U.S. workers have been getting a shrinking share of the American economic pie, and economists have been puzzled by the decline. A new paper says one factor is the rise of “superstar” companies that dominate their market sectors and don’t need a lot of workers to generate profits.

“The aggregate share of labor falls as the weight of superstar firms in the economy grows,” the paper’s authors conclude. They are economists David Autor, Christina Patterson, and John Van Reenen of the Massachusetts Institute of Technology; Lawrence Katz of Harvard University; and David Dorn of the University of Zurich. The research was released in January as a working paper for the National Bureau of Economic Research.

The paper doesn’t name the superstars but notes that “firms may attain large market shares with a relatively small workforce, as illustrated by Facebook and Google.”

The study’s implications for policy are ambiguous. Donald Trump has promised to lift the wages of ordinary Americans but has also spoken positively about the contributions of U.S. tech giants–not just Facebook and Google but also Apple, Microsoft, Amazon.com, and others. “I’m here to help you folks do well,” he said at a summit with tech chief executive officers in December. Superstar companies can make the pie bigger, so squelching them probably wouldn’t make workers better off.

“It’s not immediately obvious what the policy would be” to reverse the downward trend for labor, Autor said in an interview.

The superstar theory is a fresh take on why workers’ share of gross domestic product has been falling since the 1980s, particularly since the 2000s. An earlier theory holds that American workers’ pay is ground down by competition from foreign labor. Another is that capital, such as machinery and software, has gotten cheaper relative to human beings. 

Autor and his fellow authors say superstar companies, because they’re big, can defray fixed labor costs such as headquarters staff over a bigger base of revenue and profits.  But why are there more such companies now than in the past? One theory they discuss is that new “competitive platforms,” such as the ability to compare prices on the internet, make it easier for the best companies to set themselves apart. Or it could be the proliferation of “information-intensive goods” such as software, which require relatively few people to produce in volume.

Using data from 676 industries in six sectors in the Economic Census, the authors find that the share of revenue controlled by the top four companies in an industry rose on average from 38 per cent in 1982 to 43 per cent in 2012 in the manufacturing sector; from 24 per cent to 35 per cent over the same period in finance; and from 15 per cent to 30 per cent in the retail trade. Concentration also rose in services and wholesale trade while falling slightly in utilities.

Next, the authors showed that the labor share fell the most in the industries with the greatest increases in concentration. They found no evidence that the superstars’ gains were ill-gotten. The increasing concentration seemed to be a sign of business success, not lobbying: The industries in which concentration increased the most, they found, were the ones that had the strongest growth in workers’ productivity.

Their conclusion: Technology and markets “increasingly concentrate rewards among firms with superior products or higher productivity–leading to better quality or lower costs–thereby enabling the most successful firms to control a larger market share.”

And the employees of those companies? They can make out well, too. There just aren’t that many of them.

Bloomberg News