This 2004 McKinsey Quarterly article reports that, back in 1999,
three-quarters of CFOs always or almost always use[d] IRR when evaluating capital projects.
The same article warns against the pitfalls of the internal rate of return (IRR). Although using the latter makes sense if the project only consists of an initial investment and a final cash inflow, things start to get hairy when the project includes interim cash flows:
- The definition of the IRR is not mathematically sound because, for such projects, there can be multiple IRR values. Some people say that, in that case, you should pick the smallest of those values (rationale?). However, Excel's IRR routine does not unconditionally converge to that smallest value, as it uses some rootfinding method that requires an initial guess, which influences the result returned.
- An underlying assumption of the IRR is that interim cash flows can and will be reinvested at the same rate as the IRR, which is rarely the case. This rather strong assumption introduces distortion that can
make bad projects look better and good ones look great.
For those reasons, the article recommends staying away from the IRR as much as possible and preferring other metrics such as the Net Present Value (NPV) and the Modified Internal Rate of Return (MIRR).
I don't work in the world of finance (I've got an engineering background), but the few finance guys who I rub shoulders with use the IRR seemingly without being aware of its limitations. Hence my questions (all related, really):
- Has the article's message sunk in since 2004?
- Or do (industry) people still routinely use IRR for capital budgeting?
- Do you know of more recent reports on the popularity (or lack thereof) of the IRR?