Price Volatility Modeling for the Lumber Futures Market: A Generalized Autoregressive Conditional Heteroskedasticity-Mixed Data Sampling Approach

Abstract

Lumber is one of the most essential forest products in the United States. During the first year of the COVID-19 pandemic, lumber prices almost quadrupled, and fluctuations reached record levels. Although market experts have pointed to various drivers of such high price volatility, no firm conclusions have been drawn yet. Using the generalized autoregressive conditional heteroskedasticity-mixed data sampling (GARCH-MIDAS) framework, this study assesses the potential drivers of lumber price volatility, with predictors including the Google Trends Web Search Index, housing starts, US lumber production quantity, and VIX index, representing public attention, housing demand, lumber supply, and macroeconomic concerns, respectively. We have found that housing demand is the key driver of lumber price volatility, followed by public attention. It is worth noting that US lumber supply and macroeconomic concerns have played a modest role in explaining lumber price volatility. Also, forecasting lumber price by using the housing demand variable substantially outperforms others. Market participants, including lumber mills, wholesalers, and home builders can get valuable information from the housing market to manage lumber price risk. Study Implications: The findings of this study can be used to improve hedging strategies, design option pricing formulas, and setting margin requirements. Critical information for price risk management on the lumber market can be gained by lumber market participants from the housing market. For forest management decisions by landowners, giving close attention to housing market would provide valuable information on the appropriate time for timber harvesting, because changes in the housing market affect lumber price that will indirectly affect the demand for timber, which is the most important factor of production for lumber mills.