Cluster 04
Market stability
When many agents like this one populate a market, what happens?
Allocative efficiency of a continuous double auction populated by the model, and how that efficiency collapses under shared anchors or mixed strategic conditions.
Cluster score · market stability
Healthy markets become near-empty markets when buyers anchor.
Allocative efficiency of a continuous double auction populated by each model, across treatments. Anchoring collapses efficiency by 80+ percentage points. A specific bias warning to every agent recovers near-baseline.
Double-auction efficiency by treatment and model
Allocative efficiency (% of competitive-equilibrium gains realised) when the entire market is populated by one model. The 'Anchored' treatment shows how individual anchoring bias cascades into market-level collapse — and how shared debiasing prompts almost fully recover.
Source · market_simulation · all treatments
All dimensions in this cluster
Double-auction baseline efficiency
Allocative efficiency of a continuous double auction populated entirely by this model with no behavioral nudges. Higher = closer to competitive-equilibrium realisation of gains from trade.
market_simulation · baseline mean_efficiency
Efficiency under shared anchors
Allocative efficiency when every buyer in the market is given the same anchor in their context. Higher = the market mechanism absorbs the bias rather than amplifying it.
market_simulation · anchored mean_efficiency
Efficiency when all agents debiased
Efficiency when every agent in the market receives a specific bias warning. Higher = market recovers to near-baseline; gap from baseline indicates limits of prompt-level debiasing.
market_simulation · all_debiased mean_efficiency