See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/24127736

The Internal Rate of Return and Project Financing

Article · August 2002

Source: RePEc

CITATIONS READS

0 3,552

2 authors:

Jean-Baptiste Lesourd Eph Clark

Aix-Marseille Université 176 PUBLICATIONS   1,484 CITATIONS   
47 PUBLICATIONS   541 CITATIONS   
SEE PROFILE
SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Securities bubble and banking context View project

An Alternative Measure of The world Market Portfolio View project

All content following this page was uploaded by Eph Clark on 14 May 2016.

The user has requested enhancement of the downloaded file.

 The Internal Rate of Return
under Equity-Related Hypotheses
A New Framework for Definition, Uniqueness and Consistency

J.B. Lesourd
GREQAM-CNRS, Université de la Méditerranée, Aix-Marseille, France
E-mail : [email protected]

and

E. Clark
University of Middlesex Business School, London, UK
E-mail : [email protected]

January 2003

1
1. Introduction

A number of techniques for the appraisal of investment projects are available (Brealey and
Myers, 2000; Luenberger, 1995). In particular, investment appraisal is generally carried out
through either the net present value (NPV) method, or the internal rate of return (IRR) method.
Both methods have strong theoretical foundations, whether for certain or uncertain future cash
flows. Simpler methods, with less theoretical background, are, however, used, including payout
time (POT) techniques, among others. The internal rate of return (IRR) was historically
introduced by Böhm-Bawerk (1889) and later by Keynes (1936). It rests on the fact that, for
given cah flows, the NPV can be considered as a function of the discount rate i. The calculation
of the IRR thus rests on determining a positive root (i > 0) of the equation NPV = 0. This root is
interpretable as a rate of profit, and may be compared to the rate of profit r that is prevails on
capital markets for similar investments.

As shown by number of authors, there are some obvious problems with the IRR method. In
particular, the equation NPV(v, i) = 0 (where v is the sequence of net cash flows at times 0, 1, 2
… t, … ,T) can have several acceptable roots, so that the IRR is not, in general, uniquely
defined. A number of works have thus been devoted to that problem. These works aim at
understanding the meaning of the set of IRR’s, if our equation has more than one acceptable
root. They include, in particular, the early paper of Lorie and Savage (1955), the theoretical
paper of Cantor and Lippman (1983), examining the IRR if imperfect capital markets prevail,
and some more recent theoretical contributions (Cantor and Lippman, 1996; Promislow and
Spring, 1996; Bidard, 1999). In another recent article, Volkman (1997) examines the
relationship between the IRR and negative cash flows related to the financing of a project, and
introduces a consistent modified IRR. The IRR and the rate of growth of the investor's wealth at
a given time horizon are equal.

However, inconsistencies with the IRR method stem from several problems. In particular, as
shown in many leading finance textbooks, such as Brealey and Myers (2000), there can be no
IRR’s, or multiple IRR's if NPV(v, i) = 0 has, respectively, no feasible roots, or has multiple
roots. This can happen with "nonnormal" cash flows, which can be thought of (Volkman, 1997)
as inordinately large negative cash flows other than the initial cash flow –F0.
As is shown in this paper, a consistent use of the IRR technique involves treating this
uniqueness problem by assuming an an investment project with a given time horizon T, under
rather reasonable “no free lunch” hypotheses, and other reasonable hypotheses whereby the
project’s equity is at all times strictly positive.
Our paper is organised as follows. It first treats the possibility of having several acceptable
roots for the equation that leads to the IRR. This possibility is ruled out by assuming "normal"
cash flows. Normal cash flows are defined as cash flows possibly including negative cash flows
that can be loan-financed in such a manner that in particular the investor's equity ratio in the
project remains positive at all times. We generalise our approach to the case in which the

2
interest rate term structure is not "flat". Thus, the IRR which conforms to our (reasonable)
axioms is unique and is equivalent to the rate of growth of the investor's terminal wealth.

2. Background
An investment which is carried out by a given investor can be characterized by the vector a of
successive cash flows at times 0,1,2 … t … T. Let us, in particular, define :

- u0 = -I0 < 0 as the net initial cash flow invested, or, in other terms, the part of the initial
investment which is equity-financed, net of the proceeds of any loan;
- IT> 0 is the terminal divestment, or the final value of the project, at which the sum of all
assets available (exclusive of the last cash flow uT) can effectively be sold at time T.
- The net profit cash flows, ie net of all cash flows due to loan reimbursements interest
payments and provisions u1, u2, … , ut, … , uT, some of which being possibly negative.

We thus define the vector v of all net cash flows as v = (-I0, u1, u2, … , ut, … , uT + IT), and the
NPV function as :
T T

  (1ui)
v t = – I+ IT
NPV(v, i) = v0 + t
 (1)
t 1 (1i) t t 1
t
(1i)T
The internal rate of return (IRR) method, very often used in investment appraisal, is derived
from the NPV method. An IRR is defined as a feasible root i (i > 0) of equation NPV(v, i) = 0.
Let  = 1/(1 + i) ( > 0), or i = G() = 1/ - 1. Defining B(v, ) = NPV(v, G()), (1) becomes:
T
B(v, ) = v0 + v  =
t 1
t t 0 (2)

B(u, ) = 0 is a polynomial equation of degree T in . This equation has in general T complex

roots: it can have no acceptable roots or more than one acceptable root. Let  > 0 be the largest
feasible lower bound for , corresponding to the practical lower bound for the discount rate; let
 the least feasible upper bound for , corresponding to the practical least upper bound for the
discount rate, corresponding to the IRR of an investment in cash assets, so that   This
is the interval in which the investor can practically expect to find the set of values of for all the
projects available to our investor.

One can find simple sufficient conditions for the existence of a unique acceptable IRR. These
T
conditions are well known (Luenberger, 1995): if v0 + v t 0 (the project is productive), v0 =
t 1

-I0, < 0 , vt > 0 for any t > 0 with at least one t such that vt > 0, and finally, if B(v, ) < 0, there
exists one unique root to the equation B(v, ) = 0.

Because of Descartes’ rule of signs, a necessary but not sufficient condition for the existence of
multiple roots is the occurrence of at least one change of sign in the coefficients (v1, v2, … , vt,

3
… , vT) of B(v, ), so that at least one cash flow ut (t > 2) is negative. This is not a purely
theoretical possibility : mining companies are generally submitted to the obligation of restoring
the environment of their sites whenever they are closed, implying expenses and negative cash
flows in mining investment projects after the closure of mining sites.

Under hypotheses weaker than the aforementioned conditions concerning the negative cash
flows (other than the initial investment - I0, = v0, we can show that there still is one unique IRR.
More precisely, we define normal cash flows as being such that the equity of the project,
defined as the sum of all discounted cash flows for , is strictly positive at all times.

Under normal cash flows, some of the vt’s can be negative, which encompasses as a particular
case the previous case in which all cash flows are nonnegative, and some other quite reasonable
assumptions, we can still prove uniqueness. Let us develop other basic concepts and
hypotheses.

3. Existence and uniqueness of the discounting factor and of the IRR

First of all, we assume that there is no free lunch or no gift, meaning that, if the investment
project starts at epoch zero, positive profit cash flows that are caused by the project can only
occur at later times. This means that the project cannot lead to a positive net DCF for
inordinately high IRR’s, and, in particular, IRR’s such that 0 <  < More precisely :

(H1) There is no free lunch: B(v, ) < 0

We shall see that this hypothesis, combined with other hypotheses, implies that B(v, 0) < 0, so
that I0, = -v0 > 0: the initial investment cash flow, which is also the initial equity, has to be
strictly positive : “no free lunch” implies that a significant nonzero part of the initial investment
has to be equity-financed.
As mentioned above, Descartes’ rule of signs imposes as a necessary condition to the existence
of multiple acceptable roots of (1) the existence of one or several negative coefficients (u < 0
for one or more positive values of  ( > 0)). These negative coefficients can be financed as
secondary investments that add to some initial primary investment. A reasonable hypothesis

4
concerning these secondary investments is that their net discounted cash flow equity is positive
at all times for some value of  such that More precisely :

(H2) Let N be the set of times such that such that u< 0. For any  one can find
some  t such that :
T

 u   I  0
 t  T
t t T T (3)
t 1

T
with: t >  and ut> 0  0 <  t < 1 and 
N

t < 1; t >  and ut< 0, or t <    t = 0;   = 1, so

that equivalently, for any :

T

 u   I 

u + 0
 t  T
t t T T (4)
t   1

In other terms, all negative cash flows such as v(v< 0, > 1) can be financed out of the
following positive cash flows at the IRR corresponding to , and this is compatible with an
overall equity of the project positive at all times.

A third hypothesis is that the project is productive. More precisely, this means that it is more
productive than an investment in cash assets. Assuming that cash assets yield a small
nonnegative IRR, this means that an investor in projects that yield more than this small IRR,
otherwise this investor would remain invested in cash. We are interested only in such projects.
This hypothesis can be formulated as follows :

(H3) The project is productive: there exists some  in a neighbourhood of 1 (  < ) such

that B(v,  ) > 0.

All these (reasonable and rather weak) hypotheses have several important implications that can
be proved as preliminary lemmas.

Proposition 1 : Under (H1) and (H2), the accounting equity at time t = 0, defined as I = -u0, and
T
the present values of the discounted cash flow equities, defined as u  I 
t'  t
t'
t'
T T at all

subsequent times t >0, are all strictly positive:

5
 I0 = -u0 > 0 (5)
T

u  I 
t'  t
t'
t'
T T 0 (6)

for   [, 1] and t >0.

T
PROOF: For any t > 0, and for  = , according to (H2), v   I 
t'  t
t'
t'
T T can be considered as the

sum of strictly positive functions. More precisely :

T T

v   I 
t'  t
t'
t'
T T = (t) +  (,t)
N
' (7)

where :

T
(t) = (1t')u t'  (1T )IT
t' T
(8)
t'  t N N

if  = inf ( N).

T
(t) =  u   I 
t 1

t t
t 
T T
T
(9)

For  > , all of these functions are either identically zero, or strictly increasing, so that (5) is
true for any   [, 1].
If   [0, ], noting that
T T
B(v, ) = v t t  ITt = (1) +  (,1) ' (10)
t 1 N

implies B(v, ) < 0, while (1) is strictly increasing, so, that (1) < (1). All of
the (1) are such that (1) > 0 and (1) = 0, so that again (1) < (1); therefore
B(v,1) = -I < 0 and I > 0, the initial accounting equity as well as the present values of the
discounted cash flow equities at all times after the beginning of the project (t > 0) are strictly
positive, QED.

A straightforward consequence of this is the following :

Proposition 2 : Under (H1), (H2) and (H3), there always exists at least one acceptable root
(ie such that   [,  ]) of the equation B(v, ) = 0, and thus always one acceptable IRR.

PROOF: Since B(u,.) is a continuous function, this is a trivial consequence of (H1) and (H3).

6
Note that our (rather reasonable) hypotheses completely exclude all cases in which no IRR is
defined, which would correspond to various “free lunches” (including, for instance, the case of
a zero initial equity (I = 0), in which the initial investment would be completely loan-financed,
with an infinite leverage effect and an infinite IRR.

We now come to our fundamental proposition, in which the uniqueness of  appears as a

necessary consequence of our hypotheses. Clearly, if i = G() = 1/ - 1, so that we can define

i i  G(  ); let us define v such that vt = ut for t < T and vT = uT + IT.

We now come to our fundamental proposition, in which the uniqueness of  appears as a

necessary consequence of our hypotheses. More precisely:

Proposition 3 : Under (H1), (H2), and (H3):

(1) There exists one unique root   to the equation B(v,) = 0, defining a function
 from V = {v B/   to   . is homogeneous of degree 0,
C- and strictly decreasing.
(2) There exists one unique root i to the equation NPV(v, i) = 0, defining a function F from
V to i i . F is homogeneous of degree 0, C- and strictly decreasing.

PROOF : According to proposition 1, the project’s DCF equity is strictly positive at all times t >
0. This may be expressed through the T following inequations :

u  I 
t 1
t
t
T T 0
T

u  I 
t 2
t
t
T T 0
…
…
u TT ITT 0

Summing up all these inequations, and dividing both sides by , one obtains :

TITT 1  B 0
T

tu 
t 1
t
t 1


7
This defines the set V = {v B/  . Therefore, owing to (H1) and (H3), B(v,
) < 0 and B(v,  ) > 0, so that, B(v, .) being continuous, for any v  V, B(v, ) = 0 has one
unique root in the closed interval   For each v  V, solving B(v, ) = 0 defines some
function such that =(v). Clearly, is homogeneous of degree 0, C- and strictly
decreasing. Therefore, part (1) of proposition 3 is established.
In a similar manner, if we define NPV(v,i) as NPV(v,i) = B(v,G-1(i)), one can also readily show
that there exists one unique root i to the equation NPV(v, i) = 0, defining a function F from V to
i i . F is homogeneous of degree 0, C- and strictly decreasing, so that part (2) of proposition
3 is established.

This encompasses as a particular case the case in which vt > 0 for any t > 0 with at least one t
such that vt > 0, but this is a weaker condition since it is compatible with vt < 0 for some values
of t (t>1).
Thus, i is unique and, under our hypotheses, equal to then rate of growth of the investor's wealth
at terminal time T, which is the investor's time horizon. Since :

T
B(v, ) = v0 +  v  =
t 1
t
t
0

Dividing by , one obtains :

T
-v0 / = I01+ r)= v 
t 1
t
t -T

and the property is proved.

4. Application to efficiency measurements and generalization to a variable

term structure of interest rates

Rather than comparing several investments, and due to the uniqueness property and to the
related monotonicity property, the IRR, or IRR-derivated properties, can be used in order to
assess the efficiency of an investment, by comparison with a standard investment that yields the
return prevailing on capital markets. Assuming the term structure of interest rates to be flat, let
i0 be that standard return, and let 0 = G(i0) be the corresponding discount factor. An efficiency
indicator is defined as a positive scalar indicating simply whether the investment is better, or
worse, than the above standard investment. A consistent set of properties that can be assumed
for an indicator defined in terms of v, E(v,i0) (see eg Robert Russell, 1990) would be the
following:

8
(I1) E(v, i0) = 1  i = i0 (or equivalently  = 0).
(I2) Monotonocity : E(v’, i0) > E(v, i0)  i = i0 (or equivalently i > i0).
(I3) Scale invariance and commensurability : whatever a > 0, E(av, i0)  E(v, i0).

(I1) attributes an indicator value of 1 to any project that yields an IRR equal to i 0. (I2) means
that any project that yields an IRR greater than i0 will be ascribed an indicator larger than 1 ;
conversely, any project that yields an IRR smaller than i0 will be ascribed an indicator smaller
than 1. Finally, (I3) makes the indicator independent of the scale of the project, and of the units
of measurement of the components of v (which are monetary units).

Still under the assumption that the term structure of interest rates is flat, a first indicator, which
is derived from the IRR, would = F(v)/i0. Let us show that this indicator, which is directly
derived from the IRR as a linear function, is satisfying the above properties.

Proposition 4 : The investment efficiency indicator (v,i0) = F(v)/i0 satisfies properties (I1),
(I2) and (I3) for any v  V.

PROOF : From the definition, it is trivial to see that (v,0) = 1 for = 0, hence (I1) is
verified; due to the fact that F is strictly increasing, and homogeneous of degree zero in v, and
invariant to the units of measurement of v, (.,i0) is increasing in v, hence (I2) and (I3) are
verified.

We now come to discussing the case of a variable rate of interest term structure. In this case, the
rate of interest varies with the maturity of the loan or investment. Let i0, i1, i2, … it … iT be the
rates of interest for, respectively, 0-period (the limiting value of the rate of interest for very
short periods of time), 1-period, 2-periods, … , t-periods, … , T-periods loans. Let 0 =1, 1, 2,
… t … T be the respective corresponding discounting factors, with t = G(it). In this case, the
NPV of a project may be expressed as follows, with i being the (T+1)-period vector i0, i1, i2, …
it … iT :

T
NPV(v, i) = v0 +  (1vi )
t 1
t

t
t
(11)

9
In a similar manner, in terms of v and of the (T+1)-dimensional vector 0 =1, 1, 2, … t …
T), one can express the NPV as :

T
B(v, ) = v0 + v ( )
t 1
t t
t (12)

The generalisation of the IRR approach is, in principle, possible, but not necessarily practically
feasible. In general, the equation B(v, ) = 0 has an infinity of solutions which, within the
primal v (T+1)-dimensional space define all the hyperplanes that contain point v. This is not
sufficient to define either i or , which are vectors, in a unique manner. In matrix form,
however, one could in principle determine a unique solution in ; this would imply having
available N projects (N >T) different projects characterised by vectors of cash flows vj = (v0j ,
v1j, v2j, … , vij, … , vTj). Let VN be the corresponding matrix with the (v1j, v2j, … , vij, … , vTj) as
row vectors. Let h be the transpose of (1, (2)2… , (t)t,… , (T)T), and w be the transpose of
(v01 , v02, … , v0n, … , v0N). In matrix form, our problem can be formulated as :

VN h = w (13)

Clearly, if VN is of rank T, one can extract from it a regular square matrix VT, so that (13)
becomes :

VT h = w (14)

which is equivalent to :

h = (VT )-1 w (15)

This scheme would give a unique solution and could theoretically be useful, especially in the
context of bonds; in this context, one could in principle select T bonds with T different
maturities, in which case matrix VT is a lower triangular matrix if the row vectors are ranked by
increasing maturities, and even a diagonal matrix if all these bond are no-coupon. However, this
scheme does not work in practice because it is difficult to select N bonds with exactly the T
maturities requested, and because of small inefficiencies in bond markets which lead to
impossible systems with practically observed VN matrices. Econometric and statistical methods
which are beyond the scope of this paper are preferable to apprehend the interest rate term

10
structure in the context of bonds. Furthermore, we have no possibility of comparing single
projects under that matrix scheme. Under the hypothesis that the interest rate term structure is
known by using one of these econometric methods, we thus propose a different approach,
which is close to the IRR approach and lends itself to comparison of individual projects. Let t0
t
= be the discount factor at time t, scaled with respect to its first component 1, so that 0=
1

(1, 10, 20… , t0,… , T0) is the scaled vector of discounting factors resulting from the
exogenous interest term structure. The components of this vector will all be equal to 1 if the
T
interest rate term structure is flatClearly, in general, the quantity v0 + v (
t 1
t t0 )t will not be

equal to 0. We are thus led to the introduction of a modified function B(v, ) :
T
B(v, ) = v0 + v (
t 1
t t0 )t (16)

In this equation, we define as a suitable factor, The question that we have to solve is,
therefore, whether, for individual existing projects, one can find one unique such that B(v,
) = 0. < 1 will indicate that the project is overall at least as profitable, or more profitable
than would result from the interest rate term structure observed on capital markets. Conversely,
> 1 will indicate that the project is overall less profitable than would result from the interest
rate term structure observed on capital markets. It is clear that will be identical to the
previous definition of if the interest rate term structure is flat. We are therefore led to
reasonable axioms that are a generalisation of our previous axioms.
In this case, let  [0 ,1] be a value of a practically feasible largest lower bond for . Let  ,

with  < 1, be the least feasible upper bound for , corresponding to the practical least upper

bound for , so that   A straightforward generalisation of (H1), (H2) and (H3)can
thus be proposed. More precisely, these generalisations can be formulated as:

(J1) There is no free lunch: B(v, ) < 0.

(J2) Let N be the set of times such that such that u< 0. For any  one can find
some  t such that :
T

 u (
t 1

t t t0 )t T IT(T0)T 0 (17)

11
 T
with: t >  and ut> 0  0 <  t < 1 and 
N

t < 1; t >  and ut< 0, or t <    t = 0;   = 1, so

that equivalently, for any :

T
u ( 0) +  u (
t   1

t t t0 )t T IT(T0)T 0 (18)

(J3) The project is productive: there exists some  in a neighbourhood of 1 (with   ) such

that B(v,  ) > 0.

We are now led to the following proposition :

Proposition 5 : Under (J1), (J2) and (J3) :

(1) The accounting equity at time t = 0, defined as I0 = -v0, and the present values of the
T
discounted cash flow equities, defined as u (
t'  t
t' ) IT(T0)T at all subsequent times t >0, are
' t'0 t'

all strictly positive:

I0 = -v0 > 0 (19)

T

u (
t'  t
t' t'0 )t' IT( T0)T 0 (20)

for   [,  ] and t >0.

(2) There always exists at least one acceptable root (ie such that  [,  ]) to the equation B(v,
) = 0.

which is a generalisation of both propositions 1 and and 2 can be readily proved in exactly the
same manner.

The generalisation of proposition 3 follows as :

Proposition 6 : Under (H1), (H2), and (H3):

(1) There exists one unique root   to the equation B(v, ) = 0, defining a function
 from V = {v B/   to   . is homogeneous of degree 0, C- and
strictly decreasing.  = G(i) is identical to the previous definition of , if i is the standard
IRR, when the interest rate term structure is flat.

12
(2) There exists one unique root i to the equation NPV(v, i ) = 0, defining a function F from
V to i i . F is homogeneous of degree 0, C- and strictly increasing. i = G-1() is identical
to the standard IRR if the interest rate term structure is flat.

Again, the proof of this proposition is identical to the proof of proposition 3.

Axioms (I1), (I2) and (I3) can be generalized in the context of a variable interest rate term
structure. With these definitions, the generalised axioms can be formulated as follows :

(GI1) E(v, i1 0)= 1  i = i (or equivalently 1).

(GI2) Monotonocity : E(v’, i1 0) > E(v, i1 0)  i > i1 (or equivalently 1).
(GI3) Scale invariance and commensurability : whatever a > 0, E(av, i1 0)  E(v, i1).

In the general case in which the term structure of interest rates is variable, an indicator, which is
derived from the IRR, would = F(v)/i1. Let us show that this indicator, which is equivalent to
the one directly derived from the IRR defined in proposition 4, is satisfying the above
properties.

Proposition 7 : The investment efficiency indicator (v,i10) = F(v 0)/i1 satisfies properties
(I1), (I2) and (I3) for any v  V. It is identical to the indicator defined in proposition 4 if the
interest rate term structure is flat.

The proof of this proposition is quite trivial and identical to the proof for proposition 4.

5. Conclusion

In this paper, we propose some “no free lunch” hypotheses concerning an investment that lead
to the definition of at least one acceptable IRR, thus ruling out the case in which no IRR is
defined. Other rather reasonable hypotheses exclude the existence of multiple IRR’s. In
particular, we define “normal cash flows” that lead to a strictly positive equity at all times and
to a unique IRR. We generalise our scheme to the case in which the interest rate term structure
is variable. We also define suitable efficiency indexes that describe the departure from the
efficiency of the project as an investment, as compared to an investment under the prevailing
interests on capital markets under comparable conditions.

13
References

Arditti, F.D., and Levy, H. (1977), The Weighted Average Cost of Capital as a Cutoff Rate : A
Critical Examination of the Classical Textbook Weighted Average, Financial Management
(Fall 1977), 24-34.
Bidard, C. (1999), Fixed capital and Internal Rate of Return, Journal of Mathematical
Economics, 31: 523–541.
Brealey, R. A. and Myers, S.C. (2000), Principles of Corporate Finance (6th edn.), New York:
Irwin-McGraw-Hill, pp. 98-108.
Von Böhm-Bawerk, E. (1889), Positive Theorie des Kapitals, Iena: Fischer.
Boudreaux, K.J., and H.W. Long (1979), The Weighted Average Cost of Capital as Cutoff
Rate : A Further Analysis, Financial Management (Summer 1979), 7-14.
Cantor, D.G. and Lippman, S.A. (1983), Optimal Investment Selection with Imperfect Capital
Markets, Econometrica, 51 (4): 1121-1144.
Cantor, D.G. and Lippman, S.A. (1995), Investment Selection with a Multitude of Projects,
Econometrica, 63 (5): 1231-1244.
Jacques, J.K., Lesourd, J.B. and Ruiz, J.M. (1988), Modern Applied Energy Conservation: New
Directions in Energy Conservation Management, Chichester/New York: Ellis
Horwood/Wiley.
Keynes, J.M. (1936), The General Theory of Employment, Interest and Money, (reprinted
1964), London: Macmillan, p. 140.
Lin, S.A. (1976), The Modified IRR and Investment criterion, The Engineering Economist, 21:
237-47.
Lorie, J.H. and Savage, L.J. (1955) Three problems in rationing capital, Journal of Business,
28: 229–39.
Luenberger, D. (1995) Investment Science, Oxford : Oxford University Press.
McDaniel, W.R., McCarthy, D.E., and Jessell, K.A. (1988), Discounted Cash Flow with
Explicit Reinvestment Rate: Tutorial and Extension, The Financial Review, 23: 359-85.
Moilanen, T. and Martin, C. (1996), Financial Evaluation of Environmental Investments,
Houston: Gulf Publishing Company.
Myers, S.C. (1974), Interactions of Corporate Financing and Investment Decisions –
Implications for Capital Budgeting, Journal of Finance, 29: 1-25.
Promislow, S.D. and Spring, D. (1996), Postulates for the internal rate of return of an
investment project, Journal of Mathematical Economics, 26: 325–61.
Robert Russell R. (1990) Continuity of Measures of Technical Efficiency, Journal of Economic
Theory, 51: 255-267.
Volkman, D.A. (1997), A consistent yield-based capital budgeting method, Journal of
Financial and Strategic Decisions, 10 (3): 75-88.

14
 15

View publication stats