Optimal Hedging Strategies for the U.S. Cattle Feeder

Abstract

Multiproduct optimal hedging is compared to alternative hedging strategies as applied to a Midwestern cattle feeder. One-period feeding margin hedge ratios are estimated using weekly cash and futures price data from a simulation of a custom feedlot for 1983 -1995. Hedge ratios are estimated using the last 4 years, 6 years, or all prior data available at the moment of estimation; the ratios demonstrate less variability as the length of the underlying sample increases. Hypothesis of all hedge ratios being equal to each other, that leads to the proportional hedging model, is rejected. Means and variances of hedged feeding margins using the computed hedge ratios suggest that there is no consistent domination pattern among the alternative strategies. For the ratios computed based on all prior data available, all strategies are on the efficient frontier, leaving the hedging decision up to the agent's degree of risk aversion. All hedging strategies are shown to significantly reduce the feeding margin's means and variances compared to no hedging, with variance reduction always exceeding 50 percent. Whether a producer chooses multiproduct, single-commodity, or proportional hedge ratios is sensitive to the dataset and its size.

Determining the effectiveness of optimal time-varying hedge ratios for cattle feeders under multiproduct and single commodity settings

“…The feedlot operator's margin is established in line with a previous study by Noussinov and Leuthold (1999). It is assumed that 700-pound steers are purchased by the feedlot operator and fed with 42 bushels of corn and 100 pounds of soybean meal during four months (about 18…”

“…Thus the feedlot margin considered is the difference between the sale price of slaughter cattle and the purchasing price of corn, soybean meal and feeder cattle. The feedlot operator's margin is established in line with a previous study by Noussinov and Leuthold (1999). It is assumed that 700-pound steers are purchased by the feedlot operator and fed with 42 bushels of corn and 100 pounds of soybean meal during four months (about 18 weeks), for an approximate gain of 3.2 pounds a day.…”

“…Haigh and Holt (2000) apply a BEKK MGARCH (Engle and Kroner, 1995) model for multiproduct hedging of wheat, soybeans and international freight rates, finding gains in reducing price variability in international grain trade for out-of-sample data. Noussinov and Leuthold (1999) study multiproduct hedging in a feedlot operation setting, yet they do not apply time-varying correlations to estimate optimal hedge ratios. Conversely, Sephton (1993) estimates single time-varying hedge ratios for wheat feed, barley feed and canola from Canada using a bivariate GARCH model, with constant conditional correlations, and finds improved performance over traditional (OLS) strategies.…”

Dealing with downside risk in a multi‐commodity setting: A case for a “Texas hedge”?

“…Few multi-product hedging studies have been published since their survey. Noussinov and Leuthold (1999) consider the joint hedging of feeder cattle, live (fed) cattle, corn, and soybean meal and find that the results are sensitive to the data set used. In general, they find that hedging reduces both risk and returns, and that soybean meal hedge ratio estimates are unreliable, being several times larger than one but not significantly different from zero.…”

“…Imposing a correlation structure derived from historical data on distributions of individual variables has been used in the recent literature, both for constant ratios (e.g. Noussinov & Leuthold, 1999;Wilson, Nganje, & Wagner, 2006) and for time-varying ratios (e.g. Garcia et al, 1995;Haigh & Holt, 2000).…”