Quantifying Market Power and Business Dynamism in the Macroeconomy

Summary

"Quantifying Market Power and Business Dynamism in the Macroeconomy" is a September 2026 paper by Jan De Loecker (KU Leuven), Jan Eeckhout (UPF Barcelona), and Simon Mongey (Federal Reserve Bank of Minneapolis), published as Federal Reserve Bank of Minneapolis Staff Report 688 (DOI: 10.21034/sr.688), dated September 4, 2026. De Loecker and Eeckhout are two of the three authors of "The Rise of Market Power and the Macroeconomic Implications" (De Loecker, Eeckhout & Unger, QJE 2020), the paper most responsible for the reduced-form finding that firm-level markups rose sharply after 1980. This new paper is a different kind of exercise, not a restatement of that one: it builds a general-equilibrium model of oligopolistic competition with heterogeneous firms and endogenous entry, and structurally estimates it year by year on U.S. data from 1980 to 2023, in order to ask why markups rose and what that rise did to welfare. Because the paper is a Federal Reserve staff report — a working paper carrying the authors' own methodological rigor but not yet through journal peer review — its results should be read as strong but provisional evidence, not a settled finding.

The Core Argument and Findings

The model and its three channels. The paper models a large economy of small markets, each populated by a limited number of firms competing in oligopoly (Cournot in the baseline, with a Bertrand robustness check), subject to a fixed operating cost. Firms differ in productivity, and a pool of potential entrants decides each period whether to pay the fixed cost and produce. Within this structure, three underlying "primitives" can independently drive up equilibrium markups: (i) greater dispersion in the firm productivity distribution (σ) — a few firms pull far ahead of the pack; (ii) a rising share of fixed (overhead) versus variable costs (φ) — firms need to spend more before they produce a unit of output; and (iii) a decline in the number of potential competitors (M) — markets simply have fewer firms able to contest them. The authors show, through comparative statics, that these three channels have different and sometimes opposite implications for measurable data: a fall in the number of competitors raises markups but lowers business dynamism (the rate of job reallocation across firms), while a rise in productivity dispersion or fixed costs raises markups while raising reallocation — so a joint time series on markups, cost composition, and business dynamism can separately identify which forces were actually at work, without the model needing to define what a "market" is or measure market shares directly.

Headline result: all three channels are needed. Estimating the model on U.S. data from 1980 to 2023, the authors find that no single channel fits the data — "we then estimate that all three channels are necessary to explain these data between 1980 and 2023" (Abstract). Fixed costs rose, productivity dispersion widened, and the number of potential competitors fell, together. The paper reads the declining number of potential entrants as consistent with "an increase in the number of mergers and acquisitions" and "increasingly slack antitrust enforcement policy in the U.S. and around the world," citing the prior literature on this point; the model does not itself measure antitrust enforcement or merger activity, and this reading is offered as context for the estimated decline in M rather than as a direct empirical test.

Welfare and output: the net numbers. The paper's central quantitative result is that the identified 1980–2023 changes produced a 5 percent decline in welfare and a 6 percent decline in output (Abstract; Introduction, p. 2). This net number conceals a much larger set of offsetting forces: "substantial output gains of +15.5% occur due to reallocation toward more productive firms, but these are more than offset by more labor being tied up in fixed costs and increasing deadweight loss from markups" (Introduction, p. 2). In other words, the same forces that shifted output toward more efficient firms also let those firms extract more surplus from the shift than they would have under tougher competition.

The wedge decomposition. The authors formally decompose the net output and welfare changes into five "welfare-relevant aggregate wedges": productivity, selection, overhead, markups, and misallocation, with overhead, selection, and the markup wedge doing most of the work (Introduction, p. 2; Section 6, p. 33). Isolating the markup and fixed-cost wedges alone, the paper finds they "would have led to around a 20 percent decline in output: 12 percent from markups, and 8 percent from the rise in fixed costs" — a decline "half offset by the combined increase in productivity due to innate changes in the productivity distribution... and better selection... of firms that enter" (Section 6, p. 33). The net 6 percent output decline is thus a small residual sitting on top of much larger offsetting effects — a pattern the paper states directly: "there is a much larger decline that is partly offset by the increase in productivity... Firms in the tail have become more productive while incurring higher fixed costs. This has lead to fewer firms entering, and those that do, keep more of those productivity gains as profits, resulting in higher deadweight loss. The net effect is negative as more productive firms extract even more rents" (Section 6, p. 33). Welfare follows a parallel decomposition and declines by "about the same amount," 5 percent (Section 6, p. 33).

Fit against the raw data. The model's estimated markup series rises from a sales-weighted average of 1.21 in 1980 to 1.59 in 2023, a 38.2-percentage-point increase (Section 5, p. 26) — broadly consistent with, though not identical to, the QJE paper's separately-estimated 1.21-to-1.61 (1980–2016) series, since the two papers use different sample windows and (in this paper's case) time-varying output elasticities. The paper reports the labor share fell by 6.1 percentage points between 1980 and 2023 on raw annual data (4.7 points on a smoothed basis, versus 3.5 points implied by the model); wages relative to total factor productivity fell by 20 log points in both the model and the data; and employment (the prime-age employment-population ratio) shows a trend decline of 6.9 log points, of which the model accounts for "about three-quarters" (Section 5, pp. 26–27). None of these aggregate series were used to fit the model — they are targeted only by the three primitive parameters {M, σ, φ} and are matched as an independent validation check.

Robustness. The authors report the three-channel result and the welfare conclusion are robust to estimating under Bertrand rather than Cournot competition, and to targeting a cost-weighted rather than sales-weighted markup series (Abstract; Section 8).

Relation to the Georgist Case

Like the companion De Loecker, Eeckhout & Unger (2020) paper, this staff report matters to the Georgist case without being about land at all. Nowhere in its 70 pages, appendix included, does the paper mention land, real estate, location or rent-seeking; it is entirely silent on real estate, location, or land rent as a source of the market power or entry barriers it models. Rather than land-specific evidence, its relevance here is as a further instance of the generalization of economic rent beyond land that runs through Jan Eeckhout's other work: a persistent gap between price and cost, captured by firms holding a scarce, non-produced advantage — here, market position rather than a parcel of land. The paper's own authors use exactly that language for their result: in the conclusion, they write that welfare gains from reallocating output to more productive firms are "more than offset by these firms' use of their dominance to extract rents from customers" (Section 9, p. 39). That the paper's own authors — mainstream industrial-organization and macro economists writing for a Federal Reserve audience, with no Georgist framing — independently reach for the word "rent" to describe the wedge their model isolates is corroboration from outside the Georgist literature; it should not be overclaimed as evidence for land rent specifically, which the paper never addresses.

Beyond that parallel, the paper adds something the reduced-form 2020 QJE paper could not supply: a causal decomposition, not just a measured pattern. Where the QJE paper documents that markups rose and correlates that with declining labor share and dynamism, this paper's structural model attributes the rise to specific, separately identified mechanisms — rising overhead costs, wider productivity dispersion, and fewer potential competitors — and shows that the same forces which raise measured output (through reallocation to more productive firms) are, on net, welfare-reducing once the accompanying markup and overhead effects are counted. That is a more demanding empirical claim than "markups rose": it says the economy would be measurably better off, by the model's own welfare metric, under the pre-1980 mix of competition and cost structure, even though output from reallocation alone rose. This gives the wiki's broader rentier-economy argument (see The Rentier Economy) a structural-model counterpart to the reduced-form and national-accounts evidence it already carries — at the cost of resting on a specific, calibrated model rather than a direct measurement (see Nuances and Limits).

Nuances and Limits

• Not yet peer-reviewed. This is a Federal Reserve Bank of Minneapolis staff report dated September 2026; as of this page's last review it has not been published in a peer-reviewed journal, unlike the 2020 QJE paper. The wiki's usual pattern of tracking post-publication critiques (as on the markups page, which documents the Traina, Basu, and Benkard–Miller–Yurukoglu disputes) has not had time to develop here; none exists yet because the paper is too recent, not because the paper is uncontested.
• The number of competitors is model-inferred, not directly measured. A central design choice of the paper is that it deliberately avoids defining a "market" or counting actual competitors or market shares; instead the number of potential entrants, M, is "solved for indirectly by taking the model to the data and fitting the moments" (Introduction, p. 3) — markups, the cost-composition share, and the job-reallocation rate. This is a genuine methodological strength for avoiding contestable market definitions, but it also means the paper's central "fewer competitors" finding is an inference from the model's fit to three aggregate time series, not a direct count of firms or entrants. The paper's own gloss on why M might have fallen — rising M&A activity and weaker antitrust enforcement — is drawn from citations to other literature, not measured within this paper.
• A static, single steady-state-per-year model. The authors are explicit that they solve a static model for tractability rather than a fully dynamic one with forward-looking investment, and that a single year's estimation takes about 20 days of computing time; they note this choice cannot capture dynamic investment stories such as the rollout of a national distribution network (their own examples: Walmart, Amazon) that raise a firm's productivity over time (Section 4, p. 21). The three-channel decomposition should be read as the best fit of a deliberately parsimonious model, not as a complete causal account.
• Sample and scope. As with the underlying markup data, the exercise draws on Compustat, meaning publicly traded U.S. firms; it does not directly cover privately held firms. The paper's welfare, output, and wedge numbers are specific to this model's assumptions (nested-CES preferences, a common cross-market elasticity, a common within-market elasticity) and to the U.S. economy 1980–2023; the authors do not claim the specific magnitudes generalize to other countries or periods.
• No decomposition of why entry barriers rose, and no land component. The paper identifies that potential competition fell and attributes plausibility to M&A and antitrust trends, but does not model zoning, real-estate costs, or any other location-based barrier to entry — land is absent from the analysis entirely, as the zero-occurrence text check above confirms. Readers should not read this paper as evidence that entry barriers are land-related; the paper simply does not address the question.

Bears On

• Rent-Seeking — the paper's own language ("extract rents from customers") and its markup wedge are a modern, structurally-estimated instance of the rent concept this page defines, generalized beyond land; the model additionally shows why the wedge opened (overhead, dispersion, entry), not just that it did.
• Economic Rent — the markup wedge the model isolates is, by the standard definition, a form of economic rent; this paper supplies a causal decomposition of what drove its post-1980 rise.
• Corporate profits increasingly reflect economic rents — this paper adds a structural-model layer to that claim page's existing markup and profit-share evidence, showing the estimated rise is not explained by productivity dispersion alone but requires overhead-cost and competition changes as well.
• De Loecker, Eeckhout & Unger (2020) — markups — the companion reduced-form paper this one builds on and extends into a structural, causally-decomposed general-equilibrium setting; the two should be read together, with the 2020 paper establishing the markup pattern and this paper attempting to explain its causes and welfare consequences.
• Narrative: The Rentier Economy — a structural-model complement to that narrative's existing reduced-form and national-accounts evidence for rising rent capture outside land.

See Also

• Jan Eeckhout
• De Loecker, Eeckhout & Unger — The Rise of Market Power and the Macroeconomic Implications
• The Profit Paradox: How Thriving Firms Threaten the Future of Work
• Corporate Profits Increasingly Reflect Rents
• Economic Rent
• Rent-Seeking
• Narrative: The Rentier Economy
• Superstar Firms

Sources

1. Jan De Loecker, Jan Eeckhout & Simon Mongey (2026), "Quantifying Market Power and Business Dynamism in the Macroeconomy," Federal Reserve Bank of Minneapolis Staff Report 688, September 4, 2026. DOI: 10.21034/sr.688 · Minneapolis Fed — the full 70-page PDF (including the online appendix) was read in full — used for every quantitative claim and quotation on this page, including the model structure and three channels (pp. 2–3, 13–18), the headline welfare/output result and +15.5% reallocation figure (Abstract; p. 2), the five-wedge decomposition and the 12%/8%/20% markup-and-fixed-cost figures (Section 6, p. 33), the markup and labor-share/wage/employment fit statistics (Section 5, pp. 26–27), the "extract rents from customers" quotation (Section 9, p. 39), and the confirmation that "land" and "rent-seeking" do not appear anywhere in the text. B-claim; a Federal Reserve staff working paper, not yet peer-reviewed at the time of this page's writing.
2. Wiki: De Loecker, Eeckhout & Unger — The Rise of Market Power and the Macroeconomic Implications — internal navigation only; records the companion 2020 QJE paper's reduced-form markup measure that this staff report builds on and its own measurement critiques (Traina 2018; Basu 2019; Benkard, Miller & Yurukoglu 2025), which bear on the markup data series both papers share.
3. Wiki: Jan Eeckhout — internal navigation only; records the author's wider research program and biography.