【摘要】 车险作为我国第一财险险种，费率厘定不够精准一直是行业发展的痛点。在国外发达保险市场上，汽车碰撞安全等级及其修复性一直是重要费率厘定因子，并设有相关机构进行汽车碰撞测试。为了弥补我国在这方面的空白，中保研汽车技术研究院和中国汽车工程研究院在中国保险协会的指导下跨界合作开展低速碰撞测试，模拟事故常见的正碰和追尾，从耐撞性、维修经济性、维修比、维修费用、气囊是否直接起爆等5个维度进行汽车耐撞性及经济维修性测评，构建中国保险汽车安全指数C-IASI。通过引入C-IASI的9项指标，并建立泊松-伽马模型和Tweedie模型，探究汽车自身耐撞性和维修经济性导致的不同赔付差异与机动车辆保费的适配性。实证结果表明，加入中国保险汽车安全指数C-IASI后，商业车险纯保费预测效果有明显提升，泊松-伽马模型和Tweedie模型预测效果接近。更多还原

【Abstract】 Auto insurance is China’s largest property insurance type, and the lack of accuracy in rate determination has always been a hindrance to the insurance industry development. In a developed insurance market outside of China, the level of safety and the reparability of vehicles in the event of a car crash have always been important factors in rate determination. There are relevant institutions to conduct car collision tests as well. To fill the gap, under the guidance of the China Insurance Association, China Insurance Research Institute collaborate with China Automotive Research Institute to carry out low-speed crash tests.The typical frontal collisions and rear-end collisions are simulated, the vehicles are evaluated in terms of five aspects, i. e. crashworthiness, maintenance economy, maintenance ratio, maintenance cost and whether the airbag is directly detonated, and the China Insurance Automotive Safety Index(C-IASI) is constructed. The paper introduces the nine indicators of C-IASI and establishes the Poisson-Gamma model and the Tweedie model to study matching between the motor vehicle insurance premiums and different amounts of coverage caused by the vehicle crashworthiness and maintenance economy. The results show that by adding C-IASI,the prediction accuracy of commercial auto insurance pure premiums has been significantly improved. The prediction results of the Poisson-gamma model are close to those of the Tweedie model statistically.更多还原