Nonparametric estimation of economic losses due to adverse natural phenomena : the case of the Mexican corn production portfolio

Files
Citation
Share
Description
To measure the risk of an assets portfolio, exposed to catastrophic losses product of adverse natural phenomena, and estimate its maximum expected economic loss, it is necessary to have a wide range of scenarios, so the strong historical deviations in natural variables as extreme precipitation or the lack of it, very high or low temperatures are captured. These infrequent scenarios can be obtained through simulations based on historical data, but, in contrast with simulations made for portfolios with only financial instruments, it is not possible to assume neither statistical independence nor parametric distributions for the data. To assume no correlation and parametric distributions for the historical observations used to obtain the simulations, will result in an underestimation of the possible losses of the portfolio. --Abstract .