A0250
Title: Fully modified estimation of a quantile cointegration model with a spatial lag
Authors: Christian Haefke - New York University (United Arab Emirates)
Leopold Soegner - Institute for Advanced Studies (Austria) [presenting]
Abstract: A quantile cointegration regression problem is investigated, which includes a spatial lag. We obtain the asymptotic distribution of the quasi-maximum likelihood estimator and show that the second-order bias depends on the order of the deterministic terms. We construct both a fully modified estimator, which removes this second-order bias and a Wald-type test. The finite sample properties of our fully modified estimator are analyzed in a simulation study. Finally, our estimation procedure is used to analyze growth-at-risk for a set of countries, where the cross-country spillover effects are described by means of a spatial lag.