Transcript
ANALYSIS OF THE DULLES GREENWAY
The provision of highway infrastructure has since the middle of this century been considered a public
responsibility in the United States. Development of the Interstate Highway System (IHS) from 1956 to
1990 was fueled by direct federal funding mechanisms. The enacting legislation in 1956 authorized up
to 90% of the cost of urban, primary, and farm-market roads to be paid by the federal government.
Since inception, approximately $40 trillion has been spent on building and maintaining nearly 40,000
miles of highway systems (Miller [2000]). While the IHS program proved successful for some time,
dwindling federal resources and growing demands on the maintenance, rehabilitation, and continued
development of the system over the past two decades have placed a tremendous strain on the
condition and performance of one of the nation's most vital resources.
Consequently, public officials recently have explored the use of alternative finance and project
delivery strategies that leverage private capital for public purposes. These alternative strategies
include build-operate-transfer (BOT), design-build-finance-operate (DBFO), design-build-operate (DBO),
turnkey, and others described by Miller [1999]. Typically, these private highway project arrangements
transfer property control and cash flow rights from the public to the private domain for a specified
concession period. Property control rights may be limited to the transportation corridor or may include
parcels of land adjacent to the corridor. Cash flow rights most often involve user fees (tolls), receipts
from the lease or sale of adjacent properties, or in some cases "shadow tolls."(*) Special legislation
and "innovative" contract mechanisms generally are employed to balance the risks among the public
and private participants.
Market risk, or the eventual demand for the highway system based on established toll rates, is among
the most significant of these project risks. Private sponsors must balance toll rates with demand (or
traffic volume) to recover large capital costs, pay ongoing expenses, and service debt. The critical
issue to these private participants is generating an appropriate return on equity and ensuring that the
project is a viable concern. Indeed, it can be argued that the public sector should have an equal
interest in the private participants' profitability if privatization is to remain a viable mechanism for the
long-term sustainment of American infrastructure. To this end, the public sector often provides direct
funding subsidies or indirect funding incentives in the way of tax relief, development rights to
properties located in proximity to the highway project, and other means. These indirect incentives and
other project contingency features, such as waiting and learning before investing and making follow-on
investments if the initial investment succeeds, create investment opportunities that are often ignored
by the strict application of traditional valuation methods.
This article retrospectively examines the investment decision of a recently developed private highway
project using traditional valuation methods and a simple option valuation model. The latter model is
developed to clarify the market uncertainty confronting this project, to approximate the value of
waiting and acquiring more information to resolve this uncertainty, and to demonstrate an improved
technique for investment decision-making. Development of this model emphasizes the significance of
recognizing the risk associated with future revenue and the irreversibility of the investment decision,
both of which are characteristic of large--scale private highway projects. The article begins with a
review of asset-backed financing, traditional valuation methods and associated shortcomings, and
managerial flexibility employed in infrastructure development.
BACKGROUND Project Financing and Risk
Asset-based (or project) financing is an alternative to conventional financing that permits private
sponsors to shift financial risks from their own balance sheets to the assets of the project. Project
financing is commonly used with alternative project delivery approaches, such as BOT, DBFO, DBO,
and their variants, where private capital is at stake over a long horizon. Project financing is considered
viable when an infrastructure facility can function as an independent economic unit. Highways are an
example of infrastructure assets that can function independently to generate revenues from tolling
arrangements that offset ongoing costs, dividend payments, debt service, and investment outlays.
Private sponsors typically form a limited-life business structure (e.g., corporation, limited liability
company, partnership, joint venture) to oversee the development, financing, construction, and
operation of the facility. The sponsor company then can raise capital on a project basis by issuing
equity and debt securities that are self-liquidating from the revenues derived from project operations.
In this arrangement, a number of firms with unique expertise may unite as the project sponsor
company to distribute risks to those best able to manage them without burdening their own balance
sheets.
A critical aspect of infrastructure development is the determination, allocation, and management of
project risks. In the case of project financing, lenders, equity participants, and other stakeholders are
concerned with the ability of the infrastructure facility to provide a sufficient return for the risks they
are asked to bear. The analysis and evaluation of a project's technical feasibility, creditworthiness,
and economic viability must be addressed to the satisfaction of these stakeholders for the project to
proceed. Technical feasibility accounts for new technologies, environmental factors, construction risks,
and the operational performance of the project. Creditworthiness refers to the ability of project
revenues to adequately cover operating costs and debt service requirements over the life of the
project. It is of particular concern in asset-based financing, especially to lenders, since their only
recourse is the project and its associated assets and cash flows.
Economic viability depends mainly on the stability of profits over the life of a project. Profitability may
be compromised initially by rising capital costs during construction. During operations, profitability is
sensitive to market conditions that impact the price of, and demand for, a project's output. Price and
demand are interdependent variables that ultimately determine the revenues generated by a project.
Increasing operating costs or declining revenues reduce a project's creditworthiness and ability to
provide an attractive rate of return to equity investors. Thus, pricing strategies in private highway
tolling arrangements and eventual traffic volume present a significant uncertainty that must be
appropriately analyzed and evaluated in the initial planning stages of development.
Traditional Valuation Methods
Economic viability is the crux of the investment (or capital budgeting) decision, which seeks to find
whether a project generates benefits that are worth more than its costs. The use of discounted cash
flows (DCF) to determine net present value (NPV) is the preferred method for establishing the value
of a project that is not established in an active market, such as infrastructure projects (see Brealey
and Myers [1996], pp. 85-106). Miller and Evje [1997] demonstrate the use of these techniques in
investigating procurement strategies for infrastructure projects. The capital budgeting decision is
followed by the capital financing decision, which aims to determine appropriate levels of equity and
debt required for the project. In this way, the investment decision is simplified and evaluated as an
all-equity project prior to adding the complexity of the financing decision. Finnerty [1996] explains
the use of discounted cash flows in analyzing and planning project financing strategies.
Shortcomings of Traditional Valuation Methods
Amram and Kulatilaka [1999], Trigeorgis [1999], Dixit and Pindyck [1994], Myers [1984], and others
point to the shortcomings of discounted cash flows in valuing flexibility. The deficiencies center
around two implicit assumptions of DCF. First, investment is considered reversible. That is, the initial
investment is assumed recoverable in the event that a negative outcome occurs. In fact, large-scale
infrastructure projects are typically quite the opposite. Initial investment is generally irreversible in
that planning, design, and construction expenses are usually sunk costs. Second, if the investment is
irreversible, then it must occur immediately, as the opportunity to invest later does not exist. In effect,
these assumptions imply that managers are passive bystanders and are not able to respond to new
information. But in reality, managers often employ flexibility throughout the development process in
order to respond to new information and limit exposure to downside events, while seizing upside
opportunities. They also usually have the option to delay investment. Oftentimes the more relevant
decision is not whether to invest, but when.
Luehrman [1997] and Myers [1984] suggest that traditional valuation methods are adequate for
investment decisions regarding existing operations (or assets-in-place). In these cases, ongoing
operations generate relatively safe cash flows and are held for this reason, not for less tangible
strategic purposes. They also work well for typical engineering investments, such as equipment
replacement, where the main benefit is cost reduction. However, when capital investment creates
future growth opportunities (e.g., follow-on development if product demand is favorable) or
contingency possibilities (e.g., delay project or abandon project), DCF methods understate the value
of this flexibility. Exhibit 1 (adapted from Luehrman [1997]) illustrates this argument. In the left pane,
the investment decision is made a priori to the resolution of future outcomes. In the right pane,
uncertain outcomes are known before the investment decision is made. In this case, the investment
can be deferred or abandoned if the future state is unfavorable. Thus, future losses are averted
resulting in a higher present value.
Valuing Flexibility
Large scale, significant capital costs, and a long useful life generally characterize infrastructure
investment. As a result, development usually proceeds in a series of stages that aim to better define
project scope and discover unknown information. Moreover, flexibility often is incorporated as an
intuitive managerial approach to deal more effectively with uncertainty. Preliminary planning and
feasibility studies, such as environmental impact studies, geo-technical surveys, and traffic volume
analyses, can reveal information that may alter further investment and development decisions.
Flexible design permits infrastructure projects to adapt more readily to changing conditions such as an
increase or decrease in expected demand for the project's output. Staged construction can afford
decision-makers the opportunity to gain more information as market conditions become more certain.
In short, flexibility can effectively reduce the life cycle costs of a project by allowing a more timely and
less costly response to a dynamic environment. Flexibility adds value; it is implicitly used as a hedge
(or insurance) or coping mechanism against uncertain outcomes. However, flexibility comes at a cost
in terms of money, time, and complexity. The added value of flexibility must be weighed against its
cost. Traditional evaluation methods do not adequately support such analyses. The following case
serves to illustrate this point.
CASE STUDY: DULLES GREENWAY Project Background
In 1987, the Virginia Department of Transportation (VDOT) began planning studies for extending the
existing Dulles Toll Road from the Dulles International Airport westward to the junctions of Routes 7
and 15 in Leesburg, Virginia. The road extension would connect a growing residential population in
rural Loudoun County to expanding work centers in northern Virginia and Washington, D.C. At the
time, only two major east-west arterial roads, State Routes 7 and 50, linked the two regions. The
typical western commuter on these routes experienced increasing congestion and frequent stops at
traffic lights. The Dulles Toll Road extension aimed to greatly improve this commute and spur
economic growth in Loudoun County.
Facing a deficit of $7 billion for transportation improvement needs, the Virginia General Assembly
passed legislation authorizing the private development of toll roads in 1988. This enabling statute was
followed one year later by an application from a group of private investors to design, build, finance,
and operate the Dulles Toll Road extension. The private development group was a partnership
comprised of the Shenandoah Greenway Corporation (formed by the Bryant/Crane family of nearby
Middleburg, Virginia), Autostrade International (an Italian toll road developer and operator), and
Kellogg Brown & Root (a large American construction firm based in Houston, Texas). The developers,
who would eventually be known as the Toll Road Investors Partnership II (TRIP II), were approved by
the state and granted a certificate of authority in 1990.
Project scope and financing. Under TRIP II, the Dunes Toll Road extension became the Dulles
Greenway. Three years of planning, design, and arranging financing resulted in the state's approval
of a four-lane, limited-access highway with seven interchanges. Two future interchanges would be
added when traffic volume reached appropriate levels. Located within 250 feet of right of way, the
project was designed to accommodate future lane expansion. Electronic toll collection technology was
included in the design to maintain steady traffic flow. The project originally was scheduled to start
construction in 1989 and operations in 1992, but difficulties in securing financing and environmental
permits caused delay. Construction commenced in September 1993 and ended six months ahead of
schedule in September 1995.
TRIP II contributed up to $40 million in equity and arranged for another $46 million in lines of credit to
cover potential revenue shortfalls during operations. The Shenandoah Greenway Corporation
contributed $22 million, while the other partners accounted for the balance. The bulk of the project's
financing, $258 million in long-term fixed-rate notes, was provided by a consortium of 10 institutional
investors. The three lead debt investors were CIGNA Investments Inc., Prudential Power Funding
Associates, and John Hancock Mutual Life Insurance Company. A bank group consisting of Barclays
Bank, NationsBank, and Deutsche Bank provided a portion of construction financing and a $40 million
revolving credit. TRIP II's entire right, title, and interest in the highway project secured the financing.
Traffic demand and initial project operations. As a completely private venture, the Dulles
Greenway would provide some 40 years of cash flows to its investors and debt holders without public
subsidies. Revenues were dependent on the interrelationship between average daily traffic demand
and established toll rates. Independent consultants conducted traffic forecasts prior to construction as
a basis for planning and investment analysis. These traffic forecasts were based on a construction
start in 1989 and operations beginning in 1992. Moreover, the forecasts relied, in part, on the healthy
economic conditions of the late 1980s and did not account for the economic downturn in 1991,
particularly in the commercial and residential real estate markets. Approximately 20,000 vehicles per
day were projected for the first year of operation at a fixed toll rate of $1.50. By 1995, the daily
traffic demand was forecasted to be 34,000 based on the same toll rate. The four-year schedule slip
between actual operations and traffic projections was accounted for by using the daily ridership
forecast of 34,000 (see Pae [1995]).
Within six months of opening in late 1995, the project was in financial distress. Average daily traffic
demand was an abysmally low 10,500. Toll rates were reduced from an initial $1.75 to $1.00 by March
1996, and future toll hikes were deferred in an attempt to increase ridership. Furthermore, the state
legislature increased the speed limit on the highway to 65 mph (miles per hour). By July 1996, road
usage increased to 21,000 daily travelers, averaging 1% to 2% monthly growth. However, the net
effect on projected revenues was marginal, as decreased toll rates offset the increase in ridership.
TRIP II officials began discussions with the project's creditors in the summer of 1996 to work out a
plan for deferring debt payments and restructuring loan contracts (see Bailey [1996]).
Significant Issues
The Dulles Greenway was among the nation's first highway projects to be delivered with a design-
build-finance-operate franchise since the nineteenth century, when the U.S. commonly relied upon
the private sector for infrastructure development. Consequently, the project presents a number of
significant issues, which in retrospect are worth re-examination. These issues include the lack of
public funding participation and subsequent alignment of incentives, the unsolicited sole-source
nature of the procurement strategy, and the potential workout strategies for restructuring current
debt obligations. This case, however, focuses on revenue risks caused by uncertain traffic demand.
The initial investment decision is reconstructed and analyzed with traditional valuation methods. Then,
a simple option valuation technique is applied to illustrate its potential impact on the initial investment
decision.
INVESTMENT DECISION ANALYSIS Traditional valuation model.
The investment analysis is reconstructed from an ex ante perspective using cash flow estimates and
construction costs from financial models submitted to the state (see Toll Road Corporation of Virginia
[1993]). To simplify the analysis, pre-construction, construction, and other development costs are
combined as $279 million and assumed to occur in one year in 1995. Actual construction, including
financing, cost $326 million and was accomplished in a two-year period. Financing, taxes, depreciation,
and other costs are ignored as the focus of this analysis is on the investment decision, specifically
from the private developer's viewpoint. The investment decision from a public vantage might include
other social benefits, such as property tax increases from new development, federal and state taxes
paid by project operations, and savings in state transportation funds. The cash flows analyzed are
earnings before interest and taxes (EBIT), which occur annually over a 40-year period (see Exhibit 2).
In this case, EBIT is calculated as follows:
(1)
EBIT = Gross Revenue
- Operating Expenses
- Capital Improvements
(2)
Gross Annual Revenue =
Average Daily Traffic
x Average Toll Rate
x 365 days per year
Average daily traffic demand begins at 34,000; it is assumed to grow at a rate of 14% annually for the
first six years of operation. Demand growth tapers off to a rate of 7% per year for the remaining 34
years. The schedule of average annual toll rates begins at $2.00 and gradually rises to $3.00 by the
fifteenth year of operation. Note that the interdependence between traffic demand and toll rates is
not modeled. These assumptions roughly match those of TRIP II'S financial model. Operating expenses
include operations and maintenance costs, various fees, police costs, lease payments, and other
expenses from the TRIP II financial model. Operating expenses start at $9 million per year and grow at
a constant rate of 5% annually. Capital improvements include major planned repaying and road
widening activities. Follow-on construction of two key interchanges at a cost of $38 million is not
incorporated in the model. It is assumed that these interchanges will be added only if demand and
subsequent revenue materializes; otherwise, they are not required and are not considered in the initial
investment decision.
The next step is to estimate an appropriate discount rate for the project using the weighted average
cost of capital (WACC). According to the WACC, the return on assets (Ra), or discount rate, is a
function of the weighted average of the cost of equity (Re) and the cost of debt (Rd). The amount of
debt (D) and equity (E) are used in the project sum to determine a project value (V). In theory, this
value represents the market value of the project, and equity is backed out (E = V - D). However, in
this case, highway projects do not have a market value per se, so book values are used instead. The
cost of debt is then approximated as a composite of the rates from the multiple long-term notes used
for the project. Equation 3 illustrates the WACC and the variables defined thus far.
(3)
Ra = Re(E/V)+Rd(D/V), where V = E+D
= Re ($40M/$319M) + 10% ($279M/$319M)
The last variable, cost of equity (Re), is estimated using the capital asset pricing model (CAPM).
According to the CAPM, the cost of equity is a function of the risk-free rate of interest (Rf), the
project's equity beta (βe), and an appropriate equity risk premium (Rm - Rf). The risk free rate is
determined from yields on the 30-year bond in 1993, approximately 7.25%. The equity risk premium
is based on typical premiums (differences between equity market returns and risk free returns) for the
early 1990s, approximately 7.4%. The equity beta is then estimated using a weighted average of risk,
where the asset beta is assumed to be 0.4 for the transportation industry and the debt beta 0. The
equity beta is calculated as follows:
(4)
βe= βa (V/E) = 0.4 ($319M/$40M) = 3.2
Using the CAPM, the cost of equity is:
(5) Re = Rf + βe(Rm - Rf)
= 7.25% + 3.2(7.4%) = 30.9%
Returning to the WACC in Equation 3 and plugging in the cost of equity, the project's discount rate is
12.6%.
Traditional valuation analysis. Using the spread-sheet in Exhibit 2, the project's present value of
the net earnings is $560 million; with a capital expenditure of $279 million, the project's net present
value (NPV) can be readily determined as $281 million, and the decision to invest is clear. However,
recognizing the uncertainty of traffic demand, a simple sensitivity analysis of the project NPV to
demand, while holding other variables constant, illustrates the critical points that change the
investment decision (see Exhibit 3). Mainly, traffic demand below 19,868 results in a negative NPV and
alters the investment decision. Examining the sensitivity of NPV in combination with the project's
discount rate and traffic demand further refines the analysis.
Exhibit 4 displays varying discount rates as vertical gridlines emanating from the bottom horizontal
axis. Starting with 6%, each gridline represents a 50 basis point (0.5%) increment up to 18% on the far
right. The horizontal gridlines represent traffic demand, which increases from 10,000 at the bottom in
increments of 2,500. The shaded gray area (upper left) depicts a positive project NPV, while the white
area (lower bottom) portrays a negative project NPV. Using this figure, the discount rate can be fixed
and the corresponding vertical gridline followed up to the point where the investment decision
changes (where the gray area meets the white area). Reading across then reveals the critical traffic
demand value. For example, at 15% the critical demand value is between 22,500 and 25,000. Traffic
demand can be fixed as well and the critical discount rate determined. If the expected demand is
34,000, the decision appears insensitive to the range of discount rates considered. The key
conclusion of this analysis is that the investment decision is subject to change for demand values
below 33,244 and discount rates above 6.88%. Based on the initial assumptions of the case, the
private sponsor has a relatively large comfort zone in choosing to invest. Initial demand would have to
decline by over 41% (from 34,000 to 19,868) before the decision would be altered, and the discount
rate is inconsequential at this expected demand level.
Simple option valuation method. A new investment model is now developed using a simple
binomial tree. The intent of this model is to recognize more explicitly the uncertainty of initial traffic
demand and its effect on the value of the project. Moreover, this model is constructed to provide a
rough estimate of the value of acquiring more information to help resolve initial demand uncertainty.
The following general assumptions serve to bind the investment decision model.
• First, the option to wait to build the highway is limited to five years, because the public sector will
pursue transportation alternatives for the region with its own resources beyond that horizon.
Furthermore, the length of the concession remains the same regardless of when construction begins.
• Second, traffic demand uncertainty can be resolved (or at least narrowed) by observing demographic
growth patterns and other key variables in the regions connected by the highway project.
• Third, initial demand is the critical variable. Demand growth thereafter increases in a relatively
consistent and predictable manner. Presuming that the highway will spur economic growth and
subsequent traffic in the regions and that competing transportation means are eliminated by contract
during the concession period further supports this assumption.
• Fourth, the value of the project is represented by the present value of net earnings over the 40-
year operations period less the cost of construction, or NPV.
• Fifth, there is no direct cost to waiting. That is, the present value of net earnings and costs in five
years' time would be similar to the value today if the project were executed immediately. Inflation is
therefore ignored here for simplicity. Nonetheless, the evaluation model still caters for the opportunity
costs of asset appreciation by discounting the cash flow realized in the future by the risk-free rate.
The model sets up benefits from the ex post performance of the project in that it is now known that
the expected traffic demand for the first year of operation was overly optimistic. However, Pae [1995]
and others point to the critical fact that the private sponsors relied on an outdated traffic forecast that
did not account for an economic downturn in 1991. This forecast might easily have been updated or
further traffic analysis conducted prior to construction. Thus, it is assumed for this case that the
expected demand could have been revised ex ante to at least 20,000, which matches the private
sponsors' traffic study for the first year of operation, albeit four years out of date. The model further
assumes a high demand of 34,000 and a low demand of 10,000. Again, recognizing such uncertainty
ex ante seems reasonable, although potentially overly volatile in terms of up and down movements.
Exhibit 5 shows the basic setup of the model. The spreadsheet in Exhibit 1 is used to generate
project PVs of net earnings given the three states of demand. The PV based on expected demand
occurs in period 0 (year 1995), while the other two PVs are realized in period 5 (year 2000). The PVs
are calculated using the same project discount rate (12.6%) determined in the traditional valuation
model, and each represents 40 years of cash flows. A risk-neutral probability (p) is then calculated
using a risk-free rate of interest (assumed to be 7.25% over the five-year period) to equate the
possible PVs five years forward with the expected PV in period 0. Thus, in a risk-neutral world, where
all assets are assumed to appreciate at the risk free rate and investors' risk attitudes do not matter,
the probabilities of high demand and low demand are 0.665 and 0.335 respectively.
These risk-neutral probabilities now can be used to evaluate the value of the project embedded with
the deferral option. Because the capital expenditure to develop the project is $279 million and the PV
of net earnings in the down state (demand = 10k) is only $83 million, it is better to abandon the
project in the case of a down state instead of pursuing a negative NPV project. This scenario is
represented by Exhibit 6 in determining the expected value of the project (using risk-neutral
probabilities) as valued in period 0.
The base case NPV with the expected demand value of 20,000 is $3M [$282M - $279M]. Traditionally,
the private sponsor would have invested immediately in the project in period 0. However, this
approach ignores the additional value of waiting to build and acquiring more information to resolve
the demand uncertainty. Since the value of the project with deferral option is $132M, which is
substantially greater than the NPV of $3M from immediate project execution, it justifies that the
private sponsor should at least wait for some of these uncertainties in traffic demand to resolve.
Specifically, the difference of $129M between the two analyses is driven by the large volatility in
the traffic demand (swinging from 10,000 to 34,000) as typically the case in most transportation
development. If this volatility is minuscule, similar evaluation with the deferral option could be less
than the base case NPV and it would then be justified to start the development immediately.
Discussion of Results
The above analysis illustrates how uncertainty and choice of timing can alter an irreversible capital
investment decision. Recognizing the value of these effects necessitated looking beyond traditional
discounted cash flows and NPV. In this case, NPV is not replaced; rather, it is augmented by a very
simple model that explicitly recognizes the revenue risks caused by uncertain demand for the
project's output. Dulles Greenway's private sponsors faced an investment opportunity with the
option to exercise now or later in exchange for the value of nearly 40 years of cash flows. Given that
some probability existed for the investment to result in a loss suggests that the opportunity to delay
and acquire better information had value. From the above analysis, this value was substantial, given
that the volatility of demand is huge.
A more rigorous option valuation method may provide greater precision than the binomial model. The
exact value of the project with deferral option is even greater than what has been calculated because
the private sponsor really has the flexibility to exercise its option in years 1, 2, 3, and 4, rather than
constraining its choice to year 5. This would require solving a continuous-time model that is
substantially more complicated. However, the aim of this case was to demonstrate simply how
infrastructure investment opportunities could be treated as options and how the role of uncertainty
could be clarified.
There are several implications from this simple analysis:
• First, investment irreversibility and uncertainty create an incentive to wait and learn. In private high-
way development, resolution of uncertainty is a circular problem to some degree. That is, building the
road generates growth and subsequent demand, thereby resolving uncertainty. On the other hand,
demand uncertainty must be resolved within reasonable limits to ensure the project's economic
viability. In this case, the costs of acquiring better traffic demand information are more than justified.
• Second, delay has value, and it should be considered in the NPV analysis. That is, building the
highway immediately "kills" the option to wait, thus reducing the project's NPV.
• Third, planned highway corridors that are appropriate for privatization are worth more than might
otherwise be determined by traditional NPV.
This case illustrates other interesting option features. The private sponsors made a "modular"
investment in designing and building the highway to be easily expanded beyond its original four lanes.
Purchasing 250 feet of right-of-way provided the investors with an option to expand without further
land acquisition and environmental permitting. Furthermore, they staged construction of all
interchanges, leaving the final two for future completion. The highway itself might also be considered
a growth option in that the largest equity investor, Shenandoah Greenway, was also a large
landowner in the region. The equity investment in the highway may positively influence local land
values, thereby increasing the broader holdings of the investor. Finally, the private sponsors have an
abandonment option in that they can elect to default on debt obligations and surrender their interest
in the project. While the accompanying transaction costs may prove prohibitive, the option to limit
further losses is nonetheless valuable.
Post Mortem
Early troubles continued. Compounding the Greenway's initial financial problems, the Virginia
Department of Transportation (VDOT) began widening Route 7, a competing free road, in the summer
of 1996. TRIP II and the project's creditors were obviously upset by this government action and
protested it to no avail. Infrastructure Finance reported that many of those at VDOT who were active in
the formative stages of the toll road were, at best, ambivalent about the Greenway project (see
Bailey [1996]). Michael Crane, CEO of TRIP II, commented, "We wouldn't do it as a totally private
infrastructure project, if we had to do it again. These projects are only successful as public-private
partnerships. The developer must have the full support of the state."
Restructuring the debt. With its financial troubles mounting, TRIP II began efforts in earnest during
the latter stages of 1996 and early 1997 to restructure its debt. In March of 1997, the Washington
Post reported that the owners and creditors were close to structuring a new deal. In May, TRIP II
reached an agreement with its creditors to avert foreclosure and forego remaining quarterly interest
payments of $7 million each for the remainder of 1997 in exchange for a toll rate increase on the
facility. At this point, the project had failed to make four quarterly debt payments totaling more than
$28 million. Throughout the remainder of the year, the project continued to struggle, but traffic
volumes did increase.
In October of 1998, the owners filed a plan to restructure the project's debt with the Virginia State
Corporation Commission. The entirely private deal, fashioned by Bear, Stearns & Co., would replace
roughly $250 million of the original bank and insurance company debt with approximately $360
million in insured, zero-coupon bonds that would mature on January 1, 2036, the end date for the
franchise. The proposed plan would minimize debt service for the first 10 years to match reduced
revenue projections while increasing principal and interest payments overall by $575 million to $1.43
billion, with over two-thirds of the payments occurring in the final 20 years. In addition, the
Bryant/Crane family would exit the owners group (see PWF [1998]).
While the state reviewed the plan, TRIP II instituted in November of 1998 a "VIP Miles Frequent Rider
Program" that would pay cash rebates to riders who reached established mileage requirements. By
the following spring, in 1999, the state had granted its approval of the restructuring deal, and the
Greenway had experienced its first 40,000-vehicle weekday. Even though the facility took nearly four
years before it reached its expected initial traffic volume, the timing could not have been better for
the pending bond issue. In April, TRIP II sold approximately $320 million in mostly zero-coupon bonds
in the 144a market, thanks to a boost from insurer MBIA's Aaa rating; Standard & Poor's rated the
issue at triple-B-minus while Moody's rated it as Baa3. All rating agencies cited the project's recent
strong performance (TRIP IIa [1999]). A key condition of the issue, however, according to Fitch IBCA
(now Fitch), was the execution of releases by TRIP II partners of any claims, rights, or damages. In
exchange for claims related to construction and original financing, the owners also offered $56.9
million in first tier and $29 million in second tier subordinated, compound interest revenue bonds (see
PWF [1999]). All three original partners retained ownership of the Greenway. According to Public
Works Financing, the Shenandoah Greenway Corporation and Autostrade International Equity had
invested roughly $105 million during and after construction. In addition, partner letters of credit had
been drawn to $80 million for overdue principal and interest on the original loans.
Brighter horizons? With a new deal for the project's debt and increasing traffic and toll revenues,
the future of the Greenway is not nearly as dismal as it once was. In fact, TRIP II announced plans in
December of 1999 to expand the highway by adding an additional eastbound lane from Exit 6 at
Route 772 to the Mainline Toll Plaza (TRIP IIb [1999]). Estimated cost of the five-mile expansion is
$8.5 million. In the same month, the owners also projected that they would pay roughly $125,000 in
rebates to members of its "VIP Miles" program (TRIP IIc [1999]). Most recently, TRIP II instituted a toll
increase of $0.25 for vehicles with two axles and $0.50 for vehicles with three or more axles in April of
2000 to help pay for the expansion. Tolls for an end-to-end weekday trip on the Greenway now stand
at $2.00 for cars and $4.00 for vehicles with three or more axles; weekend trips are $1.50 and $3.00
(TRIP II [2000]).
CONCLUSIONS
Recognizing and valuing the inherent flexibility of an infrastructure project permits a more robust
approach to making investment decisions. As demonstrated in the Dulles Greenway case, the
application of a simple option valuation model is a valuable supplement to traditional financial
valuation methods and more appropriately accounts for the uncertainty and irreversibility related to
large-scale infrastructure investment. Indeed, in the Dunes Greenway case, such an analysis might
have clued its private sponsors and creditors that waiting and acquiring better traffic information was a
more prudent decision than building immediately. This option is even more valuable when the
transaction costs associated with financial distress (i.e., debt restructuring, additional credit draws,
and equity infusions) and toll rate decreases are considered.
Broader conclusions for private infrastructure development may further be induced from this case.
Flexibility is common in large-scale infrastructure investment. Flexibility describes management
opportunities that are often the key to making strategic investment decisions. The managers of real
assets intuitively act to take advantage of favorable conditions and mitigate the results of
unfavorable situations. While recognized for its strategic importance, this managerial flexibility is often
considered among the "intangibles" when analyzing financial forecasts and choosing among
investments. Traditional valuation techniques using discounted cash flows implicitly assume that real
assets are passively managed, and they fail to account for the value of this flexibility. However, where
precision is not required, relatively simple techniques can be applied to augment traditional methods
in estimating the value of project option features. The emerging application of real option valuation
may prove beneficial to both public and private infrastructure planners in the realization and analysis
of additional project value and the ultimate structuring of development strategies based on this
added value.
ENDNOTE
(*) "Shadow tolls" are a fee charged to private entities (landowners and businesses) along the
transportation corridor that benefit through increased property value and/or increased access. The
term may also apply to public entities (local or state governments) that benefit from a highway project
and subsequently guarantee or subsidize portions of that project to maintain minimum levels of
revenue streams (see Tillman [1998]).
DIAGRAM: EXHIBIT 1 Valuing Assets-In-Place versus Opportunities and Contingencies
EXHIBIT 2 Dulles Greenway Base Case Valuation
Legend for Chart:
A - Year
B - Index
C - Operating & Capital Projections Traffic per Day (# tolled
vehicles) [1]
D - Operating & Capital Projections Toll per Vehicle [2]
E - Operating & Capital Projections Gross Revenue [3]
F - Operating & Capital Projections Operating Expenses [4]
G - Operating & Capital Projections Capital Improvements [5]
H - Operating & Capital Projections Earnings Before Interest
& Taxes [6]
I - Operating & Capital Projections Capital Expenditure [7]
J - Valuation Discount Factor [8]
K - Valuation Discounted EBIT [9]
L - Valuation Discounted Capital Expenditure [10]
A B C D E
F G H
I J K
L
1995 0
(279,000,000) 1.0000
(279,000,000)
1996 1 34,000 2.00 24,820,000
(9,000,000) 15,820,000
0.8880 14,047,841
1997 2 38,760 2.00 28,294,800
(9,450,000) 18,844,800
0.7885 14,859,278
1998 3 44,186 2.00 32,256,072
(9,922,500) 22,333,572
0.7002 15,637,505
1999 4 50,372 2.25 41,368,412
(10,418,625) 30,949,787
0.6217 19,242,879
2000 5 57,425 2.25 47,159,990
(10,939,556) 36,220,434
0.5521 19,997,197
2001 6 65,464 2.50 59,735,987
(11,486,534) 48,249,453
0.4903 23,654,347
2002 7 70,047 2.50 63,917,507
(12,060,861) 51,856,646
0.4353 22,574,913
2003 8 74,950 2.50 68,391,732
(12,663,904) (3,000,000) 52,727,828
0.3866 20,382,837
2004 9 80,196 2.65 77,569,902
(13,297,099) 64,272,803
0.3433 22,062,519
2005 10 85,810 2.65 82,999,796
(13,961,954) (1,700,000) 67,337,842
0.3048 20,525,329
2006 11 91,817 2.65 88,809,781
(14,660,052) 74,149,730
0.2707 20,069,826
2007 12 98,244 2.85 102,198,275
(15,393,054) 86,805,221
0.2403 20,863,299
2008 13 105,121 2.85 109,352,154
(16,162,707) 93,189,447
0.2134 19,888,726
2009 14 112,480 2.85 117,006,805
(16,970,842) (9,400,000) 90,635,962
0.1895 17,176,864
2010 15 120,353 3.00 131,786,612
(17,819,384) 113,967,227
0.1683 19,179,019
2011 16 128,778 3.00 141,011,674
(18,710,354) 122,301,321
0.1494 18,275,979
2012 17 137,792 3.00 150,882,492
(19,645,871) 131,236,620
0.1327 17,414,365
2013 18 147,438 3.00 161,444,266
(20,628,t65) (10,400,000) 130,416,101
0.1178 15,366,923
2014 19 157,758 3.00 172,745,365
(21,659,573) 151,085,792
0.1046 15,808,200
2015 20 168,801 3.00 184,837,540
(22,742,552) 162,094,989
0.0929 15,060,225
2016 21 180,618 3.00 197,776,168
(23,879,679) 173,896,489
0.0825 14,346,824
2017 22 193,261 3.00 211,620,500
(25,073,663) 186,546,837
0.0733 13,666,457
2018 23 206,789 3.00 226,433,935
(26,327,346) 200,106,588
0.0651 13,017,649
2019 24 221,264 3.00 242,284,310
(27,643,714) 214,640,596
0.0578 12,398,985
2020 25 236,753 3.00 259,244,212
(29,025,899) 230,218,313
0.0513 11,809,111
2021 26 253,325 3.00 277,391,307
(30,477,194) 246,914,112
0.0455 11,246,732
2022 27 271,058 3.00 296,808,698
(32,001,054) 264,807,644
0.0404 10,710,606
2023 28 290,032 3.00 317,585,307
(33,601,107) 283,984,200
0.0359 10,199,545
2024 29 310,335 3.00 339,816,279
(35,281,162) 304,535,116
0.0319 9,712,412
2025 30 332,058 3.00 363,603,418
(37,045,220) 326,558,198
0.0283 9,248,119
2026 31 355,302 3.00 389,055,657
(38,897,481) 350,158,176
0.0251 8,805,625
2027 32 380,173 3.00 416,289,554
(40,842,355) 375,447,198
0.0223 8,383,934
2028 33 406,785 3.00 445,429,822
(42,884,473) 402,545,349
0.0198 7,982,095
2029 34 435,260 3.00 476,609,910
(45,028,697) 431,581,213
0.0176 7,599,197
2030 35 465,728 3.00 509,972,604
(47,280,132) 462,692,472
0.0156 7,234,370
2031 36 498,329 3.00 545,670,686
(49,644,138) 496,026,547
0.0139 6,886,781
2032 37 533,212 3.00 583,867,634
(52,126,345) 531,741,289
0.0123 6,555,636
2033 38 570,537 3.00 624,738,368
(54,732,662) 570,005,706
0.0109 6,240,175
2034 39 610,475 3.00 668,470,054
(57,469,296) 611,000,758
0.0097 5,939,671
2035 40 653,208 3.00 715,262,958
(60,342,760) 654,920,197
0.0086 5,653,432
559,725,427
(279,000,000)
Analysis Variables:
Initial Traffic Volume 34,000
Annual Traffic Growth Rate (Index 1-6) 14.00%
Annual Traffic Growth Rate (Index 7-40) 7.00%
Annual Growth Rate of Operating Exp. 5.00%
Weighted Average Cost of Capital 12.62%
Valuation Results: 280,725,427
Net Present Value
Internal Rate of Return 18.27%
GRAPH: EXHIBIT 3 Sensitivity of NPV to Traffic Demand
GRAPH: EXHIBIT 4 Sensitivity of Project NPV to Discount Rate and Average Daily Traffic Demand
DIAGRAM: EXHIBIT 5 Binomial Model for Establishing Risk-Neutral Probabilities
DIAGRAM: EXHIBIT 6 Project Value with Deferral Option
REFERENCES
Amram, Martha, and Nalin Kulatilaka. Real Options: Managing Strategic Investment in an Uncertain
World. Boston: Harvard Business School Press, 1999.
Bailey, Elizabeth. "Driving Up the Learning Curve." Infrastructure Finance, July-August 1996, pp. 9-10.
Brealey, Richard A., and Stewart C. Myers. Principles of Corporate Finance, 5th ed. New York; McGraw
Hill, 1996.
Dixit, Avinash K., and Robert S. Pindyck. Investment Under Uncertainty. Princeton, NJ: Princeton
University Press, 1994.
Finnerty, John D. Project Financing: Asset-Based Financial Engineering. New York: John Wiley & Sons,
1996.
Luehrman, Timothy A. "What's It Worth? A General Manager's Guide to Valuation." Harvard Business
Review, May-June 1997, pp. 132-142.
Miller, John B. Principles of Public and Private Infrastructure Delivery. Norwell, MA: Kluwer Academic
Publishers, 2000.
-----. "The Practical Application of Delivery Methods to Project Portfolios." Construction Management
and Economics, Vol. 17, No. 5 (September 1999), pp. 669-677.
Miller, John B., and Roger H. Evje. "Life Cycle Discounted Project Cash Flows: The Common
Denominator in Procurement Strategy." Proceedings of the 1st International Conference on
Construction Industry Development, Vol. 2, Singapore, 1997, pp. 364-371.
Myers, Stewart C. "Finance Theory and Financial Strategy." Interfaces, Vol. 14, No. 1 (January-February
1984), pp. 126-137.
Pae, Peter. "Drivers Put the Brake on Toll Road's Promise." Washington Post, December 26, 1995, pp.
A1, A18, A19.
-----. "Agreement Buys Time for Dulles Greenway." Washington Post, May 29, 1997.
Public Works Financing. "Credit Enhancer for Dulles Greenway Debt?" November 1998, pp. 6-8.
-----. "Dulles Greenway Refinancing Closed." April 1999, pp. 12-13.
Tillman, Raymond. "Shadow Tolls." Civil Engineering, April 1998, pp. 51-53.
Toll Road Corporation of Virginia. "Financial Model for the Final Financing." Application to the Virginia
State Corporation Commission for Dulles Greenway, Exhibit 3, Case No. PUA900013, 1993.
Toll Road Investors Partnership II, L.P. (TRIP IIa.) "Dulles Greenway Obtains Triple-A Bond Rating in
$350 Million Refinancing Project." Press release, April 29, 1999.
-----. (TRIP IIb.) "Dulles Greenway Constructing Additional Lanes." Press release, December 6, 1999.
-----. (TRIP IIc.) "Dulles Greenway Projects over $125,000 to Be Distributed through Frequent Rider
Plan." Press release, December 28, 1999.
-----. "Dulles Greenway Raises Toll to Aid Roadway Expansion." Press release, March 17, 2000.
Trigeorgis, Lenos. Real Options: Managerial Flexibility and Strategy in Resource Allocation. Cambridge,
MA: The MIT Press, 1999.
To order reprints of this article please contact Ajani Malik at amalik@iijournals.com or 212-224-3205.
~~~~~~~~
By Stephen C. Wooldridge; Michael J. Garvin; Yuen Jen Cheah and John B. Miller
STEPHEN C. WOOLDRIDGE is a major with the U.S. Army, Europe Regional Medical Command in
Heidelberg, Germany.
MICHAEL J. GARVIN is an assistant professor in the Department of Civil Engineering and Engineering
Mechanics at Columbia University in New York. garvin@civil.columbia.edu
YUEN JEN CHEAH is a Ph.D. candidate in the Department of Civil and Environmental Engineering at
MIT in Cambridge, MA.
JOHN B. MILLER is an associate professor in the Department of Civil and Environmental Engineering at
MIT in Cambridge, MA.
top related