A forest management risk insurance model and its application to coniferous stands in southwest Germany

The paper presents a general forest insurance model that can serve as a basis to calculate risk premiums to insure the risk of forest destruction due to single or cumulative damaging agents. The methodology and the theoretical approach include the derivation of empirical probability functions for the occurrence of forest destruction for age classes and the modelling of these probabilities with a Weibull function to obtain point estimates of probabilities. Based on that, the interval estimates of destruction probabilities were calculated using a nonparametric Kolmogorov-Smirnov statistic. The risk insurance model consists of two essential parts: the risk of the forest owner informing about the expected loss induced by a forest destruction occurrence, expressed as the net insurance premium, and the risk of an insurer, defined as risk premium, that is related to the total insured area of forest. The sum of both is the gross insurance premium. The destruction probabilities were derived from a case study of a forest enterprise in southwest Germany dominated by coniferous stands using forest maps of three consecutive decades that were digitised and intersected. The insurance values as the financial framework to demonstrate the calculation of the insurance premiums for the case study were calculated based on a stand management project that should depict the financial output of average forest stand in the survey area. The stand management project was generated with the help of a simulation run with a distance dependent growth model. The net insurance premiums that were calculated based on the insurance values and the point estimates for destruction probabilities for age classes range from (sic)0 to (sic)160. The risk premiums that were calculated based on the interval estimates decreased by more than 90% if the insured area increased from 1400 to 140,000 ha. The resulting gross insurance premiums ranged from (sic)0.77/ha (age 0, insured area 140,000 ha) to (sic)4429 (age 70, insured area 14 ha). Although the insurance model produced rather high premiums, especially for higher age classes, it was possible to detect and evaluate small differences for expected destruction rates in particular age classes with the help of the proposed procedure. Normative problems of forest valuation that may occur when calculating the insurance values are discussed. (C) 2004 Elsevier B.V. All rights reserved.