This paper introduces bootstrap neural network pure significance tests for the no cointegration hypothesis against nonlinear cointegration alternatives. The theoretical properties of the tests are discussed and a Monte Carlo investigation of their small sample properties is undertaken.

This paper corrects some points in the appendix of the paper: George Kapetanios (2003) "Bootstrap Neural Network Cointegration Tests Against Nonlinear Alternative Hypotheses", Studies in Nonlinear Dynamics & Econometrics: Vol. 7: No. 2, Article 2.

Using a new methodology that allows nonlinearities, we find frequent support for external debt sustainability in a number of Latin American countries. Our findings reverse the results for several countries, obtained with traditional unit-root tests and present a richer framework for evaluating...

This paper introduces bootstrap neural network pure significance tests for the no cointegration hypothesis against nonlinear cointegration alternatives. The theoretical properties of the tests are discussed and a Monte Carlo investigation of their small sample properties is undertaken.