Forecasting cash on a construction project

A cash flow forecast on a construction project answers one question: on any given date, how much money will have come in and how much will have gone out. The difference is what has to be funded, and it is almost always negative for most of the job.

What the naive version misses

The common approach is to take the valuation schedule, spread the contract sum across it on an S-curve, and call that the cash flow. Three things are missing, and each is material.

• Retention. A percentage of every certificate is not paid. On a three per cent retention with a five million pound contract, that is a hundred and fifty thousand pounds not arriving, half of it not until practical completion.
• The payment lag. Value is earned at the valuation date, certified some days later, and paid at the final date for payment. Under JCT that is fourteen days from the due date, which is itself seven days after the valuation date. Three weeks minimum between doing the work and being paid, and that assumes everything runs to time.
• Outgoings run on a different clock. Labour is paid weekly, materials on supplier terms, subcontractors on the subcontract cycle. None of those align with the main contract cycle.

The shape of the curve

Most projects show a widening negative position through the early and middle stages, a peak somewhere past the halfway point, and a slow recovery towards practical completion. The peak is the funding requirement, and it is the number the forecast exists to produce.

The recovery is slower than people expect, because the last certificates are small and the retention releases sit months and then a year beyond practical completion.

Where forecasts go wrong

Assuming payment on time. Late payment is common enough that a forecast built on contractual dates is a best case rather than an expectation. Modelling a fortnight of slippage on receipts is not pessimism, it is realism.

Ignoring the final account tail. The last five per cent of the contract sum frequently takes a year to convert to cash, and a forecast that shows the project cash positive at practical completion is usually wrong by the amount of the retention and the unagreed final account.

Updating it

A forecast prepared at the start and never revisited is a document, not a tool. It should be reforecast each month against actuals, and the variance explained.

The value is in the variance rather than the forecast. A month where receipts came in eighty thousand short of plan is telling you something now that the account will otherwise tell you in six months.

The one figure to know

The peak funding requirement, and the date it occurs. Everything else in the forecast supports that number, and a business that knows it can plan. A business that does not is relying on the bank to notice first.