Description

We apply Machine Learning (ML) methods with Big Data, similar to Medeiros et al. (2019) for the US, to forecast headline and core inflation of the CPI in Chile. We document that the ML methods do not consistently dominates in the inflation forecast for the Chilean case over simple and univariate linear competitors such as AR, the mean and median of past inflation, which have proven to be highl competitive. In fact, these are the best performing methods in many cases. A second contribution of this work is the construction of a large data set with macroeconomic variables related to the Chilean economy, similar to McCracken and Ng (2016), who built (and maintain) a similar data set for the US.