AACE International's Recommended Practice 40R-08 is the starting point for most serious discussions about contingency. It doesn't give you a formula. It sets out what any good contingency estimate should do, whatever method you use. This guide explains what the RP is, what it counts as contingency, its principles in plain language, and what it means for a project team that has to put a number in the budget.
What 40R-08 is
| RP 40R-08 | |
|---|---|
| Full title | Contingency Estimating – General Principles |
| Published by | AACE International, first issued in 2008 |
| Part of | The risk management section (7.6) of AACE's Total Cost Management framework |
| Covers | Estimating contingency, reserves and similar risk funds, plus time allowances in schedules |
| Does not cover | Specific calculation methods; those are in separate RPs |
| Status | A guideline of good practice, not a standard |
Think of it as the rulebook for the rulebooks. AACE has a whole family of RPs on contingency methods. 40R-08 sits above them: it groups the methods into types and sets the principles each one should be tested against. You can buy or download it from AACE's recommended practices page; members get it free.
What counts as contingency
AACE's terminology RP, 10S-90, defines contingency as an amount added to an estimate for items, conditions or events that are uncertain, and that experience says will add cost overall. In other words, contingency is expected to be spent. It is not a spare pot.
Just as important is what it leaves out. Under the 10S-90 definition, contingency does not cover:
- Major scope changes, such as a different capacity, size or location.
- Extraordinary events, such as major strikes and natural disasters.
- Management reserve.
- Escalation and currency effects.
40R-08 itself takes a slightly wider view. It talks about estimating risk impact in general, which can include reserves or insurance for catastrophic risks and ranges for a business case. But it also leaves out escalation, currency and other mainly financial risks. AACE handles escalation in its own RPs, such as 68R-11.
The principles in plain language
The core of 40R-08 is a short list of principles. Any method you develop or choose should meet them. In our words:
| Principle | What it means for you |
|---|---|
| Meet the client's objectives | Agree first what "risk" and "contingency" mean here, what output is needed, and how much risk the owner accepts. Should the budget be the most likely outcome, or something safer? |
| Serve a risk management process | The number isn't the goal. It should feed decisions and risk management. If no such process exists, set one up. |
| Fit for use | Match the method to the project type, size and complexity, the kind of risk, the project phase and estimate class, how the base estimate was made, and the skills of the people involved. |
| Start with the risk drivers | Don't put a range on a cost line until you know what drives that range. |
| Link drivers to outcomes | The method should show clearly how each risk turns into cost and time. |
| Avoid self-inflicted risk | The method itself shouldn't create errors, for example through a model that hides or double counts risks. |
| Use empirical data | Ground the results in what actually happened on past projects. |
| Use experience and competence | The people running the analysis need the knowledge to do it well. |
| Give probabilistic results | Show the outcome as a range with probabilities, in a form that helps decision makers. |
AACE's own guide to quantitative risk analysis notes that no single method meets all of these. That's why there are several method RPs, and why hybrids exist.
Two kinds of uncertainty
40R-08 asks you to look at both inherent estimate uncertainty and risk events. AACE's later guidance builds this out into two groups:
- Systemic risks come from the project system itself: how well the scope is defined, how new the technology is, how complex the project is. They are always present, rarely in a risk register, and best quantified from historical data.
- Project-specific risks are events and conditions particular to this project, such as a late permit or poor ground. They fill the risk register and may or may not happen.
A risk register alone tends to miss the first group. A pure percentage tends to miss the second. That is the practical reason behind the "fit for use" principle.
Where the method RPs fit
40R-08 points to separate RPs for the actual methods. The main ones, as AACE describes them:
| RP | Method | Best suited to |
|---|---|---|
| 41R-08 | Understanding estimate ranging | Explains range estimating and its limits; not a method on its own |
| 42R-08 | Parametric estimating | Systemic risks, using regression on historical data |
| 44R-08 | Expected value | Cost of the critical project-specific risks, with Monte Carlo |
| 57R-09 | Risk drivers on a CPM schedule | Integrated cost and schedule, when a good schedule model exists |
| 65R-11 | Expected value, cost and schedule | Extends 44R-08 to duration without needing a CPM model |
| 113R-21, 117R-21 | Hybrids | Parametric for systemic risk plus expected value or CPM for project risks |
| 118R-21, 123R-22 | Estimate ranging with Monte Carlo | Well-defined, low-complexity projects where systemic risk is small |
The common thread is that risks drive the result, and the method is chosen to fit the project and its stage.
What it means for your project team
You don't need to apply every AACE method to follow 40R-08. You do need to be able to answer a few questions:
- What is in, what is out? Write down that contingency covers uncertainty within the agreed scope, and that scope changes, escalation, currency and management reserve are handled elsewhere. See contingency vs management reserve.
- Which confidence level? Agree with the owner whether the budget is set at P50, P80 or another level. P50, P80 and P90 explained covers the choice.
- What drives the number? Each part of the contingency should trace back to named risks or to the level of scope definition, not to a habit.
- Does the method fit the stage? An early-stage estimate needs more weight on systemic risk; a late-stage estimate can lean more on the register.
- Can you check it against history? Compare the result with what similar projects actually needed.
Take the fictional Harbor Crossing Rail Tunnel: a base estimate of EUR 340.0M and a recommended reserve of EUR 42.2M, or 12.4%. On its own, "12.4%" says little. In the spirit of 40R-08, the useful version reads: EUR 42.2M at an agreed confidence level, driven by these risks, excluding escalation and scope change, and in line with comparable tunnels. That is something a board can actually decide on. How to set a P80 contingency walks through the steps.
A note on framing: a contingency estimate built along these lines is decision support. It shows a range and its chances. It does not guarantee the project will stay within budget.
Do it in Excel
The Risk Contingency Allocation Tool estimates and justifies contingency with Monte Carlo simulation and three estimating methods, splits it into project contingency and management reserve, and produces a business case report for the board. If you want the cost estimate and the risk register in one simulation, use QRCIM, which runs both together so estimate uncertainty and risk events end up in one result. For the method itself, read how to calculate contingency with Monte Carlo.
References
- AACE International. Recommended Practice 40R-08: Contingency Estimating – General Principles (2008), sample pages.
- AACE International. Professional Guidance Document 02: Guide to Quantitative Risk Analysis (rev. 2024).
- AACE International. Recommended Practice 10S-90: Cost Engineering Terminology.