Estimating economic loss from cascading infrastructure failure: a perspective on modelling interdependency

摘要

Abstract Infrastructure failure can cause significant disruption of economic activity. The size of economic loss is a direct function of the interdependencies between infrastructure and economic systems raising important questions about infrastructure vulnerability and resilience. Economic theory is important in this regard as it makes a distinction between damage to infrastructure (stock) and how this may transfer to losses in economic productivity (flow). In order to capture the economic consequences of infrastructure failure, various economic models have been proposed to represent the multimodal complex networks and capture the effects of cascading infrastructure failure. There is still no consensus on the correct approach for estimating economic loss. The method commonly known as input-output analysis has gained the most attention in recent years for its ability to model indirect or higher-order economic losses. The typical input-output approach has spawned an entire field of related models which include: the inoperability input-output model (IIM); Ghosh supply-side model; dynamic input-output models; key-linkages analysis; as well as inventory based models amongst others. Amongst the various methods used to model infrastructure failure this paper identifies the assumptions and shortcomings that must be overcome to produce better estimates of economic loss. Firstly, critical infrastructure systems are connected to the economy through both physical and economic linkages. Models need to capture both types of linkage to accurately represent how cascading infrastructure failure will lead to economic loss and then how sectoral losses may have an indirect impact on infrastructure systems. Secondly, input-output based approaches assume that the economic structure within an economy remains stable during a disaster and throughout the recovery process. New models are required that are able to capture substitution of goods and structural change within an economy. Thirdly, models of economic loss are generally deterministic in nature and thus give no indication about the uncertainty behind model-based estimates. Economic loss estimates using probability theory and methods such as Monte-Carlo simulations or fuzzy logic may prove to be important avenues for quantifying uncertainty in economic loss estimates resulting from infrastructure failure.