Lecture 21

Rate of Return Problems

Unfortunately, the rate of return method suffers from a couple of problems, not the least of
which is that it’s a pain in the neck to calculate by hand! Another problem is that the equation

PW(i*) = 0

may have multiple solutions. To see why, you have to recognize that the present worth can
always be calculated by ignoring all of the fancy interest rate factors and simply discounting
each year’s net cash flow to Time 0:

N
PW i *   A 1  i *
n
n
n0

If we let x = (1+ i*)–1 then this equation can be rewritten as

N
PW i *  A x n
n
 A0  A1x  A2 x2    ANxN
n0

which is just an Nth-degree polynomial. So when we solve PW(i*) = 0 we are really just finding
the roots of this polynomial.

The number of roots is equal to the degree of the polynomial, which is the highest exponent
in the equation. In the present worth equation, the highest exponent is equal to the number
of compounding periods in the project life!

So a project with a 10-year life and annual compounding

may have as many as 10 different i* values!

Fortunately, not all of the roots of a polynomial are real and not all of the real roots are
non-negative. Since we’re only interested in non-negative real values for i*, the number of
potential i* values is actually much less than the degree of the polynomial suggests.

Descarte’s rule of signs says that the maximum number of real roots is equal to the
number of sign changes when the polynomial is written in canonical form. For example,

x3  x2  x  1  0

has at most one real root because the sign changes just once (from positive on one side of
x2 to negative on the other).

When applied to our problems, Descarte’s rule of signs means that the maximum number of
positive real roots is equal to the number of sign changes in the net cash flow diagram.

We have to use the net cash flow diagram because, in canonical form, a
polynomial has just one term per exponent (one cash flow per year).

7
In the Pigs-R-Us problem, the net cash flow diagram changes sign just once (between the
negative investment at Time 0 and the first net revenues in Year 1), so there is just one real
root (which we found to be 16.06%):

$41K/yr $51K

0
1 2 3 4 5 6 7 8 9 10

$200K

This is not the case in the second Pigs-R-Us example:

Pigs-R-Us Example 2 (REPRISE)

Pigs-R-Us is trying to decide whether or not to automate their meat packing process.
The machine costs $200,000, has a useful life of 10 years and a salvage value of
$10,000. It will require a $60,000 overhaul after 5 years of use. The machine costs
$9000 per year to operate and maintain, but it will save the company $50,000 per
year in labor costs. Pigs-R-Us has asked you to evaluate the economics of this
purchase and tell them what do to. If their MARR is 14%, what would you recommend?

If Pigs-R-Us needs to overhaul the machine at a cost of $60,000 at the end of Year 5, there
would be 3 sign changes and therefore as many as 3 real roots:

$41K/yr $41K/yr $51K

0 5
1 2 3 4 6 7 8 9 10

$19K

$200K

Sometimes we get lucky and all but one of the real roots is negative, in which case, the one
non-negative root is the correct value. That is the case with the second Pigs-R-Us project.

If two or more of the real roots are non-negative, then NONE of them is correct.

There’s a second rule we can apply to determine whether or not there’s more than one non-
negative real root. It’s called the cumulative cash flow test or Norstrom’s Criterion.

8
Start by calculating the cumulative cash flow series for the project. Each term in this series
is just the sum of all the cash flows up to that point in the project. If there’s only one sign
change in the cumulative cash flow series and the final value is not zero, then there’s only
one non-negative real root and that root is the “real” rate of return.

NOTE: It is not necessary for the cash flows to start negative as stated in the book but
it is necessary for the final value to be nonzero, otherwise IRR = 0 will automatically
satisfy PW=0, meaning there are at least 2 potential solutions to PW=0.

If there’s more than one sign change, the test is inconclusive. There may be multiple non-
negative roots or there may not be multiple non-negative roots. You can’t say either way.

For Pigs-R-Us Example 2, the cumulative cash flow series (in $1000s) is as follows:

Year 0: -200
Year 1: -200 + 41 = -159
Year 2: -159 + 41 = -118
Year 3: -118 + 41 = -77
Year 4: -77 + 41 = -36
Year 5: -36 – 19 = -55
Year 6: -55 + 41 = -14
Year 7: -14 + 41 = 27
Year 8: 27 + 41 = 68
Year 9: 68 + 41 = 109
Year 10: 109 + 51 = 160

Let’s put these in tabular form so they’re easier to see:

0 1 2 3 4 5 6 7 8 9 10
-200 -159 -118 -77 -36 -55 -14 27 68 109 160

Since there is only one sign change (between Years 6 and 7) and the final value is nonzero,
there is only one non-negative root and we can solve for it as we would any rate-of-return
problem.

We’ll start by calculating the present worth using the 14% MARR:

5.2161 0.5194 0.2697

PW  –$200K  $41K P | A,14%,10  $60K P | F,14%,5  $10K P | F,14%,10   $1 4.61
K
1 2 3 4

This time the present worth is less than zero, which means our actual rate of return is less
than 14%, so we need to try a lower rate. The next lower rate in the textbook is 12%:

5.6502 0.5674 0.3220

–$200K  $41K P | A,12%,10  $60K P | F,12%,5  $10K P | F,12%,10  $0.83 42K
1 2 3 4

We have now bracketed the answer between two adjacent tables (12% and 14%) that provide
PW values with opposite signs and we can interpolate the final answer:

9
  0  0.8342 
i*  12%  (14%  12%)    12.11%
 14.60  0.8342  1 2 3 4

This value is less than the 14% MARR, so Pigs-R-Us should not purchase the equipment if
they are going to have to overhaul it halfway through its service life. Of course we already
knew that because the present worth was less than zero!

REINVESTMENT RATE ISSUE

Another problem with the rate of return method is that the calculated rate of return can make
a good project look better—and a bad project look worse—than it really is. The reason is that
the rate of return assumes all of the money stays internal to the project, which is why we
often refer to it as the internal rate of return.

For example, let’s say Company X has invested $500,000 in an equipment upgrade that will
save them $200,000 per year over the next 10 years. The rate of return on this project is an
excellent 38.5% per year.

But what actually happens to the $200,000 they save each year? That money will be
reinvested elsewhere in the company as retained earnings and probably earn a lot less than
38.5% per year. If the company’s MARR is 15% per year, that money will probably earn
something a lot closer to 15% than 38.5%. In fact, the standard assumption in Engineering
Economics is that the money will earn exactly the MARR.

On the other hand, let’s say Company X has to come up with $300,000 at some point to
overhaul the equipment. They may pay that bill using retained earnings from some other
project in the company. Or they may have to raise capital to make up the shortfall. Either
way, is will probably cost them a lot less than 38.5% to raise that capital.

So for good projects (i.e., ones with i* greater than the MARR), the excess cash will actually
earn something less than i* and for bad projects, the borrowed cash will cost something less
than i*. Thus the rate of return method makes good projects look better than they really
are and makes bad projects look worse than they really are. The greater the difference
between the MARR and the calculated rate of return, the greater the disparity.

EXTERNAL RATE OF RETURN

One way we can get around all of the problems above is to calculate an external rate of
return instead.

The phrase external rate of return reflects the fact that, unlike the internal rate of return,
it is calculated by assuming that excess revenue gets reinvested outside of the project and
shortfalls are covered by borrowing from outside of the project.

The rate earned by funds reinvested outside of the project is called the reinvestment rate
(ir) and is usually set equal to the MARR for the reasons mentioned above.

10
The cost of funds borrowed from outside of the project is called the borrowing rate (ib). One
obvious value for this rate is the WACC because that’s the interest rate they’ll have to pay to
raise capital. If, however, the company uses retained earnings to fund the shortfall, the MARR
is a better choice because it represents the lost opportunity cost of those retained earnings
as we discussed in Chapter 13. (If we didn’t use those retained earnings for our project, we
could have used them to fund Project X—the next one on the list—which we used to set our
MARR).

The textbook provides two methods for calculating an external rate of return:
MIRR and ROIC. We’ll only look at the MIRR method because it’s the easiest.

The modified rate of return method (or MIRR method) uses reasonable borrowing and
reinvestment rates in place of i* and it always produces a single, unique solution.

To calculate the external rate of return (EROR) using the MIRR method requires four simple
steps:

1. Draw the net cash flow diagram (one cash flow per interest period).
2. Discount all of the negative net cash flows to the project start using ib.
3. Compound all of the positive net cash flows to the project end using ir.
4. Calculate the EROR as the interest rate that equates the two.

To see how this works, let’s return to the Pigs-R-Us problem with the added equipment
overhaul in Year 5:

Pigs-R-Us Example 2 (reprise)

Pigs-R-Us is trying to decide whether or not to automate their meat packing process. The
machine costs $200,000, has a useful life of 10 years and a salvage value of $10,000.
It will require a $60,000 overhaul after 5 years of use. The machine costs $9000 per
year to operate and maintain, but it will save the company $50,000 per year in labor
costs. Pigs-R-Us has asked you to evaluate the economics of this purchase and tell them
what do to. If their MARR is 14%, what would you recommend?

For our first solution, we’ll assume that any borrowed money will come from retained earnings
that could have been invested in Project X and earned the MARR. Likewise, we’ll assume that
any excess funds could be invested in that same Project X and therefore will earn the MARR
as well. Thus, we’re assuming (for now) that

ib = ir = MARR

If you don’t have any better information to go on, this is the best you can do.

We’ll start by drawing the net cash flow diagram for the project:

11
 $51K
$41K/yr $41K/yr

0 5
1 2 3 4 6 7 8 9 10

$19K

$200K

The two negative net cash flows occur at Year 0 and Year 5, so we discount both of them to
Year 0 (the start of the project) using the 14% MARR as the borrowing rate:

P = – $200K – $19K (1.14)–5 = – $210K

This is simply the present worth of all the negative cash flows.

The positive net cash flows in Years 1-4 and 6-10 are then compounded to Year 10 using the
14% MARR as the reinvestment rate:
4.9211 4.9211
F  $41K F | A,14%, 4  1.14   $41K F | A,14%, 4 1.14   $51K  $724K
6 1

This is simply the future worth of all the positive cash flows.

Now we find the interest rate that would turn a $210K investment at Time 0 into a $724K
balance 10 years later:

$210 (1 + EROR)10 = $724

EROR = (724/210)1/10 – 1 = 13.2%

This is the external rate of return for this project under the assumption that the
borrowing rate and the reinvestment rate are both equal to the MARR.

Recall that our earlier calculation of the rate of return (the internal rate of return) was 12.1%,
which was less than the MARR. The somewhat more-realistic external rate of return is 13.2%,
which isn’t as bad when compared to the 14% MARR.

The EROR will always lie between the IRR and the MARR, regardless of whether the IRR is
higher or lower than the MARR.

Now, what would happen if we used something other than the MARR for the borrowing rate?
Let’s assume Pig-R-Us had a 10% per year WACC, so ib = 10% per year, ir = 14% per year.

The present worth of all the negative cash flows is now

P = – $200K – $19K (1.10)–5 = – $212K

12
The future worth of all the positive cash flows is the same as before:
4.9211 4.9211
F  $41K F | A,14%, 4  1.14   $41K F | A,14%, 4 1.14   $51K  $724K
6 1

Now we find the interest rate that would turn a $212K investment at Time 0 into a $724K
balance 10 years later:

$212 (1 + EROR)10 = $724

EROR = (724/212)1/10 – 1 = 13.1%

So with different assumptions as to the borrowing rate, we get a different EROR. Unlike the
IRR (if it exists), the EROR is not unique. It depends entirely on what you assume for the two
external interest rates.

Let’s go back to the Lockheed-Martin Example and redo it using this MIRR method.

Lockheed-Martin Example 2

Lockheed Martin has secured a satellite launch contract from a European communications
company that will pay them €3.9M per year for the next 8 years. To land that contract,
they invested €13M in a satellite tracking system. Of that amount, €8M was paid up front
and the remaining €5M was paid during the first year of operation. The annual operating
costs for the satellite tracking system are estimated at €0.9M per year (which includes the
cost of personnel, electricity, maintenance, etc.). At the end of the contract, it is estimated
that the equipment will have a salvage value of €0.5M. What is Lockheed Martin’s EROR
on this project if their MARR is 15% and their WACC is 10%?

First, we need to draw the net cash flow diagram:

€3M/yr €3.5M

0 1
2 3 4 5 6 7 8
€2M

€8M

Discounting the negative net cash flows to Time 0 using the WACC as the borrowing rate:

0.9091
P  8M  2M P | F,10%,1  9.82M

Compounding the positive net cash flows to Time 8 using the MARR as the reinvestment rate:

13
 11.0668
F  3M F | A,15%,7   0.5M  33.7M

Now we find the interest rate that would turn a €9.82M investment at Time 0 into a €33.7M
balance 8 years later:

9.82(1+EROR)8 = 33.7

(1+EROR)8 = 3.43

EROR = 3.431/8 – 1 = 16.7%

In the last lecture, we found that the internal rate of return for this project was 18.4%, so
the EROR is once again closer to the MARR than the IRR was.

What would happen if we set both the borrowing rate and the reinvestment rate equal to the
IRR that we calculated earlier?

P  8M  2M (1.184)1  9.69M
12.2926
F  3M F | A,18.4%,7  0.5M  37.4M

9.69(1+EROR)8 = 37.4

(1+EROR)8 = 3.86

EROR = 3.861/8 – 1 = 18.4%

This clearly shows that the IRR method is equivalent to using the IRR as both the borrowing
rate and the reinvestment rate. This implies that excess cash is reinvested in the project at
the IRR and cash shortfalls are made up by borrowing from the project at the IRR, which is
the same as saying that all of the money stays internal to the project.

Note that the EROR method takes care of several problems.

1. It provides a more-realistic return-on-investment.

2. It doesn’t require a trial-and-error solution.
3. It provides a single, unique answer (there are no multiple roots).

Keep in mind that the EROR is not necessarily “better” than the IRR; it just uses different
assumptions as to how the individual cash flows are discounted or compounded. Both methods
provide a measure of the company’s return on investment.

14