How Costly Is Rising Market Power for the U.S. Economy?
We survey the recent, active debate on market power in the U.S. economy. While typical studies on market power focused on narrow industries due to data constraints, the relevance of market power for the aggregate economy was reinvigorated by a study focusing on publicly traded firms that documented a significant rise in U.S market power since the 1980s. This article is meant to provide a bird's-eye view of the (sometimes heated) discussion on market power. Furthermore, we examine the macroeconomic consequences of a rise in U.S. market power.
For an economy to achieve efficiency, it needs to be perfectly competitive. Only under this specific case are goods and services produced and sold at the lowest possible price. In the absence of perfect competition, however, resources are not allocated optimally, and consumers face higher prices due to firms exerting their market power. As a result, high levels of market power could have devastating consequences for welfare and inequality.
While economists are fully aware of the qualitative ramifications of market power, there are surprisingly few studies that try to quantify it. In this article, we survey the recent, active debate on market power and explore the implications of rising market power for the U.S. economy. We do so by first reviewing the study that documented a significant rise in U.S. market power since the 1980s. This study ignited the debate on market power, leading to many follow-up studies and criticism.
Rising Market Power in the U.S. Economy
The lack of studies covering the extent of market power for a large part of the U.S. economy is mostly due to its difficulty of measuring it. A firm's market power is typically defined as its ability to set a profit-maximizing price above marginal cost. While prices are observable, firms' marginal costs are not.
However, the 2020 study "The Rise of Market Power and the Macroeconomic Implications" made significant progress on this front by relying on methodological innovations from the industrial organization literature. More importantly, it found that aggregate U.S. market power has risen considerably in the past few decades: Aggregate markups went from 21 percent above marginal costs in 1980 to 61 percent in 2016. Even under conservative assumptions about a firm's returns to scale, this would imply that a firm with a markup of 61 percent above marginal cost would make 38 cents of profit on every dollar of revenues. The study has led to a true renaissance of market power in academic studies and sparked many debates in policy circles.
To identify markups, this study leverages a key insight from the so-called "production approach:" A cost-minimizing firm's markup can also be captured by the "production wedge" of a flexible input.1 This wedge equals its output elasticity (that is, the percentage increase in a firm's physical output by increasing the flexible input by 1 percent), and this input's expenditure share relative to revenues.
The emphasis on "flexible" is crucial. In this case, the ratio between what a firm obtains in terms of output and what it costs must reflect its good's profit margin as reflected by its markup. In their study, which focuses on publicly listed firms in the U.S., the authors use a firm's cost of goods sold as their measure of a flexible input, leading to the following findings:
- The aggregate markup — defined as an economy's sales-weighted markups across firms — has increased dramatically from 21 percent in 1980 to 61 percent above marginal costs in 2016. That is, market power has increased by about one-third over the course of almost four decades.
- The rise in the aggregate markup has been driven by a reallocation of market share from low-markup to high-markup firms. That is, individual firms' markups have not necessarily risen over time. Aggregate trends in market power are driven by a composition effect across firms. Interestingly, the median firm's markup has not changed over time.
- Average profit rates relative to revenues — which are related to markups — have increased from 1 percent in 1980 to 8 percent in 2016.
The third observation is important: An increase in markups can also be a mere reflection of a rise in overhead costs. Under the latter scenario, firms only increase profit margins to cover these fixed costs to avoid losses, not to exploit their market power. However, this study shows that while overhead costs have risen somewhat, markups have increased disproportionately more, implying that "excess" markups did increase over the past few decades.
Macroeconomic Implications
While a rise in market power can lead to deteriorating welfare, several studies have also suggested that the recently documented rise in market power is the driving force behind other U.S. long-run (or secular) trends of interest. One of the most (if not the most) well-known secular trends is the decline in the labor share of income: From the early 1950s until the late 2010s, the share of income that accrued to workers' labor compensation and income for self-proprietors declined by 8 percent. When a firm has market power, it has an incentive to keep prices high by keeping output below competitive levels. In other words, market power leads to lower production which dampens the demand for inputs, including labor.
Even though there is a plethora of explanations for the decline in the labor share, this study argues that a rise in market power is consistent with other stylized facts documented for the U.S. economy. Under perfect competition, a decline in the labor share must imply an increase in the capital share since value added consists of labor and capital by construction. However, recent evidence (in particular, the 2020 paper "Declining Labor and Capital Shares") indicates that the share of income on capital has also declined. To reconcile these facts, economic profits must have increased, which has been supported by the data and, more importantly, is consistent with the rising market power hypothesis.
In addition, depressed labor demand due to rising market power leads to lower labor force participation and real wages. The latter occurs even when labor supply is completely elastic (leading to nominal wages being unchanged), since the price of output goods increases when market power increases. Note that these effects on labor market outcomes stem from market power for output. Hence, the documented facts can even occur in the absence of market power in input (such as labor) markets.
The macroeconomic implications of a rise in market power can also stretch to future producers. In particular, incumbent firms with high levels of market power can anchor their leader positions by imposing significant barriers to entry and/or simply buying up any future competitors (as noted in the 2023 paper "The Great Start-Up Sellout and the Rise of Oligopoly"). As a result, entrepreneurs have very little incentive to start new businesses. This seems consistent with the secular decline in firm entry that I recently discussed in my 2023 article "Why Are Startups Important for the Economy?"
In related work, the 2022 book The Profit Paradox describes how the benefits of the rise in market power have accrued to only a handful of firms and their behavior in "acquiring rivals and securing huge profits" has created "brutally unequal outcomes for workers."
As we have seen, a rise in market power affects real wages, but the 2021 paper "Kaldor and Piketty's Facts: The Rise of Monopoly Power in the United States" shows quantitatively that it also affects wealth inequality, which tends to be much higher than, say, income inequality. The intuition is straightforward: A rise in market power leads to higher economic profits and, hence, higher stock prices. It is well known that the wealthiest individuals own the most (if not the overwhelming majority of) stocks. Hence, thriving firms benefit the wealthiest in a disproportionate manner, widening wealth inequality.2
Critiques of "The Rise of Market Power and the Macroeconomic Implications" Paper
As noted earlier, the influential study "The Rise of Market Power and the Macroeconomic Implications" (which we'll refer to as the "market power study" for the rest of the article) has led to many follow-up studies investigating other aspects of U.S. market power. However, it has also drawn an equal amount of criticism. Three critiques stand out in particular.
Econometric Flaws
The 2021 paper "Some Unpleasant Markup Arithmetic: Production Function Elasticities and Their Estimation From Production Data" points out there are several econometric flaws in the market power study. The most poignant involves "deflated revenues." To obtain markups through the production approach, physical output elasticities for a flexible input are required. However, physical output is rarely observed in firm-level data, so researchers typically resort to using deflated revenues, or dividing total revenues by some price index.
The problem with proxying physical output in this way, however, is that the production approach then results in revenue elasticities. The "Unpleasant Markup Arithmetic" paper shows that the ratio between a flexible input's revenue elasticity and its revenue share must equal 1. The fact that the market power study consistently observes values above 1 must indicate that some fundamental assumptions are violated and the obtained production wedges must reflect something else than market power. This is critical since we have seen how a market power interpretation has far-reaching consequences for the evolution of the U.S. economy.
Flexibility of Cost of Goods Sold
The 2018 working paper "Is Aggregate Market Power Increasing? Production Trends Using Financial Statements (PDF)" makes a related point. It argues that the adopted input "cost of goods sold" is not flexible. This is a reasonable argument since the cost of goods sold typically contains expenses on raw materials and labor. It has been argued that labor is subject to frictions such as adjustment costs and is furthermore dynamic. As a result, it disqualifies the cost of goods sold as a flexible input.
Instead, this paper replicates the exercise in the market power study with operating expenses as a flexible input. The author's results lead to a starkly different conclusion: U.S. market power has barely moved the past few decades. More importantly, its level has stayed constant at around 1.15. While it is hard to reconcile these findings with rising profit rates, it does illustrate the fragility of the results of the market power study.
Markups Occurring Through Costs
The 2023 paper "How Costly Are Markups?" argues that aggregation of individual firms' markups must occur through costs, rather than revenues/sales (which is the baseline measure of the market power study). A sales-weighted average of markups might reflect useful information on the distribution of markups, but it is not consistent theoretically with any aggregate measure of markups. This paper shows that a cost-weighted average of markups is the aggregate object that is relevant for welfare in a rich class of models.
More importantly, the choice of aggregation also has quantitative consequences. The rise in market power seems to be much milder: It shows an increase of about 15 basis points from the early 1980s to 2016. This is in stark contrast with the increase of 40 basis points in the baseline results discussed earlier.
Welfare Costs of Markups
Despite the lively debate on market power containing many critics, most academics seem to agree that market power is present in the U.S. economy and has been moderately rising. The only question seems to be the magnitude of this rise. While a consensus on the magnitude of this rise has not been reached, some studies have already quantified the welfare costs of markups under several scenarios. Some of the studies take the numbers in the market power study at face value, whereas others adopt more conservative figures.
One study of the latter category is the "How Costly Are Markups?" paper. The authors show that the presence of markups affect welfare in three ways. First, markups induce a so-called "deadweight loss." When firms price their good(s) above marginal cost, there are some consumers who are not willing to pay. However, these same firms could have offered their goods at a lower price but still above marginal cost. This would make these consumers (who get to consume these firms' goods) and firms (who would make strictly positive profits) strictly better off. This efficiency loss (induced by the level of markups) is also referred to as the "Harberger triangle."3 This type of loss induced by the aggregate markup is referred to as a "uniform output tax" in this paper.
Second, inefficiencies can occur because resources get used by firms that aren't the most productive, inducing misallocation. The authors show that the non-optimal allocation of resources can be summarized by the dispersion in markups across firms.
Third, markups distort entry decisions of future producers. In their paper, the authors show that the aggregate markup and misallocation channels induce the bulk of markup welfare losses. Depending on the level of the aggregate markup (reminiscent of the sales-weighted average markup in the market power study), welfare losses induced by markups can be sizable. If the aggregate markup would be equal to 1.15 (which is a fairly conservative value among studies) and the dispersion of markups is moderate (for example, the interquartile range of firm-level markups is only 7 percent), then aggregate productivity would be 2.7 percent lower when compared to a social planner's outcome who maximizes welfare.
Another way of stating these losses is by considering the amount of consumption a representative consumer would be willing to give up to not face any markups. The authors show that this value is high: When the aggregate markup is 1.15, this consumption-equivalent welfare loss would be 8.7 percent of annual consumption. Note that this is an order of magnitude larger than typical calculations for the welfare cost of business cycles.4 Importantly, this paper shows that the welfare cost of markups increases disproportionately when the aggregate markup increases further. For example, the consumption-equivalent welfare loss increases to 23.6 percent when the aggregate markup is 1.25 instead.
This study adopts rather conservative values for the level and dispersion in markups when compared to the market power study. This is partly because the authors are aware of the issues of the production approach, but also because they consider only one source of variation (that is, size) that can cause differences in markups across firms. The 2020 paper "Productivity and Misallocation in General Equilibrium" takes the markups by the market power study at face value and feeds these numbers into their structural framework. As a result, the authors consider markup dispersion that can be caused by any source (not just size). As expected, their calculations show a much larger efficiency loss: Aggregate productivity would have increased by 23 percent if markups were eliminated in 1997. This is an order of magnitude larger than implied by the "How Costly Are Markups?" paper. Hence, the welfare losses induced by markups can be massive.
These studies indicate that even moderate levels of market power can be extremely costly for society. Are there any policies that can correct these distortions caused by market power? The book The Profit Paradox stresses the importance of restoring competition. Market-power-exploiting firms have been thriving due to lack of competitors.
However, the "How Costly Are Markups?" paper demonstrates this is not obvious. Subsidizing entry would definitely encourage more firms to enter the playing field, but the welfare-relevant aggregate markup (that is, the cost-weighted average markup) would barely decrease. This is because the increase in the number of firms would hurt smaller firms more than their larger counterparts.
In this paper's framework, larger firms set higher markups. While every firm's markup would decrease due to the increased competition ("direct" effect), there is a disproportionate shift towards larger firms ("indirect" effect through composition). In the end, the authors show that these direct and indirect effects roughly cancel each other out leading to only a minor fall in the aggregate markup.
For relatively low levels of the aggregate markup, they show that most of the welfare cost of markups is accounted for by the aggregate markup itself. (That is, the misallocation and entry channels matter less under this scenario.) As a result, policies that target the level of the aggregate markup can be quite effective. Under this paper's setup, firm's demand functions are parameterized and, hence, fully known. If this information is available, then size-dependent policies can be created to restore an economy's efficient outcome. Obviously, it is unrealistic to assume this information is known for all firms, but their exercise does illustrate why bluntly stimulating competition might not be effective at all.
Conclusion
Firms' market power can induce large efficiency losses. While economists have long known the implications of market power, quantitative studies on market power not restricted to small parts of the U.S. economy have been missing. The recent paper "The Rise of Market Power and the Macroeconomic Implications" argues that U.S. market power has dramatically increased since the early 1980s. While this study on market power has been heavily criticized, its contribution has put market power fully back on academic and policy radars.
Importantly, follow-up studies have demonstrated that even moderate levels of market power can induce large welfare costs. In terms of aggregate productivity, these losses can range from 3 percent to 23 percent. As a result, the problems surrounding market power seem quantitatively important. While the lack of competition can be an important source of market power, bluntly subsidizing entry to increase the playing field can be ineffective: Indirect, compositional effects can offset most of the downward pressure on markups induced by these policies. As a result, more research is required to characterize the U.S. economy's market structure to develop more nuanced size-dependent policies. The increasing availability of sophisticated microlevel data gives us hope that we can do so soon.
Chen Yeh is an economist in the Research Department at the Federal Reserve Bank of Richmond.
A flexible input is some static input that is supplied competitively and does not feature any adjustment costs. Typically, the industrial organization literature considers material inputs to be flexible.
For more on this discussion, see, for example, the 2021 articles "Who Owns Stocks? Explaining the Rise in Inequality During the Pandemic" and "Wealth Inequality in the United States."
Under linear demand and supply curves, this efficiency loss is exactly geometrically characterized by a triangle. This notion was popularized and first quantified in a more general setting by the 1966 paper "Efficiency Effects of Taxes on Income From Capital." If firms were able to charge different prices to each different consumer (i.e., perfect price discrimination), these losses would not occur. However, this does not seem to be supported by the data.
See, for example, the 2003 paper "Macroeconomic Priorities."
To cite this Economic Brief, please use the following format: Yeh, Chen. (July 2023) "How Costly Is Rising Market Power for the U.S. Economy?" Federal Reserve Bank of Richmond Economic Brief, No. 23-24.
This article may be photocopied or reprinted in its entirety. Please credit the authors, source, and the Federal Reserve Bank of Richmond and include the italicized statement below.
Views expressed in this article are those of the authors and not necessarily those of the Federal Reserve Bank of Richmond or the Federal Reserve System.
Receive a notification when Economic Brief is posted online.