A consumption function for Malta : formulation and analysis

Abstract

This dissertation sets out to give deeper insight into modelling household consumption in Malta. This is attempted by constructing and analysing two conflicting specifications of consumption modelling. These are the random walk model and the ECM. The conflict between the models stems from the fact that a pure random walk entails that any consumption changes are caused by random shocks, and maybe an element of drift. Thus, consumption cannot be forecasted under this model. The ECM on the other hand, is based on various explanatory variables which could be utilised in a forecast. The dissertation shall construct the two specifications and analyse their economic and econometric features. The study will then go on to build a forecasted series of consumption, based on the ECM, and test the strength of the forecast. If the forecast is viable, consumption can be forecasted and the random walk hypothesis is dropped. One result of the study is that consumption in Malta can be modelled as a pure random walk. This model was proven to be econometrically weaker than the ECM by diagnostic tests. Other than this, the forecasted series proved to be significant and so the random walk model is seen to be less relevant. The study also sheds light on the importance of consumer perceptions in modelling consumption. This is due to the fact that consumer confidence and price expectations were found to be the main short run drivers of Maltese consumption.