Research

Why do organizations fail to implement policies they have formally adopted, even under external legitimacy pressure? Classical institutional accounts portray policy–practice decoupling as a deliberate, top-down strategy in which leaders adopt symbolic policies without intending to change practice. We develop a bottom-up account centered on the employees responsible for implementation. Because organizational identification is weaker among lower-ranking employees, policies adopted by senior leaders may be implemented only partially below. We test this theory using 17,000 firm-years and two preregistered experiments with behavioral and attitudinal outcomes. We also introduce a scalable, text-based measure of aggregate organizational identification and formalize the mechanism in a mathematical model. Across studies, higher-ranking members identify more strongly, and identification increases implementation effort only under high legitimacy pressure. These findings explain how leaders can adopt policies intended to change practice, yet lower-ranking employees do not fully implement them.

We theorize prospective polarization as a process by which organizational members become socially divided over imagined futures before those futures materialize. Research has typically examined retrospective and present-oriented forms of polarization, treating divisions over the future as contestation rather than polarization. Yet the futures organizational members are asked to pursue can themselves become polarizing. This occurs because the future is a source of meaning that is open to projection and framing. We focus on three temporal dimensions of prospective framing—pace, distance, and rupture—that shape how imagined futures are appraised. We argue that these dimensions generate initial divergence in members’ reactions and activate prospective faultlines through which that divergence becomes socially organized, potentially escalating contestation into polarization marked by emotional intensity, moralization, and identity-based closure. Our core contribution is to theorize when imagined futures divide organizations, what form that division takes, and how it can be interrupted.

Do expert decision-makers exhibit loss aversion in high-stakes settings, and does time pressure amplify it? We examine these questions using data from the World Rapid and Blitz Chess Championships series, where professional chess players compete for substantial monetary prize incentives under different time controls, in otherwise comparable tournament settings. We measure move quality using evaluations from a chess engine and use tournament simulations to estimate players' prospective monetary gains and losses in each game. We find that both prospective gains and prospective losses are associated with better play, but the improvement in play in response to avoiding prospective losses is larger than the response to pursuing comparable gains. Moreover, this pattern of asymmetry tends to be larger in Blitz games, when decision time is tightly constrained, than in Rapid games.

Why do stakeholders often fail to detect or challenge gaps between organizational policy and practice? We develop a stakeholder-centric theory of decoupling as a collective action problem. Although stakeholders benefit from exposing noncompliance, each bears the cost of monitoring. Our formal model shows that self-interested stakeholders therefore underinvest, while prosocial stakeholders may duplicate one another’s efforts, creating blind spots that firms exploit. Delegating oversight to a regulator or watchdog can coordinate monitoring and close these gaps, but it also creates the risk of capture. We identify when centralized monitors remain vigilant, when they enable decoupling, and when even captured monitors outperform decentralized stakeholder communities. By shifting attention from firms to stakeholders, our theory shows how stakeholder motivation, monitoring capacity, coordination, and oversight integrity determine whether decoupling persists or is exposed.