Abstract
Empirical analyses of interlinkages and price dependencies in the forest-based sector support the forecast of market developments and the design of efficient utilization pathways. This article aims at analysing price cointegration between roundwood (sawlogs, pulpwood), sawmill by-products (sawdust, wood chips) and wood products (pellets, particle board) in the forest-based sector in Austria. Monthly price data for the period 2005–2019 were used for the following statistical tests: (1) The Augmented-Dickey-Fuller and Zivot-Andrews unit root tests were conducted to investigate stationarity of the data; (2) The Johansen Cointegration test was pairwise applied to price time series; (3) The Granger Causality test was used for cointegrated time series to examine which one is price leading. Furthermore, sawmill by-product prices were modelled as Vector Error Correction Models (VECM) to analyse their common behaviour. The dataset was divided to a training (2005–2017) and test (2018–2019) subset to assess the prediction accuracy of the models. The training data were used to estimate a VAR model as basis for forecasts, which were compared to the test data. Results show that sawdust prices are cointegrated and thus modelled with pellet and particle board prices. In contrast, wood chips are used for several applications and thus prices are cointegrated and modelled with prices of sawlogs, pulpwood, pellets and particle board. The comparison with the test data showed that forecasts were able to predict data from 2018 to 2019 well. However, a decrease in prices, starting in 2019 and intensified by the Covid-19 pandemic, could not be fully captured by these forecasts.
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have