Competition Law and Labour Data Concentration Concerns

1. Introduction

Labour data concentration refers to a situation in which one employer, a small group of employers, a labour-market intermediary, recruitment platform, data broker, salary-survey provider, or algorithmic platform controls a substantial amount of information concerning workers and employment conditions.

Such data may include:

• wages and salary histories;
• bonuses and benefits;
• job classifications;
• skills and qualifications;
• employment histories;
• productivity and performance data;
• recruitment and turnover data;
• worker availability;
• reservation wages;
• hiring and attrition rates;
• geographic mobility;
• non-compete information;
• employee ratings and rankings; and
• data generated through recruitment or workforce-management platforms.

Competition law becomes relevant when concentration of labour data reduces competition between employers for workers, facilitates wage coordination, creates barriers to entry, enables discriminatory or exclusionary conduct, or gives a dominant intermediary the ability to control access to workers.

The modern approach increasingly treats workers as participants in a labour market, with employers acting as purchasers of labour. U.S. competition authorities expressly recognize that information exchanges concerning compensation can create antitrust problems, particularly where competing employers receive current or prospective information.

2. Meaning of Labour Data Concentration

Labour data concentration can arise in several ways.

A. Employer concentration

A small number of employers may collectively employ a large proportion of workers in a particular geographic or occupational market.

For example, if five hospitals employ most nurses in a locality, their access to detailed information about each other's wages can materially affect competition.

B. Data-platform concentration

A recruitment or workforce platform may become the principal repository of information about:

• available workers;
• salary expectations;
• employee movements;
• skills;
• hiring intentions; and
• employer compensation.

The platform can consequently become an important competitive intermediary.

C. Salary-survey concentration

Several employers may obtain detailed compensation information from the same survey provider.

A survey can be competitively useful, but the risk increases when it provides:

• highly current information;
• firm-specific information;
• information about future compensation;
• information concerning a small number of competitors; or
• sufficiently granular occupational/geographic information.

D. Algorithmic concentration

An algorithm may aggregate employer and worker data and recommend:

• wages;
• hiring levels;
• bonuses;
• recruitment targets; or
• employee retention strategies.

The competition concern becomes particularly significant where competing employers rely upon the same algorithm and substantially similar non-public data.

3. Why Labour Data Is Economically Sensitive

Information normally facilitates competition. However, information that allows competitors to observe and predict each other's competitive conduct can also facilitate coordination.

The classic distinction is:

More transparent markets are not automatically more competitive markets.

For example, historical, aggregated and anonymized salary information may assist employers and workers without materially facilitating coordination. By contrast, current and prospective salary information identifying individual competitors may allow employers to monitor and align compensation decisions.

The U.S. authorities have specifically warned that exchanges of compensation information among competing employers can violate antitrust law when structured in a manner that harms competition.

4. Principal Competition Concerns

A. Wage coordination

The most obvious concern is that employers use shared labour data to coordinate wages.

Suppose competing hospitals receive a weekly database showing:

The information could reduce uncertainty about competitors' future conduct.

Instead of competing aggressively for nurses, employers may converge around similar compensation levels.

B. Reduction of employer competition

Competition law protects competition between employers, not merely competition between sellers of goods.

Where employers compete for workers, competition can occur through:

• wages;
• bonuses;
• flexible working arrangements;
• benefits;
• working conditions;
• training;
• promotion opportunities; and
• job security.

The DOJ and FTC expressly recognize this labour-market dimension of antitrust law.

C. Monopsony and buyer power

Labour data concentration can strengthen monopsony power.

A monopsonist is an employer or group of employers possessing substantial purchasing power in a labour market.

Data can strengthen such power because the employer can learn:

• how many workers are available;
• what workers are currently paid;
• workers' likely reservation wages;
• competing employers' hiring intentions;
• turnover probabilities; and
• employees' willingness to move.

The result can be reduced competition for workers even without an explicit wage-fixing agreement.

5. Data Granularity and Antitrust Risk

Not all labour information presents the same risk.

Lower-risk characteristics

• historical information;
• aggregated information;
• anonymized information;
• sufficiently broad occupational categories;
• independent third-party administration;
• delayed publication.

Higher-risk characteristics

• current wages;
• future salary budgets;
• employer-specific data;
• employee-specific data;
• small datasets;
• frequent updates;
• geographically narrow information;
• detailed job classifications;
• identifiable competitors.

The DOJ's historical guidance concerning hospital salary surveys illustrates this distinction: it accepted a structure involving an independent third party, aggregated averages and sufficiently old information, while warning about exchanges likely to facilitate coordination.

6. Six Important Case Laws

1. Todd v. Exxon Corp., 275 F.3d 191 (2d Cir. 2001)

This is one of the leading cases concerning exchange of salary information among competing employers.

Employees alleged that major oil and petrochemical companies exchanged detailed compensation information concerning managerial, professional and technical employees.

The information allegedly included:

• salaries;
• bonuses;
• benefits;
• starting salaries;
• relocation expenses;
• stock options;
• salary budgets; and
• future compensation information.

The Second Circuit held that the plaintiff had adequately alleged a Sherman Act §1 information-exchange claim and reversed dismissal.

Principle

The case demonstrates that a labour-information exchange can constitute an antitrust issue where:

1. the relevant labour market is properly defined;
2. the market is susceptible to coordination;
3. the exchanged information has anticompetitive potential; and
4. antitrust injury is plausibly alleged.

Importance

Todd v. Exxon is directly relevant to labour data concentration because the information itself was the mechanism through which competitors allegedly reduced uncertainty concerning compensation.

2. United States v. Utah Hospitals / Utah Hospital Association, 1994

The DOJ brought proceedings against eight Utah hospitals and related associations concerning exchanges of registered-nurse wage information.

The alleged information exchange concerned current and prospective entry wages for nurses and allegedly stabilized entry wages and limited wage increases.

The resulting consent decrees restricted agreements to exchange current or prospective compensation information except in specified circumstances.

Principle

Current and prospective labour-compensation information can be competitively sensitive.

Importance

This is especially significant because it demonstrates that labour information exchange can attract antitrust scrutiny even when there is not necessarily a conventional cartel involving the sale of goods.

3. Fleischman v. Albany Medical Center, 728 F. Supp. 2d 130 (N.D.N.Y. 2010)

Registered nurses alleged that hospitals in the Albany area:

1. suppressed RN wages; and
2. exchanged detailed, non-public compensation information.

The alleged exchanges occurred through communications, meetings and professional organizations and concerned matters such as RN pay ranges, new-graduate rates and shift differentials.

The court allowed important aspects of the litigation to proceed, including claims concerning the competitive effects of the information exchanges.

Principle

The competitive significance of labour data depends heavily upon:

• whether information is current;
• whether it is non-public;
• how frequently it is exchanged;
• how detailed it is; and
• whether the exchange can facilitate coordinated compensation.

Importance

The case illustrates the connection between data concentration and wage-setting power.

4. Cason-Merenda v. Detroit Medical Center, 862 F. Supp. 2d 603 (E.D. Mich. 2012)

Registered nurses alleged that Detroit-area hospitals exchanged compensation information and thereby harmed competition.

The information was exchanged through:

• direct employer contacts;
• healthcare organizations;
• professional meetings; and
• third-party compensation surveys. 

The court ultimately distinguished between the alleged wage-fixing theory and the information-exchange theory and examined whether the evidence established competitive harm.

Principle

An exchange of salary information is not automatically unlawful.

The court emphasized the importance of proving the competitive effects and causal relationship between the information exchange and alleged labour-market injury.

Importance

This case is particularly useful for explaining why data concentration alone is insufficient. The legal question is how the information affects competitive conditions.

5. Unger v. Albany Medical Center, No. 06-CV-0765 (N.D.N.Y. 2010)

The litigation involved allegations that hospitals conspired to depress RN wages and exchanged detailed, non-public compensation information concerning nurses.

The court examined allegations involving the exchange of information about compensation actually being paid and compensation expected to be paid.

Principle

Information about future compensation can be especially sensitive because it can reduce uncertainty concerning future competitive behaviour.

Importance

The case illustrates a major labour-data problem:

The more a dataset allows competitors to predict future wage decisions, the greater its potential competitive significance.

6. Nitsch v. DreamWorks Animation SKG, Inc., No. 14-CV-04062 (N.D. Cal. 2016)

The case arose from the broader Silicon Valley employment-conduct litigation involving technology companies and alleged restrictions on employee solicitation.

Evidence concerning salary surveys and compensation communications was relevant to allegations that companies sought to suppress employee compensation. The litigation considered evidence that companies participated in salary surveys and discussed compensation in connection with restrictions on recruiting.

Principle

Labour data exchanges can become especially problematic when combined with:

• no-poach arrangements;
• non-solicitation agreements;
• recruitment restrictions; or
• other mechanisms restricting employee mobility.

Importance

It shows that labour data should not be analyzed in isolation. A data exchange can become more significant when it operates alongside direct restrictions on hiring.

7. Related Case: Silicon Valley No-Poach Enforcement

In United States v. Adobe Systems, Inc., Apple Inc., Google Inc., Intel Corp., Intuit Inc. and Pixar, the DOJ challenged agreements restricting employee solicitation.

The companies agreed to restrictions preventing certain forms of direct recruitment between them. The DOJ stated that such agreements eliminated an important form of competition for highly skilled workers.

Similarly, the DOJ subsequently required eBay to terminate anticompetitive employee-hiring agreements.

These cases are not principally labour-data concentration cases, but they demonstrate an important complementary proposition:

labour data + restricted worker mobility can create substantially greater competition concerns than either issue considered independently.

8. National Resident Matching Program

The litigation concerning the National Resident Matching Program (NRMP) provides another important labour-market example.

Physicians alleged that the matching system and related arrangements restrained competition and contributed to below-competitive resident compensation. The litigation also raised issues concerning information sharing and the structure of the market for medical residents.

The broader lesson is that centralized labour-market infrastructure can raise competition questions where it substantially affects how employers and workers interact.

9. Data Concentration and Digital Labour Platforms

The problem becomes more complicated with modern platforms.

A recruitment platform may possess:

worker identity + skills + salary expectations + employment history + job applications + employer demand + hiring outcomes.

This creates a two-sided labour-market platform.

The platform potentially serves:

• workers on one side; and
• employers on the other.

Competition concerns can arise if the platform:

1. Self-preferences its own recruitment services

The platform may rank workers or vacancies in ways that favour affiliated services.

2. Restricts data portability

Workers may be unable to transfer their:

• ratings;
• employment history;
• credentials;
• skills profile; or
• reputation score

to another platform.

This can create switching costs and strengthen platform power.

3. Prevents interoperability

A dominant platform may restrict competing recruitment systems from accessing necessary information.

4. Uses exclusive contracts

Employers or workers may be prevented from simultaneously using competing platforms.

5. Uses data to discriminate against rivals

A dominant intermediary might use data concerning a competing recruitment platform's users or employers to disadvantage that platform.

10. Algorithmic Wage-Setting

One of the most important emerging issues is algorithmic wage coordination.

Imagine 100 competing employers provide a common algorithm with:

• employee compensation;
• vacancies;
• turnover;
• hiring rates;
• geographic supply;
• worker preferences; and
• competing employers' salary information.

The algorithm recommends compensation to each employer.

Even if the employers never communicate directly, competition authorities may ask whether the algorithm has effectively become a coordination mechanism.

The key questions include:

1. Who owns the algorithm?
2. Who supplies the data?
3. Is the data public or confidential?
4. Is the information current?
5. Are employers individually determining wages?
6. Does the algorithm use competitors' non-public data?
7. Can employers reject the algorithm's recommendation?
8. Does the algorithm monitor whether employers follow its recommendations?

Modern U.S. enforcement increasingly considers algorithmic decision-making and labour-market concentration together. The FTC and Department of Labor have expressly identified both labour-market concentration and algorithmic decision-making as areas of mutual interest.

11. RealPage as an Important Analogical Development

Although United States v. RealPage, Inc. concerns rental housing rather than employment, it is highly relevant conceptually to labour-data analysis.

The DOJ challenged the use of competitively sensitive information in algorithmic pricing. Later proposed remedies addressed restrictions concerning the use of non-public competitively sensitive data in training models and required changes to software features.

The analogy for labour markets is:

Employer data → algorithm → wage recommendation

can raise similar conceptual questions to:

Landlord data → algorithm → rental-price recommendation.

The legal analysis, however, cannot simply transfer the RealPage outcome to labour markets because the relevant market, statutory provisions and competitive effects are different.

12. Labour Data and Merger Control

Labour-data concentration can also arise through mergers.

Suppose:

Recruitment Platform A + Recruitment Platform B

control large databases containing:

• millions of worker profiles;
• salary information;
• employer demand;
• hiring histories; and
• skills data.

The merger may affect both:

Product-side competition

Competition between recruitment services.

Labour-side competition

Competition between employers seeking workers.

Data-side competition

Access to an important dataset that competing platforms cannot easily reproduce.

Therefore, merger analysis may need to consider whether the transaction creates:

• increased data concentration;
• foreclosure of competing recruitment platforms;
• reduced worker mobility;
• higher barriers to entry;
• increased switching costs;
• discriminatory access conditions; or
• greater ability to extract rents from either employers or workers.

The EU's merger framework examines concentrations under the EU Merger Regulation, and current EU practice continues to review transactions involving digital and employment-related businesses.

13. Labour Data as an Essential Competitive Input

A particularly important question is whether labour data can constitute an essential competitive input.

For example, suppose a dominant recruitment platform possesses almost all historical information concerning a specialized occupational group.

A rival platform may need that information to:

• match workers and employers;
• verify credentials;
• assess experience;
• establish worker reputation; and
• provide meaningful search results.

If access is refused, competition authorities may examine:

1. whether the data is indispensable;
2. whether duplication is realistically possible;
3. whether refusal excludes competitors;
4. whether access is technically feasible;
5. whether access would undermine legitimate privacy or security interests; and
6. whether the refusal protects a legitimate business justification.

The essential-facilities analysis must therefore be balanced against privacy, cybersecurity, confidentiality and intellectual-property interests.

14. Privacy and Competition Law

Labour data frequently contains personal information.

Consequently, competition law and data-protection law may overlap.

A competition authority may ask:

Does control over personal labour data give the undertaking market power?

A data-protection regulator may ask:

Is the collection and processing of the information lawful?

These are different questions.

A company does not automatically violate competition law merely because it possesses extensive personal data.

Conversely, compliance with data-protection law does not automatically immunize conduct from competition scrutiny.

15. Worker Data Portability

Worker data portability can be particularly important in platform markets.

A worker's profile may contain a form of economic reputation capital.

For example:

Platform A

→ 10 years of ratings
→ skills certification
→ employment history
→ completed projects
→ employer reviews

If the worker cannot transfer that information to Platform B, the worker may face substantial switching costs.

This can reinforce platform dominance.

Potential competition remedies include:

• standardized data formats;
• worker-controlled portability;
• interoperability;
• API access;
• prohibition of discriminatory access;
• transparent ranking systems; and
• restrictions on exclusive arrangements.

16. Competition Issues in Labour-Data Sharing

A competition-law assessment should consider the following matrix:

17. Rule-of-Reason Versus Per-Se Analysis

A crucial distinction is that not every labour-data exchange is automatically unlawful.

Courts have recognized that information exchanges can have legitimate efficiency benefits.

For example, aggregated compensation surveys can assist:

• workforce planning;
• recruitment;
• equal-pay analysis;
• compensation benchmarking; and
• efficient allocation of labour.

The competition question is therefore often:

Does the exchange improve market information and efficiency, or does it reduce strategic uncertainty and facilitate coordination?

The Second Circuit's approach in Todd v. Exxon demonstrates the importance of examining the market structure and competitive potential of the particular information exchange.

18. Safe-Harbour-Type Design Principles

Businesses conducting legitimate salary benchmarking can reduce risk by adopting safeguards such as:

1. Independent administrator

Use an independent third party rather than direct competitor-to-competitor exchange.

2. Aggregation

Do not disclose employer-specific compensation data.

3. Anonymization

Remove identifying information.

4. Historical data

Avoid unnecessary circulation of current or prospective compensation information.

5. Minimum participant thresholds

Avoid datasets in which information from only a few employers can be reverse-engineered.

6. Limited frequency

Avoid unnecessary real-time or highly frequent exchanges.

7. Purpose limitation

Use data solely for legitimate benchmarking.

8. Access controls

Restrict access to personnel who genuinely require it.

9. Algorithmic safeguards

Do not permit an algorithm to automatically coordinate competing employers' wage decisions.

10. Audit mechanisms

Regularly examine whether the data system is facilitating coordination.

The DOJ's treatment of hospital wage surveys illustrates the importance of independent administration, aggregation and historical data in reducing information-exchange concerns.

19. Remedies for Labour Data Concentration

Competition authorities may potentially employ several remedies.

Structural remedies

• divestiture of data assets;
• separation of recruitment and data businesses;
• prohibition of certain acquisitions.

Behavioural remedies

• non-discrimination obligations;
• data-access requirements;
• interoperability;
• data portability;
• restrictions on data use;
• prohibition on exclusive agreements.

Information-exchange remedies

• prohibition of current compensation exchanges;
• restrictions on prospective wage information;
• anonymization;
• aggregation;
• third-party administration.

Algorithmic remedies

• independent audits;
• explainability requirements;
• restrictions on competitor data;
• human override;
• prohibition on automatic wage coordination.

20. Key Legal Principles Emerging from the Cases

The cases collectively establish several important principles.

Principle 1 — Labour is a competition market

Employers compete against one another to obtain and retain workers. U.S. enforcement policy expressly recognizes competition among employers as an antitrust concern.

Principle 2 — Salary information can constitute competitively sensitive information

Todd v. Exxon is particularly important here.

Principle 3 — Current and future data are particularly sensitive

The Utah hospital matter demonstrates the concern associated with current and prospective RN wage information.

Principle 4 — Aggregation can reduce risk

The DOJ's New Jersey hospital survey guidance demonstrates the importance of aggregation, third-party administration and historical data.

Principle 5 — Data exchange is context-dependent

Cason-Merenda demonstrates that the existence of a data exchange does not by itself establish an antitrust violation; competitive effects and causation remain important.

Principle 6 — Data and mobility restrictions can reinforce one another

The Silicon Valley no-poach litigation illustrates how compensation information and restrictions on employee solicitation can operate together.

21. Conclusion

Labour data concentration is becoming an important dimension of competition law because control over labour information can influence the competitive process itself.

The principal concerns are not simply the possession of data but its strategic use.

The greatest risks arise where:

Concentrated labour data
↓
Current/future compensation information
↓
Competitor visibility
↓
Reduced uncertainty
↓
Wage coordination or reduced hiring competition
↓
Lower worker mobility and bargaining opportunities

The leading cases—particularly Todd v. Exxon, the Utah Hospitals proceedings, Fleischman, Cason-Merenda, Unger and Nitsch—demonstrate that courts and competition authorities have long recognized the competition significance of exchanging sensitive employment information. Modern recruitment platforms and algorithmic labour-market systems extend these concerns by allowing enormous datasets to be collected, combined and used to influence employment decisions.

The central competition-law inquiry is therefore not “How much labour data does an undertaking possess?” but rather:

Whether control, exchange or algorithmic use of that data materially reduces competition among employers, restricts worker mobility, facilitates coordination, forecloses rivals, or strengthens durable labour-market power.