One of the most urgent questions in economics is why pay for middle-income workers has increased only slightly since the 1970s, even as pay for those near the top has escalated.
For years, the rough consensus among economists was that inexorable forces like technology and globalization explained much of the trend. But in a new paper, Lawrence Mishel and Josh Bivens, economists at the liberal Economic Policy Institute, conclude that government is to blame. “Intentional policy decisions (either of commission or omission) have generated wage suppression,” they write.
Included among these decisions are policymakers’ willingness to tolerate high unemployment and to let employers fight unions aggressively; trade deals that force workers to compete with low-paid labor abroad; and the tacit or explicit blessing of new legal arrangements, like employment contracts that make it harder for workers to seek new jobs.
Together, Dr. Mishel and Dr. Bivens argue, these developments deprived workers of bargaining power, which kept their wages low.
“If you think about a person who’s dissatisfied with their situation, what are their options?” Dr. Mishel said. “Almost every possibility has been foreclosed. You can’t quit and get a good-quality job. If you try to organize a union, it’s not so easy.”
The slowdown in workers’ pay increases happened rather abruptly. From the late 1940s to the early 1970s, hourly compensation for the typical worker grew roughly as quickly as productivity. If the value of the goods and services that workers provided rose by 2 percent in a year, then their wages and benefits tended to go up by roughly 2 percent as well.
Since then, productivity has continued to grow, while hourly compensation largely flattened. According to the paper, the typical worker earned $23.15 an hour in 2017, far less than the $33.10 that worker would have earned had compensation kept up with productivity growth.
In the 1980s and 1990s, economists increasingly argued that technology largely explained this flattening of wages. They said computers were making workers without college degrees less valuable to employers, while college graduates were becoming more valuable. At the same time, the growth in the number of college graduates was slowing. These developments dragged down wages for those in the middle of the income distribution (like factory workers) and increased wages for those near the top (like software engineers).
The technology thesis largely relied on a standard economic analysis: As the demand for lower-skilled workers dropped, their wages grew less quickly. But in recent years, many economists have gradually de-emphasized this explanation, focusing more on the balance of power between workers and employers than on long-term shifts in supply and demand.
The idea is that setting pay amounts to dividing the wealth that workers and employers create together. Workers can claim more of this wealth when institutions like unions give them leverage. They receive less when they lose such leverage.
Dr. Mishel and Dr. Bivens argue that a decades-long loss of leverage largely explains the gap between the pay increases that workers would have received had they benefited fully from rising productivity, and the smaller wage and benefit increases that workers actually received.
To arrive at this conclusion, they examine numerical measures of the impact of several developments that hurt workers’ bargaining power — some of which they generated, many of which other economists have generated over the years — then sum up those measures to arrive at an overall effect.
For example, when surveying the economic literature on the unemployment rate, Dr. Mishel and Dr. Bivens find that it was frequently below the so-called natural rate — the rate below which economists believe a tight job market could cause inflation to accelerate uncontrollably — in the three decades after World War II, but frequently above the natural rate in the last four decades.
This is partly because the Federal Reserve began to put more emphasis on fighting inflation once Paul Volcker became chairman in 1979, and partly because of the failure of state and federal governments to provide more economic stimulus after the Great Recession of 2007-9.
Drawing on existing measures of the relationship between unemployment and wages, Dr. Mishel and Dr. Bivens estimate that this excess unemployment lowered wages by about 10 percent since the 1970s, explaining nearly one-quarter of the gap between wages and productivity growth.
They perform similar exercises for other factors that undermined workers’ bargaining power: the decline of unions; a succession of trade deals with low-wage countries; and increasingly common arrangements like “fissuring,” in which companies outsource work to lower-paying firms, and noncompete clauses in employment contracts, which make it hard for workers to leave for a competitor.
Together, Dr. Mishel and Dr. Bivens conclude, these factors explain more than three-quarters of the gap between the typical worker’s actual increases in compensation and their expected increases, given the productivity gains.
If that figure is in the right ballpark, it is a crucial insight. Underlying most of the explanations for anemic wages that Dr. Mishel and Dr. Bivens cite is the idea that wage growth depends on policy choices, not on the march of technology or other irreversible developments. Government officials could have worried less about inflation and erred on the side of lower unemployment when setting interest rates and passing economic stimulus. They could have cracked down on employers that aggressively fought unions or foisted noncompete agreements onto fast-food workers.
And if policymakers are to blame for wage stagnation, they can also do a lot to reverse it — and more quickly than many economists once assumed. Among other things, the conclusion of the paper would suggest that President Biden, who has enacted a large economic stimulus and sought to increase union membership, may be on the right track.
“One of the biggest things about the American Rescue Plan,” said Dr. Mishel, referring to the pandemic relief bill Mr. Biden signed, “is first and foremost its commitment to getting to full employment quickly. It’s willing to risk overheating.”
So is the paper’s number plausible? The short answer from other economists was that it pointed in the right direction, but may have overshot its mark.
“My sense is that things like fissuring, noncompetes have become very important in the 2000s, along with unions that have gotten to the point where they’re so weak,” said Lawrence Katz, a labor economist at Harvard who is a longtime proponent of the idea that the higher wages earned by college graduates have increased inequality.
But Dr. Katz, who has also written about unions and other reasons that workers have lost leverage, said the portion of the wage gap that Dr. Mishel and Dr. Bivens attribute to such factors probably overstated their impact.
The reason, he said, is that their effects can’t simply be added up. If excessive unemployment explains 25 percent of the gap and weaker unions explain 20 percent, it is not necessarily the case that they combine to explain 45 percent of the gap, as Dr. Mishel and Dr. Bivens imply. The effects overlap somewhat.
Dr. Katz added that education plays a complementary role to bargaining power in determining wages, citing a historical increase in wages for Black workers as an example. In the first several decades of the 20th century, philanthropists and the N.A.A.C.P. worked to improve educational opportunities for Black students in the South. That helped raise wages once a major policy change — the Civil Rights Act of 1964 — increased workers’ power.
“Education by itself wasn’t enough given the Jim Crow apartheid system,” Dr. Katz said. “But it’s not clear you could have gotten the same increase in wages if there had not been earlier activism to provide education.”
Daron Acemoglu, an M.I.T. economist who has studied the effects of technology on wages and employment, said Dr. Mishel and Dr. Bivens were right to push the field to think more deeply about how institutions like unions affect workers’ bargaining power.
But he said they were too dismissive of the role of market forces like the demand for skilled workers, noting that even as the so-called college premium has mostly flattened over the last two decades, the premium for graduate degrees has continued to increase, most likely contributing to inequality.
Still, other economists cautioned that it was important not to lose sight of the overall trend that Dr. Mishel and Dr. Bivens highlight. “There is just an increasing body of work trying to quantify both the direct and indirect effects of declining worker bargaining power,” said Anna Stansbury, the co-author of a well-received paper on the subject with former Treasury Secretary Lawrence Summers. After receiving her doctorate, she will join the faculty of the M.I.T. Sloan School of Management this fall.
“Whether it explains three-quarters or one-half” of the slowdown in wage growth, she continued, “for me the evidence is very compelling that it’s a nontrivial amount.”