Chapter 27
The Real Options Model of Land Value and
Development Project Valuation
Major references include:
• J.Cox & M.Rubinstein, “Options Markets,” Prentice-Hall, 1985
• L.Trigeorgis, “Real Options,” MIT Press, 1996
• T.Arnold & T.Crack, “Option Pricing in the Real World: A Generalized
Binomial Model with Applications to Real Options,” Dept of Finance,
University of Richmond, Working Paper, April 15, 2003 (available on the
Financial Economics Network (FEN) on the Social Science Research
Network at www.ssrn.com).
© 2014 OnCourse Learning. All Rights Reserved.
1
Exhibit 2-2: The “Real Estate System”: Interaction of the Space Market, Asset Market, & Development Industry
SPACE MARKET
SUPPLY
(Landlords)
ADDS
NEW
LOCAL
&
NATIONAL
ECONOMY
DEMAND
(Tenants)
RENTS
&
OCCUPANCY
FORECAST
FUTURE
DEVELOPMENT
INDUSTRY
ASSET MARKET
IF
YES
IS
DEVELPT
PROFITABLE
?
CONSTR
COST
INCLU
LAND
SUPPLY
(Owners
Selling)
CASH
FLOW
PROPERTY
MARKET
VALUE
MKT
REQ’D
CAP
RATE
CAPI
TAL
MKTS
DEMAND
(Investors
Buying)
= Causal flows.
= Information gathering & use.
© 2014 OnCourse Learning. All Rights Reserved.
2
Land value plays a pivotal role in determining whether, when,
and what type of development will (and should) occur.
Relationship is two-way:
Land
Value
Optimal
Devlpt
• From a finance/investments perspective:
- Development activity links the asset & space markets;
- Determines L.R. supply of space, L.R. rents.
- Greatly affects profitability, returns in the asset market.
• From an urban planning perspective:
- Development activity determines urban form;
- Affects physical, economic, social character of city.
Recall relation of land value to land use boundaries noted in Ch.5…
© 2014 OnCourse Learning. All Rights Reserved.
3
Effect of Urban Growth & Uncertainty on Land Rents & Land Values Just
Beyond the Urban Boundary
When you develop today, you give
up the option to develop tomorrow
instead. Devlpr needs to be
compensated for loss of this option.
Exhibit 5-1: Components of Land Rent Outside &
Inside the Urban Boundary, Under Uncertainty . .
.
Location Rent
Irreversibility Rent
Construction Rent
Agricultural Rent
B
Distance from
CBD
B = Urban
Boundary
Exhibit 5-2: Components of Land Value Outside &
Inside the Urban Boundary, Under Uncertainty . . .
PV of property asset includes PV of
expected future growth in rents
developed property can earn after its
development.
Location Value
Construction Cost
Irreversibility Premium
Growth Premium
Agricultural Value
Distance from CBD
B
B = Urban Boundary
© 2014 OnCourse Learning. All Rights Reserved.
4
Effect of Urban Growth & Uncertainty on Land Rents & Land Values Just
Beyond the Urban Boundary
Exhibit 5-1: Components of Land Rent Outside &
Inside the Urban Boundary, Under Uncertainty . .
.
Location Rent
Irreversibility Rent
As boundary expands, land
value just beyond boundary
can grow rapidly, due to
increase in growth premium
& irreversibility (option)
premium values, depending
on how fast the boundary is
expanding, the magnitude
of the location value rent
gradient inside the
boundary, and the
magnitude of uncertainty
(volatility) in that growth.
Construction Rent
Agricultural Rent
B
Distance from
CBD
B = Urban
Boundary
Exhibit 5-2: Components of Land Value Outside &
Inside the Urban Boundary, Under Uncertainty . . .
Location Value
Construction Cost
Irreversibility Premium
Growth Premium
Agricultural Value
Distance from CBD
B
B = Urban Boundary
Direction of bdy expansion
Direction of time flow
© 2014 OnCourse Learning. All Rights Reserved.
5
Different conceptions of “land value” (Recall Property Life Cycle theory from Ch.5)
Property Value, Location Value, & Land Value
Evolution of the Value (& com ponents) of a Fixed Site (parcel)
3.0
HBU Value As If
Vacant = Potential
Usage Value, or
"Location Value"
("U")
Value Levels ($)
2.5
Property Value
("P") = Mkt Val
(MV) = Structure
Value + Land
Value.
2.0
1.5
Land Value by
Legal/Appraisal
Defn. ("land comps
MV").
1.0
0.5
0.0
C
C
C
C
Land Value by
Econ.Defn. =
Redevlpt Option
Value. ("LAND")
Time
("C" Indicates Reconstruction times)
© 2014 OnCourse Learning. All Rights Reserved.
In Ch.27 we
focus on the
Econ.Defn.:
“LAND”
6
Different conceptions of “land value”
Property Value, Location Value, & Land Value
Evolution of the Value (& com ponents) of a Fixed Site (parcel)
3.0
HBU Value As If
Vacant = Potential
Usage Value, or
"Location Value"
("U")
Value Levels ($)
2.5
Property Value
("P") = Mkt Val
(MV) = Structure
Value + Land
Value.
2.0
1.5
Land Value by
Legal/Appraisal
Defn. ("land comps
MV").
1.0
0.5
0.0
C
C
C
C
Land Value by
Econ.Defn. =
Redevlpt Option
Value. ("LAND")
Time
("C" Indicates Reconstruction times)
© 2014 OnCourse Learning. All Rights Reserved.
Note that there
are points in
time when
three of the
four definitions
all give the
same value,
namely,
property value
= land value
defined by
either defn at
the times of
optimal
redevelopment
(construction)
on the site.
7
The economic definition of land value (“LAND”) is based on
nothing more or less than the fundamental capability that land
ownership gives to the landowner (unencumbered):
The right without obligation to develop (or redevelop)
the property.
© 2014 OnCourse Learning. All Rights Reserved.
8
This definition of land value is most relevant
Evolution of the Value (& com ponents) of a Fixed Site (parcel)
3.0
Value Levels ($)
2.5
2.0
1.5
1.0
0.5
0.0
C
C
C
Time
("C" Indicates Reconstruction times)
C
Just prior to the times when development or
redevelopment occurs on the site.
© 2014 OnCourse Learning. All Rights Reserved.
9
To understand the economic conception of land value, a
famous theoretical development from financial economics
is most useful: “Option Valuation Theory” (OVT) :
In particular, a branch of that theory known a “Real
Options.”
© 2014 OnCourse Learning. All Rights Reserved.
10
27.1 Call Options: Some Basic Background & Definitions
Financial economics definition of “Option” :
The right without obligation to obtain something of value
upon the payment or giving up of something else of value.
Option Terminology:
• “Owner” or “holder” of the option = Person having this right.
• “Underlying Asset” = The asset which is obtained by the “exercise” of the option.
• “Exercise Price” (or “strike price”) = What is given up to exercise the option.
• “Maturity” or “expiration date” = Time by when the option must be exercised or lost.
• “American Option” = Option can be exercised any time prior to maturity.
• “European Option” = Option can only be exercised on its maturity (expiration) date.
• “Call Option” = Right to buy the underlying asset upon payment of exercise price.
• “Put Option” = Right to sell underlying asset for a specified price.
• “Put/Call Parity” = Put option on X @ price Y = Call option on Y @ price X.
• “Compound Option” = An option on an option.
• “Perpetual Option” = No expiration (infinite maturity).
© 2014 OnCourse Learning. All Rights Reserved.
11
Option exercise is irreversible:
Options can only be exercised once. Then the option is lost.
OVT is a body of theory and methodology for quantitatively
evaluating options.
• For American options, this includes the problem of
specifying the conditions when it is optimal to exercise the
option: Optimal Exercise Policy.
• For the land development option, this relates to the conditions
when it is optimal to develop (or redevelop) the land.
© 2014 OnCourse Learning. All Rights Reserved.
12
Some history:
Call option model of land arose from two strands of theory:
• Financial economics study of corporate capital budgeting,
• Urban economics study of urban spatial form.
Capital Budgeting:
• How corporations should make capital investment decisions
(constructing physical plant, long-lived productive assets).
• Includes question of optimal timing of investment.
• e.g., McDonald, Siegel, Myers, (others), 1970s-80s.
Urban Economics:
• What determines density and rate of urban development.
• Titman, Williams, Capozza, (others), 1980s.
It turned out the 1965 Samuelson-McKean Model of a perpetual
American warrant was the essence of what they were all using.
© 2014 OnCourse Learning. All Rights Reserved.
13
The value of newly-built property will include the (remaining)
economic value of the land (“LAND”), but:
• This land value will now be minus its initial development
option value (which has been exercised),
• It will include the redevelopment (or abandonment) option
value.
• However, this redevelopment option value will normally be
quite small in a newly developed property (recall property
“life-cycle” chart).
Evolution of the Value (& com ponents) of a Fixed Site (parcel)
3.0
Value Levels ($)
2.5
2.0
1.5
1.0
0.5
0.0
C
C
C
Time
("C" Indicates Reconstruction times)
© 2014 OnCourse Learning. All Rights Reserved.
C
Speaking in
terms of the
economic
definition of
land value,
based on its
option value.
14
The Land Development Option Contrasted with Financial Options:
Distinguishing characteristics of the land devlpt option:
• Perpetual (no expiriation):
• More flexibility (greater value),
• Only reason to exercise is to obtain operating cash flows.
• “Time to Build” (exercise not immediate):
• Can’t observe exact at-completion mkt val of underl.asset at time
exercise decision is made (added risk in exercise decision).
• “Noisy” value observation of (even current) mkt val of underl. asset.
(“thin mkt”, recall Ch.12, also adds to risk of exercise decision):
• Possibly heterogeneous information about true value of underlying asset
(the to-be-built property): Some devlprs may be more knowledgable than
others. ( Wait longer until exercise.)
• Exercise creates new real assets that add to the supply side of the space
market (affecting mkt val of all competing properties):
• Can increase risk of not exercising (option may effectively “expire” if
demand is absorbed by competing devlpt projects).
© 2014 OnCourse Learning. All Rights Reserved.
15
What the real option theory of land development can tell us
about the “overbuilding phenomenon”
What is the “overbuilding phenomenon”?
The widely observed tendency for commercial real estate
markets to periodically become “overbuilt”, that is,
characterized by excess supply (abnormally high vacancy,
downward pressure on rents), due to excessive speculative
development of new buildings.
Recall that in Chapter 2 we discussed an explanation for this
“cyclicality” phenomenon using the “4-Quadrant Diagram”,
based on the existence of myopic behavior (not just lack of
perfect foresight, but some degree of irrational expectations)
on the part of investors and developers in the system .
© 2014 OnCourse Learning. All Rights Reserved.
16
Exhibit 2-4a: Effect of Demand Growth in Space Market:
Rent $
Space Market:
Rent Determination
D0
D1
R*
Price $
P*
Q*
Stock (SF)
© 2014 OnCourse Learning. All Rights Reserved.
Asset Market:
Valuation
C*
Asset Market:
Construction
Space Market:
Stock Adjustment
Construction (SF)
17
Exhibit 2-4a: Effect of Demand Growth in Space Market: First phase…
Rent $
Can this be a long-
Space Market:
run equilibrium
Rent Determination
Asset Market:
Valuation
D0
R1
D1
Doesn’t form a
rectangle.
R*
Price $
P1
P*
Q*
Stock (SF)
© 2014 OnCourse Learning. All Rights Reserved.
result?…
C*
Asset Market:
Construction
Space Market:
Stock Adjustment
Construction (SF)
18
Exhibit 2-4a: Effect of Demand Growth in Space Market: LR Equilibrium…
Rent $
Space Market:
Rent Determination
Asset Market:
Valuation
D0
R1
R**
D1
R*
P**
Price $
P1
P*
Q*
Q** Stock (SF)
C*
Asset Market:
Construction
C**
Space Market:
Stock Adjustment
Construction (SF)
© 2014 OnCourse Learning. All Rights Reserved.
19
Real option theory offers several explanations for why/how
overbuilding can be due to completely rational (i.e., profitmaximizing) behavior on the part of developers (landowners):
1. “Cascades”: Noisy observations of the mkt values of the underlying assets
(comparable built properties), combined with heterogeneous developer
knowledge about the “true” value, causes a follow-the-leader type effect, in
which developers wait longer than they otherwise would to develop, and
then they all rush in as soon as the first (presumably most knowledgeable)
developer reveals his knowledge by commencing development.
2. “Lumpy supply & first out of the gate”: Economies of scale in building
size, combined with finite user demand and the fact that option exercise
creates real physical capital, leads to early exercise of the development
option to preclude loss (expiration) of the option if a competitor builds first.
3. “Long-term leasing option”: The cost of having empty space in a new
building may be less than it first appears in space markets characterized by
long-term leases, as it gives the landlord a leasing option, that has value
prior to its “exercise” (in the signing of a lease contract): Volatility in the
rental mkt may bring better long-term lease deals in the future.
© 2014 OnCourse Learning. All Rights Reserved.
20
Chapter 27: Real Options & Land Value
Real Options:
Options whose underlying assets (either what is obtained or
what is given up on the exercise of the option) are real assets
(i.e., physical capital).
The call option model of land value (introduced in Chapter 5) is a
real option model:
Land ownership gives the owner the right without obligation to develop (or
redevelop) the property upon payment of the construction cost. Built
property is underlying asset, construction cost is exercise price (including
the opportunity cost of the loss of any pre-existing structure that must be
torn down).
In essence, all real estate development projects are real options,
though in some simple cases the optionality may be fairly trivial
and can be safely ignored.
© 2014 OnCourse Learning. All Rights Reserved.
21
What we’re going to try to do in these lectures
In the typical development project (or parcel of developable
land), there are three major types of options that may present
themselves:*
• “Wait Option”: The option to delay start of the project
construction (Ch.27);
• “Phasing Option”: The breaking of the project into
sequential phases rather than building it all at once
(Ch.29);
• “Switch Option”: The option to choose among alternative
types of buildings to construct on the given land parcel.
All three of these types of options can affect optimal
investment decision-making, add significantly to the value of
the project (and of the land), affect the risk and return
characteristics of the investment, and they are difficult to
accurately account for in traditional DCF investment analysis.
© 2014 OnCourse Learning. All Rights Reserved.
22
What we’re going to try to do in this lecture
Real options valuation theory (OVT) provides a framework to
rigorously account for these options in the valuation and
optimal investment decision-making regarding development
projects.
OVT is a very sophisticated and highly technical subject that
has yet to penetrate much into real world practical application.
This lecture will attempt to provide you with sufficient
knowledge and methodological tools so that you can apply
OVT reasonably effectively in real world situations regarding
the first two types of options: the Wait Option and the Phasing
Option. The Switch Option is too complex for us to attempt in
this course.
This lecture is thus an attempt to take you “beyond the cutting
edge” not only of current practice but also of current pedagogy
in graduate real estate education. Nevertheless…
You can do it! (Fasten your seatbelts and HAVE FUN!)
© 2014 OnCourse Learning. All Rights Reserved.
23
What we’re going to try to do in this lecture
We will begin by considering the Wait Option, which is the
simplest of the three types of options . . .
© 2014 OnCourse Learning. All Rights Reserved.
24
27.2 A Simple Numerical Example of OVT Applied to Land
Valuation and the Development Timing Decision
Today
Probability
Next Year
100%
30%
70%
Value of Developed Property
$100.00
$78.62
$113.21
Development Cost (exclu land)
$88.24
$90.00
$90.00
NPV of exercise
$11.76
-$11.38
$23.21
(Don’t build)
(Build)
0
$23.21
(Action)
Future Values
Expected Values
= Sum[ Probability X Outcome ]
$11.76
$16.25
(1.0)11.76
(0.3)0 + (0.7)23.21
PV(today) of Alternatives @20%
$11.76
16.25 / 1.2 = $13.54
Note: In this example the expected growth in the HBU value of the built property is 2.83%:
as (.3)78.62 + (.7)113.21 = $102.83.
What is the value of this land today? Answer: = MAX[11.76, 13.54] = $13.54
Should owner build now or wait? Answer: = Wait. (100.00 – 88.24 – 13.54< 0.)
The $13.54 – $11.76 = $1.78 option premium is due to uncertainty or volatility.
© 2014 OnCourse Learning. All Rights Reserved.
25
Consider the effect of uncertainty (or volatility) in the evolution of the built
property value (for whatever building would be built on the site), and the
fact that development at any given time is mutually exclusive with
development at any other time on the same site (“irreversibility”). e.g.:
Today
Probability
Next Year
100%
30%
70%
Value of Developed Property
$100.00
$78.62
$113.21
Development Cost (exclu land)
$88.24
$90.00
$90.00
NPV of exercise
$11.76
-$11.38
$23.21
(Don’t build)
(Build)
0
$23.21
(Action)
Future Values
Expected Values
$11.76
$16.25
= Sum[ Probability X Outcome ]
(1.0)11.76
(0.3)0 + (0.7)23.21
PV(today) of Alternatives @20%
$11.76
16.25 / 1.2 = $13.54
Note the importance of flexibility inherent in the option (“right without
obligation”), which allows the negative downside outcome to be avoided.
This gives the option a positive value and results in the “irreversibility
premium” in the land value (noted in Geltner-Miller Ch.5).
© 2014 OnCourse Learning. All Rights Reserved.
26
Representation of the preceding problem as a “decision tree”:
• Identify decisions and alternatives (nodes & branches).
• Assign probabilities (sum across all branches @ ea. node = 100%).
• Locate nodes in time.
• Assume “rational” (highest value) decision will be made at each node.
• Discount node expected values (means) across time reflecting risk.
Build: Get
113.21-90.00
= $23.21
70%
Wait Today:
PV = 16.25/1.2
= 13.54.
Choice Next Yr.:
1 Yr
Node Value =
(7.)23.21+ (.3)0 =
$16.25.
Choice
30%
Today
Don’t build:
Get 0.
Build Today:
Get 100.0088.24 =
$11.76.
Decision Tree Analysis is closely related to
Option Valuation Methodology, but requires a
different type of simplification (finite number of
discrete alternatives).
© 2014 OnCourse Learning. All Rights Reserved.
27
A problem with traditional decision tree analysis
We were only able to completely evaluate this decision because we
somehow knew what we thought to be the appropriate riskadjusted discount rate to apply to it (here assumed to be 20%).
Build: Get
113.21-90.00
= $23.21
70%
Wait Today:
PV = 16.25/1.2
= 13.54.
Choice Next Yr.:
1 Yr
Node Value =
(7.)23.21+ (.3)0 =
$16.25.
Choice
30%
Today
Don’t build:
Get 0.
Build Today:
Get 100.0088.24 =
$11.76.
But is this really the correct discount rate
(and hence, the correct decision and
valuation of the project)?...
© 2014 OnCourse Learning. All Rights Reserved.
28
Where did the 20% discount rate (OCC) come from anyway?
To be honest…
It was a nice round number that seemed “in the ballpark” for
required returns on development investment projects.
Build: Get
113.21-90.00
= $23.21
70%
Wait Today:
PV = 16.25/1.2
= 13.54.
Choice Next Yr.:
1 Yr
Node Value =
(7.)23.21+ (.3)0 =
$16.25.
Choice
30%
Today
Don’t build:
Get 0.
Build Today:
Get 100.0088.24 =
$11.76.
Can we be a bit more “scientific” or
rigorous? . . .
© 2014 OnCourse Learning. All Rights Reserved.
29
27.3.1. The Essence of Option Valuation Theory
Suppose there were “complete markets” in land, and buildings, and
bonds, such that we could buy or sell (short if necessary) infinitely
divisible quantities of each, including land and buildings like our subject
development project.
Thus, we could buy today:
• 0.67 units of a building just like the one our subject development would
produce next year that will either be worth $113.21 or $78.62 then.
And we could partially finance this purchase by issuing:
• $51.21 worth of riskless bonds (with a 3% interest rate).
Then this “replicating portfolio” (long in the bldg, short in the bond)
next year will be worth:
• In the “up” scenario: (0.67)$113.21 - $51.21(1.03) = $75.95 – $52.74 =
$23.21, or:
• In the “down” scenario: (0.67)$78.62 - $ 51.21(1.03) = $52.74 – $52.74 = 0.
Exactly Equal to the Development Project in All Future Scenarios!
© 2014 OnCourse Learning. All Rights Reserved.
30
The Essence of OVT
Recall:
These are the future scenarios, describing all
possible future outcomes.
Build: Get
113.21-90.00
= $23.21
70%
Wait Today:
PV = 16.25/1.2
= 13.54.
Choice Next Yr.:
1 Yr
Node Value =
(7.)23.21+ (.3)0 =
$16.25.
30%
Choice
Today
Build Today:
Get 100.0088.24 =
$11.76.
In the upside outcome, the project
will be worth $23.21, same as the
replicating portfolio.
Don’t build:
Get 0.
In the downside outcome, the
project will be worth 0, same as
the replicating portfolio.
© 2014 OnCourse Learning. All Rights Reserved.
31
The Essence of OVT
Thus, this “replicating portfolio” must be worth the same as the land
(the development option) today.
Suppose not:
• If the land can be bought for less than the replicating portfolio, then I
can sell the replicating portfolio short, buy the land, pocket the
difference as profit today, and have zero net value impact next year (as
the land and replicating portfolio will in all cases be worth the same
next year, so my long position offsets my short position exactly).
• If the land costs more than the replicating portfolio, then I can sell
the land short, buy the replicating portfolio, pocket the difference as
profit today, and once again have zero net impact next year.
This is what is known as an “arbitrage” – riskless profit!
In equilibrium (within and across markets), arbitrage opportunities
cannot exist, for they would be bid away by competing market
participants seeking to earn super-normal profits.
© 2014 OnCourse Learning. All Rights Reserved.
32
The Essence of OVT
In real estate, markets are not so perfect and complete to enable
actual construction of technical arbitrage. But nevertheless
competition tends to eliminate super-normal profit, so we can use this
kind of analysis to model prices and values.
Fundamentally, this approach will always equalize the expected
return risk premium per unit of risk, across the asset markets.
So, how much is the land worth in our example . . .
The replicating portfolio is:
(0.67)V(0) - $51.21
And thus must have this value.
The only question is, what is the value of V(0), the value of the
underlying asset (the project to be developed) today (time-0)?...
© 2014 OnCourse Learning. All Rights Reserved.
33
The Essence of OVT
We know that a similar asset already completed today is worth $100.00.
However, this value includes the value of the net cash flow (dividends,
rents) that asset will pay between today and next year.
Our development project won’t produce those dividends, because it
won’t produce a building until next year.
So, we need a little more analysis…
Suppose that the underlying asset (the built property) has an expected
total return of 9%.
If a similar building has a value today of $100.00, and an (ex dividend)
value next year of either $113.21 (70% chance) or $78.62 (30% chance),
then the expected value next year is (0.7)113.21+(0.3)78.62 = $102.83
(i.e., expected growth is E[gV]=2.83%).
Thus, the PV today of a building that would not exist until next year
(i.e., PV of similar pre-existing building net of its cash flow between now
and next year) is:
PV[V1] = V(0) = $102.83 / 1.09 = $94.34.
(versus V0 = $100.00 for pre-existing bldg.)
© 2014 OnCourse Learning. All Rights Reserved.
34
The Essence of OVT
Now we can value the option by valuing the replicating portfolio:
C0 = (0.67)V(0) - $51.21
= (0.67)$94.34 - $51.21
= $63.29 - $51.21
= $12.09.
Thus, our previous estimate of $13.54 (based on the 20% OCC) was
apparently not correct. The option is actually worth $1.45 less.
© 2014 OnCourse Learning. All Rights Reserved.
35
Suppose we could sell the option for $13.54
Then we could (with complete markets):
• Sell the option (short) for $13.54, take in $13.54 cash.
•Borrow $51.21 at 3% interest (with no possibility of default), thereby take in
another $51.21 cash.
• Use part of the resulting $64.75 proceeds to buy 0.67 units of a building just
like the one to be built (minus its net rent for this coming year), for a price of
(0.67)$94.34 = $63.29.
• Our net cash flow at time 0 is: +$64.75 – $63.29 = + $1.46.
• A year from now, we face:
• In the “up” outcome:
• We must pay to the owner of the option we sold $23.21 = $113.21 - $90, the
value of the development option.
• We must pay off our loan for (1.03)$51.21 = $52.74.
• We will sell our .67 share of the building for (.67)$113.21 = $75.95 cash
proceeds.
• Giving us a total net cash flow next year of $75.95 – ($23.21 + $52.74) =
$75.95 - $75.95 = 0.
• In the “down” outcome:
• We owe the owner of the option nothing, but we still owe the bank $52.74.
• We sell our .67 share of the building for (.67)$78.62 = $52.74 cash proceeds.
• Giving us a total net cash flow next year of 0.
• Thus, we make a riskless profit at time 0 of +$1.46.
• We could perform arbitrage for any option price other than $12.09.
© 2014 OnCourse Learning. All Rights Reserved.
36
27.3.1: The Arbitrage Perspective on the Option Value…
Here is another way of depicting what we have just suggested:
Today
Next Year
Development Option
Value
C = Max[0,V-K]
PV[C1] = x
“x” = unknown value,
x = P0 , otherwise
arbitrage..
C1up =113.21-90
= $23.21
C1down = 0
(Don’t build)
Built Property Value
PV[V1] = E[V1] / (1+OCC)
=
[(.7)113.21+(.3)78.62]/1.09
= $102.83/1.09 =$94.34
V1up = $113.21
V1down = $78.62
B = $51.21
B1 = (1+rf )B =
(1.03)51.21
= $52.74
B1 = (1+rf )B =
(1.03)51.21
= $52.74
P0 = (N) PV[V1] – B
= (0.67)$94.34 - $51.21
= $63.29 - $51.21 = $12.09
P1up =(0.67)113.21 $52.74
= $75.95 - $$52.74
= $23.21
P1down =(0.67)78.62 $52.74
= $52.74 - $52.74
= $0
Bond Value
Replicating Portfolio:
P = (N)V – B
The replicating portfolio duplicates the option value in all future scenarios, hence
its present value must be the same as the option’s present value: C0.
Thus, the option is worth $12.09.
© 2014 OnCourse Learning. All Rights Reserved.
37
We can now correct our decision tree:
The correct OCC was not 20%, but rather 34.4%.
We know this because this is the rate that gives the correct PV of
the option: $12.09 = E[C] / (1+E[rc]) = $16.25 / 1.344.
Build: Get
113.21-90.00
= $23.21
70%
Wait Today:
Choice
PV =
16.25/1.344 =
12.09.
Choice Next Yr.:
1 Yr
Node Value =
(7.)23.21+ (.3)0 =
$16.25.
30%
Today
Don’t build:
Get 0.
Build Today:
Get 100.0088.24 =
$11.76.
In effect, we were able to derive the
option value without knowing the OCC.
If we want to know the OCC we can
“back it out” from the option value.
© 2014 OnCourse Learning. All Rights Reserved.
38
Note:
This options-based derivation of the OCC of developable land is
completely consistent with Chapter 29’s formula for development
project OCC:
PV [CT ]
VT LT
VT
LT
1 E[rC ]T 1 E[rV ]T 1 E[rD ]T
Only in the circumstance where the option will definitely be developed
next period (e.g., in the previous example, if the construction cost were
$78.62 million instead of $90 million, the option would be worth
$18.01 million and it would be “ripe” for immediate development):
$102.83 78.62 $102.83 $78.62
$18.01
1.344
1.09
1.03
In all cases, the result is to provide the same expected return risk
premium per unit of risk across all the asset markets (land, buildings,
bonds): the equilibrium condition within and across the relevant markets.
© 2014 OnCourse Learning. All Rights Reserved.
39
Here is the corrected summary of the analysis of the development project:
The land is worth: MAX[$100.00 – $88.24, C0] = $12.09:
34.4% OCC for the option
Today
Probability
Next Year
100%
30%
70%
Value of Developed Property
$100.00
$78.62
$113.21
Development Cost (exclu land)
$88.24
$90.00
$90.00
NPV of exercise
$11.76
-$11.38
$23.21
(Don’t build)
(Build)
0
$23.21
(Action)
Future Values
Expected Values
$11.76
$16.25
= Sum[ Probability X Outcome ]
(1.0)11.76
(0.3)0 + (0.7)23.21
PV(today) of Alternatives @ 34%
$11.76
16.25 / 1.344 = $12.09
What is the value of this land today? Answer: = MAX[11.76, 12.09] = $12.09
Should owner build now or wait? Answer: = Wait. (100.00 – 88.24 – 12.09 < 0.)
© 2014 OnCourse Learning. All Rights Reserved.
40
The general formula for the Replicating Portfolio in a Binomial World is:
Replicating Portfolio = NV-B, where:
“N” is “shares” (proportional value) of the underlying asset (built
property) to purchase,
“B” is current (time 0) dollar value of bond to issue (borrow), and:
N=(Cu-Cd)/(Vu-Vd); and
B=(NVd-Cd)/(1+rf).
With: Cu = MAX[Vu-K, 0]; Cd = MAX[Vd-K, 0];
Vu, Vd, = “up” & “down” values of property to be built; K = constr cost.
In the preceding example:
N = (23.21-0)/(113.21-78.62) = 23.21/34.59 = 0.67; and
B = (0.67(78.62)-0)/1.03 = $52.74/1.03 = $51.21.
© 2014 OnCourse Learning. All Rights Reserved.
41
How did we know what values of “N” and “B” to use in the
replicating portfolio”?...
2 linear equations in 2 unknowns:
Cu = NVu – (1+rf)B
Cd = NVd – (1+rf)B
Based on definition of “replicating portfolio”:
Option and replicating portfolio must have identical values in
both possible future states of the world.
General solution is:
N=(Cu-Cd)/(Vu-Vd); B=(NVd-Cd)/(1+rf).
e.g., for our numerical example:
N = (23.21-0)/(113.21-78.62) = 23.21/34.59 = 0.67;
B = (0.67(78.62)-0)/1.05 = $52.74/1.03 = $51.21.
© 2014 OnCourse Learning. All Rights Reserved.
42
General solution is:
N=(Cu-Cd)/(Vu-Vd);
B=(NVd-Cd)/(1+rf).
Combining these formulas with the definition of the replicating
portfolio, we can expand the binomial option value as:
C = NV – B
C
C Cd
Cu C d
V (0) u
Vd Cd 1 rf
Vu Vd
Vu Vd
C Cd
(C Cd )
Cd
V (0) u
Vd (1 rf )
C u
(1 rf )
Vu Vd
Vu Vd
Vd (1 rf )
Vd (1 rf )
Cu
Cd
Cd
Cu
Cd
C
(1 rf )
Vu Vd V (0) Vu Vd
Vu Vd V (0) Vu Vd
1 Vd (1 rf ) / V (0)
1 Vd (1 rf ) / V (0)
Cd
Cu
Cd
C
(1 rf )
Vu Vd V (0)
Vu Vd V (0)
1 rf Vd / V (0)
1 r f Vd / V (0)
Cu
Cd Cd
C
V V V (0) /(1 r )
V V V (0) /(1 r )
(1 rf )
d
f
d
f
u
u
1 rf Vd V (0)
1 rf Vd V (0)
Cu 1
Cd 1 rf
C
V
V
V
(
0
)
V
V
V
(
0
)
u
d
u
d
1 rf d
1 rf d
Cu 1
Cd 1 rf
C
u
d
u
d
C pCu (1 p )Cd 1 rf , where p 1 rf d u d , p " risk neutral prob"
© 2014 OnCourse Learning. All Rights Reserved.
Hence, the
classical (Cox-RollRoss) binomial
option value
formula using “riskneutral
probabilities”…
43
27.3.2: The Certainty-Equivalence Perspective on the Option Value…
The previously described option valuation of a development project is
completely consistent with the “Certainty Equivalent Valuation” form of the
DCF valuation model presented earlier in the Chapter 10 lecture in this
course.
The general 1-period Certainty Equivalent Valuation Formula is:
E[rV ] rf
E0 [C1 ] Cup Cdown
Vup % Vdown %
CEQ0 [C1 ]
C0
1 rf
1 rf
e.g., in our example :
(.7)$23.21 (.3)$0 $23.21 $0
9% 3%
(
113
.
21
/
94
.
34
)%
(
78
.
62
/
94
.
34
)%
C0
1 3%
6%
$16.25 $23.21
$16.25 $23.21 366.67%% $16.25 $23.210.1636
120
%
83
.
33
%
1.05
1.03
1.03
$16.25 $3.80 $12.45
$12.09
1.03
1.03
© 2014 OnCourse Learning. All Rights Reserved.
44
I’m hoping you developed some intuition for the certainty equivalence valuation model
back in Chapter 10 (CD Appendix 10C). But in case not, let’s try this . . .
The certainty equivalent value next year is the downward adjusted value of the risky
expected value for which the investment market would be indifferent between that value
and a riskfree bond value of the same amount…
E[rV ] rf
E0 [C1 ] Cup Cdown
Vup % Vdown %
CEQ0 [C1 ]
C0
1 rf
1 rf
The certainty equivalent value next year is the expected value minus a risk discount.
E[rV ] rf
Cup Cdown
V
%
V
%
up
down
The risk discount consists of the amount of
risk in the next year’s value as indicated by the
range in the possible outcomes times…
times the market price
of risk.
The market price of risk is the market expected return risk premium per unit of return risk, the
ratio of…
the market expected return
E[rV ] rf
risk premium divided by the
Vup % Vdown %
range in the corresponding
return possible outcomes.
© 2014 OnCourse Learning. All Rights Reserved.
45
Thus, we can derive the same present value of the
option through two completely consistent
(mathematically equivalent) approaches:
• The “arbitrage analysis” based on the
replicating portfolio, or;
• The CEQ DCF PV model.
The latter is more convenient for computations.
These two approaches are equivalent, but can be
derived independently (recall that we did NOT
employ an arbitrage analysis to derive the CEQ
valuation model).
This makes the valuation model very robust.
© 2014 OnCourse Learning. All Rights Reserved.
46
27.3.3 What is fundamentally going on with this framework:
V1up=
$113.21
p = 0.7
PV[V1 ]
=$94.34
1-p = 0.3
V1down=
$78.62
C1up=
$23.21
p = 0.7
PV[C1 ]
=$12.09
1-p = 0.3
C1down=
$0
Underlying asset (built
property) outcome %
range:
$113.21 $78.62
37%
$94.34
Option (land) outcome
% range:
$23.21 $0
192%
$12.09
With perfect correlation between the two (land is derivative).
© 2014 OnCourse Learning. All Rights Reserved.
47
27.3.3 What is fundamentally going on with this framework:
With perfect correlation between the two (land is derivative).
Hence, relative risk exactly equals ratio of outcome ranges:
192%
5.24
37%
Option (land) is 5.24 times more risky than investment in the
underlying asset (built property).
Thus, option value must be such that E[rC] risk premium in
land is 5.24 times greater than that in built property.
Built property risk premium is: RPV = 9% - 3% = 6%.
Thus, land risk premium must be: RPC = 6%*5.24 = 31.4%.
Thus land’s E[rC] = rf + RPC = 3% + 31.4% = 34.4%.
© 2014 OnCourse Learning. All Rights Reserved.
48
27.3.3 What is fundamentally going on with this framework:
The “price of risk” (the ex ante investment return risk premium per unit of
risk) is being equated across the markets for land and built property:
E[r]
Exh.27-4
34.4%
E[RP]
9.0%
rf = 3.0%
Risk
Built Property
37% range
Devlpt Option
192% range
If this relationship does not hold, then there are “super-normal”
(disequilibrium) profits (expected returns) to be made somewhere, and
correspondingly “sub-normal” profits elsewhere, across the markets for: Land,
Stabilized Property, and Bonds (“riskless” CFs).
© 2014 OnCourse Learning. All Rights Reserved.
49
Technical aside…
Deriving the model from standard risk-adjusted discounting using the
E[C ]
E[C ]
E[C ]
CAPM
C
, based on CAPM ,
1
1
1 E[rC ]
0
C0
1 rf RPC
1
1 rf C , Mkt ( E[rMkt ] rf )
E[C1 ]
, because CORREL [rC , rV ] 100%
1 rf C ,V V , Mkt ( E[rMkt ] rf )
E[C1 ]
,
1 rf C ,V ( E[rV ] rf )
C0 1 rf C ,V ( E[rV ] rf ) E[C1 ]
C0 1 rf C0 C ,V ( E[rV ] rf ) E[C1 ]
C0
C0
E[C1 ] C0 C ,V ( E[rV ] rf )
1 rf
[ C , rV ]
E[C1 ] COV
VAR[ rV ] ( E[ rV ] r f )
1 rf
STD[ C ]
E[C1 ] STD
[ rV ] ( E[ rV ] r f )
1 rf
E[C1 ]
E[C1 ] C0
COV [ rC , rV ]
VAR[ rV ]
( E[rV ] rf )
1 rf
STD[ C ]
E[C1 ] STD
[ rV ] CORREL [C , rV ]( E[ rV ] r f )
1 rf
, because CORREL [C , rV ] 1. In a binomial world this
E[rV ] rf
E[C1 ] Cup Cdown
V % V %
( E[rV ] rf )
up
down
1 rf
1 rf
Cup C down
Vup % Vdown%
© 2014 OnCourse Learning. All Rights Reserved.
50
Back to our option valuation problem
Note that the probabilities and expected future values used in this model
are “real”, not “risk-neutral dynamics” values. i.e., (0.7)113.21 +
(0.3)78.62 = $102.83 is the real or true expected value of the to-be-built
building next year, and 70% and 30% are the true probabilities of the
“up” and “down” outcomes.
It is necessary in this formulation to pose:
• An expected total return (OCC) to the underlying asset (the E[rV] =
9% in our example), and
• The underlying asset’s cash payout rate (the E[yV] = 6% in our
example).
It is also possible to obtain an exactly equivalent solution using so-called
“risk-neutral dynamics”, in which case it is not necessary to know the
OCC of the underlying asset. However, this poses little additional
advantage in the case of real estate, and it results in a less intuitive
formulation. (And the CEQ solution is insensitive to the OCC rate
employed.)
© 2014 OnCourse Learning. All Rights Reserved.
51
So far, we have illustrated these derivations using a simple 2-state
world, a “binomial” world. In fact, this can be the basis for a very
powerful yet intuitive and transparent model to rigorously value
options…
27.4 The Binomial Option Value Model
A financial-economic “molecule”…
C1up=
$23.21
p = 0.7
PV[C1 ]
=$12.09
1-p = 0.3
C1down=
$0
Smallest, simplest element that has all the key characteristics:
Time and Risk
© 2014 OnCourse Learning. All Rights Reserved.
52
A financial-economic “molecule”…
C1up=
$23.21
p = 0.7
PV[C1 ]
=$12.09
1-p = 0.3
C1down=
$0
Stitch these molecules together in a “crystal” (binomial lattice or “tree”):
Time
© 2014 OnCourse Learning. All Rights Reserved.
53
Each node in the lattice represents one “state of the world”,
one possible value of the underlying asset (the building)
at one possible future point in time.
© 2014 OnCourse Learning. All Rights Reserved.
54
Make three such trees:
t=
0
1
2
3
4…
“V” value of
underlying
asset (bldg)
“K” value of
construction
cost
“C” value of
option to get
“V” by paying
“K”: (V-K now,
or wait…)
© 2014 OnCourse Learning. All Rights Reserved.
55
Make three such trees:
t=
0
1
2
3
4…
“V” value of
underlying
asset (bldg)
“K” value of
construction
cost
“C” value of
option to get
“V” by paying
“K”: (V-K now,
or wait…)
© 2014 OnCourse Learning. All Rights Reserved.
“Evolve”
(compute) these
lattices forward to
reflect expected
growth & volatility,
enumerate future
V & K values…
Base the “C” value
on CEQ method:
MAX[V-K,C].
Compute C lattice
backward from
right to left.
V-K today, or hold C today
based on binomial
possibilities of C next period.
56
Here is a 1-period binomial representation of our previous numerical example:
• Color-code:
– Evolution of value of underlying asset (developed property)
– Value of the option to develop (i.e., the land) at each node
– Rational (optimal) decision at each node (e.g., to develop or to wait)
Vu 113.21
NPV1develop 113.21-90=23.21
NPV1notdevelop 0
V 0 100.00,
V(0) = (.7(113.21)+.3(78.62))/1.09 = 94.34
NPV0develop =100.00-88.24=11.76
NPV0notdevelop 0.7(23.21)+0.3(0) =12.09
1+rc
C(0) = max[11.76,12.09] =12.09
Don't develop; wait,
even though NPVdevelop
=11.76 > 0
0
1-p = 0.7
NPV1u max[23.21,0] = 23.21
Develop!
1-p = 0.3
Vd 78.62
NPV1develop 78.62–90 = -11.38
NPV1notdevelop 0
NPV1d = max[-11.38.0] = 0
Don't develop !
t 0
t 1
© 2014 OnCourse Learning. All Rights Reserved.
57
Extension to multiple periods: Build a binomial tree
t
t
•
•
•
•
•
t
Start from the end
Just repeat the same two-node tree at every node
At each node, take the NPV-maximizing decision, looking only 1 period ahead!
1
If d
then the tree recombines: updown = downup
u
Do you see a function f (MV,t), “above” which you develop immediately?
© 2014 OnCourse Learning. All Rights Reserved.
58
27.4: The Binomial Option Value Model
Think of an individual binomial element (1 period, either “up” or
“down”) as like a financial economic “molecule”: the smallest, simplest
representation of the essential characteristics dealt with by financial
economics: value over time with risk.
We can have as many periods of time as we want (individual “molecules”
stitched together as in a “crystal”: as layers or rows & columns in a table,
or nodes & branches in a “tree).
Each period can represent as short a span of calendar time as we want.
We can have as many periods as we want.
Result:
Binomial “tree” can very realistically model actual evolution of values
over time.
© 2014 OnCourse Learning. All Rights Reserved.
59
Appendix 27:
27A.1: The Binomial Option Value Model
Here are the rules for constructing the underlying asset value tree.
Let:
• rV = Expected total return rate on the underlying asset (built property).
• yV = Payout rate (dividend yield or net rent yield).
• rf = Riskfree interest rate
• σ = Annual volatility of underlying asset (instantaneous rate)*.
• Vt = Value of the underlying asset at time (end of period) t, ex dividend
(i.e., net of current cash payout, i.e., the value of the asset itself based only
on forward-looking cash flows beyond time t). The asset is assumed to
pay out cash at a rate of yV every period: yV = CFt+1 / Vt+1.**
All rates are simple periodic rates:
r = i/m, where r is the simple periodic rate, i is the nominal annual
rate, and m is the number of periods per year. The implied effective
annual rate (EAR) is thus given by: 1+EAR = (1+r)m
© 2014 OnCourse Learning. All Rights Reserved.
60
Appendix 27:
27A.1: The Binomial Option Value Model
For example, in our previous illustration…
• rV = 9%
• yV = 6%
• rf = 3%
• σ = Let’s say this is 20%.
• Vt = $100 at time 0, E[Vt+1] = $102.83 at time 1.
V1up=
$113.21
p = 0.7
V0=
$100
+ CF1up = $6.79
= (.06)$113.21
E[V1] = (1.09)$100/(1.06) = $102.83.
E[V1] = (0.7)$113.62 + (0.3)$78.62 = $102.83.
E[CF1] = (0.06)$102.83 = $6.17.
E[CF1] = (0.7)$6.79 + (0.3)$4.72 = $6.17.
1-p = 0.3
$100 =
($102.83 + $6.17) / 1.09
V1down=
$78.62
+ CF1down = $4.72
= (.06)$78.62
“going-in cap rate” = $6.17 / $100 = 6.17%.
E[gV] = (1+rV)/(1+yV) = 1.09/1.06 – 1 = 2.83%
= rV – (going-in cap rate) = 9% - 6.17%.
© 2014 OnCourse Learning. All Rights Reserved.
61
Appendix 27:
27A.1: The Binomial Option Value Model
Now define the 1-period “up” movement ratio as:
u = Vup / V(0) . e.g., in our last example: u = $113.21 / $94.34 = 1.20.
For the binomial model to work, the “down” movement ratio must be the
inverse of the “up” movement ratio:
d = Vdown / V(0) = 1 / u . e.g., in our last example: d = $78.62 / $94.34 =
0.833 = 1/1.20.
The magnitude of the “up” movement is determined so that the binomial
tree will converge to a “normal” (Gaussian) distribution of periodic
returns with annual volatility σ as the period lengths approach zero (m
∞ , or T/n 0). This requires:
u 1 T / n
where T is the total calendar time in the tree (in years) and n is the
total number of periods (hence, T/n is the fraction of a year in any
one period, and m = n/T is the number of periods per year).
In our previous example: T = 1, n = 1, and σ = 20%.
© 2014 OnCourse Learning. All Rights Reserved.
62
Appendix 27:
27A.1: The Binomial Option Value Model
The probability of the “up” move, p, is determined so that the binomial
tree will converge to a normal (Gaussian) distribution of periodic returns
with a mean annual total return based on rV as the period lengths
approach zero (m ∞ , or T/n 0). [Or equivalently, an appreciation
return of approximately: gV = (1+rV)/(1+yV)-1.] This requires:
1 rV d
p
ud
1 rV 1 / 1
T /n
1 T / n 1/ 1 T / n
The probability of the “down” movement is of course just 1 – p .
Note: These are actual probabilities, not “risk-neutral” pseudo-probabilities.
They produce a tree that reflects actual real underlying value distributions.
© 2014 OnCourse Learning. All Rights Reserved.
63
Appendix 27:
27A.1: The Binomial Option Value Model
Example based on our previous illustration
1 rV d
p
ud
1 rV 1 / 1
T /n
1 T / n 1/ 1 T / n
1.09 1 1.20 1.09 0.833 0.2567
0.70
1.20 1 1.20
1.20 0.833
0.3667
The probability of the “down” movement is of course just:
1 – p = 1 – 0.7 = 0.3 .
© 2014 OnCourse Learning. All Rights Reserved.
64
Appendix 27:
27A.1: The Binomial Option Value Model
The binomial tree for the underlying asset ex-dividend values is then
constructed as follows.
For any given value node with current (observable) ex-dividend value
Vt, the subsequent “up” and “down” values in the two possible
subsequent value nodes are:
Vt up
1 uVt 1 yV 1 T / n Vt 1 yV
down
t 1
V
dVt 1 yV Vt
1
© 2014 OnCourse Learning. All Rights Reserved.
T / n 1 yV
65
Appendix 27:
27A.1: The Binomial Option Value Model
The binomial tree for the underlying asset ex-dividend values is then
constructed as follows.
For example in our previous illustration…
Vt up
1 uVt 1 yV 1 T / n Vt 1 yV
1.20 $100 1.06 $113.21
down
t 1
V
dVt 1 yV Vt
1
T / n 1 yV
0.833$100 1.06 $78.62
© 2014 OnCourse Learning. All Rights Reserved.
66
The Binomial Option Value Model
Numerical Example:
Suppose:
Expected total return on the underlying asset is 10%, with a cash payout
rate of 6%, and a riskfree interest rate of 3% (all nominal annual rates).
Option expires in T = 1 year. n = 12: There are 12 periods (beyond 0), of
one month each (1 year total). Hence:
rV = 10%/12 = .833% , yV = 6%/12 = 0.5% , rf = 3%/12 = 0.25% ;
gV = (1+rV)/(1+yV)-1 = 1.00833/1.005 – 1 = 0.0033 = 0.33%.
Suppose the volatility of the underlying asset (built property) is σ = 15%.
V0 = $100, the time 0 value of the underlying asset (as if pre-existing).
Thus: u = [1+.15*SQRT(1/12)] = 1.0433; d = 1 / u = 1 / 1.0433 = 0.9585.
Vup = u($100)/(1+.005) = $104.33/(1.005) = $103.81.
Vdown = d($100)/(1+.005) = $95.85/(1.005) = $95.37.
p = ((1+.10/12)-1/(1+.15*SQRT(1/12)))/((1+.15*SQRT(1/12))-1/(1+.15*SQRT(1/12)))
= (1.00833 – .9585)/(1.0433 – .9585) = 0.5877; hence: 1 – p = 0.4123.
© 2014 OnCourse Learning. All Rights Reserved.
67
The Binomial Option Value Model
Numerical Example (cont.)
V1up=
$103.81
p = .5877
V0=
$100
1-p = .4123
V1down=
$95.37
Note that: E[V1] = .5877(103.81) + .4123(95.37) = $100.33 = (1.0033)$100 = (1+gV )V0
Note that: V(0) = $100.33 / (1+(.10/12)) = $100.33/1.00833 = $99.50 ≠ $100 = V0
Equivalently: V(0) = V0 / (1+yV) = $100 / (1+(.06/12)) = $100/1.005 = $99.50.
© 2014 OnCourse Learning. All Rights Reserved.
68
The Binomial Option Value Model
Now consider the “down” jump from V1up , and the “up” jump from V1down (call
this value “V1,2”, because it is in the 1st row down from the top, 2nd column over
from the left in the overall binomial tree).
V1,2 = up from V1down = u($95.37)/(1+.06/12) = 1.0433(95.37)/1.005 = $99.01.
V1,2 = down from V1up = d($103.81)/(1+.06/12) = .9585(103.81)/1.005 = $99.01.
It’s the same value!
V1up=
$103.81
p = .5877
1-p = .4123
V0=
$100
V1,2=
$99.01
1-p = .4123
p = .5877
down=
V1
$95.37
This is not a coincidence.
It is a general property of the
way we have constructed the
binomial tree. (constant u, d = 1/u.)
© 2014 OnCourse Learning. All Rights Reserved.
69
The Binomial Option Value Model
We build the underlying asset value tree forward in this manner
Here is the tree up through the V0,2 , V1,2 , and V2,2 value nodes . . .
V0,2 = up from V1up = u($ 103.81)/(1+.06/12) = 1.0433(103.81)/1.005 = $107.77.
V2,2 = down from V1up = d($95.37)/(1+.06/12) = .9585(95.37)/1.005 = $90.96.
V0,2=
$107.77
p = .5877
V1up=
$103.81
p = .5877
1-p = .4123
V0=
$100
V1,2=
$99.01
1-p = .4123
p = .5877
down=
V1
$95.37
1-p = .4123
© 2014 OnCourse Learning. All Rights Reserved.
V2,2=
$90.96
70
Appendix 27:
27A.1: The Binomial Option Value Model
We build the underlying asset value tree forward in this manner
$116.58
$110.77
$105.25
$100.00
$104.20
$99.01
$94.07
$93.14
$88.49
$83.25
© 2014 OnCourse Learning. All Rights Reserved.
71
Appendix 27:
27A.1: The Binomial Option Value Model
We build the underlying asset value tree forward in this manner
$184.73
$175.52
Exhibit 27A-3: Underlying Asset Value Tree for 12 Months
$166.77
$158.45
$150.55
$143.05
$135.91
$129.14
$122.70
$116.58
$110.77
$105.25
$100.00
$104.20
$99.01
$94.07
$98.02
$88.49
$97.05
$87.62
$83.25
$78.31
$95.13
$85.03
$76.77
$69.30
$84.19
$79.99
$76.00
$72.21
$68.61
$65.19
$94.19
$89.49
$80.79
$72.94
$105.38
$100.13
$90.39
$85.89
$77.53
$73.67
$112.02
$101.13
$81.60
$117.90
$106.44
$96.09
$86.75
$82.42
$125.33
$113.15
$91.30
$131.91
$119.08
$107.50
$102.14
$92.21
$140.22
$126.59
$114.28
$147.58
$133.23
$120.28
$108.58
$103.17
$93.14
$141.63
$127.86
$115.43
$109.67
$156.88
$149.06
$134.57
$121.48
$165.11
$75.25
$71.50
$67.93
$64.55
$61.33
$67.26
$63.91
$60.72
$57.69
$60.12
$57.12
$54.27
$53.73
$51.05
$48.03
© 2014 OnCourse Learning. All Rights Reserved.
72
The Binomial Option Value Model
Here is the 12-period, monthly periods (1 year) numerical example
tree we have been working on (from Excel)
V tree (net of payout, "ex dividend" values):
Period ("j "):
0
1
2
3
4
V tree (net of payout, "ex dividend" values):
100.00
103.81
107.77
111.87
95.37
99.01
102.78
90.96
94.43
86.75
116.14
106.70
98.02
90.06
82.74
Notice the first element here, as
we previously calculated it.
5
120.56
110.76
101.76
93.49
85.89
78.91
6
125.16
114.99
105.64
97.05
89.16
81.92
75.26
7
129.93
119.37
109.66
100.75
92.56
85.04
78.12
71.77
8
134.88
123.92
113.84
104.59
96.09
88.28
81.10
74.51
68.45
9
140.02
128.64
118.18
108.58
99.75
91.64
84.19
77.35
71.06
65.29
10
145.36
133.54
122.69
112.71
103.55
95.13
87.40
80.30
73.77
67.77
62.26
11
150.90
138.63
127.36
117.01
107.50
98.76
90.73
83.36
76.58
70.36
64.64
59.38
Each node in the tree (each row, column cell in the table)
represents a possible future “state of the world”, as indicated by a
possible value of the underlying asset as of the given future time
period (month in this case).
© 2014 OnCourse Learning. All Rights Reserved.
"n " =
12
156.65
143.91
132.22
121.47
111.60
102.52
94.19
86.53
79.50
73.04
67.10
61.65
56.64
73
The Binomial Option Value Model
Although the conditional probabilities are p = .5877 and 1-p = .4123 going forward
one period (up and down) from any given node, over multiple periods the
unconditional probabilities become bell-shaped over all the possible outcomes
Tree real probabilities (p based):
Period ("j "):
0
1
2
1.0000
0.5877
0.3454
0.4123
0.4846
0.1700
3
0.2030
0.4272
0.2997
0.0701
4
0.1193
0.3347
0.3523
0.1648
0.0289
5
0.0701
0.2459
0.3451
0.2421
0.0849
0.0119
6
0.0412
0.1734
0.3042
0.2846
0.1497
0.0420
0.0049
7
0.0242
0.1189
0.2503
0.2926
0.2053
0.0864
0.0202
0.0020
Period 12
Value Probabilities
25%
20%
8
0.0142
0.0798
0.1961
0.2752
0.2413
0.1355
0.0475
0.0095
0.0008
9
0.0084
0.0528
0.1482
0.2426
0.2553
0.1791
0.0838
0.0252
0.0044
0.0003
10
0.0049
0.0345
0.1088
0.2036
0.2500
0.2105
0.1231
0.0494
0.0130
0.0020
0.0001
11
0.0029
0.0223
0.0782
0.1645
0.2309
0.2268
0.1591
0.0798
0.0280
0.0065
0.0009
0.0001
"n " =
12
0.0017
0.0143
0.0551
0.1289
0.2035
0.2285
0.1870
0.1125
0.0493
0.0154
0.0032
0.0004
0.0000
Actually, although the model converges
toward continuously-compounded
return probabilities that are normally
distributed, the asset value level
probabilities are log-normally
distributed (skewed bell shape).
15%
10%
5%
0%
$57
$62
$67
$73
$80
$87
$94
$103
$112
$121 $132
$144
$157
© 2014 OnCourse Learning. All Rights Reserved.
74
The Binomial Option Value Model
Define the tree as a table of rows and columns. The jth column is the number of periods
after the present (time 0), where j = 0, 1, 2, …, n (where n is the total number of periods).
The ith row is the number of “down” moves in the asset price since time 0, where i = 0, 1,
2, … , j . Each row, column cell (i, j) defines a “state of the world” j periods in the future.
Vi,j is the value of the underlying asset in that state.
V tree (net of payout, "ex dividend" values):
Period ("j "):
0
1
2
3
"down" moves ("i"): V tree (net of payout, "ex dividend" values):
0
100.00
103.81
107.77
111.87
1
95.37
99.01
102.78
2
90.96
94.43
3
86.75
4
5
6
7
8
9
10
11
12
4
116.14
106.70
98.02
90.06
82.74
5
120.56
110.76
101.76
93.49
85.89
78.91
6
125.16
114.99
105.64
97.05
89.16
81.92
75.26
7
129.93
119.37
109.66
100.75
92.56
85.04
78.12
71.77
8
134.88
123.92
113.84
104.59
96.09
88.28
81.10
74.51
68.45
9
140.02
128.64
118.18
108.58
99.75
91.64
84.19
77.35
71.06
65.29
10
145.36
133.54
122.69
112.71
103.55
95.13
87.40
80.30
73.77
67.77
62.26
11
150.90
138.63
127.36
117.01
107.50
98.76
90.73
83.36
76.58
70.36
64.64
59.38
"n " =
12
156.65
143.91
132.22
121.47
111.60
102.52
94.19
86.53
79.50
73.04
67.10
61.65
56.64
The ex ante probability (as of time 0) of any given state of the world i, j is given by:
j! ( j i )
i
p 1 p
prob(i, j )
i!( j i)!
where the symbol “!” indicates the “factorial” product operation: x! = 1*2*3* . . . *x.
© 2014 OnCourse Learning. All Rights Reserved.
75
The Binomial Option Value Model
In this numerical example, the Period 12 underlying asset values and
probabilities, indicated in the last column on the right in the previous
two slides, give:
E0[V12] = E0[V(1 yr)] = $104.05
which is identical to: V0((1+rV)/(1+yV))12 = V0(1+gV)12 = $100(1.0033)12 = $104.05,
And:
STD[V12]/V0 = 1 yr Volatility = ±14.99%
which is very similar to the 15% simple annual volatility assumption.
(If you’re curious, these statistics are found as follows…)
12! (12i )
p
1 p i $104.05
E0 [V12 ] Vi ,12
i 0
i!(12 i )!
where : p 0.5877, and Vi ,12 are as found in the table.
12
STD[V12 ] V0
12! (12i )
2
i
V
E
[
V
]
p
1
p
i ,12
V0 14.99%
0 12
i!(12 i )!
i 0
12
© 2014 OnCourse Learning. All Rights Reserved.
76
The Binomial Option Value Model
Value probabilities for the underlying asset (V) 12 periods into the future
Period 12
Value Probabilities
25%
20%
15%
10%
5%
0%
$57
$62
$67
$73
$80
$87
$94
$103
$112
$121
$132 $144
$157
E0[V12] = $104.05 = V0(1.0033)12 = V0(1+gV)12 .
STD[V12]/V0 = ±14.99% ≈ 15% = σ
© 2014 OnCourse Learning. All Rights Reserved.
77
The Binomial Option Value Model
For each node (cell) in the underlying asset value tree (the “V Tree” )
described previously, there will also be associated a projected value of
the “exercise price” for the option (the construction cost of the
development project).
We label this cost K.
Assuming K grows risklessly at 2%/yr nominal (0.1667%/mo), the table
of Ki, j values giving construction costs corresponding to the previous
Vi, j values is as follows
Period ("j "):
0
"down" moves ("i"): K Value Tree:
0
80.00
1
2
3
4
5
6
7
8
9
10
11
12
1
80.13
80.13
2
80.27
80.27
80.27
3
80.40
80.40
80.40
80.40
4
80.53
80.53
80.53
80.53
80.53
5
80.67
80.67
80.67
80.67
80.67
80.67
6
80.80
80.80
80.80
80.80
80.80
80.80
80.80
7
80.94
80.94
80.94
80.94
80.94
80.94
80.94
80.94
© 2014 OnCourse Learning. All Rights Reserved.
8
81.07
81.07
81.07
81.07
81.07
81.07
81.07
81.07
81.07
9
81.21
81.21
81.21
81.21
81.21
81.21
81.21
81.21
81.21
81.21
10
11
81.34
81.34
81.34
81.34
81.34
81.34
81.34
81.34
81.34
81.34
81.34
81.48
81.48
81.48
81.48
81.48
81.48
81.48
81.48
81.48
81.48
81.48
81.48
"n " =
12
81.61
81.61
81.61
81.61
81.61
81.61
81.61
81.61
81.61
81.61
81.61
81.61
81.61
78
The Binomial Option Value Model
How to value a call option using the model
While the underlying asset value and exercise price trees are
constructed going forward in time as described previously,
An option on the underlying asset is valued by working backward in
time, starting at the right-hand edge of the tree (option expiration) and
working back to the left.
The option can be valued one period at a time, at each node of the tree,
based on the option values in the two subsequent possible nodes.
Starting in the last column (expiration period j = n ) one works
backwards in time ultimately to the present (time 0 at period j = 0 ).
Each valuation in each node (i , j cell in the table) is a simple 1-period
binomial valuation using the certainty-equivalent present value model
discussed previously.
© 2014 OnCourse Learning. All Rights Reserved.
79
The Binomial Option Value Model
The general formula and procedure for call option valuation is
thus as follows
First, let:
Vi,j = The ex-dividend underlying asset value in state-of-the-world i at
time (period) j, as enumerated in the binomial value tree described
previously (e.g., built property value).
Kj = The exercise price (construction cost) at time j (known for certain in
advance). i.e., paying Kj in period j will produce in period j an asset worth
Vi,j at that time (instantaneous construction).
Then in the terminal period j = n (in which the option expires), the call
option is worth, in any given state whose underlying asset value is Vi,n :
Ci ,n Max (Vi ,n K n ,0)
© 2014 OnCourse Learning. All Rights Reserved.
80
The Binomial Option Value Model
For example, look at the top two values in the terminal column j = n = 12
of our previous tree, V0,12 = $156.65 and V1,12 = $143.91 respectively.
Label the value of the call option in each of these nodes: C0,12 and C1,12 .
Given that construction cost in month 12, K12 , is $81.61 (see previous
table), we thus have:
C0,12 = Max( $156.65 - $81.61, 0 ) = $75.04
C1,12 = Max( $143.91 - $81.61, 0 ) = $62.30
Now consider the period j = n-1 state from which these two j = n value
states are each possible, and suppose (for now) that the option cannot be
exercised prior to its maturity at period n ( “European Option” )…
For example, for the $156.65 and $143.91 values in period 12, this would
be the state in period 11 in our previous example where V0,11 is worth
$150.90.
© 2014 OnCourse Learning. All Rights Reserved.
81
The Binomial Option Value Model
V tree (net of payout, "ex dividend" values):
Period ("j "):
0
1
2
3
"down" moves ("i"): V tree (net of payout, "ex dividend" values):
0
100.00
103.81
107.77
111.87
1
95.37
99.01
102.78
2
90.96
94.43
3
86.75
4
5
6
7
8
9
10
11
12
4
116.14
106.70
98.02
90.06
82.74
5
120.56
110.76
101.76
93.49
85.89
78.91
6
125.16
114.99
105.64
97.05
89.16
81.92
75.26
7
129.93
119.37
109.66
100.75
92.56
85.04
78.12
71.77
8
134.88
123.92
113.84
104.59
96.09
88.28
81.10
74.51
68.45
9
140.02
128.64
118.18
108.58
99.75
91.64
84.19
77.35
71.06
65.29
10
145.36
133.54
122.69
112.71
103.55
95.13
87.40
80.30
73.77
67.77
62.26
11
150.90
138.63
127.36
117.01
107.50
98.76
90.73
83.36
76.58
70.36
64.64
59.38
"n " =
12
156.65
143.91
132.22
121.47
111.60
102.52
94.19
86.53
79.50
73.04
67.10
61.65
56.64
V12up=
$156.65
Max(156.6581.61, 0) =
$75.04
V0,11=
$150.90
C0,11 = ?
Max(143.9181.61, 0) =
$62.30
© 2014 OnCourse Learning. All Rights Reserved.
V12down=
$143.91
82
The Binomial Option Value Model
The value of the call option in any state is a function of its two possible
values in the subsequent period. The exact valuation formula is the same
certainty-equivalence PV formula we have presented previously:*
Ct CEQ[Ct 1 ] (1 rf )
Et [Ct 1 ] C
up
t 1
C
down
t 1
E[rV ] rf
up
down
Vt 1 % Vt 1 %
1 rf
Substituting our previously-described binomial model parameters, this
becomes:
rV rf
pCi, j1 (1 p)Ci1, j 1 Ci, j 1 Ci1, j 1
1
T
/
n
1
1
T
/
n
Ci , j
1 rf
where the probability p is as defined previously:
1 rV 1 1
T /n
p
1 T / n 1 1 T / n
© 2014 OnCourse Learning. All Rights Reserved.
83
The Binomial Option Value Model
In particular, for C0,11 , recalling our previous 1-year monthly numerical
example input parameters of: rV = 10%/12 = .833% , yV = 6%/12 = 0.5% ,
rf = 3%/12 = 0.25%, and σ = 15%, we obtain:
rV rf
C0,11 .5877C0,12 .4123C1,12 C0,12 C1,12
1 1 / 12 1 1 1 / 12
(1 rf )
0.833% 0.25%
.5877C0,12 .412C1,12 C1,12 C2,12
1.0025
104
.
33
%
95
.
85
%
0.583%
.5877(75.04) .4123(62.30) 75.04 62.30
1.0025
8
.
48
%
$69.78 $12.730.0688 1.0025
$68.91 1.0025
$68.73
© 2014 OnCourse Learning. All Rights Reserved.
84
The Binomial Option Value Model
V12up=
$156.65
C0,12 =
$75.03
V0,11=
$150.90
C0,11 =
$68.73
K12 = $81.61
K11 = $81.48
V12down=
$143.91
C1,12 =
$62.30
rV rf
C0,11 pC0,12 (1 p)C1,12 C0,12 C1,12
1 1 / 12 1 1 / 12
(1 rf )
10% / 12 3% / 12
.5877(75.03) .4123(62.30) 75.03 62.30
1 15% 1 / 12 1 1 15% 1 / 12
(1 3% / 12)
$68.91 1.0025
$68.73
© 2014 OnCourse Learning. All Rights Reserved.
85
The Binomial Option Value Model
Repeating this process within each column for i = 0, 1, 2, …, j , and then
across columns from right to left for j = 11, 10, 9, …, 0 , we eventually
obtain the value of the option as of the present time 0:
Note here the values we calculated
in the previous slides.*
Period ("j "):
0
1
2
"down" moves ("i"): Eurpoean Call Option Value Tree:
0
15.76
19.23
23.06
1
12.09
15.20
2
8.81
3
4
5
6
7
8
9
10
11
12
3
27.20
18.72
11.48
5.97
4
31.60
22.60
14.63
8.15
3.64
5
36.23
26.78
18.21
10.85
5.27
1.90
6
41.07
31.19
22.15
14.06
7.44
2.97
0.77
7
46.13
35.83
26.36
17.72
10.19
4.51
1.31
0.18
8
51.41
40.67
30.79
21.72
13.50
6.67
2.21
0.35
0.00
9
56.94
45.72
35.42
25.96
17.26
9.51
3.64
0.68
0.00
0.00
10
11
62.71
51.01
40.26
30.39
21.32
12.98
5.81
1.32
0.00
0.00
0.00
68.73
56.53
45.32
35.02
25.55
16.86
8.87
2.55
0.00
0.00
0.00
0.00
"n " =
12
75.03
62.30
50.60
39.85
29.98
20.91
12.58
4.92
0.00
0.00
0.00
0.00
0.00
The European option is worth $15.76 in the present time 0.
© 2014 OnCourse Learning. All Rights Reserved.
86
The Binomial Option Value Model
If the landowner can begin the development project (exercise the option)
at any time ( “American option” ), then the value of the land in any state
prior to option expiration is given by:
rV rf
pC
(
1
p
)
C
C
C
i , j 1
i 1, j 1
i , j 1
i 1, j 1
1 T / n 1 1 T / n
Ci , j Max Vi , j K i , j ,
1 rf
In our example, given that K0,11 = $81.48, then if the land can be
developed immediately at any time, it is worth in state 0,11 :
C0,11 Max ($150.90 $81.48, $68.73) Max ($69.42, $68.73) $69.42
The flexibility to build the project at any time prior to the end of the year
( “American Option” instead of “European Option” ), makes the option
worth more, namely, $20 at time 0 (instead of $15.76).
© 2014 OnCourse Learning. All Rights Reserved.
87
The Binomial Option Value Model
Here is the complete American Option land value tree, assuming as
before initial building value of: V0 = $100, initial construction cost: K0 =
$80, (deterministic) construction cost growth rate of 2%/yr
(0.167%/month), and that the right to ever develop the land expires
after 1 year. From Excel:
Period ("j "):
0
"down" moves ("i"): LAND Value tree:
0
20.00
1
2
3
4
5
6
7
8
9
10
11
12
1
23.68
15.24
2
27.50
18.74
10.69
3
31.47
22.38
14.03
6.88
4
35.60
26.16
17.49
9.52
4.06
5
39.90
30.10
21.09
12.82
5.90
2.11
6
44.36
34.18
24.84
16.25
8.38
3.25
0.88
7
48.99
38.43
28.73
19.81
11.62
4.92
1.47
0.25
Note the value we
just calculated, here:
8
53.81
42.84
32.77
23.52
15.02
7.27
2.41
0.46
0.03
9
58.81
47.43
36.97
27.37
18.54
10.43
3.89
0.84
0.06
0.00
10
11
64.01
52.20
41.34
31.37
22.21
13.79
6.10
1.52
0.11
0.00
0.00
69.42
57.15
45.88
35.53
26.02
17.28
9.25
2.74
0.21
0.00
0.00
0.00
"n " =
12
75.03
62.30
50.60
39.85
29.98
20.91
12.58
4.92
0.40
0.00
0.00
0.00
0.00
An example Excel spreadsheet template for this binomial option model example is
available for downloading from the course web site.
© 2014 OnCourse Learning. All Rights Reserved.
88
The Binomial Option Value Model
Note that, as we have presented it, the option (land) valuation formula
appears to be a function of nine variables or parameters*:
Ci,j = C(Vi,j , Ki,j , rV , rf , yV , gK , σ, T, n)
In general, ceteris paribus (holding all other variables and parameters
constant), option value increases as a function of:
V , rf , σ , and T.
And decreases as a function of:
K , yV , and gK
Importantly (and very interestingly), note that the option value is:
UNAFFECTED BY THE UNDERLYING ASSET
REQUIRED RETURN , rV OR THE GROWTH RATE, gV
(holding constant the current underlying asset VALUE, Vi,j , and the
payout rate yV .)
© 2014 OnCourse Learning. All Rights Reserved.
89
Option is: UNAFFECTED BY THE UNDERLYING ASSET
REQUIRED RETURN , rV OR THE GROWTH RATE, gV
Why?...
Here’s one way to see it:
Recall: “Linear Pricing” (aka “Law of One Price”…):
Arbitrage-based Valuation Certainty-Equivalence Valuation
Certainty-Equivalence Valuation Arbitrage-based Valuation
Mathematically equivalent.
Recall that Arbitrage-based Valuation did not require
knowledge of rV or gV (as long as we could observe V(0), the
current value of the underlying asset).
© 2014 OnCourse Learning. All Rights Reserved.
(holding constant the current underlying asset VALUE, Vi,j , and the
payout rate yV .)
And V(0) can be computed if we can observe V0 the current value of the
underlying asset and yV its payout ratio: V(0) = V0 / (1+yV).
Also, recall accounting relationship that must hold:
(1+rV) =(1+yV) (1+gV), rV ≈ yV + gV
90
The Binomial Option Value Model
Effect of Underlying Asset Volatility
This is the same option as before only we’ve increased the underlying
asset volatility (σ) from 15% to 25%.
Period ("j "):
0
"down" moves ("i"): LAND Value tree:
0
20.16
1
2
3
4
5
6
7
8
9
10
11
12
1
26.55
13.86
2
33.55
18.90
8.89
3
41.02
25.22
12.66
5.15
Note the increase in
value from $20.00 to
$20.16
4
49.00
32.15
17.62
7.76
2.56
5
57.52
39.55
23.91
11.40
4.14
0.99
6
66.63
47.45
30.76
16.31
6.55
1.75
0.24
7
76.34
55.88
38.08
22.60
10.08
3.04
0.46
0.01
8
86.72
64.89
45.90
29.39
15.02
5.19
0.91
0.02
0.00
9
97.80
74.51
54.26
36.63
21.30
8.62
1.78
0.04
0.00
0.00
10
109.63
84.79
63.17
44.37
28.02
13.79
3.48
0.08
0.00
0.00
0.00
11
122.26
95.75
72.70
52.64
35.19
20.01
6.81
0.16
0.00
0.00
0.00
0.00
"n " =
12
135.74
107.46
82.87
61.47
42.85
26.66
12.58
0.32
0.00
0.00
0.00
0.00
0.00
With 15% volatility, the option model called for optimal immediate
exercise at time 0.
With 25% volatility, the model indicates that the option is more valuable
held for speculation at time 0 instead of immediate exercise at that time.
© 2014 OnCourse Learning. All Rights Reserved.
91
The Binomial Option Value Model
Notice that the option value model not only values the option, but also indicates in
which states of the world it is optimal to exercise the option (build the development
project). Here are the valuation and optimal exercise trees for the option with σ back at
the original 15%
Period ("j "):
0
"down" moves ("i"): LAND Value tree:
0
20.00
1
2
3
4
5
6
7
8
9
10
11
12
1
23.68
15.24
2
27.50
18.74
10.69
3
31.47
22.38
14.03
6.88
4
35.60
26.16
17.49
9.52
4.06
5
39.90
30.10
21.09
12.82
5.90
2.11
6
44.36
34.18
24.84
16.25
8.38
3.25
0.88
7
48.99
38.43
28.73
19.81
11.62
4.92
1.47
0.25
8
53.81
42.84
32.77
23.52
15.02
7.27
2.41
0.46
0.03
9
58.81
47.43
36.97
27.37
18.54
10.43
3.89
0.84
0.06
0.00
10
11
64.01
52.20
41.34
31.37
22.21
13.79
6.10
1.52
0.11
0.00
0.00
69.42
57.15
45.88
35.53
26.02
17.28
9.25
2.74
0.21
0.00
0.00
0.00
10
11
exer
exer
exer
exer
exer
exer
hold
hold
hold
hold
hold
exer
exer
exer
exer
exer
exer
exer
hold
hold
hold
hold
hold
Period ("j "):
0
"down" moves ("i"): Optimal exercise:
0
exer
1
2
3
4
5
6
7
8
9
10
11
12
1
2
exer
exer
exer
exer
exer
3
exer
exer
exer
hold
4
exer
exer
exer
exer
hold
5
exer
exer
exer
exer
hold
hold
6
exer
exer
exer
exer
hold
hold
hold
7
exer
exer
exer
exer
exer
hold
hold
hold
© 2014 OnCourse Learning. All Rights Reserved.
8
exer
exer
exer
exer
exer
hold
hold
hold
hold
9
exer
exer
exer
exer
exer
exer
hold
hold
hold
hold
"n " =
12
75.03
62.30
50.60
39.85
29.98
20.91
12.58
4.92
0.40
0.00
0.00
0.00
0.00
"n " =
12
exer
exer
exer
exer
exer
exer
exer
exer
hold
hold
hold
hold
hold
92
The Binomial Option Value Model
Here is the same option only with σ = 25% . . .
Period ("j "):
0
"down" moves ("i"): LAND Value tree:
0
20.16
1
2
3
4
5
6
7
8
9
10
11
12
1
26.55
13.86
2
3
4
5
6
7
8
9
10
11
33.55
18.90
8.89
41.02
25.22
12.66
5.15
49.00
32.15
17.62
7.76
2.56
57.52
39.55
23.91
11.40
4.14
0.99
66.63
47.45
30.76
16.31
6.55
1.75
0.24
76.34
55.88
38.08
22.60
10.08
3.04
0.46
0.01
86.72
64.89
45.90
29.39
15.02
5.19
0.91
0.02
0.00
97.80
74.51
54.26
36.63
21.30
8.62
1.78
0.04
0.00
0.00
109.63
84.79
63.17
44.37
28.02
13.79
3.48
0.08
0.00
0.00
0.00
122.26
95.75
72.70
52.64
35.19
20.01
6.81
0.16
0.00
0.00
0.00
0.00
2
3
4
5
6
7
8
9
10
11
exer
exer
exer
exer
exer
exer
hold
hold
hold
hold
hold
exer
exer
exer
exer
exer
exer
exer
hold
hold
hold
hold
hold
Period ("j "):
0
"down" moves ("i"): Optimal exercise:
0
hold
1
2
3
4
5
6
7
8
9
10
11
12
1
exer
hold
exer
hold
hold
exer
exer
hold
hold
exer
exer
hold
hold
hold
exer
exer
exer
hold
hold
hold
exer
exer
exer
hold
hold
hold
hold
exer
exer
exer
exer
hold
hold
hold
hold
exer
exer
exer
exer
exer
hold
hold
hold
hold
exer
exer
exer
exer
exer
hold
hold
hold
hold
hold
"n " =
12
135.74
107.46
82.87
61.47
42.85
26.66
12.58
0.32
0.00
0.00
0.00
0.00
0.00
"n " =
12
exer
exer
exer
exer
exer
exer
exer
exer
hold
hold
hold
hold
hold
Notice that with greater underlying asset volatility, option exercise is held back, less
likely. (Compare the “exer” cells in this table with the previous.)
© 2014 OnCourse Learning. All Rights Reserved.
93
The Binomial Option Value Model
Period ("j "):
0
"down" moves ("i"): Optimal exercise:
0
exer
1
2
3
4
5
6
7
8
9
10
11
12
1
2
exer
exer
exer
exer
exer
3
exer
exer
exer
hold
4
5
6
7
8
9
10
11
exer
exer
exer
exer
hold
exer
exer
exer
exer
hold
hold
exer
exer
exer
exer
hold
hold
hold
exer
exer
exer
exer
exer
hold
hold
hold
exer
exer
exer
exer
exer
hold
hold
hold
hold
exer
exer
exer
exer
exer
exer
hold
hold
hold
hold
exer
exer
exer
exer
exer
exer
hold
hold
hold
hold
hold
exer
exer
exer
exer
exer
exer
exer
hold
hold
hold
hold
hold
4
5
6
7
8
9
10
11
exer
exer
exer
exer
exer
exer
hold
hold
hold
hold
hold
exer
exer
exer
exer
exer
exer
exer
hold
hold
hold
hold
hold
With 15% volatility,
immediate exercise.
Period ("j "):
0
"down" moves ("i"): Optimal exercise:
0
hold
1
2
3
4
5
6
7
8
9
10
11
12
1
exer
hold
2
3
exer
hold
hold
exer
exer
hold
hold
With 25% volatility,
delay construction.
exer
exer
hold
hold
hold
exer
exer
exer
hold
hold
hold
exer
exer
exer
hold
hold
hold
hold
exer
exer
exer
exer
hold
hold
hold
hold
© 2014 OnCourse Learning. All Rights Reserved.
exer
exer
exer
exer
exer
hold
hold
hold
hold
exer
exer
exer
exer
exer
hold
hold
hold
hold
hold
"n " =
12
exer
exer
exer
exer
exer
exer
exer
exer
hold
hold
hold
hold
hold
"n " =
12
exer
exer
exer
exer
exer
exer
exer
exer
hold
hold
hold
hold
hold
94
The Binomial Option Value Model
Risk & the OCC
What are the implications of the option model for the amount of investment
risk, and the corresponding risk-adjusted discount rate (OCC ) applicable
to the land?
For a finite-lived option, the risk and OCC will differ according to the
“state-of-the-world” (the V, and K values, and the time until option
expiration).
But recall that the certainty-equivalence valuation we are employing here
allows us to “back out” what is the OCC once we have computed the option
value. The formula to do this is obtained as follows:
Ci , j
CEQ[C j 1 ]
1 rf
E j [C j 1 ]
1 rf E[ RPCi , j ]
E j [C j 1 ]
1 OCCi , j
,
1 OCCi , j
1 r
f
E j [C j 1 ]
CEQ[C j 1 ]
© 2014 OnCourse Learning. All Rights Reserved.
95
The Binomial Option Value Model
Risk & the OCC
Thus, for any i, j state of the world in the binomial tree, we can
compute the OCC (and the implied amount of investment risk
indicated by the corresponding risk premium in the OCC).
Here is are the 1-period (monthly) OCCs for our previous numerical
example (1-yr finite option with σ = 15% and the other parameters as before)
Period ("j "):
0
1
2
"down" moves ("i"): 1-Period Option Opportunity Cost of Capital:
0
3.22%
2.84%
2.56%
1
3.98%
3.39%
2
4.94%
3
4
5
6
7
8
9
10
11
12
3
2.35%
2.97%
4.27%
5.82%
4
2.17%
2.66%
3.58%
5.34%
6.87%
5
2.03%
2.42%
3.11%
4.62%
6.40%
8.34%
6
1.91%
2.23%
2.77%
3.81%
5.86%
7.85%
10.46%
7
1.81%
2.08%
2.50%
3.27%
5.03%
7.28%
10.07%
13.54%
8
1.72%
1.95%
2.30%
2.88%
4.07%
6.61%
9.57%
13.54%
NA
9
1.65%
1.84%
2.13%
2.59%
3.45%
5.54%
8.90%
13.54%
NA
NA
10
1.58%
1.75%
2.00%
2.37%
3.01%
4.38%
7.92%
13.54%
NA
NA
NA
11
1.52%
1.67%
1.88%
2.19%
2.69%
3.65%
6.19%
13.54%
NA
NA
NA
NA
Note that the OCC is mathematically indeterminate in states of the world where the
option has a certain value of zero (in all future possible states of the world and hence in
the current state).*
© 2014 OnCourse Learning. All Rights Reserved.
96
The Binomial Option Value Model
Risk & the OCC
Here are the per annum (effective annual rates) expected returns
implied by the preceding periodic OCCs
Period ("j "):
0
1
2
3
4
"down" moves ("i"): Option Opportunity Cost of Capital per Annum (EAR):
0
46.2%
40.0%
35.5%
32.1%
29.4%
1
59.8%
49.2%
42.1%
37.0%
2
78.3%
65.2%
52.6%
3
97.2%
86.7%
4
121.9%
5
6
7
8
9
10
11
5
27.2%
33.3%
44.4%
71.9%
110.5%
161.6%
6
25.5%
30.3%
38.7%
56.6%
98.1%
147.8%
229.8%
7
24.0%
28.0%
34.5%
47.1%
80.2%
132.4%
216.2%
358.9%
8
22.7%
26.1%
31.3%
40.6%
61.4%
115.6%
199.4%
358.9%
NA
9
21.7%
24.5%
28.8%
36.0%
50.2%
91.1%
178.3%
358.9%
NA
NA
10
20.7%
23.2%
26.8%
32.4%
42.8%
67.2%
149.7%
358.9%
NA
NA
NA
11
19.9%
22.0%
25.1%
29.7%
37.5%
53.8%
105.5%
358.9%
NA
NA
NA
NA
Note: These OCCs are probably not very realistic (too high) for actual land
investment, because of our assumption here of a 1-year finite life of the
development option. In reality, land development rights typically do not expire at
the end of a year. (More on this shortly.)
© 2014 OnCourse Learning. All Rights Reserved.
97
The Binomial Option Value Model
Risk & the OCC
How do these option OCCs compare to the “Method 1” (“canonical”)
Formula for the OCC of a development project, introduced in the
Chapter 29 lecture notes?
Recall that the Method 1 Formula is as follows (where TC is the time
required for construction):
TC
TC 1 TC
V LT 1 E[rV ] 1 E[rD ]
E[rC ] T
1
TC
TC
1 E[rD ] VT 1 E[rV ] LT
Recasting this in our current nomenclature with TC = 1 yr , this
formula is:
E j [V j 1 ] K j 1 1 rf
E[rCi , j ]
1
1 rf E j [V j 1 ] 1 rV K
This formula will in fact be equivalent to the option OCC previously
computed in any state of the world where the option will definitely be
exercised in the next period. (Recall that the canonical formula assumes a
definite commitment to go forward with the development project.)
© 2014 OnCourse Learning. All Rights Reserved.
98
The Binomial Option Value Model
Comparison of “canonical” OCC versus actual option OCC (previous
numerical example, 1st 4 periods only)
Period ("j "):
0
1
2
3
"down" moves ("i"): 1-Period Option Opportunity Cost of Capital:
0
3.22%
2.84%
2.56%
2.35%
1
3.98%
3.39%
2.97%
2
4.94%
4.27%
3
5.82%
4
4
2.17%
2.66%
3.58%
5.34%
6.87%
Period ("j "):
0
1
2
3
4
"down" moves ("i"): "Canonical" ("Method 1") OCC Formula from Ch.29:
0
3.22%
2.84%
2.56%
2.35%
2.17%
1
3.98%
3.39%
2.97%
2.66%
2
Not Valid
4.27%
3.58%
3
Not Valid Not Valid
4
Not Valid
Period ("j "):
0
"down" moves ("i"): Optimal exercise:
0
exer
1
2
3
4
1
2
exer
exer
exer
exer
exer
© 2014 OnCourse Learning. All Rights Reserved.
3
exer
exer
exer
hold
4
exer
exer
exer
exer
hold
99
Problems with the Binomial Model
There are two major technical problems with the Binomial Model:
1. Discrete time & values (the real world is continuous).
2. Finite expiration of the option (land is perpetual).
Both of these can have significant effects on the option value and
optimal exercise decision characteristics.
The first problem (discreteness) can be addressed by making the periods
of time very short (m ∞, T/n 0).
Perpetual expiration can be approximated by a long time horizon, but
more accurate solution requires an entirely different type of model.
For modeling a simple option, sufficient for dealing with the Wait Option,
there is a simple solution to this problem:
A model of perpetual option value in continuous time that includes the
value of the option to delay construction as well as a solution to the
decision problem of optimal development timing . . .
© 2014 OnCourse Learning. All Rights Reserved.
100
27.5 The Samuelson-McKean Formula Applied to Land Value
as a Development Option
The simplest option valuation formula is also the first one developed
(before Black-Scholes), and the one that is most relevant to land valuation
and optimal development timing:
The Samuelson-McKean Formula
Developed by Nobel Prize winning economist Paul Samuelson and his
mathematician partner Henry McKean, at MIT in 1965, as a model of a
“perpetual American warrant”.
The Samuelson-McKean Model is consistent with the Binomial Model in
that the latter would converge to the former if we could let T ∞ and also
T/n 0. (You can imagine how big a table this would require, since the
binomial table has dimension nXn, with approximately n2/2 elements in it .)
© 2014 OnCourse Learning. All Rights Reserved.
101
To see how this works, recall our replicating portfolio model of option value
Today
Development Option Value
Built Property Value
Bond Value
Replicating Portfolio:
C(0) = (N)V(0) – B
Next Year
C(0)=x
“x” = unknown val.
C(t)UP=113.21-90
= $23.21
C(t)DOWN = 0
(Don’t build)
V(0) = V0 / (1+yV)
= $102.83/(1.09)
=$94.34
V(t)UP= $113.21
V(t)DOWN= $78.62
B = $51.21
B(t) = (1+rf )B =
$52.74
B(t) = (1+rf )B =
$52.74
C(0) = (N)V(0) – B
= (0.7)$94.34 $51.21 = $12.09
C(t)UP =
(0.7)113.21 - $52.74
= $23.21
C(t)DOWN =
(0.7)78.62 - $52.74
=0
The Replicating Portfolio = NV-B, where: N=(Cu-Cd)/(Vu-Vd); B=(NVd-Cd)/(1+rf);
and V = V(0) (not V0 )
CtUP CtDOW N DOW N
UP
V
CtDOW N
DOW N t
UP
DOW N
C Ct
C
Vt Vt
C0 tUP
V
(
0
)
V (0) B
DOW N
V
V
1
r
V
t
f
t
© 2014 OnCourse Learning. All Rights Reserved.
102
Note that as we approach continuous time (periods get very short), N becomes like the
dC
derivative of the option value with respect to the underlying asset value: N dV
dC
Thus: C NV B, dC NdV dB dV
dV dB
We can also use the Taylor Series expansion from basic calculus (supplemented by
some very advanced mathematics known as “Stochastic Calculus” ) to approximate the
change in value of a (perpetual) option over time as:
dC
dC
dV
dV 12
2C
V 2
2V 2 dt
Combining our Replicating Portfolio formula C = NV – B with the above, and looking
at changes over time (returns), we see:
dV 12 VC2 2V 2 dt
dC
dC
dV
dB 12 VC2 2V 2 dt
2
dC
dV
dV dB
2
But we also know that the riskless bond value, B, given its Replicating Portfolio value
of: B = NV – C, will change over time according to the riskfree interest rate, as:
dC
dB B rf dt dV
V C rf dt
Equating the above two expressions for dB, we obtain the following ordinary
differential equation:
2
2 2
dC
dB 12 VC2 V dt dV
V C rf dt
1
2
2V 2 VC rf V
2
2
dC
dV
rf C 0
© 2014 OnCourse Learning. All Rights Reserved.
103
rf C rf V
dC
dV
1
2
V
2
2 d 2C
dV 2
The solution to this differential equation, combined with suitable boundary
conditions ( C(V=0) = 0, C(V=∞) = V ) and the conditions of optimal
exercise (expected exercise timing so as to maximize the present value of
the option), gives the Samuelson-McKean Formula.
This works as a model for land value because, like a perpetual American
warrant, land never expires (“perpetual”), and can be developed at any
time by its owner (exercise policy is “American”).
Actually, the equation presented above ignores dividends and assumes a constant
exercise price. To allow (more realistically for construction projects) for the
underlying asset to pay dividends (property net rent) and for construction costs to
grow over time, some minor modifications must be made in the formula. These
are reflected in the model presented on subsequent slides.
© 2014 OnCourse Learning. All Rights Reserved.
104
The Samuelson-McKean Model Applied to Land Value:
Let: V = Currently observable value of built property of the type that is the
HBU for the land (underlying asset, what we have labeled V0 , not V(0) ).
σ = Volatility of (Std.Dev. of return to unlevered) individual built properties (=
“total risk”, not just systematic or non-diversifiable risk, includes idiosyncratic
risk: Typical range for real estate is 15% to 25% per year).
yV = Payout ratio of the built property (current cash yield rate, like cap rate
only net of capital improvement reserve, typical real estate values range from
4% to 12%).
yK = Construction cost “yield” rate (= rf – gK , where gK is the growth rate of
construction costs, typically approximately equal to inflation).
© 2014 OnCourse Learning. All Rights Reserved.
105
The Samuelson-McKean Model Applied to Land Value:
Then the option value (and optimal exercise) formula has three steps:
(1) The “option elasticity” [(dLAND/LAND)/(dV/V)], η (“eta”), is given by:
= {yV – yK + σ2/2 + [(yK – yV - σ2/2)2 + 2yKσ2]1/2} / σ2
(2) The option critical value (“hurdle value”) of the built property at and
above which it is optimal to immediately exercise the option (develop the
land), labeled V* , is:
V* = K / ( - 1)
(3) The option (land) value is given by:
V
, if V V *
V * -K
V *
LAND =
V K , otherwise
© 2014 OnCourse Learning. All Rights Reserved.
106
Example:
rf = 3%, yV = 6%, σ = 15%, K = $80 (with yK=1%,2%growth) , V = $95,
= {yV – yK + σ2/2 + [(yK – yV – σ2/2)2 + 2yKσ2]1/2} / σ2
= {.06-.01+.152/2+[(.01-.06- .152/2)2+2(.01).152]1/2}/.152 = 5.60.
V* = K[η/(η-1)] = $80[5.6/(5.6-1)] = $80(5.6/4.6) = $80(1.22) = $97.38.
LAND = (V*-K)(V/V*)η = ($97.38 - $80)($95/$97.38)5.6 = $15.13.
In this example,
Option Elasticity = 5.60, Hurdle Benefit/Cost Ratio = 1.22,
Land Value = $0.16 per dollar of current built property value.
Note: In applying the Sam-McK Formula, in principle V and K should be
defined based on the HBU that the site will ultimately be developed for (not
necessarily what it could immediately be developed for).*
(Obviously you wouldn’t memorize this formula! Use the downloadable file from
course web site.)
© 2014 OnCourse Learning. All Rights Reserved.
107
Here is a picture of what the Samuelson-McKean Formula looks like:
$0.70
LAND (OPTION) VALUE
$0.60
Devlpt
Optimal
$0.50
$0.40
$0.30
Land
Value
$0.20
NPV of
Construction
$0.10
$0.00
$0.00
$0.15
$0.30
$0.45
$0.60
$0.75
$0.90
$1.05
BUILT PROPERTY VALUE (V)
Option Value
$1.20
$1.35
$1.50
$1.65
V*=$1.28
"Hurdle"
Option Value at Expiration
Parameters in above chart are: σ = 15%, yV = 8%, yK = 5%.
Land value (LAND) is a monotonically increasing, convex function of the current HBU
built property value (underlying asset value). Above the hurdle benefit/cost (V/K) ratio,
the option should already be exercised, and its value is simply V-K.
© 2014 OnCourse Learning. All Rights Reserved.
108
Both the option value, and the hurdle V/K ratio, are increasing functions of
the volatility (σ) and decreasing functions of the payout ratio (y).
Exhibit 27-7
$0.70
LAND (OPTION) VALUE (L)
$0.60
$0.50
$0.40
$0.30
V* =
$1.44
$0.20
V* =
$1.28
$0.10
$0.00
$0.00
$0.15
$0.30
$0.45
$0.60
$0.75
$0.90
$1.05
$1.20
$1.35
$1.50
$1.65
BUILT PROPERTY VALUE (V)
Opt Val @ 15% Vol
Opt Val @ 20% Vol
Option Value at Expiration
Assuming: yV = 8%, yK = 5%.
The hurdle benefit/cost ratio, and the land value as a fraction of the
construction cost, are independent of the scale of the site (in the sense of the
size of the land parcel, holding HBU density constant).
© 2014 OnCourse Learning. All Rights Reserved.
109
Recall that with the Binomial Model there appeared to be a “hurdle value”
of the underlying asset above which it is optimal to exercise (develop) . . .
V tree (net of payout, "ex dividend" values):
Period ("j "):
0
1
2
3
"down" moves ("i"): V tree (net of payout, "ex dividend" values):
0
100.00
103.81
107.77
111.87
1
95.37
99.01
102.78
2
90.96
94.43
3
86.75
4
5
6
7
8
9
10
11
12
4
116.14
106.70
98.02
90.06
82.74
5
120.56
110.76
101.76
93.49
85.89
78.91
6
125.16
114.99
105.64
97.05
89.16
81.92
75.26
Here (with 15% volatility) the hurdle
value of V seems to be about $90
(until the end).
7
129.93
119.37
109.66
100.75
92.56
85.04
78.12
71.77
8
134.88
123.92
113.84
104.59
96.09
88.28
81.10
74.51
68.45
9
140.02
128.64
118.18
108.58
99.75
91.64
84.19
77.35
71.06
65.29
10
145.36
133.54
122.69
112.71
103.55
95.13
87.40
80.30
73.77
67.77
62.26
11
150.90
138.63
127.36
117.01
107.50
98.76
90.73
83.36
76.58
70.36
64.64
59.38
Period ("j "):
0
"down" moves ("i"): Optimal exercise:
0
exer
1
2
3
4
5
6
7
8
9
10
11
12
1
2
exer
exer
exer
exer
exer
3
exer
exer
exer
hold
With 15% volatility,
immediate exercise.
4
exer
exer
exer
exer
hold
5
exer
exer
exer
exer
hold
hold
6
exer
exer
exer
exer
hold
hold
hold
7
exer
exer
exer
exer
exer
hold
hold
hold
8
exer
exer
exer
exer
exer
hold
hold
hold
hold
9
exer
exer
exer
exer
exer
exer
hold
hold
hold
hold
10
11
exer
exer
exer
exer
exer
exer
hold
hold
hold
hold
hold
exer
exer
exer
exer
exer
exer
exer
hold
hold
hold
hold
hold
"n " =
12
156.65
143.91
132.22
121.47
111.60
102.52
94.19
86.53
79.50
73.04
67.10
61.65
56.64
"n " =
12
exer
exer
exer
exer
exer
exer
exer
exer
hold
hold
hold
hold
hold
With the Sam-McK Model there is an explicit formula for this
hurdle value . . .
© 2014 OnCourse Learning. All Rights Reserved.
110
The hurdle benefit/cost ratio:
(V*/K) = η / (η-1)
is an interesting measure in its own right.
It tells you how much greater the anticipated completed new
built property value (including land) must be than its
construction cost (excluding land), in order for it to be optimal
to stop waiting to develop, and immediately begin
(instantaneous) construction.
Expressing this in terms of the land value fraction of the total
development project value at the time of optimal development,
the optimal land value fraction is given by the inverse of the
elasticity: V * K
K
1 1
V*
1
V*
1
e.g., Elasticity = 3 Hurdle B/C Ratio = 1.5 Optimal Land Fraction = 33%.
© 2014 OnCourse Learning. All Rights Reserved.
111
Samuelson-McKean Implications for Optimal Development
As noted, the option elasticity also determines the hurdle benefit/cost
ratio, V*/K, at which it is optimal to immediately begin development
whenever the current value of V and K equate to this ratio*:
V*
K 1
The hurdle benefit/cost ratio is thus an inverse function of the option
elasticity: larger elasticity means a lower hurdle ratio.
© 2014 OnCourse Learning. All Rights Reserved.
112
Samuelson-McKean Implications for Optimal Development
Hurdle Ratio ( V* / K ) as a function of underlying asset volatility ( σ ):
1.50
1.45
1.35
1.30
1.25
1.20
1.15
1.10
1.05
25%
24%
23%
22%
21%
20%
19%
18%
17%
16%
15%
14%
13%
12%
11%
1.00
10%
Hurdle Ratio ( V* / K )
1.40
Underlying Asset Volatility ("sigma")
With: yV = 8%, yK = 2%.
© 2014 OnCourse Learning. All Rights Reserved.
113
Samuelson-McKean Implications for Optimal Development
Hurdle Ratio ( V* / K ) as a function of underlying asset yield ( yV ):
2.20
1.80
1.60
1.40
1.20
12%
11%
10%
9%
8%
7%
6%
5%
4%
3%
1.00
2%
Hurdle Ratio ( V* / K )
2.00
Underlying Asset Yield ("y(V)"))
With: σ = 15%, yK = 2%.
© 2014 OnCourse Learning. All Rights Reserved.
114
Samuelson-McKean Implications for Optimal Development
Hurdle Ratio ( V* / K ) as a function of construction yield (yK= rf – gK ):
1.70
1.50
1.40
1.30
1.20
1.10
10%
9%
8%
7%
6%
5%
4%
3%
2%
1%
1.00
0%
Hurdle Ratio ( V* / K )
1.60
Construction Yield ("y(K)")
With: yV = 8%, σ = 15%.
© 2014 OnCourse Learning. All Rights Reserved.
115
The hurdle benefit/cost ratio (& optimal land fraction) is:
• Greater the more volatile is the built property market (i.e., the more
uncertainty there is in the future value of built properties):
• As long as you hold the option unexercised, greater volatility gives you
greater potential upside outcomes you can take advantage of while the option
flexibility allows you to avoid the greater downside outcomes implied by the
greater volatility.
• Uncertain and volatile property markets will dampen development, as
developers wait until they can get built property values (based on space market
rents) sufficiently above the construction cost exclusive of land).
• Lower the greater is the current cash yield (akin to cap rate) being
provided by built properties:
• You only start to get the net rent the property can generate when the
building is complete, so the greater the current yield, the greater the incentive
to build sooner rather than later.
• Land value (site acquisition cost) will be a smaller fraction of total
development cost (including construction) in locations where built property
values tend to grow slower (holding risk constant, lower “g” higher “y”, as
g+y=r, recalling Ch.9).
© 2014 OnCourse Learning. All Rights Reserved.
116
Example:
In the U.S., land (site acquisition) is typically about 20% of the total
development cost in most areas of the country, but often 50% in major
metropolises on the East and West Coast. Why?
Suppose rf = 5%, and property market volatility and payout rates differ as
follows:
Big East & West
Coast Cities
Rest of U.S.
Property Mkt Volatility (σ)
20%
15%
Property Payout Rate (y)
5%
8%
Then the Samuelson-McKean Formula gives the following difference in land
value fraction of total developed property value at the time of optimal
development (based on the implied V*/K hurdle ratio):
LAND/V*
@ V=V* (optimal dvlpt)
Big East & West
Coast Cities
Rest of U.S.
46%
22%
© 2014 OnCourse Learning. All Rights Reserved.
117
The option elasticity measure, η, is also interesting in its own right.
Prior to the point of optimal exercise (when the land is still
optimally held undeveloped for speculation), the elasticity tells the
percentage change in land value resulting from a given
percentage change in built property value (for the type of
property that would be the HBU of the land).
For a “live” option (below the hurdle ratio) the elasticity is:
• Independent of the size of the land parcel (for a given HBU
density);
• Independent of the current value of the underlying asset (the
state of the property market).*
• A decreasing function of the volatility in the property
market.
© 2014 OnCourse Learning. All Rights Reserved.
118
The option elasticity relates the volatility (and risk) of the
option (the undeveloped land investment) to the volatility (and
risk) of the underlying asset (the built property market for the
HBU of the site).
Assuming riskless construction costs:
σLAND = ησV ,
Where σLAND is the volatility of the undeveloped land.
Since the option return is perfectly correlated with the
underlying asset return, the option elasticity can therefore also
be used to relate the required expected investment return risk
premium in undeveloped land to that in the HBU built property
market:
RPLAND = ηRPV .
© 2014 OnCourse Learning. All Rights Reserved.
119
Example:
Built property expected return rV = 8%, Cash yield yV = 6%
Riskfree interest rate = 4%, Built property RPV = 4%.
Construction yield yK = 2%.
(Which might be determined as the 4% riskfree rate minus a 2% likely
construction cost growth rate: yK = rf – gK = 4% - 2% = 2%.)
If built property volatility σ = 15%, then: (σ =.15, yV =.06, yK =.02)
η = 4.9.
Thus,
RPLAND = η(RPV) = 4.9(4%) = 19.7%
Expected return (OCC) on land speculation investment =
rf + RPLAND = 4% + 19.7% = 23.7%.
Based on the Samuelson-McKean assumptions, this required expected return for land
speculation would hold no matter how big or small the land parcel (for a given HBU
density), or what the current state of the built property market is, as long as σ, y, rf , and
RPV remain the same. The Sam-McK Formula is a “constant elasticity” formula.
© 2014 OnCourse Learning. All Rights Reserved.
120
Samuelson-McKean Implications for Land OCC
As noted, the option elasticity, η , gives the ratio of the land risk to the
underlying asset risk, hence the ratio of the land to underlying asset
expected risk premium in the opportunity cost of capital (in the
expected investment return):
E[ RPC ]
E[ RPV ]
We also noted that for a “live option” (not yet ripe for exercise) η is
independent of both:
• Scale (value of V or of K ), and
• Current benefit cost ratio (amount of “operational
leverage” in the construction project: V / K
In fact, η is a function of only three variables: σ, yV , and yK .
This makes the option elasticity in the Samuelson-McKean Formula a
very useful tool for understanding and quantifying land investment
risk and return requirements.
© 2014 OnCourse Learning. All Rights Reserved.
121
Samuelson-McKean Implications for Land OCC
Option elasticity ( E[RPC] / E[RPV] ) as a function of underlying asset
volatility ( σ ):
14
10
8
6
4
2
25%
24%
23%
22%
21%
20%
19%
18%
17%
16%
15%
14%
13%
12%
11%
0
10%
Option Elasticity ("eta" )
12
Underlying Asset Volatility ("sigma")
With: yV = 8%, yK = 2%.
© 2014 OnCourse Learning. All Rights Reserved.
122
Samuelson-McKean Implications for Land OCC
Option elasticity ( E[RPC] / E[RPV] ) as a function of underlying asset
yield ( yV ):
12
8
6
4
2
12.0%
11.0%
10.0%
9.0%
8.0%
7.0%
6.0%
5.0%
4.0%
3.0%
0
2.0%
Option Elasticity ("eta" )
10
Underlying Asset Yield ("y(V)")
With: σ = 15%, yK = 2%.
© 2014 OnCourse Learning. All Rights Reserved.
123
Samuelson-McKean Implications for Land OCC
Option elasticity ( E[RPC] / E[RPV] ) as a function of construction yield
(yK = rf – gK ):
9
8
6
5
4
3
2
1
10.0%
9.0%
8.0%
7.0%
6.0%
5.0%
4.0%
3.0%
2.0%
1.0%
0
0.0%
Option Elasticity ("eta" )
7
Construction Yield ("y(K)")
With: yV = 8%, σ = 15%.
© 2014 OnCourse Learning. All Rights Reserved.
124
Generalizing the Samuelson-McKean Model to allow for risky
construction costs
The Sam-McK Model can allow for risky construction costs by use of a
simple transformation: Value the option per dollar of construction cost, as
follows:*
1. Divide the current underlying asset value by the current
construction cost, replacing V in the formula with V/K, and
replacing K in the formula with 1.
2. In computing the “construction yield”, yK , use the expected
return on an asset with market risk equivalent to that of the
construction cost, instead of the riskfree rate. i.e., yK = rK – gK
= rf + E[RPK] – gK .
3. In computing σ, use the volatility of a portfolio of the underlying
asset minus the construction cost:
VAR[rV ] VAR[rK ] 2COV [rV , rK ]
( This transformation is attributed to Fisher and Margrabe* .)
© 2014 OnCourse Learning. All Rights Reserved.
125
Summarizing up to now:
• The Binomial Model can handle:
• Finite-lived development options (rights expire at a
specified future time), “American” or…
• “European” development options (construction
prohibited prior to a given future point in time)
• The Samuelson-McKean Model can handle:
• The simple “Wait Option” for a perpetual “American”
development option (typical “land value” problem).
It remains for us to address two important considerations:
• Until now we have assumed instantaneous exercise: we need to
consider the effect of construction time, aka “time to build” .
• The “Phasing Option”, in which the project is broken into
sequential phases rather than building it all at once
© 2014 OnCourse Learning. All Rights Reserved.
126
Appendix 27
27A.1.3: Time to Build
With non-instantaneous construction, when you exercise the option to
build in state of the world i, j , you don’t get Vi,j – Kj .
You get the PV of the completed project:
(Ei,j[Vj+TC])/(1+rV)TC – Kj+TC /(1+rf)TC
where TC is the number of periods it will take to complete construction.
In the Binomial Model, if you exercise the option at time 0 when the
underlying asset has current observable value V0 , then if the time to
build is 1 period you will get:
PV [V1 K1 ]
p uV0 1 yV 1 p d V0 1 yV K1 pV0,1 1 p V1,1 K1 V0 K 0
E0 [V1 ]
K
1
1 rV 1 rf
1 rV
1 rf
1 rV
1 rf 1 yV 1 yK
If the time to build is 2 periods, you will get:
p 2V0, 2 2 p 1 p V1, 2 1 p V2, 2
E0 [V2 ]
V0
K0
K2
K2
PV [V2 K 2 ]
2
2
2
2
2
1 rf 1 yV 1 yK 2
1 rV 1 rf
1 rV
2
© 2014 OnCourse Learning. All Rights Reserved.
127
Appendix 27
27A.1.3: Time to Build
For example, consider our previous underlying asset value tree
and a time-to-build of 2 months with rV = 10%/yr = 0.833%/mo,
yV = 6%/yr = 0.5%/mo ( gV = 0.33%/mo) …
V =
0,2
A decision at time 0 to build the asset
obtains an asset at month 2
that is worth at
time 0: $99.01
p=
V0=
$100
(1 gV ) V0
E [V ]
PV [V2 ]
0 22
2
1 rV
1 rV
2
$107.77
V0,1=
$103.81
1-p =
.4123
.5877
1-p =
.4123
p=
.5877
V1,1=
$95.37
(.5877) 2107.77 2(.5877)(.4123)99.01 (.4123) 2 90.96
1.008332
V1,2=
$99.01
p=
.5877
1-p =
.4123
V2,2=
$90.96
V0
(1.0033) 2 $100 $100.66
$100.00
$99.01
2
2
2
1.00833
1.005
1.00833
1 yV 2
© 2014 OnCourse Learning. All Rights Reserved.
128
Appendix 27
27A.1.3: Time to Build
Summarizing:
To account for time to build in the option model:
• In any state i, j where the option could be
exercised, replace the immediate exercise value
Vi,j – Kj with the present value as of time j of the
exercise value TC periods later (where TC is the
required construction time) : PVi,j [Vj+TC – Kj+TC], as
defined in the previous slides.
• In the Samuelson-McKean Formula, replace the
current value of the underlying asset, Vij , with:
Vij / (1 + yV)TC , and replace the exercise price Kj
with Kj / (1 + yK)TC .
© 2014 OnCourse Learning. All Rights Reserved.
129
Appendix 27
27A.1.3: Time to Build
The general effect of time-to-build is:
• The value of the option is reduced below what it
otherwise would be.
• The expected time until optimal exercise is
increased beyond what it otherwise would be
(hurdle value of Vt as measured by current
observable price of pre-existing assets is increased):
• Condition of optimal exercise, where “Vt” is current
observable price of identical pre-existing asset:
• η/(η-1) = (Vt /(1+yV)TC ) / (Kt/(1+yK)TC )
• Vt = Kt[η/(η-1)]((1+yV)/(1+yK))TC ,
• and normally: yV > yK .
© 2014 OnCourse Learning. All Rights Reserved.
130
Appendix 10C:
Certainty Equivalence Discounting
• Review of traditional RADR DCF
• Introduction to Certainty-Equivalence DCF
• Example
Reference: Geltner-Miller et al 3e, Chapter 10 Appendix C (on the CD)
© 2014 OnCourse Learning. All Rights Reserved.
131
Chapter 10: “Discounted Cash Flow Valuation”…
THE CLASSICAL (and most widely used)
RADR METHOD OF DCF
1. FORECAST THE EXPECTED FUTURE CASH
FLOWS;
2. ASCERTAIN THE REQUIRED TOTAL
RETURN;
3. DISCOUNT THE CASH FLOWS TO
PRESENT VALUE AT THE REQUIRED RATE
OF RETURN.
THE VALUE YOU GET TELLS YOU WHAT YOU MUST
PAY SO THAT YOUR EXPECTED RETURN WILL EQUAL
THE "REQUIRED RETURN" AT WHICH YOU
DISCOUNTED THE EXPECTED CASH FLOWS.
© 2014 OnCourse Learning. All Rights Reserved.
132
Chapter 10: “Discounted Cash Flow Valuation”…
E0 [CF1 ]
E0 [CF2 ]
E0 [CFT 1 ]
E0 [CFT ]
V0
2
T 1
1 E0 [r ] 1 E0 [r ]
1 E0 [r ]
1 E0[r ]T
where:
CFt = Net cash flow generated by the property in period
“t”;
Vt = Property value at the end of period “t”;
E0[r] = Expected average multi-period return (per
period) as of time “zero” (the present), also known as
the “going-in IRR”;
T = The terminal period in the expected investment
holding period, such that CFT would include the resale value of the property at that time (VT), in addition
to normal operating cash flow.
© 2014 OnCourse Learning. All Rights Reserved.
133
Chapter 10: “Discounted Cash Flow Valuation”…
Match the discount rate to the risk. . .
r = rf + RP
Disc.Rate = Riskfree Rate + Risk Premium
(Riskfree Rate = US T-Bill Yield.)
© 2014 OnCourse Learning. All Rights Reserved.
134
Appendix 10C: “Certainty-Equivalent Valuation”…
Consider the fundamental element of our valuation model:
E0 [V1 ]
PV (V1 )
1 E0 [rV ]
Accounts for both time and risk in the discount rate in the
denominator, as E0 rV rf E[ RPV ], where rf is riskless and
accounts for the time value of money, and RPV is the market’s
required risk premium in the expected return for the investment
(ex ante).
Manipulate this formula so that the denominator purely reflects
the time value of money (the discounting is done risklessly) and
the risk is completely and purely accounted for in the numerator,
by reducing the cash flow amount in the numerator to a
“Certainty Equivalent Value” . . .
© 2014 OnCourse Learning. All Rights Reserved.
135
Appendix 10C: “Certainty-Equivalent Valuation”…
E0 [V1 ]
E0 [V1 ]
PV (V1 )
1 E0 [rV ] 1 rf E[ RPV ]
A little algebra . . .
1 r
f
E[ RPV ]PV (V1 ) E0 [V1 ]
1 r PV (V ) E[ RP ]PV (V ) E [V ]
f
1
V
1
0
1
1 r PV (V ) E [V ] E[ RP ]PV (V )
f
1
0
1
V
1
E0 [V1 ] E[ RPV ]PV (V1 )
PV (V1 )
1 rf
© 2014 OnCourse Learning. All Rights Reserved.
136
Appendix 10C: “Certainty-Equivalent Valuation”…
E0 [V1 ] E[ RPV ]PV (V1 ) is the “Certainty Equivalent” value of V1 .
The risk-adjusted discount rate formulation accounts for both
time and risk simultaneously in the denominator:
E0 [V1 ]
PV (V1 )
1 rf E[ RPV ]
The certainty-equivalence formulation accounts for time only in
the denominator, and risk only in the numerator, discounting
the certainty-equivalent cash flow amount risklessly back to the
present:
E0 [V1 ] E[ RPV ]PV (V1 )
PV (V1 )
1 rf
© 2014 OnCourse Learning. All Rights Reserved.
137
Appendix 10C: “Certainty-Equivalent Valuation”…
E0 [V1 ] E[ RPV ]PV (V1 ) is the “Certainty Equivalent” value of V1 .
The risk-adjusted discount rate formulation accounts for both
time and risk simultaneously in the denominator:
E0 [V1 ]
PV (V1 )
1 rf E[ RPV ]
Risk premium
(in %)
The certainty-equivalence formulation accounts for time only in
the denominator, and risk only in the numerator, discounting
the certainty-equivalent cash flow amount risklessly back to the
present:
E0 [V1 ] E[ RPV ]PV (V1 )
PV (V1 )
1 rf
Risk discount
(in $), aka
“Haircut”
© 2014 OnCourse Learning. All Rights Reserved.
138
Appendix 10C: “Certainty-Equivalent Valuation”…
The investment market is indifferent
between a claim on the actual risky amount
of V1 one period in the future (with the
actual amount of risk involved in that),
versus a claim on the lesser amount of:
E0 [V1 ] E[ RPV ]PV (V1 )
one period in the future with no risk at all.
The “haircutted” value is called the “certaintyequivalence value”, labeled “CEQ”:
CEQ0 (V1 ) E0 [V1 ] E[ RPV ]PV (V1 )
© 2014 OnCourse Learning. All Rights Reserved.
139
Appendix 10C: “Certainty-Equivalent Valuation”…
Why do we care about this? . . .
The certainty-equivalent approach is more
general, can be applied in situations where
RADR discounting can’t be applied, including
options valuation and other derivatives
valuations (e.g., swaps, forwards, futures).
Typically, CEQ discounting useful if:
• Risk does not accumulate at a constant rate over
time (i.e., there is no single correct OCC).
• PV=0 even though future cash flow expectations
are not zero (as with futures contracts)
Consider a simple example . . .
© 2014 OnCourse Learning. All Rights Reserved.
140
Appendix 10C: “Certainty-Equivalent Valuation”…
Suppose:
Riskfree interest rate = 3%
Commercial property market total return risk premium = 6%
Property market expected total return is:
E[rV] = rf + E[RPV] = 3% + 6% = 9%
“Binomial World”…
An office building that is worth $100M today (if it already
exists, earning income) will next year have a value of either:
$113M (with 70% probability), or
$79M (with 30% probability).
Therefore, expected value of office bldg next yr is:
E[V1] = (0.70)113 + (0.30)79 = $103M
© 2014 OnCourse Learning. All Rights Reserved.
141
Appendix 10C: “Certainty-Equivalent Valuation”…
Next Year
Here is the situation …
V1 = $113 M
Prob = 70%
Today
E[V1] = $103 M
PV[V1]
Prob = 30%
Real probabilities
(not “risk-neutral”)
V1 = $79 M
Real expected values
(not “risk-neutral”)
True expected value next year (nominal) = $103 M.
© 2014 OnCourse Learning. All Rights Reserved.
142
How much is the forward claim on the office building worth today? . . .
Risk-adjusted discount rate valuation:
PV [V1 ]
E0 [V1 ]
(.70)$113 (.30)$79 $103
$94 M
1 rf E[ RPV ]
1 3% 6%
1.09
Certainty-equivalent valuation:
PV [V1 ]
CEQ0 [V1 ] E0 [V1 ] E[ RPV ]PV [V1 ] $103 (.06)$94 $97
$94 M
1 rf
1 rf
1 3%
1.03
$97 is the certainty-equivalent value of V1 whose expected value is $103.
If we know rf and either E[RPV] or PV[V1] then we can derive above.
Thus, easy to apply RADR discounting to underlying asset (due to
transparency in market for trading such built properties).
© 2014 OnCourse Learning. All Rights Reserved.
143
How much is the forward claim on the office building worth today? . . .
PV [V1 ]
PV [V1 ]
E0 [V1 ]
(.70)$113 (.30)$79 $103
$94 M
1 rf E[ RPV ]
1 3% 6%
1.09
CEQ0 [V1 ] E0 [V1 ] E[ RPV ]PV [V1 ] $103 (.06)$94 $97
$94 M
1 rf
1 rf
1 3%
1.03
If we know rf and either E[RPV] or PV[V1] then we can derive above.
Even if we just know V0 (current value of pre-existing “twin” building) and yV
(cash payout rate, or income yield, of such building)…
PV[V1] = (1 + gV)V0 / (1 + rV) = V0 / (1 + yV) ,
e.g., with 6% “caprate”, $94 = $103 / 1.09 = $100 / 1.06.
This is due to relationship between rV, gV, & yV:
(1 + rV) / (1 + yV) = (1 + gV) rV ≈ yV + gV
(This is an “accounting relationship”: Holds by definition.)
Thus, easy to apply RADR discounting to underlying asset (due to
transparency in market for trading such built properties).
© 2014 OnCourse Learning. All Rights Reserved.
144
Appendix 10C: “Certainty-Equivalent Valuation”…
Now suppose the office building does not exist yet,
but we could build it in 1 year
by paying $90M of construction cost next year.
(Construction is instantaneous: We can decide next
year whether to build or not.)
Suppose our option to build this project expires in 1
year: We either build then or we lose the right to ever
build.
Suppose market for such options is thin, lacks
transparency, so we cannot directly observe value of
option or of relevant E[RP] in discount rate…
How much is this option worth?
© 2014 OnCourse Learning. All Rights Reserved.
145
Appendix 10C: “Certainty-Equivalent Valuation”…
Next Year
Label the value of the option today PV(C1).
Vup = $113 M
Here is the situation we face…
K = $90 M
Do dvlpt, get:
Prob = 70%
$113 - $90 = $23 M = Cup
Today
Vdown = $79 M
PV[C1]
K = $90 M
Prob = 30%
Don’t build, get: $0 = Cdown
E0 [C1 ]
(.70)$23 (.30)0
$16.1
PV [C1 ]
?
1 rf E[ RPC ] 1 .03 E[ RPC ] 1 .03 E[ RPC ]
We can’t value the option using RADR discounting
(which works to value the forward claim on the building),
because we don’t know what E[RPC ] should be. (For built
property we could observe or derive RP in the mkt, but dvlpt option
market is likely much thinner, less transparent, less homogeneous.)
© 2014 OnCourse Learning. All Rights Reserved.
146
Among “derivative” assets, relative risk can be measured by volatility (or
return outcome spread range) ratio between the derivative & the underlying
asset. “Derivatives” are assets that are perfectly correlated: derivative
outcome depends entirely on underlying outcome.
(We don’t know
this yet.)
E[r]
34%
E[RP]
© 2014 OnCourse Learning. All Rights Reserved.
Invoke “linear pricing” principle (aka “Law of One Price”):
• Same “Price of Risk” must apply to all assets at a given time.
• Same expected return risk premium per unit of risk (ex ante).
• Otherwise, “not fair” (people will trade until “One Price” holds).
9%
rf = 3%
Risk
Built Property
37% range
Devlpt Option
192% range
(We don’t know
this yet.)
147
Appendix 10C: “Certainty-Equivalent Valuation”…
Price of risk in the “underlying asset” (the completed built
office bldg – what you can get by exercising the
development option) = Risk Premium / Risk (spread)…
E[ RPV ]
E[ RPV ]
9% 3 %
up
$113 $79 $94 V1 $ V1down $ PV [V1 ] V1up % V1down %
9% 3%
6%
0.162
(20%) (16%) 36%
Therefore, price of risk must be the same in the
derivative asset, the development option:
E[ RPC ]
E[ RPC ]PV [C1 ]
0.162 up
up
,
down
down
C1 $ C1 $ PV [C1 ] C1 $ C1 $
E[ RPC ]PV [C1 ] C1up $ C1down $ 0.162 $23 $00.162 $3.7
Now we have all we need to apply the certainty-equivalence
discounting formula to value the option…
© 2014 OnCourse Learning. All Rights Reserved.
148
Appendix 10C: “Certainty-Equivalent Valuation”…
Now we have all we need to apply the certainty-equivalence
discounting formula to value the option:
CEQ0 [C1 ] E0 [C1 ] E[ RPC ]PV [C1 ]
PV (C1 )
1 rf
1 rf
(.7)$23 (.3)0 $3.7 $16.1 $3.7 $12.4
$12 M
1 3%
1 3%
1.03
E[r]
34%
E[R
P]
9%
rf = 3%
Risk
Built
Property
37% range
Devlpt
Option
192% range
© 2014 OnCourse Learning. All Rights Reserved.
149
Appendix 10C: “Certainty-Equivalent Valuation”…
E0 [C1 ]
$16.1
PV [C1 ] $12
,
1 rf E[ RPC ] 1 3% E[ RPC ]
$16.1
E[ RPC ]
103% 31%
$12
E0 [C1 ]
(.70)$23 (.30)0 $16.1
PV [C1 ]
$12M
1 rf E[ RPC ]
1 .03 .31
1.34
© 2014 OnCourse Learning. All Rights Reserved.
Having computed (derived ) PV[C1], we can now back out
the implied E[RPC] and RADR for the option…
Which implies that the option has 31/6 = 5.2 times the
investment risk that the built office property has, as:
E[RPC] = 31% = (5.2)*6% = (5.2)*E[RPV].
However, if we didn’t already know the option’s risk premium (31%), or hadn’t already done the
certainty-equivalent valuation, we wouldn’t be able to use the RADR method of DCF valuation.
150
Invoke “linear pricing” principle (aka “Law of One Price”):
• Same “Price of Risk” must apply to all assets at a given time.
• Same expected return risk premium per unit of risk (ex ante).
• Otherwise, “not fair” (people will trade until “One Price” holds).
(Now we know
this.)
© 2014 OnCourse Learning. All Rights Reserved.
Among “derivative” assets, relative risk can be measured by volatility (or
return outcome spread range) ratio between the derivative & the underlying
asset…
E[r]
34%
E[RP]
9%
rf = 3%
Risk
Built Property
37% range
Devlpt Option
192% range
(Now we can
compute this.)
151
Appendix 10C: “Certainty-Equivalent Valuation”…
Underlying asset outcome spread (Vup – Vdown) / V(0):
• Up outcome: ($113 – $94) / $94 = 19/94 = +20%
• Down outcome: ($79 – $94) / $94 = -15/94 = -17%
• Spread = (+20% – (-17%)) = 37%.
Derivative asset outcome spread (Cup – Cdown) / C(0):
• Up outcome: ($23 – $12) / $12 = 11/12 = +92%
• Down outcome: ($0 – $12) / $12 = -12/12 = -100%
• Spread = (+92% – (-100%)) = 192%.
Which implies that the option has 192/37 = 5.2 times the
investment risk that the built office property has, the same
as the risk premium ratio we just computed:
E[RPC] / E[RPV] = 31% / 6% = 5.2
© 2014 OnCourse Learning. All Rights Reserved.
152
The “Law of One Price”…
E[r]
34%
E[RP]
9%
rf = 3%
Risk
Built Property
37% range
Devlpt Option
192% range
If this relationship does not hold, then there are “supernormal” (disequilibrium) profits (expected returns) to be
made somewhere, and correspondingly “sub-normal” profits
elsewhere, across the markets for: Land, Built Property,
and Bonds (“riskless” CFs).
© 2014 OnCourse Learning. All Rights Reserved.
153
Appendix 10C: “Certainty-Equivalent Valuation”…
Fundamentally, what we are doing here is making sure that the expected
return risk premium per unit of risk is the same between an investment in
the development option, and an investment in already-build (stabilized)
property. For example, for stabilized property:
RPV
RPV
9% 3%
9% 3%
6%
up
0.17
down
up
down
$113 $79 $94 V1 $ V1 $ PV [V1 ] V1 % V1 % (20%) (16%) 36%
For the development option:
RPC
RPC
34% 3%
34% 3%
31%
up
0.17
down
up
down
$23 $0 $12 C1 $ C1 $ PV [C1 ] C1 % C1 % (90%) (100%) 190%
Thus, this procedure for the valuation of the development option is
equivalent to a cross-market equilibrium condition: that the expected
return risk premium per unit of risk (or the “risk-adjusted return”)
presented by the investments should be the same between the market for
development options and the market for stabilized built property.
Otherwise, investors will buy one type of asset and sell the other type until
prices equilibrate the risk-adjusted returns across the two markets.
© 2014 OnCourse Learning. All Rights Reserved.
154
Appendix 10C: “Certainty-Equivalent Valuation”…
General formula to value derivative “C” based on
underlying asset “V”…
E[rV ] rf
E0 [C1 ] Cup Cdown
V % V %
down
up
PV [C1 ]
1 rf
e.g ., in our example :
9% 3%
113 / 94 79 / 94
(.70)$23 (.30)0 $23 $0
PV [C1 ]
1 3%
6%
$16.1 $23
36% $16.1 $23.162
1.03
1.03
$16.1 $3.7 $12.4
$12
1.03
1.03
© 2014 OnCourse Learning. All Rights Reserved.
155
Appendix 10C: “Certainty-Equivalent Valuation”…
The equivalent valuation using the risk-adjusted discount rate approach is:
C0
E0 [C1 ]
(.70)$23 (.30)$0 $16.1
$12
1 rf E[ RPC ]
1 3% 31%
1.34
Which implies that the option has 31/6 = 5.2 times the investment
risk that the office building has (and that the stock market has), as:
RPC = 31% = 5.2*6% = 5.2*RPV (= 5.2*RPS).
However, if we didn’t already know the option’s risk premium
(31%), or hadn’t already done the certainty-equivalent approach, we
wouldn’t be able to use the risk-adjusted discount rate approach.
The certainty-equivalent approach does not require prior knowledge
of RPC. It only requires knowledge of Cup , Cdown , and the “up”
probability (70% in our example).
This makes the certainty-equivalent approach very useful for option
valuation. Most development projects are essentially “options”.
© 2014 OnCourse Learning. All Rights Reserved.
156
Appendix 10C: “Certainty-Equivalent Valuation”…
Formula for derivative asset valuation with Binomial
outcome distribution (as before):
E[rV ] rf
E0 [C1 ] Cup Cdown
V % V %
up
down
PV [C1 ]
1 rf
Formula for derivative asset valuation with general
outcome distribution (continuous or discrete):
PV [C1 ]
STD[ C ]
E[C1 ] STD
[ rV ] ( E[ rV ] r f )
1 rf
Formula for underlying asset valuation with general
outcome distribution (based on CAPM):
PV [V1 ]
[V , rMkt ]
E[V1 ] COV
VAR[ rMkt ] ( E[ rMkt ] r f )
1 rf
© 2014 OnCourse Learning. All Rights Reserved.
157
The “binomial world” is more realistic and useful than
you might think, because we can define the temporal
length of a period to be as short as we want, and we
can link and branch individual binomial elements
together to make a “tree” of asset value possibilities
over time.
More generally for a cash flow T periods in the
future with expected value E0[VT], define the
certainty-equivalent amount CEQ[VT]. We have:
PV [VT ]
E0 [VT ]
CEQ[VT ]
,
T
T
1 rf E[ RPV ] 1 rf
© 2014 OnCourse Learning. All Rights Reserved.
Appendix 10C: “Certainty-Equivalent Valuation”…
1 rf
E [V ]
CEQ[VT ]
1 r E[ RP ] 0 T
f
V
T
Provided the risk accumulates continuously
through the entire time to T.
158
Intuition in the CEQ formula:
The certainty equivalent value next year is the downward adjusted value of the risky
expected value for which the investment market would be indifferent between that value
and a riskfree bond value of the same amount…
E[rV ] rf
E0 [C1 ] Cup Cdown
Vup % Vdown %
CEQ0 [C1 ]
C0
1 rf
1 rf
The certainty equivalent value next year is the expected value minus a risk discount.
E[rV ] rf
Cup Cdown
V
%
V
%
up
down
The risk discount consists of the amount of
risk in the next year’s value as indicated by the
range in the possible outcomes times…
times the market price of
risk (for same “type” of
risk, e.g., office building*)
*dvlpt option is
perfectly correlated
with building
The market price of risk is the market expected return risk premium per unit of return risk for
the same asset, the ratio of…
the market expected return
E[rV ] rf
risk premium divided by the
Vup % Vdown %
range in the corresponding
return possible outcomes.
© 2014 OnCourse Learning. All Rights Reserved.
159
Technical aside
The astute student will have observed that there appears to be a slight
difference between the certainty-equivalent valuation formula presented
in the Chapter 10 lecture notes:
E[rMkt ] rf
E0 [C1 ] Cup Cdown
Mkt % Mkt %
up
down
C0
1 rf
(1)
And the one presented here:
E[rV ] rf
E0 [C1 ] Cup Cdown
V % V %
down
up
C0
1 rf
(2)
Formula (1) assumes there is a “Market Portfolio”, and either that the
option is perfectly correlated with that portfolio, or else that the CAPM
applies and the “up” and “down” movements considered in the formula
are only those components that are perfectly correlated with the “up”
and “down” movements of the Market. Formula (2) is a more general
expression.
© 2014 OnCourse Learning. All Rights Reserved.
160
Technical aside
Here’s a slightly different perspective relating the model to the CAPM:
According to the CAPM, in a Binomial World where we are considering
only movement components of the specific underlying asset (V ) that are
perfectly correlated with the Market Portfolio (i.e., ignoring V’s
idiosyncratic or non-systematic risk), we have:
E[rV ] rf
V
up
% Vdown %
up
% Mktdown %
Mkt
E[r
Mkt
] rf V , Mkt E[rMkt ] rf
(3)
Substituting (3) into (2) we obtain (1):
Vup % Vdown %
E[rMkt ] rf
Mktup % Mktdown %
E0 [C1 ] Cup Cdown
E[rMkt ] rf
Vup % Vdown %
E0 [C1 ] Cup Cdown
Mkt % Mkt %
up
down
C0
1 rf
1 rf
E0 [C1 ]
1 rf C1 , Mkt E[rMkt ] rf
© 2014 OnCourse Learning. All Rights Reserved.
161
Recall this digression from our Ch.10 lecture on certainty-equivalent valuation
Three asides (see notes for further explanation)
The “binomial world” is much more realistic and useful than you might think,
in part because we can define the temporal length of a period to be as short as
we want, and we can link and branch individual binomial elements together to
make a “tree” of asset value possibilities over time.
In the “CAPM”:
COV [rV , rMkt ]
E[ RPV ] V risk in return%
" Beta"
E[ RPMkt ] Mkt risk in return%
VAR
[
r
]
Mkt
, Thus:
COV [V1 , rMkt ]
V0 E0 [V1 ] ( E[rMkt ] rf )V0 1 rf E0 [V1 ]
( E[rMkt ] rf ) 1 rf
VAR[rMkt ]
E0 [V1 ] 1 rf ( E[rMkt ] rf )
More generally for a cash flow T periods in the future with expected value E0[VT],
define the certainty-equivalent amount CEQ[VT]. We have:
1 rf E [V ]
E0 [VT ]
CEQ[VT ]
V0
,
CEQ
[
V
]
T
1 r E[ RP ] 0 T
1 rf E[ RPV ]T 1 rf T
f
V
T
© 2014 OnCourse Learning. All Rights Reserved.
162
