Before You Sign the Capex: 6 Rules for Approving a Capital Investment (Hurdle Rate, NPV and Payback)

Berlin Brandenburg Airport was supposed to open in 2011. It opened in October 2020, nine years late and at roughly three times its first budget. Every one of those overruns started the same way: a plan that looked sensible on paper and was approved on the assumption that it would go as planned.

That is not a public-sector problem. Bent Flyvbjerg, the Oxford professor who has built a database of more than 16,000 projects, reports in How Big Things Get Done (2023) that only 8.5% came in on budget and on time, and only 0.5% delivered on budget, on time and with the promised benefits.

Most capital investments fail long before construction starts, at the moment someone approves a spreadsheet nobody tried to break.

Our group was founded in 1890 and today runs eight business verticals in 75+ countries. In businesses like wine and mineral water, the big capital decisions look like bottling lines, tanks, warehouses and buildings: assets you pay for today and live with for twenty years. These are the six rules I would put in front of any owner before signing a capital expenditure (capex).

Why capex decisions deserve their own discipline

A capital investment is different from almost every other decision a company makes. The money goes out at once, the benefits arrive slowly, and the decision is hard to reverse. You can stop a marketing campaign on Monday. You cannot un-pour a concrete floor.

That is why I treat capex as a separate process inside capital allocation, not as a line in the annual budget. The question is never "can we afford it?" but "is this the best use of this money compared with everything else it could do?"

Rule 1: Write the decision memo before the spreadsheet

A spreadsheet answers the question you give it. If the question is "how much will the new line save?", the model will find savings. The memo forces the earlier questions:

• What problem does this solve? Capacity, cost, quality, safety, or a new market.
• What are the alternatives? Including a cheaper machine, a used one, outsourcing, leasing, and doing nothing.
• What happens if we do nothing? Sometimes the honest answer is "not much", and the project dies right there.

If the memo cannot explain the investment in one page, the spreadsheet is decoration. I ask for the memo first, and I only read the model once the alternatives are on the table.

Rule 2: Use a hurdle rate above your cost of capital, and tier it by risk

Your cost of capital is what the money costs you: the blend of what lenders charge and what owners expect to earn. The hurdle rate is the minimum return you demand before you approve a project. They are not the same number, and they should not be.

The hurdle sits above the cost of capital because forecasts are optimistic and because capital is scarce. Approving every project that clears the cost of capital leaves no room for the better project that arrives next quarter.

One rate for everything is also a mistake. I use tiers:

• Maintenance and safety: replacing what must be replaced. The question is the cheapest sound option, not the return.
• Efficiency in a business we know: a modest premium over the cost of capital.
• Expansion, new products or new markets: a clearly higher hurdle, because the forecast is weaker.

A single hurdle rate quietly subsidizes the riskiest projects with the safety of the boring ones.

Rule 3: Use NPV for value and payback for survival

In the best-known survey on the subject, John Graham and Campbell Harvey asked 392 CFOs how they evaluate projects (Journal of Financial Economics, 2001). About 76% always or almost always used IRR, 75% used net present value (NPV), and 57% used payback, the simplest method of all.

The textbooks dismiss payback because it ignores the time value of money and everything after the break-even year. They are right about the theory. For a private company, payback still answers a question NPV does not: how long is our cash exposed?

NPV tells you whether a project creates value; payback tells you whether the company survives while waiting for it. A family company with a bank covenant cares about both. I want to see a positive NPV at the right hurdle and a payback that our balance sheet can carry through a bad year.

Rule 4: Budget for the reference class, not for your plan

Flyvbjerg's central lesson is to stop forecasting from the inside. The team looks at its own plan, its own schedule and its own suppliers, and concludes that this time will be different. It rarely is.

The fix is the outside view: find a group of similar projects you or your industry have already completed, and look at how far they actually ran over in cost and time. Then apply that distribution to your plan.

For a private company, the reference class is often your own history. How late was the last warehouse? How far over budget was the last IT system? Your last three projects are a better forecast of the next one than your project manager's Gantt chart. If you have never tracked those numbers, start now: Rule 6 builds the data.

Rule 5: Stress the two assumptions that kill projects, delay and shortfall

Most capex models are sensitive to many inputs, but two do most of the damage: the benefits arrive later than planned, and they are smaller than planned. Take an illustrative example:

• A new bottling line costs €1.2 million and is expected to save €300,000 a year for eight years.
• At a 10% hurdle rate, its NPV is about +€400,000, with a four-year payback. Easy approval.
• If savings come in 20% lower, at €240,000 a year, NPV falls to about +€80,000.
• If the line starts a year late, NPV falls to about +€255,000.
• If both happen, which is exactly what Flyvbjerg's data says is normal, NPV turns to about -€36,000 and payback stretches to six years.

A project that only works if nothing goes wrong is not a good project; it is a bet on luck. Before approving, I ask for this table: base case, 20% shortfall, one-year delay, and both together. If the combined case destroys value, the project needs a cheaper design, a stronger case, or a no.

Rule 6: Approve in stages, then look back

Two habits close the loop.

The first is staged approval. Instead of releasing the whole budget at once, release it in phases, with a clear checkpoint: design, then order, then installation. Each gate is a chance to stop a project whose numbers have moved. Stopping at 15% spent is cheap. Stopping at 85% almost never happens.

The second is the post-investment review. Twelve to twenty-four months after commissioning, compare what the memo promised with what actually happened: cost, schedule, and benefits. Not to punish anyone, but to build your reference class and calibrate the people who forecast.

The companies that get better at capex are the ones that check their old forecasts. It is the same principle behind any budget that actually runs the company: plans improve only when someone compares them with reality.

Key Takeaways

• Capital investments are hard to reverse, so they deserve their own process, separate from the annual budget.
• Write a one-page decision memo, including "do nothing", before anyone builds a model.
• Set a hurdle rate above your cost of capital and raise it for riskier categories of project.
• Use NPV to measure value and payback to measure how long your cash is exposed.
• Forecast cost and schedule from similar past projects, not from the plan itself.
• Always test a delay and a shortfall together; if that case destroys value, redesign or decline.
• Release money in stages and review every major project 12 to 24 months after it starts running.

Frequently Asked Questions

Is hurdle rate the same as cost of capital?

No. The cost of capital is what your funding costs you, a blend of the interest you pay lenders and the return owners expect. The hurdle rate is the minimum return you require before approving a project, and it is normally set above the cost of capital to cover forecasting optimism and project risk.

What is a good hurdle rate?

There is no universal number. A good hurdle rate starts from your own cost of capital and adds a premium that grows with the risk of the project. Maintenance spending may need little more than the cost of capital, while a new product or a new country deserves a clearly higher rate.

How do you evaluate a capital expenditure?

Start with a written case that states the problem, the alternatives and what happens if you do nothing. Then calculate NPV at your hurdle rate and the payback period, and stress-test the result with a delay and a shortfall in benefits. Approve in stages and review the outcome after the project is running.

Which capital budgeting method ignores the time value of money?

The simple payback method. It counts how many years it takes to recover the investment without discounting future cash flows, and it ignores everything after the break-even point. It is still useful as a measure of cash exposure, as long as it is used alongside NPV rather than instead of it.

Before You Sign the Next Capex Request

Pick the largest investment on your table this year and ask for three things: the one-page memo, the four-case stress table, and the overruns of your last three comparable projects. If nobody can produce them, you have found the real project.

The same discipline applies to the biggest capital decision many owners face, whether to own or lease the building they operate from. To see how we apply long-term thinking across our businesses, discover the Manzanos Enterprises group.