Aggregation and time series implications of state-dependent consumption

Aggregation involves two issues. One is about the individual consumer behavior. The other one is about compositional distribution of population. The difficulty in studying aggregation is caused by the fact that theoretical and econometric models of individual behavior do not contain information using features. This dissertation uses the concept of state-dependent consumption, that focuses on the processes of and constraints on human activities instead of preferences, to consider aggregation and time series implications. Empirical evidence is also provided using the US data from Consumer Expenditure Survey during 1980-89.