An information-based multi-asset artificial stock market characterized by different types
of stocks and populated by heterogeneous agents is presented. In the market, agents trade
risky assets in exchange for cash. Beside the amount of cash and of stocks owned, each agent
is characterized by sentiments and agents share their sentiments by means of interactions
that are determined by sparsely connected networks. A central market maker (clearing
house mechanism) determines the price processes for each stock at the intersection of the
demand and the supply curves. Single stock price processes exhibit volatility clustering and
fat-tailed distribution of returns whereas multivariate price process exhibits both static
and dynamic stylized facts, i.e., the presence of static factors and common trends. Static
factors are studied making reference to the cross-correlation of returns of different stocks.
The common trends are investigated considering the variance–covariance matrix of prices.
Results point out that the probability distribution of eigenvalues of the cross-correlation
matrix of returns shows the presence of sectors, similar to those observed on real empirical
data. As regarding the dynamic factors, the variance–covariance matrix of prices point out
a limited number of assets prices series that are independent integrated processes, in close
agreement with the empirical evidence of asset price time series of real stock markets.
These results remarks the crucial dependence of statistical properties of multi-assets stock
market on the agents’ interaction structure.
