Given that the six components of Logistic Performance Index are highly correlated, it might also be interesting to find out which of the six LPI components are most important in predicting ImportsGDP. Another interesting question is which model is best suited for predicting ImportsGDP using the six components of the Logistics Performance Index. Fortunately, the two issues can easily be analyzed using the Big Data analytical technique called LASSO (LASSO is an acronym for Least Absolute Shrinkage and Selection Operator). The nice thing with LASSO is that it allows the empirical researcher to simultaneously perform variable selection and prediction. Owing to this property, LASSO is an important tool for Big Data analysis for which there is potentially a large number of variables which make it impractical to include all the potential predictors in one regression. For this question you will use the LASSO feature of STATA 16 to answer the following questions:

(a) Identify how many of the six components of the Logistics Performance Index that best predict ImportsGDP, in addition to the Real Exchange Rate (RealExchangeRate) as one of the predictors.

(b) Identify the best model for predicting ImportsGDP using all the six components of the Logistics Performance Index i.e. what value of the tuning parameter does LASSO select? Based on the estimate of reported by STATA does it seem as if OLS is appropriate for predicting ImportsGDP? Hint: OLS is appropriate when is (close to) 0