Testing for Nonlinear Spatial Integration in Roundwood Markets
Prior studies have frequently not supported spatial market integration and the law of one price in roundwood markets. However, these studies have relied on cointegration tests that have low power in the presence of nonlinearities. Roundwood markets may display these effects due to high transactions costs between regions. In this article, we revisited the question of spatial integration in roundwood markets by using tests that have no cointegration as a null hypothesis and threshold cointegration as an alternative. For testing, a dataset of monthly New Zealand roundwood prices from four regional markets over the period from January 1995 to April 2010 were used. Similar to standard linear cointegration tests, the null hypothesis of no cointegration could not be rejected for most of the regional price pairs. One interpretation is that regional markets are segregated. Another interpretation is that the power of these nonlinear tests, although an improvement over that of their linear counterparts, is still relatively low.
No Supplementary Data
No Article Media