| Abstract [eng] |
In the first part of this work, I tested if there is a relationship between weather conditions and crop insurance loss ratios. I tested two models. The first model was used to test the linear relationship between crop insurance loss ratios and weather conditions. Results show that this model does not satisfy the multiple regression model assumption about constant variance of residuals. In the second model the dependent variable – loss ratio – was transformed using power transformation. Then the linear relationship between transformed crop insurance loss ratios and weather conditions was tested. Results suggest that this dependency is weak. It shows that crop insurance loss ratios are affected not only by weather conditions but also by other factors. In the second part, I analyzed the distribution of loss ratios for five crops. I tested three probability distributions: gamma, log-normal and shifted Pareto. Shifted and truncated gamma, truncated gamma and truncated log-normal distribution did not fit according to the distribution of data. The results show that the truncated shifted Pareto distribution can be used for insurance loss ratio approximation. In the third part, I analyzed the intensity of indemnities and indemnity amount for one hectare. Analysis shows that there is no actual trend in average indemnity amount and average intensity for 2000 – 2005. There is no trend in standard deviation of intensity for 1995 – 2005. The results also show that there is no linear relationship between indemnity amount for one hectare and intensity of indemnities. |
