Concentrating on Market Concentration

"De Loecker and Eeckhout and other papers were arguing that market power was going up," says Nicholas Trachter, a senior economist at the Richmond Fed. "But on the other hand, many prices didn't seem to be getting higher, and firms were claiming that it was very hard for them to raise their prices because of competition. So I began trying to understand how to connect these two stories."

Along with his colleague at the Richmond Fed, Pierre-Daniel Sarte, and Esteban Rossi-Hansberg of Princeton University, Trachter began thinking about how large national firms expand into local markets. In many sectors of the economy, such as retail and services, firms compete locally rather than nationally. For example, a coffee shop in Richmond doesn't compete with a coffee shop in Seattle for customers. Likewise, restaurants in Manhattan don't compete with restaurants in San Francisco.

Using the National Establishment Time Series, they were able to study the concentration of industries at the national and local levels. They found that for many sectors of the U.S. economy, concentration was rising at the national level but was actually falling locally.

"At first we thought we might have made a mistake," says Sarte. But through a series of tests, they confirmed that the results were accurate. "Once we saw that, we thought, how can we reconcile these two trends of rising national concentration and falling local concentration?"

In a paper published in the Richmond Fed's working paper series as well as in the National Bureau of Economic Research's working paper series, they explained how both trends were being driven by large national companies. When a national chain such as Walmart or Starbucks opens a new store, the chain increases its share of the national market, which, in turn, increases national concentration. But they typically don't open new stores where one already exists. Instead, they expand into new locations, where other firms are already operating. Rather than drive those firms out of the local market, Rossi-Hansberg, Sarte, and Trachter found that large firms decreased local concentration when they opened a new store.

"So large firms grow at the national level by expanding geographically, but at the local level there is more competition because there are now more firms in the local market," explains Trachter.

To the extent that consumer markets are local, then, these findings suggest that competition may be increasing rather than decreasing. To be sure, not all industries are local. The manufacturing sector, for example, benefits from economies of scale and easy transportation of its products that make it more of a national sector than a local one. But industries that do exhibit diverging trends of national and local concentration employ roughly three-quarters of U.S. workers and account for two-thirds of all sales, making local competition important for a large segment of the economy.

"What we found is that the world is a lot more subtle than one might have been led to believe just based on the aggregate concentration data," says Sarte. "Even before we did this study, some people had a feeling that we shouldn't conclude that the U.S. economy has become less competitive just because of what we see at the national level. And I think we showed that's right."

Of course, concentration is just one possible sign of market power, and economists have been exploring other ways to measure whether the economy has become less competitive. But Trachter and Sarte view their findings as a word of caution for advocates of breaking up large firms.

"Our paper shows that market concentration is not market power," says Sarte. "There's more work to be done, but there's enough evidence here to suggest that we should at least pause and do that work before making major policy changes."