Identification of external shock impact on agri-food sector by applying outlier detection tests: case of Latvia

Outlier detection tests (e.g. Dixon’s Q-test, Grubbs’s test) are typically used to check the reliability of data measurement and the reliability of data sets. However, outliers can be caused by other factors, notably external shocks. Recent Russian invasion of Ukraine has caused an enormous geopolitical shock that has affected many countries and sectors including the agri-food sector. The intrinsic feature of the agri-food sector is frequent fluctuations in revenue and costs due to different factors. Thus, apparently extreme variations do not necessarily indicate the impact of an external shock. Moreover, the agri-food sector in Latvia has experienced several shocks previously. Therefore, the long-term historical data on the agri-food sector are highly likely affected by previous shocks, which limits the use of many methods of statistical analysis. The aim of this study is to investigate whether the variations in revenue, cost and income in the agri-food sector were probably caused by this geopolitical shock. The study demonstrates the application of outlier detection tests in order to identify whether the indicators of this sector in 2022 were affected by this geopolitical shock. Although many subsectors exhibited considerable variations in revenue, cost and income, only part of these variations were identified as outliers and thus affected by this geopolitical shock. The rest of them was rather the inherent variations of subsectors. This application of outlier detection tests proposed by the authors is not limited to the agri-food sector or the geopolitical shocks related to Russian invasion of Ukraine. They have wider scope of application. However, further research is required to develop their wider use, especially incorporation into government aid policies.