Forecasting Prices of Paper Products: Focusing on the Relation Between Autocorrelation Structure and Economic Theory Available to Purchase

In this paper we introduce a relatively recent forecasting technique, which here is used to forecast prices of Swedish forest products. The technique is, like vector autoregressive models (VAR), based on the idea that time series of prices and quantities from different sectors of the forest industry typically co-vary over time. A second aim is to test whether the Swedish forest industry act as a price-taker on the world market. If they do we show that current prices should be bad predictors of future quantities, and vice versa, which in turn implies that multivariate techniques do not necessarily yield better forecasts than univariate techniques. The results show that we can not reject the “small open economy” hypothesis, i.e. that Swedish producers are price takers. This result is founded on the econometric result that the inclusion of quantities do not significantly improve the forecasts of prices.