농작물재해보험의 최적 할인‧할증시스템 분석

다국어 초록 (Multilingual Abstract)

Crop insurance in Korea has been unprecedentedly expanded as a risk management program supported by the government since it was introduced with pilot programs of apple and pear in 2001. It covered 50 crops and marked insured rate of 27.5% as of the end of 2016. On the other hand, not many efforts were given to upgrade insurance products, where Bonus-Malus System (BMS) as a part of rate-making ranked one of the important issues to be resolved. In cop insurance farmers face production risks occurred by natural affairs such as storms, flood, drought, pest, plant disease and others. The natural affairs commonly occur at wide areas coincidently, thus affecting premium rate at county level. Sequentially, the rate updated at county level in addition to personal loss experience affects the BMS at individual level in farming sector while BMS is only affected by individual loss experience in general. This may be unfair and discourage for farmers to buy insurance products.
This study sets up a BMS in Korean crop insurance to resolve a double loaded penalty in current BMS. The BMS in this study adopts the model suggested by Frangos and Vrontos(2001) with Bayes’ theorem, where initial premium rate is determined by prior distribution but is updated with the loss experience by individual farmers with lapse of time. The concept of fairness suggested by Heras, Vilar and Gil(2002) is adopted to support the BMS developed. For the study, crop insurance data including premium, indemnity and frequency of occurrence provided by NH Property & Casualty Insurance Corporation (the insurer) are used. Among the crops insured only apple, pear and rice are chosen because the first two crops have the longest history of crop insurance and ranks the highest insured rate during the program period. Also the rice is the most important crop and has the widest cultivation area in Korean agriculture. In addition, we confine the region to main producing areas to see the regional difference of BMS.
The results are as follows: Most of all, the BMS of selected crops rapidly narrows the difference between expected risk of the insured and premium by lapse of time. This means that the premium rate reflected by the frequency and severity of occurrence based on the insured, not county level, is enough to evaluate the risk farmers face. Second, the BMS is robust even though the model is generalized by including the characteristics of the insured such as region, deductible, insured area, amount insured, etc. The premium rate determined by generalized BMS with the characteristics of the insured is rapidly getting close to the expected risk farmers have. Thus, the generalized BMS is expected to contribute to the predictability of farmers’ premium rate according to their characteristics.
Third, the individual fairness of BMS differs between current system and the system developed in this study. The BMS with current system makes the difference between premium and indemnity by the insured be bigger with lapse of time while the BMS with Bayes’ theorem in this study makes the difference be smaller. The difference is becoming very small and stabilized in 20 years after joining insurance program when the latter is applied to rate making.
The global sufficiency produced by individual fairness combined with each discrete probability also differs between the current BMS and the BMS with Bayes’ theorem in this study. According to the BMS with Bayes’ theorem, the difference between premium and indemnity is very small and becoming smaller with lapse of time while the BMS with the current system makes the difference be bigger. As a result, we know that the BMS with Bayes’ theorem appropriately evaluate the risk individual farmer has. It can also contribute to the stability of insurance programs operated by insurance companies.
The current insurance program is focused on the expansion of crops and insured areas. More sophisticated insurance programs including rate-making procedure to reduce asymmetric information are also required to increase farmers’ participation and stabilize business operation of insurance companies. The BMS with Bayes’ theorem suggested by this study is expected to contribute to the appropriate evaluation of farmers’ risk and the soundness of insurance programs. This also contribute to reduce adverse selection and moral hazard common in agricultural insurance, resulting in the soundness and sustainability of insurance programs to some degree. In addition, this study may contribute to further study of BMS with Bayes’ theorem and its application in Korean crop insurance.

목차 (Table of Contents)

• Ⅰ. 서 론 1
• 1.연구의 필요성 1
• 2.연구의 목적 4
• 3.선행연구 5
• Ⅱ. 농작물재해보험의 특수성과 보험료 산정방식 8
• 1.농업재해와 농업재해관리 수단 8
• 2.농작물재해보험의 특수성과 역할 11
• 3.농작물재해보험의 보험료 산정방식 15
• Ⅲ. 이론적 고찰 19
• 1.할인할증 시스템의 역할과 목적 19
• 2.정보의 비대칭성과 할인할증시스템 20
• 3.최적 할인할증 시스템 구축 원리 24
• Ⅳ. 자료 및 분석방법 25
• 1.자료 25
• 2.분석방법 37
• Ⅴ. 분석결과 49
• 1.사고발생빈도와 사고규모만을 고려한 최적 할인할증시스템 49
• 2.가입자의 특성을 고려한 최적 할인할증시스템 56
• 3.개별공정성과 전체적 충분성에 기초한 할인할증시스템의 공정성 평가 64
• Ⅵ. 결 론 69
• 참고문헌 72
• 부 록 76

참고문헌 (Reference)

1 최정호, "리스크와 보험", 청람, 2014

2 김석현, 서종석, "“농작물 보험료율 산정방식”", 한국농업경제학회, 농업경제연구, 42(3), 69-88, 2001

3 박대식, 최경환, "농업재해대책의 실태와 개선방향", 한국농촌경제연구원, 한국농촌경제연구원, 2001

4 김종숙, 최경환, 정명채, "농업재해보상과 작물재해보험대책", 한국농촌경제연구원, 한국농촌경제연구원, 1993

5 윤병석, 최경환, 채광석, "농작물재해보험의 성과와 정책과제", 한국농촌경제연구원, 한국농촌경제연구원, 2010

6 최경환, "미국의 작물수입보험 실태와 시사점", 한국농촌경제연구원, 한국농촌경제연구원, 2011

7 서상택, "보험과 재해지원의 역할 재정립 방안", 농업정책보험금융원, 2015

8 오만숙, "R 몬테칼로와 함께하는 베이지안 통계추론", 자유아카데미, 자유아카데미, 2012

9 정원호, "“농업수입보험제도 도입의 방향과 과제”", 농정연구, 51, 103-150, 2014

10 김대원, 최태길, 김태균, "“사과 재해보험에 대한 생산자 선호 분석”", 한국농업정책학회, 농업정책연구, 22(1), 165-180, 1995

11 백지영, "자동차보험 할인할증제도의 공정성 분석 연구", 이화여자대학교 대학원, 이화여자대학교 석사학위논문, 2005

12 정원호, "“농작물재해보험의 위험분산체계 개선방안”", 한국농식품정책학회, 농업경영정책연구, 40(2), 223-247, 2013

13 김태균, 조재환, 박준형, "“현행 농작물재해보험에서의 역선택 가능성”", 한국농식품정책학회, 농업경영정책연구, 30(4), 561-578, 2003

14 김석현, 백장선, 김태균, "“모수 비모수 분포를 이용한 사과 보험요율 추정”", 한국농업경제학회, 농업경제연구, 45(1), 147-167, 2004

15 지홍민, "“국내 자동차보험 할인할증제도의 점근 공정성 측정”", 한국리스크관리학회, 리스크관리연구, 21(2), 143-167, 2010

16 정민국, "“포아송확률분포를 이용한 가축 공제 요율 산정 연구”", 한국농업경제학회, 농업경제연구, 41(3), 79-96, 2000

17 김충실, 박재화, "“농작물재해보험 국가재보험제의 시뮬레이션 모형연구”", 한국농업정책학회, 농업경영정책연구, 33(2), 405-423, 2006

18 이주관, 정진화, "“한국 농작물재해보험 시장에서의 역선택과 도덕적해이”", 한국농업경제학회, 농업경제연구, 55(1), 29-47, 2014

19 김태균, "“재해보험과 수입보험에 대한 생산자선호 및 후생효과 비교”", 한국농업경제학회, 농업경제연구, 42(2), 33-49, 2001

20 조재린, 채원영, "농작물 재해보험으로서의 지수형 날씨보험 도입 사례와 시사점", 보험연구원, 2012

21 강수진, 정원호, "“농작물재해보험의 농가 수입안정 및 사회적 후생 효과 분석”", 한국유기농업학회, 한국유기농업학회지, 24(3), 369-383, 2016

22 김재한, 김태균, "“작물재해보험의 지역별 보험요율 결정: 지역보험의 도입 가능성 검토”", 한국농업경제학회, 농업경제연구, 43(2), 17-34, 2002

23 김진한, 민재형, "“마코프체인을 이용한 자동차보험료의 할인ㆍ할증시스템에 관한 탐색적 연구”", 한국경영과학회 학술대회논문집, 79-95, 1994

24 김성섭, 서상택, "“혼합전략하의 내쉬균형을 이용한 농작물재해보험의 최적 검증조사전략 연구”", 한국농식품정책학회, 농업경영정책연구, 41(1), 152-171, 2014