Forecasting Macroeconomic Indicators for Eurozone and Greece : How Useful are the Oil Price Assumptions?

This study evaluates oil price forecasts based on their economic significance for macroeconomic predictions. More specifically, we first use the current state-of-the-art frameworks to forecast monthly oil prices and subsequently we use these forecasts, as oil price assumptions, to predict eurozone and Greek inflation rates and industrial production indices. The macroeconomic predictions are generated by means of regression-based models. We show that when we assess oil price forecasts, based on statistical loss functions, the MIDAS models, as well as the futures-based forecasts outperform those generated by the VAR and BVAR models. By contrast, in terms of their economic significance we show that none of the oil price forecasts are capable of providing predictive gains for the eurozone core inflation rate and the Greek industrial production index, whereas some gains are evident for the eurozone industrial production index and the Greek core inflation rate. However, in all cases the oil price forecasting models, including the random-walk, generate equal macroeconomic predictive accuracy. Thus, overall, we show that it is important to assess oil price forecasting frameworks based on the purpose that they are designed to serve, rather than based on their ability to predict oil prices per se