We develop an agent-based model of a financial market which is able to jointly reproduce many of the stylised facts at different time scales. These include properties related to returns (leptokurtosis, absence of linear autocorrelation, volatility clustering), trading volumes (volume clustering, correlation between volume and volatility), and timing of trades (number of price changes, autocorrelation of durations between subsequent trades, heavy tails in their distribution, order-side clustering). Our model combines heterogeneous boundedly rational agents, endogenously activating on the basis of market events, with realistic assumptions on market microstructure. In particular, we introduce a strict event scheduling borrowed from the EURONEXT exchange. We study the model in a bottom-up fashion under alternative scenarios regarding the sophistication of agents’ strategies. These scenarios allow us to disentangle the role of microstructure characteristics from trading behaviour in the emergence of market statistical properties. Our results reveal that traders’ endogenous activation is crucial to jointly reproduce most of these properties. The ability of the model to replicate the main stylised facts of financial markets proves that it can be fruitfully used by policymakers as a test-bed for regulatory experiments aimed at improving market outcomes at different time-scales.
- Agent-based artificial stock markets
- High-Frequency Trading
- Intra-day financial dynamics
- Market microstructure
- Stylised facts