It has become widely accepted that workers are losing out in the global labor force to robots that can perform the same tasks at a lower cost. Over the past three decades, economists have tracked a decline in the share of the economy going to labor—and have assumed a perfectly equal rise in the share going to buy and maintain technology, including robots doing people’s work.
But Chicago Booth PhD candidate Simcha Barkai calculates the total annual amount companies pay for all of their capital, such as buildings, equipment, robots, and software, and finds that when presented as a share of economic output, that capital share is also declining.
In which case, where is the money going? Barkai has a theory: more is going to corporate profits.
Economists typically think of the economy as a series of trade-offs between labor (workers) and capital (investments in corporate assets). If the amount of money going to labor is falling, they surmise, it must be because more money is going to capital—in the form of such investments as factories or software. Thomas Piketty of the Paris School of Economics has famously argued in his book Capital in the Twenty-First Century that falling labor shares and rising capital shares drive global inequality.
But most economists fail to calculate the amount going to capital, relying on the underlying assumption that it includes whatever isn’t going to labor. This assumption has roots in research from the 1980s and 1990s, notably a study by Boston College’s Susanto Basu and Federal Reserve Bank of San Francisco’s John G. Fernald, suggesting that a third category, corporate profits, were very small and could be ignored. But Barkai’s work suggests the situation has changed.
Between 1984 and 2014, the labor share fell 10 percent, and the capital share fell 30 percent, Barkai calculates. And he finds that in this time period, the share going to profits increased sixfold—profits in US nonfinancial companies reached $1.4 trillion, or $17,000 per employee.
Barkai says that industries with larger increases in concentration also have larger declines in their labor share, and he offers two possible explanations. First, incumbent companies in concentrated industries who are able to compete with one another may choose not to in order to keep prices high. Second, a company dominant in an industry may be capable of producing a good or service at a lower cost than any of its competitors.
Consider a hypothetical dry cleaner that will clean a shirt for $1.25. That price covers a number of costs including transportation, chemicals, and labor. “This entire process requires people, and a decent amount of equipment,” Barkai says. A decade ago, the cleaner may have spent 80 cents on labor and 40 cents on machines, pocketing the final 5 cents.
In the past decade, however, the dry cleaner’s costs have declined, and the cleaner now spends 70 cents on people and 30 cents on machines. But because no other cleaner has managed to similarly reduce costs, the dry cleaner continues to charge $1.25, keeping 25 cents in profit. “You’d expect competition would encourage others to come in and reduce the price of laundering a shirt to $1.05,” says Barkai—but the cleaner faces no competitive pressure to reduce prices below $1.25.
While this is good for corporate profits, it’s less so for consumers and workers. Indeed, Barkai’s model suggests that in a more competitive environment, corporations would produce $750 billion to $1 trillion in additional goods and services, and wages would increase 20 to 25 percent.