My main area of interest is in Agent-Based Computational Finance (ACF) and more specifically models of fundamentalists and technical analysts.
The Fundamentalists vs Chartists model design is highly motivated by observations made on how real financial traders behave. Empirical evidence shows that, by large, there are two kinds of forecasting behaviour in the market. First, we have the fundamentalist who act as a stabilizing force. These agents base their decisions on rational expectations and rely on powerful economic concepts. They believe the price will move towards a long term equilibrium or a fundamental value. This is similar to a mean-reverting strategy. On the other hand, there are chartists, also called technical analysts. These agents forecast the future prices by modelling historical data.
An important research direction in the field is presenting the ACF with econometrics. This is the application of mathematics, statistics and computer science to economic data in order to provide empirical explanations to economic relations. More specifically, the researchers use econometrics to analyse the data generated by financial models and check if it is able to display a number of frequently observed features of real markets. The challenge is to design and calibrate a model such that the data it generates matches a series of statistical properties of real data known as “stylized facts”. This is particularly important for validating the model.
The project I am working on consists of a Heterogeneous Agent-Based Financial Model that extends the simple Fundamentalists vs Technical Analysts models as an evolving system of autonomous interacting agents.
My objectives are matching a richer set of stylized facts of real life financial markets while also provide an explanation of their existence, observing what determines agents to switch their strategies and finding a good policy at the individual level.