关键词: 牧草保险, 干旱天气指数, 产品设计, 费率厘定, 果洛州

Abstract:

By utilizing annual precipitation and forage yield per unit area data in Guoluo prefecture from 2003 to 2017, the drought weather index (DI_q) was constructed by selecting the percentage of precipitation anomaly (PA), the yield loss rate (YLR_q) was calculated by employing the Hodrick prescott filter (HP filter) method to estimate the trend yield of pastures, the correlation between the drought index and the pasture yield loss rate was fitted using a fixed-effects model, finally, to establish a loss rate expectation value model. Enabling the design of forage drought weather index insurance product and facilitating the determination of the pure insurance premium rate. The results clearly demonstrated that: (1) the risk of drought-related disasters affecting forage in Guoluo prefecture had a spatial distribution characteristic, increasing progressively from southeast to northwest. Specifically, with Maduo county being high-risk area that required focused attention on drought prevention. (2) From 2003 to 2017, the average forage yield per unit area in the six counties of Guoluo prefecture showed a fluctuating upward trend. Banma county had the highest average yield, averaging an impressive 11126.67kg·ha⁻¹ over the 15 years period. (3) There was a significant （P<0.050） linear relationship between the drought index (DI_q) and the forage yield loss rate (YLR_q), expressed by the equation YLR_q=0.311DI_q+0.035（P<0.050）. Consequently, each unit increase in the DI_q resulted in a 0.311 percentage points increase in the forage YLR_q. When the DI_q was 0, other factors lead to a 3.500% reduction in forage yield. (4) Using the EasyFit software to analyze the probability distribution models of the drought index, four distinct distribution functions were identified for the forage drought index across in the six counties of Guoluo prefecture: Generalized Pareto, Error, Johnson SB, and Generalized Extreme Value models. The Anderson−Darling (A−D) test statistics confirmed the goodness of fit for all models, the A−D test statistics were all less than 0.3. The probability of encountering mild, moderate, severe, and extreme drought conditions in each county of Guoluo prefecture were found to range between 8.750% and 30.350%, 13.000% and 34.530%, 0 and 13.790%, and 0 and 0.820%, respectively. (5) With 100% guarantee level provided by the forage drought weather index insurance product in Guoluo prefecture, the pure premium rates for the forage drought weather index insurance product in the six counties of Guoluo prefecture ranged from 2.259%−3.748%, with the highest pure rate of 3.748% in Maduo county in the northwest and the relatively lowest rate of 2.259% in Dari county. The spatial distribution pattern of pure premium rates for pasture drought weather index insurance gradually increased from southeast to northwest. Based on the spatial distribution of pastures and drought conditions in Guoluo prefecture, it was recommended that counties with higher pure premium rates, such as Maduo and Maqin, be considered as pilot counties to implement differentiated fiscal subsidy policies for insurance premiums.

Key words: Forage insurance, Drought weather index, Product design, Premium rate determination, Guoluo prefecture