Beyond procedural justice: Power, identity and the asymmetric architecture of public trust in security services

Public trust in security services is a critical social-psychological phenomenon, yet existing theoretical accounts remain fragmented and insufficiently attentive to the structural power asymmetries characterising security-public relations. This article argues that the dominant paradigm, procedural justice theory, requires fundamental revision within intergroup and institutional power dynamics. Drawing upon procedural justice theory (Why people obey the law, 1990, Yale University Press; Why people obey the law, 2nd ed., 2006, Princeton University Press), social identity and self-categorisation theory (The social psychology of intergroup relations, 1979, 33, Brooks/Cole; Rediscovering the social group: A self-categorization theory, 1987, Blackwell), the integrative model of organisational trust (Academy of Management Review, 1995, 20, 709), intergroup contact theory (The nature of prejudice, 1954, Addison-Wesley), system justification theory (British Journal of Social Psychology, 1994, 33, 1), institutional betrayal theory (American Psychologist, 2014, 69, 575) and the political trust tradition (A systems analysis of political life, 1965, Wiley; Trust and trustworthiness, 2002, Russell Sage Foundation; Annual Review of Political Science, 2000, 3, 475), the article advances three propositions. First, trust formation is asymmetric: trust erodes faster than it accrues, with the asymmetry amplified for subordinate-status groups. Second, procedural justice effects are identity-contingent: perceived fairness operates differently depending on whether the authority is categorised as in-group protector or out-group enforcer. Third, trustworthiness appraisal under institutional betrayal requires reconceptualising Mayer et al.'s integrity dimension as epistemic credibility. A four-pathway model is proposed, with power asymmetry as a structural moderator and recursive feedback producing virtuous or vicious trust spirals.