9414083 Bai This project develops a new and unified econometric theory for the testing and estimation of multiple structural changes in a variety of econometric models where changes in the parameters of the model occur at unknown times. Recent developments in economic theory as well as in empirical applications have shown the need for an econometric theory of multiple structural changes. Although the case of a single structural change has received considerable attention, research on the subject of multiple structural changes is scant. The goal of this research is to develop on a coherent basis the econometric theory of multiple structural changes by extending previous work of the principle investigator in this area. This project answers a number of questions pertaining to multiple structural changes. Among the issues addressed are: 1) Determination of the changes of regimes (periods during which the parameters of a model do not change) in a given data set; 2) Testing the hypothesis of no structural shift versus that of multiple structural shifts, And, more generally, testing k shifts versus k+l shifts; 3) Determination of critical values and construction of standard reference tables; 4) Estimation of models with multiple structural changes and determination of the criteria for estimating the shift points; 5) Solving the computational problem for multiple-shift estimation; and 6) Deriving the statistical properties of the resulting estimators and the estimated shift points.
