Don't blame it all on climate change. Every water disaster has two stories: what the water did, and what human decisions made possible. The second story is harder to tell and easier to avoid. For the institutional habits that make avoidance systematic, see Governance Cannot Reflect on Itself. For the budget consequences, see Relief Crowds Out Prevention.

Climate change is real, its effects on water systems are serious and growing — and it is routinely invoked after disasters to absorb the accountability that belongs to decades of governance failure.

When a major flood occurs, the official account follows a recognisable sequence. Rainfall totals are cited. Records are noted as broken. The language of climate change enters within hours, sometimes within minutes. The framing is not wrong — extreme events are becoming more frequent and more intense, and the connection to a warming atmosphere is well established. The moral and political case for naming this is real.

But running alongside this truth, in almost every major water disaster on record, is another story that receives less official attention. Embankments that had not been maintained for years. Encroachment on floodplains that local engineers had flagged repeatedly. Early warning messages that reached administrative centres but not the communities that needed them. Drainage systems designed for a landscape that no longer exists. Institutional mandates so fragmented that no single agency could be held responsible for any of it.

The rainfall was extraordinary. The disaster was not inevitable.

These are not competing claims. They are two separate findings, each pointing to a different set of actions. The first points toward climate adaptation, long-term infrastructure investment, and emissions policy. The second points toward maintenance funding, floodplain enforcement, institutional coordination, and early warning reform. Both are necessary. But when the first absorbs all the explanatory weight, the second quietly disappears — and with it, the accountability of every institution whose decisions shaped the outcome.

This is not a pattern of bad faith, though bad faith sometimes plays a role. It is a pattern of institutional logic. Climate change is external, systemic, and shared. No agency caused it. No agency can be blamed for it. Governance failures are internal, specific, and attributable. They point to decisions made by identifiable institutions over identifiable periods of time. The external explanation is safe. The internal one is not.

The same dynamic operates in the opposite direction, and this must be named. Sometimes climate change is minimised or dismissed — attributed to natural variability, or treated as a distant concern — when acknowledging it would require systemic changes that powerful interests prefer to avoid. The pattern is not about which direction the distortion runs. It is about the use of causal attribution as a shield, in whichever direction the shield is most convenient.

The test is not whether climate change contributed to a disaster — in most cases it did, and the contribution is growing. The test is whether the attribution is complete. Was the floodplain protected? Was the embankment maintained? Was the warning issued in time, in a form communities could act on? Were settlements permitted in areas known to be at risk? If these questions are not asked alongside the climate question, the explanation is serving institutional convenience rather than institutional learning.

Separating the two is not scepticism about climate science. It is a precondition for honest governance.

After any significant water disaster, treat the hydrological event and the governance context as two distinct findings requiring two distinct lines of inquiry. Ask explicitly: what decisions, made by which institutions over what period, determined how damaging this event was? Hold both findings simultaneously — the climate signal and the governance record — and resist any framing that allows one to absorb the other. Design post-disaster reviews that are structurally required to address both.

When the same protective mechanism operates through vocabulary rather than attribution, see The Label Is Not the Solution. When crisis creates a rare opportunity to act on the governance findings that attribution has long suppressed, see Never Waste a Good Crisis.