This project will introduce a new and general class of nonlinear varying coefficient panel data models that simultaneously incorporates endogeneity, nonstationarity and cross-sectional dependence. A new suite of efficient and robust approaches will be developed to produce accurate estimates of relationships among variables of interest, and some general large sample theory for the proposed methods will be established. This project will make a significant contribution to the study of problems of high importance in climatology, economics and finance, and enhance the existing strengths of Australian researchers in econometrics.