Bayesian and nonBayesian Techniques for Forecasting Monthly Cattle Prices
Zapata, Hector O.
This item is only available for download by members of the University of Illinois community. Students, faculty, and staff at the U of I may log in with your NetID and password to view the item. If you are trying to access an Illinois-restricted dissertation or thesis, you can request a copy through your library's Inter-Library Loan office or purchase a copy directly from ProQuest.
Permalink
https://hdl.handle.net/2142/69885
Description
Title
Bayesian and nonBayesian Techniques for Forecasting Monthly Cattle Prices
Author(s)
Zapata, Hector O.
Issue Date
1987
Doctoral Committee Chair(s)
Garcia, Philip
Department of Study
Agricultural Economics
Discipline
Agricultural Economics
Degree Granting Institution
University of Illinois at Urbana-Champaign
Degree Name
Ph.D.
Degree Level
Dissertation
Keyword(s)
Economics, Agricultural
Abstract
Econometric and time series forecasting models for monthly prices of slaughter steers 1100-1300 pounds are evaluated using mean squared error and turning point criteria. The economic model is a two-equation recursive system of supply and demand whose reduced form is used as the forecasting equation for the econometric model and is also the information set on which all other models are based. The econometric analysis is based on constant, stochastic and Bayesian estimation procedures. The univariate time series models are estimated using Box-Jenkins techniques. Vector autoregressions (VARs), classical and Bayesian, comprise the multivariate time series models. The specification of VARs is based on the Scharwz Bayesian information (SBIC), Akaike's information (AIC) and final prediction error (FPE) criteria.
Initial estimation of all models is based on the 1975-1983 period; the models are updated monthly and six-month ahead forecasts are generated during 1984-1985.
The RMSE evaluation reveals that the Bayesian VAR of order two (as selected by the AIC and FPE) model with asymmetric priors which reflect subjective decisions performs best for longer forecast horizons. Univariate ARIMA models (ARIMA(2,1,2) and ARIMA(3,1,1)) do well for short forecast horizons (one to three months ahead). The mean squared error decomposition (MSED) shows that all models become increasingly biased as the forecast horizon lengthens; the bias of the subjective BVAR being close to zero for one to three months ahead and about one-third of the bias for the ARIMA models for the other forecast horizons. All the econometric models do not perform well as evaluated by the RMSE and MSED.
The turning point evaluation indicates that the univariate ARIMAs and the subjective BVAR closely follow the movements of actual prices for the one month-ahead forecast but this deteriorates appreciably as the forecasting horizon lengthens.
Use this login method if you
don't
have an
@illinois.edu
email address.
(Oops, I do have one)
IDEALS migrated to a new platform on June 23, 2022. If you created
your account prior to this date, you will have to reset your password
using the forgot-password link below.
