The biggest question in economics today is whether capitalist competition still delivers the goods — specifically, whether it delivers them at a price not excessively above the costs of production. If competition is strong, companies can’t charge too much more than what they spent on labor, materials and overhead; otherwise, someone else would swoop in, cut prices, and nab customers. That’s the story, at least.
Last year, Jan De Loecker and Jan Eeckhout sought to demolish the above narrative in their blockbuster paper, “The Rise of Market Power and the Macroeconomic Implications.” Using data for publicly traded U.S. companies, they found that markups — the prices firms charge over the costs of production — have risen by a factor of 3.6 since 1980. Roughly speaking, where the average company used to charge $1.18 for something that cost an additional dollar to make, they now ask $1.67. The numbers were eye-popping, and the conclusion of rampant monopoly power drew widespread media attention, as well as some well-warranted skepticism.
Recently, the Jans published a followup that confirms their findings on an international scale. So does a separate IMF working paper released nearly concurrently. What’s most notable about the papers, however, is that both took the opportunity to respond to criticisms that De Loecker and Eeckhout’s original choice of data led to an overstatement of industry markups, and thus an inaccurate picture of how much we’re getting hosed by the companies that supply everything from beer to healthcare.
These new responses are worth exploring. De Loecker and Eeckhout made a splash in part because they offered such broad conclusions. Not only did they argue that markups have blown up since the Reagan era, they carried their findings over into some of the liveliest debates in economics. Everything from the decline in labor’s share of income to slowing measures of productivity growth might be explained by the markup hypothesis, the authors wrote. They set out to make waves and they did.
Their method was relatively simple. They assumed a neoclassical production function in each industry (a big ask for the heterodox-inclined), and used firms’ publicly reported sales and cost data to compute firm-level markups. Aggregating from 1950 to 2014, they arrived at this stark picture:
They also found that, within industries, markups have accelerated where concentration has increased most. Noting that the research held “profound policy consequences,” the authors concluded by offering up some bold theorizing about the macro implications. Absent the growth in markups, for instance, they estimated that the Dow Jones Industrial Average would stand at 5,500, not 22,000. They also hypothesized, contra Robert Gordon, that “properly accounting for market power, there is no productivity slowdown but instead an increase in productivity.” Here’s De Loecker with the broad strokes at ProMarket:
If you get a productivity shock in perfect competition, you’re just going to pass on that productivity gain to your consumers. Your costs go down and mechanically price goes down for everybody…But if you have a bit of market power, you’re going to hold some of those cost savings in your own pocket, and the wedge between price and cost is going to increase. If you don’t produce as much as you would in a perfect competitive outcome, you’re not going to need as much labor, you’re not going to need as much capital, and there’s not going to be that much entry. That kind of insight almost naturally gives rise to all those implications.
Given the conclusions, it’s not surprising that the Jans kicked up some dust. Rebuttals ranged from the mood affiliated — “ask yourself a simple question…in how many sectors of the American economy do I, as a consumer, feel that concentration has gone up and real choice has gone down?” offered one blogger — to the more substantive. In the latter category was a working paper earlier this year from U Chicago’s James Traina.
Part of Traina’s response deals with composition issues in the sample data used. Another par argues that measures of rising markups have less to do with market power than with underlying technological change that biases toward economies of scale. But Traina’s central critique concerns how the Jans define “costs.” He explains:
While De Loecker and Eeckhout (2017) use similar data sources and methods, I argue that they do not use the best accounting measure of variable cost. Specifically, they use Cost of Goods Sold (COGS), which measures direct costs of sales such as materials and most of labor. However, it excludes Selling, General and Administrative Expenses (SGA), which measures indirect costs of sales such as marketing and management. SGA costs are an increasingly vital share of variable costs for firms, and neglecting them meaningfully overstates both the level and growth in markups. By using Operating Expenses (OPEX), which is COGS plus SGA, I find a starkly different conclusion from the main empirical result of De Loecker and Eeckhout (2017).
The digital age demands more marketing, and technology shifts have raised costs for competent management and the like. Thus overhead costs, or SGA, should be included in the cost of production. Which seems reasonable enough. Here’s what markups look like after the Traina adjustment (blue line):
In Traina’s estimation, markups have risen since 1980, though “only modestly.” While the pattern shows “direct evidence of market power in the US economy,” the increase since 1980 falls within historical variation. The upshot: antitrust should not concern us here.
I do have one issue with Traina’s SGA assumptions. If firms are capturing increasing markup-driven rents, as in the De Loecker-Eeckhout story, we should expect management to pocket some of that lucre. Since SGA includes the salary and bonuses of the C-suite, a rise in executive compensation should be reflected by a rise in overhead, with larger increases in more oligopolistic industries. It’s no mystery that CEO pay has ballooned since the 1980s; depending on the magnitude of executive compensation within SGA costs, Traina’s findings might be consistent with increasing markups. *
In a new NBER paper that expands their original study worldwide, the Jans decline to address Traina by name. But they do take on the notion that SGA costs have driven the apparent rise in markups. To test this idea, they look at the relationship between their original measure of markups and profits. If costs were merely shifting from COGS to SGA, overall profitability wouldn’t be affected. Yet at the aggregate level they find that “higher markups lead to higher profits, and that they are not driven by higher overhead [SGA] costs.” Here’s the relationship between markups and one profitability measure, the ratio stock market value to sales:
As a response to Traina, this isn’t quite satisfying. In Traina’s story the shifting technological landscape has created mega-firms that spend more on management and marketing and also earn higher profits. The finding of an aggregate correlation between profits and markups (as measured by COGS) doesn’t dispel Traina’s concerns.
That said, the main thrust of the new De Loecker and Eeckhout paper is international: “Markups are increasing across the board, in all continents,” they conclude. Markups have risen sharply in North America and Europe and to a lesser degree in Asia. The trend is flatter in Africa and South America.
These findings are echoed in a working paper released last month by IMF economists Federico Diez, Daniel Leigh, and Suchanan Tambunlertchai. They find that since 1990, markups have risen by about 43 percent in advanced economies, and by a gentler 5 percent in emerging and developing economies. Notably, they arrive at these numbers using a different dataset than that of De Loecker and Eeckhout. They also provide evidence that the most highly concentrated industries dial back capital investment, as De Loecker and Eeckhout previously suggested.
The IMF economists also provide a more direct response to Traina. Citing Traina’s paper, they re-estimate aggregate markups, this time using SGA costs as a separate term within the markup regression (this is in lieu of adding SGA and COGS together, as Traina does). In this case they find that markups have risen globally by an average 35 percent, rather than 43 percent. Thus they conclude that “technological change associated with shifts in SGA accounts for a small part of the rise in markups.”
I think it’s fair to say that the exact measure of corporate markups remains an open question. In actual practice, however, no one interested in monopoly markups need listen to the music of the academic spheres. Take, for instance, the house view of the biggest bank in America, as expressed in the 2018 JPMorgan Chase equity market outlook.
Writing shortly after the Trump tax cuts passed, JPM analysts wrote that the prospects for a company’s stock in 2018 depended on whether its tax cut would be “competed away” by price-cutting, or “retained” by shareholders. The difference lies in whether a company has pricing power, the ability to raise markups without getting undercut by a competitor. As for the tax windfall, the JPM analysts were blunt: “Industry pricing power should be the primary determinant of how much of this upside is retained by shareholders vs. competed away and passed down to end-users.”
Alternately, you might consult the most successful investor of the twentieth century. When government financial crisis investigators interviewed Warren Buffett about his investments in the Moody’s rating agency, which is part of a powerful duopoly, he outlined his investment strategy in blunt terms:
… basically, the single-most important decision in evaluating a business is pricing power. If you’ve got the power to raise prices without losing business to a competitor, you’ve got a very good business. And if you have to have a prayer session before raising the price by a tenth of a cent, then you’ve got a terrible business.
If your business happens to be a monopoly, Buffett said, “your idiot nephew could run it.” This explains why, in 2016, after a series of mergers left the four top airlines in controls of about 80 percent of the market, Buffett abandoned his decades-long aversion to airline stocks and took a near 10 percent stake in each of the four companies.
The fact of monopoly markups has been obvious to Buffett and the big banks for years. Now the academic evidence seems to be catching up. Needless to say, we could use more investigation into the markup figures, particularly a more rigorous accounting of SGA and COGS costs. I also wonder if it’s possible to produce some kind of markups estimate that doesn’t rely on the industry-wide production function. Still, the evidence is already compelling that we’re all getting fleeced.
* There is some interest within accounting research in the phenomenon of SGA stickiness, i.e., the fact that the rise and fall in SGA costs are asymmetric to the rise and fall in output. One hypothesis surrounding the stickiness of SGA costs involves agency problems: perverse managerial incentives might be reflected in the SGA measure. Traina ignores this issue.