استخدام نماذج VAR في التنبؤ بالمساحات المزروعة بمحصول الذرة الصفراء في العراق للمدة (2011-2020)

Abstract

In many countries, maize is considered as one of the most important cereal crop in nutrition and industry. That is because this crop is essential material for forage in animal breeding projects. This study aimed to predict the maize-cultivated acreage (MCA) in Iraq for the period 2011-2020 using time series analysis based on autoregressive vector models. To achieve this aim, main supply response factors of MCA were determined which were maize price, competitive crop (cotton) price, irrigation water and total yield. Augmented Dickey Fuller (ADF) test was used for detection of unit roots among variables. The result of this test indicated nonstationarity of time series at their levels and their stationarity at their first differences, and they were all cointegrated. In order to estimate the effect of short- and long-term determinants on MCA, error correction vector was used, where the results revealed significance of variables which represented 88% of the change in MCA. The analysis of dynamic relationships among variables was based upon the fractionation of variance and impulse response functions. The variance of prediction error was fractionated for prospective 10 years using Cholesky criterion. The results indicated increasing the importance of shocks in maize price, competitive crop price and irrigation water in the explanation of fluctuations in the MCA in long-term. The impulse response was used to illustrate time track of variables as they response to different shocks in different scales using one standard deviation shock. Error correction vector model was used for prediction for the period 2011-2020 where Theil inequality coefficient and root main square error for judgment on the prediction efficiency of static and dynamic models at levels and difference. The results revealed that the dynamic models had better prediction power at difference compared to other models, and the MCA for the studied period was swinging among its normal mean