The official ocean index says this is not an El Niño. The RONI reading for April–June 2026 is +0.47, which sits below the +0.50 threshold that would classify the cycle as anything other than Neutral. If your climate exposure review stops at that headline, it records a quiet season. The sub-national water balance says otherwise.

Neutral is a label about the ocean, not about your borrowers

ENSO indices describe the state of the tropical Pacific. They are useful, and they are deliberately coarse: one number, for one ocean, for a three-month window. A risk book is none of those things. It is a set of specific places, each with its own rainfall, its own evaporative demand, and its own tolerance for a dry quarter.

The gap shows up immediately when you put the two side by side. The basin is Neutral. The land is not.

What the land is reading

SPEI-3 measures the last three months of water balance — precipitation against evaporative demand — relative to the same season in the historical baseline. Zero is normal for the time of year. Negative is drier than that baseline, and readings past −0.5 are the conventional line where a dry spell stops being noise.

Counting the grid cells past that line, in a season the headline index calls Neutral:

CountryCells with SPEI-3 < −0.5Share
Thailand272 of 71837.9%
Vietnam579 of 1,13251.1%
Malaysia272 of 69839.0%

Correction, 13 August 2026. The May 2026 figures on this page were revised on this date. The underlying drought files were recomputed on 9 August 2026, and a month-scoped re-read of them (verification run 31711356352) returns Thailand −0.376 / 37.9%, Vietnam −0.504 / 51.1%, Malaysia −0.327 / 39.0%. Earlier versions of this page carried −0.288, −0.487 / 46.6% and −0.398 / 34.4% from the original May computation, and the Thai share was withheld on 11 August pending this read; it is now restored. The cell totals — 718, 1,132 and 698 — are unchanged, and the counts past the line (272, 579 and 272) come from the same run log.

Just over half of Vietnam’s cells are past the line. Close to two in five of Thailand’s and Malaysia’s are. That is the distribution underneath a single word — and the word does not carry it.

Why an average cannot tell you this

A national average would round this away. If a third of a country is dry and the rest is not, the mean lands somewhere unremarkable, and the unremarkable number is the one that reaches the committee. The share of area past a threshold survives that rounding: it answers how much of the map is affected, which is the question a portfolio actually asks.

It also travels differently across a book. Thirty-eight per cent of Thailand is not thirty-eight per cent of any given lender’s Thai exposure. Whether those cells sit under irrigated land, rain-fed smallholders, or a reservoir catchment is what turns a share into a loss estimate — and that is a question you can only ask of sub-national data.

What this means for credit, insurance and water

Credit. A Neutral classification is not evidence that agricultural borrowers had a normal quarter. The evidence is the water balance where the collateral is.

Insurance. Parametric triggers written against a basin-scale index and losses driven by local water balance are two different distributions. A Neutral season with a third of the map past −0.5 is exactly where that basis risk shows up.

Water. Allocation decisions are made per catchment. A national or oceanic label has no catchment in it.

The next reading is not in yet

The RONI value for June–August 2026 is pending and is expected around 5 September. Until it publishes, the honest position is that the ocean signal for that window is unknown — which is a further argument for watching the land directly rather than waiting on a label to tell you whether to look.

Neutral is a statement about the Pacific. Whether your exposure had a dry season is a different question, asked in a different place, and it has to be measured there.

GlobMaps builds sub-national climate intelligence for South-East Asia — province and grid-cell level drought signals across Thailand, Vietnam and Malaysia. Learn more at globmaps.com.