Asymptotics of the signed-rank estimator under dependent observations

摘要

In this paper, we consider a signed-rank estimator of nonlinear regression coefficients under stochastic errors. These errors include a wide array of applications in economic literature such as serial correlation, heteroscedasticity, autoregression, etc. General conditions for strong consistency and T~(1/2)-asymptotic normality of the resulting estimator are provided.