ECONOMIC WEIGHTS FROM PROFIT EQUATIONS - APPRAISING THEIR ACCURACY IN THE LONG-RUN

Methods which estimate the effects on farm profit of a change in a genetic trait (economic weight) are compared. A neoclassical economic model based on a Cobb Douglas type production function, is extended to the long run and compared with a conventional profit equation for two types of trait changes. With this model, an optimal production plan can be derived for the farm which depends on input prices and the production function parameters. In the calculation of economic weights by animal breeders, costs are usually linear with respect to farm output, and so input levels and output are determined arbitrarily. Differences between methods for the estimation of both absolute and relative economic weights are shown to be small for proportional trait changes of 0.01 to 0.05. However, for proportional trait changes of 0.3, proportional differences were from 0.01 to 0.75 and from 0 to 0.6 for absolute and relative economic weights respectively, depending on the production function parameters. This dependence on the size of the trait change is attributed to the non-linearity of the production function. The effects of 'rescaling', which implicitly discounts economic weights for output increasing traits, are also evaluated. Absolute and relative 'rescaled' economic weights for 0.01 proportional trait changes are shown to differ proportionally from those estimated by the neoclassical economic model by from 0.14 and 0.12 to 0.50 respectively. For 0.3 trait changes, proportional differences were from 0.16 to 0.79 and from 0.8 to 0.43 for absolute and relative trait changes respectively. The results have important implications for conventional breeding programmes, and the economic evaluation of breeds and major genes.