The purpose of this project is to analyze the dynamic co- movements of aggregate consumption, output, inventories of finished goods and asset returns. The analysis is conducted at both a theoretical and empirical level using recent advances in the formulation and estimation of linear and nonlinear rational expectations models. The project consists of two parts. The first part discusses classes of parametric general equilibrium models which can be used to test different theories of aggregate consumption, income and asset returns within a unified framework. The project evaluates the quantitative impact of the specification error that results when agents make consumption and portfolio decisions at intervals of time that are finer than the data sampling interval. The analysis is focused on the significance of temporal aggregation bias in the context of tests of the Permanent Income Hypothesis and the Intertemporal Capital Asset Pricing Model. The second part investigates the structural determinants of inventories, sales and relative prices. For most industries, sales Granger cause inventories but inventories do not Granger cause sales. This is puzzling from the perspective of any equilibrium theory, competitive or not, in which inventory investment and sales are jointly determined stochastic processes. The project studies the possibility that these Granger causality results are artifacts of the signal extraction problem inherent in fitting vector autoregressions to inventory and sales data which are polluted by measurement errors.
