An Optimum Regression Model to Estimate Economic Values for Milk Yield, Milk Yield Persistency and Calving Interval in Dairy Cattle

Emphasis on milk yield (MY) as well as milk yield persistency (MYP) and calving interval (CI) is neces-sary to achieve more sustainable production in dairy cattle. Therefore the main objective of this study was to find an optimum regression model to estimate economic values for MY, MYP and CI. Using a determi-nistic bio-economic model, seventy five production states differing mainly in MY, MYP and CI were stud-ied. For each production state, the total revenue comprised income from sold milk, calves of one week of age and manure. Feed costs were obtained from energy requirements for maintenance, growth, lactation and pregnancy. Non feed costs included costs of net replacement, health, artificial insemination and some others which were modeled as a function of CI. Multiple regression analyses of annual profits for production state on the means of MY, MYP and CI were used to estimate the economic values. Two different regression models were used. Both models included the linear effect of MY and the quadratic effect of MYP. How-ever, in one model the effect of CI was linear (Model CIL) whilst it wasquadratic in the other (Model CIQ). Under both models, economic value for MY was positive (0.10 $ for model CIL and 0.32 $ for model CIQ) as was expected for the assumed milk pricing system. Economic values for MYP in the models had differ-ent signs (-118.2 $ for model CIL and 715.55 $ for model CIQ). Under model CIQ maximum profit was associated with a value of MYP greater than unity and was not consistent with the definition of persistency. Economic value of CI was negative under both models (-2.68 $ for model CIL and -6.36 for model CIQ). In the model CIQ, the profit function had a minimum value for CI (at 803 days) which was not consistent with the previously reported relationship between profit and CI. Estimates of economic values for MY, MYP and CI showed that the model CIL was superior to the model CIQ due to a lower number of fitted effects and increased consistency with the real situation of dairy systems.