Treating firms as large, strategic employers changes everything: a macro general-equilibrium approach that embeds micro causal estimates shows employer monopsony can generate large misallocation and welfare losses, and that policies like minimum wages and antitrust can materially improve outcomes in concentrated markets.
Labor Market Power: From Micro Evidence to Macro Consequences
The paper argues that to measure and policy-test monopsony correctly one must embed non-atomistic employer strategic behavior in a general-equilibrium model and identify key primitives with micro quasi-experiments, since atomistic micro methods materially understate misallocation and welfare losses from employer market power.
The traditional theoretical and empirical “micro approach” to studying labor market power (or monopsony) requires that firms are small and atomistic. This is at odds with the reality of labor markets in which monopsony potentially matters most. Empirically, many markets are concentrated and characterized by large, dominant employers. The actions of large employers in an occupation or industry affect local and national wages, employment and output. Employers that understand their largeness may then act strategically when hiring and setting wages, generating misallocation and harming workers. This paper advocates for a “macro approach”: (1) directly model equilibrium behavior of large employers, (2) combine macro data and empirical estimates of employers’ responses to policy changes—obtained using the “micro approach”—to estimate the model, (3) use the model to compute the aggregate costs of monopsony, and optimal policies. This approach provides new perspectives on minimum wage and antitrust policy.
Summary
Main Finding
The paper argues that studying labor-market monopsony requires a "macro approach" that models the equilibrium behavior of large, non-atomistic employers, combines macro data with micro estimates of employer responses to policy, and uses that model to quantify aggregate costs of monopsony and design optimal policies (notably for minimum wages and antitrust). This approach reveals large-employer strategic behavior generates misallocation and sizable welfare losses that micro (atomistic-firm) methods miss.
Key Points
- The standard "micro approach" assumes many small, atomistic firms; this assumption is unrealistic where monopsony matters most because many markets are concentrated with dominant employers.
- Large employers affect local and national wages, employment, and output; recognizing firm largeness changes predicted responses to policy.
- Large employers may act strategically in hiring and wage-setting (e.g., exploiting downward-sloping labor supply, exercising market power), creating misallocation across firms and occupations and harming workers.
- The proposed "macro approach" has three steps:
- Build a general-equilibrium model that explicitly represents the strategic behavior of large employers (non-atomistic firms).
- Estimate that model by combining macro-level data with micro estimates of firms’ causal responses to policy shocks (the latter obtained using quasi-experimental micro methods).
- Use the estimated model to compute aggregate costs of monopsony and to evaluate and design optimal policies (e.g., minimum wages, antitrust interventions).
- This integrated strategy produces new perspectives on policy: e.g., minimum wages can increase welfare in concentrated markets by curbing monopsonistic wage suppression; antitrust policy can mitigate labor-market power and improve allocation.
Data & Methods
- Data sources and empirical building blocks
- Macro-level data: industry- and occupation-level wages and employment (BLS, QCEW), national accounts, productivity statistics, geographic cross-sections of labor markets.
- Firm- and worker-level micro data: administrative payroll (LEHD, Social Security, ADP), matched employer-employee datasets, online job posting data.
- Concentration and market-structure measures: firm shares by occupation/region, HHI, employer entry/exit records, merger filings.
- Micro identification strategies to estimate firm responses
- Quasi-experimental designs: difference-in-differences around localized shocks (e.g., plant openings/closings), event studies of mergers, spatial or industry variation in policy exposure, IV strategies for exogenous demand or supply shocks.
- Estimation of key primitives: employer-specific labor supply elasticities, hiring costs, matching/search frictions, pass-through of cost shocks to wages.
- Structural/macroeconomic modeling and estimation
- Build equilibrium models with large employers: monopsony models with downward-sloping firm-level labor supply, models with search-and-matching and employer-specific market power, multi-firm general equilibrium with firm-specific productivity and hiring policies.
- Estimation approaches: calibration or structural estimation (GMM, simulated method of moments, likelihood-based methods) that match macro aggregates and micro causal estimates.
- Counterfactual analysis: simulate competitive (atomistic) counterfactuals, compute aggregate welfare, output, employment, and distributional effects; evaluate policy instruments (minimum wage, antitrust remedies, payroll/subsidies).
- Quantities computed
- Aggregate cost of monopsony (wage gaps, employment shortfalls, output loss, misallocation across firms).
- Optimal policy design and welfare trade-offs (e.g., optimal minimum wage level conditional on market concentration; targeted antitrust remedies).
Implications for AI Economics
- Relevance to AI-driven labor-market change
- Large AI adopters/platforms can become dominant employers or gatekeepers; their strategic hiring and wage-setting will have aggregate effects that atomistic models miss.
- AI-induced demand shifts and automation can interact with employer market power: consolidation around AI capabilities could amplify monopsony, while AI-enabled remote work or matching tools could reduce frictions and potentially weaken employer power.
- Research directions and applications
- Incorporate firm-level AI adoption and platform labor-market features into macro monopsony models to assess aggregate impacts on wages, employment composition, and inequality.
- Use micro quasi-experiments (e.g., firm-level AI deployment, platform algorithm changes, major AI-hiring events) to identify employers’ causal responses, then embed those estimates in a macro equilibrium to compute economy-wide effects.
- Evaluate policy responses to AI-driven concentration: antitrust interventions targeting AI-related mergers or data/compute monopolies, minimum wages or wage subsidies in concentrated AI-intensive sectors, and retraining/transfer programs to mitigate reallocation costs.
- Study platform-mediated gig work where algorithmic control and multi-homing affect bargaining power—apply the macro approach to quantify welfare impacts and design regulation for algorithmic management.
- Practical suggestions for empirical AI-economics work
- Combine administrative payrolls (to observe wages/employment changes around AI adoption) with firm-level AI usage data and platform metrics.
- Estimate employer labor supply elasticities conditional on AI adoption intensity; feed those estimates into general-equilibrium models to simulate economy-wide AI scenarios.
- Compare counterfactuals where AI raises market concentration vs. where AI reduces frictions and increases competition to bound policy-relevant outcomes.
If you want, I can (a) sketch a simple model structure used in the macro approach (objects, equations, equilibrium condition), (b) list concrete datasets and empirical designs to implement step (2), or (c) draft a short research plan applying this approach to AI platform employers. Which would be most useful?
Assessment
Paper Typetheoretical
Evidence Strengthn/a — The paper is a conceptual/methodological proposal rather than an empirical study; it outlines an identification strategy and data needs but does not present new empirical estimates or validation of the approach.
Methods Rigormedium — The conceptual framework recommends rigorous tools (causal quasi-experiments for primitives; structural estimation to match aggregates) but provides no empirical implementation; methodological challenges (external validity of micro shocks, identification of firm-level supply elasticities, computational demands of GE estimation) remain to be resolved in applied work.
SampleNo single sample — the paper recommends combining macro aggregates (BLS, QCEW, national accounts, productivity statistics, geographic cross-sections) with firm-/worker-level administrative data (LEHD, Social Security records, ADP payrolls, matched employer-employee panels), online job-posting/platform metrics, and market-structure data (firm shares by occupation/region, HHI, merger filings).
Themeslabor_markets adoption governance inequality
IdentificationProposes a structural identification strategy: build a general-equilibrium model with non-atomistic (large) employers, estimate key causal primitives using micro quasi-experiments (DiD/event studies around plant openings/closings, merger studies, spatial/policy variation, IV for exogenous shocks), then fit/calibrate the macro model to match aggregate moments (via calibration, GMM, simulated method of moments or likelihood) so counterfactual competitive equilibria and policy effects can be computed.
GeneralizabilityFramework is best suited to concentrated labor markets and large-employer contexts; less applicable where firms are truly atomistic., Requires high-quality micro quasi-experiments and rich administrative data which may be unavailable in some countries/sectors., Results depend on structural model specification and chosen primitives; misspecification can bias aggregate counterfactuals., Policy conclusions may vary by institutional setting (labor laws, mobility, welfare systems) limiting cross-country external validity., AI-specific dynamics (e.g., rapid skill change, platform two-sidedness) may require additional model extensions before direct application.
Claims (7)
| Claim | Direction | Confidence | Outcome | Details |
|---|---|---|---|---|
| The traditional theoretical and empirical “micro approach” to studying labor market power requires that firms are small and atomistic. Market Structure | null_result | high | assumption about firm size/atomistic nature in micro monopsony models |
0.02
|
| This micro approach is at odds with the reality of labor markets in which monopsony potentially matters most. Market Structure | negative | medium | fit between micro model assumptions and actual labor market structure |
0.01
|
| Empirically, many markets are concentrated and characterized by large, dominant employers. Market Structure | null_result | medium | market concentration / presence of large dominant employers |
0.01
|
| The actions of large employers in an occupation or industry affect local and national wages, employment and output. Wages | mixed | medium | local and national wages, employment, and output |
0.01
|
| Employers that understand their largeness may act strategically when hiring and setting wages, generating misallocation and harming workers. Wages | negative | medium | misallocation and worker welfare (e.g., wages, employment outcomes) |
0.01
|
| A “macro approach” that (1) directly models equilibrium behavior of large employers, (2) combines macro data with empirical estimates of employers’ responses (from the micro approach) to estimate the model, and (3) uses the model to compute aggregate costs of monopsony and optimal policies, is the appropriate methodological response. Fiscal And Macroeconomic | null_result | high | aggregate costs of monopsony and optimal policy prescriptions |
0.02
|
| This macro approach provides new perspectives on minimum wage and antitrust policy. Governance And Regulation | mixed | low | policy implications for minimum wage and antitrust |
0.01
|