Aller au contenu
Corporate & Investment Banking Mon parcours English

Accueil / Corporate Finance / Valeur et investissement / Module 24

24Investment decisions: NPV and IRR

How a company decides whether a project is worth the money

Comment une entreprise décide si un projet vaut l’argent qu’il coûte.

Lecture
20 min
Sections
10
Exercices
2
Questions de rappel
6
Questions de QCM
12
Simulateurs
2

Le texte de ce module est encore en anglais : la version française arrive. Les outils interactifs, eux, sont déjà en français.

What you will be able to do

  • Compute a net present value and say in plain words what the number means.

  • Build the cash flows of a project, keeping what belongs and throwing out what does not.

  • Compute an internal rate of return and name the three situations where it misleads.

  • Use payback, the profitability index and the equivalent annual cost for the jobs they are good at.

  • Challenge a business case the way an investment committee does.

1The decision rule

Before any spreadsheet, the logic. A project is a trade between today and tomorrow, and there is one fair way to score that trade.

An investment is a trade: cash today for cash later

You cannot add euros that arrive at different dates

Suppose a project costs 100 today and returns 30 a year for four years. It is tempting to say it earns 20, because 120 comes back. That sum mixes euros of four different years, and they are not the same thing. Money you receive in four years could not be invested in the meantime, and it might not arrive at all.

The fix is to translate every future amount into what it is worth today. That translation is discounting. A cash flow received in year t is divided by (1 + k) raised to the power t, where k is the return investors require for waiting and for bearing the project's risk.

Net present value is everything you get, minus everything you pay, in today's money

NPV = − I₀ + CF₁ / (1 + k) + CF₂ / (1 + k)² + … + CFₙ / (1 + k)ⁿ

I₀ is the cash spent at the start. Each CF is the cash the project produces in a given year. The word net matters: the NPV is what is left after the investors have been paid the return they require. It is the surplus.

The rule fits in one line

If the NPV is positive, do the project: it creates wealth. If it is negative, do not: it destroys wealth, even if it shows an accounting profit. If it is zero, the project earns exactly what investors require and nobody is richer or poorer.

The discount rate is a price, not a preference

It is what investors give up by choosing your project

The rate k is an opportunity cost. Investors can always put their money into other assets of similar risk. The project must beat that alternative, otherwise they are better off elsewhere. This is why the rate depends on the risk of the project, and not on who pays for it or how the company happens to be financed this year.

A safe warehouse extension and a launch in a new country should not be judged with the same rate, even inside the same company. Module CF25 shows how the rate is estimated. In this module it is given.

A friend who runs a bakery wants to buy a second oven for 20,000. She says: it will bring in 5,000 more a year for five years, so I make 5,000. Explain to her, without using the words discount, present value or NPV, why her sum is not the right one and what question she should ask instead.

      The bakery. The 25,000 arrives over five years while the 20,000 leaves today. The same 20,000 placed elsewhere would also have grown. The question is whether the oven leaves her better off than the best other use of the money at similar risk.

      The car park. How much money does each project involve, and how much better off is the town in total? And is the 22% compared with the right benchmark for a project of that risk? A high percentage on a small sum can be worth less than a modest one on a large sum.

      1

      1. A project costs 100 and returns 30 a year for four years. At 9%, its NPV is…

      2. What does an NPV of +2 million mean for Harmony?

      3. Why does the discount rate depend on the project rather than on how it is financed?

      2Which cash flows count

      The formula is easy. Almost every bad investment decision comes from what was put into it. Six rules decide what belongs in the cash flows.

      Six rules for the cash flows

      Rule 1 — Count cash, not profit

      Depreciation is not a payment, so it is not a cash flow. The machine was paid for on day one. Sales made on credit are profit today and cash only when the customer pays. A project is scored on money that moves.

      Rule 2 — Count only what changes because of the project

      The right question is always: what happens to the company's cash with the project, compared with without it? Three consequences follow.

      • Sunk costs are ignored. Money already spent, such as a feasibility study, is gone whatever you decide. It cannot influence the decision.

      • Side effects are included. If the new product takes sales from an existing one, the lost cash belongs to the project. This is cannibalisation.

      • Allocated overheads are ignored unless they really increase. Charging the project a share of head-office rent changes the management accounts, not the cash.

      Rule 3 — Price what you already own at what it could fetch

      A project that uses a building the company owns is not using it for free. If the building could be sold or rented, that foregone cash is a cost of the project. Use the market value, never the value in the accounts.

      Rule 4 — Leave the financing out

      Interest and dividends do not appear in the cash flows. The discount rate already contains the return owed to lenders and shareholders. Subtracting interest as well would charge for the money twice. It also means the verdict on a project cannot be improved by borrowing more to pay for it.

      Rule 5 — Work after tax

      The tax office takes its share of the project's operating profit, and that payment is real cash. Depreciation is not cash, yet it lowers the tax bill, so it matters indirectly. Compute tax as if the project were financed entirely with equity, to stay consistent with Rule 4.

      Rule 6 — Be consistent

      Cash flows that include inflation need a rate that includes inflation. Cash flows in dollars need a dollar rate. Mixing the two is the most common silent error in business cases.

      From the six rules to one line of numbers

      Free cash flow is what the project hands back to all investors

      FCF = EBITDA − tax on operating profit − increase in working capital − capital expenditure + disposals after tax

      EBITDA is operating profit before depreciation. The tax is computed on operating profit after depreciation. An increase in working capital is cash locked into stock and customer credit.

      2.1 Two mistakes in one business case

      A junior analyst rebuilds Harmony's case. He adds the 0.4 million study to the initial outlay. He also deducts the interest on the loan that will finance the line: 0.6 million a year before tax, so 0.45 million a year after tax, for five years. Compute his NPV at 9%. Then say which rule each change breaks, and why the second mistake is the more dangerous of the two.

      The study lowers the NPV by 0.40. Five yearly deductions of 0.45 are worth 0.45 × 3.890 = 1.75 today. His NPV is 2.12 − 0.40 − 1.75 = −0.03 million. A sound project is rejected.

      The study breaks Rule 2: it is sunk. The interest breaks Rule 4: the 9% already pays the lenders, so they are charged twice.

      The second mistake is the more dangerous because it is large, it looks prudent, and its size depends on how the project happens to be financed. The same line would look good if paid for with equity and bad if paid for with a loan.

      the Harmony table with a toggle on each line (study, interest, lost repair margin, working capital recovery, resale). The learner decides what is in or out and watches the NPV move, then compares with the correct build.

      2

      1. Which item belongs in the battery line's cash flows?

      2. Why is depreciation relevant even though it is not a cash flow?

      3. The junior analyst adds the study and the after-tax interest. His NPV becomes…

      3The other yardsticks

      Managers rarely look at NPV alone. Four other measures are in daily use. Each answers a different question, and each has a blind spot.

      IRR: the rate at which the project breaks even

      The internal rate of return is the discount rate that brings the NPV to zero

      Raise the discount rate and future cash is worth less, so the NPV falls. The rate at which it reaches zero is the IRR. It is the highest required return the project could bear. The rule: accept the project if its IRR is above the required return.

      Discount rate 6% 9% 12% 15%
      Harmony's NPV (€ million) +3.45 +2.12 +0.95 −0.08

      The NPV crosses zero just below 15%: the IRR of the battery line is 14.8%.

      Harmony's line returns 14.8% against a required 9%. For a single, ordinary project, IRR and NPV always give the same yes or no. Managers like the IRR because it is a percentage that can be compared with a cost of money. That convenience hides three traps.

      Trap 1 — IRR ignores size

      Harmony can equip a workshop in two ways, and only one can be chosen. Option A costs 2 million and returns 1.2 million a year for three years. Option B costs 10 million and returns 4.6 million a year for three years.

      At 9% Outlay IRR NPV
      Option A 2.0 36.3% +1.04
      Option B 10.0 18.0% +1.64

      A has twice the IRR. B makes shareholders 0.6 million richer. Choose B. A high percentage on a small amount is worth less than a decent percentage on a large one. To check with rates, look at the extra 8 million that B requires: it brings 3.4 million more a year, an IRR of 13.2%, still above 9%.

      Trap 2 — Some projects have several IRRs, or none

      When cash flows change sign more than once, for example a mine that must be cleaned up at the end, the equation can have several solutions. Take −4, then +25, then −25. The NPV is zero at 25% and again at 400%. At 9% it is −2.11. No single rate describes this project, but the NPV still gives a clear answer.

      Trap 3 — IRR assumes the cash can be reinvested at the IRR

      An IRR of 36% is only earned over the whole period if every euro released is reinvested at 36%. That is rarely possible. NPV makes the modest assumption that cash is reinvested at the required return.

      Modified IRR

      The modified IRR carries each positive cash flow forward to the final year at the required return, then finds the single rate that links the outlay to that final sum. It removes the reinvestment problem and always gives one answer. It is still a percentage and still ignores size.

      a slider for the discount rate from 0% to 40% that redraws the NPV curve of options A and B. The learner sees the two curves cross near 13% and understands why the ranking by NPV changes there while the ranking by IRR never does.

      Payback, profitability index and annual cost

      Payback measures how long the money is at risk

      Payback is the time needed to recover the outlay. Harmony's line has recovered 9.43 million after three years and the rest during year 4: about 3.6 years. Using discounted cash flows the answer is 4.4 years.

      Payback ignores everything that happens after the cut-off, so it punishes long projects and says nothing about value. It survives because it is simple and because it measures something real: exposure. A firm short of cash, or operating where the future is very uncertain, has good reason to care how fast the money returns. Use it as a second opinion, never as the verdict.

      The profitability index ranks projects when money is rationed

      Profitability index = NPV / initial outlay

      For Harmony's line: 2.12 / 12.0 = 0.18. Each euro invested creates 18 cents of wealth. When the budget is capped, ranking by this ratio points to the projects that create most value per scarce euro.

      3.1 Ten million and four projects

      Harmony's budget is capped at 10 million. Projects cannot be split.

      Project Outlay NPV
      P 6 1.8
      Q 4 1.4
      R 5 1.6
      S 10 2.6

      Rank the projects by profitability index. Then find the combination that creates the most value. Are the two answers the same?

      Index: Q 0.35, R 0.32, P 0.30, S 0.26. Following the ranking gives Q + R, an outlay of 9 and an NPV of 3.0, with 1 million left unused.

      The best combination is P + Q: an outlay of 10 and an NPV of 3.2. S alone gives 2.6.

      The answers differ because projects cannot be split. The index is a guide. With indivisible projects, compare the total NPV of the combinations that fit the budget.

      Equivalent annual cost compares machines with different lives

      Machine X costs 6 million, lasts three years and costs 1.0 million a year to run. Machine Y costs 9 million, lasts five years and costs 0.8 million a year. Comparing total cost is unfair because Y works two years longer. Convert each total cost into the constant yearly amount that has the same present value.

      At 9% Present value of costs Equivalent annual cost
      Machine X (3 years) 8.53 3.37
      Machine Y (5 years) 12.11 3.11

      Y is cheaper per year of service, by 0.26 million. Choose Y, provided the machine will be needed for the long run.

      The accounting rate of return belongs to reporting, not to decisions

      Average accounting profit divided by average capital invested is easy to read from the accounts. It ignores timing and depends on the depreciation method chosen. It is useful afterwards, to see whether the project delivers, because results are reported in that form.

      1. 3.2 A project has an IRR of 11%. Is it a good project?

      2. 3.3 Why can two managers computing the IRR of the same project never disagree, while they can disagree on its NPV?

      3. 3.4 A project pays back in two years. Name one thing you know and two things you do not.

      3

      1. Option A costs 2 with an IRR of 36.3%; option B costs 10 with an IRR of 18.0%. Only one can be done, at 9%. Choose…

      2. Harmony's line pays back in about 3.6 years. What does that tell you?

      3. With a budget of 10 and indivisible projects, how do you choose?

      4Judgement: what the spreadsheet cannot tell you

      The arithmetic is never the hard part. The hard part is believing the forecast.

      A positive NPV needs a reason

      In a competitive market, the normal NPV is zero

      If a project were obviously lucrative, competitors would do it too, prices would fall and the surplus would vanish. A positive NPV therefore has to come from something rivals cannot copy quickly: a patent, a brand, a cost advantage, a location, a head start. An investment committee's first question is not about the discount rate. It is: why us, and for how long?

      Forecasts are written by people who want the project

      The team that proposes a project also writes its numbers. Optimism creeps into volumes, prices and timetables. Experienced companies respond by comparing forecasts with the actual results of past projects, and by asking for a downside case. They do not respond by adding a few points to the discount rate, which punishes distant cash flows far more than near ones and hides the real doubt.

      Strategic projects still have to pay

      A project with a negative NPV is sometimes defended as strategic. That can be legitimate if it opens the door to later projects that are valuable. Then the value of those later options should be estimated, as CF26 explains. Strategic is a reason to look harder at the numbers, not to stop looking.

      Carbon and inflation go into the cash flows

      Climate constraints do not call for a new decision rule. They change the inputs. Many groups now apply an internal carbon price to a project's emissions, which lowers the cash flows of polluting options and can reverse a ranking. Inflation is treated the same way: put it into prices, wages and working capital, and use a rate that includes it.

      What companies actually do

      Surveys of finance directors in the United States and Europe show that most large companies use both NPV and IRR, and that more than half still look at payback. Smaller firms rely more on payback and on the owner's instinct. Using several measures is sensible, as long as NPV has the final word when they disagree.

      Your cousin has just joined a local council. A consultant tells the council that a new car park has an IRR of 22% and should be preferred to a tram extension with an IRR of 9%. Explain to your cousin, in everyday language, two questions she should ask before accepting that comparison.

          1. 4.1 Harmony's supplier offers a 10% price cut on battery packs if Harmony drops the project. Which line of the table changes, and in which direction does the NPV move?

          2. 4.2 Why is adding three points to the discount rate a poor way to deal with an optimistic forecast?

          3. 4.3 A manager says the line should be built because the study has already cost 0.4 million. What is wrong with the argument?

          4

          1. A project shows a large positive NPV in a competitive market. The first question is…

          2. How should a company handle an optimistic forecast?

          3. How does an internal carbon price enter an investment decision?

          Summary and cheat sheet

          Les formules du module

          Touchez un terme pour voir ce qu’il veut dire et sa valeur chez Harmony.

          Valeur actuelle nette

          VAN = − + Σ / (1 + )

          Ligne de batteries : VAN = +2,12 M€.

          Flux disponible d’un projet

          = − − −

          Indice de profitabilité

          = /

          Coût annuel équivalent

          = /

          Idea What to remember
          Net present value Present value of all cash flows minus the outlay. Positive: do it. It is an amount of wealth created.
          Discount rate Return investors could get elsewhere at the same risk. Depends on the project, not on its financing.
          Free cash flow EBITDA − tax on operating profit − increase in working capital − capex + disposals after tax.
          What to leave out Sunk costs, allocated overheads that do not change, interest and dividends.
          What not to forget Cannibalisation, opportunity cost of owned assets, working capital in and out, resale value after tax.
          IRR Rate at which NPV is zero. Fine for one ordinary project. Misleads on size, on sign changes and on reinvestment.
          Payback Time to recover the outlay. Measures exposure, not value.
          Profitability index NPV per euro invested. Guides the choice when the budget is capped.
          Equivalent annual cost Present value of costs spread evenly over the life. Compares assets with different lives.
          Harmony's battery line Outlay 12.0 · NPV +2.12 at 9% · IRR 14.8% · payback 3.6 years · index 0.18.

          The one-sentence version

          Forecast only the cash that the project changes, after tax and before financing, translate it into today's money at the return investors require for that risk, and go ahead if what comes back is worth more than what goes in.

          Key terms

          Term Plain-language meaning
          Discounting Converting a future amount into what it is worth today.
          Sunk cost Money already spent that no decision can bring back.
          Cannibalisation Sales that a new product takes from the company's own existing products.
          Opportunity cost What you give up by using a resource here instead of in its best alternative use.
          Working capital Cash tied up in stock and customer credit, net of supplier credit.
          Capital rationing A cap on the investment budget that prevents doing every good project.

          Going further

          • Vernimmen, Quiry, Le Fur — Finance d'entreprise (Dalloz), chapter on investment decisions.

          • Brealey, Myers, Allen — Principles of Corporate Finance, chapters on net present value and investment decisions.

          • Graham and Harvey (2001) — The theory and practice of corporate finance: evidence from the field, Journal of Financial Economics.

          Independent educational material. Not investment, legal or tax advice. All companies and figures in the running case are fictional.