Options and Efficiency Author(s) : Stephen A. Ross Source: The Quarterly Journal of Economics, Vol. 90, No. 1 (Feb., 1976), Pp. 75-89 Published By: Stable URL: Accessed: 20/12/2014 08:54

Options and Efficiency
Author(s): Stephen A. Ross

Source: The Quarterly Journal of Economics, Vol. 90, No. 1 (Feb., 1976), pp. 75-89
Published by: Oxford University Press
Stable URL: http://www.jstor.org/stable/1886087 .
Accessed: 20/12/2014 08:54
Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at .
http://www.jstor.org/page/info/about/policies/terms.jsp
.
JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of
content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms
of scholarship. For more information about JSTOR, please contact [email protected].
Oxford University Press is collaborating with JSTOR to digitize, preserve and extend access to The Quarterly
Journal of Economics.
http://www.jstor.org
This content downloaded from 203.129.217.52 on Sat, 20 Dec 2014 08:54:41 AM

All use subject to JSTOR Terms and Conditions
OPTIONS AND EFFICIENCY

STEPHEN
A. Ross
This paper argues that in an uncertain world options written on existing

assets can improve efficiency by permitting an expansion of the contingencies
that are covered by the market. The two major results obtained are, first,
that complex contracts can be "built up" as portfolios of simple options and,
second, that there exists a single portfolio of the assets, the efficient fund, on
which all options can be written with no loss of efficiency.
An option contract is any security whose returns are dependent

on the returns of some other underlying security (or securities).
Warrants, puts, and calls are familiar financial examples of options written on stocks, but preferred stock and subordinated debentures and even such diverse contracts as life insurance policies
could also be viewed as options. The serious study of options in
the financial literature began with the long-neglected thesis by
Bachelier and was revived in the 1960's by a number of authors
who focused in on the pricing problem, i.e., the problem of determining the equilibrium relationship between the value of an option and
the value of the stock it is written on.1 The intention of this paper
is to consider the related problem of the efficiency aspects of option
contracts.
Arrow's 2 introduction of the state-space approach to uncertainty in economics has brought the recognition that an inadequate
number of markets in contingent claims would be a source of inefficiency. In the state-space approach the random events that
might occur are subsets of elementary points or "states" in a (probability) space, and the possibility of inefficiency arises whenever
the feasible set of pure contingent claims, claims to wealth if a
single state occurs and nothing otherwise, fails to span all the state
space. An easy way to understand this is by analogy with a market
where individuals are permitted to purchase a grapefruit only if
they also buy an orange. If, by a fluke, everyone wishes to consume
* This work was supported by a grant from the National Science
Foundation. The author is grateful to the Rodney L. White Center for Financial Research for support, to the members of the Economic Theory Workshop of the University of Pennsylvania, and the July 1974 A. T. and T.
economics seminar for helpful comments.
1. For a bibliography of this literature, see John C. Cox and Stephen A.
Ross, "The Pricing of Options for Jump Processes," Rodney L. White Center
for Financial Research, Paper No. 2-75, University of Pennsylvania.
2. Kenneth Arrow, "The Role of Securities in the Optimal Allocation of
Risk-Bearing," Review of Economic Studies, XXXI (April 1964), 91-96.

76
QUARTERLY JOURNAL OF ECONOMICS
one grapefruit with one orange, this constraint has no force. Otherwise, opening separate markets would improve efficiency.
This result, however, must be qualified. On the one hand, many
of the states will be idiosyncratic to individuals, and events on these
states will be independent across individuals permitting a simplification of the efficient market structure. Malinvaud 3 has recently
confirmed this intuition and demonstrated that with large numbers
of individuals a simple insurance program in lieu of the theoretically
requisite complete contingency markets will remove this source of
inefficiency. Second, economists have now begun to consider explicitly the impact of transactions and set-up costs of an institutional sort on equilibrium and efficiency.4 If the introduction of a
contingent claims market will use more resources than it will save, in
an opportunity cost sense, by moving closer to efficiency, then within
the context of the institutional structure of the economy the absence
of the market is required for efficiency. Nevertheless, it is difficult
to believe that such costs would be so prohibitive as to prevent the
formation of nearly all contingent claims markets. Yet with the
exception of some insurance examples, contingent contracts are
difficult to find in actual markets. Even if we eliminate individualistic partitions of the state space, the number of states may greatly
exceed the number of assets, and competitive equilibrium could be
significantly inefficient.
The possibility of writing option contracts opens up new spanning opportunities. Although there are only a finite number of
marketed capital assets, shares of stock, bonds, or as we shall call
them "primitives," there is a virtual infinity of options or "derivative" assets that the primitives may generate. Furthermore, in
general, it is less costly to market a derived asset generated by a
primitive than to issue a new primitive, and there is at least some
reason to believe that options will be created until the gains are
outweighed by the set-up costs.
The main purpose of this paper will be to explore the relationship between primitive assets, derived options, and the attainment
of theoretical efficiency. We shall not be concerned explicitly with
transactions costs, but the relative cheapness of forming options, as
opposed to primitives, underlies much of the analysis. The basic
framework and definitions are presented in Section I and used to
3. E. Malinvaud, "Markets for an Exchange Economy with Individual
Risks," Econometrica, XXXXI (May 1973), 383-410.
4. See, for example, F. H. Hahn, "Equilibrium with Transaction Costs,"
Econometrica, XXXIX (May 1971), 417-40.

77
prove a theorem relating general types of options and efficiency.

In Section II some representation theorems among different types of
options are obtained. The representation theorems are used to study
the relationships between simple options and efficiency in Section
III. In Section III we prove a rather surprising result that greatly
simplifies the structure of option markets. In effect, we argue that
there is a single efficient fund or portfolio of the primitive assets
on which options can be written to enhance efficiency. Section IV
summarizes and concludes the paper.
I
In the state-space framework, commodities are viewed as functions on the underlying state space. For simplicity and without loss
of generality, we shall interpret each random vector as a security
yielding returns denominated in dollars in each state of the world
per unit investment. A typical asset x, then, is a map from the state
space f2 to the line E:
x: s-4E.
If the range of x is restricted to E+, then the market is organized so
that the asset offers limited liability. We shall assume that the
state space ? is finite, that a= {01, . . . , 64}, and that there are n
primitive assets {1, . . . , xn}. The set of primitives is assumed to
be invariant and cannot be altered, i.e., production decisions are
precluded. We shall use X to denote both the mxn state-space tableau, with entries xij, the gross return on asset j in state i, and the
set of n primitives.
Associated with X is the generated set,
Px
{Z I (aqaEEn) Z= Xa},
of derived assets attainable by forming portfolios a of the primitive
assets. In the definition of Px we permit short sales. We could, however, restrict z to be nonnegative to avoid the possibility of bankruptcy in any state. With this restriction Px is a polyhedral cone
in the positive orthant E+8, and the results below are unaltered.
If X has rank (or dimension) p (X), then Px lies in and spans a
subspace of dimension p(X). If p (X) = m, then there will exist a
matrix of portfolios A, such that
XA = Im,
i.e., X will possess a right inverse. This is equivalent to our being
able to combine the primitives so as to form a complete set of pure

78
contingent claims offering a return in only one state and zero in all
the other states.
We shall assume that each of the states is critical in the economy
in the sense that all of the states must be spanned (by contingent
claims) to attain full Pareto efficiency. A sufficient condition for
this to be true is that for each state there is some individual who
values wealth in that state (and is not satiated). In an important
sense, though, it is difficult to see how a state could appear in the
tableau without its being critical for efficiency. States are merely
elements in a minimal mathematical construct Q, chosen to be just
large enough to explain observed realizations. If a state appears in
Q., it is required to explain anticipated realizations, and as such
must be critical. The criterion for efficiency then is that there exist
assets to span all the states.
If, as is typical, there are more states than primitives, then we
cannot span all of the states, and competitive equilibrium will be
inefficient. Even though X fails to span Q, however, it may be
possible to augment the rank of X sufficiently by forming options
on the existing primitives. This possibility is the focus of this paper.
Of course, we are neglecting the consideration that the creation
of markets in new assets will be costly. In general, efficiency must
be assessed across alternative market and institutional structures.
If costs are sufficiently high, it will be inefficient to open all the
markets even if it does permit all the states to be spanned. (If
costs are low, however, unless markets have significant public goods
aspects, it is not clear why they will not be open in competition.)
Our concern, though, will be solely with whether pure, or theoretical,
efficiency is attainable. We shall use a crude ordinal notion of cost
to establish a taxonomy of options. In effect, we shall prohibit
some options as exorbitant in their resource use, and those that are
allowed will be considered as costlessly marketable.
To begin with a concrete example, consider a call option written
on an asset x. A call option promises a gross payment of
c (x;a)
max{x () -a, O}
in state of the world 0, where a is the exercise or threshold price.
Figure I illustrates the option contract. If
a> max x
0
then co(x;a)= 0, and the call option will have a zero gross rate of
return in all states. For
O<a< max x
0

79
the gross return will depend on the price of the option, which will be
determined in the equilibrium. For all (positive) prices, though,
the gross return will be proportional to c0(x;a), and this is all that
we need to know for our purposes.
aa
c~~~~~~~~~~~~~~e
( x; a)
0~
Xe
FIGURE,
If x has limited liability, then

co (x;O) = x(O) ,
i.e., a call with a zero exercise price is equivalent to. the primitive
asset on which it is written.
Similarly, we can define a.put on an asset x by its gross payment,
pa (xha) ~-max {0. a -x (0) },
where a is the exercise price of the put. Inversely to a call, for a
limited liability asset
pao(x; ;)
= 0,
and in general, as first pointed out by Kruizenga,-5

co (ha) - p (xma)
= max {x (0) -a,O} -max {O,a - x(0)}
=x(0) -a.
The following examples illustrate the use of options and some
of their limitations in permitting the attainment of efficiency.
5. Richard J. Kruizenga, in The Random
Prices, Paul Cootner, ed. (M.I.T. Press, 1964).
Character of Stock Market

80
Example 1. Let X contain a single asset x with returns in the

three states
2I
x=
3_
since p(X)
By itself X cannot span = {01,02,03},
calls on x with exercise prices 1 and 2, we have
O?I
c (x;1)
1 <3. Forming
and
0
c (x;2) =
8_
Now the rank of the augmented tableau,

I 0 0
2 1 0
_3 2 1_
is full, and the call options permit us to attain efficiency.
[x
C
c(X;l!.
c(x;2)]=
Example 2. Let the single asset in X be

x=
2
2
Now, all nontrivial call options on x have the form,

l 2-a
if a<2,
2-a
C(xha)=
3-a_
a
if 2<a<3.
_3-a_
Any augmented matrix formed with call options on x will have

the first two rows identical and be of less than full rank. The same
is true for put options as well.
The second- example illustrates an important point. By definition an option is defined on the range of the random returns. The
range defines the limited class of events that the random assets can
distinguish among, and we cannot write options that distinguish between two states in which all assets have identical returns. (Presumably, though, the states are distinguishable from an efficiency
viewpoint.) Quite generally, if we view the state tableau X as a
mapping from Q into El, i.e.,

81
X: nyEnl
then a general or multiple option M is a mapping
M: EnsuE
giving a composite mapping M(X(0)) on states. It is important to
emphasize that an option's return depends only on the return on the
underlying assets it is written on and not on which state occurred.
Letting M denote the class of general options and Ox(M) the space
spanned by X and all general options that can be written on X, we
have the following simple result.
1. The dimension of Ox (M) is full if and only if no two
rows of X are identical.
THEOREM
Proof. If two rows Oi and Oj of X are identical, then for all

multiple options
M (X,) = M (Xj),
where Xk denotes the kth row of X, and the augmented tableau
Ox(M) will not be of full rank. Conversely, if all rows of X differ,
then we can define the option G, as
Gi (Xj) =
O if j=/.i
I1 if j-i
and
[ GI (X1i)G2 (X2)
-Im
spanning all the states.
. Gm(Xm)]
Q.E.D.
Theorem 1 is somewhat obvious, but it does serve to formalize
the conjecture that a sufficient condition for spanning n is that for
any two states there be some asset that distinguishes between them.
It is clear that X is not necessarily of full rank simply because
(for all i,j) (alk)
Xik=,-1Xjk)
but when we augment X with multiple options, this condition is

sufficient to permit us to distinguish among states and span ?.
Multiple options, though, are really quite general and, in
practice, contracts written on multiple contingencies are extremely
rare. If we rule out such options as too costly, it still might be possible to augment the rank of X using simple options. By definition,
a simple option 0 maps E into E, and if x(O) is an asset, O(x(O))
is the return of the option in state 0. The class of simple options is

82
QUARTERLY
JOURNAL
OF ECONOMICS
thus quite large (essentially the class of all functions on E to E),

but, fortunately, it is sufficient to consider only puts or calls.
II
Puts and calls are basic examples of simple options. In this section
we shall show that all simple options can be thought of as portfolios
of puts and calls. Before proceeding, however, there is one minor
point to be taken up.
If there is some state 0, in which all assets give a zero return,
then there are simply no resources available in this end of the
world state. As such it is illusory to construct options that give a
positive return in such states. We shall eliminate this problem by
considering only the rank of the restricted tableau with the property
that for each state there is some asset giving a positive return. This
property will be referred to as the productivity assumption. We
shall now prove in the next two theorems that it is sufficient to consider only call (or put) options to study the power of simple options.
2. Let Nx denote the space spanned by all the simple
options written on the primitives X, and let Ox denote the
space spanned by the call (put) options that can be written on
X. It follows that
Nx=Ox.
THEOREM
Proof. Since calls are simple options, OxCNx. To prove the

converse, let yENx. It follows that (axA,)and Na such that
m
(1)
y= >
wa
y=1
where N1 is a simple option written on some primitive asset.

It suffices, then, to show that each simple option is equal to a
linear combination of puts and calls. Let x be a primitive asset on
which one of the simple options N is written. Order x so that
A basis for the space spanned by the calls on x is the set of calls
ci
~co(X ;Xi_1),
where we set xo <x1.

Now partition the states into the k subsets sj, . . ., Sk, with
indices
each, on which xi= xi for ijESk. By definition N is constant
ki
on each subset. Therefore, defining

(2)
y7jZO
83
Nl/(Wl.Wo)
for I<i<kl
N,
Nk
+1
N1-S (Xkl
+ 1 -S X)
Xkl+l-X1
yi=O for kj<i<k2,

etc., where Nj is the return on the simple option N in state i,
we see that
m
Y.
i= 1
yic .
The productivity assumption assures us that the calls are not illusory and the call options, alone, span N. A similar argument holds
for put options.
Q.E.D.
If X is restricted to limited liability assets and we are not permitted to write calls with negative exercise prices, then Theorem 2
is no longer true. Consider the following example.
Example 3. Let
0 1 0
O 1 1
1 1
Clearly X is productive, and p(X) = 3. Furthermore, all call options

on the assets (columns) are simply proportional to the asset they are
written on and, as such, they cannot augment the rank of x. Writing
a put on the first asset, though, with a unit exercise price gives a
return of
and
11
is of full rank. The example is not unique, and with limited liability
we can span the simple options by using both puts and calls.

84
3. Let Nx denote the space spanned by all the simple options written on the primitives X, and let Ox denote the space
spanned by the put and call options that can be written on X.
It follows that
Nx= Ox,
even if X is limited liability and exercise prices are nonnegative.
THEOREM
Proof. The proof is the same as that of Theorem 2 unless x1- 0,

in which case we cannot write a call with an exercise price below x1.
By the productivity assumption, however, there must exist other
assets {y1, . . . , yk1} such that
yi> 0.
This permits us to write a put on x with a positive exercise price a,
less than Xkl+1. Now, setting
N1
a
in (2) and substituting the put for c1 and removing
xl
i_ in the
formulas for yi, i> 1 permit us to span NXas in (2).

Q.E.D.
III
In this section we prove our main result, Theorem 4, that there
is a single portfolio or efficient fund of assets such that simple options on that portfolio will span the same space as general multiple
options. Having obtained the representation theorems in Section II,
we can study the class of simple options by studying puts and calls.
We shall distinguish two situations. In the first case we are assumed
to be able to write simple options not only on the primitives in X,
but also on portfolios of the primitives, i.e., the elements of Px.
In the second case we assume that options can be written only on
the primitives. As we shall see, the distinction is a meaningful one.
For any state-space tableau F, define OFto be the space spanned
by F, and all simple options that can be written on F. The first case
is treated in the following theorem.
4. A necessary and sufficient condition for p (OPx) =m
that there exists a single portfolio a such that
(3a)
Xa=b
with bi1#bj all (ij).
THEOREM

85
If limited liability provisions apply, then we also require

that
(3b)
min b6>o,
as a necessary and sufficient condition for efficiency.

Proof. Suppose that an <ab>
pair exists. Since bEPx, by
Theorem 1 considering only options written on b we can span Opx.
By Theorem 3 this can be accomplished with calls if (3b) is satisfied.
Conversely, suppose that there does not exist an <ab> pair
satisfying condition (3a). This means that for any given a, (a0,0')
(4)
Xaa=Xa a.
The set
af
{a I (a030'1,0) (Xo-Xo ) a= O}
is the union of a collection of linear manifolds,
r= U As,0,
0,'
0#0,
where
A a
{a I (XO-XO)a =O} .
If we cannot obtain a solution to (3a) for any a, then a==En. Since
a finite union of linear manifolds cannot have a dimension in excess
of the highest dimension among the component sets, (ao,o')
=
As=
En.
It follows that X =X0' and by Theorem 1, p (Opx) < m.

In the limited liability situation, if there does not exist an
<ab>
pair satisfying both (3a) and (3b), then for any given
a (316,6')satisfying (4) or 0 such that
X~aaO.
Defining
As
{a I Xoa-O},
we see that the set
aU[UoA0] =n.
If for some (0,6'), Aoo,= En, the proof is as above. If not, then a
will not contain the positive orthant. Since X- is nonnegative, AO
will be disjoint from the positive orthant, unless X0 = 0, violating
the productivity assumption and, of course, not permitting us to
span f2 with calls alone.
Q.E.D.
Theorem 4 is somewhat surprising. When we are permitted to
write options on portfolios, a necessary as well as sufficient condi-

86
QUARTERLY
JOURNAL
OF ECONOMICS
tion for efficiency is that there exists a single portfolio a with the
property that options written on it can span Q. This result permits
us to link the simple options to the more general ones.
THEOREM
5. The spaces Ox(M) and Opxare identical.
Proof. Since a simple option written on a portfolio of assets in

From Theorem 1,
Ox (M) is simply a subspace of Em:
OX(M) = {x I xEEm and xi=xj if X0,=Xaj}.
From the proof of Theorem 4, though, there exists a portfolio a
with returns b such that b = bj if and only if X0 =X0,. For any
YEOx (M), then, y is an arbitrary simple option on b and therefore
X is a multiple option on X, Opx C Ox(M).
YEOpX.
Q.E.D.
In other words, by increasing the domain of simple options to
include portfolios of primitive assets, we can span the same space
with simple options as with multiple options. (Notice, of course,
that
Ox (M) =?px (M),
so that nothing is gained by permitting multiple options to be written

explicitly on portfolios.) This result is quite important; it may
explain why so few multiple options are written in practice. If
individuals are permitted to form portfolios, then there is a single
portfolio a, such that any multiple option they might wish to write
will be equivalent to a simple option on portfolio a.
Unfortunately, though, if we do not permit simple options to be
written on portfolios, then the class of simple options is not as
powerful as that of multiple options. Consider the following example.
Example 4. Let
X=
1 1
1 2
2 1
2 2
a productive tableau. Since each of its rows is unique, there exists

a multiple option that spans Q. Equivalently a portfolio with a=
(2,1) is equivalent to an asset with returns

87
1~~~~~
2
11
226
KY0
0KK
=LT
and calls written on the portfolio will also span u2.

If we consider only simple options on X1 and X2, though, we
cannot span U. By Theorem 2 it is sufficient to consider only call
options. Augmenting X by the nontrivial calls on X1 and X2, we
have
11I0
A_= [X
|C1 |C2]
1 2 0
ol.
2 11 01
2 2 11
Since
A1+A4 =A%2+A3
A is of less than full rank.

Our goal now is to characterize those X for which Ox is of full
rank. To do this, we shall want to examine the space Oxa for an
individual asset Xi in X somewhat more closely. Let Li be a matrix
with rows that have all zeros except for a 1 and a -1 in positions
k and 1,where Xk' =Xv. To illustrate, consider the asset
xi =
1
3
2
We can take
r1 0 -1 0
L0 1
0l
0 0 -I]
If we change the 3 to a 1, we would add a row,

[O 0 1 -1 0]
or
[1 0 0 -1 0].
The order of the 1 and -1 in any row is irrelevant.
The importance of Li is that it permits us to characterize Oxi.
LEMMA1.
Oxi
{X I Lix = }.
Proof. If ycO,
then
Yk=Yl
if Xkf=X1,
which implies that

88
Ly =0. Conversely, if Ly = 0, then yk = yl, whenever Xki = Xii. By

Theorem 1, y is therefore an option on Xi.
Q.E.D.
is
Thus, the set of options simply the space of vectors orthogonal to the rows of Li. We can now characterize those X for which
Ox is of full rank.
6. The rank of Ox is full if and only if the row spaces of
.
{Li, . . , L,} are mutually orthogonal, i.e., there does not
exist (al, . . . , an) with
(5)
aiLi = a2L2 = . . = anLn#7O
THEOREM
Proof. Since we can only write options on the primitive assets,

Ox = ?Xl+ ...+
(6)
?Xn
Since Ox is the sum of linear spaces, it is a linear space itself. It
follows that p(Ox) <m if and only if (aqzEEmz#O0)
polar to Ox, i.e.,
(for all yEOx)zy=O. From (6) (for all ycOxi)zy=O. Thus, z belongs to the polar set of Ox,
(7)
Otxi {z (for all ycOxO)zy = O}
={Z|
(3a) z = aL},
where we have used Lemma 1. From (7) it follows that p(Ox) <Km
if and only if
n
Otxi +{O}
y
n
which is equivalent to finding a zE

(alal,
. . .,
n
i=j
Otxt, z#O, such that
an.)
z aLi=
. . .
=anLn.
Q.E.D.
Condition (5) is actually fairly straightforward to check in
practice. By performing column operations, each (kxm) Li can be
reduced to echelon form in k operations of the type "add column k
to column 1 if there is a row 0 with a 1 in column k and a -1 in
column 1." Condition (5) can now be verified by checking if any
unit vector appears in all of the reduced Li matrices.
Applying Theorem 6 to Example 4, we have
0
L2
0 1 -1
rtlo wt-1
to
_1g
Taking a portfolio with weight + 1 on the first row and weight- 1

89
on the second row, we obtain (1 -1 -1 1) for both L1 and L2,

which means that the matrices are not mutually orthogonal. By
Theorem 6 the rank of Ox is less than 4, as we verified earlier.
IV
This paper has studied the use of options to attain -efficiency in
competitive equilibrium in the absence of complete markets. Perhaps the most interesting characteristic of the results has been the
finding that rather simple options have considerable power to accomplish this. Arbitrary simple options are equivalent to a portfolio of call options. Most important, though, complex multiple
options are equivalent to simple options written on a portfolio of
primitive, marketed assets. These reductions should act to simplify
the use of options considerably, particularly in well-organized security markets.
Beyond this work at least two extensions seem clear. On the
one hand, there is a need to strengthen the present results by weakening the mathematical structure. It should be possible to carry
out all of the above analysis for rather arbitrary a--algebrasof events
with appropriate equivalents on null sets. This would not appear
to offer any real surprises, though. In an independent paper
Schrems 6 has shown, for example, that if n is the line, then calls
span simple options when the probability measure has a density
representation.
Much more important will be the extension of these results to
an intertemporal context. In such a context as the option pricing
literature makes clear, we have to distinguish between types of
simple options according to their characteristics over time. For
example, it is well-known that an American put that may be exercised at any time before its expiration date is a quite different instrument than a European put that can be exercised only at the
expiration date. The American call, though, is the same as the
European call.7 One conjecture suggested by this paper is that complex financial instruments are made up of portfolios of American puts
and calls.
UNIVERSITY
OF PENNSYLVANIA
6. Edward L. Schrems, "The Sufficiency of Existing Financial Instruments for Pareto-Optimal Risk Allocation," Ph.D. thesis, Stanford University,
1973.
7. See Robert C. Merton, "The Theory of Rational Option Pricing,"
Bell Journal of Economics and Management Science, IV (Spring 1973), 141-83.


Options and Efficiency Author(s) : Stephen A. Ross Source: The Quarterly Journal of Economics, Vol. 90, No. 1 (Feb., 1976), Pp. 75-89 Published By: Stable URL: Accessed: 20/12/2014 08:54

Uploaded by

Copyright:

Available Formats

Options and Efficiency Author(s) : Stephen A. Ross Source: The Quarterly Journal of Economics, Vol. 90, No. 1 (Feb., 1976), Pp. 75-89 Published By: Stable URL: Accessed: 20/12/2014 08:54

Uploaded by

Document Information

Original Description:

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Options and Efficiency Author(s) : Stephen A. Ross Source: The Quarterly Journal of Economics, Vol. 90, No. 1 (Feb., 1976), Pp. 75-89 Published By: Stable URL: Accessed: 20/12/2014 08:54

Uploaded by

Copyright:

Available Formats

Options and Efficiency

Author(s): Stephen A. Ross

This content downloaded from 203.129.217.52 on Sat, 20 Dec 2014 08:54:41 AM