Abstract

“Knowing the factors that influence the secondary refined copper supply behavior has been fundamental in generating a copper market model and developing public and corporate policies. When analyzing the explanatory variables used in the existing models in the literature, it is possible to observe high variability in estimating the parameters when modifying the availability of information or changing the observation period. Based on this, we argue that only some explanatory vari- ables will have robust estimated parameters, which means that they are unbiased, stable (i.e., they do not vary significantly when the specification of the equation or the number of observations changes), and with asymptotic convergence over time. This work defines and validates a method to select robust explanatory variables capable of quantifying the refined second- ary supply of copper (or any other variable) in a given period. Using a database with 23 explanatory variables in the period 1960–2017, we characterize the estimated parameters with high and low robustness, thus supporting the proposed hypothesis. The results obtained allowed identifying those variables with low uncertainty in estimating their parameters, with a high statistical significance, and with a low standard deviation. This allows to obtain a robust function for the secondary refined copper supply in the long term, capturing essential elements of reality”.