Book Description
Recent research has shown that recursive real-time VAR forecasts of the real price of oil tend to be more accurate than forecasts based on oil futures prices of the type commonly employed by central banks worldwide. Such monthly forecasts, however, differ in several important dimensions from the forecasts central banks require when making policy decisions. First, central banks are interested in forecasts of the quarterly real price of oil rather than forecasts of the monthly real price of oil. Second, many central banks are interested in forecasting the real price of Brent crude oil rather than any of the U.S. benchmarks. Third, central banks outside the United States are interested in forecasting the real price of oil measured in domestic consumption units rather than U.S. consumption units. Addressing each of these three concerns involves modeling choices that affect the relative accuracy of alternative forecasting methods. In addition, we investigate the costs and benefits of allowing for time variation in VAR model parameters and of constructing forecast combinations. We conclude that quarterly forecasts of the real price of oil from suitably designed VAR models estimated on monthly data generate the most accurate forecasts among a wide range of methods including forecasts based on oil futures prices, nochange forecasts and forecasts based on models estimated on quarterly data.
