private quality control systems which can qualify for less intensive public control for Salmonella and
other food related hazards by inspectors of the government (de Bakker, 2007).
This paper presents results of an analysis of an incentive system for Salmonella control in a two
segment pork supply chain. The model described here is a direct extension of a dynamic principal-
agent model developed by King, Backus, and van der Gaag (2007) that allows for explicit
consideration of the producer’s performance history in controlling serological Salmonella herd
prevalence. In this study we add the possibility of plant control measures that reduce bacteriological
Salmonella carcass prevalence. The producer incentive system determines Salmonella testing
frequencies for pigs delivered to the slaughter plant, as well as charges to the producer for testing and
penalties for substandard Salmonella control. Plant level control measures are one-time investments
made at the time the plant chooses parameters for the producer incentive system. Using cost estimates
and technical parameters based on Dutch experts and Danish data, we use the model to evaluate the
cost effectiveness of plant and farm control measures and trade-offs between prevalence reduction and
related costs and gains. We also assess the impact of ownership structure onperformance for a wide
range of prevalence threshold levels.
2. Salmonella testing procedures and control measures
Pork can become contaminated with Salmonella in different ways, and at different points in the
pork supply chain. Control measures to decrease the risk from Salmonella contamination should be
considered on the farm and at the slaughterhouse (Beloeil et al., 2004).
Two types of testing, serological and bacteriological, are used to assess Salmonella prevalence.
Serological tests determine the level of Salmonella antibodies in blood samples typically taken at
slaughter. Serological sampling of finishing herds is possible at the farm or at the slaughterhouse,