Browsing by Subject "Risk based control"

Now showing 1 - 1 of 1
  • Thumbnail Image
    Publication
    Economic analysis of organic certification systems : determinants of non-compliance and optimum control strategies
    (2012) Zorn, Alexander; Dabbert, Stephan
    Organic certification systems are prerequisite for the existence of a large-scale organic food market. Despite a well-established and generally effective control system, fraud regarding organic food that passed organic controls is detected regularly. This cumulative thesis consisting of four articles addresses current questions regarding the improvement of organic certification systems. The need for governmental supervision of an organic certification system run by private control bodies is demonstrated by a game theoretic model. A framework prepares the statistical analysis by conceptually linking factors that can influence organic control results. The case study on German supervision data from the years 2006 to 2008 reveals significant differences between private control bodies regarding the number of severe sanctions imposed, i.e. fundamental control results. These data that were collected for supervision of the control system, however, are not sufficient to explain these differences. This is due to shortcomings in the data collected. Key terms of the data are not defined and the variable definitions seem to change over time. This study concludes that there is more detailed and reliable data from organic control bodies needed to understand the determinants of non-compliance with an organic standard. Detailed data on organic farm controls from the years 2007 to 2009 were supplied by two control bodies. Theoretical considerations founded on the ?Economics of Crime? approach yield hypotheses on factors affecting non-compliance with an organic standard. The data provided by a German and a Swiss control body are analysed by two different logistic regression models. The probability of receiving a sanction (which is used as proxy for non-compliance) is estimated on farm level by using data on farm and farm production. Such an approach to assess the determinants of non-compliance has not been used previously in the literature. Given the gradual sanction system, an ordinal logistic regression model is appropriate for the analysis of the German data. Swiss data are analysed by a random effects logistic regression model. Both models confirm some of the factors contributing to the risk of non-compliance that are applied in qualitative risk assessment so far. Control results from previous years, the overall farm complexity and the farm livestock production complexity, as well as farm size are factors that increase the probability of receiving a sanction. Risks connected to specific crops or livestock types that could come along, e.g., with particular requirements of the production method cannot be confirmed across the models. The explanatory value of both models is likely to be improved by the integration of further variables, such as data on farmers? personal and financial characteristics. The heuristic model builds on the results of the econometric models. This model adopts a societal view on the control system by considering the costs of controls and the damages resulting from non-compliance with an organic standard. Monte-Carlo simulations illustrate the relationship between important parameters for optimising control strategies. These simulations show that even without fines a situation can occur where most operators comply. The different approaches to analyse control data encounter difficulties inherent to the control data. In this context, the dark figure consisting of undetected non-compliances, inhomogeneous detection probabilities linked to particular production methods, and a potential positive confirmation bias connected to the risk based control approach are especially relevant. The working hypothesis that these potential biases are distributed randomly deserves closer attention in subsequent studies. Such future analysis should be based on even more detailed data, e.g., pooling original data from different control bodies in a control system. Such a data base would allow focusing on severe non-compliances which occur only rarely. Furthermore, pooled data could be used to investigate issues that are fundamental for the supervision of a control system such as a control body effect on the detection of non-compliance. This thesis presents important results that can be consulted for further analysis of organic control systems. Beyond, the approach, the methods used, and the results obtained are of general relevance for food certification systems beyond the organic sector.