Construction contingency: how much is enough
Construction contingency is enough when it reflects the project's real uncertainty, not when it matches a generic percentage. The practical question is: what risks remain, who owns them, how likely are they, and what would they cost if they occurred?
Contingency Planning Takeaway: There is no universal contingency percentage that fits every construction project; the right amount depends on uncertainty, design maturity, risk allocation, and owner tolerance.
- A useful contingency is tied to identified risks and tracked separately from scope growth, allowances, and discretionary upgrades.
- Teams should document how contingency is set, who controls it, and when it is released or replenished.
Why a single percentage is a weak answer
Early estimates include more unknowns than late estimates. A concept budget may carry major uncertainty around scope, quantities, market conditions, site conditions, code upgrades, procurement, and design decisions. A bid-ready estimate should have fewer unknowns, but it may still need contingency for coordination risk, unforeseen conditions, owner decisions, or residual design development. The GAO Cost Estimating and Assessment Guide emphasizes reliable estimates, risk, uncertainty, documentation, and updates, which makes its cost estimating guidance useful beyond government projects.
A flat percentage can be a placeholder, but it should not be the final logic. Two projects with the same budget can need very different contingencies. A repetitive warehouse on a clean site with complete drawings has a different risk profile than a phased hospital renovation above occupied spaces. A small residential roof replacement has different uncertainty than a mixed-use adaptive reuse project with unknown utilities, structural repairs, and tenant changes.
Contingency also should not become a hidden upgrade fund. If an owner adds nicer finishes, expands the scope, or changes program requirements, that is scope growth. If a contractor missed a quantity that was clearly shown, that may be estimating error or risk depending on contract terms. If an unknown subsurface condition appears, it may be a contingency draw, owner risk, allowance, change order, or claim based on the contract.
Separate contingency from allowances and change orders
| Budget item | What it is for | How to manage it |
|---|---|---|
| Design contingency | Unresolved scope and detail development during design | Reduce as documents mature and decisions are finalized |
| Construction contingency | Known unknowns and project execution risk | Track by risk category, approval authority, and remaining balance |
| Allowance | A defined scope item with uncertain final cost | Reconcile when actual pricing is known |
| Owner reserve | Owner-controlled risk, business decisions, or broader program exposure | Keep separate from contractor-controlled funds |
| Change order | Approved modification to cost, time, or scope | Document through the contract process |
This separation prevents arguments. If contingency is buried inside a line item, teams may not know whether a draw is a true risk event or a scope change. If allowances are treated as contingency, owners may believe money is available for anything, while contractors may believe it is tied to a specific scope. Clear rules also improve decisions around specialized scopes such as steel fireproofing work because the parties can identify whether the event is part of the original risk plan.
AACE International's recommended practice on contingency estimating principles frames contingency and risk funds as part of risk management. The useful takeaway is that contingency should come from a structured view of uncertainty, not from habit alone. Teams can build that view with a risk register, estimate maturity review, sensitivity analysis, comparable project data, or probabilistic methods on larger projects.
How to decide what risks deserve funding
- List the remaining uncertainties by category: scope, design, site, utilities, permitting, procurement, weather, labor, phasing, owner decisions, and third-party approvals.
- Assign each risk an owner, probability range, cost impact range, schedule impact, and mitigation action.
- Identify risks already covered by contract terms, insurance, allowances, unit prices, alternates, or escalation clauses.
- Decide which risks need contingency, which need design resolution, and which need a separate management reserve.
- Create approval rules for drawdowns and update the remaining balance at each cost report.
This process can be simple on smaller projects. A residential owner may use a short list of likely unknowns such as deck repair, flashing damage, ventilation corrections, or code-triggered upgrades during roofing material selection. A commercial owner may need a formal risk workshop involving estimating, design, operations, procurement, safety, and finance.
Warning signs your contingency is not credible
- The number was copied from a prior project without comparing risk conditions.
- No one can explain what the contingency is meant to cover or who can approve its use.
- The same risk is counted twice, once in contingency and once in an allowance or line item.
- The contingency does not shrink or change as design information improves.
- The project has major unknowns, but the contingency is reduced only to meet a target budget.
- Field teams are discouraged from documenting risks because the budget story is politically uncomfortable.
GSA's cost estimating and cost management materials discuss principles that include budget control, cost management, and disciplined estimating practices. Reviewing public-sector cost management guidance can help owners frame contingency as an auditable project control rather than a vague cushion.
Governance matters as much as the starting amount. A contingency that anyone can spend without documentation will disappear into convenience decisions. A contingency that no one can access without a long approval chain may fail to solve real field problems. The approval process should match the project size and risk. Small draws might be approved by the project manager with supporting photos, while larger draws may require owner review, design input, and cost backup. On growing contractors, even tool replacement and rental gaps should be separated from broader tool standardization choices so contingency is not used to hide recurring operating problems.
The team should also record rejected draws, because those decisions reveal which events are scope growth, estimating misses, or contractor risk rather than true contingency events.
A contingency conversation worth having early
A good contingency conversation is specific. Instead of asking 'Is 10 percent enough?' ask 'What risks remain after design, procurement, site investigation, and contract award, and how will we track them?' That question turns a percentage debate into a risk conversation.
The next step is to create a contingency log with four columns: risk event, planned amount or range, approval owner, and actual draw. Review it monthly with the cost report. If the project is early, keep the log broad. If the project is in construction, tie each draw to a documented event. That discipline helps owners preserve funds for real uncertainty and helps contractors avoid treating contingency as a substitute for complete scope definition.
