摘要 : The real exchange rate is a key indicator of a country's trade competitiveness in the world economy. In this study, we set up a forecasting framework to generate the multi-period random walk predictions of both real broad and narr... 展开 The real exchange rate is a key indicator of a country's trade competitiveness in the world economy. In this study, we set up a forecasting framework to generate the multi-period random walk predictions of both real broad and narrow effective exchange rates of NAFTA countries (Canada, Mexico, and the US). We show that the random walk prediction errors for 2008-2016 fail to be orthogonal to changes in oil prices. Consistent with these results, we find that oil price changes accurately predict directional change in real exchange rates for up to two (three) months ahead for Canada (Mexico and the US). The accurate directional predictions imply symmetric loss, meaning that they are of value to a user who assigns similar loss (cost) to both incorrect upward and downward moves in real exchange rates. This is important since accurate predictions of upward (appreciation) and downward (depreciation) moves in the effective real exchange rate are equally crucial for policymaking. 收起