# El Niño Event Studies: Documented Market Behavior in Strong Events Research report covering the four strong/very strong El Niño events of the modern era — **1982-83 (peak ONI ~2.2), 1997-98 (~2.4), 2015-16 (~2.6), 2023-24 (~2.0)** — plus cross-event academic evidence. Every quantitative claim is sourced; the companion file `data/events/event_studies.json` contains the structured entries. Where sources give ranges or qualitative statements, we report them as such. Attribution caveats are flagged inline — this literature is full of confounds, and pretending otherwise would be malpractice. **Legend:** *Documented* = sourced historical fact for that event. *Mechanism* = general channel supported by theory/model evidence, not necessarily observed that event. *Contested* = sources disagree or attribution is disputed. --- ## 1. 1982-83 (peak ONI ~2.2) - **Peruvian anchovy / fishmeal — supply collapse (documented).** 1983 anchoveta yield <100,000 t, under 1% of the ~12.5 Mt 1970 peak, as warm water halted upwelling ([T&F 2023](https://www.tandfonline.com/doi/full/10.1080/00076791.2023.2292751)). Caveat: the stock was already depleted from 1970s overfishing. Peru's total losses were $3.28B, 11.6% of 1983 GDP ([Andina/SENAMHI](https://andina.pe/agencia/noticia-como-impacto-nino-peru-los-eventos-19821983-19971998-y-2017-943812.aspx)). - **Cocoa +22% (documented).** IMF records average price rising from $1,742/t (1982) to $2,119/t (1983) and $2,396/t (1984), with world production cut 12% below 1981/82 on West African drought and Ecuador floods ([IMF Primary Commodities](https://www.elibrary.imf.org/display/book/9781451943061/ch003.xml)). Tea nearly doubled in 4 months on drought fears — but reversed sharply in 1985 (fear-driven). - **Sugar — NO rally (documented, important).** Prices stayed at "very low levels" into 1984 amid a global glut ([World Bank](https://documents1.worldbank.org/curated/en/828081468739443509/pdf/multi_page.pdf)). The "El Niño = sugar spike" template failed outright here. - **Coffee — flat (documented).** ICO indicator $1.25→$1.28/lb (1982→83); record stocks absorbed the Côte d'Ivoire shortfall. The famous 1985-86 coffee spike was a separate Brazilian drought, not El Niño. - **US corn/soybeans up (documented, confounded).** 1983 yields −28% corn, −17% soy; but the PIK acreage-reduction program cut plantings simultaneously — weather vs. policy attribution is genuinely split ([Chicago Fed](https://www.chicagofed.org/publications/chicago-fed-letter/1988/october-14)). - **Atlantic hurricanes suppressed (documented).** 1983: 4 named storms, quietest of the satellite era ([NHC](https://www.nhc.noaa.gov/data/mwreview/1983.pdf)) — yet the one major (Alicia) still cost Texas $3B. Suppression ≠ zero risk. - **Australia — severe drought (documented).** Wheat crop 63% of average; farm production −24%; ~A$3B economy-wide loss; contributed to the 1982-83 recession ([ANZ](https://www.anz.com/Documents/FXOnline/Drought_Aug02.pdf)). - **US natural gas — mild winter (documented mechanism, unquantified price).** Winter 1982-83 was unusually mild; no price series found. Dec 1983 brought record cold — the signal reversed fast. - **Global damage:** contemporary estimates up to $14B ([World Bank](https://documents1.worldbank.org/curated/en/790851468740967268/pdf/multi0page.pdf)). Callahan & Mankin's later $4.1T five-year income-loss estimate is *contested* (see §5). ## 2. 1997-98 (peak ONI ~2.4) - **Peruvian anchovy — best-quantified impact of the event (documented).** Anchovy catch −79.6% in 1998; total fisheries landings −55% vs 1996; "extremely high fishmeal prices" ([Exalmar OM](https://www.exalmar.com.pe/wp-content/uploads/2016/05/Final-Offering-Memorandum.pdf), [Broad et al. 1999](https://www.ilankelman.org/glantz/Broadetal.1999ElNino.pdf)). - **Coffee +30% in 1997, peak 180¢/lb (documented, attribution contested).** FAO itself attributes the rally mostly to arabica tightness, NOT El Niño ([FAO](https://www.fao.org/unfao/bodies/council/cl115/W9751e.htm)). - **Palm oil output −15% Malaysia (documented, order-of-magnitude).** Sumatra/Kalimantan rainfall −25%; Indonesia banned CPO exports; price signal muddied by Asian-crisis FX ([scialert](https://scialert.net/fulltext/?doi=ajes.2016.1.8)). - **Indonesian rice +27% CPI y/y (documented).** Drought-delayed planting plus the currency crisis's import constraint; record 3.5 Mt import requirement, 2 Mt met by food aid ([FAO/WFP](https://www.fao.org/4/w8458e/w8458e00.htm)). - **IDR −70%+ (documented, NOT El Niño).** Asian financial crisis; El Niño was a food-inflation amplifier only. Same for copper and gold — down on Asian demand destruction ([UN WESS 1998](https://www.un.org/en/development/desa/policy/wess/wess_archive/1998wess.pdf)). - **US energy −24% Jan 1998 (documented).** Warmest US winter on record at the time cut heating demand ([ISWS](https://www.isws.illinois.edu/pubdoc/DCS/ISWSDCS2000-01.pdf)). - **California storms ~$850M damage, ~$1.1B insured (documented).** Mitigation after forecasts reduced losses vs 1982-83 ([Changnon via ggweather](https://ggweather.com/enso/calif_flood.htm)). - **Fires:** $4.4-4.5B (EEPSEA/WWF) — estimates range $1.6-4.5B depending on haze/health inclusion. **Global losses ~$33-36B, ~23k deaths** ([WMO](https://library.wmo.int/viewer/53800/download?file=wmo_1145_en.pdf)). - **The core caveat of this event:** the Asian financial crisis swamped weather signals in IDR, metals, and SE Asian softs. Depreciated currencies even *lowered* dollar prices of cocoa, coffee, rubber despite supply damage. ## 3. 2015-16 (peak ONI ~2.6, strongest on record) - **Sugar — strong rally (documented).** Best-performing commodity of 2015; Thai cane output forecast cut below 95 Mt (from 105.95 Mt) on El Niño-weakened monsoon; shift to global deficit ([Agweek](https://www.agweek.com/news/drought-hits-thai-sugar-next-years-crop-dying)). Note the rally started from 6-year lows. - **Palm oil +25% in 2016 (documented, lagged).** BMD CPO's strongest annual gain in 6 years, arriving 6-9 months after the late-2015 drought via the biological yield lag ([Reuters/Jakarta Globe](https://jakartaglobe.id/business/reuters-survey-new-indonesia-plantations-will-blunt-el-nino-hit-2016-palm-yields)). - **Robusta coffee (documented, forecast-based).** Reuters poll: +30% expected for 2016, biggest gain since 2010, on Vietnam/Indonesia dryness plus Brazil drought — first supply deficit in 6 years ([Fortune](https://fortune.com/2016/07/28/coffee-prices-expensive/)). - **Cocoa — supply hit, price fell (documented, counterintuitive).** Sharp Côte d'Ivoire production drops from hot harmattan, but demand/grind weakness dominated; cocoa fell through much of 2016 ([Neuberger Berman](https://www.nb.com/ar-SA/insights/a-sea-change-for-commodity-prices-why-markets-appear-to-be-underpricing-el-ni-o-risks)). - **US natural gas — the cleanest trade of the event (documented).** Record-warm winter (HDD −18%): March 2016 Henry Hub $1.71/MMBtu, lowest CPI-adjusted monthly in 40 years; propane demand −16%, heating oil −18% ([EIA](https://www.eia.gov/todayinenergy/detail.php?id=25952)). Shale glut compounded it. - **WTI — DOWN (documented, anti-consensus).** Supply glut dominated; runs counter to the IMF model finding that El Niño shocks historically raise oil slightly. Macro swamped weather. - **South African maize — record (documented).** White maize futures hit 4,781 rand/t (Jan 2016); regional maize output −27%; worst drought in ~3 decades ([Reuters/trust.org](https://news.trust.org/item/20160104134248-oxpiz/)). - **Rice — output down, prices flat (documented).** Global production −5 Mt but huge Thai/Indian stocks capped prices — stocks absorbed the shock ([ODI](https://odi.org/documents/4991/9939.pdf)). - **India:** monsoon −14% (second straight drought year), yet food inflation stayed subdued via buffer stocks/MSP — a −14% deficit need not produce CPI or INR stress. - **Global food prices DOWN despite the strongest El Niño on record (documented, important).** FAO Food Price Index −19.4% y/y July 2015, near 6-year low — well-supplied grains and cheap energy overwhelmed the supply shocks. - **Indonesia fires $16.1B (~1.9% of GDP)** ([World Bank](https://documents1.worldbank.org/curated/en/776101467990969768/pdf/103668-BRI-Cost-of-Fires-Knowledge-Note-PUBLIC-ADD-NEW-SERIES-Indonesia-Sustainable-Landscapes-Knowledge-Note.pdf)). Ethiopia: 9.7M affected, ~$2.2B humanitarian gap. ## 4. 2023-24 (peak ONI ~2.0) - **Cocoa +120% y/y (documented, multi-causal).** $5,798/t Feb 2024 (65-yr high); spot reportedly ~$12,000/t at peak. Excess rain → black pod, then dry harmattan in Côte d'Ivoire/Ghana (60% of world supply); Ghana output ~55% of average ([Fortune](https://fortune.com/2024/02/09/chocolate-cocoa-prices-record-highs-climate-change-el-nino/)). Rally began before the El Niño peak; disease and structural underinvestment co-drove it. - **Robusta record highs (documented).** London ~$4,205/t Jun 2024; Vietnam crop −20% on Central Highlands drought, amplified by thin inventories ([Tridge/Reuters](https://www.tridge.com/news/coffee-prices-today-june-20-2024-robusta-inc-fsymia)). - **Sugar — 12-year highs Q4 2023 (documented, anticipatory).** Weak Asian monsoon, India/Thailand cane shortfalls, India export curbs. The rally *anticipated* the event's peak rather than following it. - **Rice — 15-year highs (documented, policy-driven).** India banned white-rice exports Jul 2023 (~40% of world trade) amid El Niño/monsoon fears — a policy spike, not a harvest failure ([USDA FAS](https://www.fas.usda.gov/data/rice-export-prices-highest-more-decade-india-restricts-trade)). - **US natural gas — record-low 2024 annual average, $1.51/MMBtu Mar 2024 (documented, confounded).** Warm winter cut demand but record production + high storage were major co-drivers ([EIA via EnerKnol](https://enerknol.com/henry-hub-spot-natural-gas-prices-hit-record-low-last-year-eia/)). - **Panama Canal — transits cut 36→18/day (documented).** Driest Oct since 1950; expressed as slot-auction premiums and rerouting rather than a clean freight-index move ([Earth.Org](https://earth.org/panama-canal-authority-cuts-ship-crossings-further-as-el-nino-induced-drought-intensifies/)). - **Peruvian anchovy — cleanest attribution of the event (documented).** First 2023 season cancelled outright, ~$1.4B lost export revenue, fishmeal $1,450→$1,720/t ([SeafoodSource](https://www.seafoodsource.com/news/supply-trade/peru-opens-anchovy-fishery-s-north-central-zone-for-first-fishing-season-of-2024)). - **Australian wheat — crop down ~28%, prices flat (documented).** ~26 Mt vs ~36 Mt record; the feared 30%+ wipeout didn't materialize and global prices fell. No clean AUD signal; the 2023-24 fire season was mild. - **Palm oil — lagged (documented).** CPO only rallied to RM4,500-5,000/t in late 2024/Q1 2025 — the 6-9 month biological lag again, confounded by B40 biodiesel. - **Atlantic hurricanes — the big exception (documented).** 20 named storms (4th most on record): record-warm Atlantic SSTs overwhelmed El Niño shear ([Klotzbach et al., BAMS 2024](https://journals.ametsoc.org/view/journals/bams/105/9/BAMS-D-23-0305.1.xml)). - **Insurance:** 2023's $108B insured nat-cat losses were driven by US convective storms ($60B) and the Türkiye/Syria quake — explicitly NOT El Niño per [Swiss Re](https://www.swissre.com/press-release/New-record-of-142-natural-catastrophes-accumulates-to-USD-108-billion-insured-losses-in-2023-finds-Swiss-Re-Institute/a2512914-6d3a-492e-a190-aac37feca15b). ## 5. Cross-event evidence (academic / institutional) - **Brunner 2002** (via [OECD 2024](https://www.oecd.org/en/publications/economic-outlook-for-southeast-asia-china-and-india/volume-2024/issue-1_3bbe7dfe-en/full-report/component-5.html)): ~+4% real commodity prices per 1-std-dev ENSO shock; ENSO explains ~20% of commodity price fluctuations. Rice and palm oil most SOI-sensitive. - **Cashin-Mohaddes-Raissi ([IMF WP/15/89](https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2015/_wp1589.pdf)):** non-fuel commodities +5.25% after four quarters; metals and energy also rise in the model. The oil result is a demand-channel *model* effect — *contested* against the fact that crude fell hard in both 1997-98 and 2015-16 for macro reasons. - **Ubilava 2017 ([World Development](https://www.sciencedirect.com/science/article/abs/pii/S0305750X15310172)):** El Niño *lowers* wheat prices; La Niña raises them up to ~6%. Contradicts the naive "El Niño = higher grains" prior — Australian droughts don't move world wheat when US/ Canadian/Black Sea crops are fine. - **Ubilava & Holt 2013:** vegetable oils (coconut/palm kernel) show the strongest and most persistent ENSO price response in the literature. - **Coffee (Ubilava 2012; ECB):** El Niño raises robusta, depresses arabica; the average coffee effect is ~zero. Brazil's big spikes are frost/non-ENSO droughts (1977, 1994, 2021). - **ECB box 6/2023:** a strong El Niño raises global food commodity prices up to ~9%, peaking ~16 months after onset; soy/corn/rice up, wheat/coffee ~zero. Also documents the announcement effect: June 2023 soybeans +12%, corn +10% in the week after NOAA's declaration (partly option-implied). - **Australia:** farm GDP fell 0.7%-25.4% (avg −12.6%) across the last five El Niños ([NAB](https://business.nab.com.au/wp-content/uploads/2015/11/Impact-of-El-Nino-on-farm-GDP-2015.pdf)); farm GDP is ~2% of GDP, so the headline hit is small. Clean AUD/INR/IDR/BRL attributions do not exist — FX moves in these years were dominated by crises and terms-of-trade cycles. - **Global GDP (Callahan & Mankin, Science 2023):** $4.1T (1982-83) and $5.7T (1997-98) cumulative income losses over ~5 years. **Contested** — ~100x contemporary direct-damage tallies ($8-45B); Pielke Jr. has publicly called the result far-fetched. Treat as an econometric growth-drag estimate, not an event loss tally. ## 6. Synthesis: what moves reliably, what doesn't **Most reliable (moved the "right" way in 3-4 of 4 events):** 1. **Peruvian anchovy / fishmeal** — the purest El Niño trade. Catch collapses or seasons get cancelled in every strong event (1982-83, 1997-98, 2015-16 partial, 2023-24); fishmeal spikes to records. Mechanism is direct and local: warm water kills upwelling. 2. **US natural gas (winter demand)** — warm-winter demand destruction in every event with a market to observe (1997-98, 2015-16, 2023-24 all saw sharp price weakness; 1982-83 mild but unquantified). Magnitudes −20% to −50%+ at the front of the curve, but always confounded by supply (shale glut twice). Direction reliable, magnitude not. 3. **Palm oil / vegetable oils** — SE Asian drought hits fresh-fruit-bunch yields with a 6-9 month lag. Up in 1997-98 (output), 2016 (+25%), and late 2024. Strongest econometric ENSO response in the literature. Trade the lag, not the ONI print. 4. **Sugar (Asia/India cane)** — up hard in 2015-16 and 2023, but *flat-to-down in 1982-83* (glut) and weakly documented for 1997-98. Reliable only when global stocks are tight. 5. **Robusta coffee / cocoa (SE Asia + West Africa drought)** — robusta up in 1997, 2016, 2024; cocoa up huge in 1982-83 and 2023-24 but *down in 2016 despite supply damage*. Direction right most of the time, but demand-side and disease co-drivers can dominate. **Unreliable / mixed (be honest):** - **Wheat** — the academic evidence says El Niño is *bearish*-to-neutral for world wheat (Ubilava 2017): Australian/Argentine droughts don't offset benign Northern Hemisphere conditions. 2023 proved it: Australian crop −28%, global prices fell. - **Soybeans/corn (US)** — El Niño summers are often *benign* for the US Midwest; the 1983 drought spike is confounded by the PIK program, and 2015 saw record crops. The ECB's +12%/+10% June 2023 pop was an announcement/positioning effect that faded. - **Rice** — twice a non-event price-wise (1998 stocks, 2015 stocks), once a policy spike (2023 India ban). Trades on government export policy more than on yield. - **WTI/copper/gold** — no robust signal. Models say slight up; events say down (1997, 2015) because macro dominates. Metals are demand assets, not weather assets. - **FX (AUD, INR, IDR, BRL)** — no clean attribution anywhere. 1997-98 IDR was the Asian crisis; India 2015 absorbed a −14% monsoon deficit without INR stress. - **Atlantic hurricanes** — statistically real suppression (1982: 9 named, 1983: 4, 1997: 9, 2015: below-average ACE) but NOT deterministic: 2023 produced 20 named storms as warm Atlantic SSTs overrode shear. Don't short hurricane risk on an ONI print anymore. **Timing (lead/lag vs the ONI peak):** - Fishmeal reacts *immediately* (local, direct). California flooding and US winter energy are *synchronous* with the winter peak. Grains/softs react at the *growing-season* teleconnection, which can lag ONI by 3-6 months. Palm oil lags 6-9 months by plant biology. The ECB's aggregate food-price effect peaks ~16 months after onset. And increasingly, markets trade the *forecast*: sugar and grains popped on NOAA's June 2023 declaration, months before the event peaked. **Structural caveat for all four events:** price response = f(shock, stocks, policy, macro). Every clean El Niño rally in this record happened against tight inventories or deficit expectations; every "failed" signal (sugar 1982, rice 2015, wheat 2023, global food CPI 2015) happened against glut. The ONI value alone tells you the weather distribution, not the P&L.