Under normality, least squares is efficient. However, if the errors are not normal, we can gain efficiency from the assertion that higher moments do not depend on the regressors. In this paper, we show how the assumption that higher moments do not depend on the regressors can be exploited in a...

The usual cointegration tests often entail nuisance parameters that hinder precise inference. This problem is even more pronounced in a nonlinear threshold framework when stationary covariates are included. In this paper, we propose new threshold cointegration tests based on instrumental...