11. Climate and Weather
Teaching rewrite of climate science, climate policy, emissions trading (with a focus on the EU ETS), emission derivatives, and weather derivatives. Original wording; worked numbers recomputed. Prices are teaching values. Not investment advice.
Introduction
Two very different markets share this lesson:
- Emissions — a commodity created by regulation: the right to emit a tonne of CO₂. Its price depends on policy, the economy, and fuel switching between coal and gas.
- Weather — a risk, not a commodity: a mild winter cuts heating demand; a cool summer cuts ice cream and air-conditioning sales. Weather derivatives pay on measured temperature, rain, or snow.
First, the definitions. Weather is short-term, local atmospheric conditions (minutes to days). Climate is long-run patterns of temperature, humidity, and rainfall over seasons to decades. Global warming is the long-term heating since pre-industrial times (1850–1900) caused mainly by burning fossil fuels; climate change is the broader term covering human and natural warming and its effects.
This lesson covers:
- Greenhouse gases, CO₂e, the carbon cycle, and feedback loops
- Consequences, the IPCC findings, and the sceptics’ arguments
- Policy: IPCC → Rio → Kyoto (ETS, CDM, JI) → Paris
- Carbon price drivers, the EU ETS design, and cap-and-trade vs carbon tax
- Emission derivatives: spot, futures, forwards, fair value, repos, swaps, options, trading strategies
- Weather derivatives: HDD/CDD indices, futures, swaps, options, and hedging examples
11.1 The science
Greenhouse gases
Sunlight warms the surface; the surface re-radiates infrared, and greenhouse gases (about 1% of the atmosphere) trap part of it. Without them Earth would be about 30 °C colder.
| Gas | Main sources | GWP (100-yr, AR4) |
|---|---|---|
| Water vapour (H₂O) | Most abundant; follows temperature | — |
| Carbon dioxide (CO₂) | Burning fossil fuels, respiration | 1 |
| Methane (CH₄) | Coal mining, oil and gas, landfill, livestock | 25 |
| Nitrous oxide (N₂O) | Fuel combustion, fertilisers | 298 |
| HFCs / PFCs | Refrigeration, air conditioning | up to 14,800 / 12,200 |
| Sulfur hexafluoride (SF₆) | Electrical and industrial uses | 22,800 |
CO₂ equivalent (CO₂e) = mass emitted × Global Warming Potential. Example: 10 t of methane = 10 × 25 = 250 t CO₂e.
These are the IPCC AR4 values, still used in older Kyoto-era accounting. The AR5/AR6 values differ (e.g. methane ≈ 28–30, N₂O ≈ 265–273, SF₆ ≈ 23,500–25,200). Check which GWP set a scheme uses.
Emission sources (IPCC 2014): energy 35%, agriculture, forestry and land use 24%, industry 21%, transport 14%, buildings 6.4%.
The carbon cycle
flowchart LR
A[CO₂ in air] -->|photosynthesis| P[Plant starch / wood]
P -->|respiration| A
P -->|decay| A
P -->|burning, incl. fossil fuels| A
P -->|eaten| N[Animal]
N -->|respiration| A
N -->|waste → microbes| A
N -->|death / eaten| N
Every path eventually returns carbon to the air as CO₂ — hence a cycle, though not a single simple route. Fossil fuels release carbon locked away for millions of years.
Feedback loops
Warming changes cloud cover, precipitation, wind, sea level, the water cycle (more droughts and floods), and season length. Some feedbacks cool (more cloud reflects sunlight); others amplify (thawing Siberian permafrost releases methane). Man-made aerosols reflect sunlight and cool temporarily, but harm air quality.
11.2 Consequences and debate
Mars (thin CO₂ atmosphere, little methane or water vapour) is frozen; Venus (about 154,000 times Earth’s CO₂) has a runaway greenhouse hot enough to melt lead. Earth sits in between. Likely effects of a stronger greenhouse:
- Warmer on average — some regions benefit, most lose
- More evaporation and rain overall, but some regions wetter and others drier
- Rising seas: melting ice plus thermal expansion of warmer water
- Some crops grow faster with more CO₂, but the best growing areas shift
IPCC Fifth Assessment Report (2014): human influence is clear; emissions are the highest in history; continued emissions risk severe, pervasive, irreversible impacts; mitigation and adaptation are complementary; no single option is enough.
The Sixth Assessment Report (2021–2023) went further: it is “unequivocal” that human influence has warmed the climate, and warming had reached about 1.1 °C above pre-industrial levels.
The sceptics’ arguments, briefly
- Solar activity, not GHGs, drives climate (NASA disputes this)
- Consensus isn’t proof; funding and attention favour the majority view
- Past warming (1918–1940), cooling (1940s–70s — attributed by others to industrial sulfur aerosols), and a 1998–2005 plateau
- Vague language (“might”, “could”), model disagreement, selective time frames, little Southern Hemisphere data
- A 2005 UK House of Lords report questioned the objectivity of the IPCC process and the effectiveness of Kyoto-style targets
For markets, what matters is that policy — not the debate — sets the supply of allowances.
Checkpoint 1: science
Checkpoint — sections 11.1 and 11.2
1 / 5Weather vs climate — which statement is right?
Pick an answer to see if you are right.
11.3 Policy history
flowchart LR
A[1979<br/>First World Climate Conference] --> B[1988<br/>IPCC formed]
B --> C[1992<br/>Rio Earth Summit · UNFCCC]
C --> D[1997<br/>Kyoto Protocol]
D --> E[2008–2012<br/>Kyoto period]
E --> F[2015<br/>Paris Agreement]
- IPCC (WMO + UN Environment) reviews research; reports in 1990, 1995, 2001, 2007, 2014 (and 2021–23).
- UNFCCC (Rio 1992) named Annex I parties: the 1992 OECD members plus “economies in transition” (e.g. Russia).
Kyoto Protocol
Legally binding targets for Annex I countries, from −8% to +10% of 1990 levels, aiming for an overall −5% in 2008–2012. One AAU (assigned amount unit) = 1 t CO₂e. A facility that emitted 100 t in 1990 might get 92 AAUs a year — it must cut, or buy the difference.
| Mechanism | Type | How it works | Credit |
|---|---|---|---|
| Emissions trading (ETS) | Market | Buy/sell rights to emit; cap and trade | AAU / EUA |
| Clean Development Mechanism (CDM) | Project | Annex I invests in a developing country’s project; UN certifies reductions below the no-project baseline | CER |
| Joint Implementation (JI) | Project | Annex I invests in another Annex I country; allowances transfer, no net global change | ERU |
CDM projects: energy efficiency, methane capture, coal-to-gas switching, destroying industrial gases — 7,803 projects in 140 countries by 2018.
Paris Agreement (COP21, 2015)
- Keep warming well below 2 °C, pursuing 1.5 °C
- Peak emissions as soon as possible, then reach climate neutrality
- Every party sets a nationally determined contribution (NDC)
- Protect sinks; developed countries support developing ones with finance and technology
- Global stocktake in 2023 and every five years
- A new mechanism replaces CDM and JI
For 1.5 °C, CO₂ must fall about 45% by 2030 versus 2010 and reach net zero around 2050.
The first global stocktake concluded at COP28 (Dubai, 2023), calling for “transitioning away from fossil fuels”. The Paris replacement for CDM/JI is the Article 6 framework: 6.2 (country-to-country transfers) and 6.4 (the Paris Agreement Crediting Mechanism), with rules finalised at COP29 (2024).
11.4 Carbon price drivers
A different supply curve
In a normal market, supply rises with price. In a cap-and-trade scheme, supply is fixed by regulators each year (plus any eligible offsets), so the curve is vertical: price is set by where demand meets the cap.
| Feature | Effect |
|---|---|
| Fixed supply | Demand shocks move price, not volume |
| Penalty (e.g. EUR 100/t) | Discourages non-compliance |
| Banking | Allowances keep value into later periods → demand flattens at high prices rather than vanishing |
The source suggests the fine sets a ceiling — “no one would pay 110 if the fine is 100”. In the EU ETS, paying the penalty does not release you from the obligation: missing allowances must still be surrendered the following year. The fine is therefore added to the allowance cost, not a substitute for it, and is not a price cap. (A pure “pay the fine and walk away” scheme would be capped.)
| Driver | Effect on allowance price |
|---|---|
| Economic growth | More output → more emissions → higher |
| Allowance supply | Over-allocation → collapse (phase 1, post-2008); political decisions matter |
| Abatement cost | Firms abate if cheaper than buying; abatement cost anchors the price |
| Weather and fuel prices | Dear gas or cold weather → switch to coal → more allowances needed → higher |
| Penalties | Strength of enforcement |
| Low-carbon generation | More nuclear and renewables → lower; expensive transition → higher |
| Regulation | Allocation rules, offsets, banking, compliance strength |
Correlations (intuitive): carbon is positively correlated with gas and power prices, and negatively with coal (dear coal → switch to gas → less carbon demand).
Worked example: fuel-switching price
Gas at EUR 25/MWh (thermal) in a 50%-efficient plant → 50/MWh of power; coal at 10 in a 36% plant → 27.78. Emission factors: coal 0.90, gas 0.37 t CO₂/MWh.
Coal and gas cost the same when:
Carbon price = (50 − 27.78) ÷ (0.90 − 0.37) = 22.22 ÷ 0.53 ≈ EUR 41.93/t
Above this, gas is cheaper → less coal burn → less carbon demand. This links carbon to the clean spark and clean dark spreads from lesson 8.
Reading forward curves (Barclays, 2007)
The author’s argument: contango signals that long-term demand will exceed supply; backwardation signals short-term tightness but a long-term surplus. In 2007, the dirtiest fuels (coal, crude) were in contango and the cleaner one (gas) backwardated. The conclusion was that markets didn’t believe policy would shift demand away from dirty fuels. (Curve shape also reflects storage and financing costs, so treat this as one lens among several.)
Checkpoint 2: policy and drivers
Checkpoint — sections 11.3 and 11.4
1 / 5Which Kyoto mechanism produces CERs?
Pick an answer to see if you are right.
11.5 The EU Emissions Trading System
“The commodity being bought and sold does not exist: it is the certified absence of carbon emissions.” — paraphrasing The Economist, 2006
| Phase | Years | Notes |
|---|---|---|
| 1 | 2005–2007 | Trial; national caps from estimates → over-allocated, price collapsed |
| 2 | 2008–2012 | Kyoto period; ~90% free allocation; aviation from 2012 |
| 3 | 2013–2020 | Single EU-wide cap; auctioning the default for power |
| 4 | 2021–2030 | Faster cap decline; Market Stability Reserve strengthened |
Worked example: why trading lowers cost
Companies A and B each emit 100,000 t and receive 95,000 allowances → each is 5,000 t short. Allowances cost EUR 10; fine EUR 100.
| Abatement cost | Choice | Cost with trading | Cost without trading | |
|---|---|---|---|---|
| A | EUR 5/t | Cuts 10,000 t (50,000) and sells 5,000 spare (+50,000) | 0 | 5,000 × 5 = 25,000 |
| B | EUR 15/t | Buys 5,000 allowances | 50,000 | 5,000 × 15 = 75,000 |
Total cost falls from 100,000 to 50,000, and the cheapest cuts happen first.
System design
| Question | EU ETS answer |
|---|---|
| Which gases? | CO₂, N₂O, PFCs |
| Who? | Power, refineries, iron and steel, aluminium, cement, glass, ceramics, pulp and paper, acids, bulk chemicals; aviation from 2012 |
| Evidence? | EUA = 1 t CO₂e; aviation gets EUAAs (airlines can use EUAs, but others can’t use EUAAs) |
| Distribution? | Phase 1 almost all free → growing auctions on EEX and ICE (weekly); free allocation by end-February |
| Ownership? | Union Registry accounts; transfers checked by the EU Transaction Log (EUTL) |
| Cap? | EU-wide since phase 2; −2.2% a year from 2021 |
| Compliance? | Verify emissions by 31 March, surrender by 30 April; shortfall fine EUR 100/t plus make-good |
| Unused allowances? | Banking allowed indefinitely; early free allocation allowed de facto borrowing from the current year |
| Linking? | Limited CDM/JI credits allowed in earlier phases; none in phase 4 |
Primary market = auctions; secondary market = trading among participants. A New Entrants Reserve funds low-carbon projects.
Fixing the surplus
After 2008, falling output left a 2.1 billion surplus by 2013 and very low prices. Fixes:
- Back-loading: delay auctioning 900 million allowances to 2019/20 (timing only).
- Market Stability Reserve (MSR): if allowances in circulation exceed 833 million, a share is withheld from auctions; below 400 million, 100 million are released.
MSR example (12% intake rate, as in the source): 1,400 million in circulation → withhold 0.12 × 1,400 = 168 million.
- The MSR intake rate was doubled to 24% from 2019 and kept there to 2030; since 2023, allowances held in the MSR above 400 million are cancelled.
- The 2023 reform raised the cap’s annual decline to 4.3% (2024–27) and 4.4% (2028–30), targeting −62% versus 2005 by 2030.
- Maritime shipping joined in 2024; a separate ETS2 for buildings and road fuel starts later in the decade.
- The Carbon Border Adjustment Mechanism (CBAM) prices the carbon in imports of steel, cement, aluminium, fertiliser, power, and hydrogen.
- The surrender deadline moved to 30 September from the 2024 compliance cycle, and the EUR 100 penalty is inflation-indexed.
- EUA prices went from about EUR 5 (2013–17) to about EUR 100 (early 2023).
Cap and trade vs carbon tax
| Cap and trade | Carbon tax | |
|---|---|---|
| Certainty of | Quantity of emissions | Price of emissions |
| Price volatility | High if the cap is misjudged | None |
| Recession | Prices fall (less output) unless the cap adjusts | Unchanged |
| Free allocation | Windfall gains possible | n/a |
| Innovation | A breakthrough lowers permit prices, cutting the innovator’s return | Clear floor, stable return |
Checkpoint 3: EU ETS
Checkpoint — section 11.5
1 / 5Allowances trade at 30. A firm can abate at 22/t. What should it do?
Pick an answer to see if you are right.
11.6 Emission derivatives
World Bank (2019): 57 carbon pricing initiatives covering 11 Gt CO₂e (~20% of emissions); prices from USD 1 to 127/t, with 51% of covered emissions below USD 10; revenue about USD 44 billion.
By 2024–25 around 75–80 carbon taxes and ETSs covered roughly a quarter of global emissions, and revenue exceeded USD 100 billion a year. China’s national ETS (launched 2021) is the largest by emissions covered.
A buyer can buy spot and hold, buy forward, or use options.
Spot and futures
- Spot: delivery T+1, OTC or on ICE/EEX; lots of 10,000 to 100,000 EUAs.
- ICE EUA futures: 1 lot = 1,000 EUAs, priced in EUR; monthly, quarterly, and annual (December is the benchmark); expire on the last Monday of the month (earlier in December for holidays); physical delivery via the Union Registry.
OTC forwards
- Fixed-price, physically settled forward
- Forward where the seller may deliver early (and receives a discounted amount)
- Floating-price forward: date fixed, price later set from an average (e.g. of futures)
- Floating forward where the buyer chooses the volume before pricing
Fair value
EUAs are fully fungible and can be banked, so cost of carry applies: forward = spot × (1 + r × days/360).
March 2020: spot 18.40, 12-month rate −0.25%:
- Fair forward = 18.40 × (1 − 0.0025 × 365/360) ≈ 18.40 − 0.05 = 18.35
- Market forward = 18.15 — about 0.20 below fair value
A firm holding free allowances it doesn’t need for a year can sell spot at 18.40 and buy forward at 18.15. That is like borrowing at:
(18.15 ÷ 18.40 − 1) × 360/365 ≈ −1.34% — being paid to borrow.
Repo / inventory monetisation
Sell EUAs spot and buy them back forward in one package — effectively a secured loan.
Example: 1 million EUAs; 12-month forward EUR 15.00 → repay 15,000,000. The bank charges Euribor 5.00% + 0.50%:
Proceeds = 15,000,000 ÷ (1 + 0.055 × 365/360) ≈ EUR 14,207,722
The corporate borrows at Euribor + 0.50% — attractive if it usually pays more. The bank uses the forward (often more liquid than spot) to set value, and profits if it funds itself below Euribor + 0.50%.
The source prints 14,207,772; the correct figure is 14,207,722.
Swaps
Cash-settled, fixed vs floating, priced on a single day (like a one-day contract for difference). Example: Company A pays fixed EUR 14.00 on 25,000 allowances vs an index on the pricing date. Index at 16.00 → A receives (16 − 14) × 25,000 = EUR 50,000. Useful for price protection without delivery, or for taking a view.
Options
A 9-month European call on the December ICE EUA future: 30,000 allowances, strike EUR 20 (at-the-money forward), premium EUR 2.39 (35% vol) → total EUR 71,700.
- Cash-settled payoff = max(F − K, 0) × quantity; physically settled → allowances delivered to the registry.
- At expiry F = 23.50 → payoff 3.50 × 30,000 = 105,000; net of premium 33,300.
- EUA implied vols have exceeded 60% at times.
View-driven strategies
| Strategy | Trade |
|---|---|
| Cash-and-carry | Forward rich vs fair value → buy spot, finance, sell forward |
| Calendar spread | Buy near / sell far (or the reverse) on curve shape |
| Geographic | Same contract priced differently on ICE vs EEX |
| EUA / EUAA spread | EUAAs normally trade at a discount; if the discount looks too wide → buy EUAA, sell EUA |
| Volatility | Straddles, strangles (lesson 3) |
Checkpoint 4: emission derivatives
Checkpoint — section 11.6
1 / 5Spot EUA 80, 1-year rate 3% (act/360, 365 days). Fair forward?
Pick an answer to see if you are right.
11.7 Weather derivatives
“You can’t make people think it’s going to be 110 degrees in London next week.” — a weather trader (FT, 2007): weather can’t be manipulated, and it is largely uncorrelated with other assets.
Weather risk = the effect of weather on cash flow. The first deal (Enron–Koch, 1997) paid USD 10,000 per degree away from a set level.
| Insurance | Weather derivative | |
|---|---|---|
| Covers | Rare, catastrophic events (hurricanes) | Frequent, moderate events (mild winter, wet summer) |
| Payout | Must prove a loss | Paid on an independent measured index |
Users: agriculture, aviation, construction, energy, film production, retail, tourism, wholesale.
Contract ingredients: an index (average temperature, HDD, CDD, rain, snow, wind chill), a period, a weather station (e.g. Heathrow), a tick value per index point, and a strike (often from history).
HDD and CDD
Base 65 °F (18 °C outside the US). Daily average temperature T:
- HDD = max(0, 65 − T) — rises when it’s cold (heating needed)
- CDD = max(0, T − 65) — rises when it’s hot (cooling needed)
- Monthly index = sum of the daily values; CAT = sum of daily average temperatures
Example: a week at 60, 55, 58, 61, 59, 56, 62 °F → HDD = 5 + 10 + 7 + 4 + 6 + 9 + 3 = 44; CDD = 0.
CME futures
| Term | Value |
|---|---|
| Contract | Monthly CDD (or HDD), e.g. Dallas |
| Size | USD 20 × index |
| Tick | 1 point |
| Settlement | Cash |
August Dallas CDD quoted at 691 → contract value 691 × 20 = USD 13,820. A 100 °F day adds 35 points; a 60 °F day adds 0. Months-ahead pricing relies on models such as GFS (US) and ECMWF (Europe) plus climatology.
OTC features
Wider underlyings (rain, snow, wind, sunshine); can reduce basis risk between station and site; monthly, seasonal, or annual periods (HDD season Oct–Apr, CDD May–Sep); data must be cleaned; tick values negotiated (e.g. 1,000, 500, 200); payouts usually capped, as temperature has no natural limit.
Swap
Dallas August CDD, fixed 691, USD 500/point. Realised CDD = 700 → the floating payer pays (700 − 691) × 500 = USD 4,500.
The source says the floating leg is the “average” of daily CDDs. A monthly CDD index is the sum of daily values (which is why it reaches ~700); an average would be about 22.
Options
- Call payoff = tick × max(index − K, 0); put = tick × max(K − index, 0)
- Binary options pay a fixed amount if in the money
Application: cattle farmer
Losses from hot springs (slow growth) and cold autumns (young stock suffer):
- Buy a CDD call for spring
- Buy an HDD call for autumn — HDD rises as it gets colder, so a call (not a put) protects against cold
Using the 44-HDD week with strike 0 and USD 200/point → payout USD 8,800. A 91-day autumn with losses below 40 °F → HDD strike = (65 − 40) × 91 = 2,275.
Application: power utility
A Midwest utility loses USD 50,000 per °F that the summer average falls below normal (less air conditioning).
| Structure | Effect |
|---|---|
| Buy a put on average temperature | Paid when it’s cool; OTM strike lowers the premium |
| Zero-cost collar | Buy OTM put + sell OTM call: protection funded by giving up gains when hot |
| Swap | Locks in the temperature exposure both ways |
Payouts are usually capped (e.g. USD 300,000). Example: put strike 75 °F, realised 72 °F, USD 50,000/°F → 3 × 50,000 = USD 150,000 (under the cap).
Rank: climate and weather practice
Climate and weather practice
1 / 3Rank EU ETS phases in order.
OrderDrag the handle to rearrange — top should be first.
- 1Faster cap decline to 2030
- 2Trial with estimated national caps
- 3Kyoto period, aviation joins
- 4Single EU-wide cap, auction-led
Drag items into the correct order (or use ↑↓), then check.
Check: climate and weather
Climate and weather
1 / 6One EUA gives the right to emit…
Pick an answer to see if you are right.
Calculate: numeric practice
Type the number (no units; minus sign for payments).
Formula sheet
| Item | Formula |
|---|---|
| CO₂e | mass × GWP |
| Fuel-switch carbon price | (gas cost/MWh − coal cost/MWh) ÷ (EF_coal − EF_gas) |
| Fuel cost per MWh | fuel price ÷ efficiency |
| Compliance cost | min(abatement cost, allowance price) × shortfall |
| MSR intake | rate × allowances in circulation (if above 833 m) |
| Forward fair value | spot × (1 + r × days/360) |
| Implied rate | (F/S − 1) × 360/days |
| Repo proceeds | forward amount ÷ (1 + r × days/360) |
| Swap (fixed payer) | (index − fixed) × quantity |
| Call | max(F − K, 0) × quantity |
| HDD / CDD | Σ max(0, 65 − T) / Σ max(0, T − 65) |
| Weather payoff | tick × index difference |
Calculate: climate and weather
1 / 1710 t of methane (GWP 25). CO₂e in tonnes?
Enter a number (at most 0 decimal places).
Randomized practice
Practice: carbon and weather drills
1 / 6EUA call on 30000 allowances, strike 61, premium 5.87. Futures at expiry 53.12. Profit after premium (negative = loss)? Round to 2 decimals.
Enter a number (at most 2 decimal places).
Takeaways
- Weather = short-term local; climate = long-run patterns; CO₂e = mass × GWP
- Policy path: IPCC → UNFCCC → Kyoto (ETS, CDM/CERs, JI/ERUs) → Paris (NDCs, Article 6)
- Cap and trade fixes quantity; a tax fixes price; firms abate when abatement is cheaper than allowances
- Carbon rises with gas and power prices, falls with coal prices; fuel-switch price = cost gap ÷ emission-factor gap
- EU ETS: EUAs/EUAAs, auctions plus free allocation, registry/EUTL, banking, MSR, fine plus make-good
- Derivatives: spot, futures (1,000 EUAs, physical), forwards, cost-of-carry fair value, repos, one-day swaps, options
- Weather: HDD = cold, CDD = heat; tick × index; calls/puts/swaps/collars; settled on measured data, not proof of loss
Further Reading
- Next: 12. Agriculture
- 10. Bulk commodities
- 8. Electricity — clean spark and dark spreads, HDD/CDD
- 2. Derivative valuation — cost of carry
- 3. Risk management principles — volatility strategies
- IPCC Assessment Reports; European Commission EU ETS pages; World Bank State and Trends of Carbon Pricing; CME weather contract specifications
- Background: N. Schofield, Commodity Derivatives: Markets and Applications (Wiley) — used as a study outline