Reservoir Computing for Macroeconomic Modelling
SOAS lead: Sophie van Huellen
Partner institutions: University College London (UCL), University of St Gallen, Konstanz University, Ministry of Finance Athens
With the increased globalisation of financial markets, financial crises and shocks have become more frequent and contagious. Despite a mass of data on financial markets and the banking sector being available, traditional macroeconomic models account for financial influences primarily through monetary aggregates and interest rates. This narrow representation of financial markets is proving increasingly inadequate.
The project draws on the expertise of an interdisciplinary research team to explore whether recent advances in machine learning, namely reservoir computing, can be applied in a novel way to improve macroeconomic modelling and forecasting by use of high dimensional financial market data.
Time series realizations in financial domains are known to share features which make them difficult to treat with conventional time series approaches. Large volumes of data, high dimensionality, and an unfavourable signal to noise ratio render many of the standard techniques infeasible. In the macro-financial context, a frequency mismatch between slow macroeconomic data and high frequency financial market data adds to the complication. Reservoir computing is a relatively recent approach to machine learning which has performed very well in similar contexts.