Ask around and you will get the same answer: contingency runs at about 10%. It is the number in most textbooks, most templates and most financiers' expectations, and as a starting position it is fine.
It is also where the thinking usually stops, which is the problem. A 10% contingency on a budget that was never properly scheduled is not a safety margin, it is a rounding of the guess. And a production that finishes with its contingency fully intact has not necessarily managed itself well. It may simply have padded every line and called the total a triumph.
The working range
Most productions carry contingency somewhere between 5 and 10% of the total budget, and the position within that range is driven by risk rather than preference:
Toward 5% when the production is genuinely predictable: a studio-based format, a returning series where last year's numbers are known, controlled locations, a crew that has worked together, no significant weather exposure, no complex VFX.
Toward 10% on a first series, an unfamiliar format, heavy location work, difficult weather windows, animals, children, stunts, significant VFX, or any international element where currency, travel and unfamiliar local costs all add variance at once.
Above 10% is possible and occasionally sensible, but it invites a question you need to be ready for. A financier looking at a 15% contingency will reasonably ask whether the budget underneath it is soft. Sometimes the honest answer is that the production carries unusual risk and the reserve reflects it. Often the more useful response is to go back and price the known risks properly, leaving contingency for what is genuinely unforeseeable.
Where a completion bond is involved, the guarantor will have a view, and it is usually a floor rather than a suggestion.
What contingency is actually for
The test is unavoidability. Contingency covers costs that could not reasonably have been anticipated at the time the budget was locked.
Legitimately contingency:
- Weather that stops a day beyond the allowance already budgeted
- Illness or injury forcing a reschedule
- A location falling through late with no comparable alternative
- Equipment failure at a point where replacement is the only option
- A supplier going under mid-production
- Genuine price movement outside your control
Not contingency:
- A department that was never properly costed
- Scope that grew because nobody said no
- Overtime that was predictable from the schedule on day one
- A rate everyone knew was optimistic when it went into the budget
- Anything you decided not to price because pricing it would have made the topsheet look bad
That last one is worth dwelling on, because it is the most common misuse and the hardest to catch. Using contingency to absorb a known-soft line is a way of hiding a budgeting decision inside a risk reserve. It works exactly once, and it works by consuming the protection you will need later for something you genuinely could not see.
The general principle: if you could have costed it at the desk, it belongs in its own line. Contingency exists for what you could not see from there.
Who releases it
This is the part most productions handle informally and then regret.
Contingency should have a named owner, usually the producer, sometimes with financier or guarantor approval above a defined threshold. Every draw should be recorded with a date, an amount, a reason and the account it went to.
Without that, contingency does not get spent, it gets absorbed. Departments run slightly over, the overages are quietly covered, nobody makes a decision, and by week seven the reserve is gone and no one can produce a list of where it went. The cost report will show that it happened, but only in retrospect, and only as an aggregate.
Practical arrangement that works:
- Contingency sits on its own account, never distributed across other lines
- Draws are approved individually, with the reason written down
- The report shows contingency drawn to date and contingency remaining, every week
- Transfers between departments are a separate mechanism, and they do not touch contingency
The separation between transfers and contingency draws matters. Transfers move money between departments and net to zero. Contingency draws increase what the production is spending. Blurring the two makes it impossible to tell an efficient production from one that is burning its reserve.
Contingency and the estimated final cost
A question that comes up on every production: does unspent contingency count in the EFC?
Both conventions exist, and the one that causes fewer arguments is to show contingency as a separate line in the cost report rather than folding remaining contingency into the estimate to complete. That way the report answers two questions at once: what will this cost on current information, and how much protection is left if it goes wrong again.
Folding it in produces an EFC that looks reassuringly on budget right up to the moment the reserve runs out, at which point the number moves sharply for no visible operational reason. Producers dislike surprises of that shape, and financiers dislike them more.
An untouched contingency is worth a conversation
Finishing with contingency intact feels like success and is sometimes exactly that. But it can also mean the budget was padded line by line, in which case the production was capitalised above what it needed and somebody's money sat idle for six months.
The way to tell the difference is department-level variance. A production that came in on budget because every department landed close to its line was well budgeted. A production that came in on budget because half the departments ran significantly under was over-budgeted, and the contingency on top of that was belt and braces on belt and braces.
Neither is a scandal. But the second one is worth knowing about before you build next year's budget on the same assumptions, which is the argument for keeping historic actuals somewhere you can actually query them rather than in a folder of closed spreadsheets.
The short version
- 5 to 10%, positioned by risk rather than habit
- Unavoidability is the test: if you could have priced it, price it
- One named owner, one account, every draw recorded with a reason
- Show it separately in the cost report, do not bury it in the ETC
- Untouched contingency is a signal, not automatically a win
Related reading: how to build a production budget and what is a production cost report.