Steine G, Kristofersson D, Guttormsen A G
AKVAFORSK, Institute of Aquaculture Research AS, PO Box 5010, N-1432 As, Norway.
J Dairy Sci. 2008 Jan;91(1):418-26. doi: 10.3168/jds.2006-747.
In this paper, a translog profit function was applied to estimate the economic values of the traits included in the breeding goal for Norwegian Red dairy cattle. The following 10 traits are included in the breeding goal: milk, meat, mastitis resistance, fertility, calving difficulties, stillbirths, other diseases, udder, temperament, and legs. An empirical implementation that locally approximates the unknown true profit function was suggested and estimated, taking farm heterogeneity into account. The model was applied to a panel data set of 3,259 Norwegian dairy farms over the period 1999 to 2003. Panel data, also called longitudinal or cross-sectional time-series data, are multiple cases (cows, farms, countries, etc.) observed over 2 or more time periods. The data set consisted of farm-level data, including production and economic data from the farm and the estimated breeding values for each cow's sire. The estimated economic values make it possible to test whether genetic selection has been profitable for the farmer, and the extent to which the currently used economic values were optimal during the period 1999 to 2003. Although the translog profit function is quite flexible, it is rather complex, and a simplified version of the model, a Cobb-Douglas profit function, was also estimated. However, the hypothesis that this simpler function adequately describes the data compared with the full translog model was rejected. Further, the hypothesis that the estimated breeding values are profit neutral was rejected (i.e., the hypothesis that there are no interactions between input and output prices on one hand and estimated breeding values on the other). These results indicated that selection not only leads to a parallel shift in profits, but also to changes in input use. Seven of the 10 traits had a significant effect on the farmers' profit. The 3 traits that were not significant were calving difficulty, stillbirth, and other diseases. The results showed that the breeding program for Norwegian Red cattle has been fairly successful in improving farmers' profits. However, a slight modification of the breeding goal, such as a reduction in the weights for stillbirths and other diseases and an increase in the weights for meat and temperament, would increase farm profits.
在本文中，运用了超越对数利润函数来估计挪威红牛育种目标中所包含性状的经济价值。育种目标包含以下10个性状：产奶量、产肉量、抗乳腺炎能力、繁殖力、产犊困难、死产、其他疾病、乳房、性情和腿部。提出并估计了一个局部近似未知真实利润函数的实证方法，同时考虑了农场的异质性。该模型应用于1999年至2003年期间3259个挪威奶牛场的面板数据集。面板数据，也称为纵向或横截面时间序列数据，是在两个或更多时间段内观察到的多个案例（奶牛、农场、国家等）。数据集由农场层面的数据组成，包括农场的生产和经济数据以及每头奶牛父亲的估计育种值。估计的经济价值使得能够检验基因选择对农民是否有利可图，以及在1999年至2003年期间当前使用的经济价值在多大程度上是最优的。尽管超越对数利润函数相当灵活，但它相当复杂，因此还估计了该模型的简化版本——柯布 - 道格拉斯利润函数。然而，与完整的超越对数模型相比，这个更简单的函数能充分描述数据的假设被拒绝了。此外，估计育种值是利润中性的假设也被拒绝了（即一方面投入和产出价格与另一方面估计育种值之间不存在相互作用的假设）。这些结果表明，选择不仅导致利润的平行转移，还导致投入使用的变化。10个性状中有7个对农民利润有显著影响。3个不显著的性状是产犊困难、死产和其他疾病。结果表明，挪威红牛的育种计划在提高农民利润方面相当成功。然而，对育种目标进行轻微修改，例如降低死产和其他疾病的权重，增加产肉量和性情的权重，将提高农场利润。
