gfbarbieri/abm-hetero-feedback-mkt

Replication of Cont (2005) model.

Define the Agent

Initialize agent parameters:

1. Threshold is set to: $\text{threshold} \sim \mathcal{U}(0, 1)$
2. Strategy is set to: 0 (inactive).
3. Holdings are set to: 100.

Threshold:

The threshold is updated with a probability of s, a parameter in the model, explained as the average update frequency. $s = 0.1$ is equivalent to 10 trading days.

Strategy:

The strategy is updated according to: $$ \text{strategy} = \begin{cases} 1 & \text{if } |\text{signal}| > \text{threshold} \text{ and signal > 0} \ -1 & \text{if } |\text{signal}| > \text{threshold} \text{ and signal < 0} \ 0 & \text{otherwise} \end{cases} $$

Define the Market

Initialize model parameters:

1. Create number of agents, N.
2. Set common signal to 0.
3. Set price change to 0.
4. Set asset price $\text{threshold} \sim \mathcal{U}(0, 1)$
5. Choose the number of agents to track.

Algorithm:

1. Generates a common noise signal.
2. Updates agent strategies.
3. Executes trades.
4. Calculates price change.
5. Updates agent thresholds.
6. Records data.