Adaptive Markets Hypothesis

The Adaptive Markets Hypothesis (AMH), proposed by economist Andrew Lo, holds that market efficiency is not a static property of financial systems but an adaptive one — a function of the ecological conditions in which market participants operate. When environments are stable and predictable, the efficient markets hypothesis approximately holds: competition drives agents toward optimal behavior and prices toward fundamental value. When environments shift — through technological disruption, regulatory change, or macroeconomic shock — the same competitive dynamics produce maladaptive behavior: herding, panic, and behavioral bias dominate until a new equilibrium emerges.

The AMH treats financial markets as complex adaptive systems in which the strategies of participants, not just their trades, evolve under selection pressure. Strategies that worked in one regime become obsolete in the next; the survivors are not necessarily the most rational but the most robust to regime change. This evolutionary framing dissolves the false dichotomy between rationality and irrationality: both are context-dependent outcomes of the same adaptive process, just as a polar bear is superbly adapted to the Arctic and helpless in the tropics.

Market Ecology as Dynamical System

The AMH can be formalized as a dynamical system in which market regimes correspond to attractors and regime shifts correspond to bifurcations. In stable regimes — low volatility, familiar asset classes, established regulatory frameworks — the market converges to an attractor where prices track fundamentals. This is the efficient-market attractor: rational strategies dominate, arbitrage is profitable, and prices mean-revert to fair value. The basin of attraction is deep because the feedback loops are strong: successful arbitrage reinforces the strategy, which reinforces price convergence.

When the environment shifts, the attractor structure changes. A new parameter crosses a critical threshold — a regulatory change alters the payoff structure, a technological disruption creates new asset classes with no historical data, a geopolitical shock breaks the correlation structure that risk models assumed. The old attractor becomes unstable. Strategies that were optimal in the old regime are now maladaptive, not because the agents became irrational but because the fitness landscape changed. The market enters a transient phase: herding, panic, and liquidity crises are not behavioral pathologies but symptoms of the system's search for a new attractor.

This dynamical reading resolves a puzzle in the AMH literature. Lo describes markets as evolving but does not specify the dynamics of the evolution. Is it gradual, Darwinian selection of strategies? Or is it punctuated, with sudden regime shifts? The dynamical systems framework shows that both occur. Within a regime, selection is gradual: small differences in strategy performance compound over time. Between regimes, the system undergoes a bifurcation: the attractor structure changes discontinuously, and the population of strategies crashes and reconstitutes around a new fixed point.

The Connection to Causal Emergence

The AMH also connects to the theory of causal emergence in a way that has not been explored. At the micro-level — individual trades, order book dynamics, agent-level strategies — the market is computationally intractable. No observer can predict the next tick from the full micro-state. But at the macro-level — market regimes, risk factors, style factors — the system exhibits higher effective information about future states. A value