Stochastic Modeling and Estimation of the Probability of Productivity Losses

Adoption of responsible managerial decisions in the conditions of potentially large or even “catastrophic” economic losses requires, apart from expert evaluation, the obligatory use of objective scientific approaches to the assessment of the possible risks. Such methods allow one to give an objective assessment of the risks in the current or forecasted critical situation. The article proposes a stochastic model for estimation of the productivity losses probability in the crop production in the conditions of various climatic scenarios, including the “bad” and “abnormally bad” seasons. The approach is based on the use of a modified algorithm for estimating the parameters of the finite mixture of distributions to the retrospective information about the productivity of the crop in a given geographic area within the previous years. The adequacy of the model is tested based on the real data. Extensive statistical material is accumulated at Russian agricultural research institutes for implementing the proposed approach in the various soil-climatic zones of the Russian Federation.