New Agent-Based Model (EDEM) Shows Markets Drift Up Without Irrationality

A new paper from researchers Mikhail Arbuzov, Sisong Bei, and Alexey Shvets introduces the Estimated Dynamic Equilibrium Model (EDEM), an agent-based framework that rethinks how supply and demand interact. Instead of assuming equilibrium, EDEM treats market dynamics as a coupled stochastic process driven by heterogeneous, noisy agent valuations. The core technical breakthrough: when market-clearing prices are repeatedly sampled from the upper tail of noisy bid distributions and then fed back as inputs for future valuations, expected prices drift upward even when estimation errors have a strict zero mean. The authors derive this order-statistic bias in closed form for uniform bids and show that compounding it across epochs leads to exponential price growth—without requiring investor optimism or irrationality. This extends Miller's divergence-of-opinion theory to a dynamic setting, recovering Walrasian equilibrium and Miller's static premium as limiting cases.

Through controlled experiments on a simulated real-estate neighborhood, EDEM reveals six distinct market regimes—ranging from band-stability to runaway bubbles—all emerging from a single agent rule set. This diversity offers a potential explanation for contradictory findings in the empirical divergence-of-opinion literature. The paper also warns that machine-learning valuation algorithms may inadvertently amplify this statistical bias, as they often recycle noisy price signals. For tech professionals, EDEM provides a rigorous mathematical foundation for understanding persistent disequilibrium in markets, suggesting that algorithmic traders and AI-driven price forecasters need to account for this endogenous drift mechanism. The full paper is available on arXiv (2605.15472).