Short term forecasting methods of international trade variables

Abstract

Forecasting monthly changes in international trade variables (volumes and prices) is a di¢ cult task given the high variability of the series. This paper proposes a number of approaches to forecast short-term changes in trade vol- umes and prices and aims, first, at ranking various forecasting methods in terms of forecast acuracy and, second, at checking whether methods fore- casting directly aggregate variables (direct approaches) outperform meth- ods based on the aggregation of country-specific forecasts (bottom-up approaches). Overall, all methods peform better than a simple random-walk benchmark. Among the forecasting approaches used, bridge equations and di¤usion indices appear to perform the best. Moreover, direct approaches outperform bottom-up ones for volume variables, while the opposite is found for trade prices. Finally, when country-specific forecasts are adjusted to match direct forecasts at the aggregate levels (top-down approaches), the forecast accuracy is neither improved nor deteriorated (i.e. top-down and bottom-up approaches are equivalent in terms of country-specific forecast accuracy).