New research looking at the role of inter-firm and intra-firm inequality finds different results depending on the size of the firm.

The newest frontier in research about the rise of income inequality in the United States is the role of firms, or more specifically inter-firm inequality and intra-firm inequality. Inter-firm inequality is due to changes within firms rather than due to changes in workers’ education, skills or capabilities. Intra-firm inequality is about rising corporate executive pay across many companies. Almost a year ago, a paper about inter-firm inequality sparked interest in this area as it showed that almost the entirety of increased income inequality since the early 1980s was due to rising inter-firm inequality. A new version of that paper, released last week, adds more texture to our understanding of the role of firms in rising income inequality.

The newly updated paper, “Firming Up Inequality,” is by economists Jae Song of the Social Security Administration, David J. Price of Stanford University, Fatih Guvenen of the University of Minnesota, Nicholas Bloom of Stanford, and Till von Wachter of the University of California, Los Angeles. Relying on Social Security Administration data, the researchers’ top line result is this: The increase in inter-firm inequality is responsible for just under 70 percent of the rise in income inequality from 1981 to 2013, with the other 30 percent due to rising intra-firm inequality.

Thirty percent is, of course, is not insignificant, but the co-authors’ results do put inter-firm inequality squaring at the center of rising income inequality. Inter-firm inequality, by definition, includes two trends: increased firm earnings premiums (the boost to your earnings by working at a particular firm) and increased sorting of high-earning workers to high-paying firms and low-earning workers to low-paying firms. Both inter-firm inequality trends contribute about the same amount to rising income inequality, according to the results in the paper. This means the total rise in income inequality is roughly evenly split between rising intra-firm inequality, rising firm earnings premiums, and increase sorting. Interestingly, this division is very similar to results found using West German data.

But the really interesting results of updated paper have to do with firm size. Song and his co-authors run their analysis restricting their data to firms with less than 10,000 employees and those over that threshold. Firms below the 10,000-worker threshold employ about 70 percent of the U.S. workers and the larger firms employ the other 30 percent. The roles of inter-firm and intra-firm inequality are fairly different among these groups of firms.

For firms with fewer than 10,000 workers, inter-firm inequality played a larger role in increased income inequality, accounting for 87 percent of the overall increase, with 43 percent coming from increased firm earnings premiums and 33 percent from increased sorting. But for firms with more than 10,000 employees, intra-firm inequality was responsible for about half of the increase in income inequality. It’s not the sole factor, but it plays a much larger role for bigger firms than in the whole universe of firms. This rise in intra-firm inequality inside these firms is driven by two trends: declining earnings for workers in the bottom half of the firm’s pay scale and rapidly increasing pay for the top managers in these large firms.

These results are very important for thinking about the forces driving income inequality. Much of the prior research on the rise of income inequality was focused on understanding what might have increased intra-firm inequality. But now it seems that increased sorting and firm earnings premiums are just as important. Understanding why some firms can pay much better than others and why sorting has gone up (and how those factors differ by firm size) are now very important questions for researchers seeking to understand the rise of income inequality. The firm has now returned to an important role in our understanding of the labor market.

As this paper shows, data like those from the Social Security administrative have the potential to reveal nuances that were unseen until now. High-quality administrative data could really help inform our understanding of the role of firms in income inequality at more refined levels of analysis. This could be vital for interpreting the U.S. economic inequality and growth. Here’s to hoping more research comes soon.