What is an Acceptable Rate of Return for an Undiversified Investor?

Gerald T. Garvey
Peter F. Drucker Graduate School of Management
Claremont Graduate University
Claremont, CA 91711
(909) 607-9501
[email protected]

Abstract
Investors in illiquid or non-marketable assets cannot fully diversify their portfolios. As a
consequence, their discount rates reflect individual attitudes toward risk and cannot be
computed by direct application of the CAPM and related asset pricing models. Recent
research has taken an “opportunity cost” approach to the problem, arguing that an
undiversified holder of a risky asset will require at least the return that they could have
earned by leveraging the market portfolio to achieve the same level of risk. We evaluate
this approach in a model with an explicit utility function. It turns out not to be true that
the opportunity cost necessarily understates the required rate of return, unless we also
restrict the holder of an illiquid asset to invest all her liquid wealth in the market
portfolio. When the holder can also invest in a riskless asset, the opportunity cost method
actually overstates the required rate of return for investors with sufficiently low risk-
aversion. In general, however, the opportunity cost approach provides a reasonable
approximation to the exact required rate of return over a wide range of risk-aversion
levels provided the investor can also borrow and lend.

Second Draft: September 2001

This paper can be downloaded from the

Social Science Research Network Electronic Paper Collection:
http://papers.ssrn.com/abstract=281432
The valuation of illiquid assets presents special challenges to practitioners and academics.
There are two complementary ways to go beyond ad hoc “haircuts” based on history or
intuition. The first is to use comparables; to somehow locate securities that differ only in
their marketability and then track the empirical discount (see, e.g., Bajaj et al, 2000 for a
recent discussion of the costs and benefits of this approach). The alternative is to
explicitly model and evaluate the disadvantages posed by illiquidity. That is the
approach taken here, and we focus on the loss of opportunities for diversification.1

While it is clear that a rational, risk-averse investor will require a premium for lost
diversification, standard valuation frameworks rely on the assumption that investors can
freely diversify their holdings.2 When investors are restricted from trading, the discount
rate is determined by individual attitudes toward risk in addition to objective, market-
level data.3 Recent work by Smith and Smith (2000), Meulbroek (2001) and Kerins,
Smith and Smith (2001) avoid this problem and apply CAPM principles by adopting an
opportunity cost approach. Specifically, these papers reason that the undiversified holder
of a risky asset will require at least the return that she could have earned by leveraging
the market portfolio to achieve the same level of risk. Put another way, the opportunity
cost approach requires the undiversfied position to provide the same Sharpe ratio as the
market portfolio.

This approach has the appealing property of delivering preference-free results, but as the
authors recognize, it cannot exactly capture an individual’s losses due to reduced
diversification. The reason is that few investors would voluntarily choose to take on
exactly the same amount of risk if their investment were in fact tradable. In order to
assess the choices that investors would actually make, we must specify a utility or
objective function. We explicitly model a utility-maximizing investor in the setting of
Meulbroek (2001) and Kerins, Smith and Smith (2001).

1
Other discounts for minority status or for control apply even to freely tradable securities. Constantinides
(1986) and Vayanos (1998) find that transaction costs are relatively unimportant in a fully specified general
equilibrium setting (but see also Swan, 2000). Longstaff (1995) characterizes the cost of illiquidity as
foregone insider trading profits, but his model relies on the assumption that liquidity providers for traded
securities are willing to bear arbitrarily large losses to an informed trader without any compensation.
2
This is obviously the case with CAPM but is true a fortiori for arbitrage methods such as Black-Scholes
which require complete continuous hedging.
3
Hall and Murphy (2000), for example, estimate the value of a stock option to an undiversified employee
by assuming a specific utility function.
For brevity, we call the investor a “venturer” if she holds the illiquid asset and an
“investor” if she does not. Depending on the situation, the venturer could be an
entrepreneur in a start-up firm, an employee holding stock or stock options in her
employer, or an investor in an asset with high transaction costs. In all cases, we assume
that the venturer cannot trade the illiquid asset at all. We require the return on the illiquid
asset to make the venturer as well off as if she instead could freely allocate her wealth in
classical CAPM fashion, between the market and the risk-free asset.4

The drawback to our approach is that the illiquidity discount now depends on our
assumptions about the specific form of the investor’s utility function and the degree of
risk-aversion. The first countervailing benefit is that we can check how closely the
preference-free opportunity cost approach approximates the exact premium, which
requires us to know the investor’s utility function. There are two other advantages that
emerge during the analysis:

1. To implement the opportunity cost approach we must identify the total risk the
investor bears and therefore make some assumption about her other investments.
Smith and Smith (2000), Meulbroek (2000), and Kerins, Smith and Smith (2001)
assume that any wealth not tied up in the illiquid asset is invested in the market
portfolio, so that the total risk equals that of a portfolio that is X% in the illiquid
asset and 100-X% in the market. By specifying a utility function, we are able to
endogenize the allocation of liquid wealth.

2. The investor’s optimal allocation of her liquid wealth reveals her risk-aversion.
In applications, we can work backward from her holding in the risk-free asset to
impute her risk-aversion and thereby the premium that she requires from the
illiquid asset. Put another way, to implement the opportunity cost approach we
must know the fraction of wealth represented by the illiquid asset. If we also
know the extent of her holdings in either the risk-free asset or in liquid risky
assets, we can impute her risk-aversion and further improve our estimate of the
required rate of return.5

4
One could ensure that the opportunity cost approach is always correct by stipulating that the diversified
investor chooses a holding in the risk-free asset and the market that exactly mimics her risk as an venturer.
However, in so doing the entire reason for two-fund separation is also lost.
5
Strictly speaking, in the context of the model we can figure out the exact required rate of return. But the
model makes a host of assumptions that are only approximations of reality; exponential utility, normally
distributed returns, and noiseless estimates of beta, variances, and the market risk premium.
Our analysis identifies circumstances in which the opportunity cost approach is a good or
a poor approximation to the actual discount rate of a risk-averse, under-diversified
venturer. The main results are:

1. If the venturer is forced to allocate all liquid wealth to the market portfolio, we
confirm Meulbroek’s (2001) contention that the opportunity cost approach
understates the required rate of return on the illiquid asset. The bad news is that
the understatement can exceed 100% for reasonable parameter values. The good
news is that in the more general case where the venturer can choose how to
allocate her liquid wealth, the required rate of return on the illiquid asset is
significantly closer to that computed by the opportunity cost approach. The
differences shrink further still if the venturer has access to investments that have a
lower correlation with her illiquid asset than the market. Such a “hedge portfolio”
further increases her utility, and for most reasonable parameter values, reduces
her required rate of return closer to that generated by the opportunity cost
approach.

2. If the venturer has very low risk-aversion, the opportunity cost approach actually
overstates rather than understates her required rate of return. This occurs when a
risk-tolerant venturer desires more risk than is provided by the combination of the
venture of the risky venture and the market portfolio. At the opportunity cost
approach’s assumed holding of X% in the venture and 100-X% in the market, the
venturer can boost her utility by reducing her holding in the risk-free asset to
invest more in the market. This increase in utility reduces the required rate of
return on the other portion of her portfolio, i.e., the illiquid asset. While the
result is important for logical completeness, it applies only for implausibly low
degrees of risk-aversion where an investor voluntarily chooses to bear annual
volatilities on the on the order of 100% of her wealth.

The paper is organized as follows. Section I compares the opportunity cost premium to
that required by a risk-averse investor who is forced to hold all her liquid wealth in the
market portfolio. Section II turns to our own main case where the venturer has a choice
of where to allocate her non-venture wealth, and section III concludes.

I. Results when liquid wealth is invested in the market portfolio

A. Basic modeling assumptions

We analyze a single portfolio choice and normalize the venturer’s wealth to one. The
venturer is forced to hold a fraction v of her wealth in the illiquid asset and in this section
she places the remaining (1-v) of her wealth in the market portfolio. If she does not take
on the venture and is instead an investor, she is free to allocate her wealth between the
market portfolio and the risk-free asset. In the next section we make the more consistent
assumption that the venturer can allocate her liquid wealth (1-v) between the market
portfolio and the risk-free asset.

The risk-free rate is denoted rf. The expected return on the market portfolio is denoted
rm>rf and its risk is denoted σm. We refer to the illiquid asset as a venture and denote its
return by rv, its total risk by σv. The venture’s systematic risk is denoted β and is
unsubscripted because it is the only risky asset in the model other than the market
portfolio.

B. Opportunity cost approach

Since the venturer has v of her wealth tied up in the venture and 1-v in the market, her
risk is now:

σ vp = (v 2σ v2 + (1 − v) 2 σ m2 + 2v(1 − v)Cov(rv , rm ) )
1/ 2

(1)
(
= v σ + (1 − v)σ [(1 − v) + 2vβ ]
2 2
v
2
m )1/ 2

The opportunity cost approach requires the venture to provide the same return that the
investor would obtain if she chose to bear this risk level, σvp. This return equals rf + (rm –
rf)σvp/σm. The venturer receives the expected return vrv+(1-v)rm. Solving for the critical
value of rv that equates these two returns we obtain
 σ vp   σ vp   σ vp 
rm  − (1 − v)  − r f  − 1  −1
σm σm  = r + σm 
rvopp =  
 (rm − r f ) (2)
v 
m
v
 
 

C. Utility-based approach
In order to stay with the mean-variance framework, we will assume the investor has
exponential utility of the form –e-kc/k where c is terminal consumption and k is the
coefficient of absolute risk-aversion. As is well-known, the alternative is either to
abandon the mean-variance setting or to make the even less palatable assumption of
quadratic utility. The gain of using a utility-based approach is that instead of just
postulating that the venturer would choose to bear the same risk as the venture, we can
solve for her optimal exposure to the market, w, with the remainder being invested in the
risk-free asset. The weakness is that our answers depend on our assumption of her degree
of risk-aversion, k. An additional limitation is that we restrict attention to a single period.
This helps in making the assumption of constant absolute risk-aversion more palatable,
but leaves out any consideration of wealth dynamics.

In this section, the venturer is forced to hold her non-venture wealth in the market
portfolio. This means her utility is:
kσ vp2
U v = vrv + (1 − v)rm − (3)
2
In the absence of the illiquid asset, the investor can choose her weight on the market to
maximize:
kw 2σ m2
U d = r f + w(rm − r f ) − (4)
2
Solving for her optimal holding we obtain w*=(rm-rf)/kσ2m. Substituting this optimal
market exposure back into the utility function, we can express the diversified investor’s
utility as:
kw*2σ m2 (rm − r f ) 2
U * d = r f + w* (rm − r f ) − = rf + (5)
2 2kσ m2

Thus, the venture must promise a return rv* that satisfies

kσ v2 (rm − r f ) 2 kσ v2
U v = r *v − = U d* ⇒ rv* = r f + + . (6)
2 2kσ m2 2

Proposition 1 summarizes the comparison between the opportunity cost and the utility-
based approach.

Proposition 1:

The opportunity cost and the utility-based discount rates are equal if the coefficient of
absolute risk-aversion k=k*=(rm-rf)/σvpσm. For all k ≠ k*, the utility-based approach
yields a higher discount rate.

The critical level of risk-aversion k* has a natural interpretation. It is that level of risk-
aversion for which the investor would voluntarily choose to leverage the market portfolio
to achieve the same level of risk as the venture. For all other levels of risk-aversion,
Proposition 1 confirms the contention of Meulbroek (2001) and Kerins, Smith and Smith
(2001) that the opportunity cost approach provides only a lower bound on the required
rate of return for an undiversified investor.
Figures 1.1-1.3 show clearly how the results come about. An investor with high risk-
aversion (k>k*) is forced to bear risk beyond the point where she is willing to take just
the market risk premium. An investor with low risk-aversion (k<k*) is denied her desire
to take on additional risk at the market risk premium. The venture’s expected return must
compensate for these losses as well as the lost diversification.

The weakness with the result is that part of the premium does not reflect lost
diversification but rather the fact that we have denied her the ability to allocate her wealth
between the market and the risk-free asset. That is what we address in the next section.

II. Venturer can allocate liquid wealth between market and risk-free asset

A. General Results
As indicated above, we have made a rather inconsistent assumption. The venturer can
only invest her liquid wealth in the market if she does the venture. If she does not do the
venture, by contrast, she is also allowed to freely allocate her wealth between the market
and the risk-free asset. We now allow the venturer to optimally allocate her liquid wealth
between the market and the risk-free asset. In additional to internal consistency, a
practical benefit is that we can use the venturer’s investment choices to infer her risk-
aversion and thereby her required rate of return.

Formally, we still force the venturer to hold v of her wealth in the venture, but allow her
to invest we of her wealth in the market, and (1-v-we) of her wealth in the risk-free asset.
We can write her utility as:
k 2 2
U v = r f + v(rv − r f ) + we (rm − r f ) −
2
(v σ v + weσ m2 ( we + 2vβ ) ) (7)

Her optimal market holding satisfies the first-order condition:

rm-rf -kσm2[we +βv]=0 (8)

The optimal market holding for the venturer can be written as:

we* = (rm-rf)/kσm2 - βv. (9)

An immediate concern that arises is short-sales constraints, and indeed there is nothing in
(9) that guarantees either the optimal market holding or the risk-free holding (1-v-we*)
are nonnegative. Table 1 illustrates these choices for various degrees of risk-aversion for
the case where the illiquid asset is 40% of the investor’s wealth, the volatility of the
illiquid asset is 50% with a beta of 1, the volatility of the market portfolio is 20%, and the
risk-free rate and the market risk premium are both 6%. Hall and Murphy (2000) use
values of risk-aversion in the neighborhood of three to simulate the value of an option to
an undiversified executive, based on estimates in Friend and Blume (1978). Investors
with this degree of risk-aversion will not desire to short either asset. Of course, there
remains the fundamental problem of identifying a reasonable level of risk-aversion.

We can provide some more insight into the reasonable level of risk-aversion by
indicating the portfolio choices that the investor would make. Specifically, we can
impute the risk-aversion parameter from the venturer’s holding of the market and the
venture along with the correlation between the two;

k = (rm-rf)/σm2(w*e+βv) . (10)

The dependence of the results on different parameter values is quite transparent. The raw
volatility of the illiquid asset is irrelevant, while investment in the market falls linearly in
either beta or the amount of wealth tied up in the venture. It is also worth noting that for
the reasonable parameter values in Table 1, the critical level of risk-aversion k* where
the opportunity cost is exactly correct is as approximately 1.11. The reason is quite
straightforward. The combination of the market and the illiquid asset has more risk than
the market portfolio, so if the diversified investor is to choose to bear this level of risk,
she must short the risk-free asset and leverage the market portfolio. This is not an
appealing choice for a reasonably risk-averse investor. However, it is also worth noting
that allowing her to borrow and lend at the risk-free rate if she holds an illiquid asset does
not seem any less realistic than the opportunity cost method’s required assumption that
the investor can leverage the market portfolio to achieve the same level of risk as he
venture. Specifically, Table 1 reminds us that the venturer and the investor have exactly
the same desired holding of the risk-free asset when beta is one. For betas less than one,
the desired holding of the risk-free asset is actually greater when she holds the illiquid
asset.

Proposition 2 turns from the qualitative results to assess the claim of Meulbroek (2001)
and Smith and Smith (2000) that the opportunity cost approach understates the required
rate of return.

Proposition 2

The opportunity cost approach strictly overstates the required rate of return at k=k* if
the returns on the illiquid asset are not perfectly correlated with the market.

The reason for the result is illustrated in Figure 2. In Proposition 1 and Figures 1.1-1.3,
the rate of return rv* gives the venturer the utility associated with the point the point (σvp,
vr*v +(1-v)rm). By allowing the venturer to allocate her liquid wealth between the market
and the risk-free asset, she gains access to the entire dashed frontier in Figure 2. At the
level of risk-aversion k*, her personal price of risk is exactly equal to the market price of
risk, (rm-rf)/σm at the risk level σvp. But so long as the venture is not perfectly correlated
with the market, the dashed frontier is everywhere steeper than the market price of risk,
meaning that her own total risk increases by less than σm when she increases her holding
in the market and decreases her holding in the risk-free asset. An investor with risk-
aversion k* takes on more risk, thereby increasing her utility and reducing the required
rate of return on the venture.

B. Key determinants of discount rates

B.1. Risk-aversion

The last observation driving Proposition 2 also implies that the opportunity cost and the
utility-based approach tend to be closer together when we allow the venturer to choose
her holding of the risk-free asset. For most reasonable values of risk-aversion, the
investor would choose significantly less risk than the venturer. While the venturer is not
allowed to reduce her holding in the venture, she also values the ability to scale back her
investment in the market portfolio in favor of the risk-free asset. When she can do so, her
required rate of return on the venture drops, as illustrated in Figure 3. More generally,
allowing the venturer to choose her holding in the risk-free asset can only increase her
utility and reduce the required rate of return on the venture. We now turn to the
quantitative significance of the ability to diversify.

We begin with the effect of alternative degrees of risk-aversion. Figure 4 summarizes the
results of our three alternative techniques for ascertaining the required rate of return on
the illiquid venture. By design, the opportunity cost technique does not depend on risk-
aversion, and consistent with the arguments of Smith and Smith (2000) and Meulbroek
(2001), it generally understates the required premium. This is not always the case,
however, once we take the internally consistent approach of allowing the investor to
choose between the market and the risk-free asset whether or not she holds an illiquid
asset.

For the selected parameter values, at risk-aversion levels below approximately 1.2, the
opportunity cost overstates the required rate of return. As a practical matter, this is an
extremely low level of risk-aversion, at which the investor would choose to short the risk-
free asset and ramp up her exposure to the market. When the investor has higher and
more reasonable degrees of risk-aversion, the opportunity cost method appears to
dramatically understate the venturer’s required rate of return if we do not allow her to
borrow and lend. The understatement is significantly smaller, sometimes cut in half, if
we do allow her to choose her own most preferred holding in the market and the risk-free
asset regardless of whether or not she holds an illiquid asset.

B.2. The correlation between the market and the illiquid asset

While the venturer is unable to fully diversify her portfolio, she can achieve some
diversification benefits when the market and her illiquid asset are imperfectly correlated.
Figure 5 fixed the venturer’s risk-aversion at three and plots the required rate of return
under our three alternative approaches, allowing the correlation between the market and
illiquid asset to vary negative one to one. The first result is that when the correlation is
large and negative, the opportunity cost approach overstates the required rate of return
even when the venturer is highly risk-averse. The reason is that the opportunity cost fixes
the venturer’s investment in the market at (1-v) of her wealth. When the market is highly
negatively correlated with the venture, the venturer desires a larger holding in the market
because doing so can even reduce her risk while also providing her with a positive risk
premium.

As the correlation between the market and the venture increases from negative one, the
required rate of return from the opportunity cost approach increases because the risk σvp
is increased. That is, the risk of the opportunity cost approach’s assumed investment of v
in the venture and (1-v) in the market increases as the correlation between the two assets
increases. The increase is more dramatic when the venturer can choose her holding in the
market, which is reflected in a steep increase in the required rate of return when the
venturer can choose how to invest her liquid wealth. This steep increase results in the
opportunity cost and the utility-based premium coming closer together as the correlation
between the market and the venture increases from negative one. When the correlation
between the two approaches zero, we enter the area where the opportunity cost approach
understates the required rate of return.

This understatement does not increase without limit as the diversification afforded to the
venturer by the market portfolio falls. To see why, consider the case where the market
and the venture are perfectly correlated. At that point, the opportunity cost premium is
exactly equal to the premium demanded by a diversified investor, simply because the
market and the venture have identical risk characteristics. The utility-based approach
generates an extremely large risk-premium if the venturer is not allowed to choose her
holding in the market portfolio, because for the assumed parmeter values the venturer
bearing far more risk than she wishes to bear at the market price of risk, (rm-rf)/σm. This
effect disappears when we allow her to choose the allocation of her liquid wealth, this
effect disappears. When the venture is perfectly correlated with the market and the
venturer can adjust her other exposure to the market, she attains exactly the same risk-
return profile as the fully diversified investor. The practical effect is to again narrow the
gap between the discount rate implied by preference-free opportunity cost approach and
that obtained if we knew the investor’s exact risk preferences.

D. Alternative investments beyond the market portfolio

As stressed throughout, the discount rates implied by the opportunity cost approach
appear more reasonable as we increase the venturer’s ability to choose how to allocate
her liquid wealth. To this point, we have still restricted the venturer to choose between
the market and the risk-free asset. This is a restriction, because the venturer would
generally wish to hold a different set of risky assets than would a fully diversified
investor. Specifically, she would wish to place a relatively heavy weight on assets that
have a low correlation with her venture. A full assessment of this issue would require
specifying the variance-covariance matrix between the venture and all the assets in the
economy and is beyond the scope of this paper. Qualitatively, however, freeing up the
investor’s choice of risky assets will generally increase the apparent accuracy of the
opportunity cost approach. For reasonable levels of risk-aversion and other parameters,
the opportunity cost approach seems to understate the required rate of return. If the
venturer is allowed a richer set of risky investments, her utility can only be increased,
which reduces the exact required rate of return closer to that provided by the opportunity
cost approach.

Table 2 provides some quantitative insight into the above argument. We give the
venturer access to a “hedge portfolio” in addition to the market portfolio. We assume the
hedge portfolio has a beta of one with the market, so that its expected rate of return is
simply rm. The first and obvious key feature of the hedge portfolio is its correlation with
the venture. The second is the risk of the hedge portfolio. Since it has a beta of one and
the market portfolio is efficient, the hedge portfolio must have a risk greater than that of
the market, where the latter risk is fixed at 20%. We have fixed the correlation between
the market and the venture at 50% and so the hedge portfolio is only valuable when it has
a lower correlation with the venture.

We begin with the case where the hedge portfolio is uncorrelated with the market, and
allow the risk of the hedge portfolio to vary from 30% to 100%. Qualitatively, the
presence of the hedge portfolio reduces the gap between the opportunity cost and the
utility-based premium, but the effect is modest. We then turn to the case where the hedge
portfolio has a correlation of –0.5 with the venture and thus provides her with an extreme
amount of risk-reduction. Not surprisingly, this reduces the required rate of return even
below that implied by the opportunity cost approach. To achieve such a high negative
correlation a hedge portfolio would likely consist of a short position in competitors of the
venture. We finally turn to the more realistic case where the hedge portfolio has a small
but positive correlation with the venture. In this case, the hedge portfolio makes
essentially no difference to the results reported thus far. When the risk of this portfolio
increases above 35%, the venturer prefers to use the market portfolio for diversification.

III. Conclusions
When an investor is forced away from a diversified portfolio by illiquidity, her discount
rate will inevitably reflect her own personal attitude towards risk. The opportunity cost
approach provides an estimate of such a discount rate without requiring knowledge of
individual risk attitudes. We show that this estimate is reasonable when the undiversified
investor can choose how to allocate the liquid portion of her wealth.
Proof of Proposition 1
Under the utility-based approach, the venture must provide a rate of return that satisfies:

*
kσ vp2 * rm2
U v = vr v + (1 − v)rm − =U =
d
2 2kσ m2
 (rm − r f ) 2 kσ vp2  (A1)
 + − (1 − v ) r 
 2kσ 2 2
m
 1  (rm − r f ) 2 kσ vp2 
*
⇒ rv =  m  = rm + r f − rm + + 
v v  2kσ m2 2 

We can therefore express the difference between the two discount rates as:

1 (rm − r f ) 2 kσ vp2 σ vp  
rv* − rvopp =  r f − rm + + −  − 1 ( r − r ) 
v  
m f
2kσ m2 2 σ
 m  
(A2)
1  (rm − r f ) kσ vp σ vp 
2 2

=  + − (rm − r f ) 
v  2kσ m 2
2 σm 


This difference is exactly zero at k=k*. To show that it is strictly positive otherwise, it
suffices to show that the difference is minimized at k=k*. This in turn follows from the
facts that (i) the derivative:
∂ (rv* − rvopp ) 1 (rm − r f )
2
σ 2pv
= [− + ] = 0 if k = k * , (A3)
∂k v 2k 2σ m2 2

and (ii) from the fact that k* is in fact a global minimum by the second-order condition:
2
∂ 2 (rv* − rvO ) 1 (rm − r f )
= >0 (A4)
∂k 2 v k 3σ m2

Proof of Proposition 2
We now have to solve for the return on the venture that equates the venturer’s maximized
utility with and without the venture. The most convenient way to do this is to re-express
her expected utility as an investor using the fact that w* = w*e + βv. Specifically, we can
write her expected utility as

U = r f + (rm − r f )( w + βv) −
* * (
kσ m2 we*2 + 2 we* βv + β 2 v 2 ) (A5)
d e
2
Now, if the venturer does do the venture, we can write her utility as:

U * *
= r f + w (rm − r f ) + v(rv − r f ) −
( )
k v 2σ v2 + we*σ m2 [ we* + 2 βv]
(A6)
vp e
2
Now solve for the value of rv that equates U*d and U*vp. Separating out the risk and
expected return terms we have rvv- rmβv = 0.5kv2[σv2- β2σm2]. Finally, we can express
the required return on the venture as :

rv = rf+β(rm-rf) + 0.5kv[σv2- β2σm2]. (A7)

Now note that since r*v equates U*vp and U*d and U*vp strictly increases in rv, an
increase in Uvp will strictly decrease r*v. Second, note that if the venturer chooses to
deviate from holding 1-v in the market and bearing risk σvp when her risk-aversion is k*,
she will strictly increase her utility Uvp by so doing.

Recall that k* is defined as that level of risk-aversion that would lead the venturer to
choose the market exposure 1-v and the risk level σvp if she faced the market price of
risk, -∂(expected return/)/∂w/∂(risk)/∂w = (rm-rf)/σm. But when the venture is imperfectly
correlated with the market, an increase in her market holding provides her with enjoys
compensation for risk that is strictly greater than (rm-rf)/σm. Therefore at k=k*, she will
take on more risk and increase her utility above U*d. The reason she faces a trade-off
more favorable than the market price of risk is that her expected return is rf+wrm+vrv so
∂(expected return/)/∂w = rm. But her risk is now:

σ vp = (v 2σ v2 + σ m2 [ w 2 + 2wvβ ])
1/ 2
(A8)

If the venture is perfectly correlated with the market, β=σv/σm and σvp = vσv+wσm and
∂(risk)/∂w = rm so the venturer faces exactly the market price of risk and will maintain
her holding in the market at 1-v. For all β<σv/σm, the venturer faces less than the market
price of risk and will increase her holding in the market and more importantly her utility,
U*vp.
References
Amihud, Y. and H. Mendelson (1986) “Asset pricing and the bid-ask spread”, Journal of
Financial Economics 17, 223-49.

Bajaj, M., D. Denis and A. Sarin (2001), “Firm value and marketability discounts”,
working paper

Blume, M and I Friend (1978) The Changing Role of the Individual Investor A 20th
Century Fund Report, New York, Wiley and Sons.

Constantinides, G. M. (1986) “Capital market equilibrium with transactions costs”,

Journal of Political Economy 94, 842-62.

Kerins, F., R. L. Smith and J. K. Smith (2001) “Opportunity cost of capital for venture
capitalists and venturers”, working paper, Claremont Graduate University

Hall, B. and K. J. Murphy (2000) “Stock options for undiversified executives”, working
paper, Harvard Business School

Hertzel, M. and R. L. Smith (1993) “Market discounts and shareholder gains for placing
equity privately”, Journal of Finance 48, 459-86.

Longstaff, F. M. (1995) “How much can marketability affect security values?”, Journal
of Finance 50, 1767-74

Meulbroek, L. K. (2001) “The efficiency of equity-linked compensation: Understanding

the full cost of awarding executive stock options”, Financial Management, Summer, 5-
30.

Pratt, S. (1989) Valuing a Business, Irwin, Homewood Ill.

Smith, J. K. and R. L. Smith (2000) Entrepreneurial Finance, Wiley, NY.

Swan, P. L. (2000) “Does illiquidity rather than risk-aversion explain the equity premium
puzzle?”, working paper, University of Sydney

Vayanos, D. (1998) “Transaction costs and asset prices: A dynamic general equilibrium
model”, Review of Financial Studies 11, 1-58.
 Figure 1.1: Required rates of return with high risk-aversion

Exp.
Return

U*d,
k<k*
vrv*+
(1-v)rm

Vropp +
(1-v)rm

σv

Risk
 Figure 1.2: Required rates of return when k=k*

U*d,
Exp. k=k*
Return

vropp+
(1-v)rm

σv

Risk
 Figure 1.3: Required rates of return with low risk-aversion

U*d,
k>k*
Exp.
Return

vrv*+
(1-v)rm

vropp+
(1-v)rm

σv

Risk
 Figure 2: How the opportunity cost approach overstates the required rate of return when
k=k*

U*d,
Exp. k=k*
Return
Feasible risk-
return pairs when
rv=r* v

vropp +
(1-v)rm

vrv*+
(1-v)rm

σvp
σvp(w*e)

Risk
 Figure 3: Why allowing for choice of risk-free investment reduces the apparent error in
the opportunity cost discount rate for high risk-aversion

Exp.
Return

vr*v+(1-v)rm,
no risk- U*d,
free asset k<k*

vrv*+(1-v)rm,
risk-free asset

Vropp +
(1-v)rm

σvp(w*e) σvp

Risk
 Table 1: Portfolio choices for sigma(v)=50%, v=40%, sigma(m)=20%, beta=1, rm-
rf=rf=6%

Investor's
Coefficient of weight on Investor's
absolute risk- Venturer's weight on Venturer’s weight on market weight on risk-
aversion market portfolio risk-free asset portfolio free asset

0.5 2.60 -2 3 -2.00

1 1.10 -0.5 1.5 -0.50

1.5 0.60 0 1 0.00

2 0.35 0.25 0.75 0.25

2.5 0.20 0.4 0.6 0.40

3 0.10 0.5 0.5 0.50

3.5 0.03 0.57 0.43 0.57

4 -0.03 0.63 0.38 0.63

4.5 -0.07 0.67 0.33 0.67

5 -0.10 0.7 0.3 0.70

 Figure 4: Required rates of return when sigma(v)=50%, sigma(m)=20%,
v=40%, beta=1, rf=rm-rf=6%

0.6

0.5

0.4
required rate of return

r*v if can choose rf

0.3 opp cost prem
r*v if can't choose rf

0.2

0.1

0
0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2 2.2 2.4 2.6 2.8 3 3.2 3.4 3.6 3.8 4 4.2 4.4 4.6 4.8 5
Coefficient of absolute risk-aversion
 Figure 5: Required rates of return when sd(v)=40%, sd(m)=20%,
v=40%, rf=rm-rf=6%, k=3, and corr(rv,rm) ranges from -1 to 1

0.4

0.35

0.3

0.25
Required rate of return

0.2
r*v if can choose rf
0.15 opp cost prem
r*v if can't choose rf
0.1

0.05

0
2

8
.8

.6

.4

.2

1
-1

0.

0.

0.

0.
-0

-0

-0

-0

-0.05

-0.1
Correlation between market and venture
Table 2: Required rates of return when the hedge portfolio has β=1, Sigma(v)=50%,
Sigma(m)=20%, rf = rm-rf = 6%, k=3, Corr(rm,rv)=0.5.

Sigma(h) Corr(rv,rh) Opp. Cost Required return Required return

required return when market is when hedge
(%) alternative (%) portfolio is also
available (%)

30% 0 18 24.8 23.8

50% 0 18 24.8 24.1

70% 0 18 24.8 24.4

100% 0 18 24.8 24.6

30% -0.5 18 24.8 14.3

50% -0.5 18 24.8 17.4

70% -0.5 18 24.8 18.6

100% -0.5 18 24.8 19.3

30% 0.25 18 24.8 24.6

50% 0.25 18 24.8 24.8

70% 0.25 18 24.8 24.8

100% 0.25 18 24.8 24.8