U.S. industries are more highly concentrated now than ever before, although market concentration at the product level has been decreasing over time in the most highly concentrated industries, according to new research.
Using a compilation of respondent-level data from the annual "Survey of the American Consumer" by market research firm MRI-Simmons and data from Kantar Media about advertising expenditures, researchers in a working paper, published May 3 by the National Bureau of Economic Research, explored trends in U.S. market concentration between 1994 and 2019. Of the industries included in the data sample, 42.2% were classified as "highly concentrated" according to U.S. Horizontal Merger Guidelines, which is much higher than previous results.
The researchers also discovered, however, that contrary to established production data indicators, product-market concentration has been decreasing over time, especially in highly concentrated industries, as companies produce a wider variety of products.
According to C. Lanier Benkard, co-author of the study and a professor of economics at the Stanford Graduate School of Business, the topic of market concentration has been central to academic and policy roundtables, with academia warning that concentration can lead to negative macroeconomic consequences such as declines in investment, productivity and labor as a share of national income. Joseph E. Stiglitz, in a January 2019 article for Columbia Business School, referred to the growing concentration of market power as a threat to the economy and a contributing factor to inequality, which also "allows dominant firms to exploit their customers and squeeze their employees, whose own bargaining power and legal protections are being weakened."
"Many people (academics, politicians and the press) have claimed that U.S. antitrust policy is failing and that this failure is negatively affecting the macroeconomy," Benkard told The Academic Times. "We were motivated by the belief that concentration had been measured incorrectly, and we were curious to see what trends in concentration would look like if they were measured correctly. We also had an inkling that the answer would be different, though it turned out we were wrong about how."
Benkard said he and his colleagues anticipated the main difference between their research and existing work that used U.S. Census Bureau data would be driven by local versus national differences, with similar outcomes observed across various industries. Instead, they discovered that national and local trends were approximately the same irrespective of industry, with a more influential factor being differences in aggregation. At the more precise, product-market level, researchers found results that contradicted existing aggregations of census data, suggesting that concentration has been decreasing over time rather than increasing.
Benkard says these differences could be a result of companies producing a wider variety of products than they ever have before.
"Large companies like Johnson & Johnson and Procter & Gamble have grown larger mainly by producing more different products, not by growing in their existing product markets," he said. "At higher levels of aggregation, this looks like increased concentration, but within individual product markets such as soap, toothpaste or Band-Aids, there is now more competition."
Although it is likely that multiple forces are at work, an interaction between product markets might be able to explain both of these trends: If companies that are most successful in single-product markets enter each others' "home" product markets, this will reduce market concentration at the product level and increase industry concentration. According to a January 2021 working paper published by the U.S. Bureau of Labor Statistics, the cross-market component can explain 99% of a national increase in concentration, reflecting the expansion of multi-market firms, despite local concentration also increasing by 34% from 1992 to 2012.
Over 80% of brands across most sectors included in the study were identified as belonging to a larger company. Ownership of brands in the data set varied a great deal, with the largest owner possessing 253 brands, although 76.3% of brand owners were associated with only one. Some of the largest brand owners in the manufacturing sector were Procter & Gamble, Kraft Heinz, Unilever, Johnson & Johnson and Clorox. The largest nonmanufacturing brand owners included Visa, State Farm and Blue Shield.
This study is not the first to speak to the importance of increasing market concentration in the U.S. In 2019, for example, a study published in the Review of Finance found that firms in industries with the largest increases in product-market concentration had higher profit margins and more profitable merger-and-acquisition deals, but did not have significantly increased operational efficiency, indicating that, "Market power is becoming an important source of value."
Where Benkard and his colleagues diverge from prior literature is their attempt to measure product-market concentration on an economy-wide scale, which Benkard says has been difficult due to a lack of necessary data in the past. The researchers for this working paper relied on a consumer survey to fill in those data gaps, but Benkard cautioned that the use of survey responses for the study results in several potential weaknesses. For one thing, consumers typically fill out market research surveys from memory, potentially affecting the reliability of results.
"We checked the data against other sources for a few industries and found that it matched quite well, but one might still worry about how accurate the data are for the industries that we could not cross-check," Benkard said. "The second weakness is that, because it is a consumer survey, it covers only products consumed by consumers. It does not cover 'intermediate goods,' which are mainly used by businesses as inputs to producing other products."
The working paper, "Concentration in product markets," published May 3 by the National Bureau of Economic Research, was authored by C. Lanier Benkard and Ali Yurukoglu, Stanford Graduate School of Business and National Bureau of Economic Research; and Anthony Lee Zhang, University of Chicago.