축산관측사업 정밀도 제고 및 정책연계 프로그램 구축 연구 :한우 수급전망모형 구축연구

Abstract

최근 들어 동시다발적인 자유무역협정(FTA) 추진 등으로 축산물 시장개방의 확대가 가속화되고 있으며, 쇠고기 가격 약세, 사육두수 과잉 문제, 국제 곡물가격의 급격한 변동 등이 지속되면서 한우 농가의 경영 압박요인으로 작용하고 있다. 이러한 어려움을 타개하기 위하여 생산자와 정부는 한우 수급과 가격안정을 위한 다각적인 노력을 기울이고 있으나 선제적 대책 마련을 위한 중장기 한우 시장전망 자료의 부재로 어려움을 겪고 있다.

농축산물 수급전망자료는 농가의 합리적인 의사결정, 농가소득 안정, 정부의 수급조절정책 수립 및 시행을 위한 기초자료로 활용도가 높지만, 생산 및 수요의 변동성이 큰 농축산물의 경우 신뢰성 있는 예측정보를 확보하기 매우 어렵다. 시장 예측정보의 평가는 다면적 측면에서 이루어질 수 있으나, 예측정보의 질적 가치를 결정하는 가장 중요한 평가 요소는 예측의 정확성이다. 따라서 현재 한우 농가들이 겪고 있는 어려움을 극복하는 데 도움이 될 수 있는 선제적 정책수단을 마련하기 위해서는 정확도가 높은 중장기 시장정보의 확보가 시급하다.

본 연구에서는 2년 이상의 한우 사육두수와 가격 전망이 가능한 수급전망모형을 구축하였으며, 본 모형의 결과는 향후 한우 농가의 경영목표 수립 및 변경 등의 의사결정과 한우 수급안정정책을 수립하기 위한 기초자료로써 사용될 수 있을 것으로 기대된다.

Life cycle of Hanwoo (native Korean cattle) is longer than other livestock. In general, Hanwoo is slaughtered at the age of more than two years. Therefore, a policy for market stabilization should be implemented at least two years before that an exceed-supply or -demand is anticipated. Currently, this is difficult to implement a policy prior to detecting market problems due to the lack of future market information. Therefore, this study aims to build a model for forecasting Hanwoo market for the total number of raised Hanwoo, the number of slaughtered Hanwoo for each sex, and of raised Hanwoo in terms of each sex and age, the wholesale price of Hanwoo for each sex, the farm prices of calves 6~7 months old for each sex, the farm prices of adult cattle (600 kg) for each sex. We identify the structure of beef market and the factors affecting the market supply and demand based on the previous research and the opinions from experts in the various fields related to Hanwoo and beef market. Equations composing the forecasting model are estimated by Excel and statistical softwares such as STATA and EViews. Forecasting error of each equation is examined by the measure of Mean Absolute Percentage Error (MAPE). The estimated results are reviewed by experts in this field. The model is composed of a supply part, a demand part, and a price part. The model for the supply part is constructed based on the ecological equations which reflect livestock raising periods and features of production thereof. The model for the demand part is constructed based on the factors including seasonal consumption that affect beef consumption. The model for the price part is composed of equilibrium price (wholesale price) and other price equations. The equilibrium price is derived from the market clearing condition and other prices are estimated using a statistical software. Individual equations which are components of the supply part of the model are estimated in the manner to minimize the sum of the absolute deviations (least absolute deviation) by using Excel’s solver. Other equations are estimated using a statistical software. MAPE is then used as the measure of verifying the prediction accuracy of the individual equations. The results of prediction accuracy verification reveal high prediction capability of the estimated equations. In particular, the average MAPE of the estimated equation for the number of raised Hanwoo is only 1.12%. The average MAPE of the other equations estimating the number of animals is also very low (1.17%~2.42%), with the exception of the estimated equation for the number of raised bulls older than 2 years old (the average MAPE: 4.94%). The average MAPE of the estimated equations for the number of slaughtered Hanwoo for each sex are 6.04% and 4.93% for female and male, respectively, which are a relatively low prediction accuracy in comparison with the estimated equations for the number of raised Hanwoo for each sex and age group. It is because the number of slaughtered Hanwoo is affected by various external factors including farmer’s profits, expected returns in the future, and government’s policy for beef cattle market. On the other hand, the number of raised animals is mainly determined by means of a standardized management system. The average MAPE of the price equations is also very low (3.34%~6.20%) which implies that the estimated equations have high prediction accuracy.

Researchers: Jae-Sung Cho, Woo-Jin Song, Yong-Geon Lee, Jung-hyun Yoon Research period: 2013. 8. ~ 2014. 3. E-mail address: kor0025cho@krei.re.kr