Abstract Optimal allocation of resources to the management of biosecurity risk, threatened species conservation or natural hazards such as bushfires is imperative—because program budgets are usually finite and, therefore, constrained. However, effectively dividing resources among management activities to achieve the greatest benefit remains a fundamental challenge of public policy, especially when benefits are difficult to quantify in monetary terms. Choosing an appropriate analytical approach and computational method for solving the decision problem is important if intending to produce meaningful guidance for resource managers. We searched the resource allocation literature to identify and analyse optimisation approaches that matched our generic decision problem: to allocate resources to border biosecurity risk controls when benefits are dependent. We focused on the biosecurity literature but also considered different decision contexts, such as threatened species conservation and bushfire risk management, which also seek to protect assets from hazards. The biosecurity resource allocation literature lacked fitting replicates of our decision problem because it concentrates on post‐border controls that are often assumed to be independent. Import risk analysis proved to be more suitable in that biosecurity risk analysts have developed quantitative models to cater for the randomness and dependence of processes along import pathways. However, both research areas simplify models by considering only specific species‐commodity interactions. Resource allocation modelling in biosecurity and other sectors needs to embrace the complexity of decision contexts and realistically describe the processes involved, treating dependencies as such without making reductionist assumptions. We contend that import risk analysis (i.e. reducing a risk to below the Appropriate Level of Protection in the least trade restrictive manner) and biosecurity resource allocation (maximising the risk reduction of a fixed budget) are two sides of the same coin and advocate the use of Bayesian Networks for effectively conceptualising these complex relationships. For border biosecurity management, it means appreciating that risk controls have dependent benefits in real world scenarios, for example, the risk reduction achieved by one risk control will affect the risk reduction potential of subsequent risk controls, and that multiple hazards affect multiple assets. This shift in thinking will increase confidence in the results of decision support tools and lead to more effective resource allocation decisions in practice.
Dodd et al. (Thu,) studied this question.