Expectation formation in Agent Based Models of financial, credit and good markets
by Prof. Giulia Iori, City University London
Unlike conventional macroeconomic models which stress forward looking behaviour by far-sighted and rational, often representative, agents at the expense of the “plumbing” (i.e. the inter-connections) of an actual economy, ABMs have the advantage of simplifying behavior at the individual level by assuming that agents follow given but evolving rules-of-thumb, and this allows them to explore the multiplicity of agent types and their set of inter-connections in far greater detail. ABM typically assume large populations of heterogeneous agents, specifying agents individual behaviour, the environment, and modes of interaction. Agents are not only heterogenous and interacting but also adaptive; they have different circumstances, different histories and adapt continuously to the overall situation they create. To handle real-world features, it is essential to permit agents to engage in comprehensive forms of learning that include inductive reasoning (experimentation with new ideas) as well as aspects of reinforcement learning, social mimicry, and forecasting of future events. In ABM, agents can range from passive automatons with no cognitive function, to active data-gathering decision makers with sophisticated learning capabilities.
In this talk I will review some mechanisms of agents learning and expectation formation that provide the micro-fundation of economic and financial ABMs, focusing on their stabilizing/destabilizing effects on the resulting macro-dynamics of financial, credit and goods markets.