Does Partial Pay Transparency Reduce Pay Inequality? An Experimental Investigation

As organizations increasingly adopt pay transparency policies, a key challenge is determining what compensation information to disclose. However, little is known about the effectiveness of disclosing only parts of total compensation, such as salary, while excluding other components, such as benefits. Using an online experiment on Prolific, this paper uncovers a concerning outcome: when decision makers are incentivized to allocate bonus earnings between workers, partial transparency systematically increases the influence of component equality on the allocation choices. This causes decision makers to focus more heavily on ensuring equality within the transparent component. This focus is significantly amplified for the transparent component relative to a nontransparent pay component, and this tendency persists even when doing so worsens overall pay inequality. We further show that social preferences play a central role in generating this asymmetry, whereas monetary incentives for decision makers primarily exert a level effect that shifts allocation behavior without altering the asymmetric pattern. In addition, decision makers’ beliefs about the importance of visible relative earnings differences help explain variation in allocation behaviors. In contrast, this asymmetric focus on achieving component equality disappears when either all or none of the compensation is transparent. Our results suggest that firms implementing partial transparency may inadvertently reinforce pay disparities rather than reducing them. These findings highlight the need for organizations to carefully design transparency policies to avoid distortions in fairness perceptions and pay allocation decisions. This paper was accepted by Marie Claire Villeval, behavioral economics and decision analysis. Funding: X. Pan gratefully acknowledges financial support from Bryant University [Summer Research Grant]. Supplemental Material: The data files are available at https://doi.org/10.1287/mnsc.2025.01032 .