بررسی کارایی انواع الگوهای ARCH در پیشبینی و الگوسازی فرآیند قیمت گوشت مرغ

Abstract

Introduction: Poultry is an important commodity for household consumption. In recent years, price fluctuation for this commodity has caused an uncertain condition for consumers and poultry prices over the past two years has changed a lot. This has caused many changes and uncertainty in a purchase decision. Analysis of changes and volatility modeling can be a great help to predict the poultry prices and great facilities in creating appropriate policies in future. The prices of staples such as poultry consumption basket is highly variable because much of the protein is necessary for daily energy are supplied in this way to households. So when the price of chicken which has been changed over the past two years and has always been in the press and media attention, has been selected in this study. Fluctuations in price of chicken have caused a surge in consumer expectations and contributed in volatility of chicken price. Materials and Methods: In this study ARCH models have been used for daily price of poultry of Iran’s market and this was investigated for2012-13and2013-14.BecauseARCH models can model the impact of heterogeneous variance over time in time series data then the variance of time series, which is limited in time, has no time limit. Many time series are more complex than a linear patterns, thus, non-linear models are of particular importance in Economic Sciences and Econometrics. Accordingly, Engle presented that ARCH model can model the heterogeneous variance components of the error term. That is a disturbing element and modeling can help to examine and explore the relationship between the components can be found disturbing. Basically, these models fit the data to a cluster and periodic oscillations with high volatility and low volatility associated with the period. In this study, we used several different models like ARCH, GARCH, IGARCH, and TGARCH. The distribution of the error term of the model also followt-student distribution. This study shows that the heterogeneous variance exists in error term and indicated by LM-test. Results and Discussion: Results showed that stationary test of the poultry price has a unit root and is stationary with one lag difference, and thus the price of poultry was used in the study by one lag difference. Main results showed that ARCH is the best model for fluctuation prediction. Moreover, news has asymmetric effect on poultry price fluctuation and good news has a stronger effect on poultry price fluctuation than bad news and leverage effect doesnot existin poultry price. Moreover current fluctuation does not transmit to future. One of the main assumptions of time series models is constant variance in estimated coefficients. If this assumption has not been, the estimated coefficients for the correlation between the serial data would be biased and results in wrong interpretation. The results showed that ARCH effects existed in error terms of poultry price and so the ARCH family with student t distribution should be used. Normality test of error term and exam of heterogeneous variance needed and lack of attention to its cause false conclusion. Result showed that ARCH models have good predictive power and ARMA models are less efficient than ARCH models. It shows that non-linear predictions are better than linear prediction. According to the results that student distribution should be used as target distribution in estimated patterns. Conclusion: Huge need for poultry, require the creation of infrastructure to response to demands. Results showed that change in poultry price volatility over time, may intensifies at anytime. The asymmetric effect of good and bad news in poultry price leading to consumer's reaction. The good news had significant effects on the poultry market and created positive change in the poultry price, but the bad news did not result insignificant effects. In fact, because the poultry product in the household portfolio is essential, it should not fluctuate. When poultry imports decline, as well as the dependence of the poultry price to world prices declines therefore lower fluctuations of world prices are transmitted to domestic prices. Expanding poultry farms, cold storage and balance the raw materials market, would lead to less fluctuations in the poultry price industry and can be effective.