Optimal Capital Structure Model for BOT Power Projects

in Turkey
Sandalkhan Bakatjan, A.M.ASCE1; Metin Arikan2; and Robert L. K. Tiong, M.ASCE3

Abstract: Interest in the Build/Operate/Transfer 共BOT兲 scheme for infrastructure projects has been growing rapidly, and numerous
projects have been implemented around the world. Through BOT projects, a government reallocates the risks and rewards in the
Downloaded from ascelibrary.org by New York University on 05/11/15. Copyright ASCE. For personal use only; all rights reserved.

development of large infrastructure projects to the private sector. One key aspect to the successful implementation of the BOT concept in
any country is the raising of finance by project sponsors. Financial engineering techniques and capital structuring skills are required to find
the proper mix of debt and equity and to achieve successful financing for the proposed project. The objective of this paper is to present
a simplified model to determine the optimum equity level for decisionmakers at the evaluation stage of a BOT hydroelectric power plant
共HEPP兲 project in Turkey, which takes place immediately after the completion of the feasibility study. The resulting model is the
combination of a financial model and a linear programming model that incorporates an objective of maximizing the return of the project
from the equity holder’s point of view. To show versatility of the model, a real case study is conducted. Thus, this research is concerned
with the determination of an equity funding level in BOT project finance. There are different equity levels found in BOT HEPP projects,
and there is a need for such a model to determine optimal capital structure, which would assist the project sponsors to ensure that the
equity level necessary for optimal capital structure is available prior to the project implementation stage.
DOI: 10.1061/共ASCE兲0733-9364共2003兲129:1共89兲
CE Database keywords: Build/Operate/Transfer; Financial management; Linear programming; Project management; Turkey; Models.

Introduction is required in making a project viable under the constraints of

financial reality.
In a Build/Operate/Transfer 共BOT兲 project, the project sponsor, In this context, a prototype program, OPTIMUM, for decision-
often a contractor, is responsible for, among other things, the makers has been developed by using macros, built-in functions,
raising of finance that is necessary for implementing the project. and Visual Basic for Application 共VBA兲 modules of Microsoft
This exercise involves quantitative economic analysis, and the EXCEL’97 共Bakatjan 2000兲. The Solver within EXCEL contains
sponsor’s financial adviser will prepare the financial model and a program to solve the linear programming models. This devel-
cash flow analysis involving the projected revenue streams, capi- opment enhances the usefulness of the model by providing a user
friendly template for data entry. The introduction and extensive
tal expenditures, and financial charges, etc.
application of spreadsheet software are ‘‘great boons to financial
During the project evaluation stage, after the completion of the
analysts and financial engineering’’ as stated by Marshall and
feasibility study, rigorous financial analysis needs to be con-
Bansal 共1992兲. It is worthy to note that the program evaluates a
ducted. It is important to determine the optimal capital structure, project from the equity holder’s 共often, contractor兲 point of view
which is important for the successful raising of finance. If the because Jenkins and Marchesini 共1999兲 have shown that when
active project sponsor is not capable of injecting the necessary analysis is undertaken from different perspectives, the results of
equity to achieve the optimal debt-to-equity ratio, then they an appraisal of a power project differ.
should search for additional passive investors until the equity The motivation for the model development stemmed from the
level is enough to provide the optimal capital structure 共Dias and realization that the determination of equity funding levels in BOT
Ioannau 1995兲, or other financing instruments should be imple- project finance is still not well researched. In addition, there are
mented such as mezzanine finance. Hence, financial engineering different equity levels implemented in BOT hydroelectric power
plant 共HEPP兲 projects. There is the need for such a model to
1 determine the optimal capital structure to assist the project spon-
Proposal Manager, ATLAS Construction Inc. Co., Cukuranbar Mah. l
Cad. 61. Sok. No:36 Ankara, Turkey. sor in the project evaluation stage in determining whether the
2
Assistant Professor, Dept. of Civil Engineering, Middle East equity level necessary for optimal capital structure would be
Technical Univ., Ankara 06531, Turkey. available or not.
3
Associate Professor, School of Civil and Environmental Engineering,
Nanyang Technological Univ., Singapore.
Note. Discussion open until July 1, 2003. Separate discussions must Formulation of Financial Model
be submitted for individual papers. To extend the closing date by one
A financial model is a mathematical expression of relationships
month, a written request must be filed with the ASCE Managing Editor.
The manuscript for this paper was submitted for review and possible among financial components. It is used to support decision mak-
publication on February 27, 2001; approved on February 20, 2002. This ing in project evaluation. The project viability is analyzed from
paper is part of the Journal of Construction Engineering and Manage- the equity holders’ perspective in the project.
ment, Vol. 129, No. 1, February 1, 2003. ©ASCE, ISSN 0733-9364/ The first step in any investment evaluation is to gather the
2003/1-89–97/$18.00. appropriate information on the project costs and calculate the cash

JOURNAL OF CONSTRUCTION ENGINEERING AND MANAGEMENT / JANUARY/FEBRUARY 2003 / 89

J. Constr. Eng. Manage. 2003.129:89-97.

 兺冋 册
flows generated by that project. In the simplest terms, the cash c j
flow is the difference between the money coming in and the EDC⫽
j⫽1
Aj 兿
k⫽0
共 1⫹␪ k 兲 ⫺A j for j⫽1,2, . . . ,c (3)
money going out of the investment project. ‘‘Every investment
opportunity can be fully described by the cash flow that it gener-
where A j ⫽cash flow for jth year of construction; c⫽the con-
ates’’ 共Marshall and Bansal 1992兲. The certainty of cash flow
struction duration in a year; ␪ o ⫽0; and ␪ k ⫽the escalation rate
determines risk associated with the investment opportunity.
In order to calculate the value of a project over a number of for the kth year, the time period. 共According to Articles 5 and 12
years, after the estimation of the cash flows, one needs to take of the concession agreement in BOT hydropower projects in Tur-
into account the time value of money. It is a basic principle of key, escalation rates are equal to the United States Consumer
finance that money has time value. Price All Item Index 共CPI兲 change rate, because all prices are in
U.S. dollars. Hence, the forecast escalation is taken as ␪ k
⫽4.1%).

冋 册
Assumptions and Theoretical Framework
c j j

兺 兿 兿
Downloaded from ascelibrary.org by New York University on 05/11/15. Copyright ASCE. For personal use only; all rights reserved.

The following are the assumptions for the model: IDC⫽ 共 1⫺e 兲 A j 共 1⫹r 兲 G⫺ j⫹1
共 1⫹␪ k 兲 ⫺A j 共 1⫹␪ k 兲
j⫽1 k⫽0 k⫽0
1. The financing of a project is raised by a combination of
equity and debt. The availability of funds is assumed to be for j⫽1,2, . . . ,c (4)
unlimited, because there is no shortage of funds to raise debt
or equity in the power sector, instead there is a lack of bank- where e⫽the equity fraction of the current value cost; G⫽the
able projects 共Malhotra 1997兲. In addition, the project in grace period of debt, which is equal to c; and r⫽the interest rate
general is a simple and pure investment with a single internal of the loan. Hence,
rate of return 共IRR兲. In other words, the net cash flow
changes its sign only once. Thus, the net cash flow during c j

the construction period is negative and positive during the TPC⫽e 兿 共 1⫹␪ k 兲 ⫹ 共 1⫺e 兲
兺 k⫽0
j⫽1
Aj
operation period.

冋兺 册
2. A loan is available from one source or from multiple sources c j
with the same term of annual equal instalments. Because
revenues peak quickly for power projects, it is common to
⫻
j⫽1
A j 共 1⫹r 兲 c⫺ j⫹1
兿
k⫽0
共 1⫹␪ k 兲
employ this form of repayment 共Walker and Smith 1995兲.
Moreover, it is assumed that upfront and commitment fees, for j⫽1,2, . . . ,c (5)
which are usually 0.5–1.5% of the loan 共Wynant 1980兲, are
included in the committed loan amount for the sake of sim- It is a common practice that equity drawings during construc-
plicity. tion are calculated as a portion of TPC 共Bakatjan 2000兲. In other
3. There is a grace period of the loan which is equal to the words, the financing cost 共IDC and EDC兲 as well as BC of a
construction duration. Generally, the grace period is equal to project are shared between investors and lenders. Thus

冋 册
the construction duration because of the nonrecourse or the
j j
limited recourse financing nature; debt repayment depends
only on the project’s revenue. E i ⫽e eA j 兿
k⫽0
共 1⫹␪ k 兲 ⫹ 共 1⫺e 兲 A j 共 1⫹r 兲 c⫺ j⫹1 兿
k⫽0
共 1⫹␪ k 兲
4. Land expropriation cost can be included in the Base Cost
共BC兲 of the project as an additional cost.
for j⫽1,2, . . . ,c (6)
5. The cash flows during construction are preestimated.
6. There are no value added tax 共VAT兲, corporate, and income where E j ⫽equity drawing in ith year of construction.
taxes. The only applicable tax is witholding tax of 11% 共in- After the completion of construction, revenue is generated
cluding surcharge兲, which is the most common tax in inter- from electricity sales during the operation period, m, which is
national lending.
fixed as 20 years for BOT HEPPs. 共This is stated in Article 14,
7. The unit prices of electricity are a declining function during
Tender Specification for BOT HEPP Projects by the Ministry of
the loan repayment period and a constant value after the loan
maturity. Energy and Natural Resources.兲 The net annual cash available in
8. Complete depreciation of the Total Project Cost 共TPC兲 is current value given by NCAi , can be estimated as
allowed during the operation period.
NCAi ⫽PBITi ⫺TAXi ⫹DEPi ⫺D i for i⫽1,2, . . . ,m (7)
Theoretical Framework where PBITi ⫽Profit Before Interest and Tax; TAXi ⫽Tax; DEPi
Ranasinghe 共1996兲 has developed a simplified model to calculate ⫽Depreciation; and D i ⫽Annual Debt Installment for ith year.
TPC for infrastructure projects in developing countries, which is There is an 11% withholding tax including surcharge appli-
the starting point of the financial model developed in this section. cable for interest. As assumed, there are no income and corporate
TPC⫽BC⫹EDC⫹IDC (1) taxes.
where BC⫽the base cost or constant value cost of the project TAXi ⫽ 共 PBITi ⫺INTi 兲 ⫻0.11 for i⫽1,2, . . . ,m (8a)
estimated at market prices of a predetermined year; EDC⫽the
cost escalation during construction; and IDC⫽the interest during where INTi ⫽interest to be paid in the ith year. The writers as-
construction. sumed annual equal installments of debt, D i ; therefore
c
r 共 1⫹r 兲 N
BC⫽ 兺 Aj
j⫽1
for j⫽1,2, . . . ,c (2) D i ⫽ 共 1⫺e 兲 xTPC⫻
共 1⫹r 兲 N ⫺1
for i⫽1,2, . . . ,m (8b)

90 / JOURNAL OF CONSTRUCTION ENGINEERING AND MANAGEMENT / JANUARY/FEBRUARY 2003

J. Constr. Eng. Manage. 2003.129:89-97.

 DPRi ⫽D i ⫻ 共 1⫹r 兲 ⫺ 共 N⫺i⫹1 兲 for i⫽1,2, . . . ,m (8c) The lowest tariff, U 2 , is defined from the condition of financial
viability of the project; PBITi ⭓0 at the end of the concession
共White et al. 1989兲. period. Thus
DPRi ⫽debt principal for ith year 共payment for the principal debt
at ith year兲, and N⫽debt repayment period. As a rule, N•m. 共 P 20xU 2 ⫺OM20⫺DEP20兲 ⫽0
INTi ⫽D i ⫺DPRi for i⫽1,2, . . . ,m (8d) ⬖ (12e)
Substituting Eq. (8b兲 and (8c兲 into 共8d兲 will result OM20⫹DEP20
U 2⫽
r 共 1⫹r 兲N P 20
INTi ⫽ 共 1⫺e 兲 TPC⫻ ⫻ 关 1⫺ 共 1⫹r 兲 ⫺ 共 N⫺i⫹1 兲 兴
共 1⫹r 兲 N ⫺1 Therefore, the highest tariff, U 1 , will be
for i⫽1, . . . ,m (8e) U a v ⫻m⫺U 2 ⫻ 共 m⫺N 兲
U 1⫽ N for i⫽1,2, . . . ,m (12f)
Substituting Eq. (8e兲 into Eq. (8a兲
兺 0.95共 i⫺1 兲
再
Downloaded from ascelibrary.org by New York University on 05/11/15. Copyright ASCE. For personal use only; all rights reserved.

r 共 1⫹r 兲 N
TAXi ⫽ PBITi ⫺ 共 1⫺e 兲 TPC⫻ i⫽1
共 1⫹r 兲 N ⫺1
Operation and Maintenance cost includes OM material and

冎
⫻ 关 1⫺ 共 1⫹r 兲 ⫺ 共 N⫺i⫹1 兲 兴 ⫻0.11
spare parts cost, personnel salaries, indirect costs, and insurance
cost. Because the civil works portion of TPC largely depends on
geologic and hydrologic condition of the project site, OM cost is
for i⫽1,2, . . . ,m (9) a function of Electromechanical Cost 共EMC兲 of the project. It is
Depreciation is a noncash expense: it only reduces taxable observed that OM is usually 3– 4% of EMC in practice.
income and provides an annual tax advantage equal to the product Substituting Eqs. 共10兲 and 共12a兲 into Eq. 共11兲
of depreciation and the 共marginal兲 tax rate, but it does not lead to TPC
a cash outflow from the company. PBITi ⫽U i P i ⫺OMi ⫺ for i⫽1,2, . . . ,m (13a)
m
The most common method for depreciation is straight-line de-
preciation. Under this method, annual depreciation equals a con- Consequently

冋
stant proportion of the initial investment. In this model, it is as-
TPC
sumed that TPC can be depreciable in its entirety. Thus NCAi ⫽0.89共 U i P i ⫺OMi 兲 ⫹0.11 ⫺ 共 1⫺e 兲 TPC
m
TPC
DEPi ⫽

PBITi ⫽R i ⫺OMi ⫺DEPi

m
for i⫽1,2, . . . ,m

for i⫽1,2, . . . ,m
(10)

(11)
⫻
r 共 1⫹r 兲 N
共 1⫹r 兲 N ⫺1 册
关 0.11⫹0.89共 1⫹r 兲 ⫺ 共 N⫺i⫹1 兲 兴

for i⫽1,2, . . . ,m (13b)

where R i ⫽annual revenue, and OMi ⫽annual operation and
maintenance costs. The viewpoint of equity holders is focused on the main project
metrics, internal rate of return and Net Present Value 共NPV兲. The
R i ⫽U i P i for i⫽1,2, . . . ,m (12a)
IRR and NPV are the most common and fundamental economic
U i ⫽unit price of electricity 共U.S. cents/kW•h, which is the elec- decision criteria employed in practice 共Lohmann 1988兲.
tricity sale price of the project company to TEA, Turkish Elec- The NPV from the equity holder’s point of view, is
tricity Authority兲. P i ⫽net annual energy production 共kW•h兲 de- c m
termined in the feasibility study. 共It is agreed by the parties that Ei NCAi
the payment basis in the Power Purchase Agreement 共PPA兲 by
NPV⫽⫺ 兺
j⫽1
⫹ 兺
共 1⫹d 兲 j⫺1 i⫽1 共 1⫹d 兲 i⫹c
TEA is based on this capacity, without considering whether this
level of electricity generation is achieved or not.兲 for i⫽1,2, . . . ,m for j⫽1,2, . . . ,c (14)
There are two types of unit price: the first is tariff during the where m⫽concession period; c⫽construction duration; E j
loan repayment period, the other is tariff after the loan maturity. ⫽equity drawing in ith year of construction; NCAi ⫽net annual
There is a practice of using different tariffs for different years cash available in jth year of operation; and d⫽the discount rate.
during the loan repayment period. In this model, we assume the To calculate the present value of an investment project, we
following declining function of tariffs during the loan period for discount the expected future cash flows by the rate of return of-
the sake of simplicity in calculations. Hence fered by comparable investment alternatives. This rate of return is
U i ⫽U 1 ⫻0.95共 i⫺1 兲 for i⭐N (12b) often referred to as the discount rate, hurdle rate, or opportunity
cost of capital. It is the return forgone by investing in the project
U i ⫽U 2 for i⬎N for i⫽1,2, . . . ,m (12c) rather than investing in securities. The selection of the discount
where U 1 ⫽the highest electricity tariff for the first year of con- rate is one of the crucial aspects of engineering economic analysis
cession; and U 2 ⫽the lowest electricity tariff for the last year of 共Park and Sharpe-Bette 1990兲. Marshall and Bansal 共1992兲 de-
concession. fined the discount rate as the opportunity cost of money to the
party considering the investment. Hence, from the equity holder’s
Therefore
point of view, the discount rate is the interest rate generated by
兺 i⫽1
m
Ui U 1 ⫻ 兺 i⫽1
N
0.95共 i⫺1 兲 ⫹U 2 x 共 m⫺N 兲 investing the capital in a capital market 共Jackson 1996兲. Birge and
U av⫽ ⫽ Zhang 共1999兲 stated that ‘‘the market price of risk is the premium
m m
that investors must receive over the risk free rate to incur the
for i⫽1,2, . . . ,m (12d) market risk.’’ In addition, according to Lohmann and Baksh

JOURNAL OF CONSTRUCTION ENGINEERING AND MANAGEMENT / JANUARY/FEBRUARY 2003 / 91

J. Constr. Eng. Manage. 2003.129:89-97.

 共1993兲 ‘‘the reduction in risk of ruin attained by . . . increasing the Debt Service Coverage Ratio 共DSCR兲: the ratio of annual
the-risk adjusted discount rate.’’ Because the investment cost and cash available at hand to annual total debt service. It is calculated
revenue are in current United States dollars term 12% discount as
rate 共if United States bond yields are 9% per annum in 2000, then
3% risk premium per annum is accepted兲, they are generally ac- PBITi ⫹DEPi ⫺TAXi
cepted in the BOT HEPP projects in Turkey from the equity hold- DSCRi ⫽ for i⫽1,2, . . . ,m (15a)
Di
er’s point of view.
The lender’s main criteria for a project’s financial viability is Hence

DSCRi ⫽
0.89共 U i P i ⫺OMi 兲 ⫹0.11 冉 TPC
m
⫹ 共 1⫺e 兲 TPC⫻
r 共 1⫹r 兲 N
共 1⫹r 兲 N ⫺1
⫻ 关 1⫺ 共 1⫹r 兲 ⫺ 共 N⫺i⫹1 兲 兴 冊 for i⫽1,2, . . . ,m (15b)
r 共 1⫹r 兲 N
Downloaded from ascelibrary.org by New York University on 05/11/15. Copyright ASCE. For personal use only; all rights reserved.

共 1⫺e 兲 TPC⫻
共 1⫹r 兲 N ⫺1

Formulation of Linear Programming Model U 1 because now some financial difficulties are faced by TEA, due
to the previously signed concession agreements and PPAs with
Linear Programming 共LP兲 models are extensively used in capital high initial and average tariffs. It is obvious that electricity pur-
budgeting 共Park and Sharpe-Bette 1990兲. The objective is to chases at a higher price than the electricity tariff charged to end
maximize IRR, ‘‘the best way to compute a rate of return for an customers will require additional sources of funds for TEA to
investment...’’ 共Marshall and Bansal 1992兲. compensate for its losses. This, in turn, will overburden the gov-
In this model, optimal capital structure is the mix of debt and ernment budget. Hence, the upper limit for U 1 should be less than
equity that maximizes IRR from the equity holder’s point of view, the highest tariff to the end-user, U 1 ⭐10 cents/kW•h.
with the following constraints: This is a ‘‘win-win’’ solution for a BOT HEPP project: to find
1. Minimum equity amount allowed by legislation is 20%. an optimal equity level that maximizes equity holder’s IRR with-
共This is stated in Article 6 of Decree No. 94/5907 of the out sacrificing the other parties’ interest.
Council of Ministers related to the implementation of the The first constraint is a legal constraint; the subsequent two
BOT Law No. 3996.兲 constraints determine the financial viability of the project from
2. IRR must be greater than the discount rate. In other words, the equity holder’s point of view. The remaining two are con-
NPV must be positive.
straints set for lenders and purchasers of the product to feel com-
3. PBIT should be always greater than zero, for financial viabil-
fortable in this deal.
ity of the project.
Generally, IRR, DSCR, and the electricity tariff are not linear
4. Average DSCR should be at least equal to 1.50. Koh et al.
functions of equity. However, they can be approximated as linear
共1999兲 stated that DSCR in the range of 1.10 to 1.25 is
functions of equity up to the correlation coefficient square of
bankable, 1.30 to 1.50 is satisfactory and comfortable, and
0.97, which is reasonable.
above 1.50 is preferable. Interviews with the managers of
Consequently, the linear programming model is
some private power companies in Turkey show that the pre-
ferred minimum average DSCR by international financial au- Objective function: Maximize IRR⫽ f 共 E 兲 (16)
thorities is 1.50, due mainly to the current country credit
rating of Turkey. 共At the end of 1999, Turkey’s foreign cur- Constraints: Subject to e⭓0.20 (17)
rency long-term sovereign credit rating was affirmed by
IRR⭓12 or NPV⭓0 (18)
Standard and Poor as ‘‘B,’’ and outlook on the long-term
rating has been revised to positive to stable, this reflects the for i⫽1,2, . . . ,m
possibility of an upgrade.兲
5. Average electricity tariff should be not be greater than 5.0 for j⫽1,2, . . . ,c
cents/kW•h and the highest tariff should be less than or m

兺 DSCRi ⭓1.50⫻m
equal to 10.0 cents/kW•h, which is the upper economic limit
of the purchase price for TEA, without additional finance. (19)
i⫽1
The unit price of electricity in Turkey is at the same level as
in developed countries, on average, 7– 8 United States cents/ for i⫽1,2, . . . ,m
kW•h. According to the recent survey by TEA, electricity
losses during distribution could reach about 20%. For com- U a v ⭐5 Cents/kW•h (20)
parison purposes, the global average is 10%. Approximately U 1 ⭐10 cents/kw•h
30% of the unit price is general expenses and profit for TEA.
Thus, the average unit sale price of electricity should not An automated computer spreadsheet solution is suitable to
exceed 5 cents/kW•h from TEA’s perspective, U a v model the formulation of the objective function and constraints
⭐5 cents/kW•h. for LP and solve the problem. OPTIMUM, a simplified financial
In the early years of BOT development, the government pur- and LP model using Microsoft EXCEL 97’s 共a popular commer-
sued the policy of accepting high tariffs to boost development. cial spreadsheet application兲 built-in functions, macros, and VBA
However, it is clear now that there should be an upper limit for modules, has been developed by Bakatjan 共2000兲. In the model,

92 / JOURNAL OF CONSTRUCTION ENGINEERING AND MANAGEMENT / JANUARY/FEBRUARY 2003

J. Constr. Eng. Manage. 2003.129:89-97.

Downloaded from ascelibrary.org by New York University on 05/11/15. Copyright ASCE. For personal use only; all rights reserved.

Fig. 2. Input data for base case

a. Loan repayment period 共N, years兲,

b. Loan interest rate 共r, %兲.
4. Operation
a. Estimated annual production ( P i , GW•h兲,
b. Average unit price of electricity, cents/kW•h 共this is particu-
larly important to win the bid兲.
Based on the above inputs, the program automatically deals
with the changing debt-to-equity ratio. Using Eq. (12e兲, U 2 ,
hence, U 1 , then TPC, NPV, IRR, and DSCR are calculated for
each debt-to-equity ratio, and their graphs are plotted as the linear
Fig. 1. Flow chart diagram for OPTIMUM functions of equity.
The equity drawn during construction is taken as negative, and
the dividend is taken as positive in the IRR equation. The solution
links between input data, the financial model, and the LP model is found using the EXCEL built-in function, which employs an
are established by writing macros in EXCEL and a small program iterative method that clearly depends on starting values. Theoreti-
written in VBA to ensure automation in. After the completion of
the input data, the program develops the financial model, formu-
lates the LP objective function, and constraints automatically. Table 1. The Lowest and Highest Tariff Calculations at 20 and 60%
Subsequently, it solves LP. Equity
20% 60%
Tariff calculation Equity Equity
Input Data and Working Principle of Model
Lowest tariff for last year of concession
The model inputs are Annual depreciation, 共thousands United States 8,519 7,821
1. Estimated base cost 共in thousands of U.S. dollars兲 dollars兲
a. Civil works, Annual O and M cost 共thousands United States 790 790
b. Electromechanical works, dollars兲
c. Connections, if any 共there may be a need to establish con- Total, DEP⫹OM 9,309 8,611
nection line共s兲 to the main transmission grid兲, Annual energy production, P 共GW•h兲 405.8 405.8
d. Additional cost 共engineering, insurance, expropriation cost, Lowest tariff, U 2 ⫽(DEP⫹OM)/ P 共cent/kW•h兲 2.29 2.12
etc.兲. Highest tariff for last year of concession
2. Construction duration and cash flows Average tariff 共cent/kW•h兲 to win the bid 4.75 4.75
a. Construction duration 共c兲 in years, 8.025 8.025
Sum of the coefficients, 兺
10 (i⫺1)
0.95
b. Cash flows during construction 共as % of BC for each year兲. i⫽1
Highest tariff, U 1 共cent/kW•h兲 8.98 9.19
3. Loan terms

JOURNAL OF CONSTRUCTION ENGINEERING AND MANAGEMENT / JANUARY/FEBRUARY 2003 / 93

J. Constr. Eng. Manage. 2003.129:89-97.

 Table 2. Comparison of Interest During Construction and Total Project Cost at 20 and 60% Equity 共in thousands of United States Dollars兲
Construction period 共year兲 1 2 3 4 Total
20% Equity
Cash flow based on base cost 16,570 36,455 39,770 39,770 132,565
Debt amount 共80%兲 13,257 29,164 31,816 31,816 106,053
Interest during construction 6,153 10,461 7,537 3,735 27,887
Escalation during construction — 1,495 3,328 5,095 9,918
Total project cost (TPC⫽BC⫹IDC⫹EDC)⫽170,370
Cash flow during construction 共A兲 21,296 46,852 51,111 51,111 170,370
Debt drawing during construction 共80%兲 17,038 37,481 40,889 40,889 136,297
Equity drawing during construction 共20%兲 4,259 9,370 10,222 10,222 34,073
60% Equity
Cash flow based on BC 16,570 36,455 39,770 39,770 132,565
Downloaded from ascelibrary.org by New York University on 05/11/15. Copyright ASCE. For personal use only; all rights reserved.

Debt amount 共40%兲 6,628 14,582 15,908 15,908 53,026

Interest during construction 3,076 5,231 3,769 1,868 13,944
Escalation during construction — 1,495 3,328 5,095 9,918
Total project cost (TPC⫽BC⫹IDC⫹EDC)⫽156,427
Cash flow during construction 共A兲 19,554 43,017 46,928 46,928 156,427
Debt drawing during construction 共40%兲 7,821 17,207 18,771 18,771 62,570
Equity drawing during construction 共60%兲 11,732 25,811 28,157 28,157 93,857

cally, there may be as many solutions as the power of the respec- can be compared. Due to commercial confidentiality, the name
tive polynomial; however, it is solved for a local solution close to and location of the project will not be mentioned. The base cost of
the assumed discount rate. the project is $132,565,000, with a civil works cost of
Then, using EXCEL’s built-in Solver, the LP is solved, and the $95,370,000, and an electromechanical cost of $26,333,000, in-
optimal equity level, and hence, the optimal capital structure is cluding contingencies of 10% for civil works and 5% for EMC.
determined. The flow chart diagram for OPTIMUM is shown in There are also interconnections with the transmission grid, with a
Fig. 1. Based on the result, decisionmakers can decide whether to cost of $3,092,000 and an additional cost of $7,770,000 for engi-
go ahead by themselves, take new sponsors to raise the equity, or neering, insurance, expropriation, and working capital costs. The
use mezzanine finance for optimal capital structure immediately duration for construction is estimated as 4 years. It is planned that
after the feasibility study. Hence, the model gives great advantage 12.5% of the total construction works is to be completed in the
for promoters to start negotiation with other potential sponsors, first year, and 27.5, 30, and 30% in the following years. Accord-
i.e., financial institutions, both in terms of time and money. ing to the feasibility report, the annual net energy production is
405.8 GW•h. 共This is not real net production, but it is the base for
Case Study payments that will be entered in PPA.兲 To win the bid, the average
unit price of electricity is forecast as 4.75 cents/kW•h.
This section illustrates the application of OPTIMUM to a real-life The base case is the estimation based on the projected cash
project. The project is at the negotiation stage; hence, the results flows 共Ross et al. 1995兲. Hence, the optimal capital structure of

Fig. 3. TPC as function of equity

94 / JOURNAL OF CONSTRUCTION ENGINEERING AND MANAGEMENT / JANUARY/FEBRUARY 2003

J. Constr. Eng. Manage. 2003.129:89-97.

Downloaded from ascelibrary.org by New York University on 05/11/15. Copyright ASCE. For personal use only; all rights reserved.

Fig. 4. NPV as function of equity

the project is determined using the forecast input data. Input data equations and correlation coefficient squares 共R-squares兲 are
for the base case is shown in Fig. 2. shown in the figures.
After the completion of the input, the lowest and the highest As expected, TPC is a linear function of equity with negative
tariffs, TPC, NPV, IRR, and DSCR are calculated for each debt- slope, because less debt in capital structure means less interest
to-equity ratio from 80–20 to 40– 60. Calculation of the highest during construction, accordingly, TPC is a declining function of
and the lowest tariffs and comparison of IDC and TPC at 20 and equity. As a result, more equity means less TPC, thus, less total
60% equity is given in Tables 1 and 2. investment cost for a project. This is one of the reasons why the
Finally, TPC, NPV, IRR, and DSCR are plotted as the func- government favors high equity. With the discount rate of 12%,
tions of equity level 共Figs. 3, 4, 5, 6兲. The corresponding linear NPV is also a declining function of equity, with positive value for
all equity levels for this particular case. Hence, it is not a critical
Table 3. Result of Linear Programming Model constraint for the optimal capital structure in this case. IRR is also
Equity IRR NPV
a declining function of equity. That is why the sponsor of the
Model 共%兲 共%兲 共thousands of United States Dollars兲 DSCR project tends to keep equity as low as possible to increase its
return.
LP 31.69 14.94 7,888.61 1.50 The only ascending function of equity is DSCRa v . As equity
Financial 31.69 14.74 7,810.30 1.47
increases, debt obligation decreases; hence, DSCRa v increases.

Fig. 5. IRR as function of equity

JOURNAL OF CONSTRUCTION ENGINEERING AND MANAGEMENT / JANUARY/FEBRUARY 2003 / 95

J. Constr. Eng. Manage. 2003.129:89-97.

Downloaded from ascelibrary.org by New York University on 05/11/15. Copyright ASCE. For personal use only; all rights reserved.

Fig. 6. DSCRa v as function of equity

Table 4. Debt Principal and Interest Calculations at Optimal Capital Structure (E⫽31.69%)
Debt repayment period, N 共year兲 1 2 3 4 5 6 7 8 9 10
Annual equal debt payment, D 18,487 18,487 18,487 18,487 18,487 18,487 18,487 18,487 18,487 18,487
Debt principal 7,128 7,840 8,624 9,487 10,435 11,479 12,627 13,890 15,279 16,806
Interest chargea 11,359 10,647 9,863 9,000 8,052 7,008 5,860 4,597 3,208 1,681
Note: Compounding interest factor⫽0.163.
a
On average debt beginning and end of period.

Table 5. Debt Service Coverage Ratio Calculation at Optimal Capital Structure (E⫽31.69%)
Debt repayment period, N 共year兲 1 2 3 4 5 6 7 8 9 10
Profit before interest and taxes 27,579 25,745 24,004 22,349 20,774 19,281 17,860 16,513 15,231 14,013
Depreciation 8,315 8,315 8,315 8,315 8,315 8,315 8,315 8,315 8,315 8,315
Tax 1,784 1,661 1,556 1,468 1,399 1,350 1,320 1,311 1,323 1,357
Total cash available (PBIT⫹DEP⫺TAX) 共C兲 34,110 32,399 30,763 29,196 27,690 26,246 24,855 23,517 22,223 20,971
Debt interest payment 11,359 10,647 9,863 9,000 8,052 7,008 5,860 4,597 3,208 1,681
Debt principal payment 7,128 7,840 8,624 9,487 10,435 11,479 12,627 13,890 15,279 16,806
Total debt repayment 共D兲 18,487 18,487 18,487 18,487 18,487 18,487 18,487 18,487 18,487 18,487
DSCR (C/D) 1.85 1.75 1.66 1.58 1.50 1.42 1.34 1.27 1.20 1.13
Note: Average DSCR⫽1.47.

Table 6. Financial Cash Flow Statement for Optimal Capital Structure from Equity Holder’s Point of View 共Thousands of United States Dollars兲
Year 1 2 3 4 5 6 7 8 9 10 11 12 20
Electricity tariff 共cent/kW•h兲 共average 4.75兲 9.04 8.59 8.16 7.75 7.36 7.00 6.65 6.31 6.00 5.70 2.24 2.24 2.24
Annual energy production, P 共GWh兲 405.8 405.8 405.8 405.8 405.8 405.8 405.8 405.8 405.8 405.8 405.8 405.8 405.8
Revenue 36,684 34,850 33,109 31,454 29,879 28,386 26,965 25,618 24,336 23,118 9,106 9,106 9,106
Depreciation 共linear兲 8,315 8,315 8,315 8,315 8,315 8,315 8,315 8,315 8,315 8,315 8,315 8,315 8,315
O and M expenses 3% of EMC 790 790 790 790 790 790 790 790 790 790 790 790 790
Profit before interest and taxes 27,579 25,745 24,004 22,349 20,774 19,281 17,860 16,513 15,231 14,013 — — —
Interest charge on debt 11,359 10,647 9,863 9,000 8,052 7,008 5,860 4,597 3,208 1,681 — — —
Profit before tax 16,220 15,098 14,141 13,349 12,722 12,273 12,000 11,916 12,023 12,332 — — —
Withholding tax 11% 1,784 1,661 1,556 1,468 1,399 1,350 1,320 1,311 1,323 1,357 — — —
Net profit 14,436 13,437 12,585 11,881 11,323 10,923 10,680 10,605 10,700 10,975 — — —
Depreciation 8,315 8,315 8,315 8,315 8,315 8,315 8,315 8,315 8,315 8,315 8,315 8,315 8,315
Cash flow 22,751 21,752 20,900 20,196 19,638 19,238 18,995 18,920 19,015 19,290 8,316 8,316 8,316
Debt principal 7,128 7,840 8,624 9,487 10,435 11,479 12,627 13,890 15,279 16,806 — — —
Total cash available for shareholders 共NCA兲 15,623 13,912 12,276 10,709 9,203 7,759 6,368 5,030 3,736 2,484 8,316 8,316 8,316
共6,821兲 共15,007兲 共16,372兲 共16,372兲 15,623 13,912 12,276 10,709 9,203 7,759 6,368 5,030 3,736 2,484 8,316 8,316 8,494
Note: Optimum equity, E⫽31.69%; NPV@12%⫽7,810.90; IRR⫽14.74%.

96 / JOURNAL OF CONSTRUCTION ENGINEERING AND MANAGEMENT / JANUARY/FEBRUARY 2003

J. Constr. Eng. Manage. 2003.129:89-97.

 Table 7. Average DSCR Constraint and Optimal Capital Structure As a result, more equity means less TPC, and thus less total
Average DSCR Optimum equity level 共%兲 IRR 共%兲 investment cost for a project. This is one of the reasons why
government favors high equity.
1.50 31.69 14.94 • With the discount rate of 12%, NPV is also a declining func-
1.45 30.24 15.03 tion of equity with positive value for all equity levels for this
1.40 28.80 15.13 particular case. Hence, NPV is not a critical constraint for
1.35 27.35 15.22 optimal capital structure in this case. Internal rate of return is
1.30 25.91 15.31 also a declining function of equity. That is why the sponsor of
1.25 24.46 15.41 the project tends to keep equity as low as possible to increase
its return.
• The only ascending function of equity is the average debt ser-
Thus, a high DSCRa v requirement by lenders results in high eq- vice coverage ratio. As equity increases, debt obligation de-
uity for the project. creases, hence DSCRa v increases. Thus, a high DSCRa v re-
Accordingly, the LP model is quirement by lenders results in high equity in the project.
Downloaded from ascelibrary.org by New York University on 05/11/15. Copyright ASCE. For personal use only; all rights reserved.

Objective function: Maximize IRR⫽⫺0.065E⫹16.992 It can be concluded that the major constraint for the optimal
Subject to: E⭓20 capital structure is DSCR. According to Standard & Poor, it is
⫺46.93E⫹9,375.87⭓0 possible to upgrade Turkey’s current credit rating if fiscal adjust-
ment and disinflation measures in the International Monetary
0.035E⫹0.404⭓1.50 Fund standby agreement are implemented. If Turkey’s rating is
where E is the equity as percentage of TPC. upgraded, the DSCR requirement will be lowered, and the opti-
The model automatically employs the EXCEL Solver to solve mum equity could fall below 25%, near the legal minimum re-
the LP. The LP result is provided in Table 3, and is to be com- quirement.
pared with the financial model’s result at optimum equity. Conse-
quently, for this particular project, equity for optimal capital References
structure is 31.69%.
Bakatjan, S. 共2000兲. ‘‘Financial engineering for BOT power projects: A
There are slight differences between NPV, IRR, and DSCR
simplified model for optimal capital structure,’’ MS thesis, Middle
calculated by the financial model and the linear programming
East Technical University, Ankara, Turkey.
model due to the linear approximation that is negligible. Birge, J. R., and Zhang, R. Q. 共1999兲. ‘‘Risk-neutral option pricing meth-
Calculations of debt principal 共DPR兲, debt interest 共INT兲, and ods for adjusting constrained cash flows.’’ Eng. Econom., 44共1兲, 36 –
DSCR at the optimal capital structure (E⫽31.69%) are given in 49.
Tables 4 and 5. It is worthwhile to note that these parameters, as Dias, A., Jr., and Ioannou, P. G. 共1995兲. ‘‘Debt capacity and optimal
well as TPC, NPV, and IRR, are automatically calculated for each capital structure for privately financed infrastructure projects.’’ J.
debt-to-equity ratio by OPTIMUM. Constr. Eng. Manage., 121共4兲, 404 – 414.
A cash flow statement from the equity holder’s point of view is Jackson, T. 共1996兲. ‘‘Risk, time and the discount rate.’’ The 1996 Annual
given in Table 6 for optimal capital structure. It should be noted Meeting of the Society for Risk Analysis-Europe.
that in the last row the equities drawn during the construction Jenkins, G. P., and Marchesini, M. 共1999兲. ‘‘An analysis of electricity
generation and tariff options in Ghana.’’ Dev. Discuss. Paper No. 702,
period are shown in parentheses because they are negative.
Harvard Institute for International Development, Cambridge, Mass.
It can be concluded that the major constraint for the optimal Koh, B. S., Wang, ShouQing, and Tiong, L. K. R. 共1999兲. ‘‘Qualitative
capital structure is DSCR. The main constraints for Turkey’s rat- development of debt/equity model for BOT infrastructure projects.’’
ings are a fragmented political environment, very limited fiscal Proc., Int. Conf. on Constr. Proc. Reengrg. (CPR-99), 501–512.
flexibility, and vulnerability to external shocks. According to Lohmann, J. R. 共1988兲. ‘‘The IRR, NPV and the fallacy of the reinvest-
Standard & Poor, it is possible to upgrade Turkey’s rating if fiscal ment rate assumptions.’’ Eng. Econom., 33共4兲, 303–330.
adjustment and disinflation measures in the IMF 共International Lohmann, J. R., and Baksh, S. N. 共1993兲. ‘‘The IRR, NPV and payback
Monetary Fund兲 standby agreement are implemented. If Turkey’s period and their relative performance in common capital budgeting
current credit rating is upgraded, then the DSCR requirement will decision procedures dealing with risk.’’ Eng. Econom., 39共1兲, 17– 47.
be lowered; hence, optimum equity will fall below 25% 共Table 7兲, Malhotra, A. K. 共1997兲. ‘‘Private participation in infrastructure: Lessons
from Asia’s power sector.’’ Fin. and Dev., 34共4兲, 33–35
near to the legal minimum requirement.
具http://www.worldbank.org/fandd/典 共Nov. 10, 1999兲.
Marshall, J. F., and Bansal, V. K. 共1992兲. ‘‘Valuation relationships and
applications.’’ Financial engineering, New York Institute of Finance,
Conclusion Allyn and Bacon Inc., New York, 63–103.
Park, C. S., and Sharp-Bette, G. P. 共1990兲. Advanced engineering
In this research, a prototype program OPTIMUM for decision- economy, Wiley, New York.
makers is developed. The program evaluates a project from the Ranasinghe, M. 共1996兲. ‘‘Total project cost: A simplified model for deci-
equity holder’s 共often, contractor兲 point of view. As a result, the sion makers.’’ Constr. Manage. Econom., 14, 497–505.
program allows decisionmakers to make early decisions for capi- Ross, S. A., Randolph, W. W., and Jordan, B. D. 共1995兲. ‘‘Project analysis
tal structuring, hence association structuring to get the optimal and evaluation.’’ Fundamentals of corporate finance, 3rd Ed., Richard
D. Irvin Inc., Ill., 266 –273.
capital structure. To show the applicability of the model, a real
Walker, C., and Smith, A. J. 共1995兲. Privatized infrastructure: The Build
case study is conducted. Operate Transfer approach, Thomas Telford, London.
The following conclusions are drawn from the case study: White, J. A., Agee, M. H., and Case, K. E. 共1989兲. Principles of engi-
• As expected, TPC is a function of equity with negative slope. neering economic analysis, 3rd Ed., Wiley, New York.
Less debt in capital structure means less interest during con- Wynant, L. 共1980兲. ‘‘Essential elements of project finance,’’ Harvard
struction, accordingly, TPC is a declining function of equity. Business Rev., May-June, 1980, 165–173.

JOURNAL OF CONSTRUCTION ENGINEERING AND MANAGEMENT / JANUARY/FEBRUARY 2003 / 97

J. Constr. Eng. Manage. 2003.129:89-97.