We test the economic theory that price differences in spatially separated markets will be equalized through arbitrage activity using time series data on housing frame lumber prices in different regions of the United States. Wildfires, hurricanes, and other extreme weather events, as well as large swings in regional housing demand, can create shocks to geographically different lumber markets. The degree to which different lumber producing markets are interrelated is of interest to better understand how prices will respond to such shocks. Linked markets will realize adjustments that keep prices as equal as possible while allowing for differences directly related to the transfer costs necessary for arbitrage activity. Overly simplified tests will seem to support a lack of price transmission between markets or lag times between price adjustments that seem to contradict standard theory. This paper implements non-linear threshold models and non-parametric estimation techniques to demonstrate a more detailed price-linkage relationship that is better supported by economic theory.