ISSUES IN THE ECONOMETRICS OF POLICY EVALUATION (CONSUMPTION)

Over the past decade, the methodology underlying economy-wide econometric models has been subjected to increasing attack from many fronts; we will deal with the so-called Lucas critique in this study. There are two fundamental elements to Lucas\u27 attack. Firstly, the equilibrium business cycle models tend to imply a very limited role for stabilization policy, but the role of stabilization policy will only be a side issue of the present study. The second element is the way prior restrictions are used in traditional models, principally for reasons of identification and tractability. Unless expectation, which play a crucial role in the theoretical framework of all models, are formed in a very specific way, these restrictions imply that the estimated parameters are not structural but rather are estimated for partially reduced forms. Lucas, Sargent and others propose an alternative methodology which is designed to estimate parameters invariant to policy changes. There is no dispute surrounding whether the criticisms of orthodox econometric models are valid or whether it can cope with the problems involved. The unresolved issues concern how relevant the criticisms are in practice, and whether the alternative methodology constitutes an improvement over existing methods. These issues are addressed in this dissertation. Some of the relevant theoretical considerations for this debate are outlined in Chapter 2. This chapter narrows the scope of the discussion for the empirical analysis to follow into a few crucial aspects of the problem. The remainder of the study aims to provide some evidence of how important in practice the Lucas critique is. This is tackled from two angles. Chapters 3 to 5 present a detailed investigation of the effect of personal taxes on aggregate consumption using not only a standard life cycle framework, but also a decision rule from a complete optimization problem using rational expectations. The differential effects of permanent and temporary changes in personal taxes are examined under the alternative formulations. The models are subjected to a variety of specification tests, including an analysis of their behavior near changes in tax policy. Chapter 6 provides a limited study of the importance of different specifications of perceptions within a complete macro model. In particular, the case of a sharp increase in oil prices within the MPS model of the U.S. economy is analyzed both using the model as is, and using it with alternative assumptions for the key lags withing the wage price sector. The modelling of perception lags is the key element in the Lucas critique