The increase in fraudulent incidents has resulted in a change of focus in theory related to investigating food fraud from risk mitigation to vulnerability reduction. While the literature provided a framework from which to identify possible areas of vulnerability to food fraud, it does not provide a methodology that would allow food producers and processors to quantify and assess their vulnerability to food fraud more accurately. This paper developed a holistic approach to analyse food fraud vulnerability factors using a Bayesian Network (BN) approach based on the Decernis food fraud database incidents. Food fraud incidents related to seafood, dairy, alcoholic beverages and meat products were reviewed as they cover more than 50 % of the Decernis incident records (as in June 2018). 580 cases of food fraud were included in the development of the BN model. SPSS Modeler 18.2 was used to construct two BN model, and fraud vulnerability factors were directly linked to seven criteria of the country of origin, country of detection, year, food fraud types, product types, the weight of evidence, and types of adulterants. Possible food fraud vulnerability factors related to each case retrieved from Decernis database, combined Barrier Analysis technique and Routine Activity Theory, and additional sources from the literature. Two Bayesian Networks of Tree Augmented Naïve (TAN) and Markov were selected to determine the most reliable holistic model. Based on the analysis result, the TAN model assessed the vulnerability to food fraud with a higher accuracy rate of 86 %. The country of origin (76 %), food product types (9 %), types of adulterants (counterfeiting) (9 %), and country of detection (China) (6 %) were the main predictors of food fraud vulnerability factors. This model helps authorities in border protection, policymakers, and quality assurance agencies assess fraud vulnerability for a range of food products for incoming (imported) food products, mainly if they know the country of origin and type of food products.
|Published - 19 Jul 2021