When two competing investments both look worthwhile but you can only pick one, the choice between NPV and IRR for mutually exclusive projects almost always resolves in favor of net present value. Pick the project with the higher NPV. Internal rate of return is useful as a sanity check and a communication tool, but when the two metrics point to different projects, NPV is the one that reliably maximizes firm value. The rest of the analysis is really about understanding why they disagree so you can explain the choice.
What Mutually Exclusive Actually Means
Projects are mutually exclusive when accepting one eliminates the other. The constraint might be physical, like a single buildable lot that can hold a warehouse or a manufacturing facility but not both. It might be budgetary, where a board authorizes $500,000 for an equipment upgrade and two competing systems each cost $400,000. It might be strategic, where two product lines would cannibalize each other and only one gets launched.
The important point is that both options may be individually attractive. Each can have a positive NPV and an IRR above the cost of capital. The question isn’t screening out losers. It’s ranking winners. That distinction is what exposes the difference between the two metrics, because a screening tool only has to answer yes or no, while a ranking tool has to get the order right.
How the Two Metrics Rank Projects
NPV discounts each future cash flow back to today at the firm’s cost of capital, sums those present values, and subtracts the initial investment. The result is a dollar figure representing the value the project adds above the return required on the capital it uses. Between two mutually exclusive projects, the rule is simply: take the higher NPV. If Project A comes in at $150,000 and Project B at $125,000, A wins because it adds $25,000 more to firm value. The size of the initial outlay doesn’t enter the decision directly. A $2 million project with an NPV of $200,000 beats a $500,000 project with an NPV of $180,000, even though the smaller one looks more efficient per dollar.
IRR is the discount rate that drives a project’s NPV to zero. It’s an annualized percentage yield. For a standalone accept-or-reject call, the rule is clean: if IRR exceeds the cost of capital, the project earns more than its financing costs. For mutually exclusive projects, the instinct is to grab the higher IRR. That instinct works often enough to be dangerous, because a percentage return says nothing about how many dollars are actually being created. When the two projects differ in size or in the timing of their cash flows, IRR and NPV can rank them in opposite orders.
Why the Two Methods Disagree
The conflict traces to one of two causes, sometimes both.
Differences in Scale
Suppose Project A costs $100,000 and returns $140,000 in one year, while Project B costs $1,000,000 and returns $1,250,000. A’s IRR is 40%; B’s is 25%. IRR picks A. But at a 10% cost of capital, A’s NPV is about $27,300 and B’s is roughly $136,400. NPV picks B, which creates roughly five times more wealth. A high percentage return on a small base doesn’t compete with a moderate return on a much larger base when the goal is dollars added.
Differences in Timing
Even when two projects have identical initial costs, the shape of their cash flow streams can flip the ranking. A project that delivers most of its cash early tends to post a higher IRR, because the math rewards quick payback. A project weighted toward later years may show a lower IRR but a higher NPV, since those larger future flows still add up to more total value once discounted. The effect gets stronger when the cost of capital is low, because distant cash flows lose less value at modest discount rates.
The Crossover Rate
The crossover rate is the discount rate at which both projects produce the same NPV. Below it, one project has the higher NPV; above it, the other does. You find the crossover by taking the difference in each year’s cash flow between the two projects and computing the IRR of that incremental stream. If your firm’s cost of capital sits below the crossover rate, the project with heavier late cash flows usually wins on NPV. If it sits above, the project with faster payback usually wins on both metrics, and the conflict disappears.
Plotting each project’s NPV against a range of discount rates makes this visible. The two curves intersect at the crossover rate, and your firm’s actual cost of capital tells you which project dominates. It’s one of the more useful diagnostics in capital budgeting and one of the least used in practice.
Why NPV Gets the Final Word
When the two metrics conflict, financial theory sides with NPV, and the reasoning comes down to what each formula implicitly assumes about intermediate cash flows.
IRR calculates a rate of return that effectively treats every dollar of cash thrown off during a project’s life as if it can be reinvested at that same internal rate. If a project’s IRR is 30%, the math behaves as though interim cash flows earn 30% until the project ends. For most firms that’s unrealistic. A company with a 9% cost of capital is unlikely to find reinvestment opportunities yielding 30%. NPV discounts everything at the cost of capital, which reflects the firm’s actual opportunity cost. Some academics push back that neither formula literally contains a reinvestment assumption in its mathematics, but the practical effect stands: IRR tends to overstate the appeal of projects with high internal rates, and NPV gives a more grounded number.
The second reason is denomination. NPV is measured in dollars, and the point of capital budgeting is to maximize shareholder wealth in dollar terms. A 15% return on $10 million creates more value than a 25% return on $1 million. NPV shows that difference. IRR hides it.
When IRR Breaks Down Entirely
Standard IRR assumes a conventional cash flow pattern: an initial outlay followed by positive inflows. When a project has negative cash flows partway through or at the end, such as a major overhaul expense in year three or environmental cleanup at termination, the IRR equation can produce two or more mathematically valid answers. A project might simultaneously carry IRRs of 8% and 42%, with neither one more correct than the other.
The cause is mathematical. The IRR equation is a polynomial, and polynomials with sign changes in the cash flow stream can have multiple roots. When that happens, the metric is unusable for that project, and any attempt to compare it to another project’s IRR is meaningless. NPV has no such trouble. However often cash flows switch direction, NPV produces one unambiguous dollar figure at any given discount rate.
The Modified Internal Rate of Return
If you want a percentage-based metric without IRR’s reinvestment problem, the modified internal rate of return is the standard fix. MIRR compounds all positive cash flows forward to the end of the project at a specified reinvestment rate, usually the firm’s cost of capital, and discounts all negative cash flows back to today at the financing rate. It then solves for the single rate that links those two figures over the project’s life.
Because MIRR uses an externally set reinvestment rate rather than the project’s own return, it produces lower and more conservative numbers than IRR. A project with an IRR of 25% might show a MIRR of 16% when the reinvestment rate is set at a 9% cost of capital. That 16% is closer to what the firm will actually earn once you account for what interim cash can realistically do.
MIRR also solves the multiple-IRR problem, since its structure eliminates sign changes and always yields a single answer. For ranking mutually exclusive projects, MIRR agrees with NPV more often than plain IRR does. NPV still breaks the tie when they diverge.
Profitability Index and Capital Rationing
One boundary worth naming: the profitability index (PI) solves a different problem. PI divides the present value of future cash flows by the initial investment. A PI of 1.3 means every invested dollar produces $1.30 of present value. Values above 1 create value, values below 1 destroy it, and a PI of exactly 1 breaks even.
PI earns its keep when capital is rationed and several positive-NPV projects compete for a fixed budget. Rank by PI, fund top down, and stop when the money runs out. That approach maximizes total value per dollar deployed. But PI shares IRR’s weakness of ignoring absolute scale. A small project with a high PI can outrank a large project with a lower PI even though the large one adds more total wealth. When only one of two mutually exclusive projects can proceed and capital isn’t the binding constraint, PI isn’t the tool for the job. NPV is.