How to choose the window
- List the window options. The choice is normally the latest single month, a three-month average, or a twelve-month rolling average.
- Describe how the cost actually moves. Decide whether the underlying cost makes real step-changes, jitters as noise with no trend, or drifts slowly.
- Use a short window for fast-moving costs. A one or three-month window fits when the cost genuinely swings fast and both sides want the price to track it closely, accepting larger and more frequent adjustments.
- Use a long window for predictability. A twelve-month rolling average fits when both sides value a smooth, predictable price over precision and will absorb a slow average.
- Match the window to the input, not a default. A step-change input is poorly served by a twelve-month average that lags a year behind. A noisy input is poorly served by a one-month window that just amplifies the noise.
What an averaging window is
An index publishes a new value every month. A clause has to decide how many of those months to use before applying an adjustment: the latest single value, a 3-month average, a 12-month average, or something else. That choice is the averaging window.
Why it changes the number
A single-month value reacts immediately to whatever happened that month, including a one-off spike that reverses the next month. A longer average smooths that noise out, but reacts slower to a real, sustained cost change. Two contracts reading the same index with different windows will calculate two different adjustments on the same day.
A worked example
Using copper wire and cable (BLS WPU10260314) as an example, take three clauses that all compare December 2025 to December 2024 on a $300,000 scope. Each reads both ends through its own window.
| Window | Base | Current | Ratio | Adjustment |
|---|---|---|---|---|
| Latest month | 412.990 | 510.784 | 1.237 | $71,000 |
| 3-month average | 419.690 | 494.585 | 1.178 | $53,500 |
| 12-month average | 417.961 | 466.993 | 1.117 | $35,200 |
Same index, same two dates, three different bills. The index was rising through 2025, so the longer the window, the more of the rise it left out. In a falling market the order would flip.
Here is the one-month reading on the latest data.
Short window, fast reaction
A 1-month or 3-month window moves quickly. Use it when the underlying cost genuinely swings fast and both sides want the price to track reality closely, even if that means more frequent, larger adjustments.
Long window, smooth reaction
A 12-month rolling average moves slowly and rarely swings hard in either direction. Use it when both sides value predictability over precision, and would rather absorb a slow average than chase every short-term spike.
Match the window to the input
Match the window to how the underlying cost actually behaves, not to a default. An input that moves in real step-changes, like a new tariff or a supply shock, is poorly served by a 12-month average that takes a year to catch up. An input that jitters without a real trend is poorly served by a 1-month window that just amplifies the noise.