Modelling Aggregate Consumption Growth with Time-Varying Parameters

Using the Family Expenditure Survey (FES) data for the United Kingdom (UK), the paper specifies and estimates a ’complete’ Hilden- brand Kneip (HK) model of consumption, extending earlier efforts that were ’partial’ in nature. As the estimated parameters in the ’partial’ HK model are time varying, the paper provides empirical evi- dence that their movements over time reflect a near unit root process. To estimate the ’complete’ HK model, the paper specifies a simple OLS model of the ’remainder term’ in the ’partial’ HK model. The remainder term in the partial HK model, which as per theory should be influenced by unobservable variables like expectation formation of households, is found to be affected by housing prices. The complete model is found to explain movements in consumption better than the partial model. Results based on bootstrap suggest that given the sampling error in the FES data, the overall fit of the model should be considered as good.