Perspectives on mutual dependence in plant health

Summary

Analyses of the economic impact of, and possible risk mitigation strategies against, pests often fail to account for spatial and economic dependencies among the evaluated decision-making units, and heterogeneity in the environment they operate in. Awareness of the mutual dependence of actors, regions and countries is critical for proper management of pests. In this thesis, I develop methodological approaches to account for the spatial nature of pest populations and the mutual dependence of farmers, countries, and markets to contribute to a more informed discussion on plant health policies in Europe.

Chapter 2 provides an empirical approach to measure spatial spillover effects of decision-making units’ characteristics with managerial performance of neighbors. This allows relaxing the assumption that decision-making units operate in isolation from their peers. The model is applied to data from the Farm Accountancy Data Network. The data comprises expenses, revenues, balance sheet positions, and farm characteristics for 75 Dutch arable crop farms which are observed over six years. The results show that managerial performance of decision-making units is related to neighbors’ characteristics such as the degree of farm specialization, received subsidies, insurance payments, and age. However, how they associate is found to depend on the definition of the neighborhood. The results imply that analyses of optimal pest control should be expanded beyond individual farm-level.

Chapter 3 provides an integrated framework that derives insights from climatic suitability, spread modelling, and economic modelling. A pest spread model is developed such that environmental heterogeneity, point of pest introduction, and host locations are fundamental components for the realized dispersal over time. The economic model captures heterogeneity of different cropping systems and countries, includes temporal effects through losses in investments, and highlights economic dependence among growers due to price responses following changes in aggregate supply. The model is applied to the invasive species Xylella fastidiosa subspecies pauca to compute impacts to olive growers in Europe, with a focus on Italy, Greece, and Spain. For Italy, across the considered spread rates the potential economic impact over 50 years ranges from 1.9 billion to 5.2 billion Euros for the economic worst-case scenario, in which production ceases after orchards die off. If replanting with resistant varieties is feasible, the impact ranges from 0.6 billion to 1.6 billion Euros. Even under slow spread rates and the ability to replant with resistant cultivars, economic impact to olive growers from further spread of Xylella fastidiosa subspecies pauca is expected to be sizable (0.6 billion Euro) and warrants strong regulatory response.

Chapter 4 translates results of the spatially explicit pest spread model to suit the needs of partial equilibrium models. The chapter shows how global sensitivity analyses can be informative in the context of partial equilibrium models. The model is applied to the invasive species Xylella fastidiosa subspecies pauca to compute impacts on producers and consumers of olive oil in Europe, with a particular focus on Italy, Greece, and Spain. I find that most of the potential future impact of Xylella fastidiosa subspecies pauca in the olive oil market would fall on consumers because of higher prices following reductions in supply. The analysis highlights that the problem of invasive pests should be contextualized as a societal challenge as opposed to one that affects only producers. The chapter stresses the fact that consumers are beneficiaries of pest control.

Chapter 5 shows that a joint analysis of several hundred pests can produce hotspot maps with a high accuracy. A machine learning model is trained on a dataset covering 248 invasive species to map risk of new pest introduction in Europe as a function of climate, soils, water, and anthropogenic factors. Due to the considerable time and labor requirements for species-specific analyses, there is no information on area-specific suitability for establishment or introduction for many hazardous species. The joint analysis of several species could be one approach for addressing this knowledge gap. Furthermore, joint analyses of various pests could help to identify weak-links and thereby inform collective control. Results show that the BeNeLux states, Northern Italy, the Northern Balkans, and the United Kingdom, and areas around container ports such as Antwerp, London, Rijeka, and Saint Petersburg are at higher risk for introductions. However, harmonized, systematic, species survey data comprising also true absences are required to further validate and improve these maps.

The thesis, as a collection of these articles, contributes to the literature by providing methodological approaches which (i) capture spatial dependencies, (ii) account for environmental and economic heterogeneity at the level of granularity feasible under the available data, (iii) acknowledge the temporal nature of pest spread and economic impact in perennial hosts, (iv) highlight the actionable insights sensitivity analyses can generate, and (v) propose cost-effective modelling strategies to address the absence of risk maps for many invasive species.