This document discusses the advantages and disadvantages of using the internal rate of return (IRR) versus net present value (NPV) for capital budgeting. While IRR and NPV often yield the same results, NPV is generally superior because it can properly handle situations where IRR cannot, such as projects with changing discount rates over time, multiple positive and negative cash flows, or an unknown discount rate. Specifically, in these situations IRR may produce inaccurate or multiple results, whereas NPV discounts each cash flow separately to account for varying conditions. Therefore, the document concludes that NPV is preferable for evaluating most real-world projects, though IRR remains popular due to its simplicity of reporting a single number.

1. Which is a better measure for capital budgeting, IRR or NPV?
In capital budgeting, there are a number of different approaches that can be used to
evaluate any given project, and each approach has its own distinct advantages and
disadvantages.
All other things being equal, using internal rate of return (IRR) and net present
value(NPV) measurements to evaluate projects often results in the same findings.
However, there are a number of projects for which using IRR is not as effective as
using NPV to discount cash flows. IRR's major limitation is also its greatest strength:
it uses one single discount rate to evaluate every investment.
Although using one discount rate simplifies matters, there are a number of situations
that cause problems for IRR. If an analyst is evaluating two projects, both of which
share a common discount rate, predictable cash flows, equal risk, and a shorter time
horizon, IRR will probably work. The catch is that discount rates usually change
substantially over time. For example, think about using the rate of return on a T-bill
in the last 20 years as a discount rate. One-year T-bills returned between 1% and
12% in the last 20 years, so clearly the discount rate is changing.
Without modification, IRR does not account for changing discount rates, so it's just
not adequate for longer-term projects with discount rates that are expected to vary.
(To learn more, read Taking Stock Of Discounted Cash Flow, Anything But Ordinary:
Calculating The Present And Future Value Of Annuities and Investors Need A Good
WACC.)
Another type of project for which a basic IRR calculation is ineffective is a project
with a mixture of multiple positive and negative cash flows. For example, consider a
project for which marketers must reinvent the style every couple of years to stay
current in a fickle, trendy niche market. If the project has cash flows of -$50,000 in
year one (initial capital outlay), returns of $115,000 in year two and costs of
$66,000 in year three because the marketing department needed to revise the look
of the project, a single IRR can't be used. Recall that IRR is the discount rate that
makes a project break even. If market conditions change over the years, this project
can have two or more IRRs, as seen below.

2. Thus, there are at least two solutions for IRR that make the equation equal to zero,
so there are multiple rates of return for the project that produce multiple IRRs. The
advantage to using the NPV method here is that NPV can handle multiple discount
rates without any problems. Each cash flow can be discounted separately from the
others.
Another situation that causes problems for users of the IRR method is when the
discount rate of a project is not known. In order for the IRR to be considered a valid
way to evaluate a project, it must be compared to a discount rate. If the IRR is
above the discount rate, the project is feasible; if it is below, the project is
considered infeasible. If a discount rate is not known, or cannot be applied to a
specific project for whatever reason, the IRR is of limited value. In cases like this,
the NPV method is superior. If a project's NPV is above zero, then it is considered to
be financially worthwhile.
So, why is the IRR method still commonly used in capital budgeting? Its popularity is
probably a direct result of its reporting simplicity. The NPV method is inherently
complex and requires assumptions at each stage - discount rate, likelihood of
receiving the cash payment, etc. The IRR method simplifies projects to a single
number that management can use to determine whether or not a project is
economically viable. The result is simple, but for any project that is long-term, that
has multiple cash flows at different discount rates, or that has uncertain cash flows in fact, for almost any project at all - simple IRR isn't good for much more than
presentation value.