Rational expectations models of macroeconomics that incorporate asymmetric information typically use a linear-quadratic technical framework because individuals' predictions or decisions are then linear (and hence tractable) functions of observable variables. If individuals have asymmetric information and obtain information by observing endogenous variables -- prices, aggregate output or the like -- then it becomes technically difficult to solve such models for the endogenous aggregates. The standard method, the method of undetermined coefficients, is limited to special cases. The investigator developed two promising new solution methods and in this project he expands them to a more general setting. He will show that all linear rational expectations models with asymmetric information are expressible as special cases of a general structure in which economic, informational, and statistical structure is captured by specifying the dimensions and elements of matrices of analytic functions. Any economic structure can be rigorously modelled and empirically estimated by specifying the details of these matrices. Solving this model requires factoring the matrix. The first approach designs the model so that the matrix is comprised solely of exogenous elements -- which are easily factored -- and then perturbing the model to achieve the solution of the economically interesting model. The second approach uses a contraction mapping fixed point method. The model will integrate the informational externality micro literature with the rational expectations literature. This paves the way for further analysis, particularly of the welfare characteristics of alternative policy regimes, using what we know from the micro literature. The investigator has already shown how informational externalities lead to aggregate fluctuations that mimic business cycles for special cases of stochastic driving processes. This project investigates how the persistence, amplitudes and co-movements of economic aggregates are affected by such externalities when the driving processes are more general.
