Contingency in a construction budget: how much is enough
Spring & Co, Stockport. Construction & property consultants since 1992
Carry 5% to 10% of the construction cost on a new build with a settled design, and 10% to 15% on a refurbishment or a conversion. Below 5% you are not holding a contingency, you are hoping. That is the short answer. The longer one is about where the money actually goes, and why the same percentage saves one project and sinks the next.
Two schemes in the same year, both set at 5%. On the new build the allowance held with a little to spare. On a mill conversion on the Stockport side of the M60, the same 5% was gone before the frame was up, because the existing structure and the made ground underneath it had opinions nobody had costed. Same number, two different outcomes. The percentage was never the point.
What is contingency actually for?
Contingency is money set aside for things you know will come up but cannot price yet. The RICS New Rules of Measurement (NRM1) calls it a risk allowance: the amount added to the base cost estimate for items that cannot be precisely predicted, so the estimate reaches a realistic cost limit. It is not a slush fund, and it is not hidden profit. It is priced uncertainty, sitting in the budget where you can see it.
So it is not there to pay for the things you later decide you want. A bigger plant room, an upgraded specification, a change of use halfway through: those are variations, and they need their own money. The day contingency starts quietly funding scope, you have lost your only buffer against the risks it was meant to cover.
How much should I actually carry?
The right figure moves with how much you know. Early on you know almost nothing, so the allowance is high. By the time the design is fixed and tendered, most of the guesswork has turned into priced work and it comes down. During construction, with the drawings frozen, it comes down again.
| Stage or project type | Typical contingency | Why |
|---|---|---|
| Feasibility or early estimate | 15% to 25% | The design barely exists, so nearly everything is still a risk |
| New build, design settled at tender | 5% to 10% | Most unknowns are now priced work |
| Refurbishment or conversion | 10% to 15% | You inherit a building whose secrets you cannot fully see |
| Older or structurally uncertain building | up to 25% | Hidden structure, past alterations, unknown ground |
| During construction, design frozen | 3% to 5% | Only genuine surprises are left |
These are ranges we see across UK projects, not a rule you can apply blind. A simple warehouse on a clean greenfield plot can sit comfortably at the bottom of its band. A Victorian building being brought back into use can burn through the top of its band and ask for more. Read the building, not the table.
If you are pricing a scheme now and are not sure the risk line stacks up, send us the cost plan and we will pressure-test it before you commit funding.
Refurbishment and conversion are where thin contingencies come to die, and around here that matters more than most places. A large share of Greater Manchester's development land is former industrial ground, and many of its best conversion opportunities are old mills and mid-century commercial blocks. That stock arrives with made ground, buried foundations, asbestos, tired services and structure that has been altered three times since the war and documented none of those times. None of it shows up until you open the building, and all of it wants paying for.
Why one line called "contingency" is close to useless
Here is an opinion plenty of consultants will not thank me for: a cost plan that carries contingency as a single percentage line is doing you a disservice, and most of them do exactly that. NRM1 splits risk into four categories for a reason, and I want to see all four priced on their own.
- Design development risk: an unclear brief, gaps between disciplines, details that are not resolved yet.
- Construction risk: ground conditions, existing services, contamination, defects you find on opening up.
- Employer change risk: the variations you introduce, in scope or in timing.
- Employer other risk: acceleration, early handover, funding constraints, the commercial pressures around the job.
Split like that, the number tells a story. A scheme carrying most of its risk in the construction category is telling you the ground and the existing structure are the danger, so that is where the survey money goes first. One loaded with employer change risk is telling you the brief is not settled and the drawings will move. A single blended percentage hides all of that behind one comfortable-looking figure.
What the optimism bias data says about getting it wrong
Government worked out long ago that professionals underestimate costs by default, and put a number on it. The HM Treasury Green Book, in its supplementary guidance on optimism bias, tells public bodies to uplift capital estimates at business-case stage by up to 24% for a standard building and up to 51% for a non-standard one. Those are upper bounds, meant to fall as real, project-specific risk gets identified and retired.
You do not have to adopt those figures to take the point. If the Treasury assumes a fresh building estimate can be a quarter light, then a 3% contingency on an early number is not caution. It is optimism with a spreadsheet.
Where the money actually gets eaten
Most of the contingency I have watched vanish went to one of three places. Ground nobody could see until it was dug. Existing structure that did not match the drawings. And the quiet one, contingency spent simply because it was sitting there unused at month four and the job had found other things to do with it.
That last one is a management failure, not a pricing failure, and it is the easiest of the three to prevent.
How we hold contingency on a monitored scheme
On the projects we monitor, contingency is ring-fenced and it stays ring-fenced. It is drawn only against a risk that has actually materialised, with a paper trail behind each release, and the remaining balance is reported at every valuation so nobody is guessing how much cover is left. That discipline is a large part of what our project monitoring and auditing and fund monitoring and due diligence work exists to do, especially where a lender needs to know their money is only paying for the works in front of it.
Getting the figure right in the first place is a cost management and quantity surveying job, and it starts before the design is finished, not after the tenders come in. Set the allowance against real, named risks for your building and your ground, split it the way NRM1 does, and it stops being a round number somebody guessed and becomes a tool you can run the job with.
Before you commit funding to a build, it is worth having someone independent check that the contingency actually covers the risks your project carries. Talk to Spring & Co about your scheme and we will tell you where we think the budget is exposed.