Behavioural biases and evolutionary dynamics in an agent-based financial market

This research is devoted to the study of financial market dynamics in a framework
which combines agent-based modelling and concepts from behavioural finance.
The thesis explores, in an agent-based financial market model, the interlinkage
between investor heterogeneity, bounded rationality, behavioural biases
and the aggregate market dynamics.
We develop a dynamic equilibrium model of a financial market in the presence
of heterogeneous, boundedly rational investors. The model combines a
performance-driven strategy-switching mechanism of an adaptive belief system
(Brock and Hommes, 1998) and an evolutionary finance model (Evstigneev, Hens
and Schenk-Hopp´e, 2011). A key feature of this new model is that it contains
a combination of passive and active learning dynamics. Passive learning refers
to the market force by which wealth accumulates on investment strategies which
have done relatively well. Active learning refers to the switching behaviour by
which investors actively move their wealth into strategies which have performed
well in the recent or distant past. This thesis extends the literature by examining
the joint effect of passive and active learning in relation to the evolutionary
dynamics of financial markets.
By drawing in concepts from behavioural finance, we focus on the micro-level
modelling of various heuristics and behavioural biases which may affect investors’
active learning and financial forecasting, such as overconfidence, recency bias,
sentiment, etc. We quantify the macro-level market impact of these behavioural
elements and study the evolutionary prospects of market dynamics.
We show that the interaction between passive and active learning is crucial to
understanding the market selection of dominant strategy or the survival of different
strategies. Investors’ bounded rationality and behavioural biases in active
learning and financial forecasting play an important role in shaping the market
dynamics. Our findings point to the causes of the persistence of market inefficiencies
and a variety of stylised facts of financial market. The added value of
drawing together agent-based modelling and behavioural finance on the study of
financial markets dynamics is demonstrated.