Trust at Organisational Speed
Why the dynamics that shape nations over centuries play out in your organisation over quarters — and why most of what you've read about building trust at work gets the mechanism wrong.
Abstract: Organisations are communities in Putnam's sense — they run the same self-reinforcing trust dynamics that shape regional economies, but at radically compressed timescales. The dominant practitioner literature (Covey, Lencioni) correctly intuits that trust compounds, but builds on a category error: treating trust as behavioural output rather than cognitive state. The Castelfranchi-Falcone model corrects this by decomposing trust into competence, disposition, and opportunity beliefs held by the trustor, not performed by the trustee. When this framework is applied at organisational scale, three consequences emerge. First, the equilibrium dynamics Putnam documented across regions over decades play out within teams over quarters, because interaction frequency, stakes, and network density are all elevated while exit costs are lower. Second, power asymmetry — absent from both Putnam's peer-community model and the practitioner literature — creates asymmetric defection costs that make low-trust equilibria structurally more stable in hierarchical organisations. Third, behavioural interventions calibrated for relationship timescales are too slow; structural interventions that change the environment for all interactions simultaneously are the only ones that operate at community speed. This is the first of two articles. The second examines how AI specifically perturbs the trust infrastructure described here.
Robert Putnam spent decades studying why some regions of Italy prospered while others stagnated. The answer was not geography, resources, or policy. It was social capital — the networks of trust and reciprocity that allow communities to cooperate effectively. The correlation between per-capita trust scores and regional economic performance was so strong (0.98 Pearson, confirmed computationally in recent work [1]) that it was difficult to attribute to anything other than a fundamental mechanism.
Putnam studied these dynamics across regions and nations, over decades and centuries. The processes were visible precisely because they were slow. A region didn’t collapse into distrust overnight. It drifted, over generations, into a self-reinforcing cycle: defection bred distrust, distrust bred shirking, shirking bred exploitation, exploitation bred disorder. Or the reverse — cooperation bred trust, trust bred further cooperation, and social capital accumulated. Two equilibrium basins. Once a community settled into one, escaping was extraordinarily difficult.
This article makes a simple claim: organisations are communities in Putnam’s sense. They run the same equilibrium dynamics. But they run them at radically compressed timescales — and that compression changes everything about what leadership means.
I. The unit of analysis is the community, not the individual
The most widely read books on organisational trust — Stephen Covey’s The Speed of Trust, Patrick Lencioni’s The Advantage, even Amy Edmondson’s work on psychological safety — share a common starting point. They treat trust as something individuals build through behaviour. Be credible. Be vulnerable. Be consistent. The prescription radiates outward from the self: if enough individuals behave in trustworthy ways, the organisation becomes trustworthy.
This gets the direction of causation backwards.
Trust, in the Castelfranchi-Falcone socio-cognitive model [2], is not a behaviour. It is not even a signal you emit. It is a mental state — a set of beliefs the trustor holds about the trustee. Specifically, three beliefs: that the other party is competent to do what is needed (competence belief), that they are disposed to act in one’s interest (disposition belief), and that structural conditions allow them to act (opportunity belief). Trust produces trusting behaviour, not the other way around. You cannot engineer the mental state by performing the behavioural outputs. And critically, the same behavioural output is interpreted differently depending on the community’s existing trust state. A leader who shows vulnerability in a high-trust team is perceived as courageous and authentic. The same act in a low-trust team is read as tactical manipulation or weakness. The basin filters the signal. This is why behavioural prescriptions that work beautifully in healthy organisations can backfire catastrophically in damaged ones — the receiver’s interpretive frame, not the sender’s intent, determines the trust outcome.
This matters because it shifts the unit of analysis. If trust were behaviour, the individual-level prescription would make sense — change the behaviour, change the outcome. But trust is a cognitive state held within relationships, and relationships exist within communities. The community’s aggregate trust capital determines whether individual trustworthy behaviour can even be expressed and recognised.
Consider a team member with genuinely good intentions operating in an organisation where information is hoarded, commitments are routinely broken by leadership, and credit is claimed by those with power. The structural conditions for trust — what Castelfranchi calls practical opportunity beliefs — are absent. It does not matter how credible or vulnerable this person is. The community-level dynamics override individual-level signals.
Organisations have every structural feature necessary for trust emergence in the Putnam sense: repeated interaction, reputation accumulation, shared stakes, information asymmetries, and — crucially — power dynamics. They are not merely analogous to communities. They are communities, with all the self-reinforcing dynamics that entails.
II. The compression
Putnam’s subjects could not see the phase transition happening in real time. The drift from high-trust civic engagement to low-trust institutional decay in American communities took half a century. By the time he published Bowling Alone [3], the transition was already deep into the vicious cycle. The diagnostic was retrospective.
Organisational leaders do not have this problem. Or rather, they have the opposite problem: the dynamics run fast enough to observe, but also fast enough to be catastrophic before the observation translates into action.
What determines the speed of Putnam’s equilibrium dynamics? At minimum: interaction frequency, stakes per interaction, network density, and information flow speed. In a nation, these variables are low — people interact with strangers infrequently, individual stakes are small, networks are sparse, and information propagates slowly. The equilibrium converges over generations.
In an organisation, every one of these variables is elevated. A team of eight interacts daily. Each interaction carries meaningful professional stakes. The network is dense — everyone knows everyone. Information, including information about defection, travels instantly. The same dynamics that take decades at regional scale take quarters at team scale.
But the compression is not merely a speedup of the same process. There is a structural difference that makes organisational dynamics even more volatile than Putnam’s regional model would predict: exit costs. In a region or nation, exit is expensive — you cannot easily leave your community, your language, your social network. High exit costs lengthen the shadow of the future, which Axelrod showed is precisely what sustains cooperation [4]. In an organisation, exit is comparatively cheap. People quit. And the easier it is to leave, the shorter the effective time horizon for any individual actor, which game theory tells us pushes toward defection. The combination of compressed dynamics and low exit costs means that organisational trust equilibria are not just faster versions of regional ones — they are structurally less stable. The vicious cycle, once entered, accelerates faster because the cooperators leave first.
This compression has two consequences that the practitioner literature has not reckoned with.
First, phase transitions become observable in real time. A leader who understands the equilibrium framework can detect which basin their organisation is moving toward. The early signals are legible: are people sharing information or hoarding it? Are commitments being met or quietly renegotiated? Are new hires being onboarded into a culture of candour or one of self-protection? These are not personality traits. They are equilibrium indicators. And at organisational speed, the window between detection and the point of no return is measured in months, not years.
Second, interventions must operate at community speed. Covey’s behavioural prescriptions — the 13 behaviours, the five waves of trust — are calibrated for relationship timescales. Years of consistent behaviour building credibility over time. This is sound advice for a personal relationship. It is too slow when equilibrium dynamics run at organisational timescales. By the time a leader has demonstrated consistent trustworthy behaviour over two years, the team may have already settled into the low-trust basin and rebuilt its norms around self-protection.
Structural interventions — changes to information flow, accountability mechanisms, decision-making processes, incentive structures — operate at community speed because they change the environment in which every interaction occurs, simultaneously, for everyone. Behavioural interventions change one person’s outputs and hope the signal propagates. At compressed timescales, the structural approach is not just more efficient. It is the only one fast enough.
III. What the practitioner literature gets right — and where it falls short
Let me be clear: the instincts in these books are largely correct. Covey is right that trust accelerates coordination and reduces friction costs. Lencioni is right that the dynamics are self-reinforcing — that organisational health compounds. Edmondson is right that psychological safety is a prerequisite for learning and performance. These are not wrong observations. They are, in many cases, independent rediscoveries of dynamics that Putnam documented at civilisational scale and that Castelfranchi formalised computationally.
The gap is mechanistic. When you build on intuition rather than mechanism, your interventions can be right for the wrong reasons — or wrong in ways you cannot diagnose.
Three specific gaps:
The direction of causation. Covey’s model begins with self-trust and radiates outward through relationship trust, organisational trust, market trust, and societal trust. This is motivationally appealing but mechanistically backwards. Trust is not a signal you emit. It is a belief state formed in the receiver based on evidence, context, and prior beliefs. You can influence the evidence available to the receiver, but you cannot control the belief formation process. Castelfranchi’s model makes this distinction precise: the trustor’s beliefs about the trustee’s competence, disposition, and opportunity are updated through interaction, not through the trustee’s self-presentation [2].
The absence of structural opportunity. Neither Covey nor Lencioni engages seriously with what Castelfranchi calls practical opportunity beliefs — the trustor’s assessment of whether structural conditions allow the trustee to act on their competence and good intentions. A manager may be competent and well-disposed, but if the organisation’s incentive structure punishes the behaviour the trustor needs, trust cannot rationally form. This is not a failure of character or credibility. It is a structural constraint. The practitioner literature treats organisational structure as context rather than as a primary variable in trust formation. Putnam would recognise this immediately — his entire framework is about how institutional structures create or destroy the conditions for trust.
The neutrality of power. This is the most consequential gap. Both Covey and Lencioni write as though power asymmetry is incidental to trust dynamics. It is not. Power asymmetry fundamentally alters the cost structure of defection. In a peer-to-peer community — Putnam’s civic associations, Axelrod’s iterated prisoner’s dilemma — defection is costly because the defector faces reciprocal punishment. In a hierarchical organisation, defection by those with power is structurally cheaper. A senior leader who breaks a commitment, hoards information, or claims credit faces fewer immediate consequences than a junior team member who does the same.
Axelrod’s insight [4] is that cooperative equilibria require defection to be immediately and predictably costly. When power asymmetry lowers the cost of defection for some actors, the conditions for sustained cooperation are undermined. This creates an asymmetric vulnerability: the powerful can defect without immediate consequence, but their defection degrades trust across the entire community — because everyone observes it and updates their beliefs accordingly.
The hypothesis I would advance — and which I believe deserves formal investigation — is that power asymmetry makes low-trust equilibria more stable and harder to escape. Not because powerful people are less trustworthy, but because the risk structure is asymmetric: when a senior leader defects — breaks a commitment, hoards information, claims credit — the community bears the cost while the leader often retains the upside. When a junior team member does the same, the consequences are immediate and personal. This asymmetric risk distribution means defection is locally rational for those with power even when it is globally destructive, making defection the dominant strategy at the top of the hierarchy precisely where its effects on the community are largest. This is a Putnam vicious cycle with an asymmetric accelerant.
IV. The diagnostic value
If organisations are communities running Putnam dynamics at compressed timescales, then organisational leadership is, in part, equilibrium management. The leader’s job is not to be trustworthy (though that helps). It is to maintain the structural conditions under which the high-trust equilibrium is self-sustaining.
This reframing has practical consequences.
Hiring is community composition. Every addition to a team changes the interaction dynamics, the reputation network, and the information flow patterns. A hire who is individually excellent but structurally disruptive to the trust equilibrium is a net negative. This is not about “culture fit” — a concept too vague to be useful. It is about whether the new interaction patterns the hire introduces push the community toward or away from the cooperative equilibrium.
Reorganisations are equilibrium shocks. When you restructure a team, you are not merely changing reporting lines. You are destroying existing trust networks and forcing new ones to form from zero. The Castelfranchi belief components — competence, disposition, opportunity — must all be rebuilt through new interactions. If the interaction frequency and stakes are high enough, the new equilibrium forms quickly. If not, the team may settle into a low-trust holding pattern that persists indefinitely. Most organisations treat reorganisation as an organisational design problem. It is equally a trust dynamics problem.
Metrics can be equilibrium indicators. The standard lagging indicators of trust collapse — turnover, disengagement, productivity decline — arrive after the phase transition is already underway. Leading indicators are harder to measure but more valuable: information sharing patterns, commitment fulfilment rates, the gap between public statements and private behaviour, the speed at which new members are integrated into candid communication. These are not personality diagnostics. They are community health metrics.
Structural interventions outperform behavioural prescriptions. Transparency by default (open calendars, shared documents, public decision logs) changes the information environment for everyone simultaneously. Accountability mechanisms that apply equally regardless of seniority address the power asymmetry problem directly. Decision-making frameworks that make the reasoning visible — not just the outcome — preserve the signals that trust beliefs depend on. None of these require anyone to be vulnerable, authentic, or credible. They change the structural conditions under which trust forms, which is faster and more reliable than hoping individual behaviour will propagate.
V. What this framework does not yet address
I want to be honest about the gaps.
The compression argument — that Putnam dynamics run faster in smaller, denser communities — is intuitively strong and directionally supported by existing research at team scale (Edmondson [5]), organisational scale (Fukuyama [6]), and regional scale (Putnam [7], De Meo et al. [1]). But the time constant itself has not been formalised. What precisely determines how fast the equilibrium dynamics converge in a community of a given size, density, and interaction frequency? This is a question worth answering rigorously.
The power asymmetry hypothesis — that hierarchy makes low-trust equilibria more stable — is, at this stage, a hypothesis. It is grounded in Axelrod’s framework and consistent with organisational experience, but it has not been tested empirically in the way I would want before claiming it as established.
And the question of leading indicators — what exactly are the early signals of a phase transition from high-trust to low-trust equilibrium? — is practically the most important question this framework raises, and it is entirely open. It is also, I suspect, answerable empirically, which makes it a research agenda rather than a speculation.
These gaps are not a weakness of the argument. They are what makes it a research programme rather than a finished theory. Let me be explicit about both the known limits and what needs to come next.
VI. Known limits and future work
The scale invariance bridge is proposed, not proven. This article argues that the same Putnam equilibrium dynamics operate at team scale, organisational scale, and regional scale — differing in time constant but not in mechanism. Empirical work exists at each scale independently: Edmondson’s psychological safety research at team level [5], social capital and organisational performance studies at firm level, and Putnam’s original work at regional and national level [7]. A recent research agenda from HEC Paris notes that the role of social capital and trust for organisations “remains underexplored” — confirming the gap but not filling it. What is missing is a formal demonstration that the dynamics are mechanistically continuous across scales, rather than superficially similar. The compression argument is intuitively strong and directionally supported, but it has not been tested with the rigour I would want before claiming it as established. A longitudinal study measuring trust equilibrium convergence rates across communities of different sizes and interaction densities would be the right next step.
Granovetter’s objection is unaddressed. Putnam’s regional dynamics emerge from weak ties across large populations. Organisational dynamics are driven by strong ties in small networks. It is possible that these produce qualitatively different equilibrium behaviour, not merely faster versions of the same process. The weak-tie/strong-tie distinction may mean that the basin structure itself differs at different scales — not just the convergence speed. I believe the equilibrium framework still holds because the underlying mechanism (reciprocity, reputation, defection cost) operates in both regimes, but this is an assertion that needs formal investigation.
The power asymmetry hypothesis is the least supported claim. The argument that hierarchy makes low-trust equilibria more stable — because power lowers the cost of defection for those who hold it — is grounded in Axelrod’s framework and consistent with organisational experience. But it has not been tested empirically. It is the claim in this article most likely to be wrong in its specifics, even if the general direction (that power asymmetry modifies trust dynamics) is almost certainly correct. Existing critiques of Putnam note that his framework underplays power dynamics and structural inequality, but none have formalised the interaction between hierarchy and equilibrium stability in the way I am proposing. This needs its own dedicated investigation.
Edmondson’s work deserves deeper engagement. Psychological safety is arguably the closest existing empirical programme to what this article describes — team-level trust dynamics measured through observable behaviour. I have cited Edmondson as support but not engaged with the substance of her findings or shown precisely where the Putnam-Castelfranchi framework extends, refines, or reframes her conclusions. That engagement is owed and will be part of future work. The short version: Edmondson’s psychological safety maps most closely to a specific configuration of Castelfranchi’s opportunity beliefs (B-PrOp) — the perception that structural conditions permit honest expression. The framework offered here embeds that within a broader equilibrium model that also accounts for competence and disposition beliefs. But this claim needs to be developed carefully, not asserted in passing.
Leading indicators of phase transition remain unidentified. This is the most practically important open question. If organisational leaders can observe Putnam dynamics in real time — which the compression argument implies — then they need instruments for doing so. What specifically are the early signals that a team or organisation is shifting from the high-trust to the low-trust basin? The standard lagging indicators (turnover, disengagement, productivity decline) arrive after the transition is well underway. Candidates for leading indicators — information sharing velocity, commitment fulfilment rates, synchronous-to-asynchronous communication ratios, speed of candid integration of new members — are plausible but untested. Identifying and validating these indicators is an empirical project worth undertaking.
What this framework does not attempt. This article does not model trust formation at the individual cognitive level — that is Castelfranchi and Falcone’s contribution, and I build on it rather than extending it. It does not address cross-cultural variation in trust dynamics, which Putnam himself documented extensively and which would add significant complexity to the organisational application. And it does not yet address the specific perturbations that AI introduces to the trust infrastructure described here — that is the subject of the next article.
The claim is not that we have all the answers. The claim is that the right unit of analysis is the community, not the individual; that the right theoretical framework is Putnam’s equilibrium dynamics, not behavioural prescription; and that the compression of those dynamics at organisational scale creates both a diagnostic opportunity and an intervention imperative that the current practitioner literature does not adequately address.
In the next article, I will examine what happens to these dynamics when AI enters the loop — because the trust infrastructure described here is exactly what is now under pressure.
References
[1] De Meo, P., Prifti, Y. & Provetti, A. (2025). Trust Models Go to the Web. ACM Transactions on the Web, 19(2). doi:10.1145/3715882
[2] Castelfranchi, C. & Falcone, R. (2010). Trust Theory: A Socio-Cognitive and Computational Model. Wiley.
[3] Putnam, R.D. (2000). Bowling Alone: The Collapse and Revival of American Community. Simon & Schuster.
[4] Axelrod, R. (1984). The Evolution of Cooperation. Basic Books.
[5] Edmondson, A.C. (2018). The Fearless Organization: Creating Psychological Safety in the Workplace. Wiley.
[6] Fukuyama, F. (1995). Trust: The Social Virtues and the Creation of Prosperity. Free Press.
[7] Putnam, R.D. (1993). Making Democracy Work: Civic Traditions in Modern Italy. Princeton University Press.
[8] Govier, T. (1997). Social Trust and Human Communities. McGill-Queen’s University Press.
[9] Covey, S.M.R. (2006). The Speed of Trust: The One Thing That Changes Everything. Free Press.
[10] Lencioni, P. (2012). The Advantage: Why Organizational Health Trumps Everything Else in Business. Jossey-Bass.
[11] Prifti, Y. (2026). A Markov Framework for AI Trust and Societal Outcomes. SSRN. doi:10.2139/ssrn.6390618
[12] Levine, T.R. (2019). Duped: Truth-Default Theory and the Social Science of Lying and Deception. University of Alabama Press.