Antitrust and Competition Policy
Antitrust is the structural counterpart to progressive taxation: where tax policy redistributes the outcome of market concentration, competition policy attacks the source. Market concentration has risen sharply across U.S. and European industries since the 1980s (De Loecker et al. 2020; Philippon 2019), with markups, profit shares, and industry-level Herfindahl indices all reaching multi-decade highs. This concentration translates almost directly into wealth concentration through equity ownership of dominant firms.
The historical track record is encouraging. The 1911 Standard Oil breakup, the 1982 AT&T divestiture, and the (failed but precedent-setting) 1982 IBM case each correlated with periods of expanded innovation and lower price-cost margins. The post-Bork “consumer welfare” framing that dominated U.S. antitrust from 1980–2020 produced a permissive regime under which most large horizontal mergers proceeded; the 2021–2024 Khan-era FTC, EU Digital Markets Act, and renewed U.S. Big Tech cases mark the largest reversal in 40 years, though outcomes are still pending.
The case for antitrust as an inequality lever is stronger when one accepts that labor-market monopsony — concentrated employer power over wages — is roughly as important as product-market concentration. Estimates (Azar et al. 2020; Yeh et al. 2022) suggest monopsony depresses wages by 10–25% relative to a competitive benchmark, with the largest effects in low-wage labor markets.
Lever inventory
Section titled “Lever inventory”| Lever | Target | Best-case effect on inequality |
|---|---|---|
| Structural breakup | Dominant firms in tech, finance, telecom | Large (one-time) |
| Merger blockade | Horizontal mergers above thresholds | Large (cumulative) |
| Non-compete bans | Labor market monopsony | Moderate (wage growth) |
| Interlocking directorate enforcement | Cross-firm coordination | Moderate |
| Sector-specific regulation (utilities) | Natural monopolies | Moderate |
| Common-ownership scrutiny | Asset managers’ aggregate stakes | Uncertain (active research) |
Chicago School vs. New Brandeisian
Section titled “Chicago School vs. New Brandeisian”U.S. antitrust law has been governed since the early 1980s by the consumer welfare standard introduced by Robert Bork’s The Antitrust Paradox (1978) and adopted by the federal courts in a series of decisions through the 1980s and 1990s. Under this standard, a business practice is anticompetitive only if it can be shown to raise consumer prices or reduce output in the short run. Mergers, vertical integrations, exclusive dealing, and predatory pricing claims that cannot meet this narrow test are generally permitted, even when they substantially increase market power.
The standard’s intellectual genealogy is the Chicago School — primarily Aaron Director, Bork, Richard Posner, and Frank Easterbrook — which argued that markets are largely self-correcting, that monopoly is rarely durable absent government protection, and that aggressive antitrust enforcement risks “false positives” that harm consumers more than the underlying conduct does. The framework’s victory in the courts was substantial enough that DOJ monopolization cases dropped from 221 (1955–1979) to 22 (1980–1997) — a roughly 90% decline that the Lancieri/Posner/Zingales analysis (cited in Power Concentration Mechanisms) attributes mostly to political-economy capture using Chicago doctrine as a vehicle, not to genuine intellectual conversion.
The New Brandeisian critique — named for Justice Louis Brandeis, who championed broader anti-monopoly enforcement in the early 20th century — emerged from the Open Markets Institute (Barry Lynn), the Roosevelt Institute, and Lina Khan’s seminal Yale Law Journal note “Amazon’s Antitrust Paradox” (2017). The New Brandeisian argument is that the consumer-welfare standard is too narrow on three margins:
- It ignores labor-market and supplier-side harms. Monopsony power over workers and suppliers can raise corporate profits without raising consumer prices and so is invisible to the consumer-welfare test.
- It ignores political-economy harms. Concentrated market power produces concentrated political power, which then re-shapes regulation and enforcement (the feedback loop in Power Concentration Mechanisms).
- It ignores innovation harms. Dominant firms can suppress nascent competitors through acquisition, predatory pricing, and platform-leveraging in ways that don’t raise prices today but reduce future innovation and competition.
The New Brandeisian framework reached institutional power with Khan’s appointment as FTC chair (2021–2025), Jonathan Kanter at the DOJ Antitrust Division, and Tim Wu in the White House. Its policy expressions include the 2023 FTC/DOJ Merger Guidelines, the FTC non-compete rule, and the major Big Tech cases discussed below.
Rising market concentration and markups
Section titled “Rising market concentration and markups”The empirical case for renewed antitrust rests on consistent evidence that market power has risen substantially since 1980 across U.S. and European industries. The strongest single piece of evidence is De Loecker, Eeckhout & Unger (2020, QJE, “The Rise of Market Power and the Macroeconomic Implications”), which uses firm-level financial data to estimate aggregate markups (the gap between price and marginal cost) across U.S. publicly traded firms 1955–2016.
The headline findings:
- Aggregate markups rose from 21% above marginal cost in 1980 to 61% by 2016 — a tripling.
- The increase is concentrated in the upper tail. The median firm’s markup is roughly unchanged; almost all the rise comes from the top quintile of firms in each industry.
- Average profit rates rose from 1% to 8% of sales — an order-of-magnitude change.
- Within-industry reallocation toward high-markup firms explains most of the aggregate trend, consistent with rising dominance of leading firms within each sector rather than a broad-based shift.
The macroeconomic implications matter for inequality: De Loecker et al. argue that rising market power can quantitatively account for the declining labor share, the declining capital share (yes, both), the decline in business dynamism, and a meaningful fraction of the post-1980 inequality rise. Philippon (2019, The Great Reversal) makes a complementary argument with European-vs-U.S. comparative data: U.S. industries have become substantially more concentrated than European counterparts since 1990, and consumer prices for many products are now lower in Europe than in the U.S. as a result.
Critics argue that rising markups reflect intangible-capital investment (software, brands, patents) that is mismeasured in standard accounts; that the rise in “concentration” partly reflects superstar productivity rather than market power; and that geographic-market definitions matter (national HHI is rising while local-market HHI is more stable in many sectors). These critiques have substance but do not overturn the core finding of significant rises in markups and profitability concentrated in dominant firms.
Labor-market monopsony
Section titled “Labor-market monopsony”The most important development in 2010s–2020s antitrust scholarship is the recognition that monopsony — concentrated employer power over labor — is roughly as economically important as monopoly power over consumers, and has been almost entirely ignored under the consumer-welfare standard.
Azar, Marinescu & Steinbaum (2022, JHR, “Labor Market Concentration”) used U.S. online vacancy data to map labor-market concentration at the local-occupation level and produced findings that have become central to the policy debate:
- The average U.S. local labor market has an HHI of ~3,157, equivalent to ~3.2 effective recruiting employers.
- 60% of U.S. local labor markets are “highly concentrated” by DOJ/FTC merger-review standards (HHI > 2500).
- A move from the 25th to 75th percentile of concentration is associated with a 17% decline in posted wages (panel IV estimate).
- Application elasticities are near zero in all but the most densely populated labor markets — implying that ~80% of the U.S. workforce works in a labor market with meaningful monopsony power.
Yeh, Macaluso & Hershbein (2022, AER, “Monopsony in the U.S. Labor Market”) extended this with Census manufacturing micro-data and estimated that wages are roughly 35% below marginal product on average — implying a substantial wedge between worker productivity and worker pay that is invisible to consumer-welfare analysis.
Policy implications:
| Lever | Mechanism | Status |
|---|---|---|
| Non-compete bans | Removes contractual barrier to job-switching | FTC rule 2024 (enjoined; on appeal) |
| No-poach agreement enforcement | Treats as per-se Sherman Act violation | DOJ enforcement since 2016 |
| Wage-fixing prosecutions | Criminal antitrust | Active under Biden DOJ |
| Merger review with labor focus | Considers monopsony in merger HHI | 2023 Merger Guidelines incorporate |
| Right-to-organize reforms | Counters monopsony with labor power | See Labor Power |
| Information-sharing prohibitions | Wage benchmarking ban | DOJ guidance 2016 |
The labor-monopsony evidence is the single most important reason why antitrust has come to be seen as an inequality intervention rather than purely a consumer-protection intervention.
Historical case studies
Section titled “Historical case studies”| Case | Year | Outcome | Lesson for inequality |
|---|---|---|---|
| Standard Oil | 1911 | Broken into 34 successor companies | Effective; durable |
| AT&T | 1982 | Divested into 7 “Baby Bells” | Enabled competitive telecom and internet expansion |
| Microsoft | 2001 | Settled with conduct remedies; no breakup | Conduct remedies less effective than structural |
| IBM (1969–1982) | 1982 (dropped) | DOJ withdrew case after 13 years | Long cases face political-economy attrition |
| Visa/MasterCard | 2003 | Conduct remedies on exclusive dealing | Limited durable effect |
| Google (1st antitrust) | 2013 | FTC closed without action | Consumer-welfare standard at work |
The Standard Oil 1911 breakup is the canonical example of effective structural relief. The 34 successor companies competed with each other and broke the trust’s pricing and distributional power; the Rockefeller fortune diversified across the successors but no longer represented monopoly control. The AT&T 1982 breakup is the clearest 20th-century example: the divestiture of the Bell system into seven regional operating companies enabled the competitive long-distance market that became the commercial internet. The Microsoft 2001 settlement is the cautionary tale: the consent decree imposed conduct remedies (interoperability, browser-bundling restrictions) without structural separation, and Microsoft remained dominant in the operating-system layer for a decade after.
The pattern across cases is that structural remedies (breakup) are much more durable than conduct remedies (behavioral restrictions). Conduct remedies require ongoing enforcement against a sophisticated and well-resourced respondent; structural remedies create new competitive structures that police themselves.
Current cases (status as of 2026)
Section titled “Current cases (status as of 2026)”The 2021–2024 period saw the largest reactivation of U.S. antitrust enforcement since the AT&T era, and the 2024 EU Digital Markets Act marks the parallel European approach. Status as of early 2026:
| Case | Filing year | 2026 status |
|---|---|---|
| DOJ v. Google (Search) | 2020 | Liability ruling Aug 2024 (Google illegally monopolized search); remedies phase ongoing |
| DOJ v. Google (Adtech) | 2023 | Trial concluded; ruling pending |
| FTC v. Meta (Instagram/WhatsApp) | 2020 | Trial began April 2025; seeking divestiture |
| FTC v. Amazon | 2023 | District court allowed central monopolization claims to proceed Oct 2024; trial scheduled |
| DOJ v. Apple | 2024 | Filed; pre-trial |
| EU DMA: Apple, Meta, Alphabet | 2024 invest. | Apple fined €500M (App Store steering); Meta fined €200M (pay-or-consent); Alphabet investigation continuing |
The Khan-era FTC was succeeded by Andrew Ferguson under the second Trump administration in 2025, with mixed signals about continuity of the existing cases. Khan’s stated concern on her departure was that the incoming administration might let pending cases settle on terms favorable to the defendants (“sweetheart deals”). The early 2026 read on whether this is happening is mixed: the Amazon and Meta cases have continued moving forward, while several smaller investigations have been quietly closed.
The EU’s Digital Markets Act is operating on a different theory entirely — ex ante obligations on designated “gatekeepers” rather than ex post litigation over individual conduct. The April 2025 Apple and Meta fines (the first DMA enforcement actions) were modest in dollar terms (€500M and €200M against companies with hundreds of billions in market cap) but established that the regulatory architecture works. The DMA’s potential is in the cumulative effect of ongoing compliance pressure rather than in headline fines.
The structural reform question for the next decade is whether the post-2021 reactivation produces durable institutional and doctrinal change or whether it ebbs back to the post-Bork equilibrium under different administrations. The institutional changes (FTC capacity expansion, DOJ Antitrust Division staffing, the 2023 Merger Guidelines, the EU DMA) are more durable than the individual cases. The fundamental doctrinal question — whether the consumer-welfare standard is replaced or merely supplemented by broader frameworks — remains open and will be resolved at the appellate level over the next 5–10 years.
Objections it must clear
Section titled “Objections it must clear”Antitrust confronts the mechanical-not-extractive objection most directly. That objection holds that concentration and high returns reflect benign forces — superstar technology, firm scale, the most capable manager running the largest firm — rather than extraction. Antitrust accepts that some scale is productive and insists on the distinction the objection elides: between returns to genuine efficiency and returns to market power. A firm that is large because it is good earns a competitive return; a firm that is large because it has foreclosed rivals or suppressed wages earns a rent. The price-of-growth objection similarly assumes intervention trades growth for fairness — but unaddressed monopoly and monopsony are themselves a drag on output and employment, so curbing them is pro-efficiency and pro-distribution at once.
Distributional effects: gender and race/ethnicity
Section titled “Distributional effects: gender and race/ethnicity”The clearest distributional channel is labor-market power. Employer monopsony depresses pay most for workers with the fewest outside options, who are disproportionately women and workers of color (Council of Economic Advisers 2016, Labor Market Monopsony); Naidu, Posner and Weyl (2018) argue antitrust should treat wage suppression as an injury in its own right. Non-compete and no-poach agreements, concentrated in low-wage, high-turnover sectors, lock those same workers out of the job-switching that drives raises — Krueger and Ashenfelter (2018) found no-poach clauses in a majority of major franchise contracts. Curbing these practices raises wages precisely where the workforce is most female and most non-white.
On the product-market side, monopoly pricing operates as a regressive transfer: markups on necessities fall hardest on lower-income households, which are disproportionately Black and Hispanic, while the equity and ownership that capture monopoly profits are held overwhelmingly by white households. Antitrust enforcement that lowers markups and curbs wage suppression therefore narrows both the income and the wealth gap, though, like the other class-based levers, it does not by itself reach the inherited component of the racial wealth gap. The workers with the fewest outside options, where monopsony bites hardest and non-competes bind most tightly, are disproportionately women of color, so curbing employer wage-setting power reaches a compounded position that a race-only or gender-only lens would treat as two separate problems. The comprehensive strategy sets out that intersectional test.