Understanding the gaps in the data

“The calculations tend to be biased in favour of insured losses: insurers have comprehensive data about their clients’ insured assets, and these form the backbone of loss estimates. When it comes to assessing the value of uninsured losses, the data are more patchy and hence the estimates less certain.

Likewise, those countries and regions in which there is greater insurance coverage have a greater proportion of more reliable data on insured assets, and so the overall estimate is more reliable.

There is also a skew towards big events that are widely reported, while we know that small events, while rarely reported, accumulate significant losses.

Data are not only collected by insurance companies. Municipalities, national governments, international organisations, and non-governmental organisations all gather information about disaster losses.

However these data are often not comparable, and may reflect inherent biases resulting from how they were collected and for what purposes. In some cases, for example, recording a higher number of victims could influence the amount of aid assigned; or conversely, a lower number of victims could absolve authorities from blame.

Taken together, the complexities in recording, reporting and compiling disaster impact records frequently result in data sets that are fragmented, disjointed or incomplete, particularly in databases with a global reach..”

https://www.undrr.org/explainer/uncounted-costs-of-disasters-2023