If you want to stop a huge crisis, sometimes the best move is for the people in charge to actually give up some of their power.
March 26, 2026
Original Paper
Chaos-Bound Autonomy: A Cross-Domain Simulation Study Of Bounded Governance, Metric Divergence, And Systemic Stability
SSRN · 6453058
The Takeaway
Simulations of migration systems and financial markets show that the most stable outcomes occur when institutional authority is 'non-increasing' as risk rises. Traditional heavy regulation and 'clamping down' during high-risk periods actually increased the probability of catastrophic swings and slowed down the system's ability to self-correct.
From the abstract
Complex socio-technical systems exhibit recurring instability when autonomous agents optimize locally within institutional environments whose feedback loops fail to detect accumulating systemic risk. We introduce and empirically test the Chaos-Bound Autonomy (CBA) framework-a dynamical systems governance architecture that formalizes bounded autonomy as a tuple F = (Ω s , U s , ρ, A, E), wherein system authority is non-increasing in composite risk. Using agent-based simulations calibrated to empi