THESIS FORECASTING FED CATTLE PRICES: ERRORS AND PERFORMANCE DURING PERIODS OF HIGH VOLATILITY Submitted by
FORECASTING FED CATTLE PRICES: ERRORS AND PERFORMANCE DURING PERIODS OF HIGH VOLATILITY Livestock and other commodity prices have displayed considerable volatility in the past ten years. In this environment, price forecasts play a key role in producers' business planning and risk management decisions. The object of this study is to evaluate fed cattle price forecasting performance and errors during this volatile period. Price forecast models are developed using autoregressive, vector
... , vector autoregressive, and vector error correction frameworks. Forecast performance is compared to the live cattle futures market. Results emphasize the importance of simplicity relative to forecast accuracy. Autoregressive and vector autoregressive methods appear the most useful, with autoregressive models typically being the most accurate of the time series methods. Time series models are significantly more accurate than futures predictions at the one-month horizon. Futures are about as accurate or more accurate at all other horizons, especially as forecast horizon increases, although differences are not significant. Time series methods still provided valuable information relative to futures-based predictions at the two-to six-month horizons. Results suggest forecast errors are related to shocks occurring after the forecast, consistent with market efficiency. Shocks related to market currentness, or the relative supply and demand conditions of the non-storable commodity, appear the most important to fed cattle price forecasting errors.