A new study by economists at Princeton University and the Federal Reserve Bank of Richmond finds that while national market concentration is clearly increasing, local concentration is not. In fact, local concentration has been declining rapidly for the last 25 years when measured at the city, county, or ZIP code level.
[Note: The views expressed herein are those of the authors and do not necessarily represent the views of the Federal Reserve Bank of Richmond or the Federal Reserve System.]
Many product markets are local. Think about how an establishment entry in L.A. affects the choices of someone in New York. Clearly, for most retail services, a new store in L.A. has no impact whatsoever. It clearly does not affect the available retail choices of New York residents, nor does it affect the competition between stores to attract new patrons. The reason is obvious: For most of what we buy, we simply go to a local store. Traveling somewhere else to buy it is way too costly. Thinking about it almost sounds ridiculous.
Not all industries are like this. In some manufacturing industries (for example, furniture producers or producers of tobacco products), there are only a few establishments that sell their products everywhere. Retailers of these products might still be local, but the manufacturing plants themselves have a national market. It may conform to basic intuition that many industries face national markets. Isn’t this the epoch of the death of distance after all? No. Most employment and sales are in industries where consumers travel little to buy the good or service from the producer. Manufacturing is small anyway, but even within manufacturing there are industries, like the ready-mixed concrete industry, that are extremely local.
The US economy has been concentrating at the national level for the past 25 years. This means that the top producers in an industry have become bigger. Think about Walmart and their increasing share of the discount retail service industry, or Starbucks and their increasing share of coffee shops. Examples abound. These trends are uncontroversial and have been measured with detailed data many times. But does such concentration mean that competition in these industries has decreased? Do the top firms exploit their large shares to exert their market power on consumers via higher prices, lower quality, or other characteristics of the product they sell?
In a recent paper (Rossi-Hansberg, Trachter and Sarte 2018), we use detailed data on firms and their establishments for most industries in the United States for the last 25 years and show that, although national market concentration is clearly increasing, local concentration is not. In fact, local concentration has been declining rapidly for the last 25 years when we measure concentration at the city, county, or ZIP code level. Figure 1 illustrates these results for counties when we use different levels of industry aggregation to measure national and county level concentration. As the figure illustrates, independently of whether we measure concentration for very specific industries, or industries that include broad classes of products, the declining trend in local concentration is evident.
To understand these diverging trends in national and local concentration, it is useful to explore the role that the market leaders in each industry, the big guys, have played in those trends. Clearly, market leaders are responsible for the increase in national concentration. This is the standard story. Perhaps surprisingly at first glance, though, top firms have also been responsible for the decline in local concentration. If we eliminate from the calculation the establishments of the top producer in an industry, local concentration falls less rapidly. How can this be? You probably have guessed it already. All of us have experienced this in our neighborhoods. Large firms have increased their national shares by opening establishments in new locations. These new establishments now compete with whoever was offering the local good or service. As a result, concentration falls.
Of course, it still could be the case that, as large firms progressively open more establishments, they eliminate all the local competition, thereby increasing local concentration. This increase does happen in some industries, but in most of them (and on average in the US economy) it does not. In the industries where both national and local concentration are increasing, we see exactly this phenomenon. Four years after entry of a top firm establishment, local concentration is above the level of local concentration before the top firm establishment entry. However, these industries amount to less than 20 percent of employment and sales. In contrast, in industries with diverging trends, which account for about half of the US economy, entry of a top-firm establishment decreases concentration on impact, and afterwards concentration remains at the lower level for at least the next seven years. This is what explains the diverging trends.
Two more comments are in order. First, given these results, it is hard to imagine that top firms in industries with diverging trends are exerting increasing product-market power. It is the nature of their growth, through geographic expansion, that imposes the need to compete at the local level. There are no mechanisms that link decreasing concentration at the relevant local market level with increasing market power. None. Of course, this does not mean that firms cannot exploit their large national market shares at all. They can exploit them by exerting monopsony power in inputs markets, or by capturing political institutions to change the rules in their favor or to get big bailouts when times are tough.
Second, we use a somewhat non-standard dataset to make our claims. It is the NETS dataset, which is privately collected, rather than public data collected by the Census. Barnatchez, Crane and Decker (2017) has compared these datasets and concludes that “at the narrow cell level (geography by industry by establishment size), we find correlations between NETS and official sources that are reassuringly strong.” We also find very large correlations across datasets when we compare employment counts across locations. In fact, county-level correlations of employment are consistently around 0.99 throughout the 25-year period. So we are confident that the trends we find are not an artifact of the dataset used.
Big guys can be bullies, but not in local product markets.
Esteban Rossi-Hansberg is a professor or economics at Princeton University. Pierre-Daniel Sarte is a senior advisor in the research department of the Federal Reserve Bank of Richmond. Nicholas Trachter is a senior economist in the research department of the Federal Reserve Bank of Richmond.
Barnatchez, K., Crane, L. and Decker, R. (2017), “An Assessment of the National Establishment Time Series (NETS) Database”, Board of Governors of the Federal Reserve System (U.S.), Finance and Economics Discussion Series 2017-110.
Rossi-Hansberg, E., Sarte, P. D. and Trachter, N. (2018), “Diverging Trends in National and Local Concentration”, NBER, Working paper No. 25066.
Disclaimer: The ProMarket blog is dedicated to discussing how competition tends to be subverted by special interests. The posts represent the opinions of their writers, not necessarily those of the University of Chicago, the Booth School of Business, or its faculty. For more information, please visit ProMarket Blog Policy.