Skip Header

Research on Plant Dynamics in the Manufacturing Sector

Mon Oct 03 2016
Lucia Foster and Scott Ohlmacher, Center for Economic Studies
Component ID: #ti2133036147

Why do some manufacturing plants grow and thrive while others falter? Economists using U.S. Census Bureau plant-level microdata have approached this complex question from at least three different angles. First, they looked at microeconomic patterns at the plant level. Second, they examined the growth, survival and exit of manufacturing plants throughout the business cycle. Finally, they documented the long-term, secular trends in manufacturing.

Component ID: #ti291897879

Each of these is the subject of recent papers by researchers using Census Bureau microdata from the Annual Survey of Manufactures (ASM), the Census of Manufactures (CM), and the Longitudinal Business Database (LBD). While this blog highlights these papers, extensive literature on this subject using Census Bureau microdata exists (many of these papers can be found in the CES Working Paper series.)

Turning first to the microeconomic patterns, the growth, survival and exit of manufacturing plants depends upon their profitability. Profitability is influenced by many factors internal and external to the plant. One important component of this profitability is the plant’s productivity. Empirical evidence using Census Bureau plant-level data reveals that there are differences in productivity in manufacturing plants even within the same narrowly defined industries. Differences in location and production technology are two possible reasons productivity at manufacturing plants vary, even within the same industry.

The Census Bureau collects information on plant characteristics through the ASM and the CM. However, even when we set controls for relevant characteristics, important differences in productivity remain. According to the productivity literature, Syverson (2011) [PDF] notes that when using the same measured inputs, a manufacturing plant at the upper end of productivity distribution is able to produce almost twice as much output as a manufacturing plant in the same industry at the lower end of productivity distribution. In discussing possible reasons for these differences, Syverson comments that managers have long been thought to be an important factor, but without data, their importance has been speculative.

The Management and Organizational Practices Survey (MOPS), a supplement to the ASM, is intended to partly fill the data gap by collecting information on these practices. Evidence from the MOPS suggests that management practices are correlated with productivity in manufacturing plants. Bloom et al. (2013) [PDF] find that “structured” management practices related to monitoring, targeting and incentives are tightly linked to better performance (including higher productivity). These “structured” practices include monitoring a large number of high-frequency key performance indicators (KPIs), setting realistic production targets, making sure that all levels of the organization at the plant are aware of KPIs and targets, and setting bonus, promotion and dismissal incentives based on those targets. While structured management practices are associated with positive outcomes, many plants do not adopt these practices. Researchers are now looking into why there are differences in management practices at plants even within the same firm.

Brynjolfsson and McEhleran (2016) [PDF] find that the adoption of intensive data-driven decision-making and an increased allocation of decision-making to front-line production workers (versus manager-centric decision-making) is associated with large gains in productivity for plants in industries that are generally capital intensive and utilize “continuous-flow” operations.

The research above focuses on the supply side of profitability, but the demand side is also important. The challenge here is that the Census Bureau does not collect microlevel information on prices. However, researchers have been able to create proxies for the demand side in a limited sample of manufacturing plants for which the Census Bureau collects both revenue and physical output. Using this sample, Foster, Haltiwanger and Syverson (2016) show that much of the growth of plants is dependent on the demand side. They find that new manufacturing plants have higher physical productivity and lower revenue productivity compared to their more mature counterparts, reflecting that new plants set prices low in order to build up their market and grow.

In terms of business cycles, Foster, Grim and Haltiwanger (2016) examine the growth, survival and exit dynamics of manufacturing plants during recent cycles. Regardless of the overall economic conditions, it is generally true that plants that are more productive grow and thrive, while lower productivity plants shrink and exit. In most recent business cycle downturns, this process of reallocation from less productive plants to more productive plants is accelerated. However, they find that in the Great Recession, this reallocation of economic activity from least productive to more productive weakened relative to other downturns. Since this was especially pronounced for young plants, they hypothesize that credit constraints impacted the reallocation. Researchers are also using the MOPS to look at the management and organizational characteristics of manufacturing plants that are able to better weather business cycles.

Finally, researchers have used Census Bureau microdata to better understand long-term trends in manufacturing. Using plant-level data is critical to understanding these trends due to changes in industry classification schemes. Without controlling for these changes, it is unclear what is due to changes in underlying economics activity at plants versus changes in classification.

Pierce and Schott (2016) focus on the decline in U.S. manufacturing employment from 2000 to 2007. They examine the link of this decline to China’s accession to the World Trade Organization (WTO) in 2001. By using the Longitudinal Business Database (a Census Bureau research dataset) and the CM, they examine the response of manufacturing plants while controlling for changes in classification. In addition to changes at the external margin (including within firm relocation of production outside the United States), they find evidence of capital deepening of U.S. manufacturing plants that continued in operation during this period.

One topic that figures in Pierce and Schott’s work is the impact of uncertainty on manufacturing plant’s decisions. They cite anecdotal evidence that uncertainty concerning China’s trade status leading up to its accession to WTO impacted manufacturing plant’s planning decisions. The second wave of the MOPS, which is currently in collection, includes a section on uncertainty. We look forward to research using the MOPS that will enable us to better understand the impact of uncertainty on manufacturing plants’ growth, exit and survival.

  Is this page helpful?
Thumbs Up Image Yes    Thumbs Down Image No
Comments or suggestions?
No, thanks
255 characters remaining
Thank you for your feedback.
Comments or suggestions?
Back to Header