How to build a copper budget with best, base and worst-case price scenarios
A single copper price in the budget is almost certainly wrong by the end of the year. The useful question is how wrong it could be, and whether the business can absorb that. Copper set record highs in September 2026 and sits around US$14,350/t at the end of October (Business Recorder/Reuters), while serious forecasters disagree on 2027 by more than US$2,000/t. That is exactly the situation in which a three-scenario budget earns its keep.
The method in brief: use the forward price for the budget period as the base case, set best and worst cases from copper's volatility over the budget horizon (for example the 10th and 90th percentiles of the expected annual average price), apply the scenarios only to volume that is not yet priced, and show currency as a separate sensitivity. With annualised volatility around 24%, a 2027 average-price budget set from late October 2026 has a P10-P90 range of roughly 20% below to 22% above the base price. In the worked example below, that range is worth about EUR 275,000 on 84 t of copper, even after 40% of the volume has been fixed.
MetalAlert estimates the probability of specified LME price changes over defined horizons. Scenario budgets are one of the most natural places to use that kind of information, because a budget is a statement about a range of outcomes, and the weights you put on that range should change when the market does.
Start with the exposure, not the price
A scenario budget is only as good as the tonnage behind it. You need annual net copper tonnes, split into what is already priced (fixed-price contracts, existing hedges) and what is open, with the months in which the open volume will be priced. If you do not have that yet, work through how to calculate copper exposure from your bill of materials first. This article uses the same illustrative company: about 84 t of net copper exposure a year.
Choosing the base-case price
There are three common candidates, and they answer different questions.
The forward curve (recommended for the base case)
The LME forward price for each month of the budget year is the price you could lock in today. Using it as the base has two advantages. It is market-consistent, so it does not embed your own view on direction. And it is hedgeable: if management dislikes the base case, they can fix it. Ask your broker or bank for the monthly strip covering your budget months and weight it by your monthly priced tonnes.
Note the shape of the curve. In late October 2026 the front of the LME curve was in backwardation, with cash trading about US$200/t above three-month, the widest since late August (copper.com.au, 29 October 2026). In a backwardated market, prices for later months are lower than spot, so a budget set at spot will look conservative against the forward curve.
Analyst consensus (a cross-check, not the base)
Published forecasts are useful for understanding the argument, but the spread is wide. In October 2026, Goldman Sachs forecast a 2027 LME average of US$15,400/t, while Macquarie, BMO and BMI were around US$13,500-13,600/t for the 2026 average and the World Bank's 2027 forecast was reported at US$13,000/t (copper.com.au summary). The International Copper Study Group expects the refined surplus to widen to about 368,000 t in 2027 (Reuters via Capital.com). Picking one bank's number as the base imports that bank's view. If you do use consensus, show it alongside the forward price and explain the difference. Our copper price forecast for the next 6-12 months sets out the competing views.
Current spot price
Simple and easy to explain, but it ignores both the curve and the fact that the budget year has not started. It is acceptable for a small business with short pricing periods; for anyone with material exposure, the forward curve is better.
Setting best and worst cases from volatility
The best and worst cases should describe plausible outcomes, not extremes or wishes. Volatility gives a disciplined way to do that.
What volatility to use
NYU Stern's V-Lab, which fits a GARCH model to the S&P GSCI Copper Spot Index, put copper's annualised volatility at about 24.3% at the end of September 2026 (NYU V-Lab). Implied volatility from LME or Comex options, available from your broker, is an alternative that reflects what the options market expects. The rest of this article uses 24% as a round figure; check the current level before you use it, because volatility itself changes.
Converting volatility to a horizon
Volatility scales with the square root of time. A one-standard-deviation move over horizon t (in years) is roughly σ x √t. Around a base price of US$14,300/t, and treating price changes as lognormal:
| Horizon | 1-sigma move | Range (about 68% of outcomes) |
|---|---|---|
| 3 months | 12% | US$12,680 - 16,120/t |
| 6 months | 17% | US$12,070 - 16,950/t |
| 12 months | 24% | US$11,250 - 18,180/t |
These are statistical ranges, not forecasts. Roughly one outcome in three falls outside them, and copper has fatter tails than the lognormal model assumes: tariff decisions, mine accidents and demand shocks produce jumps.
Why the annual average has a narrower range
Most manufacturers do not buy all their copper on one day. If you buy evenly through 2027, your cost depends on the average price across the year, and an average moves less than the price at any single date. A simulation of 2027 monthly prices from a late-October 2026 starting point, using 24% volatility and no drift, gives these percentiles for the 2027 average (our calculation, rounded):
| Percentile of 2027 average | LME price (US$/t) | vs base |
|---|---|---|
| P10 (best case for a buyer) | 11,400 | -20% |
| P50 (median) | 14,100 | -1% |
| P90 (worst case for a buyer) | 17,500 | +22% |
The simulated mean is about US$14,300/t, equal to the starting forward price; the median sits slightly lower because of the skew, which is why we keep the forward price as the base case. The standard deviation of the annual average comes out at about 17%, compared with 24% for a 12-month point price. Notice the asymmetry: the worst case sits further from the base than the best case, because prices cannot fall below zero but can rise a long way. A symmetric "plus or minus 15%" budget understates the upside risk for a buyer.
Should you add a stress case?
Many finance teams add a fourth case beyond P90 to test liquidity and covenant headroom, for example the 12-month two-sigma level (about US$23,000/t using the figures above). It should not drive the operating budget, but it answers the question "what would break us?".
Weighting the scenarios
A three-point budget invites the question "which one do we plan for?". A probability-weighted figure gives a single number that still respects the range. A rule of thumb from the oil and gas industry for P10/P50/P90 estimates is Swanson's 30-40-30 rule: weight P10 at 0.3, P50 at 0.4 and P90 at 0.3 to approximate the mean of a moderately skewed distribution (John D. Cook on Swanson's rule, originally Hurst, Brown and Swanson in the AAPG Bulletin, 2000).
For the percentiles above, the weighted average is very close to the simulated mean of about US$14,300/t. That is no coincidence: with no drift, the expected average equals the forward price. The point of weighting is not to find a better forecast but to make sure the budget figure reflects the skew rather than the most optimistic reading of it.
If you have your own probability estimates (for example, an estimate that copper is more likely than usual to rise by a given amount over the next six months), you can adjust the weights. Be explicit when you do so, and record why. Changing weights quietly is how a budget turns into a forecast nobody can audit.
Including hedged volumes and fixed-price contracts
Only open volume moves with the scenarios. Anything already priced, through supplier fixed-price contracts or financial hedges, is budgeted at its fixed price in every scenario. This is where hedging shows its value in a budget: it narrows the range. For setting the share to fix, see how much of our annual copper volume to hedge. If you use derivatives and want hedge accounting under IFRS 9, involve finance and your auditors early; the documentation requirements affect how hedges appear in the budget and the accounts.
Including currency
Copper is priced in US dollars. For a budget in EUR or SEK, the exchange rate is a second, separate exposure. Keep it as a separate sensitivity line rather than folding it into the copper scenarios: the two are not reliably correlated, and combining them into a single "worst case" overstates the joint probability. Approaches to the currency side are covered in how to hedge copper bought in USD with a EUR or SEK budget.
Worked example: 2027 copper budget
Illustrative figures only. The company, volumes and hedge position are invented; the base price and volatility are round figures close to late-October 2026 market levels.
Assumptions
- Net copper exposure 2027: 84 t (from the exposure calculation example).
- Already priced: 40%, or 33.6 t, at an average of US$14,300/t (fixed-price supplier contracts, including 7.4 t under an existing motor contract, plus forward hedges).
- Open: 60%, or 50.4 t, priced evenly through the year.
- Base price: US$14,300/t (illustrative 2027 forward average). Best and worst: P10 and P90 of the simulated 2027 average, rounded.
- Budget exchange rate: EUR/USD 1.12.
- Copper metal element only. Cathode premiums (Aurubis has set US$347/t for European deliveries in 2027, per Reuters via Kitco) and fabrication charges are separate budget lines.
Copper cost by scenario
| Scenario | 2027 avg LME (US$/t) | Priced 33.6 t (US$) | Open 50.4 t (US$) | Total (US$) | Total (EUR) | vs base (EUR) | If nothing were priced (EUR) |
|---|---|---|---|---|---|---|---|
| Best (P10) | 11,400 | 480,480 | 574,560 | 1,055,040 | 942,000 | -130,500 | 855,000 |
| Base (forward) | 14,300 | 480,480 | 720,720 | 1,201,200 | 1,072,500 | 0 | 1,072,500 |
| Worst (P90) | 17,500 | 480,480 | 882,000 | 1,362,480 | 1,216,500 | +144,000 | 1,312,500 |
| Weighted 30/40/30 (of rows above) | 1,205,736 | 1,076,550 | +4,050 |
The P10-P90 range is EUR 274,500 with 40% priced, against EUR 457,500 if nothing were priced. The effective average copper price ranges from US$12,560/t in the best case to US$16,220/t in the worst.
Currency sensitivity (base copper price)
| EUR/USD | Copper cost (EUR) | vs budget rate |
|---|---|---|
| 1.07 | 1,122,600 | +50,100 |
| 1.12 (budget) | 1,072,500 | 0 |
| 1.17 | 1,026,700 | -45,800 |
Presenting the copper budget to management
Management needs four things, on one page:
- The budget figure and how it was set. "Base case at the forward price; weighted figure EUR 1.08m."
- The range and what drives it. "P10-P90 of EUR 0.94m to EUR 1.22m, driven by 50 t of unpriced copper."
- The sensitivity. "Each US$1,000/t on the open volume is about EUR 45,000; each 5 cents on EUR/USD is about EUR 46,000-50,000."
- The decision on the table. Whether to fix more volume, accept the range, or pass more of it to customers. For the margin consequences, see how a US$1,000/t copper move hits gross margin, and for the full format, how to draft a CFO briefing on copper risk.
Avoid presenting the best case as a target. It is a 1-in-10 outcome for a buyer.
When to re-forecast
Set rules in advance rather than reacting to headlines:
- On a fixed calendar, typically quarterly, aligned with the financial forecast.
- When the price leaves the band. A monthly one-sigma move at 24% volatility is about 7% (24% / √12). A move beyond that, or a monthly average outside the P10-P90 range for the remaining period, is a reasonable trigger.
- When volatility changes materially, for example a sustained rise from 24% to 30% widens the bands by a quarter.
- When exposure changes: a large order, a new product, a contract renewal or a new hedge.
- On known binary events, such as a decision on US refined copper tariffs.
Each re-forecast should roll the base to the current forward curve for the remaining months, keep priced volume at its fixed price, and recalculate the range for the shorter remaining horizon. The range narrows as the year progresses, which is itself useful information for management.
FAQ
What copper price should we use for the 2027 budget?
Use the LME forward prices for the months in which your 2027 volume will be priced, weighted by tonnage. Show analyst forecasts as context, and set best and worst cases from volatility rather than from the highest and lowest forecasts you can find.
How wide should the best and worst-case range be?
For a calendar-year average set two months before the year starts, P10-P90 at 24% volatility is roughly 20% below to 22% above the base. For shorter horizons or short pricing periods, use σ x √t for the relevant horizon.
Should the budget assume the price we expect or the price we can lock in?
For the base case, the price you can lock in. It is the only number that management can turn into a certainty if they choose to.
Sources
- NYU Stern V-Lab: S&P GSCI Copper Spot Index GARCH volatility analysis (accessed October 2026)
- Business Recorder/Reuters: Copper slips on concerns over AI demand, high US inventories (29 October 2026)
- copper.com.au: Copper weekly brief, 29 October 2026
- Reuters via Capital.com: ICSG sees global refined copper surplus widening in 2027 (14 October 2026)
- Reuters via Kitco: Aurubis raises 2027 European copper premium to record US$347 (14 October 2026)
- John D. Cook: Swanson's rule of thumb
Where MetalAlert fits
A scenario budget sets the range; purchasing decisions during the year are about where within that range you are likely to end up. If you want to see how probability estimates for specified LME copper price moves over defined horizons could inform your scenario weights, re-forecast triggers and timing of open volume, MetalAlert is built for exactly that question. The estimates describe likelihoods, not certainties, and they work alongside your hedging policy rather than replacing it.