CONSUMERS' expenditure forms a major component of the Gross National Expenditure and the parameters of this function play a crucial role in the mechanism of income generation in a macro-model. Because of their importance, a considerable amount of care and attention should be devoted to the estimation of these parameters and their probable movement. This consideration becomes especially important in mediumand long-range forecasting because over a long-run horizon the consumer expenditure parameters may be strongly influenced by socio-economic factors such as female participation in the labour force, education level of the population, urbanization, etc. In this paper we make a beginning in the direction of estimating parameters of the consumption function which are shifting over time due to these socio-economic factors. The importance of taking into account the shifting parameters in a macro-equation can also be emphasized from another point of view. It is becoming increasingly clear that macroeconomic analysis and policy will be seriously inadequate in a medium-run horizon unless attention is paid to changes in the composition of aggregative units. One line of attack is, of course, to have disaggregative models. However, quite often the detailed time-series data are not available on all the variables concerned. Even when the data are available, they are somewhat unreliable at a disaggregate level and in statistical estimation the ratio of noise to signal tends to be high in such disaggregate equations. Moreover, there are often some overall macro-constraints which are difficult to take into account in disaggregate models. It is, therefore, desirable as a supplement (if not as an alternative) to the disaggregated models, to have a macro-equation capture the compositional effects as far as possible. Our DaDer is one example of an attempt to capture the influence of changing population composition in a macro-equation, by introducing shifting parameters. For purposes of estimating these shifting parameters, we make combined use of cross-section and time-series data. The lagged reaction path of consumption due to income and price changes is also considered, and we use the Almon technique (1965) for this purpose.