Professional Documents
Culture Documents
IRR As A Financial Indicator
IRR As A Financial Indicator
IRR As A Financial Indicator
Miroslaw M. Hajdasinski
School of Engineering, Laurentian University,
Sudbury, Ontario, Canada
In their recent paper Tang and Tang (2003, pp. 69–78) revive a long-
standing controversy—net present value (NPV) versus internal rate of
return (IRR)—by characterizing the NPV as an economic indicator and
the IRR as a financial one. The paper implies that this distinction justifies
ranking financial alternatives by ranking their IRRs. In the current arti-
cle, it is argued that the direct IRR ranking does not necessarily provide
the same evaluation environment—and therefore a fair comparison—for
each alternative involved, and that the incremental ranking approach
is needed to remedy this shortcoming. The article also points out that
Tang and Tang’s numerical examples of simple projects with one sign
change in their cash flow patterns do not address the problem of multi-
ple IRRs, which consequently renders Tang and Tang’s ranking approach
dysfunctional. It is demonstrated that the concept of a true rate of return,
substituting for the non-performing IRR and applied in conjunction with
the incremental approach, provides an adequate tool for ranking mutually
exclusive projects or a project’s technical or financial alternatives.
In their recent paper, Tang and Tang (2003) join the decades-long
discussion on the merits and validity of the internal rate of return (IRR)
as a viable alternative to the net present value (NPV) project-evaluation
criterion. They argue that, notwithstanding the persistent criticism of the
IRR, which has been voiced primarily in academia (Brigham et al., 1994;
Hirshleifer, 1958; Rapp, 1980; Solomon, 1963), this criterion is still
sound and useful, provided that it is properly interpreted. To remedy the
185
P1: GIM
The Engineering Economist TJ1099-06 May 27, 2004 18:21
186 M. M. Hajdasinski
perceived problem of the IRR and the NPV being inadequately defined,
Tang and Tang offer their own definitions of these two criteria. According
to them, the “IRR gives the private investor’s point of view and the NPV
the society’s point of view.”
However, regardless of the point of view, the mathematics of the
NPV/IRR relationship remains the same, and so do the well-documented
problems resulting from this relationship (Fisher, 1930; Fleischer, 1966;
Hirshleifer, 1958; Mao, 1966). In what follows, it is demonstrated, using
Tang and Tang’s and other numerical examples, that a direct comparison
of the IRRs of various project-financing alternatives for the purpose
of ranking is not a recommended approach. The proper approach will
be revisited, and the limitations of the IRR applicability will again be
emphasized.
The considerations will be conducted with the assumption of certainty
and a perfect capital market with identical investment and borrowing
rates referred to simply as the market interest rate MARR—minimum
attractive rate of return for investment projects, and maximum attrac-
tive rate of return for borrowing projects. In their paper, Tang and Tang
implicitly embrace this assumption by adopting the same interest rate
for calculating the NPV. Also, for the sake of consistency with the for-
mat used in the above-mentioned paper, an annual discrete time period,
annual cash flows (CFs), and annual discrete interest rates expressed in
percentages are assumed without any loss of generality of the ensuing
considerations.
This CFP reflects the first, and basic, project financing alternative (PFA),
where an investment of $10,000 is financed in full by the equity capital,
producing an IRR of 23.38%.
The second PFA considered is created by leveraging Project (1) with
a market borrowing project:
which results in the following leveraged CFP (Tang & Tang, 2003,
Table 3, p. 72):
generating an IRR of 113.1%. While the IRRs of Projects (1), (3), and
(4) are increasing as the equity for project financing declines due to
financial leverage, the NPVs of all of these projects are identical, thus
indicating that all projects are equally attractive from the economic point
of view, as Tang and Tang refer to it.
Note, in this context, that a post-leverage CFP, like CFP (3), or (4), is
one that is composed by combining two individual projects: the original
unleveraged PFA (1) and a market borrowing project that reflects the ex-
ternal financing arrangements. The borrowing project’s NPV calculated
at the borrowing MARR is, by definition, equal to zero, since the market
interest rate of 10% used for discounting is the same as the rate of return
(ROR) of the borrowing project. As a result, the NPV of the combined
project, being the sum of the original, unleveraged CFP (1)’s NPV and
of the zero-NPV of the borrowing project, will always be identical with
the NPV of the original Project (1). This is true for any possible CFP of
the borrowing project (as any market borrowing project will, by defini-
tion, produce its NPV equal to 0), and is, therefore, evident without an
elaborate mathematical proof.
Notwithstanding the NPV identity of the three PFAs (1), (3), and (4),
Tang and Tang arrive at a conclusion, based on a direct comparison of the
three alternatives’ IRRs on equity, that the higher the degree of market
leverage in a PFA, the more desirable this alternative is from, again using
Tang and Tang’s terminology, the financial point of view.
However, when comparing two PFAs by means of the IRR, Tang and
Tang do not provide the same evaluation environment for both alter-
natives. For example, if PFA (1) were to be compared with PFA (3),
then it would obviously have to be assumed that the investor has the
necessary amount of $10,000 of equity that is required to finance the
alternative with the higher equity consumption—in this case, PFA (1).
P1: GIM
The Engineering Economist TJ1099-06 May 27, 2004 18:21
188 M. M. Hajdasinski
which is a mirror image of borrowing Project (2), and has the same
IRR of 10%. The case of a return pattern different from that defined by
investment Project (5) will be addressed in the next segment of this text.
Next, combining CFPs (3) and (5) yields a resultant CFP that is iden-
tical with CFP (1), thus producing the same IRR = 23.38%. As a con-
sequence, PFA (3) turns out to be IRR-equivalent with PFA (1), which
demonstrates that a direct comparison of two IRRs earned by the same
project on unequal equity investments may produce misleading results.
A similar conclusion would be reached if the same analysis were applied
to comparing PFA (1) with PFA (4), thus proving that, if $10,000 of eq-
uity is available, all three PFAs are equivalent—to use Tang and Tang’s
nomenclature and ranking approach—not only from the economic, but
also from the financial, point of view.
Further, if the investor did not have $10,000, but only $4,000 of eq-
uity capital available for the project, PFA (1) would become financially
infeasible, and therefore irrelevant, and only the feasible alternatives (3)
and (4) could be meaningfully compared. To rank them in the manner
P1: GIM
The Engineering Economist TJ1099-06 May 27, 2004 18:21
demonstrated above will require market borrowing of $6,000 for PFA (3),
and of $9,000 for PFA (4). This would leave PFA (4) with $3,000 of free
equity that, if invested in the market, would produce a market investment
project:
190 M. M. Hajdasinski
defined by CFPs (1) and (3). The incremental CFP [(1)–(3)], exempli-
fying the difference between PFAs (1) and (3), turns out to be identical
with CFP (5). Since the IRR of this CFP is equal exactly to the MARR
= 10%, this indicates that neither of the two compared Projects, (1) and
(3), benefits from the hypothetical incremental project; hence, neither is
dominant over the other. Comparing now, in the same fashion, Projects
(1) and (4), an identical result is obtained, implying that Projects (1),
(3), and (4) are equivalent, thus confirming the NPV-based ranking.
The incremental approach can be applied to ranking mutually ex-
clusive alternatives of the same project or to ranking different mutually
exclusive projects. By evaluating the differences between two competing
projects, the incremental approach allows the comparison of projects, or
project alternatives, with different CFPs. A special class of such CFPs
are those having identical investments because, according to popular
belief, projects, or alternatives of projects, with such CFPs can be ranked
by directly ranking their IRRs.
To show that this belief is a misconception, let us revisit the com-
parison of PFAs (1) and (3), along with the complementary market in-
vestment of $6,000 of the equity that was earlier “liberated” through the
leverage of Project (1) and was assumed to produce CFP (5). However,
this assumption, although entirely legitimate, represented only one of
an infinite number of possible market return patterns. Let us consider
two additional, somewhat extreme, albeit possible, market investment
patterns reflected by the following CFPs (7) and (8):
Combining each of these CFPs with CFP (3) results in the following
CFPs (9) and (10), respectively:
yielding an IRR of 27.58%. All three CFPs, (1), (9), and (10), have an
identical investment of $10,000, and if the IRRs of CFPs (9) and (10)
were now to be directly compared with the IRR = 23.38% of CFP (1), this
would suggest that different market return patterns on the $6,000 market
P1: GIM
The Engineering Economist TJ1099-06 May 27, 2004 18:21
investment can make PFA (3) either inferior to PFA (1), if the returns
are defined by CFP (7), or superior to it, if the returns are reflected by
CFP (8).
However, an application of the incremental approach shows that nei-
ther is true, as all three return patterns (5), (7), and (8) produce equivalent
CFPs (1), (9), and (10). Indeed, the incremental alternative reflecting the
difference between CFPs (1) and (9) is the borrowing project:
192 M. M. Hajdasinski
will emerge, defining a TRR of 18.29%. This rate is smaller than any
of the IRRs, 23.38, 20.23, and 27.58%, of Projects (1), (9), and (10),
respectively, because the reinvestment rate of 10% is smaller than each
of the three IRRs. The above TRR is identical for all three projects
because, in this case, it is a genuine ROR earned on those projects’
identical investments, rather than on their unrecovered balances.
It should also be noted that the identical NPVs of CFPs (1), (9), and
(10) will not change when their intermediate returns are reinvested for
the purpose of the TRR calculation. Consequently, the transformed CFP
(13) has the same NPV as the three CFPs. This is due to the fact that the
reinvestment rate is the same one that is used for calculating the NPV.
It has already been pointed out that an IRR itself is actually not the
interest rate that it is commonly assumed to be, and that the practice of
ranking PFAs by directly comparing their IRRs cannot be conceptually
justified. In addition, a comparison of the IRRs of two PFAs can easily
run into serious interpretational problems if the original, fully equity-
financed projects are more diverse than those used in Tang and Tang’s
simple numerical examples.
To illustrate, consider the following equity-financed project:
Because the NPV, at the MARR = 10%, of borrowing Project (15) is,
by definition, equal to 0, the NPVs of Projects (14) and (16) are identical,
and equal to $159.83. Furthermore, CFP (16), similar to the previous CFP
(14), also produces three IRRs: −28.91%, 8.21%, and 3370.70%, which
again calls for a sound approach to the determination of Project (16)’s
representative IRR indicator. Using the Tang and Tang methodology, this
financial indicator could then be compared with a similar one selected
or derived from the three IRRs of Project (14). Unfortunately, Tang
and Tang do not address in their considerations the so-far-unresolved
problem of the financial interpretation of multiple IRRs, which has,
over the years, generated strong interest among researchers (Arrow &
Levhari, 1969; Cannaday, Colwell, & Paley, 1986; Flemming & Wright,
1971; Hajdasinski, 1987; Hazen, 2003; Jean, 1968; Kirshenbaum, 1964;
Lorie & Savage, 1955).
In contrast with the indeterminable representative IRRs of CFPs (14)
and (16), the unique TRRs of those CFPs can easily be determined.
Choosing the external rate of return (ERR) (White, Case, & Agee, 1976)
as an NPV-compatible TRR concept, the transformed CFPs derived from
CFPs (14) and (16) are obtained by reinvesting in the market, at the
MARR = 10%, each CFP’s year-two CF until the end of the CFP’s
lifetime.
This leads to the following two transformed CFPs:
194 M. M. Hajdasinski
generated from CFP (16). While the NPVs of the two CFPs are, for
reasons explained earlier, identical ($159.83), and are the same as the
NPVs of their parent CFPs, the TRRs produced by CFPs (17) and (18)
are different from one another and equal to 10.26% and 10.35%, re-
spectively. If compared directly, they would indicate the superiority of
the leveraged Project (18), and thus contradict the outcome of the NPV
evaluation. Initially, this may look surprising, since a TRR, unlike an
IRR, does represent an actual ROR on investment. However, one also
needs to realize that the investment patterns in Projects (17) and (18) are
quite different, which explains the differences in their respective returns.
To properly rank Projects (17) and (18) by means of the TRR criterion
(or any other NPV-compatible one), the incremental approach is called
for. Defining the generic differences between CFPs (17) and (18) leads
to the following incremental CFP:
which, after reinvesting CF 500 for one year, will change into a trans-
formed CFP:
Project (20)’s ROR of 10%, tantamount with the TRR of Project (19),
is the same as the MARR, which indicates that neither of the two Projects,
(17) and (18), dominates the other. This judgment is also supported by
the zero-valued NPV of Project (20), which demonstrates again that
ranking mutually exclusive projects by directly comparing their TRRs is,
generally, not a sound approach. The only exception when this approach
can be applied in the NPV-compatible fashion is when the transformed
projects to be ranked exhibit identical patterns of investments in the
case of ranking investment projects, or identical patterns of borrowings,
if borrowing projects are being ranked. In all other cases the incremental
approach should be applied.
the discrete interest rate i tends toward a finite value. In reality, however,
this function approaches a vertical asymptote located at i = −100%,
tending either toward −∞, if the last non-zero CF in the project’s CFP
is negative, or toward +∞, if that CF has a positive sign.
The misleading NPV-function profiles also suggest that the NPV-
function tends toward −∞ as i tends toward +∞. In reality, however, as
i tends toward +∞, the NPV-function approaches a horizontal asymp-
tote that is defined by the value of the time-zero CF in the project’s
CFP. Using “free-style” graphs for conceptual illustration purposes is a
common, helpful, and accepted practice; however, such graphs should
always reflect the basic features of the functions involved.
CONCLUSION
The above comments point out that, regardless of whether the IRR
is interpreted as a financial or economic indicator, its generic defini-
tion remains the same; hence, the same must also remain the IRR-
based project-evaluation, and project-ranking principles, whatever this
criterion’s intepretation. Accordingly, the ranking of projects, or their
financial alternatives, by ranking their IRRs is, generally, conceptually
unsound. The correct approach to ranking competing mutually exclusive
alternatives, using the NPV, or any NPV-compatible evaluation criterion
(Hajdasinski, 1993), including the IRR, is the incremental approach.
Other issues associated with the use of the IRR as an economic or
financial evaluation tool are its actual interpretation, and its ambiguity
when a project produces multiple IRRs within the economically mean-
ingful range (−100%, +∞) of the discrete interest rate. In this context,
it is reiterated that, contrary to its popular interpretation, the IRR does
not define the ROR on a project’s investment, but rather one on this in-
vestment’s unrecovered balance (Bernhard, 1962; Brigham et al., 1994).
To circumvent the interpretational and operational problems caused
by the IRR, several ROR indices have been proposed in the technical
literature to reflect a project’s TRR. One of them, the ERR (White, Case,
& Agee, 1976), was chosen to demonstrate its IRR-like application. It
was also emphasized that, except for special cases, the ranking of mutu-
ally exclusive projects by means of this NPV-compatible IRR substitute
should be done by means of the incremental approach.
REFERENCES
Adler, M., “The true rate of return and the reinvestment rate.” The Engineering
Economist, Vol. 15, No. 3, 1970, pp. 185–187.
P1: GIM
The Engineering Economist TJ1099-06 May 27, 2004 18:21
196 M. M. Hajdasinski
Arrow, J.K. and D. Levhari, “Uniqueness of the internal rate of return with variable life
investment.” Economic Journal, Vol. 79, 1969, pp. 560–566.
Baldwin, R.H., “How to assess investment proposals.” Harvard Business Review, Vol. 37,
No. 3, 1959, pp. 98–99.
Bernhard, R.H., “Discount methods for expenditure evaluation—A clarification of their
assumptions.” The Journal of Industrial Engineering,” Vol. 18, No. 1, 1962, pp. 19–
27.
Brigham, E.F., A.L. Kahl, W.F. Rentz, and L.C. Gapenski, Canadian Financial Man-
agement, 4th ed. Toronto: Harcourt Brace & Co., 1994.
Bussey, L.E., The Economic Analysis of Industrial Projects. Englewood Cliffs, N.J.:
Prentice-Hall, 1978.
Cannaday, R.E., P.F. Colwell, and H. Paley, “Relevant and irrelevant internal rates of
return.” The Engineering Economist, Vol. 32, No. 1, 1986, pp. 17–38.
Fisher, I., The Theory of Interest. New York: Macmillan, 1930.
Fleischer, G.A., “Two major issues associated with the rate of return method for cap-
ital allocation: “The Ranking Error” and “Preliminary Selection.” The Journal of
Industrial Engineering, Vol. 17, No. 4, 1966, pp. 202–208.
Flemming, J.S. and J.F. Wright, “Uniqueness of internal rate of return: A generalization.”
Economic Journal, Vol. 81, 1971, pp. 256–263.
Hajdasinski, M.M., “On relevant and irrelevant internal rates of return.” The Engineering
Economist, Vol. 32, No. 4, 1987, pp. 347–353.
Hajdasinski, M.M., “Payback period as a measure of profitability and liquidity.” The
Engineering Economist, Vol. 38, No. 3, 1993, pp. 177–191.
Hajdasinski, M.M., Comments on “Using heuristics to evaluate projects: The case of
ranking projects by IRR.” The Engineering Economist, Vol. 42, No. 2, 1997, pp. 163–
166.
Hazen, G.B., “A new perspective on multiple internal rates of return.” The Engineering
Economist, Vol. 48, No. 1, 2003, pp. 31–51.
Hirshleifer, J., “On the theory of optimal investment decisions.” The Journal of Political
Economy, Vol. 66, No. 4, 1958, pp. 329–352.
Jean, W.H., “On multiple rates of return.” The Journal of Finance, Vol. 23, No. 1, 1968,
pp. 187–191.
Kirshenbaum, P.S., “A resolution of the multiple rate-of-return paradox.” The Engineer-
ing Economist, Vol. 10, No. 1, 1964, pp. 11–16.
Lorie, J.H. and L.J. Savage, “Three problems in rationing capital.” Journal of Business,
Vol. 28, 1955, pp. 229–239.
Mao, J.C.T., “The internal rate of return as a ranking criterion.” The Engineering
Economist, Vol. 11, No. 1, 1966, pp. 1–13.
National Association of Accountants. Return on Capital as a Guide to Managerial
Decisions, Research Report No. 35, New York, 1959, p. 58.
Rapp, B., “The internal rate of return—A critical study.” Engineering Costs and Pro-
duction Economics, Vol. 5, 1980, pp. 43–52.
Solomon, E., “The arithmetics of capital budgeting decisions.” Journal of Business,
Vol. 29, No. 2, 1956, pp. 124–129.
Solomon, E., The Theory of Financial Management, New York: Columbia University
Press, 1963.
P1: GIM
The Engineering Economist TJ1099-06 May 27, 2004 18:21
Tang, S.L. and H.J. Tang, “The variable financial indicator IRR and the constant eco-
nomic indicator NPV.” The Engineering Economist, Vol. 48, No. 1, 2003, pp. 69–78.
White, J.A., K.E. Case, and M.H. Agee, “Rate of return: An explicit reinvestment rate ap-
proach.” Proceedings of the 27th Annual Conference of the AIIE, St. Louis, Missouri,
May 18–21, 1976, pp. 47–53.
BIOGRAPHICAL SKETCH
MIROSLAW M. HAJDASINSKI holds a B.Sc. and an M.Sc. in mining engineering from the
Technical University of Mining and Metallurgy (AGH) in Cracow, Poland. He earned
his Ph.D. in mine economics from the same university. Subsequently, Dr. Hajdasinski
conducted research at the Technical University Clausthal, Germany (Department of
Mining Engineering and Mine Economics), and at the Technical University Eindhoven,
The Netherlands (Department of Industrial Engineering and Management Science).
Late in 1982, he joined Laurentian University of Sudbury, Ontario, Canada, where he
currently works as a professor in the School of Engineering. Dr. Hajdasinski’s research
interests embrace project evaluation techniques and applications of engineering economy
and operations research to mine planning and evaluation, as well as to mine-design
optimization. He has published numerous articles in international journals and presented
papers at national and international conferences on the above subjects. Dr. Hajdasinski
is a professional engineer registered in the Province of Ontario.