Competition and monopoly in the U.S. economy: What do the industrial concentration data show?

Abstract

A recent series of academic studies, think-tank reports, and news articles shows widespread attention to rising industrial concentration and market power in the U.S. economy. In this paper, we focus on concentration in the U.S. nonfinancial corporate sector to make three contributions to the literature. First, we use examples from the debate on industrial concentration to show that there are often-divergent predictions in the theoretical literature surrounding the expected consequences of concentration and monopolization for nonfinancial firms. Second, we use industry-level concentration data to describe recent trends in average concentration. We show that, while concentration increases across the majority of industries after the late 1990s, the retail and information-services sectors are particularly key for understanding recent trends in average industrial concentration. Third, we link our industry-level analysis with firm-level data to describe the relationship between industrial concentration and nonfinancial corporations’ profitability, markups, and investment. Consistent with the ambiguities in the theoretical literature, we find that these relationships are not uniform: while some highly concentrated industries confirm standard expectations with high markups, high profitability, and low investment rates, other highly concentrated industries earn lower-than-average markups and profits, suggesting that – in some industries – increased concentration and intensified competition may go hand in hand.

Keywords

Industrial concentration market concentration monopolization competition capital accumulation profitability

Introduction

Industrial concentration and monopolization in the U.S. economy has attracted significant attention in the last few years. A recent literature emphasizes increasing concentration ratios and market power over the last two decades, and argues that this increase is a significant factor behind a number of major macroeconomic trends, including rising corporate profits, slowing investment, declining labour share of income, and growing income and wealth inequalities (e.g. Barkai, 2016; CEA, 2016; De Loecker and Eeckhout, 2017; Grullon et al., 2019; Gutiérrez and Philippon, 2017b; Kurz, 2017; Stiglitz, 2016). Rising concentration is variously explained as the result of lax antitrust enforcement, developments in information technology systems, and the emergence of winner-take-all industries in which superstar firms bring significant productivity increases. Nonetheless, interpretations of the existing evidence on the degree of competition are mixed. Shapiro (2018) argues that the evidence is insufficient to establish declining competition within many industries. Crouzet and Eberly (2018) contend that, when excluding retail, the aggregate rise in concentration largely disappears. There are three important issues that are insufficiently addressed by this literature. First, the theoretical literature on industrial concentration and monopolization is discussed little, and a straightforward story in which increased industrial concentration drives higher profitability and lower investment is presumed. Second, most analyses emphasize average across-industry concentration ratios and changes in these averages, with less attention to whether rising average concentration reflects a widespread change across industries or a shift only within specific sectors. Third, it is unclear whether the profit rates, markups, and investment rates of firms in highly concentrated industries differ systematically from those in less-concentrated industries.

In this paper, we focus on these three issues by linking results from the theoretical literature on concentration and monopoly power to data describing the relationship between concentration, profitability, and investment. First, in the ‘Approaches to competition and monopoly’ section, we use examples from the theoretical literature to show that the expected relationships between industrial concentration and profitability and investment vary both across and within neoclassical, Keynesian, and Marxian approaches. By comparing these approaches, we show that much of the theory regarding the impact of concentration on firm outcomes is unsettled, and that these approaches generate often-divergent predictions about the implications of industrial concentration and monopolization. While we identify a ‘standard’ story – wherein high concentration is indicative of low competition and allows firms to earn high profits, charge high markups, and perhaps even reduce investment expenditures – the theoretical discussions suggest the relationships are less straightforward.

Second, in light of the divergent conclusions in the theoretical literature, we turn to industry- and firm-level data to describe the relationship between concentration, profitability, and investment. In the ‘Trends in industrial concentration since 1997’ section, we use industry-level concentration data to document aggregate concentration growth across the U.S. economy, identify highly concentrated industries, and explore if rising average concentration reflects a widespread change across industries or primarily a shift within specific sectors. In addition to showing an increase in average concentration that is robust to different measures of concentration and levels of industrial classification, we document an important sectoral dimension to rising concentration. Specifically, both the retail and information sectors play important roles in driving the recent rise in, and the current level of, average concentration. In contrast, manufacturing industries tend to have lower-than-average concentration. Both growth in information services and a decline in manufacturing over this period, therefore, put upward pressure on aggregate concentration, suggesting that structural change has contributed to rising average concentration.

Third, in the ‘Concentration, profitability, and investment’ and the ‘Capital and intangible intensities’ sections, we link our industry-level analysis to firm-level data to ask whether the profitability, markups, and investment rates of firms in highly concentrated industries differ systematically from those of firms in less-concentrated industries. The relationships we show between concentration and profitability, markups, and investment suggest that links between the level of concentration and the nature of competition in an industry are not straightforward. We show that, while firms in industries with above-average concentration have higher profitability than those in below-average concentration industries, there are key differences within the above-average concentration group. In particular, higher profitability across above-concentration industries is not due to firms in industries with the highest concentration levels. Instead, the highest profit rates accrue to firms in industries falling in the middle of the concentration distribution. This finding adds to an earlier literature that establishes a weak link between profitability and concentration (Keil, 2017a; Melmiès, 2016) by suggesting this weak relationship reflects, specifically, lower profitability in the most concentrated industries. Similar heterogeneity appears with markups: among firms in high-concentration industries, only those in specific industries – mainly information-services firms – have above-average markups. Turning to investment, we show a sharp fall in average investment during the early 2000s, when a large share of concentration growth occurred. Firms in low-concentration industries do not, however, have higher investment rates on average, although average investment in the most concentrated industries has fallen below that of mid-concentration firms in recent years. Finally, in the ‘Capital and intangible intensities’ section we explore if fixed-capital intensity or intangible-asset intensity help clarify these trends. The evolution of intangible assets in particular suggests that, for some industries, rising market power may be expressed through intangibles, even with lower industrial concentration (Orhangazi, 2019).

Together, these findings run in part against the common perception that firms in highly concentrated industries have higher profitability, higher markups, and lower investment. In the ‘Discussion and concluding remarks’ section, we discuss these results and suggest that, rather than a single ‘standard story’, at least three cases describe the relationship between concentration, profitability, and investment: (i) Industries with high (low) concentration and high (low) markups, high (low) profitability, and low (high) rates of investment, which confirm common expectations. Information-services firms may be an example of this case. (ii) Industries, like those in retail, that are highly concentrated, but have low markups, low profitability, and average investment. Notably, this case suggests that the degree of concentration need not capture the degree of competition, and that, in some industries, increased concentration may go hand in hand with intensified competition. (iii) Industries with medium-levels of concentration, high profitability, and above-average investment rates. This case suggests that concentration is not the only reflection of market power, and firms with lower market shares may acquire market power (for example, through intangible assets) to increase their markups and profitability, while continuing to invest above average. These three cases suggest that there is still much room and, in fact, need for theoretical and empirical studies on the current dynamics of industrial concentration and competition in the U.S. economy.

Approaches to competition and monopoly

Competition occupies a central role in different explanations of how capitalist systems work since classical economics. In this section, we outline basic theoretical approaches to competition and monopoly in the neoclassical, Marxian, and Keynesian literatures. While we do not attempt to exhaustively review each strand of the literature, we use examples from each tradition to outline important themes in each approach to competition and monopoly, with an emphasis on predictions regarding profitability and investment.

Neoclassical theory is built around the case of perfect competition in markets with a large number of firms, each with small enough shares that no single firm can affect the market price. Firms, therefore, simply decide on the level of output, given the market price and their cost structure, to maximize profitability. Firms in perfectly competitive markets earn ‘normal’ profits, defined as returns sufficient to pay all inputs at least their returns in alternative employment, i.e. their opportunity cost. Monopoly is juxtaposed against this ideal(ized) set-up. For monopolies, output decisions directly change total supply and hence the market price. By limiting output, monopolies increase the market price above marginal cost and earn ‘monopoly profits’, or ‘rents’, defined as returns to inputs in excess of the amount necessary to keep them in operation. Perfect competition maximizes both consumer and producer surpluses, while monopoly generates deadweight losses. Monopolies charge higher prices, limit output, and reduce consumer welfare. In turn, imperfect competition describes the cases between these two polar states. Under imperfect competition, both competitive and monopolistic forces are combined in determining market prices. Even in markets with rival producers, oligopolies or price-fixing cartels can charge higher prices with small losses in sales. The two common versions of the oligopolistic market structures are Cournot- and Bertrand-type models. While in Cournot competition oligopolistic firms decide on production quantity, in Bertrand competition oligopolistic firms decide on prices. Like monopolies, imperfect competition reduces the efficiency of resource allocation, especially in the Cournot-type market structures, and causes wasteful rent-seeking activities (Pepall et al., 2014; Tirole, 1988).

Perfect competition is sometimes depicted as a Darwinian process in which firms are forced to innovate in order to remain competitive for survival. Against this claim, others argue that monopoly rents may be allocated to promote innovation, especially if market power provides firms with incentives for research and development that promotes long-term economic growth. In particular, a monopolistic firm’s incentive for investment and innovation comes from the desire to protect and continue their monopoly rents. This is in line with the Schumpeterian approach where competition is a dynamic process of differentiation among firms in which technological innovation plays the central role. Therefore, in Schumpeter, neither is monopoly associated with stagnation nor does competition necessarily generate dynamism. In fact, according to Schumpeter, monopolization may increase innovation to defend monopoly positions. Proponents of the Darwinian approach argue that survival is a stronger incentive than protecting monopoly rents, and that the lack of market discipline in monopolistic markets can lead to managerial slack and agency problems resulting in sub-optimal levels of investment and innovation. As such, different strands each expect higher pricing power and profitability for monopolies, but the implications for investment are contested.

For the early post-war period, mainstream thinking and policy in the U.S. was dominated by a negative view of monopolies and industrial concentration. Promoting competition was generally accepted as the best policy to increase social welfare and has been a central tenet of antitrust laws since the Sherman Act (Stucke, 2013). The consensus changed in the late 1970s, with Bork’s (1979) consumer-welfare metric for antitrust. This metric contends that the broad application of antitrust law to promote competition can harm consumer welfare; therefore, antitrust enforcement should be limited to cases when monopoly power reduces consumer welfare (Shapiro, 2018). A recent note by the U.S. government lays out this position, arguing that increased concentration

…would not necessarily imply a failure of competition law or enforcement. Increasing concentration is apt to occur as a result of two distinct, albeit similar, natural forces. First, when success and failure are random events, markets become concentrated over time. Second, when success and failure are driven by relative degrees of innovation and efficiency, markets also become more concentrated. Firms that serve their customers’ interests much better than rivals can gain substantial market share as a result of a healthy competitive process. (OECD, 2018: 6)

The recent literature on concentration further exemplifies the tension in the mainstream view between competition and economic efficiency. Some argue that the recent increase in concentration is not only associated with higher markups and lower investment (Gutiérrez and Philippon, 2017a), but that it also reduces consumer welfare, increases rents, decreases business dynamism, and harms resource allocation (De Loecker and Eeckhout, 2017). Others argue that concentration is the efficient byproduct of technological change, as productivity differentials drive a reallocation of demand towards the highest-productivity firms, and/or as rising productivity differences lead to concentration (Autor et al., 2020, 2017).

Within the post-Keynesian literature, oligopolistic competition is considered a natural state of affairs with limited discussion of its efficiency and no calls for promoting competition. Robinson (1933) developed an early theory of imperfect competition, around the same time that Chamberlin developed his theory of monopolistic competition. Kalecki (1971) saw perfect competition as ‘a most unrealistic assumption not only for the present phase of capitalism but even for the so-called competitive capitalist economy of the past centuries: surely this competition was always in general very imperfect’ (158).¹ Furthermore, an increase in the average degree of monopoly could reduce the wage share and increase the capital income share (22, 63). Profit margins are determined by the intensity of competition, usually referred to as the degree of competition: the lower the degree of competition, the higher the profit margin. The degree of competition not only determines profit margins but, at the aggregate level, also determines profit and wage shares.

Central to this approach is the premise that almost all markets have administered pricing (e.g. Lavoie, 2015). Thus, administered pricing need not indicate the existence of oligopolies, as it also exists in markets with intense competition if there are a limited number of competitors. In turn, competition need not necessarily occur through pricing, but also takes place when firms attempt to reduce unit costs to achieve larger profit margins than their competitors (127). In fact, competitive activities around investment, advertising, research and development, production process, and production decisions can lead to such significant cost differences across firms that many are driven from the market (Lee, 2013: 169). As such, ‘[c]ompetition is a dynamic process, not an end-state or a static situation’ (Lavoie, 2015: 127) and too much competition can easily be risky for firms: ‘Thus the post-Keynesian position on the value of competition is rather ambiguous’, but shares similarities with Schumpeter (1943) who states that ‘perfect competition is not only impossible but inferior, and has no title to being set up as a model of ideal efficiency’ (639). A final branch of the post-Keynesian literature regards profit margins as correlated not to the degree of market power, but instead as directly connected to the internal financing requirements for investment (Melmiès, 2016: 154). Accordingly, ‘managers choose a profit margin that takes market competition into account but also yields a targeted profit rate needed to finance the growth of the firm’ (156, emphasis in original). This point follows from Eichner (1983), who argues that markups are mostly determined by an industry’s growth rate multiplied by its incremental capital-output ratio. Firms seek higher markups when they need funds to expand investment and weigh the cost of higher markups against lower long-term sales (138).

Marxian approaches see dynamics of competition as central to the workings of the capitalist system. In an often-quoted passage, Marx (1990 [1867]) argues that competition is the driving force of the system as

…the development of capitalist production makes it necessary constantly to increase the amount of capital laid out in a given industrial undertaking, and competition subordinates every individual capitalist to the immanent laws of capitalist production, as external and coercive laws. It compels him to keep extending his capital, so as to preserve it, and he can only extend it by means of progressive accumulation. (739)

However, competition is not a static process and leads to increased concentration and centralization² of capital:

The battle of competition is fought by the cheapening of commodities. The cheapness of commodities depends, all other circumstances remaining the same, on the productivity of labor, and this depends in turn on the scale of production. Therefore the larger capitals beat the smaller. (Marx, 1990 [1867]: 777)

Competition leads to monopolization, which creates barriers to entry on the one hand and increased competition in the other industries on the other:

It will further be remembered that, with the development of the capitalist mode of production, there is an increase in the minimum amount of individual capital necessary to carry on a business under its normal conditions. The smaller capitals, therefore, crowd into spheres of production which large-scale industry has taken control of only sporadically or incompletely. Here competition rages in direct proportion to the number, and in inverse proportion to the magnitudes, of the rival capitals. It always ends in the ruin of many small capitalists, whose capitals partly pass into the hands of their conquerors, and partly vanish completely. (Marx, 1990 [1867]: 777)

For the Monthly Review School, the rise of giant corporations and the concomitant monopoly pricing power result in increased profits that cannot be profitably re-invested as monopolization leads to overaccumulation, chronic overcapacity, and a scarcity of profitable investment outlets (Baran and Sweezy, 1966; Sweezy, 1946 [1970]). This process generates a stagnation tendency as investment weakens with respect to total available funds for investment, and this tendency can only be averted by increased sales effort, military spending, and financialization. Foster and McChesney (2012), similarly, contend that the tendency towards monopolization in the 21st century increased corporate profitability. While this approach sees a continuous decline in competitive intensity, Brenner (2006) argues that expanded and excess competition in world markets, coming especially from German and Japanese firms, led to declining profitability among U.S. firms. In theoretical terms, both approaches expect a positive link between the degree of monopolization and profitability, while the monopoly capital approach also posits a negative relationship between monopolization and investment.

One can also identify a more dialectical approach within Marxian economics. While Dumènil and Lèvy(1993) argue that price-taking firms are ‘a fiction derived from the neoclassical analytical apparatus’ (76), Christophers (2016) contends that the notion of ‘monopoly stage of capitalism’ is equally fictional, ‘albeit … emanating from a very different analytical source’ (11). These contributions recognize that both excess competition and excess monopolization can generate problems for profitability and capital accumulation. While capitalism would lose its dynamism without adequate competition, excess monopoly power can also jeopardize investment and growth dynamics through scarcity of supply and limitations on investment and output. In Harvey (2002), capitalism ‘organically’ comes to such a balance, although this process is not fully explained, whereas Christophers (2016) argues that rules and regulations, and especially antitrust and intellectual property regulations, are the primary lever to that end:

When capital has become sufficiently overcentralized and monopolistic to threaten its own successful, profitable reproduction, antitrust law has been called upon to help restore the necessary degree of balance. This balance will never be perfect and at rest; in a dialectical relation, such as that between monopoly and competition, it never can be. When the dangerous excesses has been of competition, by contrast, IP law has come to the rescue. (11–12)

Crotty (2017), similarly, emphasizes that ‘[f]or a core industry to remain viable and be dynamically efficient, its key firms must strike a balance between competition and cooperation. Too much competition can destroy the industry rather quickly, while too little will ensure its long-run decline’ (254). He identifies two competitive regimes in the post-war U.S. economy: The co-respective competition regime of 1950–1970, characterized by high profitability, low uncertainty, financial robustness, and long-term investment strategies; and a regime of coercive competition since 1980, characterized by lower profitability, stagnant aggregate demand and forced capital-deepening, labour-saving, cost-cutting investment. As such, debt-financed investment increased during the 1980s, despite declining profitability due to intensified competition. While it is possible to see high investment in both regimes, this investment may vary between capital widening and capital deepening types.

To summarize, mainstream approaches associate declining competitive intensity with higher pricing power and profitability, while the impact on investment is ambiguous. While similar ambiguity exists in post-Keynesian theories, competition need not be captured by the degree of concentration, markups may depend on cost-structure competition, and increased profitability due to monopoly power may generate further investment financed by these profits. In Marxian approaches, the monopoly capital school associates increased monopolization with rising profits and falling investment, while other strands contend that, while monopolization may increase profitability, excess monopolization may also undermine the conditions of high profitability. Similarly, less intense competition allows firms to invest in long-term capacity building whereas more intense competition may lead to capital-saving investment. Each approach expects that declining competition generates increasing profits (at least for a while), whereas predictions regarding investment are more ambiguous.

Trends in industrial concentration since 1997

Data and definitions

We draw on two sources of data in our empirical analysis. First, we describe concentration using the U.S. Census Bureau’s Economic Census for 1997, 2002, 2007, and 2012, which reports CRn ratios measuring the revenue share of the n largest firms in an industry.³ We also use industry-level revenue from the census.Industry definitions are based on the six-digit North American Industry Classification System (NAICS) codes: the first two digits denote the broad sector, the third digit denotes the subsector, and the fourth digit denotes the narrower industry group. We focus on three-digit classifications, although our main conclusions are robust to four-digit definitions.⁴ We use these data in ‘Sectoral and industry-level patterns in the level of concentration’ and ‘Sectoral patterns in the evolution of concentration’ sections to describe the evolution of concentration at the aggregate level and across industries.

Second, we merge these data with an annual panel of firm-level balance sheet and income statement data for publicly traded corporations from Compustat. To clean the Compustat data we eliminate duplicates and drop firms with negative sales, total assets, or capital. We also drop firms incorporated outside the U.S., firms in finance and real estate, and those in sectors for which the census does not provide concentration data.⁵ Drawing on the discussion above, we calculate firm-level measures of the after-tax profit rate, markup, and investment rate, as well as capital and intangible intensity (see Appendix Table A1 for definitions). Finally, we assign concentration statistics from the census to all firms in each three-digit industry in the fiscal year in which the census was conducted (1997, 2002, 2007, or 2012), and to a two-year band before and after the census year (e.g. 1997 CR4 data are assigned to firms in 1995–1999).⁶ Doing so yields a firm-level panel of 107,867 observations between 1995 and 2014. In the ‘Concentration, profitability, and investment’ and the ‘Capital and intangible intensities’ sections, we use these data to describe relationships between concentration and profitability, markups, investment, capital intensity, and intangible intensity.

Two factors drive our use of census-based CRn ratios to measure concentration. First, the Herfindahl–Hirschman Index is only available from the census for industries in the manufacturing sector, whereas CRn ratios are available for all major economic sectors. Second, census-based measures have key advantages over similar, annual concentration measures calculated with Compustat (Ali et al., 2008; Keil, 2017b). Most notably, the census covers both public and private firms, whereas Compustat only includes publicly listed companies. The census also accounts for foreign companies’ domestic sales and excludes domestic companies’ foreign sales. Finally, because census measures are constructed at the establishment level, whereas Compustat data are consolidated at the corporate level, the classification of conglomerates is more precise in the census, which groups each division’s sales with the standalone firms in that industry (Grullon et al., 2019).

Sectoral and industry-level patterns in the level of concentration

Table 1 presents weighted means and medians of CRn ratios for the top 4, 8, 20, and 50 firms in an industry over time and shows that a substantial increase in concentration took place after 1997, primarily during the early 2000s. This rise in concentration is evident in both the weighted mean and the median across all CRn measures. The average CR4 ratio across three-digit industries increases 26.8% over this period. Much of this increase takes place between 1997 and 2002, when the mean CR4 ratio increases from 14.6 to 18.1%, then remaining relatively steady at 18.3%. This trend is consistent, for example, with a peak in the number of U.S. public firms in 1997, which was followed by a substantive increase in delisting rates among seasoned firms, driven largely by merger activity (e.g. Doidge et al., 2017; Kahle and Stulz, 2017).

Table 1.

Change in concentration ratios over time across three-digit industries.

	1997	2002	2007	2012	Per cent change (1997–2012)
Median
CR4	14.3	16.8	17.3	16.3	14.0
CR8	21.5	25.0	24.7	24.7	14.9
CR20	31.2	37.2	36.4	36.1	15.7
CR50	41.5	47.0	46.7	47.5	14.5
Weighted mean
CR4	14.6	18.1	18.2	18.5	26.8
CR8	20.6	24.8	25.1	26.1	27.1
CR20	29.6	34.1	34.5	36.0	21.9
CR50	39.3	43.4	44.0	45.6	16.1

Source: Authors’ calculations from the Census Bureau.

Note: Weighted means are across-industry averages weighted by industry revenue. See ‘Data and definitions’ section for variable definitions.

Do specific industry-level patterns underlie this increase in aggregate concentration? While an across-industry increase in concentration has been clearly established in the existing literature, less attention has been paid to the specific industries and sectors where concentration has increased. We, therefore, turn to the industrial composition of rising concentration to, first, describe the set of industries with high or low concentration and, second, consider the extent to which rising average concentration reflects a widespread change across industries versus a shift within specific sectors. We start by identifying high- and low-concentration three-digit industries. For both theoretical and empirical reasons, we focus primarily on the level of, rather than growth in, concentration. From a theoretical perspective, the level of concentration, rather than the change, captures the degree of market power within an industry. Consider, for example, the hypothesis that firms in highly concentrated industries charge higher markups. This hypothesis describes a persistent feature of market power in highly concentrated industries, not a transient feature of the period when concentration rises. In contrast, analyses focused on changes in concentration, as in much of the recent literature, suggest that industries charge higher markups during the period when concentration rises, with the resulting implication that – after settling at a higher level of concentration – this market power disappears.

In turn, the most concentrated industries do not come to the forefront of empirical analyses of concentration growth. Consider, for example, the retail sector, which is comprised of 12 three-digit industries. All 12 retail industries become more concentrated after 1997, with CR4 growth ranging from 32.2% (general merchandise stores) to 212.5% (motor vehicle parts and dealers). Motor vehicle parts and dealers have, in fact, the highest per cent CR4 growth among three-digit industries. However, because of its very low initial CR4 ratio (1.6% in 1997), it remains one of the least concentrated industries in 2012. In contrast, while general merchandise stores have the lowest CR4 growth among retail industries, it is the second-most-concentrated industry in 2012. The reason lies in a much higher initial level of concentration (55.9%). Thus, while more than 70% of three-digit industries become more concentrated between 1997 and 2012, understanding the industrial and sectoral nature of monopoly power in the U.S. economy requires identifying the industries with high versus low levels of concentration.

We begin by classifying each industry as having high, medium, or low concentration in each census year: First, we categorize each three-digit industry as having above- or below-average concentration by comparing its CR4 ratio in each census year to the (weighted) average CR4 ratio across three-digit industries in 2012 (18.5%). In other words, we use a time-invariant cutoff of 18.5% to distinguish ‘high’ from ‘low’ levels of concentration, while allowing industries that become more or less concentrated to switch between the high and low category over time. Second, we account for the wide range of CR4 ratios among above-average concentration industries (from 19.4 to 88.0% in 2012) by dividing above-average concentration industries into high- and mid-concentration groups. We do so using the midpoint of CR4 ratios within the above-average concentration group (53.7%). Last, because the 2012 CR4 ratio for one industry (hospitals) lies just below this cutoff (53.3%), and is followed by a substantial drop in the CR4 ratio of the next most-concentrated industry (to 46.3%), we also assign hospitals to the high-concentration group. Via this process, we designate each three-digit industry as having a high, medium, or low level of concentration in each census year. Specifically, we classify industries with CR4 ratios greater than or equal to 53.3% as high-concentration industries, those with CR4 greater than or equal to 18.5% but less than 53.3% as mid-concentration industries, and those with below-average CR4 (less than 18.5%) as low-concentration industries (see Appendix Tables A2 and A3 for lists of these industries).

Clearly, these cutoffs can be defined in various ways and it is important to note that the conclusions drawn in the discussion below are robust to alternative cutoffs. The discussion is also robust to moving hospitals to the mid-concentration group. In other words, the discussion below does not hinge on these specific cutoffs; instead, they are simply an organizing device to summarize patterns across industries and over time.⁷

Seven of the 67 three-digit industries are designated ‘highly concentrated’ in 2012. Of these seven industries, the top six also experienced concentration growth between 1997 and 2012. In turn, our classification designates 25 industries as mid concentration and 35 as low concentration. Notably, these above- and below-average concentration industries lie in different sectors of economic activity: above-average concentration industries are more commonly in the retail and information sectors, whereas low-concentration industries lie primarily in manufacturing and wholesale. To facilitate this sectoral interpretation, Panel A of Table 2 summarizes the sectoral composition of low-, mid-, and high-concentration industries in 2012. Columns 1–3 report the number of three-digit industries within each concentration group and two-digit sector, as well as these industries’ share of sample revenue. To contextualize each two-digit sector’s size in the broader sample, Column 4 then presents this information for the full sample. For example, Column 3 of Panel A indicates that, of the 35 low-concentration industries in 2012, more than one third are in manufacturing. These 13 low-concentration manufacturing sectors comprise 13.9% of total sample revenue. In turn, Column 4 shows there are a total of 21 three-digit manufacturing industries, earning 23.1% of total revenue. Thus, 60.1% of manufacturing revenue accrues to low-concentration industries.

Table 2.

High-, mid-, and low-concentration industries in 2012 and 2007 by two-digit sectors of activity.

NAICS	Two-digit sector	(1) High CR4		(2) Mid CR4		(3) Low CR4		(4) Full sample
NAICS	Two-digit sector	N	% of rev	N	% of rev	N	% of rev	N	% of rev
Panel A: 2012
22	Utilities	0	–	0	–	1	2.2	1	2.2
31–33	Manufacturing	0	–	8	9.3	13	13.9	21	23.1
42	Wholesale	0	–	1	2.6	2	29.5	3	32.1
44–45	Retail	3	4.1	7	7.2	2	5.8	12	17.1
48–49	Transportation and warehousing	2	1.0	4	0.6	3	1.4	9	3.0
51	Information	1	2.2	4	2.4	1	0.4	6	5.0
54	Prof., sci. and technical services	0	–	0	–	1	5.9	1	5.9
56	Administrative and support services	0	–	1	0.3	1	2.4	2	2.7
61	Educational services	0	–	0	–	1	0.2	1	0.2
62	Health care and social assistance	1	0.4	0	–	3	3.7	4	4.1
71	Arts, entertainment, recreation	0	–	0	–	3	0.8	3	0.8
72	Accommodation, food services	0	–	0	–	2	2.9	2	2.9
81	Other services	0	–	0	–	2	0.9	2	0.9
	Total, 2012	7	7.8	25	22.4	35	69.8	67	100
Panel B: 1997
22	Utilities	0	–	0	–	1	3.1	1	3.1
31–33	Manufacturing	0	–	7	11.0	14	17.5	21	28.5
42	Wholesale	0	–	0	–	2	30.2	2	30.2
44–45	Retail	1	2.5	6	6.4	5	9.4	12	18.3
48–49	Transportation and warehousing	1	0.3	3	0.5	5	1.5	9	2.4
51	Information	0	–	3	3.3	0	–	3	3.3
54	Prof., sci. and technical services	0	–	0	–	1	4.3	1	4.3
56	Administrative and support services	0	–	1	0.3	1	1.9	2	2.2
61	Educational services	0	–	0	–	1	0.1	1	0.1
62	Health care and social assistance	1	0.3	0	–	3	2.8	4	3.1
71	Arts, entertainment, recreation	0	–	1	0.0	2	0.6	3	0.6
72	Accommodation, food services	0	–	0	–	2	2.6	2	2.6
81	Other services	0	–	0	–	2	1.2	2	1.2
	Total, 1997	3	3.1	21	21.6	39	75.4	63	100

Source: Authors’ calculations from the Census Bureau.

Note: Columns 1–3 list the number (N) and revenue share (% of rev) of all three-digit industries in each two-digit economic sector for the high-, mid-, and low-concentration industries listed in Tables A2 and A3. Revenue share is total revenue of all three-digit industries within both a particular concentration group and two-digit sector relative to total sample revenue in 2012 (Panel A) and in 1997 (Panel B). Column 4 lists the total number of three-digit industries and each sector’s revenue share. The row totals for ‘N’ and ‘% of rev’ in Columns 1–3 equal the values in Column 4. See ‘Data and definitions’ section for variable definitions and ‘Sectoral and industry-level patterns in the level of concentration’ section for details on the classification of industry groups.

Table 2 highlights the importance of the retail and information sectors among above-average concentration industries. Retail comprises approximately one third of above-average concentration industries and 38.5% of their revenue. Narrowing in on the high-concentration industries, Column 1 shows that three of the seven highly concentrated industries are in retail. These three industries comprise 52.6% of highly concentrated industries’ revenue in 2012. A lot of well-known firms are in these industries.For example, general merchandise stores, which has the second-highest level of concentration in 2012 and is the largest high-concentration industry, includes Walmart, Target, and Costco. In the mid-concentration group, Amazon is the top firm by total assets among non-store retailers (Appendix Table A2, row 21).

Table 2 also shows that the vast majority of information-services activity in 2012 took place in above-average concentration industries.Five of the six information-related industries have above-average concentration: telecommunications (e.g. AT&T and Verizon), broadcasting, other information services (e.g. eBay and Alphabet Inc.), motion picture and sound recording industries (e.g. Time Warner), and publishing industries (e.g. Microsoft). These five industries have a small weight in the overall economy (4.4% of total revenue), but dominate the information sector, constituting 92% of information-related revenue. Furthermore, one of these industries – telecommunications – lies in the high-concentration group, with a 2012 CR4 ratio of 55.8%. Telecommunications is notable in that it accounts not only for almost half of information-services revenue, but also has a heavy weight among the group of highly concentrated industries, accounting for 28.2% of these industries’ revenue. As we discuss below, telecommunications firms also stand out for high markups. Finally, note that the only information-related industry in the low-concentration group in 2012 (data processing, hosting, and related services) is among the most concentrated of these industries (see Appendix Table A3).

At first glance, Columns 1 and 2 of Table 2 also suggest that transportation and warehousing, and manufacturing are strongly represented among above-average concentration industries. Not only are two of the seven high-concentration industries in transportation and warehousing, but the airline industry also registers remarkable CR4 growth of 176.6% after 1997. However, the transportation sector accounts for a small share of the economy (2.9% of sample revenue), and only 5.2% of above-average concentration industries’ revenue. Finally, while approximately one quarter of above-average concentration industries are in manufacturing, a larger share of manufacturing activity falls within the low-concentration group, in terms of both the number of industries and revenue.

Specifically, low-concentration industries lie, in revenue terms, primarily in manufacturing and wholesale, with almost 20% of these industries’ revenue deriving from manufacturing and 42.2% from wholesale.⁸ Even more importantly, these shares are large relative to the size of the manufacturing and wholesale sectors: 60.1% of manufacturing and 91.9% of wholesale revenue accrues to low-concentration industries. Particularly in the case of manufacturing, this observation suggests another important sectoral dimension to rising concentration: because manufacturing industries are less concentrated than average, the declining share of manufacturing within the economy (from 17.5 to 13.9% of revenue between 1997 and 2012) implies upward pressure on the weighted average level of concentration across three-digit industries. This evidence of relatively low concentration in manufacturing raises questions about the implications of international competition for the dynamics of domestic concentration measures as manufacturing industries are likely to be more affected by imports than in sectors like retail.

Last, Table 2 indicates that these high-, mid-, and low-concentration industry groups have different weights in the economy. High-concentration industries constitute 7.8% of sample revenue and mid-concentration industries constitute 22.4%, whereas the low-concentration group constitutes 69.8% of our sample. The small weight of highly concentrated industries raises questions about whether concentration ratios themselves explain macroeconomic trends in high markups and profitability or low investment. We consider these issues further in the ‘Concentration, profitability, and investment’ and the ‘Capital and intangible intensities’ sections.

Sectoral patterns in the evolution of concentration

The discussion up to this point isolates sectoral differences in the level of concentration in 2012, wherein retail and information technology industries tend to have above-average concentration, while manufacturing and wholesale play outsize roles among low-concentration industries. To conclude the industry-level description of rising concentration, we consider if these sectoral patterns changed after 1997.

Table 2 begins by comparing the sectoral composition of low-, mid-, and high-concentration industries in 1997 and 2012 by reproducing Panel A for the first year of our sample (1997) in Panel B.⁹ Together, Panels A and B of Table 2 show that – consistent with rising average concentration after 1997 – the number of high- and mid-concentration three-digit industries rises from 3 to 7 and 21 to 25, respectively. Furthermore, Table 2 suggests that the retail and information sectors are important not only for describing the level of concentration in 2012, but also the increasing incidence of highly concentrated industries since 1997. There is both a net increase of three retail industries among mid- and high-concentration industries and an increase in the share of retail-related revenue accruing to industries with mid- or high-concentration, from 48.6 to 66.1% – even as the overall size of retail is relatively stable (falling slightly from 18.3 to 17.1% of revenue). Both the number and revenue share of information-services industries also increase among the mid- and high-concentration groups, from three in 1997 (with 3.3% of revenue) to five by 2012 (4.6% of revenue).¹⁰

Concurrent changes in the number of low-, mid-, or highly concentrated industries within a sector and in sector size make it difficult, however, to isolate a sector’s role in driving aggregate changes in concentration from Table 2 alone. To further clarify if specific sectors underlie rising average concentration, we therefore present two final sets of calculations in Table 3. First, we record the average change in concentration across all three-digit industries within each two-digit sector (Panel A). Second, we calculate the per cent change in average CR4 between 1997 and 2012 when sequentially excluding each of the 13 two-digit sectors (Panel B). Panel B allows us to consider each sector’s contribution to the aggregate increase in concentration while accounting for that sector’s change in concentration and size. We summarize this information in the final column of Panel B, which records the ratio of the per cent change between 1997 and 2012 when excluding the sector (for utilities, for instance, 27.4%) to the per cent change across all sectors (26.8%). When this ratio is equal to 1, it suggests the excluded sector does not exert substantial pressure on average concentration (due either to small size or because its change in concentration lies near the average). When the ratio is less than 1, it indicates that the overall increase in average concentration shrinks when excluding the sector in question, such that the sector’s inclusion yields a larger increase in average concentration. In other words, the excluded sector helps explain rising average concentration. If the ratio is greater than 1, excluding the sector in question yields a bigger change in concentration than observed for the full sample, such that the excluded sector depresses average concentration growth.

Table 3.

What sectors drive the increase in concentration? (Evolution of CR4 ratio when sequentially dropping sectors of economic activity).

		Average CR4, 1997 (%)	Average CR4, 2012 (%)	Per cent change (1997–2012)	(Per cent change with sector excluded)/(overall per cent change)
Panel A	Two-digit sector:
1.	Utilities	14.7	15.0	2.1	–
2.	Manufacturing	19.8	21.6	14.5	–
3.	Wholesale	7.9	11.0	17.9	–
4.	Retail	19.1	31.0	87.1	–
5.	Transportation and warehousing	21.8	30.9	62.6	–
6.	Information	32.0	41.7	69.2	–
7.	Prof., sci. and technical services	4.2	4.1	−0.7	–
8.	Administrative and support services	8.4	10.9	16.9	–
9.	Educational services	5.5	6.9	9.6	–
10.	Health care and social assistance	10.6	10.5	−0.9	–
11.	Arts, entertainment, and recreation	8.0	7.9	−0.9	–
12.	Accommodation and food services	9.0	8.9	−0.8	–
13.	Other services	5.3	5.0	−2.4	–
Panel B	Excluded two-digit sector:
1.	Utilities	14.6	18.6	27.4	1.02
2.	Manufacturing	12.5	17.6	40.4	1.51
3.	Wholesale	17.5	22.1	26.0	0.97
4.	Retail	13.6	16.0	17.3	0.65
5.	Transportation and warehousing	14.4	18.2	25.8	0.96
6.	Information	14.0	17.3	23.5	0.88
7.	Prof., sci. and technical services	15.1	19.4	28.9	1.08
8.	Administrative and support services	14.8	18.8	27.1	1.01
9.	Educational services	14.6	18.6	26.9	1.00
10.	Health care and social assistance	14.7	18.9	28.1	1.05
11.	Arts, entertainment, and recreation	14.7	18.6	27.1	1.01
12.	Accommodation and food services	14.8	18.8	27.5	1.03
13.	Other services	14.7	18.7	26.7	1.00
	All sectors	14.6	18.5	26.8

Source: Authors’ calculations from the Census Bureau.

Note: Panel A records the weighted average CR4 across three-digit industries in each two-digit sector in 1997 and 2012, and the per cent change from 1997 to 2012. Panel B records the average CR4 ratio across three-digit industries in all sectors, but excluding the listed sector, in 1997 and 2012, and the per cent change from 1997 to 2012. The final column of Panel B records the ratio of the per cent change from 1997 to 2012 when excluding the sector in question, to the per cent change for all sectors (26.8%). The bottom row shows weighted average CR4 across all three-digit industries. See ‘Data and definitions’ section for variable definitions.

Table 3 highlights three main points. First, there is a broad-based increase in concentration across the economy, with average CR4 increasing in 8 of 13 sectors after 1997. Furthermore, the five sectors with declining CR4 not only experience small-scale declines in concentration (particularly in comparison to sectors with rising concentration), but are also small in revenue terms (see Table 2). Thus, there is a notable within-sector dimension of rising concentration in the U.S. economy.

Second, Table 3 reiterates the role of retail and information services in rising average concentration. Average CR4 across retail industries grows a striking 87.1%, from 19.1% in 1997 to 31.0% in 2012, followed by information (with a 69.2% increase in concentration), and transportation and warehousing (with a 62.6% increase). Furthermore, Panel B highlights that, when excluding retail, average concentration growth falls from 26.8% (for the full sample) to 17.3%. The ratio of the per cent change in average CR4 when excluding retail to the per cent change across all sectors is, accordingly, substantially less than 1 (0.65). Put differently, by including retail we observe a substantially greater increase in concentration. As the size of the retail sector is quite stable, these calculations point to rising concentration within retail (rather than growth of an already-concentrated sector). Information, which has become an increasingly important sector of activity in the US economy, similarly puts upward pressure on overall concentration growth, although to a smaller extent than retail.

Third, Table 3 gives insight into sectors that offset rising average concentration. Of particular note is manufacturing: excluding manufacturing yields far greater growth in average CR4 (40.4%, as compared to the observed 26.8% across all industries). Manufacturing is, in fact, the only sector whose exclusion yields a value in the final column of Table 3 substantially larger than 1. Thus, while the CR4 ratio across manufacturing industries does rise, from 19.8 to 21.6%, it rises less quickly than average. These observations build on the discussion of manufacturing from Table 2, above. On the one hand, manufacturing industries tend to have below-average concentration, which has risen since 1997, but less quickly than the average. These trends offset the increase in aggregate concentration. On the other hand, as manufacturing shrinks, the weight of this relatively low-concentration sector falls, pushing up average concentration across the economy. Finally, aside from retail, information, and manufacturing, all remaining sectors have values in Column 4 close to 1, indicating they do not exert substantial pressure on the full-sample trend.

Concentration, profitability, and investment

We now turn to the relationship between concentration and firm-level profitability, markups, and investment. In line with the divergent predictions of the theoretical literature, we show relationships that are more complex than a standard narrative equating high concentration with high profit rates, high markups, and low investment suggests, indicating that – in some industries – higher concentration comes with intensified competition.

Figure 1(a) begins by dividing all firms into industries with above-average concentration and below-average concentration, and plotting the weighted average profit rate across firms in each group. It shows the expected pattern, wherein firms in more concentrated industries have higher average profitability than firms in less-concentrated industries. Specifically, over these years, firms in above-average concentration industries have an average profit rate of 31.1% versus 26.7% in below-average concentration industries.

Figure 1.

Average profit rates in low-, mid-, and highly concentrated industries.

However, Figure 1(b), which disaggregates above-average concentration industries into the mid- and high-concentration groups from the ‘Sectoral patterns in the evolution of concentration’ section, shows that Figure 1(a) masks important heterogeneity within these groups. Most importantly, firms in the most concentrated industries are not, on average, the most profitable.¹¹ Instead, the highest average profitability accrues to firms in mid-concentration industries (31.8% across all years), whereas the profitability of firms in the highest-concentration industries (29.1%) lies close to that of firms in low-concentration industries. Notably, this non-monotonic pattern does not simply reflect unobserved industry-specific factors obfuscating across-industry averages within these three groups: even when accounting for industry fixed effects, the correlation between profitability and concentration is close to zero. For example, a simple OLS regression describing the raw correlation between average profitability and industry CR4 is small, but negative and statistically significant (−0.002 with a robust standard error of 0.004). In turn, the within-industry correlation, which also includes industry fixed effects that absorb industry-specific level differences, becomes somewhat positive but statistically insignificant (0.002, with a robust standard error of 0.0024).

This weak relationship is consistent with the early empirical literature on concentration, which finds a ‘paradoxically’ statistically weak and economically small link between concentration and profitability (see Schmalensee, 1988 for a review of early empirical studies). While a recent paper by Grullon et al. (2019) does find a positive correlation between profitability and concentration (using concentration measures based on Compustat), other studies continue to find a relationship that is unstable over time and space and that sometimes disappears in multivariate studies (Keil, 2017a, 2019; Melmiès, 2016). In turn, two points are particularly interesting about our results here: (1) they suggest a non-linear/non-uniform link, rather than no link at all; (2) the industry description in the ‘Trends in industrial concentration since 1997’ section shows that highly concentrated industries lie in specific sectors (namely retail), in which undercutting pricing strategies may cut into profit margins. In discussing the ‘concentration-profitability paradox’, Keil (2017a), similarly, emphasizes that maintaining market power can be costly in terms of the excess capacity and low pricing strategies needed to discourage entry (e.g. Spence, 1977).¹²

To further investigate this issue, we turn to simple markups.¹³ Unlike other variables in the ‘Concentration, profitability, and investment’ and the ‘Capital and intangible intensities’ sections, however, markups are sensitive to the cutoffs defining low-, mid-, and high-concentration industries, suggesting variation in markups across the distribution of concentration ratios. To highlight this heterogeneity in markups across the distribution of concentration ratios we, therefore, show average markups across firms in each decile and ventile of CR4 ratios (rather than across low-, mid-, and high-concentration groups) in Figure 2. The top panel of Figure 2 highlights that the highest-markup firms fall at the lower bound of the high-concentration group, with CR4 ratios between 50 and 60%. While this could be considered evidence of high markups among firms in high-concentration industries, we also observe the lowest markups among firms within the group where CR4 is greater than 80%, although – notably – only one industry (couriers and messengers) falls into this decile.¹⁴ There is also a second cluster of high-markup firms in industries with CR4 ratios between 10 and 20%, driving up average markups within the low-concentration industries. Aside from these two groups, markups are quite constant across concentration levels. The same analysis with finer increments reiterates this pattern, with the highest markups accruing to firms in industries with CR4 ratios from 55 to 60% and 15 to 20%.

Figure 2.

Average markups across deciles of the CR4 ratio.

Which industries charge these high markups? Across firms with CR4 ratios from 55 to 60%, high markups reflect two industries: health and personal care stores, and telecommunications. Similarly, while there is a larger and more heterogeneous group of industries with CR4 ratios between 15 and 20%, information services again stand out among this group of firms. Three industries in this group have notably high markups: publishing industries (except internet); data processing and hosting; and electronic equipment, appliance, and component manufacturing. The first two are in information services and, while the third falls in manufacturing, it also has a technology orientation. In contrast, while retail industries tend to have above-average concentration, retail firms’ markups are generally below average. As such, even as markups increase within most sectors over this time, the overall relationship between concentration and average markups is strongly mediated by the specific industries that are, or are not, highly concentrated.

Next, we compare the investment rates of firms in high-concentration industries to those in lower-concentration industries. Investment is defined as capital expenditures relative to the fixed capital stock (the net stock of property, plant, and equipment). Again using the cutoffs from the ‘Sectoral patterns in the evolution of concentration’ section, Figure 3(a) first shows average investment for firms in industries with above- and below-average concentration ratios. This figure shows that investment rates are, in fact, higher for firms in above-average-concentration industries, averaging 19.1% over this period, as compared to 13.5% among firms in low-concentration industries. Notably, however, there is a sharp decline in average investment rates of firms in above-average concentration industries during the early 2000s, when the bulk of the increase in concentration took place. In turn, disaggregating the above-average-concentration group into high-concentration and mid-concentration groups (Figure 3(b)), shows an increase in investment among firms in mid-concentration industries in the late 1990s, followed by a sharp decline in the early 2000s. For the rest of the period, investment rates of firms in all industrial concentration groups largely move together, while they are higher for the firms in the high-concentration group in the 2000s and in the mid-concentration group in the 2010s.

Figure 3.

Average investment rates in low-, mid-, and highly concentrated industries.

Capital and intangible intensities

The ‘Concentration, profitability, and investment’ section shows that more concentrated industries do not necessarily have higher profitability or markups or lower investment rates. One potential explanation for comparatively lower profitability in high-concentration industries could lie in capital intensity. If these industries are more capital intensive, relative to low- and mid-concentration industries, then their profitability could be pulled down by the large denominator. There are economic reasons to expect these industries have higher capital intensity: Steindl (1952), for example, argues that firms in concentrated markets can maintain more excess capacity as a barrier to entry and the ‘average cost of larger equipment with excess capacity is smaller than average cost of smaller equipment with full capacity. So that long run cost curve declines’ (10). Could this mean that firms in our most-concentrated group have larger capital stocks, such that profits measured relative to capital are lower? Figure 4 explores this possibility by comparing average capital intensity, defined as fixed capital divided by output, across firms in low-, mid-, and high-concentration industries.¹⁵ Figure 4(a) shows that, contrary to the hypothesis laid out above, capital intensity in above-average concentration industries is lower than in below-concentration industries, and is declining over time. Figure 4(b), furthermore, shows thatmid-concentration industries generally have lower capital intensity and, while capital intensity in the high-concentration group rises from 1995 to 2000, this group’s capital intensity then falls sharply below that of the low-concentration group.

Figure 4.

Average capital intensity in low-, mid-, and highly concentrated industries.

If capital intensity does not explain the relatively lower profitability of highly concentrated industries, can it instead explain higher profitability in low/mid-concentration industries? One notable possibility is that firms in certain industries increasingly hold intangible assets, like intellectual property rights (e.g. patents, copyrights, trademarks, brand names) that may increase their market power and, hence, profitability (Orhangazi, 2019; Pagano, 2014). We, therefore, explore if firms in low-concentration industries tend to make greater use of intangible assets. We define intangible-asset intensity using Compustat estimates of intangible assets net of goodwill relative to fixed capital, and describe the evolution of intangible intensity over time for firms in industries with different levels of concentration. While Figure 5 does not show a clear distinction between low- and high-concentration industries in terms of intangible intensity, we observe that average intangible asset use is now higher in all industries compared with the second half of the 1990s. Therefore, it is quite possible that even in industries with low-concentration ratios, firms may be gaining market power through increased use of intangibles and hence earning higher profits.

Figure 5.

Average intangible intensity across firms in low-, mid-, and highly concentrated industries. Source: Authors’ calculations from Compustat and Census.

Discussion and concluding remarks

A few concluding remarks are in order. First, an increase in average concentration has taken place across U.S. industries between 1997 and 2012, with much of it occurring in the late 1990s and the early 2000s. While this average increase in concentration has an important within-industry dimension, wherein a majority of industries become more concentrated after the late 1990s, a notable share of concentration growth is in fact driven by industries in the retail and information-services sectors. Second, our theoretical discussion captures considerable ambiguity in the expected relationships between concentration and profitability or investment. In turn, our empirical findings, which do not show a uniform relationship between the level of industrial concentration and profitability, markups, or investment rates, are in line with this theoretical ambiguity. Highly concentrated industries are not the most profitable (instead, mid-concentration industries earn the highest profit rates) and, with a couple of sector-specific exceptions – namely, in information services – do not have the highest markups.

Third, the absence of a uniform andmonotonic relation suggests at least three different cases. (i) First, there are industrieswhose behaviour is consistentwith a 'standard' story inwhich firms in industries with high (low) levels of concentration have high (low) profit rates, high (low) markups, and low (high) rates of investment. In this case, high concentration is indicative of low competition, allowing firms in highly concentrated industries to capture monopoly profits. The analysis suggests that information-services firms may fit within this case. This case is particularly notable given the increasing centrality of the information-services sector to economic activity in the U.S. (ii) In turn, there are highly concentrated industries in which firms have, on average, low profitability, low markups, and average investment. Retail, as well as some highly concentrated industries in the transportation sector (e.g. airlines), may fit this scenario. In notable contrast to the ‘standard’ story, this scenario suggests that there are cases in which increased monopolization goes hand in hand with intensified competition. As such, while it has been argued that increased concentration can push down prices when rising concentration derives from productivity increases, this case suggests it is also quite possible that higher concentration pushes down prices by generating more intense competition among large, dominant firms. This possibility may be especially likely to play out in the absence of a co-respective competitive regime in these industries, wherein firms compete intensely to guarantee or increase their market shares, and/or to drive out competitors. If, in particular, the threat of entry is high, then this state of intense competition can curtail high-concentration industries’ markups and profitability. Furthermore, the irreversibility of investment, whichmakes exit costly, means that monopolization can trap firms with below-average profit rates. (iii) Last, the set of low- and mid-concentration industries in which firms earn high profits and markups suggest that industrial concentration is not the only form of market power. In these industries, firms with small market shares may still have market power, for instance through intangible assets, that allows them to increase their markups and profitability. The different experiences of industries in these three cases suggest that key aspects of the dynamics of rising concentration may be missed by aggregate-level analyses, and point to a role for industry- and sector-specific analyses of changing concentration in the U.S.

Finally, it is relevant to note that at least two shortcomings of the data may also be important in explaining the trends we observe. These limitations, in turn, suggest directions for future research. First, while the industry-level census data does account for the domestic market share of foreign firms, we do not explicitly consider the effects of international competition. If the degree of international competition is high in certain industries, then markups may be low despite high domestic concentration, as higher imports may compensate for higher domestic concentration. In fact, rising international competition may also drive increased domestic concentration, as smaller units may be unable to withstand to international competition. Similarly, demand constraints may prevent some firms in high-concentration industries from charging higher markups. Second, industry-level concentration data may not always delineate relevant markets from the perspective of competition, as these markets could be regional/local or product-based (rather than industry-based). In particular, firms in some mid-concentration industries with high markups and profitability may have geographically local monopolies. Similarly, while we assign each firm to a single industry, large firms may be important players in product markets spanning multiple industries. Industry-level measures of concentration ratios may, therefore, underestimate these firms’ degree of market power.

Footnotes

Acknowledgements

The authors would like to thank Enes Işık and Çağdaş Yalçınkaya for research assistance; Cecilia Rikap, Harry Konstantinidis, Armağan Gezici, the participants of the ‘Profit rate and economic cycles’ panel at the 2019 Eastern Economic Association conference, and two anonymous referees for their valuable comments on earlier drafts of this paper.

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) received no financial support for the research, authorship, and/or publication of this article.

Notes

References

Ali

Klasa

Yeung

(2008) The limitations of industry concentration measures constructed with Compustat data: Implications for finance research. Review of Financial Studies 22(10): 3839–3871.

Autor D, Dorn D, Katz LF, et al. (2020) The fall of the labor share and the rise of superstar firms. Quarterly Journal of Economics 135(2): 645–709.

Autor

Dorn

Katz

et al. (2017) Concentrating on the fall of the labor share. American Economic Review 107(5): 180–185.

Baran

Sweezy

(1966) Monopoly Capital. New York: Modern Reader Paperbacks.

Barkai

(2016) Declining Labor and Capital Shares. Stigler Center for the Study of the Economy and the State New Working Paper Series no. 2. Chicago, IL: Stigler Center for the Study of the Economy and the State.

Bork

(1979) The Antitrust Paradox: A Policy at War with Itself. New York: Free Press.

Brenner

(2006) The Economics of Global Turbulence. London: Verso.

CEA (2016) Benefits of Competition and Indicators of Market Power. Washington, DC: Council of Economic Advisers.

Christophers

(2016) The Great Leveler: Capitalism and Competition in the Court of Law. Cambridge, MA: Harvard University Press.

10.

Crotty

(2017) Capitalism, Macroeconomics and Reality: Understanding Globalization, Financialization, Competition and Crisis. Cheltenham: Edward Elgar.

11.

Crouzet

Eberly

(2018) Intangibles, investment, and efficiency. AEA Papers and Proceedings 108(5): 426–431.

12.

De Loecker

Eeckhout

(2017) The rise of market power and the macroeconomic implications. CEPR Discussion Papers no. 12221. mimeo.

13.

Doidge

Karolyi

Stulz

(2017) The U.S. listing gap. Journal of Financial Economics 123(3): 464–487.

14.

Dumènil

Lèvy

(1993) The Economics of the Profit Rate: Competition, Crises and Historical Tendencies in Capitalism. Cheltenham: Edward Elgar.

15.

Eichner

(1983) The micro foundations of the corporate economy. Managerial and Decision Economics 4(3): 136–152.

16.

Foster

McChesney

(2012) The Endless Crisis: How Monopoly-Finance Capital Produces Stagnation and Upheaval from the USA to China. New York: Monthly Review Press.

17.

Grullon

Larkin

Michaely

(2019) Are U.S. industries becoming more concentrated? Review of Finance 23(4): 697–743.

18.

Gutiérrez

Philippon

(2017a) Declining competition and investment in the U.S. NBER Working Paper no. 23583. Cambridge, MA.

19.

Gutiérrez

Philippon

(2017b) Investment-less growth: An empirical investigation. Brookings Papers on Economic Activity Fall: 89–169.

20.

Harvey

(2002) The art of rent: Globalization, monopoly and the commodification of culture. Socialist Register 38: 93–110.

21.

Hoberg

Phillips

(2016) Text-based network industries and endogenous product differentiation. Journal of Political Economy 124(5): 1423–1465.

22.

Kahle

Stulz

(2017) Is the U.S. public corporation in trouble? Journal of Economic Perspectives 31(3): 67–88.

23.

Kalecki

(1971) Selected Essays on the Dynamics of the Capitalist Economy, 1933–1970. Cambridge: Cambridge University Press.

24.

Keil

(2017a) Explaining the concentration-profitability paradox. Review of Political Economy 29(2): 209–231.

25.

Keil

(2017b) The trouble with approximating industry concentration with Compustat. Journal of Corporate Finance 45: 467–479.

26.

Keil

(2019) Is there a causal effect of concentration on persistent profitability differentials? Industrial and Corporate Change 28(2): 241–257.

27.

Kurz

(2017) On the formation of capital and wealth: IT, monopoly power and rising inequality. Available at: https://dx-doi-org.web.bisu.edu.cn/10.2139/ssrn.3014361 (accessed 11 June 2019).

28.

Lavoie

(2015) Post-Keynesian Economics: New Foundations. Cheltenham: Edward Elgar.

29.

Lee

(2013) Competition, going enterprise, and economic activity. In: Moudud

Bina

Mason

(eds) Alternative Theories of Competition: Challenges to the Orthodoxy. London: Routledge, pp.160–173.

30.

Marx

(1990 [1867]) Capital 1. London: Penguin Books.

31.

Melmiès

(2016) Effects of competition upon profit margins from a Post Keynesian perspective. In: Jo

Todorova

(eds) Advancing the Frontiers of Heterodox Economics. London: Routledge, pp.154–172.

32.

Nell

(1992) Transformational Growth and Effective Demand: Economics After the Capital Critique. London: Macmillan.

33.

OECD (2018) Market concentration – Note by the United States, DAF/COMP/WD(2018)59.

34.

Orhangazi

(2019) The role of intangible assets in explaining the investment-profit puzzle. Cambridge Journal of Economics 43(5): 1251–1286.

35.

Pagano

(2014) The crisis of intellectual monopoly capitalism. Cambridge Journal of Economics 38(6): 1409–1429.

36.

Pepall

Richards

Norman

(2014) Industrial Organization: Contemporary Theory and Empirical Applications. Malden, MA: Wiley.

37.

Robinson

(1933) The Economics of Imperfect Competition. London: Macmillan.

38.

Schmalensee

(1988) Inter-industry studies of structure and performance. In: Schmalensee R and Willig RD (eds) Handbook of Industrial Organization. Amsterdam: North Holland, pp.951–1009.

39.

Schumpeter

(1943) Capitalism, Socialism and Democracy. New York: Harper.

40.

Shapiro

(2018) Antitrust in a time of populism. International Journal of Industrial Organization 61: 714–748.

41.

Spence

(1977) Entry, capacity, investment and oligopolistic pricing. The Bell Journal of Economics 8(2): 534–544.

42.

Steindl

(1952) Maturity and Stagnation in American Capitalism. New York: Monthly Review Press.

43.

Stiglitz

(2016) The new era of monopoly is here. The Guardian. Available at: https://www.theguardian.com/business/2016/may/13/-new-era-monopoly-joseph-stiglitz (accessed 11 June 2019).

44.

Stucke

(2013) Is competition always good? Journal of Antitrust Enforcement 1(1): 162–197.

45.

Sweezy

(1946 [1970]) The Theory of Capitalist Development. New York: Monthly Review Press.

46.

Tirole

(1988) The Theory of Industrial Organization. Cambridge: MIT Press.