Year-end bonuses for CEOs and top managers may explain a significant amount of the rise in wage inequality in the last 40 years.

In a new working paper posted by the National Bureau of Economic Research, Melanie Wallskog, an assistant professor of finance at Duke University’s Fuqua School of Business, found that since the 1980s, as firm productivity went up, so did pay inequality, largely because of performance-based bonuses that disproportionately benefit the top earners.

“We originally thought that more productive firms may have lower inequality,” Wallskog said. “But what we see is that, even though in productive firms all employees do better, it's the people at the top who benefit the most.”

Wallskog and co-authors Nick Bloom of Stanford University, Fed economist Scott Ohlmacher, and U.S. Census Bureau economist Cristina Tello-Trillo found support for this pattern in both publicly traded and private companies, although the effect is almost twice as strong for public companies.

Leveraging three U.S. Census Bureau datasets

Decades ago, CEOs at top companies made 15-20% more than the typical worker, but that ratio seems to have “skyrocketed” to more than 300% in recent years.

“There is some big explosion there,” Wallskog said.

The researchers wanted to examine whether highly productive firms were contributing in some ways to the rise in pay inequality.

To test their hypothesis, they studied data on millions of US employees from major U.S. Census Bureau datasets, such as the Longitudinal Employer Household Dynamics (LEHD) and the Longitudinal Business Database (LBD).

With the LEHD dataset, they tracked the quarterly pay of employees (which includes wages, bonuses and stock options) as well as the firm they were working for and other demographic data. They then ranked the employees on a scale 1 to 100, from the lowest earning percentile (1) to the highest (100).

In order to see whether firms with higher average pay were also the most productive, they drew from the LBD database to measure firm productivity as revenue per worker.

By matching the two datasets, they found a) that more productive firms tend to have higher average pay – meaning that everyone is better off — but b) that the higher earners benefit more than anybody else, showing that more productive firms have higher pay inequality.

To establish what drives this link between productivity and pay inequality, the researchers then leveraged the quarterly records on employee pay in the dataset, which showed evidence of top earners receiving more bonus pay at more productive firms.

“You could have thought that productivity would be raising all ships equally,” Wallskog said. “But what we found is that, although even the lowest paid people seem to be getting something, it’s the top people who receive the largest bonuses.”

With productivity almost doubling between 1980 and 2013 (2013 is the last available year for some of the data), the researchers estimated productivity may have accounted for up to 40% of the growth of pay inequality over the same period.

The researchers also found that the link between productivity and pay inequality is almost twice as strong in publicly traded companies than in private ones, another indication of the role of performance-based bonuses, which are more prevalent in public firms.

Testing the findings with further datasets 

The researchers tested the robustness of their findings by comparing their results with other datasets limited by sector, like the Census Bureau’s Management and Organization Practices Survey (MOPS), which is focused on manufacturing, or Compustat’s Execucomp, which gathers data on executive pay.

They all confirmed that larger, more productive companies drive pay inequality through “structured” incentive schemes that reward the higher earners more than anyone else.

The paper also found further support for a causal link between productivity and pay inequality by showing that when certain industries benefit from external, “lucky” shocks (for example, the oil industry benefiting from rising gas prices), pay also goes up, especially for the top earners.

Wallskog added that the large firms examined in this research (100+ employees) may even be more unequal than they appear because (as she found in previous work) some large companies outsource lower-paid jobs — security and cleaning services, for example — which are then “clustered” outside of their firm.

Is it possible to address inequality without tamping down productivity?

Rising productivity is a good thing, and linking pay to performance is also generally positive, Wallskog said.

At the same time, incentive-based pay may not work for everybody, she added, because these kinds of pay schemes are also volatile (they follow the ups and downs of the company’s revenues), and lower earners cannot afford that kind of uncertainty.

“You should use incentive-based pay for the people whose performance you can measure and whose jobs are extremely tied to the productivity of the firm, and to people who can afford to have volatility in their income, “ Wallskog said.

If regulators want to tamp down inequality, they should focus on the lowest paid people, she said, with policies directly impacting their earnings, such as minimum wage or unemployment insurance programs.

“You want to make sure the social safety net is there,” she said.