Cost risk analysis and contingency

Contingency is the money a project holds for the uncertainty in its estimate and the risks it carries. Setting it as a flat percentage is common and indefensible; setting it from a cost risk analysis ties it to the actual uncertainty, gives it a confidence level, and lets it be managed down as risks retire. Here is how cost risk analysis works, and why correlation and the schedule matter so much.

What contingency is for

Contingency covers the costs that will almost certainly be spent but cannot yet be identified line by line: the spread around the estimate and the risks in the register, within the project's defined scope. It is not for scope changes, which are funded through change control, and not a cushion for an estimate known to be low. Management reserve, held above the project for unknown unknowns, is separate.

How a cost risk analysis works

1. Give the uncertain cost elements ranges: rates, quantities, productivity, prices, often by cost type or account.
2. Include the register's risks with their cost impacts.
3. Link cost to the schedule: labour and equipment that cost per hour or per day cost more when the work takes longer. Time-dependent cost is often the largest single source of overrun.
4. Simulate thousands of iterations; each gives a total cost, and together they give a distribution with its P50 and P80.

Contingency at a confidence level is then the difference between that percentile and the base estimate: P80 minus base is the contingency for an 80% chance of not overrunning.

Correlation: the input most analyses get wrong

If every cost line varies independently, highs and lows cancel and the total looks tightly bounded. Real projects do not behave that way: one labour market sets every trade's rates, one productivity problem slows every crew, one escalation index moves every material. Modelling those shared factors, one draw for all the lines they affect, widens the cost range to what experience shows, and a P80 that ignores them understates the contingency needed.

Joint confidence of cost and schedule

A project can meet its cost and miss its date, or the reverse. Joint confidence is the share of iterations that meet both a date and a cost; it is always lower than either alone. Plotting each iteration's finish against its cost shows the trade-off and is the clearest picture of what a commitment really means.

Managing contingency down

• Allocate contingency to the risks and accounts it is held for, so it can be traced.
• When a risk is retired, release its share; when one occurs, draw it. Keep a ledger, approved like any budget change.
• Re-run the analysis at each major update: the remaining contingency should still cover the remaining risk at the chosen confidence.

In DeltaSys

The DeltaSys Estimator sets contingency by risk analysis on the estimate, at a confidence level, and allocates it to the risks by their expected value. In the Planner, cost uncertainty is on by default for cost-loaded projects: rates and productivity vary with recorded basis, correlated cost factors move whole trades or the whole project together, labour and equipment costs follow the simulated durations, and the results include the cost distribution, joint confidence and probabilistic cash flow. The contingency ledger tracks every draw and release against the register.

Questions

Primavera and P6 are trademarks of Oracle Corporation; Microsoft Project is a trademark of Microsoft Corporation; PMBOK is a trademark of the Project Management Institute; AACE is a trademark of AACE International; Last Planner is a registered trademark of the Lean Construction Institute. DeltaSys is not affiliated with or endorsed by them. Standards are described in our own words; see the publications themselves for their text.