농업부문 전망모형 구축 연구

Abstract

한국농촌경제연구원이 한국 농업부문의 계량경제학적 부분균형모형인 ASMO (Agricultural Simulation Model)를 1995년에 개발하여 운용한지 10여년이 지났다. 그간 이 모형은 농업 거시지표들의 중장기 전망에 널리 이용되어왔으며, 매년 통계자료 갱신과 모형 정비가 지속적으로 이루어져왔다. 날로 확대되는 농식품 수입 등 시장환경 변화에 따른 세부품목별 효과분석이 요구되고, 그에 대응한 정책의 효과분석에 대한 연구 수요도 늘어나고 있다. 그러나 ASMO는 정책실험 보다는 중장기전망에 주목적을 두고 개발된 모형으로서, 개발 당시 통계자료 미비로 인하여 국내산 위주의 품목군별로 수급 추정이 되도록 짜여져 있어, 세부 품목들의 수급구조분석과 정책실험 등에 한계가 있었다. 이 연구의 목적은 한국농촌경제연구원의 한국 농업 부분균형모형인 ASMO(Agricultural Simulation Model)를 갱신하여 한국 농업거시지표들의 중장기 전망과 정책실험능력을 개선하였으며, 새로운 모형을 KASMO(Korea Agricultural Simulation Model)로 명명하였다.KASMO는 ASMO의 기본틀인 거시경제, 농업요소, 재배업, 축산, 농업총량의 5개 부문을 유지하면서, 다음의 사항들을 수정하였다.첫째, 첫째, 재배업부문에 있어서 기존의 품목군별 경지배분모형을 개별품목별 재배면적반응함수로 전환하였다.둘째, 1990년대 중반의 ASMO 개발 당시 수출입실적 자료제약으로 수입수요함수가 추정되지 못하여 품목별 수급모형을 국산청산가격과 수입가능가격 중 낮은 가격이 국내가격이 되는 주도가격구조로 구축하였다.셋째, ASMO에서는 총공급량이 역수요함수에 대입되어 전량 소비되는 청산가격이 도출되며, 이 가격이 차기 재배면적을 결정하는 축차(recursive) 방식이었다. KASMO에서는 이를 수급동시균형방식으로 전환하였다. 넷째, ASMO는 회계년도 기준의 연도별 시계열자료를 이용하였다. KASMO에서는 곡물과 과일에 대하여 유통연도 기준으로 전환하여 수급함수의 설명력을 높였다. 다섯째, 농업요소와 농업총량부문에 있어서, 농가소득과 농가인구 등이 국제 유가, 환율, GDP 성장률 등에 민감하게 반응하도록 방정식 구조를 개편하였다.마지막으로, KASMO의 재배업부문은 곡물, 채소·과채, 과일, 특용·기타작물의 하부모형으로 세분류되고 축산부문과 함께 품목간 재배면적 경합성 및 수요 대체성 등의 관계에 의해 상호 연결되어 있다.

KREI-ASMO (Korea Rural Economic Institute - Agricultural Simulation Model) was developed by the Korea Rural Economic Institute in 1995, and has been used for mid- and long-term outlook as well as to analyze external shocks or various alternative policies for the Korean agricultural sector. The model is a partial equilibrium model that adopts a structural/recursive simulation method. KREI-ASMO includes five sub-models; 1) sub-model for forecasting of macro-economic variables, 2) agricultural input price sub-model, 3) plants sub-model, 4) livestock sub-model, and 5) sub-model for agriculture macro variables. This 2-years research project aims to version-up the KREI-ASMO, and we named the new model KASMO(Korean Agricultural Simulation Model). Main features of KASMO are; 1) replacing the supply and demand estimation for 18 large commodity groups to 45 specific commodities, 2) finding the equilibrium prices through simultaneous equilibrium of supply and demand instead of recursive method, 3) switching calendar year average to marketing year average of supply and demand for grain and fruits, 4) switching econometric equation to Cohort method to forecast agricultural population, and 5) each commodity group submodel can be operated independently, and aggregated and feedbacked easily in the Excel spreadsheet. Ex post simulation of KASMO was performed to test the fitness for the sample period of 2000∼06. The average root mean squre percentage errors (RMSPE) are as low as 2% for the equations in the submodel of macro-economic variables, agricultural input, and the agricultural macro variables, so that the model fitness seems quite good. The RMSPEs of the plant submodel may be considered as acceptable. The average RMSPEs are less than 5% for the grain acreage equations except for wheat, less than 10% for the grain demand equations except for corn, and less than 3% for the grain import demand equations except for wheat. The average RMSPEs are less than 10% for the demand and supply equations of vegetables except for the import demand equations of cabbage and chinese cabbage, acreage equation of fresh pepper, and demand equation of eggplant. The RMSPEs of the fruits and special crops are poor. The average RMSPEs are less than 10% for the demand and supply equations of fruits except for the import demand equations of grape and persimmon, acreage equations of grape and tangerine, and demand equations of pear, peach and persimmon. The average RMSPEs are 10% for the acreage demand and supply equations of special crops except for the import demand equations of ginseng, green tea and mushrooms, acreage equation of green tea, and demand equation of flowers. The RMSPEs of the livestock submodel are poor. While the average RMSPEs are less than 3% for the supply equation, those are more than 20% for the meat demand equations and 15% for the meat import equations. Baseline forecast to the year 2018 was performed under the following assumptions; - Tariff and tariff rate quota (TRQ) of 2005 stand still, except the TRQ of rice increased until 2014, and tarrified thereafter. - Average exchange rate of 2008 of 1,050 won per US dollar, and the year average international crude oil price 105 US dollar per gallon continues. - GDP growth rate is quoted from Global Insight forecast. - World grain price is quoted from the FAPRI forecast. The agricultural population is forecasted to decrease 3% per annum. While the nominal agricultural production will increase 0.8%, the value added of agriculture will decrease 0.1% per annum. Income per household from farming will increase 1.1% per annum, and non-farm income will increase 3.5%. Two sensitivity analyses are performed to test the model stability. One is exchange rate shock, and the other is FTA shock.If exchange rate will be 1,200 won per US dollar, which is 14.3% higher than the baseline, agricultural production in 2018 will inelastically increase 3.9% than the baseline as the agricultural import will decrease. And the value added will increase only 1.7% as the main intermediate goods like oil will be imported more. If we assume the Free Trade Agreement between Korea and US will start from 2010, agricultural production in 2018 will decrease 3.5% than the baseline as the agricultural import will increase by tariff cuts. And the value added will decrease 4.7% as there will be less demand for intermediate goods than baseline. The result is quite consistent with the recent analyses.Researchers: Myung-Hwan Kim (Ph.D.), Oh-Bok Kwon (Ph.D.), Young-Soo Cho (Ph.D.), Dae-Seob Lee (Ph.D.), Tae-Hun Kim (Ph.D.), and Sang-Mi ParkE-mail address: kimkim@krei.re.kr, obkwon@krei.re.kr, yscho@krei.re.kr, daeseob@krei.re.kr, taehun@krei.re.kr, psm801@krei.re.kr