A0733
Title: Instrumental variable estimation and inference for spatial autoregressive geographically weighted quantile regression
Authors: Vivian Yi-Ju Chen - National Chengchi University (Taiwan) [presenting]
Abstract: Past years have witnessed significant advancements in spatial modelling techniques that allow simultaneously dealing with spatial heterogeneity in the regression coefficients and the spatial autoregressive lag in the response variable. Spatial autoregressive geographically weighted quantile regression (GWQR-SAR) is one such technique that has recently been devoted to the literature for conducting spatial quantile-based analysis. GWQR-SAR is proposed as a new estimation method, termed instrumental variable quantile estimation. The associated inference properties are also derived, which offer a covariance matrix estimate that is simpler to construct compared to the existing method. To strengthen the theoretical framework of GWQR-SAR, bootstrap tests are further developed to identify constant parameters, as well as the semiparametric modelling framework. The proposed methodologies are then evaluated through simulations. Lastly, an empirical example is given to illustrate the application of the approach.
