Contingency and management reserve both sit on top of the base estimate, and on many projects they end up as one pot. That's a mistake. They cover different things, belong to different people and should be released under different rules. This guide explains the difference and shows how to size and split them.
The short answer
| Contingency | Management reserve | |
|---|---|---|
| Covers | Identified risks and estimate uncertainty within the agreed scope | Events nobody has identified yet |
| Held by | The project manager | The sponsor or the board |
| Part of the cost baseline? | Yes | No, it sits above the baseline |
| How it's sized | From the risk register and the estimate ranges, usually by simulation | By policy, or as a top-up to a higher confidence level |
| Released | By the project manager, within agreed rules | Only by a decision of the sponsor or board |
In PMI terms, contingency deals with the "known unknowns": risks you've written down, even if you don't know whether they'll happen. Management reserve deals with the "unknown unknowns". Both are part of the total budget, but only contingency is part of the baseline the project is measured against.
Why one pot goes wrong
When contingency and management reserve sit together under the project manager, two things tend to happen.
First, the pot gets spent early. Early problems are real and visible, while the later ones are still on paper. By the time the large risks in the second half of the project arrive, there's little left. Then the project goes back to the board for more money, which is exactly what the reserve was meant to prevent.
Second, nobody can tell whether the project is on track. If the reserve absorbs every setback without anyone seeing it, the reports stay green until the money runs out. Separate pots, each with its own drawdown log, show where the money goes and how much is left for which purpose.
How to size each part
Contingency comes from the risks you know about. Put ranges on the cost items, add the risk events from the register, and run a Monte Carlo simulation. The contingency is the gap between the base estimate and the confidence level your organization budgets at, often P80. How to set a P80 contingency walks through the steps.
Management reserve is harder, because by definition you can't list what it covers. Organizations use one of two approaches:
- A policy amount, such as a fixed percentage of the base estimate, set by the board for each type of project.
- A top-up to a higher confidence level. The project budget is set at P80, and the management reserve covers the step from P80 to P90 or P95. That links the reserve to the same simulation instead of a rule of thumb.
A third approach is a fixed split of one total reserve, for example 70% contingency and 30% management reserve. It's simple and easy to explain, and it works well as long as everyone agrees which risks fall in which part.
A worked example
Take the fictional Harbor Crossing Rail Tunnel from our sample business case. The base estimate is EUR 340.0M. The analysis recommends a total reserve of EUR 42.2M, 12.4% of the base, and splits it 70/30:
| Part | Amount | Held by | For |
|---|---|---|---|
| Contingency | EUR 29.6M | Project manager | Risks in the register |
| Management reserve | EUR 12.7M | Sponsor / steering committee | Risks nobody has identified yet |
| Total reserve | EUR 42.2M |
The project manager can now act on a known risk, say a delay in the tunnel boring machine delivery, without asking the board each time. The board keeps EUR 12.7M for what nobody saw coming, and sees it requested as soon as something like that happens.
Rules for releasing the money
Splitting the reserve only helps if the release rules are clear. Write down, before the project starts:
- What contingency may pay for: only risks and uncertainty within the agreed scope. A scope change is not a risk; it needs its own decision and its own budget.
- Who approves what: for example, the project manager up to a set amount per risk, above that together with the sponsor.
- When management reserve moves: only on a sponsor or board decision, with a short note on what happened and why the contingency doesn't cover it.
- What happens to unused contingency: when a risk passes without occurring, its share is released back, not quietly reused for something else.
Common mistakes
- Counting risks twice. Allowances already in the base estimate, such as a growth allowance on quantities, shouldn't come back as a risk in the register.
- Paying scope changes from contingency. It looks convenient, but it empties the pot meant for the risks, and the change is never decided on properly.
- Reporting unused contingency as savings halfway through the project, while the large risks are still ahead.
- No management reserve at all, so every unforeseen event becomes a request for more budget.
Do it in Excel
The Risk Contingency Allocation Tool works out the reserve in three ways, combines them into one recommendation and splits it into project contingency and management reserve, with a business case for the board that includes the drawdown rules. To see how the confidence level drives the amount, read P50, P80 and P90 explained, or run the simulator on our homepage.