Capital Expenditure Forecasting: The Annual CapEx Number Is Not the Forecast

The CapEx budget says $4 million.

The cash forecast also says $4 million.

So far, so good.

Then the project that was supposed to start in February starts in June, equipment requires a deposit nobody modeled, and a construction invoice crosses the year-end boundary.

The annual total may still be right.

The forecast is not.

Capital expenditure forecasting is usually less about the approved amount than the timing of the project behind it.

Separate approval from forecast

An approved project is permission to spend.

It is not evidence that the spend will occur exactly as budgeted.

I like keeping project approval status separate from forecast timing.

A project can be approved and delayed. It can be partially committed. It can be canceled. Scope can change.

When those states are hidden inside one annual number, Finance learns about the cash movement late.

Forecast projects, not one CapEx percentage

For material capital spending, I prefer a project schedule.

Project name. Owner. Approved amount. Amount committed. Expected timing. Cash-payment timing. In-service date where depreciation matters. Remaining forecast.

Not every laptop needs its own line.

Materiality should decide the detail.

Recurring small equipment can use a simpler run rate while large projects get explicit schedules.

Cash timing and accounting timing are different

This is where CapEx can confuse an otherwise clean model.

A deposit may leave cash before the asset is placed in service. Depreciation begins according to accounting treatment, not necessarily when the vendor is paid. A project may accumulate construction-in-progress before becoming a depreciating asset.

The forecast should respect those differences.

Otherwise the P&L, balance sheet and cash forecast can each contain a different version of the same project.

Do not let unused budget become automatic future spend

A project comes in $200,000 under budget.

That does not mean another department has discovered $200,000 of free money.

Budget authority and forecast expectation are different concepts.

If the project is complete, the forecast should release the unused amount unless management deliberately reallocates it.

This sounds obvious. Annual planning processes have a way of making unspent budget feel like a natural resource that must be extracted before December.

Maintenance and growth CapEx tell different stories

I like distinguishing spending required to maintain current operations from spending intended to create new capacity or growth.

Replacing aging equipment is different from opening a new facility.

The distinction helps management understand what can be delayed and what delay would actually cost the business.

It also improves scenario planning. A downside case may defer expansion projects while preserving critical maintenance.

Projects need trigger dates

The easiest CapEx forecast to maintain is one that knows what event should cause Finance to update it.

Contract signed. Purchase order issued. Deposit paid. Construction begins. Delivery moves. Project goes live.

Those are better signals than waiting for the monthly actual to appear.

If Finance only updates CapEx after invoices post, the forecast is describing cash movement after the business has already committed to it.

Reforecast the remaining project, not the original project

Once a project starts, the original budget becomes a reference point.

The forecast should focus on what remains.

Actual spend to date plus committed spend plus the current estimate to complete.

If the project is overrunning, show it. If timing moved, move it. If scope changed, update the expected total.

This is the same discipline behind a useful rolling forecast: preserve the target, but let the outlook change.

The CFO should be able to see the commitments

Cash surprises often begin before cash moves.

A signed contract, purchase order or approved project can create a future obligation that is invisible in current actuals.

A good CapEx forecast gives management visibility into that committed pipeline.

How much is already spent? How much is committed? What remains discretionary? When is cash expected to leave?

The annual CapEx number answers almost none of those questions.

It is a budget.

The project schedule is the forecast.

The first question I ask is what management is actually approving

“We approved $4 million of CapEx” can mean several things.

It can mean the board authorized a ceiling. It can mean projects totaling $4 million have business cases. It can mean purchase orders are already issued. Or it can mean somebody typed $4 million into the annual plan because that was the number available in October.

Those are not the same commitment.

I want the forecast to distinguish authorization, commitment and expected spend.

Otherwise an approved project can sit in the model for nine months looking inevitable while the business has not ordered a thing.

A project schedule gives the number a story

For every material project, I want enough information to explain the movement.

What is being purchased or built? Who owns it? What problem does it solve? What has been approved? What has been contracted? What has been spent? What remains? When does the owner actually expect the milestones to occur?

That does not require a 40-column project-management system inside FP&A.

It requires enough structure that a $700,000 movement does not arrive in cash with the explanation “Oh, right. The equipment came in.”

CapEx is one of those areas where a little operational context saves a lot of financial surprise.

Deposits make annual thinking especially dangerous

Large equipment and construction projects often have payment schedules that look nothing like a smooth monthly allocation.

Twenty percent on signing. Another payment at fabrication. Another on delivery. Retainage after completion.

If Finance takes a $1.2 million project and divides it by twelve, the annual number looks beautifully controlled while the cash forecast is fiction.

I would rather have an ugly project schedule that reflects the contract.

Cash cares about dates more than aesthetics.

Estimate to complete is where the forecast becomes useful

Once work starts, I want the conversation to shift.

The original budget remains important for accountability, but the forecast needs the current estimate to complete.

Suppose a $2 million implementation has spent $900,000. Another $500,000 is contractually committed. The project owner now expects $800,000 of additional work.

The question is no longer whether the original $2 million budget was reasonable.

The current outlook is $2.2 million unless something changes.

Hiding that because the approved budget still says $2 million does not protect the budget.

It just delays the conversation.

CapEx can make EBITDA look calmer than cash feels

This is one reason I want FP&A to understand the accounting treatment rather than treating capital spending as a separate spreadsheet.

A business can make a major cash investment without taking the same amount through current-period operating expense.

Management looking mainly at EBITDA can therefore feel better than the bank account does.

That is not an argument against EBITDA.

It is an argument for connecting the income statement, balance sheet and cash flow.

A large capital program is exactly where three-statement thinking stops being academic.

Depreciation should follow the operational reality too

For a meaningful capital program, I want the forecast to understand when assets are expected to go into service.

That affects depreciation timing and sometimes other operating assumptions.

A project delayed six months may move cash, depreciation and the capacity or savings the project was supposed to create.

If the business case assumed a new facility would add output in July and it opens in December, the forecast problem is larger than CapEx timing.

The operational benefit moved too.

I separate maintenance, compliance and growth

Not all CapEx has the same degree of choice.

Some spending keeps the existing operation functioning. Some satisfies safety, regulatory or contractual requirements. Some creates new capacity or supports growth.

That distinction becomes very useful when management needs to conserve cash.

“Cut CapEx 20%” sounds decisive.

It is much less useful if nobody knows which 20% can actually move without damaging operations.

In a downside scenario, I want to know what is deferrable, what is cancellable, and what simply has to happen.

Watch the project that is always one quarter away

Every capital plan seems to have one.

The project was supposed to start in Q1. Then Q2. Now Q3 looks more realistic. The annual forecast still contains the full spend because removing it feels too aggressive.

Eventually December is expected to perform feats of construction that would impress HGTV.

I prefer to force the timing conversation.

What milestone has to happen next? Has it happened? If not, what is the realistic earliest spend date?

Forecasting should reflect the project we have, not the project the original budget remembers.

CapEx reviews should include project owners

Finance can reconcile invoices.

Finance usually cannot tell from the ledger whether an engineering project is six weeks behind schedule.

The project owner knows that.

I like a regular review of material projects where Finance brings actual spend and commitments, and the owner brings schedule, scope and estimate-to-complete information.

It does not need to be a two-hour committee.

For many companies, a focused monthly review of the handful of material projects is enough.

The value is getting operational information into the forecast before it becomes accounting history.

Variance analysis needs more than “timing”

CapEx variance explanations have a tendency to become one word.

Timing.

Fine. Which timing?

Was the project delayed? Was the invoice delayed? Was the purchase order issued later? Did the vendor change terms? Did scope move? Is the spend still expected at all?

“Timing” is useful only if Finance can say where the money moved and why.

Otherwise it is the capital-project version of “miscellaneous.”

What I want the CFO to see

For material CapEx, I want a view that answers a few questions without archaeology.

What was approved? What has been spent? What is contractually committed? What is the current estimate to complete? What remains discretionary? When will cash leave? What operational milestone or benefit is tied to the project?

That is much more useful than one annual CapEx line.

The annual line tells me the destination management approved months ago.

The project schedule tells me where we are now.

And for forecasting, now is the part I care about.

CapEx forecasting also needs an owner for the benefit

There is another side of capital spending that Finance can lose once the purchase order is approved.

Why were we spending the money?

A growth project usually promised something: additional capacity, lower unit cost, faster throughput, new revenue, fewer contractors, better reliability.

I like keeping that expected benefit connected to the project, particularly for large investments.

If the opening date moves six months, the benefit probably moves too.

If scope is reduced, perhaps the benefit changes. If the project costs 20% more, the return changes even if the operating benefit does not.

This does not mean FP&A needs to reopen every business case monthly.

It means the capital forecast should not treat the cash outflow as the only thing that happened.

Large CapEx is usually an operating decision wearing accounting clothes.

The forecast should make cancellation visible

Projects have a strange tendency to remain in forecasts after the business has quietly stopped believing they will happen.

Nobody wants to remove them because they are technically still approved.

So the dollars move to the right every month like a suitcase nobody wants to claim.

I prefer an explicit status: active, delayed, on hold, canceled, complete.

If a project is on hold, show it separately from expected spend. If it is canceled, release it. If management later revives it, put it back with a new realistic schedule.

That sounds almost offensively simple.

It also prevents the fourth quarter from accumulating every project Finance was afraid to remove during the first three.

December has enough going on.

It does not need to construct three facilities and install an ERP because the annual budget said so.

One final thing: I would keep the original approved budget visible beside the current forecast rather than overwriting it. Management needs both truths. The budget tells us what we expected when the decision was approved. The forecast tells us what we believe now. Losing either one makes the conversation worse.

Accountability does not require pretending the original estimate is still the forecast.