Economic evaluation of Holstein sire selection strategies for herds in Italy, Mexico, The Netherlands and the United States using stochastic simulation

Abstract

STOCHASTICALLY SIMULATED DAIRY HERDS WITH GENETIC, ECONOMIC AND MANAGERIAL PARAMETERS FOR MILK, FAT AND PROTEIN PRODUCTION IN ITALY, THE NETHERLANDS AND THE UNITED STATES, AND FOR MILK YIELD IN MEXICO FOR INVESTMENT HORIZONS OF 10 AND 20 YEARS WERE USED TO EVALUATE SIRE SELECTION STRATEGIES. ONE TO TWENTY PROGENY AND PEDIGREE-EVALUATED SIRES THAT WERE COMMERCIALLY AVAILABLE FROM US AI UNITS IN JANUARY OF 1996, AND GENETIC TRENDS, WERE USED AS A BASIS FOR SELECTION ON EXPECTED PROFIT EACH YEAR. THE USE OF 20 RANDOMLY CHOSEN YOUNG TESTING SIRES WITH LOW SEMEN COST WAS ALSO EVALUATED. AVERAGE PROFIT, LOWER 95% CONFIDENCE LIMIT OF PROFIT (LCL95), AND UTILITY (PROFIT ""0.06 €” VARIANCE OF PROFIT) WERE OBTAINED ON THE BASIS OF 1 000 REPLICATES. SIMULATIONS USING ONE SIRE PER YEAR ALWAYS GAVE THE MAXIMUM AVERAGE PROFIT. THE NUMBER OF SIRES FOR MAXIMUM RESPONSE FOR LCL95 WERE SMALLER IN COUNTRIES WITH GREATER PROFITS AND AN INVESTMENT HORIZON OF 20 YEARS. FOR UTILITY, THE NUMBER OF SIRES FOR MAXIMUM RESPONSE TENDED TO BE 10 TO 20 IN MOST SITUATIONS. USE OF EITHER SELECTED PROGENY-EVALUATED, OR YOUNG SELECTED SIRES, WAS SUPERIOR TO USE OF RANDOMLY CHOSEN YOUNG SIRES FOR PROFIT, AND LCL95 FOR PROFIT, AT YEAR 20 IN ALL COUNTRIES AND HERD SIZES STUDIED, BUT WAS GENERALLY INFERIOR FOR UTILITY AT YEAR 10. THE EFFECT OF HERD SIZE ON OPTIMUM DECISIONS WAS SMALL, ALTHOUGH LCL95, AND, ESPECIALLY, UTILITY WERE SUBSTANTIALLY LOWER FOR HERDS OF LESS THAN 100 COWS. A VALUE OF ""0.06 TIMES THE VARIANCE OF PROFIT IN THE CALCULATION OF UTILITY SEEMS UNNECESSARILY LOW FOR RESTRAINING RISK. OPTIMUM SELECTION OF DAIRY SIRES FOR AI DEPEND ON ECONOMIC AND MANAGERIAL CONDITIONS AND ON THE DEGREE OF RISK AVERSION, AS WELL AS THE INVESTMENT HORIZON. HENCE, OPTIMUM STRATEGIES FOR SIRE SELECTION WILL DIFFER AMONG COUNTRIES AND EVEN AMONG PRODUCERS WITHIN A COUNTRY.