Measuring the crisis protection gap

Author: Conor Meenan, Lydia Poole and John Ward

Photo credit: Viacheslav Lopatin / Shutterstock

The Centre for Disaster Protection recently published findings from a multi-year research exercise to understand whether it is possible to quantify how much money should be pre-arranged to protect low-income and crisis-vulnerable countries and populations in response to future crises. 

This research explored data and modelling methods which could be applied to estimate the future crisis financing needs of national or international responders, initially focussing on immediate response costs. We looked at crises which are most impactful in lower-income and fragile settings, including drought and displacement events. 

We tested technical approaches in the context of various potential use cases, ranging from global-level monitoring to more detailed analysis that can inform the design of specific risk financing instruments. 

This blog outlines our rationale for this research and highlights five key findings. 

What is the crisis protection gap, and why does it need to be measured?

We define the ‘crisis (financing) protection gap’ as the difference between the total expected contingent liabilities of national or international responders (i.e. the costs they incur in responding to crises) and the amount of funding available to meet these costs through pre-arranged financing. 

In other words, the crisis protection gap serves as a measure of how much money could be arranged in advance by a government or humanitarian organisation so that future crisis responses are fully covered using pre-arranged financing instruments (PAF). 

At a national level, this risk information could support a range of financial planning use cases, including response planning, development of risk financing strategies, and instrument selection and design. At a global level, comparative estimates of crisis protection gaps could help inform the allocation of international funding towards PAF, set targets, trace progress towards reducing protection gaps, and highlight geographies or risks that are systematically under-protected. 

Crisis Protection Gap Research Guiding Star

There are many types of crisis-related costs and various ways to approach the modelling of crisis protection gaps depending on the intended use of the risk information. To focus the research, we used a set of questions as a ‘guiding star’ to investigate the characteristics of the crisis protection gap: what does the risk information describe? What are the horizons and scope of the forecast? And, what does the risk information enable?

Our review of existing tools found that there are risk information platforms that provide risk monitoring at the global or national level, as well as tools that offer risk information for specific risk types or specific costs. However, when we began the research, no publicly available datasets provided globally comparative estimates of pre-arranged financing gaps as defined by the guiding star. 

Five Research Insights

We explored a range of potential use cases and developed demonstration analyses to test possible approaches and limitations. The research produced a range of findings and technical insights – five key highlights are presented below: 

1. Data and methods are available: Crisis information is already good enough to start working with. It is increasingly possible to generate forward-looking estimates of crisis protection costs.

The research demonstrates how crisis protection needs and costs can be estimated by combining information on exposure, crisis events and response costs. Detailed spatial datasets relating to exposed populations, crisis hazards, and response cost information are more comprehensive and available than ever. This data can already be used to form initial estimates of the size, likelihood, and distributions of crisis protection needs for some types of crisis and response costs at a global level. 

2. Model choices matter: Measuring crisis protection costs and gaps requires design choices that should be made explicit.

The use case for the crisis protection gap information needs and costs to be clearly articulated before modelling choices are made. The preferences of the end-user should be transparent and explicitly reflected in the modelling. Otherwise, there is a risk that the modelling design choices will create a view of crisis protection needs and costs that is either inappropriate to the use case or present a view of risk that does not align with the value judgements of the model users. 

The effects of using different value assumptions are explored in the paper using a global analysis of tropical cyclone risk. The analysis highlights how the view of risk changes substantially depending on whether value is assigned based on economic productivity or is linked explicitly to crisis protection needs and costs relating to low-income populations.  

3. Exposure data as a glue: a common view of exposure can promote consistency and comparability across crisis protection gap analyses.

Many use cases for crisis protection costs and gaps will require information that allows comparisons between crisis types in different geographic.  

If comparability is important for the use case, but this is not prioritised in the modelling approach, there is a risk that protection gap estimates will reflect differences in technical approaches and model assumptions much more than they reflect the distribution of crisis protection needs. 

In these use cases, a common population-based exposure data layer can promote comparability across crisis protection gap analyses for different crisis types in different locations. 

4.   The technical challenge is difficult but solvable – however, important conceptual and modelling challenges remain.

Although the work demonstrates that estimating crisis protection costs and gaps is increasingly feasible, several challenges remain where continued work will be required. 

First, for each specific crisis type, there can be challenges in calibrating what it means for people to be ‘affected’ by a crisis. This calibration challenge is difficult for complex risks in data-scarce environments, but this technical challenge can be addressed with assumptions that make sense for a given use case. 

A second area relates to difficulties in capturing (time-varying) drivers of vulnerability and response costs. For example, food price inflation, emergence of conflict, or the occurrence of prior crises can affect the thresholds at which a person is ‘affected’ to the point where a funding response is warranted. 

A third key area concerns the need to improve understanding of the numbers and locations of crisis-vulnerable people. This includes ensuring that data sets fully account for displaced people and refugees if they are not reflected in underlying census or survey data. 

The main challenge in developing the demonstration analysis relates to the availability of crisis response cost information… 

5.   In search of costs – better information on crisis-related response costs would be of considerable value.

Ensuring access to cost data then using it wisely to undertake robust and credible costing analysis appears to be the biggest challenge the international community will need to overcome if it wants to develop its understanding of crisis protection costs. 

The research explored the feasibility of top-down and bottom-up costing approaches, using historical humanitarian appeal and response cost data as a proxy for methods that could be applied more broadly to extend also to the contingent liabilities of national responders. 

The costing analysis in the research highlights just how challenging it is to access reliable cost information and, therefore, how challenging it is to estimate crisis protection needs. Higher quality information on the costs of humanitarian or government response activities would be of considerable value to exercises that seek to estimate the crisis protection gap and support learning on best practices relating to costs and financial planning in the sector. 

Building towards a view of global crisis protection gaps

Technical and practical challenges make modelling protection gaps difficult for all crisis types at a global level.  

However, the research identified that recent improvements in the availability of high-resolution demographic data, combined with physical modelling techniques from engineering and reinsurance sectors and cost functions derived from humanitarian response data, can start to form a clearer picture of the relative size and likelihood of future crisis-related costs. 

In practice, this task is substantially more feasible now than it would have been a decade ago. There are initiatives such as the Global Risk Modelling Alliance (GRMA) and Global Resilience Index Initiative (GRII). These initiatives are working to address similar data gaps identified in this research. The newly established Global Shield against Climate Risks aims to 'close urgent protection gaps’ by financing activities and instruments that increase the use of PAF. A recent paper developed by the World Bank sets out a methodology for ‘Counting People Exposed to, Vulnerable to, or at High Risk From Climate Shocks’. These initiatives highlight the growing feasibility and demand for risk information that can support policy and financial planning decisions that serve the most vulnerable globally.  

Dive into the report for more detailed analysis, and get in touch with any thoughts or comments!