Modelling the volatility of Maltese currency in circulation

Abstract

The aim of this dissertation is that of studying the behaviour of the mean and the volatility of currency in circulation. This dissertation will mainly focus on the theory of conditional heteroscedastic models which are implemented when volatility clustering is present. Volatility clustering in finance means that turbulent periods tend to be followed by turbulent periods while tranquil periods tend to be followed by other tranquil periods. In 1982, Engle introduced a regression model with disturbances following an autoregressive conditional heteroscedastic (ARCH) process. A useful generalisation of this ARCH model is the GARCH parameterization, which was introduced by Bollerslev (1986). Taking into consideration Engle, Bollerslev and Horvath, we shall give a brief overview of the ARCH(p) model and consider in more detail the theory regarding the GARCH(p,q) model. Next, an extensive time series analysis is applied to the study of the Maltese currency in circulation. First, we shall try to implement Box-Jenkins methodology and traditional methods of seasonal decomposition to estimate mean behaviour. Following this, due to the limitations of ARIMA models in estimating volatility, ARCH and GARCH models will be implemented to the residual series of our data with the intention of finding a model which explains the volatility of our time series.