Accepted Manuscript

Title: Internal Rate of Return: Good and Bad Features, and a

New Way of Interpreting the Historic Measure

Author: Duncan A. Mellichamp

PII: S0098-1354(17)30251-X
DOI: http://dx.doi.org/doi:10.1016/j.compchemeng.2017.06.005
Reference: CACE 5836

To appear in: Computers and Chemical Engineering

Received date: 5-10-2016

Revised date: 18-4-2017
Accepted date: 3-6-2017

Please cite this article as: & Mellichamp, Duncan A., Internal Rate of Return: Good
and Bad Features, and a New Way of Interpreting the Historic Measure.Computers and
Chemical Engineering http://dx.doi.org/10.1016/j.compchemeng.2017.06.005

This is a PDF file of an unedited manuscript that has been accepted for publication.
As a service to our customers we are providing this early version of the manuscript.
The manuscript will undergo copyediting, typesetting, and review of the resulting proof
before it is published in its final form. Please note that during the production process
errors may be discovered which could affect the content, and all legal disclaimers that
apply to the journal pertain.
 DAM: March 31, 2017

Internal Rate of Return:

Good and Bad Features, and a New Way of Interpreting the Historic Measure

Duncan A. Mellichamp, Professor emeritus

Department of Chemical Engineering
University of California
Santa Barbara, CA USA 93106
[email protected]

HIGHLIGHTS for: Internal Rate of Return – The Pluses and Minuses, etc.

 IRR—most-used profitability measure, but nonlinear, non-analytic & w/o para-

meters.

 Attempts to make IRR more useful/realistic actually have confused the situation.

 Despite IRR’s issues, a new measure, NPV%, can help yield more information.

 E.g., effect of Enterprise ROR on NPV% can be interpreted wrt IRR (both fcns of
ROIBT).

 But, better approach is to use NPV% directly and obtain its many advantages.

Abstract: IRR, a widely used profitability measure, is the Discount Rate that yields Net Present
Value (NPV) = 0 for a stream of positive and negative cash flows, at least one of each sign and
with no explicit financing payments. A big disadvantage is lack of parameters, such as a project
finance rate or the enterprise rate (ER), i.e., Return on Investment of the overarching investment
group to serve as a measure of opportunity cost. The coupled metrics proposed earlier by the
author— NPVproject and NPV%--do not suffer these disadvantages, so IRR is analyzed in terms
of NPV% . Useful information can be obtained from a projection of IRR values onto the NPV%,
ER plane revealing the sensitivity of IRR to risk under meaningful operating conditions.

Keywords: Internal rate of return; Hurdle rate; Profitability measure; Enterprise rate of return;
Modified internal rate of return; Business risk

1. Introduction.
The Internal Rate of Return (IRR) has been used for years by economists and engineers to
estimate the profitability (or potential profitability) of projects. Its definition is rooted in
procedures of Discounted Cash Flow (DCF), a methodology that is utilized to “weight” cash
flows occurring at the “present time” in some rational way so as to represent their value
relative to “future” cash flows in later years. When coupled with Net Present Value (NPV),
IRR forms the necessary second measure of profitability. NPV is scaled, i.e., with units of
dollars, while IRR is un-scaled or normalized, with units of % or %/time. Two such

1
 DAM: March 31, 2017

measures are required to distinguish between projects that may appear to be about equal
in terms of profitability but are of different size or scope. Whether the designer prefers a
higher profitability or a lower scale of investment can then be considered.

Process Engineers interested in plant design/costing {Douglas [1988], Edgar et al [2001],

Peters et al [2003], Ross et al [2005], Seider et al [2008], Towler and Sinnott [2007], Turton
et al [2003] and Wells & Rose [1986]} have long used a variation of the IRR, now generally
referred to as the “hurdle criterion.” In this approach, a company will impose a significantly
high hurdle rate before approving any conceptual design, in other words, will require that a
potential process possess a minimum value of IRR as its discount rate, then choose the
conceptual process alternative that yields the highest value of NPV above this hurdle.

Process engineers are generally satisfied with this approach; and economists view IRR/NPV
as a mature area, another way of saying that the field is “dead” as far as need for additional
research. Still, the reexamination of the use of NPV methods made earlier by the author
[Mellichamp, 2013] and [Mellichamp, 2016] indicated that much more could be
accomplished by defining/using a different normalized profitability measure (the
normalized NPV, referred to as NPV%). It is useful to look at the IRR concept again in light of
the NPV% measure, to see what it does and does not provide as an alternative.

First, it is important to note that IRR is strictly defined and used only to determine whether
a plant or project will be profitable enough to a company (the Enterprise) to build it. The
definition specifically does not involve the concept of financing. Thus, for companies large
enough to have a group that focuses on financing plants/projects, it is only after a design
engineer or group evaluates that it is potentially profitable enough to the company to
justify constructing and operating it for its anticipated lifetime that the “finance group”
considers what alternative to us— to take on an outside financier, borrow money, sell a
bond issue, sell shares of stock—and then how much of any of these is required.

2. The Good Features Underlying IRR.

Apparently, the original idea of IRR was to calculate the annual return on a purchase made
with a single-payment of cash, then sold n years later for cash, for example, the acquisition
of a fine painting. How much it appreciated or depreciated each year that it was owned, the
so-called IRR, was simply the nth root of the quotient of sales and purchase values, thus

 Sales Price 
1/ n

IRR   1
Purchase Price 
 (1)

Notice that no cash flows occur except at times “going in” and “coming out” of the purchase.

Internal Rate of Return represents a similarly simple concept, a single number that does not
depend in any way on financing issues (whether money is used to buy the object, if not,
what rate of interest might be paid and other terms of a loan, etc.). Investors historically

2
 DAM: March 31, 2017

liked the concept and a simple interpretation appears to have been attached to it from the
outset, namely, that IRR represents the upper limit on the rate of interest that could be paid
to purchase an object for later sale and still make a profit.

Later, multi-year projects were considered in which the costs of a project were paid out
over several years, then expenses were recovered via after-tax profits earned over a
subsequent period of several years. Net Present Value methods were developed to
compensate for the different times at which the cash flows occurred, and it seemed natural
to find a single number that could characterize an entire stream of cash flows for analysis.

However, the important issue is not whether a project will make a profit, but prospectively
whether it will make enough money in early years to return the capital invested, and over
its full assumed lifetime, more than enough to justify the intrinsic risks involved. Not every
investment returns all the capital invested; some don’t return any portion of it. An investor
wants to obtain both sunk capital and sufficiently enough beyond that to compensate (at
least statistically and over a long period of experience) for the known and unknown risks
the investments will be exposed to.

In the process world, plant designers would like to know, well before building and starting-
up a plant, just how profitable the designed project must be to justify its intrinsic risks. For
this purpose, the IRR measure is not directly applicable as can be shown. Before getting so
specific, the calculation of IRR for a textbook project example, from Douglas (1988)1 is
illustrated in the Supplemental Material. There the project example exhibits IRR = 22.1%,
as shown using classical spreadsheet methods in Tables S2 and S3.

If one can conjecture a “symmetric growing/declining account,” i.e., one that charges 22.1%
interest on loans and pays 22.1% interest on deposits, it is shown in the Supplementary
Material (Table S4) that a sequence of loans taken out during the Construction Period and
then paid back via Profits after Taxes during the Operations Period, exactly sums to 0.
Thus, IRR can have a practical/meaningful interpretation only if a real financial account
that pays interest at the IRR rate is available to deposit/withdraw the in/outflows of cash.
Or the Enterprise, if it finances the project from internal funds, would have to be earning
money at the rate of 22.1% for the Douglas example to be meaningful. Because rates
calculated for IRR often are higher than commercial rates, this requirement almost never is
met. In the process design community it mostly is unrecognized and always is ignored.

We note that IRR as a measure of profitability does exhibit several good features: (1) How
to optimize the desirable result is obvious: merely maximize IRR. Also, (2) many individual

1
Parameters for Douglas’ Example 2.5-1: Discounted Cash Flow (Supplementary Material). The notation, i.e., FC =
Fixed Capital, etc. is spelled out in detail in the Notation section.
FC = 100,000,000 a-3 = 0 b+1 = 0.60
WC = 20,000,000 a-2 = 0.15 b+2 = 0.90
SU = 10,000,000 a-1 = 0.35 b+3 = 0.95
SV = 3,000,000 a0 = 0.50 TR = Tax Rate = 48%

3
 DAM: March 31, 2017

practitioners/design groups have over time built up much operating experience with IRR
(i.e., a suitable value to use for the profitability criterion, often called the “Hurdle Rate,” as
is true with any familiar metric (measure).

Despite its widespread use, when considered as an interpretable metric of prospective

profitability, IRR unfortunately exhibits several problems. Of these, some issues can be
mitigated, some even eliminated:

 IRR is a non-linear function of the cash flows. At high values of Discount Rate
(equivalently, of IRR), much less value will be attached to cash flows farther into the
future. Thus, a processing plant that uses a noble metal catalyst clearly will have
initial fixed costs much higher than one using a traditional catalyst. But it also will
necessarily exhibit a lower IRR (whether the catalyst can be salvaged or not)
because the value of the catalyst at the time of decommissioning will be much lower,
when discounted.

 Also, how might the concept of a “risk-justified IRR” be introduced in the absence of
a realistic interpretation of the IRR quantity itself? The lack of a meaningful physical
interpretation for discounted cash flows, particularly with high discount rates,
makes this concept difficult. One can require a higher IRR (“Hurdle Rate”) for a
riskier project/product on an ad hoc basis, and define

Risk Cushion = IRR – Hurdle Rate (2)

but none of these quantities is simply related to ER given the non-linear expression
for IRR. The issue is: quantitatively how large a Risk Cushion is sufficient when
dealing with a non-analytical and non-linear function?

 In a similar way, how does one evaluate the benefit of using external financing or of
using more/less costly financing? These important quantities are not involved in the
IRR measure. Similarly, it is well known that inflation affects interest rates, but how
does one account for the costs/benefits of higher/lower inflation costs analytically?
 IRR is a non-dimensional (%/year) rather than a scaled quantity ($ or $/year). Thus,
two plants exhibiting very different values of NPV but nearly the same values of IRR
will be rated essentially the same. Clearly, for the higher NPV case the profitability
(returns) to the company will be better, perhaps significantly better. Consequently,
IRR most often is used along with a dimensional measure such as NPV, itself.

 In general, the measurement of profitability requires the use of two separately

calculated metrics, one un-scaled (non-dimensional) and the other scaled
(dimensional) to get around this problem. Whereas NPV and NPV% are calculated
together and are closely and analytically related, IRR has no similar analog. The
“trial-and-error approach” required in calculating IRR from NPV precludes the
application of analytical methods, as done in the former case [Mellichamp (2013)].

4
 DAM: March 31, 2017

For the example examined in Supplementary Tables S1 – S3, IRR = 22.1% has real meaning
only if the intermediate cash flows generated in each year can be reinvested at the IRR
Value as shown in Table S4. That this assumption is implicitly invoked when one makes
traditional DCF calculations, i.e., uses a DCF table, is unrecognized by many users. But
achieving such a high ROI regularly via routine investment is not a likely prospect. If such a
fund were available, the Enterprise would be better off increasing its assets year-over-year
at this high rate by investing them directly.

Economists/financiers have known for years of problems interpreting the IRR. Though its
basis is simple, and therefore attractive, it yields unrealistic or un-interpretable results for
quantitative purposes. This characteristic, plus too few degrees of freedom to represent
most practical financial situations, motivated the development of a so-called “modified
IRR”. MIRR occasionally is used instead of IRR. But it has its own problems, as noted below.

3. The so-called Modified IRR (MIRR): An Incorrectly Defined Related Measure

Many business applications have utilized a modification to the IRR that presumptively deals
with the issue of having all cash flows originate or terminate in an account paying/charging
at the IRR rate to maintain physical reality and mathematical rigor. The Modified IRR (MIRR,
Kierulff, 2008) is an attempt to return to a situation more like the original defining
relationship for IRR (in which a single payment is made to purchase the investment and a
single payment is returned to the buyer following its sale).

Several definitions of MIRR are available in print. All appear to exhibit a common error,
namely that the MIRR definition conflicts with the fundamental defining relation for
discounted cash flow computations. The error is subtle and deserves a focused research
note for explanation (in preparation). However, one must be careful because the incorrect
way of defining MIRR has been picked up in the function available in at least one widely-
used spread sheeting method, e.g., Excel, and therefore has acquired a certain false
authenticity.

4. A New Way of Interpreting IRR, providing a Constructive Path to Broader Use.

At this juncture, having questioned several fundamental issues with the IRR metric, one
might ask whether it is possible to take advantage of new knowledge available from the
author’s earlier NPV% developments, to make better use of an obviously flawed but still
widely applied metric.

Note that in this and other representations, the Enterprise Rate is defined as the after-tax
rate of return, averaged year over year2, that the company (the Enterprise) is experiencing
at the time it considers a project for potential design and construction. Many traditional
DCF evaluations such as those shown in the Supplementary Material (Tables S1 – S4)

2 Often using a multiple-year moving average to smooth results.

5
 DAM: March 31, 2017

assume that internal (Enterprise) funds will be used for construction costs and are “paid
out” in the year spent. Operating cash flows are returned to the Enterprise in each year of
the assumed Operating Period and summed at their after-tax DCF value until shutdown of
the plant. Whether internal or external funds are used, ER is an appropriate Discount Rate
for both the construction and operations period. It penalizes the Enterprise operations
appropriately for the lost internal investment opportunities that result from building the
subject project via discounting of its future profits.

In the original development discussed by the author [Mellichamp (2013)], an explicit

(external loan) financing was used to cover construction costs, discounting and
accumulating these to obtain the Total Capitalized Investment (i.e., including intermediate
interest). A bond issue then was used to finance (pay off) the construction costs. Operating
cash flows returned to the company were invested at the Enterprise Rate, so the
appropriate Discount Rate remains the Enterprise Rate. One can obtain a significant
reduction in the ROIBT required to justify a plant project (construction and operation) by
delaying payment of construction costs in this way until end of the project, a time when
they are discounted significantly. The largest increase in NPV of the proposed project will
be obtained when FR << ER [Mellichamp (2016)].

To help the analysis here, it is useful to recall the design spreadsheet developed earlier to
calculate NPV%. A re-look provides a clue to a constructive way to enhance information
conveyed by the IRR. The spreadsheet used to illustrate the NPV% methodology for the first
example was given as Table 3 in the original [Mellichamp (2013)], here substantially
reproduced as Table A.1 in the Appendix. But now external financing is not employed so FR
= 0%. The choice of CR = ER = 8% retains the appropriate Discount Rate during the
Construction and Operating Periods. Thus, costs of construction are assumed to be paid
using internal funds at rate CR = ER during the construction period; and the accumulated
expenses, including internal interest are “booked” just as operations commence [EOY(0)].

Note that this example, with FC = 42.3 and ProfitBT = $33.6MM chosen arbitrarily, yields ROIBT
= 61.1% and NPV% = 12.3%. If one wished to determine the conditions necessary to yield a
certain value of NPV%, say 10%, Excel’s Goal Seek routine could be used to search for the value
of ProfitBT, or alternatively FC, that is required. One of the main results of the earlier papers was
to show that NPV% is a unique linear function of ROIBT. Once the financial parameters (CR, FR,
TR), construction cost schedule represented by the ai, and profitability ramp-up schedule given
by the bj have been selected, NPV% is then solely a function of ROIBT. Here, however, we have
used Goal Seek to determine that IRR = 28.0%.

Surprisingly, the conclusion is that under these same conditions (with all other parameters
selected) IRR itself is a function of ROIBT only. In summary, using Table A.2, similarly
developed for the case when construction costs are paid off immediately at end of
construction, but here using Excel’s IRR function, one can quickly obtain values of
required
ROIBT that correspond to desired values of IRR for any chosen set of design

6
 DAM: March 31, 2017

parameters. In Table A.2, IRR is retained at a value of 28% by keeping ROIBT = 61.1%, but
NPV% is independently set to 5% by adjusting the discount rate (ER = CR) to 13.5%.

Thus, a key result of this present study is that IRR is dependent only on ROIBT, once design
and financial parameters have been chosen.

Question 4.1: Can IRR methodology be extended, perhaps by inference, to evaluate its
implicit dependence on ER?

Background. Though IRR and NPV% both measure profitability, IRR is a function of a single
operational variable (ROIBT) once the designer has set the key design parameters
( aWC , aSU , and aSV ), the tax rate (TR), and the construction/operations parameters [i.e.,
capital allocation schedule (ai) and ramp-up of profit following start-up (bj)].

In principle, NPV% has been defined previously as a function of four variables (ROIBT, ER,
CR, and FR) once the design variables are chosen. Since external financing is not considered
in this study of IRR, CR and FR are not used (except as a surrogate for discount rate, such as
with CR in the spreadsheet of Tables A.1 and A.2). Thus, NPV% here is treated as a function
of only two variables—ROIBT and ER.

The effect of this small difference in number of parameters is significant. Whereas ER is not
reflected at all in the definition of IRR, the specific ER earned by the Enterprise over time will
have a large effect on the needed profitability of any new process, as shown in the author’s
earlier work. So while IRR is parametrically deficient by two degrees of freedom compared to
NPV%, in the present study, and this deficiency shows up in the inability of the measure to reflect
realistically the effect of varying ER in an application, we show now that enough information can
be inferred from the analogous NPV% situation to partially compensate for this deficiency ... at
the minimum.

In the remaining work, the spreadsheet in Table A.2 is used exclusively, with IRR and NPV%
established at values convenient for plotting or other purposes. It is important to keep in mind
that IRR can only be set directly by varying ROIBT. For example, the Excel Goal Seek function
can be used to find a value of ROIBT (ProfitBT and/or FC) to fix IRR at a particular value for
plotting, etc. Once that is done, NPV% still is subject to changes in ER, allowing the designer to
utilize that one extra degree of freedom usefully.

It is convenient here to pre-specify several process designs that can be distinguished relative to
each other in terms of potential profitability. For this purpose, the parameters in Table 1 serve to
differentiate via one-year, two-year, and three-year construction projects.

Table 1. Parameter Sets for Three Design/Construction Cases Evaluated for Profitability

7
 DAM: March 31, 2017

Summarizing: Because there is an established relationship (the spreadsheet in Tables A.1 or A.2)
that links IRR with NPV% through ROIBT, it is possible to map the otherwise undefined relation
between these two metrics as a function of ER. Thus, while IRR is not specifically a function of
ER, one can infer such a relationship via the links between NPV%, ER, and ROIBT.

First, we plot the functional relationships showing NPV% (linear) and IRR (nonlinear) against
ROIBT. We note the significant nonlinearity that must be present in some interior relationships, as
well (see Figure 1).

Figure 1. NPV% and IRR vs. ROIBT

Now, using the spreadsheet (A.2), we find IRR values corresponding to each desired value of
NPV% as a function of ER for each of the three sets of Construction/Operations parameters in
Table 1. The results in Table 2 are obtained.

Table 2. IRR Required to Achieve Various NPV% Levels as a Function of ER.

Just as was shown in Mellichamp (2013), in which Pay-Out Time was placed on the ordinate of a
plot of selected values of NPV% vs ER, though POT strictly speaking was not a function of ER,
we can do the same thing for IRR, as shown in Figure 2. Here, we continue to consider the
unfinanced case, therefore deal with NPV% as a function of two variables (ROIBT and ER). The
ordinate values of NPV% (equivalent via the mapping of ROIBT) have been left off Figure 2 to
emphasize the underlying implied relationship between IRR and ER.

Figure 2. Internal Rate of Return “versus” Enterprise Rate with NPV% as Parameter.
(25% Hurdle Rate also shown. Lowest line presents IRR = ER, the minimum condition of interest.)
In summary, for these three sets of data, there is an inferable relationship between NPV% and
IRR, each of which is provided in Figure 2. Note the “unsurprising result” that IRR can be
interpreted in terms of NPV% = ƒ(ER). But, in fact, two real surprises do appear in this plot: One
is that IRR is scarcely a function of the Construction/Operations parameters. But also the
ordering of parameter sets – 1<2<3 for the smaller values of NPV%, changes at larger values to
3<2<1; the “cross-over appears to occur approximately at NPV% = 10.

However, the degree to which the separate Construction/Operations parameter cases “cluster
together” for each value of NPV% in Figure 2 is a surprise. The clear implication is that IRR is
not a sensitive measure of profitability, at least not for differences in the fine detail of fixed
capital allocation schedules or profit expectations over just a several year period. This feature
and the significant dependence on ER are discussed below. But first, alternative data (see Table
3) are obtained that better show the effect of changing ER on imputed NPV%.

Table 3. IRR Corresponding to NPV% Values as Function of ER.

Financial analysis here utilizes the “Adequate” construction/operating parameters.

8
 DAM: March 31, 2017

Based on the results in Table 3, one can argue that choosing a single value of “Hurdle Rate”
(e.g., 25% or any other value) as a criterion for a build/no build decision significantly
oversimplifies the profitability situation. The problem is that IRR contains too little information,
by itself, to reflect the Enterprise’s own profitability, ER.

To illustrate that issue more clearly, only the middle set of construction/operation parameters
(the “OK” or “Adequate” values that represent a two-year investment period) are employed in
constructing Figure 3.3

Figure 3. IRR Levels Required to Achieve Particular Values of NPV% Plotted vs. ER
(Here, values of IRR have been mapped onto the NPV% vs. ER plane
for the case of the “Adequate” Construction/Operating Parameters)

Table 3 contains the same information as in Table 2; but only the intermediate (“Adequate”)
parameters case is plotted in Figure 3 to reduce confusion. In this case, the information plotted is
identical, but was obtained and arranged differently for viewing purposes. Specifically, it is
desired to show how NPV% plots against ER using values of IRR as parameter in order to
illustrate the NPV% (profitability) bands more clearly.

Analyzing Figure 3, one sees that any horizontal line represents a constant value of NPV%. Any
value of IRR, e.g., hurdle rate (for example, the 25% hurdle rate shown as a heavier line) serves
as an upper bound on all values of NPV% that correspond to lower values of IRR (i.e., IRR <
25%). Thus, if IRR = 25% is achievable and ER = 8% represents the operating condition of the
Enterprise, then 0 < NPV% < 7.5 represents achievable values of the profitability measure. If IRR
= 30% is achievable, and ER =8%, the operating range increases to about 10.5% (0 < NPV% <
10.5).

One can conjecture that individual companies have worked out, empirically and over time, what
is needed in their own manufacturing environment to provide successful design decisions. Thus,
referring to Figure 3, a company investigating a design with the “Adequate” (or “OK”) design
parameters and operating with an internal rate of return ER = 8% might choose a 25% hurdle rate
in which case they will still have (perhaps without knowing it) a profitability cushion of 0 -
7.5%. A company employing the same hurdle rate but operating with ER =6% would have a
profitability cushion up to nearly 10%.

At the other extreme, a company operating with ER =16% would have a risk margin of only
2.5% for a candidate new plant that meets the 25% hurdle rate condition. This last company

3Analysis utilizing the spreadsheet can be done in several ways (e.g., ROIBT can be found to
set IRR to a desired value, or to set NPV%; or ROIBT can be set independently). Similarly,
discount rate ER can be set to meet alternative objectives. Thus, the individual procedure
used to develop each figure is spelled out in notes attached to the corresponding data table.

9
 DAM: March 31, 2017

would have to achieve a hurdle rate of 40% in order to obtain a degree of security as high as
NPV% = 7.5%; or a hurdle rate of 45% to achieve a range equivalent to 10%, etc. Of course,
these figures are approximate (limited to the precision of the plots and the ability to know ER
accurately), and designers likely will be completely unaware of their situation absent an analysis
similar to that provided here.

Note that even very high levels of IRR imply that as ER increases (higher profitability of the
entire Enterprise), eventually only low values of the NPV% profitability measure actually are
feasible. Also, closer inspection of the two figures (2 and 3) seems to indicate that the
“placeholder values” of NPV% proposed by the author [2013} may be somewhat too high, in the
absence of external financing, unless a more reasonable Tax Rate (~30%) is in effect. In
particular, for the reasonable (“Adequate”) parameters, a hurdle value (IRR) of 50% would be
required to achieve NPV% = 20, if ER =10% and TR = 48%.

These results emphasize the “Catch 22 situation” in which companies that produce high-
risk/high-return products (such as pharma) and generate a high Enterprise Rate find themselves:
having had to achieve a high IRR (hurdle rate) just to maintain their high net profitability,
perhaps even higher in order to be able to sustain ER in the face of the inevitably higher failure
rates characteristic of such product lines. The ability to utilize external leveraging mitigates that
situation, as noted earlier.

At this point, a second question arises, namely:

Question 4.2. Can a screening methodology be developed that usefully distinguishes relatively
profitable from unprofitable processes using IRR as the direct measure.
Or perhaps can IRR be augmented in some manner to accommodate different values of ER?
Background. Figure 2 indicates that IRR can be used reasonably well to separate process
candidates by their underlying profitability. And using NPV% methods developed earlier, one can
even assess the need to attain specific values of inferred NPV%, as in Figure 3. Without such
information, IRR used as an objective function leads to processes being lumped together on some
basis that is nearly independent of the minor perturbations introduced by a multiple year
design/construction period (ai ≠1) and delays after Start-up of ProfitBT that can lead to lags in
required
achieving the design values (bj ≠1). But what values of ROIBT would be needed to obtain
the values of IRR seen in the relatively insensitive results of Fig 3? And is the “basic measure” of
profitability, ROIBT, and its linearly-related surrogate, NPV%, equally insensitive to relatively
minor perturbations in the investment schedule (the ai) and to similarly minor perturbations in
the ProfitBT schedule (the bj)?

Results. For purposes of answering Question 4.2 and its corollary issues, we start with the two
extreme sets of Construction/Operating parameters [Sets 1 (“Best”) and 3 (“Marginal,”

10
 DAM: March 31, 2017

required
as defined in Table 1)]4 and obtain results for ROIBT vs ER, for values of constant NPV%
(NPV% = 2.5, 5, 7.5, 10, 12.5 and 15), in Table 4 and Figure 4.

Table 4. ROIBT Required to Achieve Various NPV% Levels as a Function of

Enterprise Rate, ER

Figure 4. versus Enterprise Rate for Two Design Cases in Table 4

NPV% values (2.5%, 5.0%, ... , 15%) are shown for both Best and Marginal parameters)

The results in Figure 4 are surprising in the sense that ROIBT clearly is sensitive to the variations
in minor parameters, not at all like what was seen with IRR in Figure 2. In particular, a great deal
of overlap is seen as ER increases, a property that one would expect in a sensitive measure of
profitability. Here the increasing profitability requirement imposed via NPV% level clearly is at
cross purposes with both increasing ER and, as one would expect, the decidedly
favorable/unfavorable effects of the capital allocation (ai) and ProfitBT (bj) schedules embedded
in the two very different sets of parameters.

Question 4.3. Why is sensitivity to these minor changes in cash flows “missing” in Figure 2 but
not in Figure 4?

required
Background. To answer this question, ROIBT was evaluated as a function of IRR for all
three parameter cases to determine what is responsible for converting (“mapping”) the relatively
insensitive results in Fig 2 (IRR vs. ER) to the similar but much more sensitive results in Figure 4
required
(ROIBT vs ER). Thus, we find numerical relations between ROIBT and IRR values
corresponding to each of the three test cases previously analyzed. The results in Table 5 are
highly interesting, but certainly not unexpected in that the “missing sensitivity” must show up
here, i.e., in the relationships linking ROIBT and IRR. Note there is only one relation for each set
of constant parameters because IRR is independent of ER.

Table 5. Relationship Between ROIBT and IRR for Construction/Operating Parameters:

1--"Best" 2--"Adequate" 3--"Marginal”

We note that the three functional relations are different from case to case. Further, the
degradation in profitability performance between parameter sets 1 and 2 and between parameter
sets 2 and 3 is clearly reflected in these comparative results.

Graphing the three non-linear curves, as in Figure 5, shows visually what is happening—the
sensitivity that plant designers normally wish an objective function to possess in order to be able

4The middle (“Adequate”) set has been left out here to avoid the resulting confusion due to
multiplicity of lines. Viewing only the two extremes provides sufficient information to see
how much ROIBT is affected by variations in ai and bj parameters.

11
 DAM: March 31, 2017

to measure Profitability accurately turns out to be attenuated in the case of IRR. But then it is
concentrated into the ROIBT vs IRR relationships as a form of compensation. All three of these
curves are non-linearly related to ER, even though that dependence is only inferred in the case of
IRR; and the progressive change in profitability one expects to see by considering the three
parameter sets (1 > 2 > 3) becomes apparent. We conclude that IRR, in addition to its other
problems, is quite insensitive to certain cash flow variations. One must be careful if these are
structurally present.

Figure 5. ROIBT Required to Achieve Specified IRR for all

Three Construction/Operating Parameter Sets

Question 4.4: What to do?

Many firms/many engineers have learned how to use IRR and have come to depend on it for
making industry financial decisions. A large set of experience, rules-of-thumb, and mythology
have grown up around its use. Since IRR is defined by a very concise but inflexible relationship;
it has no extra internal degrees of freedom. On the other hand, it requires no extra information
such as discount rate, finance rates, etc. to use.

From the results presented in Figures 2 and 4, one sees that IRR (in contrast with NPV%) is:

1. Not functionally related to ER, an important missing characteristic because

profitability of a prospective project should be evaluated relative to how profitable
the overarching company’s operations are in its present business environment;

2. Not capable of distinguishing among minor but important differences in annual

cash flows. These can easily represent the difference between profitable and
unprofitable designs; and

3. Insensitive to several important characteristics one normally wants to include in a

financial evaluation. Surprisingly, these include key measures such as ROIBT, NPV%, etc.

The advantage of IRR, namely its requirement for no additional information beyond a projected
cash flow schedule over the lifetime of a project, is its biggest positive. The lack of additional
parameters in its defining relation is its underlying negative ... it is not possible to parameterize
IRR’s properties directly in terms of other important business variables, such as the company’s
ongoing profitability (measured by ER), except indirectly using the method provided here, nor to
estimate directly the advantages that might accrue through use of external financing, etc. The
work in this paper provides a partial way around these problems.

5. Conclusions

One can conjecture that IRR has grown up in many different business environments and has been
adapted to use in each one. Thus, a company will have established a “Hurdle Rate” that works
for its own applications, meaning that it will be set at a value that reflects its particular
Enterprise Rate. It also will reflect some sort of standard expectation for the timing of design,

12
 DAM: March 31, 2017

construction, and start-up cycles. And it will have included a suitable (even if unidentified value
of profitability cushion, cf. NPV%min) to reflect experience and the need to cover the relative
losses experienced in past design projects that didn’t work out to be as profitable as projected.

Obviously this historical approach can work suitably and it will continue to work so long as there
are no big changes in the financial assumptions underlying a particular design
procedure/analysis. But, if these do change—for example, the company decides to move into a
new business area where profitability expectations are certain to be different, or if the underlying
business risk changes—then there are two existing ways of dealing with the new situation:

1) Ignore the changes, i.e., follow Admiral Farragut’s well-known command;5

2) Move consciously to an adaptive or evolutionary mode where new experience(s) can

change the rules-of-thumb that have been applied in the past.

A third way is more soundly based on the new fundamental profit measures uncovered in this
and earlier related work by the author. These suggest that a company take a new look at how
profitability is being measured internally. With this approach, one will:

3a) Either attempt to use some of the insights that the present paper has developed to
understand better how inferred information from an NPV% analysis might be used to
incorporate underlying profitability changes better, and to adjust IRR expectations based
on them, etc., or

3b) Adopt a more fundamental view of profitability measurement (such as to use the
linked, sensitive measures—ROIBT, NPV, NPV%—as discussed in Mellichamp (2013)).

In the present paper, several key ideas introduced previously by the author, in particular, the use
of external financing to reduce basic profitability requirements, are not discussed at all. How
these developments might apply to process designs that use IRR as the fundamental profitability
measure is the subject of ongoing research.

Acknowledgments
The always challenging conversations with my colleague, Michael Doherty, have provided the
sort of devil’s advocacy for IRR methods that one needs to keep questioning the status quo.
These special discussion times are gratefully acknowledged. Also, the mostly written discussions
with Christopher Burk, a chemical plant design TEA practitioner who makes his living doing

5“Damn the torpedoes, full speed ahead.” Attributed to Admiral David G. Farragut during
the Civil War Battle of Mobile Bay (August 1864): (His reference to torpedoes was actually
to fixed mines anchored in the water rather than ones using the modern definition of the
word. However, the corporate philosophy behind his command to the fleet is the same.)

13
 DAM: March 31, 2017

analysis the right way (Burk Engineering LLC, Salt Lake City) have been very helpful,
especially in the computational area.

14
 DAM: March 31, 2017

Appendix A.
Modifying Original Spreadsheet to Calculate IRR; Use of Excel IRR Function

In considering modification of the original spreadsheet [Table 3 in Mellichamp (2013)],

first note that it can be modified to calculate IRR. But we want to retain all the apparatus for
calculating NPV% in order to be able to exploit the relationship between it and IRR that
involves ROIBT.

The original spreadsheet can be modified to calculate IRR, implementing the classical
definition of IRR by making the following three minor changes (to obtain Table A.1):

 Set FR = 0 (assumes that no financing is used when calculating IRR).

 Set CR = ER (thus ER is used here as the discount factor in calculating IRR).
 Set the discount factor that multiplies the cost of final pay-off of the bond issue in
the original spreadsheet to 1.0 (meaning that the pay-off of any construction loan or
of any internal debt accrued by the Enterprise during construction) occurs at the
end of Year 0.6 In the ROIBT situation discussed in this paper, there is no bond issue
and no bond payments so the PV calculations for bond payments in Table A.1 can be
ignored.
 Search for the value of ER that sets any of the NPV quantities to 0. That value of ER is
the defined value of IRR.

Although not required here, we leave intact all the “spreadsheet apparatus” for calculating
NPV%. The reason why will be obvious from the next following development (Table A.2).

A search for the value of ER (the Discount Rate) that sets NPV = 0 in the original example
yields a value of 28.0%. This is the unique value of IRR corresponding to ROIBT = 61.1%,
when all appropriate financial parameters are as chosen.7

In revising the original spreadsheet to calculate IRR, it would be good practice actually to
show the payment of the construction capital at the beginning of the operations period.
Even though there is no difference in results so long as the discount factor is changed to
1.0, it may later be confusing to see the entry at the bottom of the table. But in the next
following development, exactly where this cash flow is located does make a difference.

Table A.1. Risk-Based Profitability Analysis: Internal Rate of Return Via Cash Flow Table
Using T&E (Goal Seek).

6 In the former case, CR would be set at the commercial loan rate rather than ER. Note that
although the definition of IRR assumes that external financing is not used, calculation of IRR
is valid in either of these two cases.
7 Note that the precision of the Goal Seek search is enhanced if a larger function of NPV, say

NPV(0), is set to 0 rather than the smallest function, NPV% .

15
 DAM: March 31, 2017

Table A.2 is a more useful modification of the spreadsheet, one that was developed to provide
IRR directly (via the Excel IRR function) while retaining the ability to calculate NPV%
simultaneously. To employ the Excel IRR function, one must place all of the annual cash flows
into a vector, each cell representing the cash flows in one year summed together. The IRR
function utilizes this vector and is calculated “continuously” whenever the spreadsheet is active. 8

Table A.2. Payoff at End of Construction. Excel IRR Function finds IRR Directly (Cash
Flows Must Be in Vector).

Note that Table A.2 simultaneously provides the IRR (= 28.0%) and NPV values (e.g.,
NPVproj = $34.6MM for these choices of PBT and FC, and design/finance parameters.

The value of IRR calculated via the explicit cash flow vector, 28.0%, is the same as that
obtained by Trial & Error (Goal Seek) in Table A.1. However, in this case we have been able
to retain intact the parts of the spreadsheet that are used to calculate NPV% and show that
it and IRR can have independent values. Thus, in Table A.2, IRR = 28% is shown to
correspond to a value of NPV% = 5.0% if ER = 13.5%. In the work reported here, the ability
to link particular values of IRR to nonzero values of NPV% is specifically exploited.

Note that Table A.2 is not designed to analyze financed projects as in the original reference.
But financing has nothing to do with calculation of IRR as we see from use of the Excel IRR
function in A.2 and also via the modifications made in Table A.1. It is important to note that
both IRR and NPV% are functions of ROIBT, but the latter metric is also a function of ER, CR,
and FR. In this version of the table (A.2), only ER is utilized to represent DR.

8The ordering of cash flows must place the construction costs in the top vector element(s)
where they will be appropriately weighted by the internal discount factor(s).

16
 DAM: March 31, 2017

Notation
Comment: The cumbersome notation here is used partly to avoid ambiguity and partly
because “spelled-out subscripts and superscripts” provide more transparency (irony). In
line with practice in some areas of finance, multiple-letter (generally two-letter) acronyms
make the mathematics easier to read. In some cases, full-word variables (always Italicized)
and full-word equations are used (all with apologies to mathematics purists).

ai Fraction allocation of Fixed Capital in Year i (i = 0, -1, -2, -3)

bj Fraction allocation of ProfitBT in Year j (j = 1, 2, 3, etc.)

C Total Production Costs (Raw materials, utility and labor costs, etc. but not
financing costs)
CF Cash Flow

CR Interest Rate on construction loan. For large loans, multiple banks often
are used.

Di Depreciation allowed and taken in Year i [= 0.1(FC+ SU) if project

operating lifetime is 10 years and depreciation is “straight-line.”]
ER Enterprise Rate, i.e., the long-term rate of earnings generated by the
company. A classic measure of ER is the Average Book Value of
Company Assets, as maintained by stock analysts. A five-year moving
average would be appropriate.

FC Fixed Capital (= Direct Costs + Indirect Costs)

IRR Internal Rate of Return (Discount Rate that makes NPV0 = 0)

N Plant anticipated lifetime (= NConstruction + NOperations)

NPV Net Present Value (the sum of discounted Cash Flows over a project’s
entire lifetime, i.e., design/construction/operations.)

NPV0 NPV at the start of process operations [i.e., at end of Year(0)]

NPV proj NPV at the time a firm decision is made to construct, e.g., EOY(-1),
EOY(-2), … depending on the length of design/construction period.
Thus NPVproj  (1 DR)NConstruction NPV0
 norm
NPV proj NPV proj/TCI ( NPV proj normalized by capital requirement, TCI)
norm/ ann
NPV proj  norm
NPV proj ( NPV proj annualized by total lifetime of the project, i.e., the
design/construction/operating period, and normalized by TCI)
  

  
17
 DAM: March 31, 2017

norm/ ann
NPV% Shorthand for NPV proj defined in terms of the project initiation date,
not the startup date, thus based on NPV proj not NPV 0 ! ]
Note that NPV% is a linear function of ROIBT . Therefore its inverse is
 proportional
 to POT and serves as a surrogate for short-term risk.
 
NPV%min Minimum value of NPV% that justifies intrinsic risk.
 

PBT Profit Before Taxes = Total Revenues – Total Production Costs = R – C

  used: Annual Profits, Gross Profits, or Net Earnings.)
(Other names
R Total Revenues (from operations) = Production Rate x Sales Price
POT Pay-Out (or Pay-Back) Time

ROIBT ROI before taxes based on TI = FC + SU + WC (calculated directly from

design cost estimations.)
SU Start-Up Capital (SU is assumed to be partially recovered along with FC
 via depreciation during each year of the project lifetime.)
SV Salvage Value of the plant (Recovered at end of the project’s lifetime.)

TI Total Investment (= FC + WC + SU)

TCI Total Capitalized Investment (TCI ≠ TI = FC + WC + SU), i.e., includes
construction interest costs, if any.
TR Tax Rate (The incremental rate paid for federal+state+local taxes
expressed as a fraction.)
WC Working Capital (recovered at the end of project lifetime along with SV.
Taxes are assumed to be paid only on SV at that time.)

Greek Letters

WC ,  SU ,  SV Factored estimate (multiple of Fixed Capital, e.g., WC = WC FC)

 

18
 DAM: March 31, 2017

Literature Cited

Douglas, J.M., Conceptual Design of Chemical Processes, McGraw-Hill, New York, NY, 1988.

Edgar, T.F., Himmelblau, D.M., and Lasdon, L.S. Optimization of Chemical Processes (2nd Ed.),
McGraw-Hill, New York, NY, 2001.

Kierulff, H., “MIRR: A Better Measure,” Business Horizons, 51, 321-329, 2008.

Mellichamp, D.A., New Discounted Cash Flow Method: Estimating Plant Profitability While
Compensating for Business Risk/Uncertainty, Computers and Chemical Engineering, 251-
263, 2013.

Mellichamp, D.A., The Advantage of Using External Financing (Leverage) to

Design/Build/Operate a New Chemical Plant, Computers and Chemical Engineering, 147-
150, 2016.

Peters, M.S., Timmerhaus, K.D, and West, R.E. Plant Design and Economics for Chemical
Engineers (5th Ed.), McGraw-Hill, New York, NY, 2003.

Ross, S.A., Westerfield, R.W,, and Jaffe, J.F., Corporate Finance (7th Ed.), McGraw-Hill, New
York, NY, 2005.

Seider, W.D., Seader, J.D., Lewin, D.R., and Widagdo, S., Product & Process Design Principles:
Synthesis, Analysis, and Evaluation (3rd Ed.), John Wiley and Sons, New York, NY, 2008.

Towler, G. and Sinnott, R.K., Chemical Engineering Design: Principles, Practice and
Economics of Plant and Process Design, Butterworth-Heinemann, Elsevier, New York, NY,
2007.

Turton, R., Bailie, F.C. Whithing, W.B., Shaelwitz, J.A. Analysis, Synthesis and Design of
Chemical Processes, Prentice-Hall, Upper Saddle River, NJ, 2003.

Wells, G.L., and Rose, L.M. The Art of Chemical Process Design, Elsevier, Amsterdam, 1986.

19
 DAM: March 31, 2017

Figure 1. NPV% and IRR vs. ROIBT

20
 DAM: March 31, 2017

Figure 2. Internal Rate of Return “versus” Enterprise Rate with NPV% as Parameter.
(25% Hurdle Rate also shown. Lowest line presents IRR = ER, the minimum condition of interest.)

21
 DAM: March 31, 2017

Figure 3. IRR Levels Required to Achieve Particular Values of NPV% Plotted vs.
ER
(Here, values of IRR have been mapped onto the NPV% vs. ER plane for the case of
“Adequate/OK” Construction/Operating Parameters)

22
 DAM: March 31, 2017

Figure 4. Return on Investment Before Taxes vs ER for Two Design Cases in Table 4
NPV% values (2.5%, 5.0%, ... , 15%) are shown for both Best and Marginal parameters)

23
 DAM: March 31, 2017

Figure 5. ROIBT Required to Achieve Specified IRR for All Three

Co
nstruction/Operating Parameter Sets

24
 DAM: March 31, 2017

Table 1. Three Parameter Sets: Design/Construction Cases Evaluated for Profitability

Parameters

CR F TR a-3 a-2 a-1 a0 b+1 b+2 b+3 b+4

ParameterC
ase R et seq

“Best” — — 48% 0 0 0 1 1 1 1 1

“Adequate” — — 48% 0 0 0.5 0.5 0.8 0.9 0.95 1

“Marginal” — — 48% 0 0.2 0.5 0.3 0.5 0.8 0.95 1

Notes: * CR and FR are not used in this study, beyond defining them to serve as the Discount Rate in
calculation of IRR in Table A.1 in order to use the spreadsheet from previous studies.
* Loan pay-off at End of Construction corresponds exactly to the definition of IRR if no interest
payments are included in the cash flows.
* “Best,” “Adequate,” and “Marginal” are placeholder names. The intent here is to formulate
hypothetical project designs (with one-, two-, and three-year design/construction periods and
correspondingly poorer ramp-up of profits) and to use them for comparison purposes.

25
 DAM: March 31, 2017

Table 2. IRR Required to Achieve Various

NPV% Levels as a Function of ER*

1. "Best" Operating Parameters

NPV%
ER, % 2.5% 5% 7.5% 10% 12.5% 15%
0 4.2 8.1 11.8 15.3 18.7 21.9
4 9.0 13.6 18.0 22.2 26.2 30.2
8 13.8 19.3 24.5 29.5 34.3 39.1
12 18.8 25.2 31.3 37.2 43.0 48.7
16 23.9 31.3 38.4 45.4 52.3 59.0
20 29.0 37.6 45.9 54.1 62.2 70.2

2. "Adequate" Operating Parameters

NPV%
ER, % 2.5% 5% 7.5% 10% 12.5% 15%
0 4.2 8.0 11.2 14.7 17.7 20.6
4 9.1 13.7 17.8 21.7 25.4 28.8
8 14.1 19.6 24.5 29.2 33.5 37.7
12 19.3 25.7 31.6 37.1 42.3 47.2
16 24.6 32.1 39.1 45.6 51.7 57.5
20 30.0 38.9 47.0 54.5 61.7 68.5

3. "Marginal" Operating Parameters

NPV%
ER, % 2.5% 5% 7.5% 10% 12.5% 15%
0 4.2 7.9 11.7 14.3 17.0 19.6
4 9.2 13.8 17.8 21.4 24.7 27.8
8 14.4 19.8 24.6 28.9 32.9 36.6
12 19.8 26.2 31.9 37.0 41.7 46.0
16 25.3 33.0 39.6 45.6 51.0 56.1
20 31.0 40.0 47.8 54.7 61.0 66.9

* Each element in table is the value of IRR

"equivalent" to the NPV% heading --
equivalent in the sense that the same value
of ROIBT yields both results at the given ER.

Note: Table 2 was developed by:

(1) Placing a desired value of ER into the spreadsheet,
(2) Then finding the value of ROIBT required to produce a desired value of NPV% using Goal Seek.
The corresponding value of IRR is simply read from the Excel vector.

26
 DAM: March 31, 2017

Table 3. IRR Corresponding to NPV% Values as Function of ER.

(Fin
ancial analysis utilizes the “Adequate” construction/operating parameters.)
Value of Enterprise Rate, ER (%)
ROIBT IRR 0 4 8 12 16 20 24
151.5 60 21.9 16.1 11.9 8.8
135.6 55 26.6 19.1 13.9 10.2 7.4
120.3 50 23.0 16.4 11.8 8.5 6.1
105.6 45 28.1 19.7 13.9 9.8 6.9
91.6 40 23.8 16.4 11.4 7.8 5.3
78.3 35 29.2 19.7 13.4 9.0 6.0 3.9
65.7 30 24.0 15.9 10.5 6.8 4.3 2.5
54.0 25 19.0 12.2 7.7 4.7 2.6 1.2
43.1 20 14.4 8.9 5.2 2.8 1.1
32.9 15 10.2 5.8 2.9 1.0
23.8 10 6.4 3.0 0.8
15.7 5 3.0 0.5

Notes: *
Main table data are NPV% values corresponding to each value of ER .

* In
the first two columns, ROIBT (%) determines the paired value of IRR (%).

*
IRR = 0.0% corresponds to ROIBT = 4.9%.

*
While IRR is strictly defined and not a function of ER, for any choice of IRR, one ER

value corresponds to the max value of risk parameter (NPV%) that can be
obtained.

27
 DAM: March 31, 2017

Table 4. ROIBT Required to Achieve Various NPV% Levels as a Function of

Enterprise Rate, ER.

Set 1: "Best" Design/Operating Parameters

ER,
% NPV%=2.5% NPV%=5% NPV%=7.5% NPV%=10% NPV%=12.5% NPV%=15%
0 13.5 18.8 24.1 29.4 34.7 40.0
4 20.0 26.8 33.6 40.4 47.1 53.9
8 27.2 35.7 44.2 52.7 61.2 69.7
12 34.9 45.4 55.9 66.3 76.8 87.3
16 43.2 55.9 68.6 81.3 93.9 106.6
20 52.0 66.9 82.2 97.4 112.5 127.7

Set 3: "Marginal" Design/Operating Parameters

ER,
% NPV%=2.5% NPV%=5% NPV%=7.5% NPV%=10% NPV%=12.5% NPV%=15%
0 15.7 22.4 29.2 35.9 42.7 49.4
4 25.0 34.8 44.5 54.3 64.1 73.8
8 36.3 50.1 63.9 77.6 91.4 105.2
12 49.8 68.9 87.9 106.9 125.9 145.0
16 65.9 91.6 117.4 143.2 168.9 194.7
20 84.8 119.0 153.3 187.6 221.8 256.1

Notes: *
The appropriate set of design/construction parameters is input to spreadsheet.
*
The selected value of ER is placed in spreadsheet.
* Then Excel’s Goal Seek Function is used to find ROIBT (i.e., either changing ProfitBT or FC or
both) to provide the desired value of NPV %.

28
 DAM: March 31, 2017

Table 5. Relationship Between ROIBT and IRR for

Construction/Operating Parameters:

1--"Best" 2--"Adequate" 3--"Marginal”

IRR ROIBT ROIBT ROIBT

60 108.5 151.5 215.7
55 99.1 135.6 188.8
50 89.8 120.3 163.6
45 80.5 105.6 140.3
40 71.4 91.6 118.9
35 62.4 78.3 99.2
30 53.6 65.8 81.2
25 45.1 54.0 65.0
20 36.9 43.1 50.4
15 28.9 33.0 37.6
10 21.5 23.8 26.5
5 14.6 15.7 16.9
0 8.2 8.5 9.0

29
 DAM: March 31, 2017

Table A.1. No-Loan Situation. Construction Costs Are Accrued with Interest Rate = ER & Booked by Enterprise at
Start-Up.

Fixed Capital and Profit_BT are the two independent variables. Internal Rate of Return Is Obtained via T & E Search
on ER.
FR is set equal to 0. ER is taken as the Discount Rate, and CR is set equal to ER. IRR is value of ER found by T&E
that sets NPV(0).

All dollar amounts in table

represent millions of dollars.
Construction & Operations
Periods are in years.

Profit_BT
= 33.6
Nconstru
Construction Tax
Rate 28.0% Rate 48% ction 2 Using Capitalization =
Enterp
Noperatio
rise Yield Tot.Cap.In TI=FC+W
Finance Rate Rate 28.0% ns 10 s v. C+SU
ROI_
BT = 55.2% 61.1%
Fixed
Capital 42.3 a-3 0.00
alpha_Workin 0.8 TI=FC+W
g Capital 20% a-2 0.00 b_1 0 C+SU 55.0
alpha_Start- 0.9
Up Capital 10% a-1 0.50 b_2 0
TCI=Total
alpha_Salvag 0.9 Capital
e Value 3% a0 0.50 b_3 5 Invested 60.9

Capital In (+) Discount Discounted

Ye DesignConstr or Out Factors Cash

30
 DAM: March 31, 2017

ar uction Period (-) Flows

Fixed Capital
-3 in Y-3 0.0 2.099 0.0
Fixed Capital
-2 in Y-2 0.0 1.639 0.0
Fixed Capital
-1 in Y-1 -21.2 1.280 -27.1
Fixed Capital
0 in Y0 -21.2 1.000 -21.2
Working
0 Capital -8.5 1.000 -8.5
Start-Up
0 Capital -4.2 1.000 -4.2

Total of
Capital
0 Outlays -60.9
(=Sum of
Constr.
DCFs)
Total Capital
Investment

Depreci Prof
Profit Bond ation it Cash
Afte
r
Operations Before Tax Flow
Period Taxes Financing Allowed es s
Construction
0 Costs, TCI 60.9 -60.9 1.000 -60.9
11.
1 26.9 0.0 -4.7 6 16.2 0.781 12.7
13.
2 30.2 0.0 -4.7 3 18.0 0.610 11.0
14.
3 31.9 0.0 -4.7 2 18.8 0.476 9.0
15.
4 33.6 0.0 -4.7 1 19.7 0.372 7.3

31
 DAM: March 31, 2017

15.
5 33.6 0.0 -4.7 1 19.7 0.291 5.7
15.
6 33.6 0.0 -4.7 1 19.7 0.227 4.5
15.
7 33.6 0.0 -4.7 1 19.7 0.177 3.5
15.
8 33.6 0.0 -4.7 1 19.7 0.139 2.7
15.
9 33.6 0.0 -4.7 1 19.7 0.108 2.1
15.
10 33.6 0.0 -4.7 1 19.7 0.085 1.7
Working
10 Capital 8.5 8.5 0.085 0.7
Salvage
10 Value 1.3 0.7 0.7 0.085 0.1

Tot
al
WC & Total Prof NPV-proj
SV Profit Bond Interest Total it Total NPV(0) [NPV(0)
Afte
r Discounte
All Figures Recov Before Depreci Tax Cash d to EOY(-
Represent ery Taxes Payments ation es Flow x)]
PV of
Operations== 45. <--- -
> 0.8 101.6 0.0 15.2 0 0.0 --> 0.0 0.0
Total
Bond Capital All Three
Recover
Repayment y NPVs NPV Increase per Year
Total
0.0 15.9 Cash Must Have normalized/annualized
Flow NPV_p
as Same NPV(0) roj
Net Present Sum Value Avg. Avg.
of To be
Value of Bonds NPVs Correct

32
 DAM: March 31, 2017

Over y Over z
0.0 0.0 <--- Years Years
0.0% 0.0%

x=
Nconstruction
y=
Nooperations
z = Nconstruction +
Noperations

33
 DAM: March 31, 2017

Table A.2. No-Loan Situation. Construction Costs Are Accrued at Rate ER & Booked by
Enterprise at Start-Up.
Excel IRR Function Finds IRR Directly (Cash Flows Must Be in Vector)
All dollar amounts in
table represent millions
of dollars. Construction
& Operations Periods
are in years.

Profit_B
T= 33.6
Nconstru
Construction Tax
Rate 13.5% Rate 48% ction 2 Using Capitalization =
Noperati
Enterpri Yield Tot.Cap.In TI=FC+W
Finance Rate se Rate 13.5% ons 10 s v. C+SU
ROI_
BT = 58.1% 61.1%
Fixed Capital 42.3 a-3 0.00
alpha_Working 0.8 TI=FC+W
Capital 20% a-2 0.00 b_1 0 C+SU 55.0
alpha_Start-Up 0.9
Capital 10% a-1 0.50 b_2 0
TCI=Total
alpha_Salvage 0.9 Capital
Value 3% a0 0.50 b_3 5 Invested 57.8

Discounte
Capital In (+) Discount d
Ye DesignConstructi or Out Cash
ar on Period (-) Factors Flows

Fixed Capital in
-3 Y-3 0.0 1.462 0.0

34
 DAM: March 31, 2017

Fixed Capital in
-2 Y-2 0.0 1.288 0.0
Fixed Capital in
-1 Y-1 -21.2 1.135 -24.0
Fixed Capital in
0 Y0 -21.2 1.000 -21.2
0 Working Capital -8.5 1.000 -8.5
0 Start-Up Capital -4.2 1.000 -4.2

Total of Capital
0 Outlays -57.8
(=Sum of Constr.
DCFs)
Total Capital
Investment

Depreci Prof
Profit Bond ation it Cash
Afte
r
Operations Before Tax Flow
Period Taxes Financing Allowed es s
Construction
0 Costs, TCI 57.8 -57.8 1.000 -57.8
11.
1 26.9 0.0 -4.7 6 16.2 0.881 14.3
13.
2 30.2 0.0 -4.7 3 18.0 0.776 13.9
14.
3 31.9 0.0 -4.7 2 18.8 0.684 12.9
15.
4 33.6 0.0 -4.7 1 19.7 0.603 11.9
15.
5 33.6 0.0 -4.7 1 19.7 0.531 10.5
15.
6 33.6 0.0 -4.7 1 19.7 0.468 9.2
15.
7 33.6 0.0 -4.7 1 19.7 0.412 8.1
15.
8 33.6 0.0 -4.7 1 19.7 0.363 7.2

35
 DAM: March 31, 2017

15.
9 33.6 0.0 -4.7 1 19.7 0.320 6.3
15.
10 33.6 0.0 -4.7 1 19.7 0.282 5.6
10 Working Capital 8.5 8.5 0.282 2.4
10 Salvage Value 1.3 0.7 0.7 0.282 0.2

Tot
al
WC & Total Prof NPV-proj
SV Profit Bond Interest Total it Total NPV(0) [NPV(0)
Afte
r Discount
All Figures Recov Before Depreci Tax Cash ed to
Represent ery Taxes Payments ation es Flow EOY(-x)]
PV of 75. <---
Operations==> 2.6 169.1 0.0 24.8 2 44.5 --><-- 44.5 34.6
Total
Bond Capital All Three
CFs from each Year Recover
Totaled and In Vector Repayment y NPVs NPV Increase per Year
Total Total
Year CF 0.0 27.1 Cash Must Have normalized/annualized
Flow NPV_
as Same NPV(0) proj
-3 0.0 Net Present Sum Value Avg. Avg.
of
NPV To be
-2 0.0 Value of Bonds s Correct
Over y Over z
-1 -21.2 0.0 44.5 <--- Years Years
0 -33.8 7.7% 5.0%
1 16.2
x=
2 18.0 Nconstruction
y=
3 18.8 Nooperations
4 19.7 z = Nconstruction +

36
 DAM: March 31, 2017

Noperations
5 19.7
6 19.7
7 19.7
8 19.7
9 19.7
<--
[Cells
I41+I42+
10 28.8 I43]

IRR by Excel
Function=IRR(D5
2:D65)--> 28.0%

37