innovation plants both

Suggested Citation:"13 Colocation of Production and Innovation: Evidence from the United States." National Academies of Sciences, Engineering, and Medicine. 2022. Innovation, Global Value Chains, and Globalization Measurement: Proceedings of a Workshop. Washington, DC: The National Academies Press. doi: 10.17226/26477.
×

on the coast or are they in less attractive locations? Fort responded that this is an interesting idea and something that the authors will investigate.

Another audience member wondered if overseas R&D centers of U.S. multinational enterprises appear in the data and if nonpatenting innovation appears. Fort responded that in their findings, a shift in innovation towards nonmanufacturing firms would be even stronger if they were able to include nonpatenting innovation. Nikolas Zolas, one of the paper’s coauthors, has worked on alternate measures of innovation, such as trademarks, that are more common among the service sectors. In future work, the coauthors may try to include this measure in their analysis, as well as other measures from the BEA censuses. Another audience member wondered what role mergers and acquisitions play in the colocation results. Fort responded that this is a good idea and something that the authors will investigate.

Suggested Citation:"13 Colocation of Production and Innovation: Evidence from the United States." National Academies of Sciences, Engineering, and Medicine. 2022. Innovation, Global Value Chains, and Globalization Measurement: Proceedings of a Workshop. Washington, DC: The National Academies Press. doi: 10.17226/26477.
×

This page intentionally left blank.

Suggested Citation:"13 Colocation of Production and Innovation: Evidence from the United States." National Academies of Sciences, Engineering, and Medicine. 2022. Innovation, Global Value Chains, and Globalization Measurement: Proceedings of a Workshop. Washington, DC: The National Academies Press. doi: 10.17226/26477.
×
Page 99
Suggested Citation:"13 Colocation of Production and Innovation: Evidence from the United States." National Academies of Sciences, Engineering, and Medicine. 2022. Innovation, Global Value Chains, and Globalization Measurement: Proceedings of a Workshop. Washington, DC: The National Academies Press. doi: 10.17226/26477.
×
Page 100
Suggested Citation:"13 Colocation of Production and Innovation: Evidence from the United States." National Academies of Sciences, Engineering, and Medicine. 2022. Innovation, Global Value Chains, and Globalization Measurement: Proceedings of a Workshop. Washington, DC: The National Academies Press. doi: 10.17226/26477.
×
Page 101
Suggested Citation:"13 Colocation of Production and Innovation: Evidence from the United States." National Academies of Sciences, Engineering, and Medicine. 2022. Innovation, Global Value Chains, and Globalization Measurement: Proceedings of a Workshop. Washington, DC: The National Academies Press. doi: 10.17226/26477.
×
Page 102
Suggested Citation:"13 Colocation of Production and Innovation: Evidence from the United States." National Academies of Sciences, Engineering, and Medicine. 2022. Innovation, Global Value Chains, and Globalization Measurement: Proceedings of a Workshop. Washington, DC: The National Academies Press. doi: 10.17226/26477.
×
Page 103
Suggested Citation:"13 Colocation of Production and Innovation: Evidence from the United States." National Academies of Sciences, Engineering, and Medicine. 2022. Innovation, Global Value Chains, and Globalization Measurement: Proceedings of a Workshop. Washington, DC: The National Academies Press. doi: 10.17226/26477.
×
Page 104
Suggested Citation:"13 Colocation of Production and Innovation: Evidence from the United States." National Academies of Sciences, Engineering, and Medicine. 2022. Innovation, Global Value Chains, and Globalization Measurement: Proceedings of a Workshop. Washington, DC: The National Academies Press. doi: 10.17226/26477.
×
Page 105
Suggested Citation:"13 Colocation of Production and Innovation: Evidence from the United States." National Academies of Sciences, Engineering, and Medicine. 2022. Innovation, Global Value Chains, and Globalization Measurement: Proceedings of a Workshop. Washington, DC: The National Academies Press. doi: 10.17226/26477.
×
Page 106
Suggested Citation:"13 Colocation of Production and Innovation: Evidence from the United States." National Academies of Sciences, Engineering, and Medicine. 2022. Innovation, Global Value Chains, and Globalization Measurement: Proceedings of a Workshop. Washington, DC: The National Academies Press. doi: 10.17226/26477.
×
Page 107
Suggested Citation:"13 Colocation of Production and Innovation: Evidence from the United States." National Academies of Sciences, Engineering, and Medicine. 2022. Innovation, Global Value Chains, and Globalization Measurement: Proceedings of a Workshop. Washington, DC: The National Academies Press. doi: 10.17226/26477.
×
Page 108
Next: 14 Global Value Chain Measurement Methodology: Challenges and Prospects »
Innovation, Global Value Chains, and Globalization Measurement: Proceedings of a Workshop Get This Book
×
Buy Paperback | $26.00 Buy Ebook | $20.99
MyNAP members save 10% online.
Login or Register to save!
Download Free PDF

In recent decades, production processes of intermediate and final products have been increasingly fragmented across countries in what are called global value chains (GVCs). GVCs may involve companies in one country outsourcing stages of production to unrelated entities in other countries, multinational enterprises (MNEs) offshoring stages of production to units of the MNE overseas, or both. GVCs can also involve completely independent companies merely sourcing their parts from whichever upstream company may be the most competitive, with no control arrangement necessarily involved. The changing global trade environment and the changes in firms' behavior have raised new and more complicated issues for policy makers and have made it difficult for them to understand the extent and operations of GVCs and their spillover effects on national and local economies.

To improve the understanding, measurement, and valuation of GVCs, the Innovation Policy Forum at the National Academies of Sciences, Engineering, and Medicine convened a workshop, "Innovation, Global Value Chains, and Globalization Measurement" May 5-7, 2021. This proceedings has been prepared by the workshop rapporteurs as a factual summary of what occurred at the workshop.

  1. ×

    Welcome to OpenBook!

    You're looking at OpenBook, NAP.edu's online reading room since 1999. Based on feedback from you, our users, we've made some improvements that make it easier than ever to read thousands of publications on our website.

    Do you want to take a quick tour of the OpenBook's features?

    No Thanks Take a Tour »
  2. ×

    Show this book's table of contents, where you can jump to any chapter by name.

    « Back Next »
  3. ×

    ...or use these buttons to go back to the previous chapter or skip to the next one.

    « Back Next »
  4. ×

    Jump up to the previous page or down to the next one. Also, you can type in a page number and press Enter to go directly to that page in the book.

    « Back Next »
  5. ×

    Switch between the Original Pages, where you can read the report as it appeared in print, and Text Pages for the web version, where you can highlight and search the text.

    « Back Next »
  6. ×

    To search the entire text of this book, type in your search term here and press Enter.

    « Back Next »
  7. ×

    Share a link to this book page on your preferred social network or via email.

    « Back Next »
  8. ×

    View our suggested citation for this chapter.

    « Back Next »
  9. ×

    Ready to take your reading offline? Click here to buy this book in print or download it as a free PDF, if available.

    « Back Next »
Stay Connected!