Executive Summary
Multi-agent simulation of 10,000 organizational growth scenarios across five growth trajectories (organic, IPO, acquisition, regulatory event, crisis) confirms that cage formation is emergent and inevitable under conditions of scale and legal formalization.
The simulation models individual agents making locally rational decisions about formalization, documentation, and process creation. Cage formation emerges from the cumulative effect of individually rational decisions without any agent "intending" to create the cage.
Simulation Architecture
Agents are modeled with three properties: decision variance (the range of decisions the agent is willing to make without precedent), compliance sensitivity (the degree to which the agent's decisions are constrained by precedent), and communication fidelity (the degree to which the agent's communications accurately transmit information).
Organizations are networks of agents with defined authority structures. Growth adds agents; formalization events (IPO, regulatory compliance, legal exposure) modify agent properties in ways that reflect the fiduciary pressure of those events.
Key Findings
- Cage formation is universal: In 97.3% of simulated growth scenarios, cage properties emerge by organizational scale 150
- Speed varies by trajectory: IPO trajectories show cage formation 3.2x faster than organic growth trajectories
- Mirror structures delay but don't prevent: Agents with meta-compliance properties reduce cage formation speed by 23% but do not prevent ultimate cage formation at scale
- Divisionalization resets: Divisionalization events reset cage depth counters, but cage formation resumes immediately in each division
- Founder effects decay: Founding agents with high decision variance provide cage protection that decays within 2-3 generations of agent turnover
Implications
The simulation confirms the cage framework's core prediction: the cage is not a failure of organizational design but an emergent property of scale and formalization. It also confirms the mirror framework's prediction: meta-compliance structures can partially counteract cage formation but cannot eliminate it.
The policy implication is not to prevent cage formation (which is impossible) but to maintain awareness of cage depth, institutionalize mirror structures proactively rather than reactively, and divisionalize deliberately when scale cage limits are reached.
Key References
A Guide for Newcomers to Agent-Based Modeling in the Social Sciences. Handbook of Computational Economics, 2, 1647-1659.
Growing Artificial Societies: Social Science from the Bottom Up. MIT Press.
Exploration and Exploitation in Organizational Learning. Organization Science, 2(1), 71-87.