Uses, Abuses, and Alternatives to the Net-Present-Value Rule

Author(s): Stephen A. Ross

Source: Financial Management, Vol. 24, No. 3 (Autumn, 1995), pp. 96-102
Published by: Wiley on behalf of the Financial Management Association International
Stable URL: http://www.jstor.org/stable/3665561 .
Accessed: 08/09/2013 23:33
Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at .
http://www.jstor.org/page/info/about/policies/terms.jsp

.
JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of
content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms
of scholarship. For more information about JSTOR, please contact [email protected].

Wiley and Financial Management Association International are collaborating with JSTOR to digitize, preserve
and extend access to Financial Management.

http://www.jstor.org

This content downloaded from 142.103.160.110 on Sun, 8 Sep 2013 23:33:17 PM

All use subject to JSTOR Terms and Conditions

Issues
Contemporary

and
Alternatives to
Net-Present-Value
Rule

Uses,

Abuses,

the

StephenA. Ross
StephenA. Ross is SterlingProfessor of Economicsand Finance at YaleSchool of Management,Yale University,New Haven, CT.

No student can leave the introductoryfinance course

without having mastered the net-present-valuerule. It is
the meat of most textbooks and lies at the core of what
financialacademicsthinktheyhave to offerCFOs,corporate
treasurers, investment bankers, and practitioners of all
stripes.In fact, it is not uncommonto spend a considerable
amount of time in class making sure that the student
understands all the wrong ways of thinking about
investment decision making-from the IRR rule to the
payback period. Wrong, of course, because they don't
coincide with the NPV rule.
The simpleststatementof the NPV ruleis thatyou should
discard projects with negative NPVs and undertake all
projectswith positive NPVs. If we arebeing careful,we add
the caveat that a positive recommendationto take a project
shouldonly be made if takingon the projectdoesn't prevent
us from undertakingsome otherproject.
Like all good rules,this one containsmuchtruth,but I am
going to take the contraryview. I have become convinced
that it is time to revisit the usefulness of NPV and to
reconsiderjust how much stock we want to place in it.
Perhapsmost surprising,I will arguethe meritsof alternative
rules-modified versions of NPV-that seem to endurein
practicedespite theirconflict with the NPV rule. In general,
though, I believe that we now have superiorways to make
investmentdecisions.
Section I describes the problems with the way we
currentlyuse the NPV rule. This section-and much of this
paper--draws heavily on my work with Jon Ingersoll as
reportedin Ingersoll and Ross (1992). Section II offers a
simple example that shows the ubiquitous need for
alternativesor modifications to the NPV rule. Section III
M

This paper was the FMA Keynote Address at the 1994 FMA Annual
Meeting, October12, 1994. The authoris gratefulto Jon Ingersolland the
Editorsfor theirhelpfulcomments.All errorsarehis own.

examinessome of the adhoc rulesused in practiceandmakes

a case that they may not be as undesirable as accepted
wisdom holds. Section IV outlines a general approachto
making wise investmentdecisions as an alternativeto the
NPV rule andprovidesa practicalrule-of-thumbadjustment
to the NPV rule. Section V briefly summarizeswhat the
paperhas said on these matters.

I. The Good, the Bad, and the Ugly

of the NPV
A firm is considering a major investment. The project
involves the completionof a majorpower source,and it will
cost $100 million upfront. One year after making the
investment,the firm will be able to liquidateits stake in the
projectfor $110 million. The projectis big enough so that
the top managementwill make the decision.
The management are all graduates of the top
managementprograms, and they have a firm allegiance
to the utilizationof the best availablefinancialtheory in their
decision-making. They are certain that the project is
absolutelyriskless. The marketagreeswith theirassessment
and will finance the projectat riskless interestrates.
A quick look at the latest results from the bond market
reveals thatthe currentyield curveis flat at 10.3%,i.e., short
rates, long rates, and all intermediaterates are 10.3%. A
back-of-the-envelopecalculationrevealsthatthe projecthas
an NPV of approximatelynegative $300,000. Armed with
this information,managementrejectsthe investment.
Dejectedathavingto turndowntheproject,butconvinced
thatthey have made the rightdecision, managementis soon
to get some good news. An independentinvestorapproaches
them andoffers to buy the rightsto the project.When asked
why she would pay anything for a worthless project, the
investoranswersthatshe subscribesto the greaterfool theory

Financial Management,Vol. 24, No. 3, Autumn1995, pages 96-102.

This content downloaded from 142.103.160.110 on Sun, 8 Sep 2013 23:33:17 PM

All use subject to JSTOR Terms and Conditions

RULE
TO THENET-PRESENT-VALUE
ROSS / USES, ABUSES,ANDALTERNATIVES

and believes that the project can be "flipped" at a profit.

Managementis unconvincedbut is more than happy to sell
the rights to the project for $10,000 and move on to more
profitableopportunities.
In fact, the investor, true to her word, just a few
days later is able to flip the project for $30,000, netting
a quick $20,000 profit. The purchaser is a sharp real
estate developer who weatheredthe 1980s and with few
opportunities in development is casting about for other
activities.As luck would have it, one monthafterpurchasing
the projectthe one-yearinterestratesuddenlydrops50 basis
points, from 10.3%to 9.8%.Computingthe NPV at the new
lowered interest rate results in approximately$200,000.
Without a second thought, the ex-developer finances the
investmentcollateralizingthe loan with the proceeds from
the project and pockets about $200,000. Deducting the
$30,000 he paid the dealmaker, the developer nets a
$170,000 profitfrom a one-monthholding.

A. The Good:Rejectingan InvestmentWhenIt

ShouldBe Rejected
The $100 million project generates $110 million one
year later. With interest rates at 10.3%, the same $100
million grows to $110.3 million when investedin a one-year
discount bond. Investing in the projectis dominatedby the
opportunities available in the capital markets, and the
managers' finance training hasn't let them down-yet.
The key point to keep in mind is that capital market
alternatives are always freely available and undertaking
them does not alterthe set of alternativesthatare open to an
investor.Hence, if a projectis dominatedby a capitalmarket
alternative,then thereare no otherfinancingconsiderations
that wouldjustify takingon the dominatedproject.

B. The Bad:Rejectingan InvestmentWhenIt

ShouldBe Accepted
Selling the project for $10,000 was tantamount to
rejecting the investment. Even though the NPV of this
projectis negative, $10,000 seems like a bargainbasement
price for a $100 million project.Suppose that interestrates
were exactly 10%. Simply because the NPV calculation
yields a zero for NPV, does anyonereallybelieve the project
is worthless?Whatis wrong with the NPV analysis?
This project is more thanjust a one-time investment.It
also includes the rights to the investment.Simply because
currentinterest rates don't justify making the investment
doesn't mean thatthis will always be the case. Nor does the
fact that the yield curve is flat precludethe possibility that
interestrates could fall below 10%and bring the NPV into
the positive range. The rights to the projectare the rights to

97

the interest rate option inherent in the project. Any such

projecthas such rights.When the investmentis undertaken,
it will have positive NPV. Since nothingforces the holderto
take the project when the NPV is negative, it will only be
undertakenat positive NPVs. Thus, the holder profits from
declines in the one-year rate and has limited liability if
interestrates rise. This projectis equivalentto a call option
on a one-year bond. Simply because the option isn't
in-the-moneytoday doesn't mean thatit is worthless.

C. The Ugly:Acceptingan InvestmentWhenIt

ShouldBe Rejected
The subtlesterrorof all in the applicationof the NPV rule
is acceptingthe project,i.e., makingthe investment,simply
because the NPV is positive. What if the interestrate were
9.999%? Do we undertakethe investment to realize the
gain of $1,000? What is wrong with this application of
the NPV rule?
Actually,nothing.Whatis wrong is the generalfailureto
seriously consider the caveat to the NPV rule, namely,
undertakethe project as long as doing so doesn't interfere
with the ability to take on a competingproject. Undertaking
the project implies that we are not taking on the project
tomorrowor next week or at any otherfuturetime.
Every projectcompetes with itself delayed in time. This
is the essence of the problemwith applying the NPV rule,
and it is anotherway to understandoptionality.In a capital
budgeting context with a budget constraint,undertakinga
project means taking on that feasible combination of
projectsthat maximizes the NPV. Clearlywith interestrate
uncertainty,we tradeoff the value of taking on the project
today against the lost opportunitycost of foregoing the
option to undertakethe project at some later date when
interestratesare morefavorable.
This same reasoning can also resolve the problem of
rejecting the project when it should be accepted, i.e., of
selling the rightsto the projecttoo low. Selling the projectto
the dealmaker is not only selling today's project, it is
also the sale of all the potential futureprojects.Weighing
each such projectby the probability(suitablyrisk adjusted,
i.e., the martingale probability) and taking the expected
value we get the value of those futureprojects.

D. TakingOptionalityinto Account
Ingersoll and Ross (1992) use a specific process for the
dynamics of interest rate movements to develop an exact
formulafor the value of the projectwhen viewed with all of
its optionalityintact,and they determinethe interestrate at
which it is optimalto exercise the option and undertakethe
investment.Thereis no reasonto display the formulasfrom

This content downloaded from 142.103.160.110 on Sun, 8 Sep 2013 23:33:17 PM

All use subject to JSTOR Terms and Conditions

98

/ AUTUMN1995
FINANCIAL
MANAGEMENT

thatpaper;suffice it to say thatthereare at least two lessons

to be learnedfrom the exact analysis.
First, not surprisingly,for our projectthe interestrate at
which it is optimal to exercise is a rate below 10% at
which the project is well in-the-money and generates
adequate NPV to compensate for the loss of the option.
Second, naturally, the volatility of interest rates is a
significantdeterminantof the value of the projectand of the
cutoff rate at which the project is undertaken.Generally
speaking,the higher the volatility, the more valuable is the
optionalityand, therefore,the more valuable is the project
and the lower the optimalinterestrate at which it should be
exercised.
It is fascinatingthatthis, the simplestinvestmentexample
possible, contains so much in the way of optionality and
carriessuch a potentialfor misleading analysis. Of course,
projects are never this straightforward.An investment
that is not undertaken today cannot be warehoused
forever.Overtime it changes-often in uncertainways-and
aftera certainperiod,it may no longerbe available.But what
is the more realistic polar case? That of a project that
disappears altogether if it's not undertakenat this exact
moment,or one thatcan be undertakenwithoutalterationat
any time in the future?
Cast in this starklight, both appearto be extremes. But
given a choice, I believe the infinitely lived projectis much
the preferredcanonicalexample.As shown in Ingersolland
Ross (1992), most of the option value is developed in a
relatively short period of time. Since most projects simply
don't go away if they are not undertakenright now, the
perpetuityexample is a good approximationto reality.
There is a long literatureon investment projects with
embedded options. Almost all actual projects have
optionality inherent in the cash flows themselves. Often
investments open up the possibility of profitable future
opportunitiesto invest. These are important issues, but
they are relativelyclear conceptually,and they are separate
from the issues raised here. The value of any projectcomes
from three sources. First, from its in-the-money-value,
which is simply its NPV if the option were to be exercised
today. Second, fromthe value of the embeddedoptionsbuilt
into the project itself. We are talking about the third and
ubiquitoussource of value. Every project,whetheror not it
explicitly contains options, always is an option on the
movementof capitalcosts and prices.
In assessing investmentvalue and in makinginvestment
decisions, we must now recognize the impact of this
of projects.
sourceof value on even the most straightforward
As a practicalmatter,Ingersoll and Ross (1992) show that
this source of value is generally large enough to have a
significantimpact.

II. Another Example

The above reasoning extends to nearly all investment
decision-making;it certainlyloses none of its force when
uncertaintyis introduced.When cash flows are random,we
have learned to value them by applying a modified
version of the NPV rule. We take the expected cash flows,
apply a risk adjustment, and discount the resulting
certainty-equivalentflows. A not entirelyinaccuratevariant
discounts the expected cash flows at risk-adjustedcosts of
capital,but it is preferableto applythe risk adjustmentto the
cash flows using the martingaleor risk-adjustedexpectation.
Not all valuation problems with random cash flows,
though, requirethe full artilleryof derivativepricing.Ross
(1978) displayeda richclass of problemsthathadimmediate
and simple solutions. For example, suppose that T periods
fromthe initialinvestment-say $ 1-the payoff on a project
will be proportional to the value of some marketed
asset, S (where S is inclusive of reinvested dividends).
Typically, S could be a stock or a marketindex, and P will
be the proportionalityconstant.Thus if $1 is investedat time
t, the payoff will be St + T at time t + T.
Since the asset itself is equivalentto a claim on the value
St + T at time t + T, it follows without resort to any fancy
derivativepricinganalysisthatthe currentvalue of thepayoff
PSt +T is simply PSt. But, it is certainlynot the case thatthe
rightto undertakethis projecthas a currentvalue of PSt- 1.
Just as in the risk-freeexample of Section I, insofar as it is
possible to delay making the initial investment,the project
has an option value.
In fact, clearlythe currentvalue of this projectis simply
the value of a call option with an exercise price of 1 and a
maturityequal to the length of time that the projectcan be
delayed. Even thoughthe cash flows involve no optionality
in and of themselves,the abilityto delay confers optionality
on the project.As before, the projectvalue is enhancedby
the value of the option, which, in turn, derives from the
opportunityto profitfrom changingfuturevaluationsof the
project'spayoff.
It is worth noting that to some extent this problem of
neglected optionality is mitigated by the flow through of
inflationaryexpectationsto cash flows. In a simple Fisherian
world, the nominal discount rate is the real rate of interest
plus the expected inflation rate. If cash flows in the
futureare expected to increasewith inflation,then changes
in the inflationrate will leave the NPV unaltered.Thus, the
optionality associated with financing will only be with
respect to uncertaintyin the real rate of interest.Of course,
to the extentthatthe cash flows on a projectarenot perfectly
proportionalto the pricelevel thatis embodiedin the interest

This content downloaded from 142.103.160.110 on Sun, 8 Sep 2013 23:33:17 PM

All use subject to JSTOR Terms and Conditions

TOTHENET-PRESENT-VALUE
RULE
ROSS / USES, ABUSES,ANDALTERNATIVES

rate,the impactof uncertaintyin the interestratewill not be

limited solely to uncertaintyin real ratesof interest.

Ill. HurdleRates and Other Rules

Finance scholars have always been puzzled by the
durabilityof a host of investmentrules that seem to survive
and even thrive despite their obvious shortcomings.Ross,
Westerfield, and Jaffe (1993) reporton a numberof these,
including the payback period, the IRR, and the hurdle
rate rule. The latter requires an investment to not
merely have a positive NPV but to have a sufficiently
positive NPV. Often, this is put in the form of requiringthat
the IRR exceed the currentmarket rate of interest by an
additionalamount,say 3%.
Such rules are interesting for a number of reasons,
and not surprisingly,they have attractedmuch attention.
Asymmetric informationargumentsfor their existence are
probablythe most prevalentrationalizations.For example,
Antle and Eppen (1985) and Antle and Fellingham (1990)
have arguedforcefullythatthe incentivesof managerswithin
hierarchiesare such as to rewardthem for amassing more
control over corporateresources.These studies argue that
firms requireprojectsto satisfy hurdlerates in excess of the
interestcosts in the capitalmarketsto serveas a brakeon this
tendencyto overinvest.Alternatively,high hurdleratesmay
simply be a practicalway to deal with uncertainty.
These theoriesall have merit,butourlook at the NPV rule
suggests that anotherexplanationis at work. Figure 1 plots
the NPV of a projectalong with the option adjustedNPV, or
OANPV, which is the simple NPV plus the value of the
interest rate options inherent in the right to delay the
project. Since the project can be delayed indefinitely, the
OANPV is never zero no matter how large is the
interestrate and negative is the NPV. At an interest rate
equalto the OAIRR,the value of theserightsis zero because
the option is sufficiently in-the-money that the gain from
delaying to preserve the option is just offset by the loss
in value from not immediatelyexercisingit andrealizingthe
positive NPV. As can be seen, the OAIRR lies below
the IRR so as to insure that currentexercise is sufficiently
valuable.The differencebetweenthe IRR andthe OAIRRis
the extra amountthat must be addedto the currentinterest
rateto find the hurdlerate.The projectwill be undertakenif
the IRR exceeds this hurdle rate. The exact hurdle rate
adjustmentdepends on interest rate volatility (or cost of
capitalvolatility) and on the exact cash flow structureof the
project, but as a practical matter within wide families of
potential projects, some such adjustmentis sensible and
improves on the simple use of the NPV or the IRR rules. I
will explore this matterin greaterdetail in the next section.

99

It would be difficult to make a similar defense of the

payback period rule. In the simplest form of the
payback rule, an investment is accepted or rejected
depending on whether the cash flows add up to the initial
investmentby some specified time period, e.g., three years.
Typically the projectsto which it is appliedinvolve a single
upfrontinvestmentand a subsequentstreamof positive cash
flows. Typically, too, the rule manifestsitself in industries
like entertainment where projects can be delayed for
meaningfulperiods of time.
In such cases, often the firm has made infrastructure
investments that realize their returnthrough a stream of
projects with excess returns for all rates within broad
historicalranges. The limitationon the firm is not so much
capitalas it is the capacityto applylimitedreservesof human
andmanagerialcapital.While this is quite differentfromthe
sort of optionality we have analyzed, nonetheless it does
embody the same principlethat undertakingan investment
at any point in time is at the cost of foregoing the option to
undertake other projects that may become available. In
effect, while projectsarenotbeing delayedto takeadvantage
of lower interest rates, they may be put on hold to take
advantageof superioralternativeprojects.

IV. A GeneralAlternativeto the

NPVRule and a Simple Rule
of Thumb
The lesson from this analysis of the NPV rule is that we
must treat all investment problems as option valuation
problems. Consider a project with cash flows of c(t) over
time. Forease of exposition,we will assumethatthese flows
are deterministicand leave the extension to random cash
flows for later.
If P(t) denotes the currentvalue of a pure discountbond
paying $1 t years from now, then the NPV of a projectif it
is undertakentoday is given by

NPV =

p(t)c(t)dt

Ignoringthe possibility that the projectmay change if it

is delayed, the above NPV formula tells us what will be
realized at any time that the project is undertaken.The
decision to undertakethe project,then,is a decisionto realize
this NPV and to forego the opportunityto realize a higher
NPV if interestratesshould happento fall.
Evaluatingthis trade-offbetween the currentrealization
and the potential future value is a complex mathematical
problem. It is dealt with in detail in Ingersoll and Ross
(1992). In essence, though,if we are willing to make some

This content downloaded from 142.103.160.110 on Sun, 8 Sep 2013 23:33:17 PM

All use subject to JSTOR Terms and Conditions

100

FINANCIAL
MANAGEMENT
/ AUTUMN1995

Figure 1. NPV and OANPV

OANPV

OAIRR

IRR

Value
ofrightto delay

Interest
Rate

NPV

reasonableassumptionsaboutthe behaviorof interestrates,

then the solution to the problemcan be characterizedin an
intuitivelyappealingfashion.
Assume, first, that there is some rate, perhapsthe short
rateof interestor, maybe,the five-yearrate,thatcan be used
to characterizethe movementof interestrates.As an intuitive
matter,clearly the solution will be to undertakethe project
whenever this relevant interest rate falls below some
appropriatehurdlerate, r*. Letting, IRR denote the internal
rate of returnfor the project,then it's clear that an optimal
choice of r*will be less thanIRR. In effect, we will demand
thatthe projectbe in-the-moneyto give us enoughadditional
value from undertakingit to offset the lost opportunityof
waiting for a possible dropin interestrates.
For any choice of r*, the formula for the value of the
project, including the option to delay choosing it, will be
given by the expected discounted value over all possible
future interestrate paths. If we let NPV(r*)denote the net
presentvalue of the projectwhen the interestrate is r*, then
we can computethe value of the projectincludingthe value
of the option to delay by following a simulationprocedure.
First, we simulatea futureinterestratescenario.Second, we
markthe first time in this scenariothatthe interestrate falls
to r*. Third, we compute the discountedvalue of NPV(r*)
along the interestratepathfor this scenarioup to the marked

time. In other words, if the marked time is t, then we

computethe pathcontribution,
fr(s)ds

e- S

NPV(r*)

If the pathnever crosses r*, then its contributionis zero.

We repeatthis process n times and averagethe resulting
contributions,contribution(j):
(1)I contribution(j)
This sum is an approximation to the exact expected
discountednet presentvalue of the projectwith the exercise
rule, r*,
r(s)ds

r(s)ds

VALUE = E(e1-4
NPV(r*))= NPV(r*)E(e1-f)
whereI is therandomtime at whichthe interestratepathfirst
hits r*.
While the above procedureis appropriatefor large and
complex projects,it is useful to have an approximationfor
simpler decisions. For simple point input and point output
projects, Ingersoll and Ross (1992) developed an analytic
formulafor the value includingthe optionto wait. Using this
formula,an approximationis availablefor determiningthe
cutoff hurdlerate at which a more generalprojectshouldbe

This content downloaded from 142.103.160.110 on Sun, 8 Sep 2013 23:33:17 PM

All use subject to JSTOR Terms and Conditions

ROSS / USES, ABUSES,ANDALTERNATIVES

TOTHENET-PRESENT-VALUE
RULE

undertaken.Let T denote the durationof the project,and let

rTdenote the yield on a zero coupon bond with maturityT.
The appendix shows that the optimal yield at which to
undertakea projectis approximatedby

101

V. Conclusion

For most investments, the usefulness of the NPV rule

is severely limited. As a formal matter, it applies only
in those cases where the investment opportunityinstantly
rT = IRR - a/disappearsif it is not immediatelyundertaken.In fact, the
vast majorityof investmentshave a not insignificant time
where a?J- is the standarddeviation of the interest rate
period over which they may be undertaken,and this implies
at level r.
that they have an embedded optionality on their own
An interestrateof 6%andanannualproportionalstandard
valuation that is exercised when the initial investment is
= 0.2 and o(dr) = 0.2 x 6% =
deviationof 20% (i.e., oY(dr/r)
made. We must take very seriously the caveat to the NPV
120 bp) produces
that it applies only in cases where an investment does
not preclude some alternativeinvestment, because every
a = 0.2 x 0.06/4=0.06 = 0.049
investmentcompeteswith itself delayedin time. It is not that
This would resultin a differencebetweenthe optimalcutoff the NPV rule is
wrong,ratherit is somewhatirrelevant,and
and the IRR of
at best, it must generallybe modified to be useful.
Because nearly all investments involve the option to
r*(T)- IRR = -o/2- = -0.035
undertake them when financing alternatives are more
or 3.5%.
favorable, in general, the preferredway to deal with such
Notice that the optimal cutoff is lower than the IRR investment decisions is to treat them as serious options
which capturesthe sense that the option to undertakethe on the financing environment.As we have shown, when
project has to be in-the-moneyto be exercised. As an IRR evaluating investments, optionality is ubiquitous and
rule, this would meanthatwe undertakethe projectwhen the unavoidable.If modem finance is to have a practicaland
IRR exceeds the marketrateof interestby 3.5%.Notice, too, salutaryimpact on investment-decisionmaking, it is now
thatthis is a significantdifferencein thatit is on the orderof obliged to treat all major investment decisions as option
half of the interestrateitself.
pricingproblems. U

References
Antle, R. and G. Eppen, 1985, "CapitalRationingandOrganizationalSlack
in CapitalBudgeting,"ManagementScience (February),163-174.

Ingersoll,J. and S. Ross, 1992, "Waiting to Invest,"Journal of Business

(January),1-30.

Antle, R. and J. Fellingham, 1990, "CapitalRationingand Organizational Ross, S., R.W. Westerfield,and J.F. Jaffe, 1993, CorporateFinance, 3rd
Slack in a Two Period Model," Journal of Accounting Research
ed., Homewood,IL, Irwin.
(Spring), 1-24.
Ross, S., 1978, "A Simple Approachto the Valuationof Risky Assets,"
Journalof Business (July),453-475.

This content downloaded from 142.103.160.110 on Sun, 8 Sep 2013 23:33:17 PM

All use subject to JSTOR Terms and Conditions

102

FINANCIAL
/ AUTUMN1995
MANAGEMENT

Appendix
This appendix uses the results of Ingersoll and Ross where ro is the instantaneousrate at which the projecthas a
(1992) to derivean approximationfor the optimalhurdlerate zero NPV
at which an investmentshouldbe undertaken.
ro
1nI
We will startwith point input-pointoutputprojectsand
b(T)
let I denote the ratio of the point investment to the point
outputat time T. The interestratedynamicsare given by the and
Ito equationfor the shortrate,r,
V

dr = Xrdt + aTrdz

(2

We can recast this solution in termsof the cutoff hurdle

where z is a Brownianprocess. This can be thought of as
rate
on a T-periodbond andthe IRRfor the project.The IRR
either the actual dynamics of the interest rate or the risk
is
given by
neutraldynamicswith risk adjustmentcoefficient, k.
Underthis assumption,the termstructureof interestrates
IRR =nI
can be solved and the price of a zero coupon bond is given
by:
Hence, the hurdlerateexpressedin termsof a T-periodbond
p(r,T)= e-b(T)r
yield is:
where

Sb(T)IRR

rT= T
2(er - 1)
(y - X)(eY -1) + 2y
b(T)=
and
= XA2+ 202
7
The IngersollandRoss formulafor the optimalshort-term
interestrate at which to exercise the option to undertakethe
projectis given by:
r* b(T))r
+?
=r?

(v l

Il(v - b(T)
= IRR Tn
v

To simplify the computations,we assume that the local

expectationshypothesis holds and that there is no interest
rate risk premiumper se, i.e., we set X = 0. Approximating
the hurdle rate for small choices of a, through a Taylor
expansion,produces
rT= IRR- (J
In general,this approximationis valid for well-behaved
projects,i.e., projectswith a single IRR. Interestingly,since
the approximationis independentof time, T, thereis no need
to estimatethe durationof the project.

This content downloaded from 142.103.160.110 on Sun, 8 Sep 2013 23:33:17 PM

All use subject to JSTOR Terms and Conditions