Does animal welfare influence dairy farm efficiency? A two-stage approach

This article investigated how process-based animal welfare indicators (PAI) affected the technical efficiency of German dairy farms. A sample of 115 North-Rhine Westphalian dairy farms was used to estimate their technical efficiency with data envelopment analysis. A censored regression model was then applied to quantify the effects of PAI on technical efficiency. The results indicated that in particular a higher percentage of cow losses, a higher replacement rate, and a longer calving interval had, at their respective mean, a negative marginal effect on the technical efficiency of the sample farms. In contrast, a lower age of first calving, a higher in-milk performance, and a higher somatic cell count were positively correlated with technical efficiency. Some of the PAI followed a polynomial trend (i.e., their influence on technical efficiency did not have a constant sign, and levels for minimum/maximum technical efficiency were present). The minimum efficiency score at constant returns to scale was obtained when farmers had cow losses of 0.4%, a calving interval of 430d, and a cell count of 146,000 per milliliter. However, maximum technical efficiency was obtained at a milk yield of 9,796 kg per cow and year. The corresponding amounts in case of technical efficiency under variable returns to scale were at a similar level, except that milk yield showed a positive linear influence on technical efficiency. Moreover, technical efficiency under variable returns to scale was positively correlated with the fat content of milk. The lowest level of technical efficiency was reached at a fat content of 4.1%. Subsequently, we found that efficient dairy farms did not always correspond with recommended values concerning animal welfare criteria. Finally, the results showed that the assumption of a monotone effect direction of PAI on farm efficiency was inappropriate, and that this issue would need to be addressed in future research.