Between 2005 and 2015, over 1.5 billion people were affected by disasters globally, resulting in approximately US$1.3 trillion in economic losses. Disaster risk models (DRMs) are crucial for assessing potential future impacts and designing mitigation interventions. However, the common reliance on economic loss quantification in DRMs often overlooks the disproportionate impact on vulnerable populations. Decades of research consistently demonstrate that disasters affect different groups unequally based on factors like income, social status, age, race, gender, and disability. This inequitable impact necessitates incorporating equity concerns into DRMs to ensure effective and just disaster risk management. This research addresses two central questions: (i) How extensively and in what ways do major international development organizations account for equity in their disaster risk assessments? and (ii) What are the implications of neglecting equity considerations in risk assessments? The study aims to improve the design of disaster mitigation efforts to prevent unintentional exacerbation of existing inequalities.

Disaster risk management (DRM) globally depends heavily on risk modeling. Disaster risk is understood probabilistically as a function of hazard, exposure, and vulnerability. Risk assessments aim to provide comprehensive estimates of potential disaster costs to guide risk reduction measures such as building retrofits, land-use planning, and disaster insurance programs. A critical issue is that traditional risk assessments, particularly those relying on asset losses, fail to capture the disparate impacts of disasters on various populations. Focusing solely on asset losses may lead policymakers to prioritize risk reduction in wealthier communities, overlooking the vulnerability of those with fewer assets and greater difficulty recovering. Furthermore, asset loss measurements are incomplete representations of the total impact, as vulnerable groups often experience disproportionate consumption losses, slower recovery times, and forego crucial resources like food, healthcare, or education. While ex-post disaster assessments highlight the unequal distribution of impacts across various demographic categories, ex-ante risk analyses used to guide policy often neglect these disparities. This research addresses this gap by developing a typology to categorize different approaches to integrating equity into risk modeling.

The study developed a typology of approaches for incorporating equity into risk modeling, ranging from Type 0 (no equity considerations) to Type 5 (welfare loss models). This typology was used to analyze 69 risk assessments conducted between 2010 and 2021 by major international development organizations: the World Bank, Asian Development Bank, PreventionWeb, and Inter-American Development Bank. Data was obtained from publicly available repositories. The assessments were categorized based on their approach to equity inclusion. Additionally, a case study was conducted. A pre-existing earthquake risk assessment for Nepal (Type 0 approach) was re-analyzed using equity-sensitive methods (Types 3-5) to demonstrate the disparities revealed by these approaches. For the re-analysis, fatality rates were disaggregated by gender, age, disability, caste, and income. A welfare loss approach using equity weights was applied to redistribute earthquake losses based on income levels to reflect the varying welfare impacts of the same loss across different income groups. Data sources for the Nepal case study included census data and previously published research on earthquake fatality rates and building vulnerability.

The analysis of 69 risk assessments revealed that only a minority (approximately 30) addressed equity issues. Of these, only a small fraction employed quantitative methods for evaluating differential impacts. The distribution of assessment types was as follows: Type 0 (no equity consideration): 39; Type 1 (qualitative description): 11; Type 2 (index-based models): 8; Type 3 (disaggregated exposure): 8; Type 4 (disaggregated vulnerability): 0; and Type 5 (welfare loss): 3. Key parameters considered were economic status (29 assessments), gender (16), age (15), while race (7) and disability (6) received less attention. The Nepal case study demonstrated significant disparities in fatality rates when disaggregating by gender, age, disability, caste, and income. The Type 4 approach (disaggregating by both exposure and vulnerability) revealed substantial differences in fatality rates across economic quantiles (twice as high for the bottom quantile compared to the top) and age groups. The Type 3 approach (disaggregated exposure) showed stark differences in fatality rates across caste types and income groups, primarily due to geographical clustering of certain castes. Applying the Type 5 approach (welfare loss) reversed the trend observed in the original assessment. The original analysis showed that losses increased with average income, concentrating in urban areas. The welfare-loss analysis showed that the poorest districts faced greater welfare losses, which decreased with increasing average district income, highlighting the limitations of solely relying on asset loss for cost-benefit analysis.

The findings highlight the widespread inadequacy of current disaster risk modeling practices in addressing equity issues. Most studies fail to account for differential vulnerability, and very few incorporate welfare loss models. The study demonstrates that straightforward methods and existing data can be used to uncover hidden disparities. While more sophisticated equity-aware models are needed, even simple techniques significantly improve the accuracy and fairness of risk assessments. The current methods often prioritize risk reduction in wealthier communities because of higher asset values, neglecting areas with greater welfare losses. Disaggregating risk by exposure or exposure and vulnerability improves risk assessment results and improves the identification of at-risk groups, allowing for more effective allocation of emergency aid and more tailored risk management interventions. Improving models also supports other stages of the disaster cycle beyond risk mitigation. Data limitations are often cited as a barrier, but the study shows that even simplified analyses with limited data can be informative. The study emphasizes the significance of accounting for not only distributional equity (as demonstrated) but also procedural and recognition equity to enhance the inclusiveness and effectiveness of DRMs.

The study emphasizes the critical need for improved methods to account for equity in disaster risk modeling. The analysis shows that current practices largely neglect equity considerations, leading to incomplete and potentially misleading results. The case study illustrates how equity-sensitive approaches can reveal significant disparities, leading to more effective disaster risk management interventions. International development organizations are urged to adopt equity-aware approaches to ensure more just and effective risk assessments.

The study focused only on ex-ante risk assessments. The assessment of equity was limited to distributional equity; procedural and recognition aspects were examined but less comprehensively reported. While the Nepal case study provides valuable insights, findings might not be generalizable to all contexts. Data availability varied across assessments and regions, potentially influencing the results. The Type 5 analysis used a simplified approach due to data limitations.