Certification is an essential feature in organic farming, and it is based on inspections to verify compliance with respect to European Council Regulation-EC Reg. No 834/2007. A risk-based approach to noncompliance that alerts the control bodies to activate planning inspections would contribute to a more efficient and cost-effective certification system. An analysis of factors that can affect the probability of noncompliance in organic farming has thus been developed. This article examines the application of zero-inflated count data models to farm-level panel data from inspection results and sanctions obtained from the Ethical and Environmental Certification Institute, one of the main control bodies in Italy. We tested many a priori hypotheses related to the risk of noncompliance. We find evidence of an important role for past noncompliant behavior in predicting future noncompliance, while farm size and the occurrence of livestock also have roles in an increased probability of noncompliance. We conclude the article proposing that an efficient risk-based inspection system should be designed, weighting up the known probability of occurrence of a given noncompliance according to the severity of its impact.
Keywords: Count data models; organic farming; risk-based inspection.
© 2014 Society for Risk Analysis.