CMStatistics 2018: Start Registration
View Submission - CFE
Title: Out-of-sample performance of nonlinear models in commodities international price differential forecasting Authors:  Nicola Rubino - University of Barcelona (Spain) [presenting]
Abstract: An analysis of a group of small commodity exporting countries' price differentials relative to the US dollar is presented. Using unrestricted self exciting threshold autoregressive models (SETAR), we model and evaluate sixteen national consumers' price index (CPI) differentials relative to the US dollar CPI. Out-of-sample forecast accuracy is evaluated through calculation of mean absolute errors measures on the basis of monthly rolling window and recursive forecasts and extended to three additional models, namely a logistic smooth transition regression (LSTAR), an additive nonlinear autoregressive model (AAR) and a simple neural network model (NNET). Our preliminary results confirm presence of some form of non linearity in the majority of the countries analyzed. The parsimonious AR(1) model does not appear to perform any worse than any nonlinear model in the rolling sample exercise. However, its validity in terms of a long run equilibrium driven by purchasing power parity is undermined by the results of the recursive estimates and the outcome of the Diebold-Mariano type tests, which more generally favor the Heckscher commodity points theory. As a policy advice to commodity exporting countries, we find no apparent reason to suggest commodity export price pegging as a generalized foreign exchange policy.