ENSOID26 is the calibrated system dynamics model behind this paper. Set the ENSO intensity and the policy levers, and it runs the three waves of propagation — direct losses, indirect channels, and the macroeconomic result — so you can see which assumptions drive the numbers. Opens in a new tab.
Abstract
Indonesia enters the 2026 dry season facing an emerging super El Niño whose economic consequences still lie largely ahead of it. This paper develops a calibrated system-dynamics model to trace how such a shock propagates through the economy in three successive waves: the direct, biophysical losses in rain-bound sectors; the indirect, pecuniary channels of price, health, demand, and public finance; and the macroeconomic indicators that record the cumulative toll.
Simulated across the ENSO intensity range, the model's central finding is that the indirect wave exceeds the direct one — under a strong-to-super event the direct losses subtract on the order of 0.7 to 0.9 percent of GDP and the indirect channels a further 1.0 to 1.4 percent, so that the two together cost roughly 1.7 to 2.3 percent of GDP, while an accounting confined to scorched fields and failed harvests captures less than half the true toll.
In the aggregate this shaves 0.5 to 0.7 points off growth, lifts inflation by 3.5 to 4.4 points, widens the fiscal deficit by about half a percentage point, stresses between 2.8 and 3.5 million agricultural livelihoods, and tips 1.1 to 1.5 million people below the poverty line. The aggregate hit is absorbable; its danger is distributional, falling on the food-heavy household budget, the informal farm job, and an already-stretched fiscal position.
Policy cannot change the size of the shock but can substantially shape its incidence — through food-price stabilisation, adaptive social protection, fire prevention treated as macroeconomic policy, and prudent relief financing, all of which must be in place before the September–October peak. The reported figures are scenario outputs rather than forecasts; the model's contribution lies in revealing the structure of propagation and the points at which it can be interrupted.
The system dynamics model on which this paper is based, ENSOID26, is available here.