Contingency Estimating — General Principles

What contingency is

Contingency is the bridge between an estimate (which is necessarily a point in time, based on incomplete information) and reality (which always brings surprises). Because every estimate carries uncertainty — especially early ones, as you saw with the estimate classes in Cost Estimating — a responsible budget includes a buffer sized to that uncertainty. Contingency is that buffer, and risk analysis is how you size it.

What contingency does — and doesn't — cover

The single biggest source of contingency disputes is confusion about what it's allowed to pay for:

• ✓ Contingency covers — Estimating inaccuracy and incomplete design; Minor scope growth within the defined scope; Identified risks that may materialize; Normal variability in quantities and rates; The "known unknowns" of the project
• ✗ Contingency is NOT for — Major scope changes (use a change order); Escalation (estimated separately — Lesson 58R-10); Management reserve / true unknowns; Covering up a poor base estimate; A general "padding" to feel safe

Ways to estimate contingency

RP 40R-08 lays out a spectrum of methods, from crude to rigorous. The right one depends on the project's size, phase, and the maturity of your risk practice (Lesson 122R-22):

• Predetermined percentage — A flat % of the base (e.g., 10%), often by estimate class. Simple, fast — but blind to the specific project's risks. The Level-1 approach.
• Expected value — Sum each identified risk's probability × impact across the risk register. Ties contingency directly to the risks — the subject of the next lesson.
• Parametric / regression — A model relating contingency to project drivers (complexity, definition level), built from historical data. Covered in Module 5C.
• Monte Carlo simulation — Model uncertainty with distributions and simulate thousands of outcomes to read contingency at a chosen confidence level. The rigorous end — Module 5C.

Using contingency well

Size contingency from the project's actual risks using a method matched to its scale and phase. State clearly what it covers and at what confidence level. Then manage it as the project runs: draw it down deliberately as risks pass or materialize, track what's left against the risks still open, and report it transparently. Contingency that's never reviewed becomes either a hidden overrun or a slush fund — both failures of control.

Ten things to remember

1. Contingency is funds (and time) for uncertain-but-expected events.
2. It's the bridge from an uncertain estimate to reality.
3. It funds known unknowns within scope — and is expected to be spent.
4. Not for scope changes, escalation, or true unknowns (those are separate).
5. Contingency is not padding — calculated from real risks, not guessed.
6. Methods range from flat % to expected value to Monte Carlo.
7. Match the method to the project — more rigor isn't always better.
8. State what it covers and at what confidence level.
9. Manage it through the project — draw down deliberately, report transparently.
10. Contingency is where risk analysis becomes a number.