Resumen:
This article investigates the statistical properties of the economic complexity data in a group of 29 OECD countries, testing for their degree of persistence by looking at the order of integration of the series from a fractional integration viewpoint. The results vary substantially depending on the assumptions made on the error term; in particular, mean reversion is only found in the case of Chile if the errors are uncorrelated; however, if autocorrelation is permitted, mean reversion is found in a group of ten countries, namely, Australia, Canada, Colombia, France, Greece, Ireland, Israel, Norway, New Zealand and South Korea. Robust estimation is conducted by means of non-parametric methods and non-linear structures, which are also incorporated in the model. Policy implications are derived at the end of the article.
