Estimation and Inference in Weakly Identified Models

In the context of socio-economic modelling observations often provide insufficient information to adequately address questions of interest. An important class of models where this phenomenon is observed is that of weakly identified models. In such models many standard statistical methods have been found to be extremely unreliable. The broad aim of this project is to develop more reliable methods of estimation and inference specifically designed to handle the problems associated with weakly identified models. A twofold strategy of improving existing techniques where possible, and developing alternative methods otherwise, will be adopted. This will result in improved techniques of modelling that will make more efficient use of sample data.