Hunter, John
(1990)
Dynamic modelling of expectations with particular reference to the U.K. labour market.
PhD thesis, London School of Economics and Political Science.
Abstract
The thesis deals with the analysis of dynamic econometric models which includes Time-Series methods. Conditional modelling and the General to Specific approach combined with Destructive testing. The modelling strategy used is dependent on the requirements of the modeller, whether he needs to Forecast, to derive policy or to produce results to support or deny a particular theory. Expectations introduce dynamics into econometric specifications and rational or consistent expectations models in particular have number of representations, which depend on the form of the inter-temporal optimisation problem and the method of solving for the expectations. Here we use the Vector AutoRegressive(VAR) form to estimate predictions of variables which are exogenous, an Errors-in-Variables method to produce initial estimates of structural parameters and a recursive systems approach to estimate the backward-forward representation. Vector autoregressive models of manufacturing wages, output prices, manufacturing inventory accumulation and vacancies are estimated using a general modelling strategy to derive predictions and one step ahead forecasts. These results are then fed into a structural model of output and employment which is estimated using a recursive estimation technique that solves out the endogenous expectations and then replaces the exogenous ones using the Wiener-Kolmogorov prediction formula. Finally we discuss generalisations of the first order rational expectations model to produce first order euler conditions which bear a closer correspondence to estimated error correction models with which they are related. The inter-temporal optimisation problem is extended to deal with lags and leads on exogenous variables, non-separabilities and lags on adjustment costs. Local and Global Identification conditions are presented for all of the models in the study.
Actions (login required)
 |
Record administration - authorised staff only |
Download statistics
Download statistics