El Niño and the Budget: brief for the Treasurer
Companion to the background note el_nino.md, which carries the climate evidence, the regional detail and the claim ledger. This brief covers only the fiscal and macroeconomic question.
The answer in one paragraph
El Niño is relevant to the Budget, but less than the historical record implies and through a different channel than most people expect. The revenue side is small: agriculture is about 2% of GDP, and a 21% fall in the winter crop is already in ABARES’ published forecasts — booked before the event was declared, and caused by an autumn drought rather than by the Pacific. On our central downside the El Niño adds a further 15% or so from here, meaning more than half the peak-to-trough production loss is already in the numbers. The spending side is where the exposure sits, and the largest single item is not drought assistance — it is bushfire. The Black Summer cost the Commonwealth more than $2.5 billion in recovery funding, against a few hundred million for a typical drought programme. Heat and fire risk are the parts of an El Niño that are reliably delivered; the drought is not. Nothing needs to be provisioned before MYEFO, but the decision inputs arrive in September and October.
1. What is likely to happen
The Pacific event is close to settled: the US Climate Prediction Center puts an 81% chance on a very strong El Niño in October–December (NOAA Climate Prediction Center, 2026). What is not settled is how much reaches Australia. On 24 events since 1951 the correlation between event strength and Australian spring rainfall is effectively zero, so the size of the Pacific anomaly is not the variable to plan on. The variable that matters is the Indian Ocean Dipole, which determines whether the drying reaches the southern cropping belt, and which will not resolve until September.
That gives three outcomes. The weights are our judgement, informed by the historical base rate and the current model guidance.
Scenario A — the north and east stay dry, the south holds (≈45%). No sustained positive Indian Ocean Dipole. Queensland, the Northern Territory and northern NSW have a dry spring, as they do in nearly every El Niño; Victoria, South Australia and Western Australia finish below but close to normal — on the historical record Victorian spring rainfall in El Niño years without a positive IOD averages 90% of normal, and South Australia 81%. This is the 1997–98 pattern. Winter crop lands near the current 54.5 Mt forecast. Incremental effect on the Budget: immaterial.
Scenario B — positive Indian Ocean Dipole confirms, the southern spring fails (≈40%). A positive IOD takes Victorian spring rainfall from 90% of normal to 68%, and South Australia from 81% to 60%. The southern states supply 40.3 Mt of the 54.5 Mt national forecast (WA 21.5, Vic 9.7, SA 9.1). A failed finishing spring in that belt historically costs 15–25% of yield. A 20% haircut removes about 8 Mt, taking the national crop to roughly 46–48 Mt. Gross value of agricultural production falls to somewhere near $90 billion rather than the forecast $98.3 billion.
Scenario C — compound event (≈15%). Positive Dipole plus a hot, dry spring across both halves of the country, on the 2002–03 or 2006–07 pattern. A 35% haircut on the southern crop plus further northern losses takes the national crop below 40 Mt. This is also the scenario in which fire risk is highest, because it pairs the rainfall deficit with the heat.
Note what all three have in common: even Scenario C leaves Australian agriculture well above where it sat in 2002 or 2006 in absolute terms. ABARES describes the current 54.5 Mt forecast, already down 21%, as still 4% above the ten-year average and the seventh largest winter crop on record (ABARES, 2026b). The sector is falling from a high base, not into a hole.
2. What it does to the economy
Much of the effect is already published. ABARES’ June 2026 forecasts for 2026–27 (ABARES, 2026a):
ABARES forecasts for 2026–27, published June 2026
| Measure | Forecast | Change |
|---|---|---|
| Gross value of agricultural production | $98.3bn | −5% |
| Agricultural export value | $74.8bn | −9% |
| Crop production value | $50.9bn | −8% |
| Crop export value | $37.0bn | −10% |
| Livestock and livestock product value | $47bn | −2% |
| Average broadacre farm business profit | — | −70% |
Source: ABARES Agricultural Commodities Report, June Quarter 2026 (ABARES, 2026a)
Three points for the macro forecast.
The farm profit number is the one that matters, and it is not a GDP number. A 70% fall in average broadacre farm business profit reflects revenue falling onto a fixed and rising cost base — fuel and fertiliser prices are elevated. It is the right indicator for farm-sector tax receipts and for rural credit stress. It is not the right indicator for GDP, where the relevant measure is farm value added.
The historical GDP rule of thumb overstates the effect today. Treasury found the 2002–03 drought cut farm GDP by 24.8% and subtracted 0.9 percentage points from growth (Australian Treasury, 2004). That arithmetic implies farm GDP was around 3.6% of the economy at the time. Agriculture’s value-added share is now closer to 2%. The same proportional collapse today would subtract roughly 0.5 percentage points, not 0.9. Scenario B is a materially smaller shock than 2002–03 and would plausibly cost 0.1 to 0.3 percentage points of GDP growth; Scenario C, 0.3 to 0.5. Scenario A costs nothing measurable.
Prices offset volumes, partially. ABARES expects crop export value to fall 10% while volumes fall further, “as lower volumes are expected to offset higher prices”, and domestic grain prices to rise (ABARES, 2026a). Nominal GDP and export receipts therefore fall by less than production does. Two qualifications: global grains and oilseeds production is running at a record high, which caps how far world prices can move on an Australian shortfall; and El Niño is a global phenomenon that also disrupts Southeast Asian rice, the Indian monsoon and southern African maize, which works the other way. On balance the price offset is real but should not be assumed large.
There is also a mild domestic inflation channel through food prices, relevant to indexed payments and to the monetary policy outlook rather than to the fiscal aggregates directly.
3. What it does to the Budget
Ordered by expected fiscal magnitude, largest first.
1. Natural disaster spending — the largest and least predictable exposure. This is the item that should hold the Treasurer’s attention. The 2019–20 fire season cost the Commonwealth $2 billion through the National Bushfire Recovery Fund plus a further $691.2 million through existing disaster support mechanisms — recovery payments and allowances for individuals (Australian National Audit Office, 2021). That is roughly five times the entire Future Drought Fund commitment of $519.1 million over eight years (Prime Minister of Australia, 2026). Fire risk is the part of the El Niño signature that is reliably delivered: heat correlates with event strength where rainfall does not, and this event is forecast to be very strong. The exposure is contingent, lumpy, and concentrated in the December–February window — after MYEFO.
2. Drought assistance. Farm Household Allowance is demand-driven and will rise in Scenarios B and C; it is a fortnightly payment and scales with the number of eligible households rather than with the size of the production loss. Discretionary programmes are the larger variable: the 2018–19 response included $297 million over four years for the Drought Communities Programme Extension and $148.5 million for a round of household support. A Scenario B response would plausibly sit in the low hundreds of millions; Scenario C higher, and with political pressure for more.
3. Tax receipts. Company and personal income tax from the farm sector falls with the 70% profit decline, but off a base that is a small share of total receipts. There is a partly offsetting timing effect: Farm Management Deposits are deductible when made and assessable when withdrawn, and drought triggers early-access provisions. Drought years therefore pull forward taxable income from the deposit pool, which cushions the receipts line in the year of the drought and costs the Budget later. This is a stabiliser that works in the government’s favour on 2026–27 timing.
4. Regional consumption and GST. Second-order, and partly offset by drought assistance flowing back into the same regional economies.
5. Water. Urban supply is far less exposed than during the Millennium drought because of desalination capacity that did not exist in 2002. The Commonwealth exposure is more likely to be regional and remote town supply, which is small in dollar terms.
4. Decision timetable
| Date | Event | What it settles |
|---|---|---|
| Mid-August 2026 | CPC diagnostic discussion, 13 August | Whether the Pacific event is still strengthening |
| Late August – September | Dipole Mode Index sustained above or below +0.4 °C | Scenario A versus B/C — the single highest-value observable |
| Early September | AFAC Spring Seasonal Bushfire Outlook | First fire-season assessment incorporating the declared El Niño |
| September | ABARES September crop report | First crop forecast incorporating the El Niño |
| November–December | Harvest | Actual production |
| December 2026 | MYEFO | Natural point to reflect the outcome; the crop will be known |
| December – February | Fire season | The contingent spending exposure |
| May 2027 | Budget | Full picture, including any recovery funding |
The practical implication: nothing requires provisioning before MYEFO on the drought side, because the crop outcome will be known by then. The fire-season exposure runs the other way — it crystallises after MYEFO and before the Budget, which is the standard pattern for disaster funding and is normally handled through contingency rather than forward estimates.
5. What we would tell you not to do
Do not scale the fiscal response to the Pacific index. The evidence in the companion note is that event strength does not predict Australian rainfall. A response calibrated to “the strongest El Niño since 1950” would have been wrong in 1997, when Australian spring rainfall came in near normal.
Do not treat the country as one unit. Northern NSW and southern Queensland are already in drought and their crop is already written down 37% and 38% respectively; Victoria is down 4% and South Australia is flat (ABARES, 2026b). Assistance calibrated nationally would over-serve the south and under-serve the north.
Do not read the −21% crop forecast as an El Niño effect. It is not. It is the consequence of an autumn drought in the northern cropping zone that preceded the event, and it was published on a May climate outlook. The El Niño’s incremental contribution is still ahead of us and lands mostly on the southern crop.
6. Bottom line for the Budget
On a probability-weighted basis the incremental hit to 2026–27 GDP growth from here is of the order of 0.1 to 0.2 percentage points, with a tail to 0.5. The revenue effect is small in absolute terms and partly self-cushioning through Farm Management Deposit withdrawals. The spending exposure is dominated by a contingent fire-season liability that history suggests can run to $2 billion or more in a bad year, and that is the item worth stress-testing rather than the drought line.
The most useful thing that can be done between now and MYEFO is to watch one number — the Dipole Mode Index through September — and to make sure the fire preparedness and disaster funding machinery is ready, because that is the part of this event that is not conditional.
Sources and claim ledger: see el_nino.md and el_nino_ledger.md. Scenario weights and the production ranges in Section 1 are judgements, not model output; the arithmetic behind them is shown in the text. GDP-share arithmetic in Section 2 is derived from the Treasury 2002–03 estimate and current national accounts shares and is an order-of-magnitude calculation, not a Treasury-equivalent estimate.