Insured losses by year
Global insured natural-catastrophe losses, $ billions, on Swiss Re’s current basis.
About this data
2025 came in at $107bn — down 24% on 2024, but still the sixth consecutive year above $100bn. The long-run trend is a 5–7% annual real increase, and Swiss Re puts roughly a one-in-ten chance on a $300bn peak year.
Insured losses by year
Global insured natural-catastrophe losses, $ billions, on Swiss Re’s current basis.
Data table
| year | usd_b | series | source_ref | value_basis |
|---|---|---|---|---|
| 2024 | 141 | annual_losses | swissre-2025-pr | Swiss Re restated 2024 basis, USD 141bn (see provider-spread chart for the USD 137bn originally published) |
| 2025 | 107 | annual_losses | swissre-2025-pr | USD 107bn insured nat-cat losses in 2025 — sixth consecutive year above USD 100bn |
Methodology & sources
Last updated: Aug 14, 2026Methodology
Insured losses only — the portion of catastrophe damage covered by insurance, always far below total economic loss (Gallagher Re put 2024 economic losses at $417bn against $154bn insured; Munich Re at $320bn against $140bn). The difference is the protection gap. THE 2024 NUMBER IS CONTESTED, and two of the four figures are Swiss Re’s own. $137bn is what Swiss Re originally published and remains by far the most widely quoted. $141bn is the restated basis Swiss Re itself uses when reporting 2025 at $107bn as "a 24% decline" (107 ÷ 141 − 1 = −24.1%, which only works from 141). Munich Re independently lands at $140bn. Gallagher Re’s $154bn reflects a wider event scope, not a different view of the same events. The annual trend chart uses Swiss Re’s CURRENT basis ($141bn for 2024) so it is internally consistent with the 2025 figure beside it. The provider-spread chart exists so that choice is visible rather than silent. OMITTED: the commonly cited $118bn for 2023 is a TREND value, not an actual, and is not plotted alongside actuals. CAVEAT: the 2025 forecast-vs-actual chart compares a trend projection ($145bn) with an outcome ($107bn). The 35% overshoot is not a modelling failure — sigma projects a trend line, not any particular year’s weather, and one quiet hurricane season moves the outcome far more than the trend does. Re-verified 2026-08-10.
Sources
Comparisons are informative, not definitive. See each source for definitions and limits.