The research consists of two topics, both related to the development of business cycle models which are formulated from micro-economic principles. The first part develops a class of stochastic growth models that combine the most appealing features of both the representative agent and overlapping generations models. Like the representative approach, the modelling concepts (consumption, income, etc.) allow an econometrician to directly match up theoretical constructs with real world data from aggregate time series. But this model will be much less restrictive than the representative agent paradigm. The second part uses a popular model of periodic phenomena to explain business cycles. The model was first used by Duffing in his study of a pendulum that is forced by a regular motion and it has not previously been studied by economists. Duffing's model, when applied to business cycles, suggests that seasonal fluctuations in tastes or endowments may be the source of apparently random fluctuations at business cycle frequencies. This model has the advantage of explaining the erratic movement in economic time series, rather than assuming it, and unlike previous applications of chaotic dynamics to economics, it is not restricted to systems which change stability as one moves away from stationary equilibrium.
