Quantifying the Announcement Effects in the U.S. Lumber Futures Market

Abstract

The impact of information release from public report announcements has been widely investigated in many commodity markets, but little attention has been paid to the lumber market. In this paper, we use generalized autoregressive conditional heteroskedasticity (GARCH) models to examine the effect of two housing market reports, namely the New Residential Construction (Housing Starts) and the New Residential Sales reports, on the U.S. lumber futures market from 2000 to 2017. Our results suggest that the housing starts report indeed affects lumber market volatility, while the new home sales report exerts a minor impact. We further find that the effect of the two reports decreases with inventory levels and differs depending on the nature of the news. When the level of inventory is low, larger-than-expected housing starts has a more substantial effect than lower-than-expected housing starts. During periods of abundant stocks, however, lower-than-expected housing starts increase the volatility more than larger-than-expected news. For the new home sales reports, we find that while lower-than-expected sales do not affect lumber price volatility, larger-than-expected sales do increase the volatility.