Construction cost overruns are not a failure of execution. They are a failure of budgeting math. The contingency line on most development pro formas is a round number, 5 or 10 percent of hard costs, chosen because it looks prudent, not because it was sized against the specific risks that blow budgets: material escalation, change orders, and labor. When the overrun arrives, it does not respect the round number. It respects the exposure, and the exposure was never measured. That gap is why the line runs dry before the building tops out.
The thesis is that contingency is treated as insurance and priced as a placeholder. It is neither. It is a reserve that has to survive the difference between the market you underwrote and the market you build in.
Key Takeaways
Construction cost overruns are structural, not incidental. A contingency set as a flat 5 to 10 percent of hard costs is a placeholder, not a reserve sized against escalation, change orders, and labor.
The Producer Price Index for inputs to new nonresidential construction rose 8.4 percent year over year in May 2025, the largest jump since the pandemic, per the Associated General Contractors of America. A contingency budgeted in a calm year cannot absorb an 8 percent escalation year.
The Turner Building Cost Index rose 4.87 percent year over year in the first quarter of 2026, per Turner Construction, meaning even a "normal" market moves the number faster than a typical 5 percent contingency assumes.
Contract structure decides who eats the overrun. A guaranteed maximum price caps the owner's exposure and shifts overrun risk to the contractor; a cost-plus contract with no ceiling leaves the owner holding it.
A hard cost overrun of 12 percent against a 7 percent contingency can cut a merchant developer's profit by more than 10 percent, because the uncovered overage lands directly on total project cost and compresses the development spread.
Why do construction cost overruns exceed the contingency line?
Construction cost overruns exceed the contingency line because the line is sized as a percentage convention rather than a risk estimate. A developer writes 5 or 10 percent of hard costs into the budget because that is what the loan underwriter expects to see. That number carries no information about material escalation, the completeness of the drawings, or how tight the labor market is on the day crews mobilize.
Three drivers break budgets. Material escalation is the first. The Producer Price Index for inputs to new nonresidential construction rose 8.4 percent year over year in May 2025, the sharpest increase since the pandemic, according to the Associated General Contractors of America, with aluminum mill shapes up 28 percent and diesel fuel roughly doubling. A 5 percent contingency does not survive an 8 percent escalation year, and escalation compounds over a two or three-year build.
Change orders are the second. Every design gap, owner revision, and field condition that differs from the drawings converts into a change order carrying both direct and schedule cost. Labor is the third: a shortage of skilled trades raises the price of every hour on site and stretches the schedule, which raises carry. The hard costs a contingency is meant to protect are the line items most exposed to all three.
What actually drives a cost overrun, and how is each risk mitigated?
The drivers of a construction cost overrun fall into a short list, and each has a specific mitigation that a flat contingency does not provide. Escalation is hedged with fixed-price material buyouts and escalation clauses. Change orders are controlled with design completeness before the GMP is set. Labor risk is managed with early subcontractor commitment. A contingency covers the residual, not the whole exposure.
The point of separating the sources is that each one has a different owner and a different tool. Lumping them into one percentage hides which risk is unpriced.
Overrun source | What it looks like | Primary mitigation | Who should carry the residual |
|---|---|---|---|
Material escalation | Steel, aluminum, fuel, and freight rising over a multi-year build | Fixed-price buyouts, escalation clauses, early procurement | Owner, unless locked by contract |
Change orders | Design gaps, field conditions, owner revisions | Design completeness before GMP, disciplined scope control | Owner for owner-driven scope; contractor for errors |
Labor and trade shortage | Higher hourly cost, schedule slippage, overtime | Early subcontractor commitment, realistic schedule | Shared, depends on contract type |
Schedule and carry | Interest and overhead accruing on delay | Realistic timeline, milestone tracking | Owner via the construction loan |
Design and coordination | Incomplete drawings, clashes discovered in the field | Constructability review, later GMP lock | Owner and design team |
Most residual risk defaults to the owner. That is the argument for pricing contingency against the sources rather than a round percentage. A project with complete drawings and locked material pricing carries a fraction of the escalation and change-order risk of one that broke ground on 60 percent construction documents. Sizing the same 5 percent to both ignores the difference. Entitlement uncertainty compounds this, which is why entitlement risk lives before you break ground and should be resolved before the budget locks.
How much contingency is enough, and how do you size it?
A representative contingency for hard costs runs in the range of 5 to 10 percent, with the specific number driven by design completeness, market volatility, and contract structure, not by convention. A project with finished drawings, a signed guaranteed maximum price, and locked material pricing sits at the low end. A project with incomplete design, cost-plus exposure, and a volatile materials market belongs at the high end or beyond.
The honest way to size contingency is to anchor it to the escalation the indices are actually reporting, then add for the project's specific unknowns. When the Turner Building Cost Index is climbing 4.87 percent year over year, as Turner Construction reported for the first quarter of 2026, and the AGC's input PPI has posted an 8.4 percent annual jump, a contingency built on a calm-market assumption is already behind. Engineering News-Record's construction cost index and the government's Producer Price Index for construction materials exist to give a developer a live read on the number rather than a guess.
Risk profile | Design completeness | Contract type | Representative hard cost contingency |
|---|---|---|---|
Low risk | Full construction documents | Signed GMP with locked buyouts | Lower end of 5 to 10 percent |
Moderate risk | Substantially complete design | GMP with open buyouts | Middle of the range |
Elevated risk | Design-build or fast-track, drawings still developing | Cost-plus with a soft cap | Upper end of the range or higher |
These ranges are representative, not prescriptive. The discipline is to derive the number from the project, then check the total budget against a live cost index before locking the construction loan, because the lender will fund to the budget you signed, not the one the market forces on you later.
How does a hard cost overrun erode the development spread? A worked example
A hard cost overrun erodes the development spread because the uncovered overage lands on total project cost, which is the denominator of yield on cost. Raise the denominator and the yield falls; a lower yield on cost against an unchanged market cap rate is a thinner spread. The overrun does not just cost money, it compresses the entire reason to take development risk.
Work the numbers with stated inputs. Assume a project with a total budget of 100 million dollars, of which hard costs are 70 million. The developer sets contingency at 7 percent of hard costs, which is 4.9 million dollars. Stabilized net operating income is projected at 6.5 million dollars, so the yield on cost is 6.5 percent. Comparable stabilized assets trade at a 5.0 percent market cap rate, giving a development spread of 150 basis points.
Now apply a 12 percent hard cost overrun. Twelve percent of 70 million is 8.4 million dollars of added cost. The 4.9 million contingency absorbs part of it, leaving 3.5 million dollars uncovered. That 3.5 million is added to total project cost, which rises from 100 million to 103.5 million.
New yield on cost: 6.5 million divided by 103.5 million equals 6.28 percent.
New development spread: 6.28 percent minus 5.0 percent equals 128 basis points, down from 150.
Value view: at a 5.0 percent exit cap, 6.5 million of NOI is worth 130 million. Created value falls from 30 million (130 minus 100) to 26.5 million (130 minus 103.5).
The overrun of 3.5 million erased 3.5 million of profit, cutting the merchant developer's created value by roughly 12 percent, and it compressed the spread by 22 basis points before a single tenant signed. The contingency was not wrong to exist. It was wrong to be a round number when the escalation risk was measurable. The pro forma assumptions buyers should challenge start with the ones a developer never stress-tested, and the contingency line is the first of them.
Frequently Asked Questions
What is a typical construction contingency percentage?
A typical construction contingency for hard costs runs in the representative range of 5 to 10 percent, with the exact figure driven by design completeness, contract type, and market volatility. Full drawings and a signed guaranteed maximum price justify the low end. Incomplete design, cost-plus exposure, and a volatile materials market push it to the high end or beyond.
Who pays for a construction cost overrun, the owner or the contractor?
It depends on the contract. A guaranteed maximum price caps the owner's cost and makes the contractor absorb overruns above the cap, unless a change order raised the ceiling for added scope. A cost-plus contract with no ceiling leaves the owner holding the overrun. Contract structure, not good intentions, decides who eats the number.
Why do material costs make construction cost overruns worse?
Because material prices move faster than budgets assume and compound over a multi-year build. The Associated General Contractors of America reported the input Producer Price Index for new nonresidential construction up 8.4 percent year over year in May 2025, the largest since the pandemic. A contingency set in a calm year cannot absorb that escalation once concrete is being poured.
Conclusion
Construction cost overruns are structural, and the contingency line fails because it is priced as a placeholder instead of a reserve. A round 5 or 10 percent carries no information about the escalation the indices are reporting, the change-order exposure baked into incomplete drawings, or the labor market on mobilization day. Size contingency against the sources, check the budget against a live cost index from the AGC, Turner, or ENR before locking the loan, and pick a contract structure that puts the residual risk where it belongs. A developer who quantifies the reserve keeps the development spread. A developer who rounds it discovers the shortfall after the concrete is set.