Firms, workers, and the changing rules of competition

August 4, 20269 papers · 20 min read

This is Franklin’s first full NBER issue. It follows the approved scope: firm dynamics, productivity, labor, trade, econometrics, and the economics of AI and automation, with selective industrial-organization and policy evidence. Nine papers passed a deliberately low-noise cutoff from 31 official NBER records. The summaries below use NBER’s abstracts—not the full papers—and preserve a direct link to every source.

NBER week ending 31 July · 31 papers reviewed · 9 included

01NBER Working Paper 35542Official NBER abstract

Winners and Losers: Competition, Creative Destruction, and Labor Income Risk

Brice C. Green, Leonid Kogan, Dimitris Papanikolaou, and Lawrence D.W. Schmidt

Firm dynamicsInnovationLabor economics

Question: How does technology-driven creative destruction affect the earnings and employment risk of workers at incumbent firms?
Approach: The authors combine U.S. administrative data with an endogenous-growth model featuring monopsonistic labor markets and worker heterogeneity. They compare how workers respond to profit losses caused by rivals’ innovations and to gains from innovation at their own firms.
Findings: The abstract reports that losses caused by rival innovation produce disproportionately large declines in incumbent workers’ earnings and a higher likelihood of job destruction, relative to the gains associated with their own firm’s innovation. Top workers are more exposed than the average worker. In the model, creative destruction concentrates downside risk among high-income workers while increasing upward mobility for lower-income workers.
Why it matters: Innovation policy changes not only aggregate growth but also who bears labor-income risk. The paper links product-market competition, firm outcomes, and the distributional consequences of technological change.
Caveat: Franklin reviewed the official NBER abstract, not the full paper. The construction of the innovation shocks, identification details, and the model’s assumptions and welfare calculations remain unassessed.

02NBER Working Paper 35552Official NBER abstract

The Equilibrium Impact of Credit Frictions: Evidence from Default Risk Using Firm-Level Data

Timothy J. Besley, Peter John Lambert, Isabelle A. Michalski-Roland, and John Van Reenen

Firm dynamicsCredit frictionsProductivity

Question: How much do credit frictions reduce output, wages, and productivity once firm responses and economy-wide equilibrium effects are taken into account?
Approach: The paper builds a micro-to-macro model of heterogeneous firms and sectoral production in which perceived default risk summarizes firms’ financing constraints. It combines the framework with U.K. administrative firm data and S&P default-risk measures from 2004–2019.
Findings: The abstract reports that relaxing credit frictions raises output by 25 percent and wages by 23 percent. Most gains come from capital accumulation. Holding wages fixed would overstate the output effect, while ignoring fixed-capital misallocation would understate it. Credit frictions also alter firm size, misallocation, and productivity growth.
Why it matters: The paper shows why financing constraints cannot be evaluated by adding up firm-level effects: wages, capital accumulation, and resource allocation adjust throughout the economy.
Caveat: Franklin reviewed the official NBER abstract, not the full paper. The counterfactual magnitudes depend on the model, the default-risk measure, and maintained assumptions that are not evaluated here.

03NBER Working Paper 35554Official NBER abstract

A Theory of Firm Wage Dynamics

Marc de la Barrera and Masao Fukui

Firm dynamicsWagesSearch frictions

Question: How do wage contracts evolve as firms grow, age, and experience productivity shocks in a labor market with search frictions?
Approach: The authors integrate a Burdett–Mortensen wage-posting environment with firm dynamics. Dynamic wage contracts, equal treatment of workers, and search frictions link a firm’s wage path to the distribution of its marginal surplus.
Findings: The model predicts that wages are more closely related to firm growth than firm size, tend to decline over a firm’s life cycle, and respond more strongly to productivity shocks in the short run. The abstract also reports that a hiring subsidy targeted at firms high on the job ladder can increase business dynamism.
Why it matters: The framework provides a unified way to think about wage setting, worker mobility, firm growth, and policy in dynamic labor markets.
Caveat: This is a theoretical contribution. Franklin reviewed the official NBER abstract, not the derivations, calibration, empirical fit, or sensitivity of the policy result.

04NBER Working Paper 35561Official NBER abstract

The Anatomy of Tariff Pass-through into Consumer Prices

Mary Amiti, Sebastian Heise, and David Weinstein

TradeTariffsConsumer prices

Question: How did the 2025 U.S. tariffs pass through to consumer prices, and how much came through imported goods, imported inputs, and domestic markups?
Approach: The authors compare goods with different exposure to the tariff increases while holding aggregate conditions fixed, then decompose the estimated price response into direct and indirect channels.
Findings: The abstract reports that 26 percent of the tariff increase passed through to consumer prices. Direct effects on foreign varieties account for 64 percent of that response; imported-input costs and domestic markups account for the remaining 36 percent. Direct effects appear quickly, while indirect effects take roughly nine to twelve months.
Why it matters: The timing and composition of pass-through determine who ultimately bears tariffs and why their full consumer-price effect may not be visible immediately.
Caveat: Franklin reviewed the official NBER abstract, not the full paper. The comparison design, decomposition assumptions, product coverage, and longer-run responses remain unassessed.

05NBER Working Paper 35540Official NBER abstract

Predictive Risk Scores in the Public Sector: Experimental Evidence from Child-Protection Investigations

E. Jason Baron, Arkadev Ghosh, and Richard Lombardo

AlgorithmsPublic sectorField experiment

Question: What happens when child-protection supervisors receive an algorithmic risk score alongside case records while retaining discretion over decisions?
Approach: The study randomizes access to a predictive risk score across 4,752 child-welfare referrals over 14 months in Northampton County. It measures responses in placements and services, later maltreatment referrals, and racial disparities.
Findings: According to the abstract, access to the score increased placements and services for the highest-risk cases, changed little for lower-risk cases, and reduced subsequent maltreatment referrals. The authors report no evidence that the intervention widened racial disparities.
Why it matters: The experiment offers rare evidence on an algorithm used as decision support rather than as an automatic decision-maker, making the interaction between prediction and human discretion central.
Caveat: The evidence comes from one county and one institutional setting. Franklin reviewed the official NBER abstract, not the outcome definitions, implementation details, subgroup precision, or broader generalizability.

06NBER Working Paper 35559Official NBER abstract

Replaceable but Employed: Automation and the Meaning of Work

Joshua S. Gans

Economics of AIAutomationMeaning of work

Question: Can automation reduce the value workers derive from their jobs even when they remain employed?
Approach: The paper develops a theoretical model in which workers value both output and the extent to which that output depends on their own contribution. A credible machine alternative can therefore change the meaning of work without replacing the worker.
Findings: The model shows that demonstrating a machine substitute weakens the worker’s sense of being necessary. If wages adjust fully, compensation rises; if adjustment is incomplete, workers bear part of the loss. An outside developer may create a “meaning externality” simply by making the alternative salient, and this effect differs from improving the machine’s quality.
Why it matters: The paper identifies a channel through which AI can affect worker welfare before jobs or tasks disappear—one that standard measures of employment and wages may miss.
Caveat: This is a theoretical mechanism, not an estimate of how large the effect is in workplaces. Franklin reviewed the official NBER abstract, not the model’s derivations, assumptions, or empirical applicability.

07NBER Working Paper 35560Official NBER abstract

Long-Run Effects of H-1B Immigration on the U.S. Economy

Ran Abramitzky, Leah Platt Boustan, Ahmet Gulek, and Jens Hainmueller

ImmigrationLabor economicsProductivity

Question: What were the long-run effects of the late-1990s H-1B expansion on workers, industries, and productivity in the United States?
Approach: The authors use differences across industries in exposure to the 1999–2003 expansion of the H-1B cap. They trace income effects for natives and earlier immigrants, along with spillovers through supply chains and patenting.
Findings: The abstract reports higher incomes for natives and pre-existing immigrants in more exposed industries, including gains for non-STEM workers. Effects extend downstream through supply chains but not upstream, which the authors interpret as evidence of higher productivity rather than a simple labor-supply shock. They find no direct increase in patenting and emphasize improved task execution as the mechanism.
Why it matters: The paper connects skilled immigration to broad workplace and production effects, rather than evaluating it only through the wages of directly competing STEM workers.
Caveat: Franklin reviewed the official NBER abstract, not the full paper. The exposure design, identifying assumptions, outcome construction, and interpretation of the downstream pattern remain unassessed.

08NBER Working Paper 35544Official NBER abstract

Two Selection Problems, One Bias Term: Experimental Sign-Up Is Treatment Choice in Disguise

John A. List

EconometricsExperimentsExternal validity

Question: Are bias from voluntary entry into an experiment and bias from voluntary treatment choice fundamentally the same selection problem?
Approach: The paper uses a Roy-model foundation to show that both problems arise from sorting on gains and share an inverse-Mills-ratio bias term. It then uses a simple calibration and discusses reporting alternatives when nonparametric identification is unavailable.
Findings: The abstract argues that experimental sign-up and treatment choice are mathematically equivalent selection problems. Its calibration suggests that bias in a typical experiment can be large, and it proposes three ways researchers can report results when the selection problem cannot be solved fully.
Why it matters: Random assignment inside a study does not by itself guarantee that the result generalizes to people who chose not to enter. Treating recruitment as an explicit selection problem can improve experimental design and reporting.
Caveat: Franklin reviewed the official NBER abstract, not the derivation, calibration choices, nonparametric arguments, or the practical performance of the proposed reporting alternatives.

09NBER Working Paper 35565Official NBER abstract

Common Ownership and Collusion

Vincent Abraham, Florian Ederer, and Catarina Marvao

Industrial organizationCommon ownershipCartels

Question: How is common ownership related to firms’ participation in legal cartels and to the duration of those cartels?
Approach: The authors combine population data on legal cartels in Sweden with hand-collected ownership records. They construct firm-pair profit weights to measure how strongly one firm internalizes another firm’s profits and relate those weights to cartel behavior over time.
Findings: The abstract reports that higher profit weights are associated with cartel participation, predict future participation, and coincide with longer cartel duration. Weights fall after cartels dissolve. The relationship is concentrated where the internalizing firm colludes and its rival does not, suggesting that common ownership complements the firm’s own participation but substitutes for a rival’s participation.
Why it matters: The evidence sharpens the competition-policy question: ownership links may affect not only unilateral incentives but also which firms join and sustain coordinated arrangements.
Caveat: The abstract reports associations and predictive patterns, not a causal research design. Franklin did not review the full paper’s ownership measures, legal-cartel setting, specifications, or alternative explanations.

A quiet record, week by week.

August 4, 2026Firms, workers, and the changing rules of competition9 papers