Boards keep asking whether to publish pay. The better question is when and in what order. In a year when global employee engagement fell to 20 percent, with an estimated 9 percent of world GDP lost to disengagement, the cost of clumsy transparency is not academic (Gallup, 2025). Some disclosures prompt people to dig in and try harder. Others invite envy and exit. The evidence points to a simple lever that most firms ignore: the comparator you make salient.
Why the comparator is the incentive
A rare field experiment inside a large commercial bank shows how stark the difference can be. Employees systematically misjudged pay at the top. On average they underestimated their manager’s salary by 14.1 percent, and their absolute error was 28 percent. By contrast, their beliefs about peer pay were closer, with an absolute error of 11.5 percent and no systematic bias (Cullen and Perez-Truglia, 2022). When randomly provided with information that their manager earned more than they had thought, employees worked more. A 10 percent increase in perceived manager pay raised hours worked over the next 90 days by 1.5 percent. The study’s other finding is the mirror image: lateral comparisons tend to demotivate. Being confronted with the fact that a colleague at the same level earns more is not a spur to effort in the way that learning the boss earns far more can be (Cullen and Perez-Truglia, 2022).
Two cautions matter. First, the study is unusually strong for management research, combining administrative data, incentivized surveys, and pre-registered information treatments. Second, it was conducted in one bank in Southeast Asia. The directional lessons travel, but the exact magnitudes should not be universalized (Cullen and Perez-Truglia, 2022). The core managerial takeaway stands: transparency is not neutral. It works through the comparison it makes top of mind.
The managerial context is brittle
If transparency can cut both ways, today’s climate makes the cut deeper. Manager engagement has fallen sharply since 2022, with the steepest annual drop between 2024 and 2025. Only 22 percent of managers report being engaged, compared with 79 percent in best-practice organizations, a gap that signals design choices rather than fate (Gallup, 2025). Roll out cross-peer pay tables into a brittle system and you heighten status anxiety without a clear path to rise. Start by showing the shape of advancement above the team and the effect is different. A visible next rung, and who earns it, clarifies what to strive for.
This is why the order of disclosure matters. When employees first learn how pay escalates with greater scope or skill, they often infer a fairer tournament and invest more. When the first reveal is that a cubicle mate earns more, the salient lesson is that the game is capricious. The same numbers induce opposite behaviors depending on which comparison dominates.
In flatter teams, choose the vertical
The standard advice to “make pay transparent” assumes a manager-subordinate ladder. Many firms are flattening instead. Self-managing organizations decentralize authority formally and eliminate the conventional boss-subordinate relationship. Their distinctive feature is the absence of managerial hierarchy as the default way to coordinate work (Lee and Edmondson, 2017). The classic hierarchy can deliver reliable execution in stable settings, but it also entrenches status differences that stifle cross-functional problem solving and people’s development.
In these flatter structures, the comparator needs to be designed, not assumed. The upward reference point should be a role, a circle, or a skill tier with explicit criteria, not a person. The spirit is consistent with the evidence. Upward comparisons motivate when they signal a credible, attainable path to higher contribution and pay. Lateral comparisons demotivate when they shift attention to status gaps with no clear remedy. Publishing skill-tier pay bands and the evidence required to progress preserves the motivational effect of upward comparisons without turning every teammate into a rival. That is an interpretation informed by the vertical-versus-horizontal asymmetry in the bank experiment, applied to a context where “boss” is not the anchor (Cullen and Perez-Truglia, 2022; Lee and Edmondson, 2017).
Make transparency safe to learn from
Even the best sequencing can falter if people do not feel safe to discuss what they see. Psychological safety, the shared belief that a team is safe for interpersonal risk taking, is associated with learning behaviors that in turn mediate performance outcomes (Edmondson, 1999). Pay talk is high risk. If employees expect blowback for questioning disparities or asking how to progress, disclosure will produce silence in public and cynicism in private.
The practical move is to pair each stage of disclosure with a learning ritual. When you publish manager pay ranges and the criteria attached to them, convene sessions that walk through real promotion cases and the evidence trail required. When you publish role-based bands, train managers and team leads to have structured development conversations tied to those criteria. If you decide to provide any peer-related analytics, aggregate them and create channels where employees can safely ask, “What would move me from this band to the next?” The point is not to avoid discomfort. It is to convert it into inquiry and forward motion, the pattern that psychological safety enables (Edmondson, 1999).
A rollout that earns effort, not envy
Sequence the reveal to match how incentives work.
- Phase 1: Upward clarity. Publish ranges for manager and next-level roles, plus the evidence required to reach them. In settings without bosses, publish role or skill-tier ladders. The expectation is an effort bump, consistent with the bank experiment’s upward-comparison effect, with the caveat that context varies (Cullen and Perez-Truglia, 2022; Lee and Edmondson, 2017).
- Phase 2: Role bands with criteria. Release pay bands by role family and location together with rubrics, exemplars, and decision rights. This builds a credible vertical without making peers the unit of comparison.
- Phase 3: Aggregated peer context, if at all. Share distributions and pay policies at the cohort level. Avoid lists of named individuals. Provide recourse processes and coaching so people can act on what they learn rather than stew about it.
Done this way, transparency strengthens, rather than strains, a fragile engagement system. It sets a vertical to climb, reduces status-driven friction, and frames pay as a learning problem. The evidence does not argue for secrecy. It argues for design. In a year when the cost of disengagement is measurable in trillions, that design discipline is not a nice-to-have. It is the difference between publishing a spreadsheet and producing a strategy (Gallup, 2025).
Sources
- Cullen and Perez-Truglia (2022). How Much Does Your Boss Make? The Effects of Salary Comparisons, Journal of Political Economy. In a pre-registered field experiment, upward comparisons increased effort while horizontal comparisons demotivated, and employees systematically underestimated manager pay. https://doi.org/10.1086/717891
- Lee and Edmondson (2017). Self-managing organizations: Exploring the limits of less-hierarchical organizing, Research in Organizational Behavior. Defines SMOs and explains how managerial hierarchy reinforces status differences that can hinder development, informing how to design comparators in flat structures. https://www.sciencedirect.com/science/article/abs/pii/S0191308517300059
- Edmondson (1999). Psychological Safety and Learning Behavior in Work Teams, Administrative Science Quarterly. Shows that psychological safety is associated with team learning behaviors that mediate performance, a precondition for productive pay transparency. https://journals.sagepub.com/doi/10.2307/2666999
- Gallup (2025). State of the Global Workplace 2025. Reports global engagement at 20 percent, large productivity costs, and declining manager engagement, underscoring the stakes of getting transparency design right. https://www.gallup.com/workplace/659610/state-global-workplace.aspx