<p>Disaster Risk Indices (DRIs) are widely employed to integrate hazard, exposure, and vulnerability into a single quantitative measure for disaster risk assessment, prioritisation, and resource allocation. Although considerable attention has been devoted to their construction and application, comparatively little attention has been paid to their behaviour as derived measurable quantities whose characteristics are governed by the statistical properties of their constituent variables. In the present study, Disaster Risk Indices are examined from a measurement science perspective, and a variability-based analytical framework based on logarithmic variance propagation is developed to investigate how hazard, exposure, and vulnerability contribute to the resulting index. It is shown that, in multiplicatively aggregated systems, components exhibiting larger variability act as implicit weighting factors and exert disproportionate influence on the overall index. The framework is applied to the Disaster Risk Index adopted by the Sixteenth Finance Commission of India for allocations under the State Disaster Response Fund (SDRF) and State Disaster Mitigation Fund (SDMF). Empirical analysis demonstrates that exposure, represented primarily by population, exhibits substantially greater variability than hazard and vulnerability, resulting in strong dependence of both the index and associated allocations on population, with nearly 65% of the variation in allocations explained by exposure alone. To address this limitation, an Enhanced Disaster Risk Index (EDRI) is proposed in which hazard is parameterised using physically interpretable components representing intensity, frequency, and fragility. The enhanced formulation increases hazard variability, improves differentiation among hazard-prone regions, and produces a more balanced representation of disaster risk while preserving the original multiplicative structure and philosophy of the allocation framework. Comparative analysis indicates systematic redistribution favouring regions characterised by elevated hazard and fragility, particularly the Himalayan and north-eastern states. Although intended primarily as a proof-of-principle framework, the proposed approach demonstrates that the behaviour of composite risk indices depends not only on their mathematical formulation and assigned weights but also on the variability structure of their inputs, thereby establishing variability as a governing design parameter in composite indicators. The study provides a transparent and measurement-oriented framework for improving disaster risk assessment and contributes to the broader paradigm of measurement-based governance and the application of the Aswal Model to complex societal systems.</p>