Salary data from placement records: privacy and antitrust limits

Why placement salary data raises antitrust questions#

Placement salary data raises antitrust questions because a staffing firm sits between many employers that compete to hire the same people. Its records show what each client pays, what it offered, what candidates accepted and how rates moved. Shared carelessly, that information can help competing employers align pay instead of competing on it.

US antitrust enforcers treat agreements among competing employers to fix wages, or not to hire each other's workers, much like price-fixing agreements. Exchanging current or future compensation information can also raise concerns without any explicit agreement, and a third party that passes such information between competitors can be seen as the channel for it. Agency positions on information exchanges have shifted over time, so older safe-harbor thresholds you may have seen should not be relied on without a current view from antitrust counsel.

Why it raises privacy questions too#

Placement salary data raises privacy questions because an individual's pay is personal information, and placement records tie it to a named candidate, a named employer and a date. Recruiter notes often add salary history and expectations, which some jurisdictions restrict employers from asking about or relying on.

Removing names does not solve the problem on its own. When a role, a region and a period describe only a few placements, a published figure can reveal what a particular person earns, and colleagues or the employer may recognize it immediately. Pay transparency laws in some states also change what employers disclose in postings, which affects how your figures will be read alongside public ranges.

Do and don't table: aggregation, recency and publication#

The do and don't table below summarizes practices that generally lower risk when salary information from placements is used outside the firm. It is a starting point for a discussion with counsel, not a safe harbor.

How risk changes by use#

Risk changes with what the firm does with salary data, so assess each use on its own. Using your own placement history to price your own services sits at one end; passing current rate information between competing employers sits at the other.

The middle uses, such as published salary guides and licensing, are where design choices matter most. A guide built on historical, aggregated figures with a clear method note looks very different from one that slices pay by named employer.

Licensing deserves its own thought because the recipient is not a reader but a system. A model trained on employer-level pay records could later answer questions about what a specific company pays, turning a one-time license into an ongoing channel for current wage information. Removing or banding compensation and restricting permitted use address that risk at the source.

Checks before you publish or license#

The checks before publishing or licensing are mostly about structure: how many sources sit behind each figure, how old the data is and whether any cell is small enough to point to a person or an employer. Record each check, because a written method is what counsel and a careful buyer will want to see.

Small-cell testing borrows a standard privacy idea. Latanya Sweeney's 2002 k-anonymity model asks that every record in a release look identical to at least k-1 others on the fields an outsider could use to single someone out. Applied to a salary guide, every published cell of role family, region and period should rest on enough placements, and enough separate employers, that no single offer or client can be read back out of it.

• 1Write down the purpose and audience for the data before building anything.
• 2Have antitrust and privacy counsel review the design, not just the final output.
• 3Set a minimum number of employers and placements behind every published figure and suppress cells below it.
• 4Use only completed historical periods and state them clearly.
• 5Remove salary history and candidate expectations from source records.
• 6Run a small-cell check across role, region and period, counting both placements and distinct employers in each cell.
• 7Publish a method note explaining sources, aggregation and suppression.
• 8For licensing, write permitted-use terms that bar resale as a benchmarking product.

Illustrative: a finance and accounting staffing firm plans two uses#

Illustrative: a fictional finance and accounting staffing firm wants to publish a salary guide and is also exploring licensing its placement workflow records. Its records include pay and bill rates for each placement, offers and counteroffers, and recruiter notes that sometimes record a candidate's prior salary.

Counsel advises building the guide only from completed periods, with ranges by role family and broad region and a minimum source count per cell, and suppressing any cell that falls short. For the licensing package, pay and bill rates are removed, offer outcomes are kept as accepted or declined with reason categories, client identity is replaced by industry and size, and salary history is stripped from notes. The firm approves both designs separately.

How SourceX handles compensation fields#

SourceX treats compensation as a high-sensitivity field. In the SourceX five-step transaction, Rights checks client contracts and candidate notices, and Preparation removes or bands pay and bill rates under rules the supplier approves. SourceX does not give antitrust advice; the supplier's counsel assesses that question for each deal.

The SourceX Evidence Packet records how compensation was treated, the permitted use agreed with the buyer and the supplier's release authorization, so any later question about what salary information left the firm has a written answer.