The U.S. manufacturing sector, particularly the high-technology space, has long been viewed as a center of innovative activity. The sector accounts for a disproportionate share of patenting and R&D spending, and, for decades, its productivity growth outpaced that of the broader economy. Whereas total factor productivity (TFP) — which is a measure describing how efficiently firms combine capital, materials, and labor to produce output — increased by 0.8 percent on average annually between 2009 and 2023 for the private economy as a whole, the manufacturing sector’s TFP declined by 0.1 percent on average per year. What accounts for the manufacturing sector’s recent slow growth?

Enghin Atalay and Nicole Kimmel of the Philadelphia Fed, along with Ali Hortaçsu and Chad Syverson of the University of Chicago, address this question in their paper, “Why Is Manufacturing Productivity Growth So Low?” Atalay and his coauthors tackle the question from two angles. First, taking existing productivity measures at face value, they ask if any industries within the manufacturing sector disproportionately account for the sector’s stagnation. Second, they examine whether official TFP growth statistics capture the underlying growth in manufacturing productivity. Specifically, they compare these statistics with an alternative method that accounts for quality improvements by adjusting for discrepancies across price indexes.1 Understanding manufacturing productivity is of great interest, they write, because gains in manufacturing productivity drive long-term economic growth and improvements in living standards.

The authors begin by showing that, using official statistics from the Bureau of Labor Statistics, the trajectory of the entire manufacturing sector can be linked to a group of industries that manufacture computers and other electronic products. This industry grouping accounted for nearly all the rapid manufacturing TFP growth from 1987 through 2009 as well as the slowdown in manufacturing TFP growth after 2009.2

Do these conventional statistics fully capture manufacturing productivity growth? To investigate, the authors look to industry deflators, a key ingredient in measuring productivity: The deflators convert nominal output into real output by accounting for changes in prices. But when products improve over time, the deflators must also account for changes in quality. If a smartphone’s sticker price rises by 5 percent from one year to the next, but its camera resolution, memory, and other features also improve, how much has its quality-adjusted price really increased? Conventional industry deflators used to measure productivity rely on producer-facing price indexes. But consumer-facing indexes, the authors argue, may more fully capture quality improvements for manufactured goods.

Overall, consumer-facing price indexes indicate substantially lower inflation than producer- and import-facing measures. These differences appear within manufacturing, not in services, and are concentrated in durable goods industries, especially those experiencing the largest quality improvements, like computer and electronic product manufacturing. When quality improvements are understated, real output growth is also understated, which means manufacturing productivity growth appears slower than it actually is.

The authors find that TFP growth for durable manufactured goods was understated by a sizable 1.4 percentage points annually between 1997 and 2023. TFP growth for nondurable manufactured goods — which include household supplies, office supplies, clothing, and food and beverages — was understated by 0.35 percentage point annually. Conversely, they find that TFP growth outside of the manufacturing sector — which includes finance, health, professional services, and transportation — was slightly overstated.

After correcting for TFP mismeasurement, the authors conclude that U.S. manufacturing productivity has continued to grow since the late 2000s, although at a slower pace than in previous decades. They estimate that manufacturing TFP growth averaged 0.5 percent annually between 2009 and 2023, with 1.6 percent growth in durable manufacturing and no change in nondurable manufacturing. This compares with manufacturing TFP growth of 2.0 percent annually between 1997 and 2009. Productivity mismeasurement was greatest in computer and electronic products, which comprise the same industries that tend to experience the largest product quality adjustments. Still, mismeasurement, they write, “is pervasive throughout manufacturing.”

The paper does not overturn the basic fact that manufacturing productivity growth has slowed. The earlier boom in computer and electronic product manufacturing has faded, and that industry grouping no longer pulls up aggregate manufacturing productivity as it once did. But Atalay and his coauthors show that, once unaccounted-for quality improvements are incorporated, manufacturing appears less stagnant than conventional measures suggest.

The manufacturing sector has played a critical role in the economy: Not only does it account for a disproportionate share of corporate patents and R&D spending, but it also plays a pivotal role in national security, in global trade, and in generating high-quality jobs for people without a college degree. Given this key role, policymakers have long sought to support innovation and investment in manufacturing.3 Better measurement may change not only how economists understand past productivity growth but also how to evaluate past governmental efforts to support the manufacturing sector.

  1. The views expressed here are solely those of the author and do not necessarily reflect the views of the Federal Reserve Bank of Philadelphia or the Federal Reserve System.
  2. The authors focus on TFP rather than labor productivity or other productivity measures, writing that “this choice is motivated by the fact that an industry’s TFP is more closely linked to its marginal cost of production and, as a result, its output price.”
  3. The share of computer and electronic product manufacturing in total manufacturing rose from 9.4 percent in 1987 to 12.3 percent in 2000 but has since fallen to 5.5 percent as of 2023. Thus, some of the documented slowdown in manufacturing productivity is likely attributable to the declining share of this high-productivity-growth industry in total manufacturing output.
  4. These programs include SEMATECH (established in 1987), the American Recovery and Reinvestment Act of 2009, and the CHIPS and Science Act of 2022.