Can you use oil and natural gas prices to forecast the norwegian krone?

Abstract

This thesis studies if we can use a commodity prices, namely oil and gas to forecast the NOK/USD exchange using data at a monthly frequency. We analyze the commodities predictive ability through two main models. The first model uses contemporaneous commodity prices as a predictor, while the second uses lagged commodity prices. We use these models to create out of sample forecast one step ahead. Our results are compared to a random walk benchmark model using the Diebold and Mariano test statistic. We find little evidence of oil and gas prices being able to forecast the exchange rate. Only the model with contemporaneous oil prices produced significantly better results than the random walk benchmark.