9

Corporate Capital Structure vs. Project

Financing

9.1 Introduction
This chapter discusses project finance. From my experience, when speak-
ing about project finance people think about different things. Often they
are confused because they do not understand why project finance should
be discussed as a separate topic since it is, seemingly, a part of almost
every other topic including general topics like investments, net present
value, etc. We will learn in this chapter that project financing has a very
special meaning in finance and is often related to terms like non-recourse
debt, limited recourse debt, asset-backed securities, and many others. To
begin, let us note that project financing has been used in many important
projects around the globe including many historical projects. Below we
will review some of them.
Lyonett du Moutier (2010) describes the famous Eiffel Tower construc-
tion, which was the world’s tallest structure at its completion in 1889.
Apparently, a project finance model was used. Public authorities awarded
a concession to the project’s private sponsor. The sponsor was financing it
with equity and limited-recourse debt. Lyonett (2010) a­ nalyzes how spe-

© The Editor(s) (if applicable) and The Author(s) 2016 183

A. Miglo, Capital Structure in the Modern World,
DOI 10.1007/978-3-319-30713-8_9
184 Capital Structure in the Modern World

cific provisions in the Eiffel Tower project contracts helped reduce agency
costs, which are often very substantial in large projects.
The World Bank’s project finance and guarantees group describes on
its website (2005) the development of the Nam Theun 2 hydropower
project.1 Since 1995, it has been a top priority for the Government of
Laos. Electricité de France of France, Italian-Thai Development Public
Company Limited of Thailand, and Electricity Generating Public
Company Limited of Thailand sponsored the project. It is the largest ever
foreign investment in Laos, the world’s largest private sector cross-border
power project financing, the largest private sector hydroelectric project
financing, and one of the largest internationally financed projects in Asia.
The project has been structured as a limited recourse financing which,
the article argues, allows for an efficient allocation of multiple risks that
are usually involved in large construction projects including completion
risk, commercial and political risks, geological risk, and risk of timely and
within budget completion.
In a December 2015 article by Matthew Amlôt we find that the
National Central Cooling Company PJSC (Tabreed) announced its
completion of the signing of an AED 192.5 million long-term limited
recourse project finance facility with Emirates NBD for the district cool-
ing plant it is developing for Dubai Parks and Resorts.2 Jasim Husain
Thabet, Tabreed’s Chief Executive Officer, said, “Tabreed’s approach to
financing new projects is to utilize long-term project financing wherever
possible which we believe is the best way of financing our assets and max-
imizing value for all stakeholders. This loan facility with Emirates NBD
is a landmark deal as it is amongst the first district cooling transactions
completed in the UAE under a limited recourse project finance structure
for a Greenfield development.”
In 2004, project financing in the United States made up 10–15 % of
total capital investment.3 In 2007 the total volume of project financ-

1
http://siteresources.worldbank.org/INTGUARANTEES/Resources/Lao_NamTheun2_Note.pdf
2
Amlôt (December 16, 2015).
3
Esty (2004).
 9 Corporate Capital Structure vs. Project Financing 185

ing reached 44,476.30 billion (Gardner and Wright 2012). According

to Corielli, Gatti, and Steffanoni (2010), the compound annual growth
rates across all global markets from 1994 to 2006 were 23 % for project
finance bank loans and 15 % for project finance bond issuances. Such
outstanding growth trends are also found by other authors (Kleimeier
and Versteeg 2010). With its rapid growth, project finance became one
of the most important sources of finance in the financial market. Even
the global financial crisis of 2008 has had a less harmful effect on project
finance transactions than on syndicated loans as a whole. To be precise,
the global lending volume of syndicated loans decreased by 44 % com-
pared to 2007. At the same time, the amount of project finance transac-
tions declined by only 10 % and is still signaling promising growth trends
(Bonetti et al. 2010).
While project financing has had a long history dating back to Roman
times, the 1980s saw the emergence of project financing as a popular
capital structure choice. More recently, it has become a global phenom-
enon spreading not only to developed countries but to less developed
countries as well. In 1985, project financing was used in approximately
80 countries. In 2003, this number increased by 75 % to 140 coun-
tries.4 According to Standard & Poor’s predictions, presented at the
20th Annual Project Finance Hot Topics Conference (November 1,
2011), the following 20 years should present a tremendous change in
the composition of project finance deals. Among the main reasons for
those changes are increasing privatization processes and the urge for
enormous investments into infrastructure development in low income
countries.
Generally speaking, project financing is the financing of a project by
a sponsoring firm, where the cash flows of the specific project are desig-
nated as the source of funds to repay the loan. The project is incorporated
as a separate legal entity, and the assets and cash flows are segregated
from those of the sponsoring firm. The firm repays this new entity’s debt
obligations solely from the cash flow generated by the operation of the
project.

4
Hainz and Kleimeier (2012).
186 Capital Structure in the Modern World

In this sense, the debt of a project-financed venture is known as

“non-­recourse” debt because the creditors have no recourse to the
assets of the parent company. If, on the other hand, a firm jointly
incorporates the new project and issues standard corporate debt to
raise the initial investment cost, the lenders will usually have first
recourse to any assets generated by the entire company. These are not
necessarily attributable to the returns on the new project itself. Non-
recourse debt allows the firm’s managers to “protect” the remaining
assets of the firm because the lender is only entitled to the assets of the
project-financed company.
The theories discussed in this chapter differ in their explanation of
the use of project financing. They look to agency costs, asymmetric
information, or corporate control considerations as the primary rea-
sons to use project financing. The models that focus on agency costs
look primarily at how separating the debt claims of old and new credi-
tors leave shareholders’ payoffs unaltered and thus reduce the agency
costs of underinvestment or overinvestment. The models that look
at asymmetric information look primarily at the lack of information
for investors and how project financing may signal information to
investors and mitigate the mis-evaluation of securities issued by the
firm.
Existing empirical literature has discovered the following stylized
facts (Esty 2003; Kleimeier and Megginson 2001): (i) project compa-
nies are characterized by high capital expenditures, long loan periods,
and uncertain revenue streams; (ii) project companies use very high
leverage compared to corporate financed companies; (iii) project com-
panies have very concentrated debt and equity ownership with debt
being non-recourse to the sponsoring companies; (iv) project finance
requires finance executives to disclose more propriety information than
they would have to disclose under corporate finance; (v) structuring a
project finance company legally, rather than financing the project as
part of the corporate balance sheet, is more expensive and takes much
more time.
 9 Corporate Capital Structure vs. Project Financing 187

9.2 Moral Hazard Models

The central concept around which most discussions on the optimality of
project financing have been developed is the underinvestment problem.
As Myers (1977) argues, the latter can arise because of, for instance, a
debt overhang (see Chap. 5). Theoretical papers that will be discussed in
this section show that project financing can both solve underinvestment
problems and prevent them from arising.
Berkovitch and Kim (1990) look at moral hazard problems and exam-
ine how project financing works to effectively overcome the underinvest-
ment incentives of entrepreneurs. They develop a theoretical model with
the primary goal of investigating the effects of seniority rules and restric-
tive covenants on the incentives associated with risky debt.
Berkovitch and Kim’s model is set up as a two-period model. An invest-
ment project is undertaken in the first period by issuing debt. A firm’s
managers will decide whether to accept or reject the new project in the
second period. Berkovitch and Kim show the presence of underinvest-
ment incentives by examining the effect that the new project will have
on the market value of the old debt. When cash flows from a new project
accrue to existing creditors, as opposed to shareholders, the market value
of the old debt increases. This causes the shareholders to reject the project.
By using project financing, the firm separates the new investment from
its current assets, as well as the current debt obligations, and removes the
underinvestment problem because old creditors do not accrue any ben-
efits and there are no wealth transfers.
The following illustrates Berkovitch and Kim’s main insights into proj-
ect financing. There are two dates ( t = 0,1 ). The interest rate is equal to 0.
Suppose a firm owns the following project (project-in-place). At t = 1 , cash
flow from the project depends on the state of nature. In the good state G
(that occurs with probability p1) it equals R1 > 0 . In state B (bad state) it
equals R2 , R1 > R2. At t = 0, the firm is given the opportunity to make an
investment K, which generates C1 > 0 in state G and C2 in state B, C1 > C2
. This investment has a positive NPV, i.e.

p1C1 + (1 − p1 ) C2 − K > 0 (9.1)

188 Capital Structure in the Modern World

A irm with
debt and
Shareholders
assets-in-place
decide whether
receives
to accept or
information
reject the
about a new
project
investment
opportunity

If the project is
accepted, the Earnings are
irm issues realized and
new debt to distributed to
inance the claimholders
project

Fig. 9.1 Sequence of events

The firm issued senior debt with total amount D such that

R1 + C1 > D > R2 + C2
(9.2)

This means that in the good state the project’s payoff exceeds the
amount of debt but not in the bad state. The sequence of events is as in
Fig. 9.1.

Proposition 9.1 The usage of project financing (non-recourse debt) can

solve a debt overhang problem.

Without the new project the shareholders’ expected profit is

p1 ( R1 − D )
(9.3)

If the firm invests in the new project, the shareholders’ expected profit
is: p1 ( R1 + C1 − D ) − K . This is equal to the firm’s expected cash flow (the
total earnings in state G reduced by the payment to the senior creditors
 9 Corporate Capital Structure vs. Project Financing 189

multiplied by the probability of success) minus the payment to the new

outside investors. Whether the financing of the new project is done using
junior debt or equity, the expected payment to the investors should be
equal to the cost of the investment. Comparing the shareholders’ payoff
without the investment and with the investment we find that the firm
will invest if p1C1 − K > 0 .
If p1C1 − K < 0 the new project (even though it has a positive NPV) will
not be undertaken. This is similar to the debt overhang problem because
the presence of existing risky claims reduces the incentive for existing
shareholders to invest in that project because it increases the value of the
existing debt and not the value of shares, which is the essence of the debt
overhang problem.
Now suppose that the firm uses project financing (non-recourse debt).
In this case, the debtholders’ payoffs depend only on the returns from the
new project and not on the returns from the assets already in place. The
face value d can be found from:

K = d p1 + (1 − p1 ) C2 (9.4)

If B is realized, the debtholders get the cash flow from the new proj-
ect (C2). The shareholders’ expected payoff (note that the shareholders
get nothing if B is realized) is: p1 ( C1 − d + R1 − D ). This is greater than
(9.3) because it follows from (9.1) and (9.4) that d < C1. Hence, the
­shareholders’ payoff with the new project is higher than it is without
it. So the project will be undertaken and it will be financed with non-
recourse debt.
The second situation is based on overinvestment: the shareholder’s
incentive to invest in negative NPV projects (similar to the asset substitu-
tion effect from Chap. 4). Consider the same firm with assets in place,
as described previously. This time, assume that the firm has a senior debt
with a face value

D > R1 (9.5)

The new project reduces the probability of G occurring by Δp. Also

assume that C2 = 0 and
190 Capital Structure in the Modern World

∆pR2 − ∆pR1 + ( p1 − ∆p ) C1 < K (9.6)

This implies that the change in the firm’s expected earnings (left side of
this inequality) is less than the amount of the investment, which means
that the new project has a negative NPV.

Proposition 9.2 The usage of project financing can solve overinvestment

problems.
Suppose that to finance the new project the firm issues a standard
subordinated debt. Assuming risk neutrality and a risk-free interest rate
of zero, the face value, d’, of this debt can be found from the following
equation: K = d ′ ( p1 − ∆p ) + (1 − p1 + ∆p ) ( 0 ). This means that if G is real-
ized, the new debtholders will receive the face value of the debt; if, how-
ever, B is realized, the firm’s cash flow will be R2 + C2 , which (as follows
from 9.4)) is less than the face value of the senior debt D, leaving the new
creditors with nothing.
The shareholders’ expected payoff without the new investment is 0.
They get nothing even in state G as follows from 9.5). With the new
project, it will be ( p1 − ∆p ) ( R1 + C1 − D − d ′ ) . In some cases it is positive
(see an example below) so, the project can be undertaken even when its
NPV is negative.
Now suppose that the firm uses project financing. In this case, the
debtholders’ payoffs depend only on the returns from the new project
and not on the returns from the assets already in place. Since the NPV is
negative, the firm will not be able to finance the project.
To summarize our analysis of these two cases. Without the ability to
issue non-recourse debt the firm is not able to solve the debt overhang
problem. Secondly, by issuing subordinated debt the firm will be able to
invest in projects with negative NPVs to the expense of the firm’s overall
value. Setting a new project as a new independent project and issuing
non-recourse debt will eliminate the incentive to invest in projects with
negative NPVs. Project financing, or financing with non-recourse debt,
can also help other agency problems related to debt financing. It will also
help to solve free cash flow or empire-building problems (Chap. 7). This
is especially important for large projects with potentially large amounts
 9 Corporate Capital Structure vs. Project Financing 191

of cash. Finally, it can also help increase managerial incentives (Chap. 7)

compared to only equity financing.

Example 9.1
Consider a firm with assets in place, an investment opportunity avail-
able, and outstanding senior debt with a face value of D = 5000. Existing
assets can produce: 8000 in state G (good state) with probability 1/2
and 3000 in state B with probability 1/2. The new project requires an
initial investment of 1500, and the firm will finance this investment by
issuing new debt. A new project’s cost is 1500. The new project generates
2200 in state G and 1000 in state B.
First note that the new project has a positive net present value (NPV)
because 1 / 2 ∗ 1000 + 1 / 2 ∗ 2200 > 1500 . Now suppose that to finance the
new project the firm issues a standard subordinated debt. The face value,
d’, of this debt can be found from the following equation:

1500 = d′ * 1 / 2 + 1 / 2 * 0

Therefore d′ = 3000. The shareholders’ expected payoff without the

new investment is 1 / 2 ∗ ( 8000 − D) = 1500. With the new project, it
will be 1 / 2 ∗ ( 8000 + 2200 − D − d′ ) = 1100. Since the expected payoff
­without the new project is more than that with the project, it will not
be undertaken.
Now suppose that the firm uses project financing (non-recourse debt).
The face value d can be found from:

1500 = d ∗ 1 / 2 + 1 / 2 ∗ 1000

Here d = 2000 . The shareholders’ expected payoff (note that the share-
holders get nothing if B is realized) is: 1 / 2 * ( 8000 − D + 2200 − d ) = 1600.
This is greater than 1500 (the shareholders’ expected payoff without the
new investment), and thus the project will be undertaken and it will be
financed with non-recourse debt.
Now assume that the firm (with the same initial project) has a senior debt
with a face value of D = 10, 000. The new project requires an initial invest-
192 Capital Structure in the Modern World

ment of 100. The new project generates 4000 in state G and 0 in state B but
reduces the probability of G occurring by 30 %. Note that the new proj-
ect has a negative NPV because 0.3 ∗ 3000 − 0.3 ∗ 8000 + 0.2 ∗ 4000 < 100
(see condition 9.6)).
Suppose that to finance the new project the firm issues a standard
subordinated debt. The face value, d’, of this debt can be found from the
following equation:

100 = d ′ * 0.2 + 0.8 * 0 (9.7)

This means that if G is realized, the new debtholders will receive the
face value of the debt; if, however, B is realized, the firm’s cash flow is
3000 + 1000 = 4000 , which is less than the face value of the senior debt,
leaving the new creditors with nothing. From 9.7), d ′ = 500 .
The shareholders’ expected payoff without the new investment is 0.
With the new project, it will be 0.2 ∗ (12, 000 − 10, 000 − 500 ) = 300 . Since
the expected payoff with the new project is more than that without the
project, it will be undertaken.
Now suppose that the firm uses project financing (non-recourse debt).
In this case, the debtholders’ payoffs depend only on the returns from the
new project and not on the returns from the assets already in place. Since
the NPV is negative, the firm will not able to finance the project.
John and John (1991) analyze project financing in the context of the
underinvestment problem in the spirit of Myers (1977). They also take
taxes into consideration. It is shown that a project financing arrange-
ment, where the debt is optimally allocated to the sponsor firm and the
new venture, increases the project’s value by reducing agency costs and
increasing the value of the tax shields (compared to the case of straight
debt financing).
There are several empirical papers that look for the existence of proj-
ect financing in capital structure and how it has worked to mitigate the
moral hazard problem.
For instance, Kensinger and Martin (1988) noted that because of
the number of large capital expenditures businesses are making, moral
 9 Corporate Capital Structure vs. Project Financing 193

hazard dilemmas are of paramount concern. Project financing helps

reduce agency costs, particularly by reducing underinvestment incentives.
Kensinger and Martin (1988) also noted that project finance could play
a role in a company’s restructuring because it pushes companies toward
greater specialization and decentralization. Similarly, project finance has
the power to act as a catalyst for a change in the governance of business
activities because it allows companies to segregate cash flows from newly
adopted projects; this protects both the sponsoring firm and the manag-
ers from defaulting on it.
Esty (2003) observes that in the United States, the majority of proj-
ect financed ventures can be seen in the natural resource and infrastruc-
tural development sectors. Project financing is being used to fund power
plants, toll roads, mines, pipelines, telecommunications systems and
developments in the oil and gas sectors. These large investment projects
with tangible assets are very susceptible to agency conflicts because of
the number of people who are involved. Project companies help deter
opportunistic behavior.
Brealey, Cooper, and Habib (1996) also credit the growing impor-
tance of project financing in infrastructure to its ability to mitigate some
agency problems. Contractual and financing arrangements that project
financing can offer are very important; their importance is magnified in
infrastructure projects because of their scope, duration, and implications
for so many people.
Hainz and Kleimeier (2006) develop a double moral hazard (an effort
by a bank and an effort by a firm) model in which the bank’s incentive
to mitigate risk is highest with a non-recourse project finance loan, while
full-recourse loans are best for the firm. It is suggested that the use of
project finance increases with both the political risk of the country in
which the project is located and the lender’s influence over this political
risk exposure. Furthermore, the use of project finance should decrease
as the economic health and corporate governance provisions of the bor-
rower’s home country improve.
194 Capital Structure in the Modern World

9.3 Asymmetric Information Models

This section demonstrates that project financing can help solve asym-
metric information problems in finance. As Chap. 3 argued, asymmet-
ric information between a firm’s insiders and outsiders can create several
problems for the firm. Good quality firms may not be able to efficiently
signal the quality of their projects and thus their stocks may become
undervalued. Also, asymmetric information problems are perhaps the
reason that firms refuse to invest in projects with positive NPVs.
We first demonstrate that project financing can help high quality firms
signal their quality. Brennan and Kraus (1987) show that an efficient
separating equilibrium in a financing game, where firms have private
information about their future cash flows, is impossible if cash flows are
ordered by the first-order dominance condition. Therefore under every
type of equilibria, issuing securities will be costly.
To illustrate the Brennan and Kraus (1987) idea, consider a company
that is raising funds for an investment project. The investment cost is
B. There are two types of firms. A type j firm’s earnings are denoted Rj,
j = 1, 2 . R1 is distributed according to the distribution function F and
the density function f and R2 is distributed according to the distribution
function G and the density function g. A firm’s type is not known to the
public.
The timing of events in this situation is as shown in Fig. 9.2.
Suppose type 1’s earnings first-order dominate type 2’s earnings.
Suppose type j issues a security that can be described as sj(R), i.e. when
the firm’s earnings are R, the security holder will get s(R). An efficient
signaling equilibrium is when every type of firm receives their first-best
value in equilibrium and no firm has any incentive to deviate from its
equilibrium strategy or mimic another type.

Proposition 9.3 If Project 1’s cash flow distribution first-order dominates

Project 2’s cash flows and sj(R) is non-decreasing, an efficient signaling equi-
librium does not exist.

The shareholders’ payoff is R − s ( R ) when a firm’s earnings equal R.

Suppose that an efficient signaling equilibrium exists. In this case each
 9 Corporate Capital Structure vs. Project Financing 195

A irm's type
Earnings are
is detrmined. Securities
realized and
The irm are sold to
distributed
chooses its outside
to
capital investors
claimholders
structure

Fig. 9.2 Sequence of events

type gets their first-best value (expected earnings from the project minus
the investment cost). So for type 1 it is ER1 − B and for type 2 it is ER2 − B .
Consider a type 2 firm mimicking a type 1. The difference between the
shareholders’ expected payoff in this case and their equilibrium payoff is

∞ ∞ ∞
∆T =
−∞
∫ ( R − s ( R ) ) g ( R ) dR − ∫ ( R − s ( R ) ) g ( R ) dR = ∫ ( s ( R ) − s ( R ) )
1
−∞
2
−∞
2 1

g ( R ) dP

∞ ∞
We also know that
∫ s1 ( R ) f ( R ) dR = B and
−∞
∫ s ( R ) g ( R ) dR = B.
−∞
2

∞ ∞
We will show that ∆T > 0 or that ∫ s1 ( R ) g ( R ) dR < ∫ s ( R ) f ( R ) dR = B.
1
−∞ −∞
∞

Consider ∫ s ( R ) ( f ( R ) − g ( R ) ) dR.
−∞
1

Let v = s1 ( R ) and du = ( f ( R ) − g ( R ) ) dR . Then

∞
∆T = s1 ( R ) ( G ( R ) − F ( R ) ) |∞−∞ − ∫ s1′ ( R ) ( F ( R ) − G ( R ) ) dR
−∞
∞

= ∫ s1 ( R ) ( G ( R ) − F ( R ) ) dR
′

−∞

This is positive since s1′ ( R ) ≥ 0 and G ( R ) ≥ F ( R ) for any R.

Brennan and Kraus (1987: 1227) assume that a firm’s cash flow is indi-
visible. Below we show that when cash flow is divided into two parts (two
196 Capital Structure in the Modern World

projects) and firms are allowed to issue securities with the projects’ con-
tingent payoffs rather than only with total cash flows contingent payoffs,
a separation may exist even if the firm’s total cash flows are ordered by
the first-order dominance condition. Possible interpretations are project
financing or financing with non-recourse debt, issuing asset-backed secu-
rities, and firms’ spin-offs. In all of these cases the firm creates securities
that represent claims on specific projects (assets).
To illustrate the idea, consider the following example. A firm has two
projects available, k = 1, 2. The return of project k is Rk. The project’s suc-
cess depends on the firm’s type j, j = h, l. Rk equals 1 with probability θkj
and equals 0 otherwise.
As in Brennan and Kraus, Rh first-order dominates Rl, which means
that the probability that Rh = 0 is less than the probability of Rl = 0 and
that the probability that Rh equals 0 or 1 is less than that of Rl. It implies:

θ h1θ h 2 > θl1θl 2 (9.8)

( h1 ) ( − θh2 ) < (1 − θl1 ) (1 − θl 2 )
1 − θ 1 (9.9)

The contracts are as follows: the creditors of Project k get dk if Project

k is successful and 0 otherwise.
The investor budget constraint for type j and Project k is: d jk kj = bk .
Hence,

d jk = bk / θ kj (9.10)

The non-deviation condition for l (i.e. the condition for which type l
will choose not to mimic type h) can be written as

(1 − dh1 )θl1 + (1 − dh2 )θl 2 ≤ (1 − dl1 )θl1 + (1 − dl 2 )θl 2

Taking into account (9.10) it becomes:

(1 − b1 / θ1h )θl1 + (1 − b2 / θ2h )θl 2 ≤ (1 − b1 / θ1l )θl1 + (1 − b2 / θ2l )θl 2

 9 Corporate Capital Structure vs. Project Financing 197

After simplifactions we get:

 θ   θ 
b1  1 − l1  ≤ b2  1 − l 2  (9.11)
 θ h1   θh2 

If this condition is satisfied, a separating equilibrium may exist where

the high profit type signals its quality by using project financing.
Obviously if θh1 > θl1 and θl 2 < θh 2, condition (9.11) fails. However, if
θ h1 > θl1 and θl 2 > θ h 2 or θ h1 < θl1 and θl 2 < θ h 2, it may work. Note that the
latter does not necessarily contradict (9.8) and (9.9).

Example 9.2
Suppose that θh1 = 0.7, θh2 = 0.3, θl1 = 0.2, θl 2 = 0.6 and b=
1 b=1 0.2. We
can check that Rh first-order dominates Rl because conditions (9.8) and
(9.9) hold. From (9.10): dh1 = 2 / 7, dh2 = 2 / 3, dl1 = 1 and dl 2 = 1 / 3. The
non-deviation condition for l (i.e., the condition for which type l will
choose not to mimic type h) can be written as

(1 − 2 / 7) * 0.2 + (1 − 2 / 3) * 0.6 ≤ (1 − 1) * 0.2 + (1 − 1 / 3) * 0.6

After simplifications we get 5 / 35 ≤ 6 / 30 which holds.
In fact, condition (9.11) suggests a large variety of possible strategies
for type h to signal its type when it performs better than l in only one
project. See Miglo (2010) for more analysis.
Many consider Myers and Majluf (1984) to have pioneered asymmet-
ric information models. According to Myers and Majluf (1984), under-
investment occurs because of asymmetric information about the value of
assets-in-place and the consideration of the investment project. When
insiders know more about the true value of a firm than outside investors,
potential creditors must rely on market signals to assess a firm’s value.
Therefore, if investors know less than insiders about the true value of the
firm’s assets in place, equity may be mispriced.
In their model, Myers and Majluf (1984) show how outside investors
will value a firm’s shares as the average of two (or more) possible values.
198 Capital Structure in the Modern World

There is a return distribution for these shares that is common knowl-

edge to all potential investors. Because of this averaging, “bad” firms will
have overpriced shares and “good” firms will have underpriced shares.
Therefore, in equilibrium, only bad firms will choose to issue equity.
Creditors are assumed to be rational players and realize this. Asymmetric
information, in this instance, has a negative implication for the good type
of firm because it will prefer not to invest in positive net present value
projects rather than be subject to underpricing their shares.
We now proceed with a quick example of their model (see Chap. 3 for
more details). Suppose there are two firms with an investment opportu-
nity available. The project costs $60 and will bring earnings of $70. The
value of their assets in place is initially either $150 or $40. Management
will know the true state of the firm, but the outside market will not. They
will only know that there is a 50 % chance of either. We can show (anal-
ogously to Chap. 3) that an equilibrium where both firms issue shares
and undertake the new project does not exist. Only the low value firm
will invest because the payoff from investing is higher than from not
investing. For this firm, the shares are overpriced and management will
obviously choose to invest in the growth opportunity. The high value
firm, on the other hand, will choose to not invest.
Indeed suppose that an equilibrium exists where both firms issue
equity to finance the new project. Let α be the fraction of equity sold
to the new investor. The new investor’s expected earnings should cover
the investment cost: α * ( 70 + 0.5 * ( 40 + 150 ) ) = 60 . This implies that
α = 12 / 31. The earnings of the high value firm shareholders then equal:
26
(1 − α ) * ( 70 + 150 ) = 134 . Since this is smaller than 150, the sharehold-
31
ers will not be interested in issuing shares for the new project.
This exemplifies the adverse selection problem. Because of asymmetric
information, firms with higher values will pass on positive return proj-
ects. The result holds if, instead of issuing equity, firms were allowed to
issue standard corporate debt.
The underinvestment problem could be avoided if the firm were to
have the ability to issue riskless debt (this assumption is often unrealistic).
While Myers and Majluf (1984) do not actually study the implications
 9 Corporate Capital Structure vs. Project Financing 199

of project financing, they do “hint” that the underinvestment problem

can be resolved by a spin-off of its assets-in-place into a separately incor-
porated company.5 In the example above, both types of firms may issue
non-recourse debt with face value 60, which would solve the underinvest-
ment problem. The new investor will finance the new project with non-­
recourse debt. The face value of debt will be 60. Earnings from the new
project (which equal 70) will cover the amount of debt. Furthermore, we
can show that in some cases the same insight holds if we consider asym-
metric information about the amount of earnings from the new invest-
ment and not about the value of assets-in-place as in the above example.
An interesting question is why projects that are financed with non-­
recourse debt often have a high degree of risk and why the amount of
investment is usually very large. Miglo (2010) considers different sce-
narios where firms choose between non-recourse debt and conventional
debt. To illustrate the idea, consider a model similar to the one described
above (where θkj is the probability of success of project j for a type k firm)
but now assume that the firm will use corporate financing (regular debt
with recourse with face value D) for the first project and non-recourse
debt for the second project (face value d). If the first project is successful,
the investors are paid in full. If it fails and the second project succeeds,
two situations are possible. If 1 − D ≥ d the investors of project 1 are paid
in full, otherwise they get 1 − d . For the second project, investors are paid
in full if the second project is successful and get nothing otherwise.
The promised payment for project 2 is d = b2 / θh 2 . Let Pk(x) be the
probability that the total cash flow of firm k equals x. For example, we
have Ph ( 0 ) = (1 − θh1 ) (1 − θh 2 ).
b1 b b1
If 1 ≥ + 2 then D = . In this case the investors (firm
1 − Ph ( 0 ) θ h 2 1 − Ph ( 0 )
h) of both projects are paid in full if the firm’s earnings equal either 1 or 2
(but not 0). If l mimics h, its payoff is: Pl (2 ) (2 − D − d ) + Pl (1) (1 − D − d ) .
This should not be greater than it’s equilibrium value θl1 + θl1 − b1 − b2 .
The non-deviation constraint for l is then:

5
Myers and Majluf (1984: 22).
200 Capital Structure in the Modern World

 1 − Pl ( 0 )   θ 
b1  1 − ≤ b 1 − l2 (9.12)
 1 − P ( 0 )  2  θ 
 h   h2 

If we consider the case when θh1 < θl1 and θl 2 < θh 2 then condition (9.12)
can be interpreted as follows. The likelihood that this condition is satis-
fied (and successful utilization of project finance by firm h) increases with
an increase in the amount of investment in project 2 (b2) maintaining that
b1 is sufficiently small. Secondly, if θl2 is sufficiently smaller than θh2 and/
or if 1 − Pl (0 ) is sufficiently close to 1 − Ph (0 ) . This means that a separat-
ing equilibrium exists when the asymmetry regarding the firm’s overall
quality (probability of default in both projects) is sufficiently small while
the asymmetry regarding the project, for which the firm issues project-­
contingent securities, is large. A similar condition can be obtained for the
b1 b
case 1 < + 2.
1 − Ph ( 0 ) θ h 2

Example 9.3
Suppose that θh1 = 0.3, θh2 = 0.6, θl1 = 0.6, θl 2 = 0.2, b1 = 0.1 and b2 = 0.4 .
We can check that Rh first-order dominates Rl because conditions (9.8)
and (9.9) above hold.
As follows from our previous analysis: d = b2 / θh 2 . So d = 2 / 3 . Also
b1 b1
D= = = 5 / 36. The non-deviation condition
1 − Ph ( 0 ) 1 − (1 − θ h1 ) (1 − θ h 2 )
for l (i.e. the condition for which type l will choose not to mimic type h)
can be written as (1 − 2 / 3 − 5 / 36 ) ∗ 0.6 + (1 − 2 / 3) ∗ 0.2 ≤ 0.2 . After sim-
plifications we get 11 / 60 ≤ 0.2 which holds.
We can also see that h cannot do it the other way around, i.e. use
project financing for project 1. We have: d = b1 / θh1. So d = 1 / 3 and
b2
D= = 5 / 9. The non-deviation condition for l (i.e. the condi-
1 − Ph ( 0 )
tion for which type l will choose not to mimic type h) can be written as

(1 − 1 / 3 − 5 / 9) ∗ 0.2 + (1 − 2 / 3) ∗ 0.6 ≤ 0.2 . After simplifications we get

2 / 9 ≤ 0.2 , which does not hold.
 9 Corporate Capital Structure vs. Project Financing 201

One interpretation of these results is project financing or financing by

non-recourse debt. Project financing differs from corporate financing in
two ways: (1) the creditors do not have a claim to the profit from other
projects if the project fails while corporate financing gives this right to
the investors and; (2) it typically has priority on the cash flows from the
project over any corporate claims. The case described in the beginning
of Sect. 9.3 can thus be interpreted as financing both projects with non-­
recourse debt. The second case can be seen as financing project 2 with
non-recourse debt and financing project 1 with standard corporate debt
when the creditors’ payoffs depend on the firm’s total earnings (after the
project 2 creditors are paid) and are not attached uniquely to the earnings
from project 1).
From (9.12), the existence of a separating equilibrium is probable
when the amount of investment financed by non-recourse debt is suf-
ficiently large with regard to the corporate investment and the uncer-
tainty regarding the performance of projects financed by non-recourse
debt is greater than that of projects financed by corporate debt. Also,
as follows from (9.12), the uncertainty regarding the performance of
projects financed by non-recourse debt is greater than that of projects
financed by corporate debt. There exists some evidence consistent with
the spirit of the above predictions. Brealey, Cooper, and Habib (1996);
Esty (2003, 2004); McGuinty (1981), and Nevitt (1979) argue that
non-recourse debt is typically used for financing large, capital-inten-
sive, projects and that the leverage ratio of project companies is typi-
cally larger than that of parent companies. Also, this is consistent with
the evidence that project financing is usually used for financing risky
projects (see for instance: Esty (2004); Flybjerg et al. (2002); McGuinty
(1981); Merrow et al. (1988); Miller and Lezard (2000), and Nevitt
(1979)).
Shah and Thakor (1987) analyze optimal financing in the presence
of corporate taxation. In their model, projects have the same mean of
return, the owners have private information about risk, and investors may
acquire (costly) information about the parameters of a firm’s risks. If the
benefits from information production are relatively big, project financing
is optimal because the cost of screening a separately incorporated project
202 Capital Structure in the Modern World

is low. Alternatively, project financing can result in higher leverage and

provide greater tax benefits. This is because, under corporate financing,
leverage is below the optimal level. In the absence of bankruptcy costs,
the first-best financing method is “pure” debt. However, firms reduce
leverage in order to provide a credible signal about risk.
Similarly, the results of this section can be applied to asset-backed
securities (ABS). Suppose that the firm issues ABS to finance the first
project. If the project fails then the creditors (or the holders of ABS)
do not have any legal rights of recourse to the assets of the firm. In
addition, there exists a bankruptcy remoteness condition. If the parent
company fails it cannot use the assets of the project company. Therefore,
formally, this debt is analogous to the case of non-recourse debt issued
for both projects in the model. ABS are now used by many corporations
as a financing method. The standard explanation in existing literature is
that these securities exist primarily for regulatory reasons (for instance,
banks trying to avoid minimal capital requirements). However, recent
empirical literature (Calomaris and Mason 2004) argues that secu-
ritization seems to be motivated more by reasons related to efficient
contracting.
Finally, note that Schipper and Smith (1983) found that 72 out of
93 firms in their sample of spin-offs involved parent companies and
subsidiaries with different industry memberships (cross-industry spin-
offs). The financing strategies discussed here can be interpreted as a
spin-off because they contain financing by non-recourse debt and the
creation of an independent company respectively. The likelihood that
firms in separate industries have divergent performance (see comments
for condition 9.9)) is higher than it is for firms in the same industry.
Thus, Schipper and Smith’s (1983) results are consistent with the spirit
of this chapter.

9.4 Other Models

The paper written by Chemmanur and John (1996) looks at managerial
considerations and the threat of a takeover as incentives for the use of
 9 Corporate Capital Structure vs. Project Financing 203

project financing. They assume that retaining control of a firm provides

the greatest benefit to entrepreneurs and that they will make their capital
structure decisions keeping this in mind. For instance, if a firm’s cur-
rent manager were to issue equity to raise funds, there is a possibility
that the majority of it will be purchased by another market participant;
which jeopardizes the manager’s position as the controller. However, by
issuing only corporate debt, the firm accrues the bankruptcy costs associ-
ated with high debt levels. Therefore, entrepreneurs will always aim to
maximize the present value of their control benefits and the security ben-
efits from lower debt levels. Non-recourse debt will allow the manager
to minimize overall debt levels and reduce bankruptcy costs. Whether
project financing will be preferred to normal debt depends on the value
of the control benefits and their degrees of risk.
In the previous two sections, the models reviewed either discussed
project financing as a solution to agency costs or as a solution to asym-
metric information. Habib and Johnsen (1999), however, combine these
two and look at how non-recourse secured debt, when combined with
ex-ante debt contracts, can solve agency costs in spite of information
asymmetries. More specifically, they examine the role that non-recourse
secured debt plays in ensuring the optimal allocation of ownership over
secured physical assets across various possible states of the world. In this
case, with non-recourse secured debt, if the borrower defaults, the lender/
issuer can seize the collateral, but the lender’s recovery is limited only to
this collateral.
Habib and Johnsen (1999) apply the results of their model to aircraft
financing in order to demonstrate the benefits of non-recourse debt over
those of ex-post bargaining. They choose to look at the aircraft financing
industry because it is made up of a large variety of financing contracts
and, more importantly, the industry is subject to exogenous shocks such
as wars or industry deregulation and these affect the state of the world.
This makes the need for a redeployer plausible because the optimal use of
assets may change frequently. Similarly, it creates an investment environ-
ment subject to risk.
The implication of Habib and Johnsen’s (1999) model is that the rede-
ployment value of a firm’s assets should have an important influence on
204 Capital Structure in the Modern World

its optimal capital structure. This helps explain why some firms have vir-
tually no debt while others have quite a lot. While there are many other
considerations behind a firm’s capital structure choice, this model helps
shed more light on the subject. It shows how the structure of project
financing contracts can help reduce agency costs and moral hazard dilem-
mas that plague firms even when there are informational asymmetries
between entrepreneurs and creditors.
As mentioned, project financing uses a complex series of contracts to
“distribute the different risks presented by a project among the various
parties involved in the project.”6 They are created to dictate the manage-
ment’s response to any number of issues. The large number of contracts
clearly reduces agency costs because there is less worry that the manag-
ers will not be acting in the best interests of the firm as a whole. With
corporate debt, there are a number of problems with creating contracts
to specify actions of shareholders and managers in response to market
stimuli. As Brealey, Cooper, and Habib (1996) note, there are simply too
many circumstances to consider.
It is very important that these contracts anticipate what may happen
and provide a solution. Some of the risks that a project finance company
may be subjected to include technological risk, construction risk, com-
pletion risks, inflation risk, environmental risk, regulatory risk, and polit-
ical and country risk. For example, the telecom sector showed a decline
in 2006 likely due to the technological acceleration and development in
the sector.7 The contracts are necessary to properly allocate the risks to
the counterparty best able to control and manage them. This is another
one of the primary reasons why firm managers will choose to separately
incorporate a new investment with project financing.
Marty and Voisins (2007) noted that since the shareholders may be
likely to sell off their stake well before the expiry date of the concession,
a professional evaluation of the project financed companies at different
stages of the project’s life seems increasingly important.
Kleimeier and Versteeg (2010) argue that project finance is designed
to reduce transaction costs arising from a lack of information on possible

6
Brealey, Cooper and Habib (1996: 28).
7
Gatti (2008).
 9 Corporate Capital Structure vs. Project Financing 205

investments and capital allocation, insufficient monitoring and enforce-

ment of corporate governance, risk management, and the inability to
mobilize and pool savings. Their empirical analysis of 90 countries, from
1991 to 2005, confirms that project finance is a strong driver of eco-
nomic growth in low income countries with high transaction costs.
Gatti et al. (2013) study negotiations of financial packages between
sponsors and lenders and their costs. It was found that lenders indeed rely
on the network of contracts as a mechanism to control agency costs and
project risks. This network of nonfinancial contracts limits the manage-
rial discretion of project sponsors, to make cash flows better verifiable
for lenders, and to reduce the negative impact of unexpected events on
project cash flows. It was also found that lenders are reluctant to reduce
the price of the credit if sponsors are involved as project counterparties in
the relevant contracts.
In Lyonett du Moutier (2010), a positive agency theory model is built
through a detailed analysis of property rights exchanged in the trans-
actions needed to build the Eiffel Tower. The analysis of the original
financial documents and contracts for the Tower reveals how specific
provisions in the contracts combined to reduce agency costs all around.
Thus a model derived from positive agency theory may help us better
understand modern project finance.
Hainz and Kleimeier (2012) analyze the design of loan contracts for
financing projects in countries with high political risk. They argue that
non-recourse project finance loans and the participation of development
banks in the loan syndicate help mitigate political risk. Their results also
show that if political risk is higher, then project finance loans are more
likely to be used, and development banks are more likely to participate
in the syndicate.

Questions and Exercises

1. Non-recourse debt can be used to solve a debt overhang problem.
2. The main difference between non-recourse debt and secured debt is
that the riskiness of secured debt is firm-specific while the riskiness of
non-recourse debt is project-specific.
206 Capital Structure in the Modern World

3. Consider a firm with assets in place, an investment opportunity avail-

able, and outstanding senior debt with a face value D = 8000 . Existing
assets can produce: 20,000 in state G (good state) with probability
1/2, 2000 in state B with probability 1/2. The new project requires an
investment of 3000, and the firm will finance this investment by issu-
ing new debt. The new project generates 5000 in state G and 2000 in
state B.

(a) Find the NPV of new project.

(b) Suppose that to finance the new project the firm issues a standard
subordinated debt. Will the new project be undertaken?
(c) Now suppose that the firm uses project financing (non-recourse
debt). Same question.
(d) Now assume that the firm has a senior debt with a face value
D = 15, 000 . The new project requires an initial investment of
1000. The new project generates 4000 in state G and 0 in state B
and it also reduces the probability of G occurring by 10 %.
Find the NPV of the new project.
(e) Suppose that to finance the new project the firm issues a standard
subordinated debt. Will the new project be undertaken?
(f ) Now suppose that the firm uses project financing (non-recourse
debt). Same question.

4. The economy consists of two different types of firms. Each firm has
two projects: they produce 1 if they are successful and 0 otherwise.
The probability of success for firm j, j = 1, 2 and project k, k = 1, 2 is
given by θjk. We have θ11 = 0.75, θ12 = 0.25, θ22 = 0.6, θ 21 = 0.25. Both
projects require an investment 0.2. Outside investors don’t know the
type of the firm (there is 50 % of each type of firms in this economy)
but they can observe the interest rate for each contract signed by each
firm.

(a) Discuss if a separating equilibrium is possible where type 1 uses

corporate debt (one contract).
(b) Analyze the situation where both firms use project financing for
each project.
 9 Corporate Capital Structure vs. Project Financing 207

References
Amlôt, M. (2015, December 16). Tabreed finalises AED 192.5m loan facility
for district cooling plant for Dubai Parks and Resorts. CPI financials. http://
www.cpifinancial.net/news/post/34011/tabreed-finalises-aed-192-5m-loan-
facility-for-district-cooling-plant-for-dubai-parks-and-Resorts
Berkovitch, E., & Kim, E. (1990). Financial contracting and leverage induced
over and under investment incentives. The Journal of Finance, 3, 765–794.
Bonetti, V., Caselli, S., & Gatti, S. (2010). Offtaking agreements and how they
impact the cost of funding for project deals. A clinical case study of the
Quezon Power Ltd Co. Review of Financial Economics, 19(2), 60–71.
Brealey, R., Cooper, I., & Habib, M. (1996). Using project finance to fund
infrastructure investments. Journal of Applies Corporate Finance, 9(3),
25–38.
Brennan, M., & Kraus, A. (1987). Efficient financing under information asym-
metry. Journal of Finance, 42(5), 1225–1243.
Calomiris, C., & Mason, J. (2004). Credit card securitization and regulatory
arbitrage. Journal of Financial Services Research, 26, 5–27.
Chemmanur, T., & John, K. (1996). Optimal incorporation, structure of debt
contacts and limited-recourse project financing. Journal of Financial
Intermediation, 5, 372–408.
Corielli, F., Gatti, S., & Steffanoni, A. (2010). Risk shifting through nonfinan-
cial contracts. Effects on loan spreads and capital structure of project finance
deals. Journal of Money, Credit and Banking, 42(7), 1295–1320.
Esty, B. (2003). The economic motivations for using project finance. Working
paper. Harvard Business School.
Esty, B. (2004). Why study large projects? An introduction to research on proj-
ect finance. European Financial Management, 10, 213–224.
Flyvbjerg, B., Holm, M., & Buhl, S. (2002). Underestimating costs in pub-
lic works projects. Journal of the American Planning Association, 68,
279–295.
Gardner, D., & Wright, J. (2012). Project finance. In T. Rhodes (Ed.),
Encyclopedia of debt finance (2nd ed.). London: Euromoney Learning
Solution. Ch. 13.
Gatti, S. (2008). Project finance in theory and practice. London: Academic.
Gatti, S., Kleimeier, S., Megginson, W. L., & Steffanoni, A. (2013). Arranger
certification in project finance. Financial Management, 42(1), 1–40.
208 Capital Structure in the Modern World

Habib, M. A., & Johnsen, D. B. (1999). The financing and redeployment of

specific assets. The Journal of Finance, 2, 693–720.
Hainz, C., & Kleimeier, S. (2006). Project finance as a risk management tool in
international syndicated lending. Discussion paper no. 183. GESY.
Hainz, C., & Kleimeier, S. (2012). Political risk, project finance, and the partici-
pation of development banks in syndicated lending. Journal of Financial
Intermediation, 21(2), 287–314.
John, K., & John, T. (1991). Optimality of project financing: Theory and
empirical implications in finance and accounting. Review of Quantitative
Finance and Accounting, 1, 51–74.
Kensinger, J., & Martin, J. (1988). Project financing: Raising money the old
fashioned way. Journal of Applied Corporate Finance, 3, 69–81.
Kleimeier, S., & Megginson, W. (2001). An empirical analysis of limited
recourse project finance. Working paper. University of Oklahoma.
Kleimeier, S., & Versteeg, R. (2010). Project finance as a driver of economic
growth in low-income countries. Review of Financial Economics, 19(2),
49–59.
Lyonnet du Moutier, M. (2010). Financing the Eiffel Tower: Project finance and
agency theory. Journal of Applied Finance, 20(1), 127–141.
Marty, F., & Voisins, A. (2007). The role of financial intermediaries in UK
Private Finance Initiatives. Working paper.
McGuinty, T. (1981). Project organization and finance. Cleveland: The Cleveland-­
Cliffs Iron Company.
Merrow, E., McDonnell, L., & Arguden, R. (1988). Understanding the out-
comes of megaprojects: A quantitative analysis of very large Civilian Projects.
RAND Corporation publication series #R-3560-PSSP.
Miglo, A. (2010). Project financing versus corporate financing under asym-
metric information. Journal of Business and Economics Research, 8(8),
27–43.
Miller, R., & Lessard, D. (2000). The strategic management of large engineering
projects. Cambridge, MA: MIT Press.
Myers, S. C. (1977). Determinants of corporate borrowing. Journal of Financial
Economics, 5, 147–175.
Myers, S., & Majluf, N. (1984). Corporate financing and investment decisions
when firms have information that investors do not have. Journal of Financial
Economics, 13, 187–221.
Nevitt, P. (1979). Project financing. London: Euromoney Publications.
 9 Corporate Capital Structure vs. Project Financing 209

Schipper, K., & Smith, A. (1983). Effects of recontracting on shareholder

wealth: The case of voluntary spinoffs. Journal of Financial Economics, 12,
437–467.
Shah, S., & Thakor, A. (1987). Optimal capital structure and project financing.
Journal of Economic Theory, 42, 209–243.