Internal Rate of Return (IRR), Explained: The Discount Rate That Zeros Out NPV, the Hurdle-Rate Rule, and the Traps

Two equipment upgrades each promise $12,000 of total savings on a $10,000 investment. The first pays back steadily over four years; the second pays almost nothing until a big payoff in year four. Same profit on paper — but the first deal is clearly better, because money in your hands today is worth more than money promised years from now. The question is how much better, and whether either deal beats the simpler alternative of not investing at all. The internal rate of return, or IRR, answers both questions with a single percentage — but only if you know what that percentage assumes.

This guide explains what IRR actually measures, works through a hand-calculated example so the mechanics are transparent, shows how to apply the hurdle-rate decision rule, and then covers the two traps that make IRR mislead even careful users: the reinvestment assumption and the multiple-IRR problem.

What IRR Actually Measures​

The internal rate of return is the discount rate that makes the net present value (NPV) of a project's cash flows equal to zero. That definition is precise but abstract, so here is the intuition: IRR is the annualized growth rate your money would have to earn to make the project exactly break even in present-value terms. It is the project's implied interest rate — the rate at which the present value of everything flowing in exactly offsets everything flowing out.

Think of it as answering, "If this project were a bank account, what interest rate would it be paying me?" A project with an IRR of 12 percent behaves, in return terms, like money compounding at 12 percent a year. That framing is what makes IRR so popular: unlike NPV, which answers in dollars, IRR answers in a percentage you can compare directly against your loan rate, your required return, or another project's IRR.

The "internal" in the name matters. The rate comes entirely from the project's own cash flows — no outside interest rate is fed into the calculation. That is both the appeal and, as you will see, the source of most IRR mistakes.

A Worked Example You Can Verify​

Suppose you are considering a $10,000 packaging machine expected to save $3,000 a year in labor for four years. The cash flows are:

• Today: −$10,000
• End of year 1: +$3,000
• End of year 2: +$3,000
• End of year 3: +$3,000
• End of year 4: +$3,000

To find the IRR, you ask: at what discount rate does the present value of those four $3,000 inflows exactly equal the $10,000 outflow? There is no algebraic shortcut — you bracket it by trial and error.

At a 5 percent discount rate, the present value of the inflows is about $10,638, so NPV is +$638. Because NPV is positive, the true IRR must be higher than 5 percent (a higher discount rate shrinks present values). At 10 percent, the present value of the inflows is about $9,510, so NPV is −$490. Now NPV is negative, so the IRR sits between 5 and 10 percent.

Narrowing further: at 8 percent the inflows are worth about $9,936 (NPV −$64), and at 7.7 percent they are worth almost exactly $10,000. So the IRR is approximately 7.7 percent. The machine behaves like money earning 7.7 percent a year.

In practice you will never do this by hand. Every spreadsheet has an IRR function that takes a range of periodic cash flows, and an XIRR function for cash flows on irregular dates — use XIRR whenever your inflows do not land on a neat annual schedule, because the plain IRR function assumes equal periods and silently gives the wrong answer otherwise. But knowing the trial-and-error logic behind the function is what lets you sanity-check its output instead of trusting a black box.

The Hurdle-Rate Decision Rule​

A bare IRR number means nothing until you compare it against your hurdle rate — the minimum return you require, usually your cost of capital. The rule is simple:

• If IRR exceeds your hurdle rate, the project creates value. Accept it.
• If IRR falls below your hurdle rate, the project destroys value. Reject it.
• If they are equal, the project exactly breaks even in present-value terms.

For the packaging machine at 7.7 percent IRR: if you would finance it with a 5 percent equipment loan, it clears the hurdle and creates value. If your only funding is a 12 percent line of credit, it fails — you would be borrowing at 12 percent to earn 7.7 percent, locking in a loss with every payment.

Choosing the hurdle rate deserves care. A common choice is the interest rate on the debt that would fund the project, possibly with a risk premium added for uncertain cash flows. Larger businesses use a weighted average of their debt and equity costs. Whatever you pick, apply it consistently — a hurdle rate you adjust project by project to bless deals you already like is not analysis, it is decoration.

One caution: the accept-or-reject rule works cleanly for a single project evaluated on its own. The moment you compare two projects against each other, IRR's ranking can disagree with NPV's — which brings us to the traps.

Trap 1: IRR Ignores Scale​

Consider two mutually exclusive uses of the same warehouse space. Project A needs $1,000 and returns $1,400 in one year — an IRR of 40 percent. Project B needs $100,000 and returns $115,000 in one year — an IRR of 15 percent. Ranked by IRR, A crushes B. But A makes you $400 while B makes you $15,000.

This is the scale problem: IRR is a rate, and rates say nothing about the dollars at stake. A 40 percent return on pocket change will never beat a 15 percent return on serious capital. Similarly, IRR can misrank projects with different timing — a short project with a high IRR can look better than a longer project that creates far more total value.

The fix is straightforward: use IRR for accept-or-reject decisions on individual projects, but use NPV — actual dollars of value created — to choose between mutually exclusive alternatives. When the two metrics disagree on ranking, NPV is the one that measures what ends up in your pocket.

Trap 2: The Reinvestment Assumption​

Here is the assumption buried inside every IRR calculation: all interim cash flows are reinvested at the IRR itself. When your packaging machine throws off $3,000 at the end of year 1, the 7.7 percent figure assumes you immediately put that $3,000 to work earning another 7.7 percent until year 4.

For ordinary projects that assumption is roughly harmless. But it becomes absurd for high-IRR projects. A marketing campaign with a 45 percent IRR implicitly assumes every dollar it generates mid-year gets reinvested at 45 percent — a rate you almost certainly cannot reproduce. The project looks better than it is because the math grades your future reinvestment decisions as brilliantly as the project itself.

The standard fix is the modified internal rate of return (MIRR). MIRR lets you name a realistic reinvestment rate — typically your cost of capital — and compounds interim inflows forward at that rate instead of at the IRR. The result is always a single, conservative figure that reflects what you can actually earn on cash freed up early. Most spreadsheets offer a MIRR function taking the cash flows plus a finance rate and a reinvestment rate. For any project with a strikingly high IRR, compute MIRR before you celebrate: if the story survives a realistic reinvestment assumption, it is real.

Trap 3: Multiple IRRs (or None at All)​

Standard projects have conventional cash flows: money goes out first, then comes back — one sign change, from negative to positive. But some projects change sign more than once. A mining operation might run: initial investment (negative), years of operating profit (positive), then a large site-cleanup cost at the end (negative again). A retrofit with a mid-life overhaul has the same shape.

Each extra sign change can produce an additional IRR. Mathematically, finding IRR means solving a polynomial, and a cash-flow stream with two sign changes can have two different discount rates that both zero out NPV — or, in some cases, no rate that does. Your spreadsheet will return one of them (whichever its iteration lands on) without warning you the other exists. Basing a decision on "the IRR" when there are two of them, say 5 percent and 38 percent, is meaningless.

There is no way to salvage IRR here — the metric itself is undefined for the decision you face. Discount the cash flows at your hurdle rate and read the NPV, or compute MIRR, which always returns a single value because it separates the financing rate from the reinvestment rate. Whenever a project's cash flows flip between outflow and inflow more than once, skip IRR entirely.

Tracking Realized vs. Projected Returns in Your Books​

IRR analysis happens before you spend the money, but its value compounds when you check your forecast against reality afterward. That requires recording each project's actual cash flows separately so you can compute a realized IRR a year later and learn whether your estimates run optimistic. Tagging project cash flows with dedicated subaccounts makes this a bookkeeping routine rather than a forensic exercise.

Say you bought the packaging machine above. The purchase and the labor savings it produces might be recorded like this:

2026-01-15 * "Packaging machine purchase" #project-packaging-machine
  Assets:Equipment:Packaging-Machine      10000.00 USD
  Assets:Checking                         -10000.00 USD
 
2026-12-31 * "Year 1 labor savings from packaging machine" #project-packaging-machine
  Expenses:Labor:Packaging                 -3000.00 USD
  Income:Cost-Savings:Packaging-Machine     3000.00 USD

With every project-tagged posting in one place, pulling the actual inflow series at year-end is a single query, and comparing realized IRR against the projected 7.7 percent tells you whether your savings estimates — and your next project's business case — deserve your trust. Projects whose realized returns chronically trail projections are usually misestimated in the same direction, and one honest post-mortem is worth a dozen prettier forecasts.

Common Mistakes Checklist​

Before you act on an IRR, run through these:

1. No hurdle rate. An IRR without a comparison rate is trivia. Always name your cost of capital first.
2. Ranking mutually exclusive projects by IRR. The highest rate can be the smallest pile of dollars. Rank by NPV.
3. Equal-period function on irregular cash flows. If inflows do not arrive on schedule, IRR misleads; use XIRR with actual dates.
4. Ignoring the reinvestment assumption. High IRRs flatter projects whose early cash cannot plausibly be reinvested that well. Cross-check with MIRR.
5. One sign change assumed, never verified. Scan the cash-flow series: if it flips between negative and positive more than once, IRR may be multiple or meaningless — use NPV.
6. Nominal vs. real mismatch. Comparing a nominal IRR (which includes inflation) against a real hurdle rate, or vice versa, shifts every decision. Keep both in the same terms.
7. Forgetting financing direction. For a borrowing decision the cash flows are reversed — money in first, payments after — and the rule flips: you want the IRR (your effective borrowing cost) below your hurdle, not above it.

Your Next Step​

Pull the last capital purchase or project you approved on gut feel or simple payback, reconstruct its actual cash flows from your books, and compute both its IRR and its NPV at your borrowing rate. If the two metrics tell different stories about whether it was worth doing, you have just found the gap this guide closes — and your next proposal gets both numbers before it gets your signature.