BUSINESS (NO PLAGARISM A+ WORK, ON TIME)

Research . Technology Management32 0895-6308/09/$5.00 © 2009 Industrial Research Institute, Inc.

VALUING RISKY PROJECTS WITH REAL OPTIONS

The value of high-risk but potentially high-return technology projects can be calculated with Boeing’s new real-option algorithm.

Scott Mathews

OVERVIEW: This article provides technologists with the business-case methods and tools to calculate the value of projects involving risky new technology or mar- kets but that potentially offer higher returns in the long run. A typical business case using NPV analysis is pre- sented for a new product, an air freighter. NPV is fi rst extended to multi-scenario analysis and then to a “what- if” model using Monte Carlo simulation. Finally, a real- option value for the air freighter is calculated. Based on the same concept as traded fi nancial options, a real op- tion is a contingent investment in “real” physical assets such as a corporate technology project. Boeing’s new real-option value algorithm, the Datar-Mathews Meth- od, is both intuitive and transparent. It gives technology managers an investment and risk-modeling tool they can incorporate into strategic thinking and contingency planning.

KEY CONCEPTS: real options, investments, risk, Datar-Mathews Method, strategic decisions.

Project fl exibility is inherently valuable. Profi tability in- creases with the ability to change the direction of a proj- ect as knowledge is accumulated during the design phase. However, traditional project valuation-modeling methods, such as net present value (NPV), or, as it is oc- casionally termed, discounted cashfl ows (DCF), do not appropriately value fl exibility or quantify risk. A major shortcoming of NPV analysis is that it fails to recognize that management has fl exibility to alter the path of a project, and thus increase overall project value.

In the capital markets sophisticated fi nancial options ac- curately capture the tradeoff between return and risk to refl ect fair market value for securities worth billions of dollars. An options trader would lose his shirt using NPV in the capital markets, so why aren’t options more com- mon in the corporate world? Until recently, option tech- niques were considered too complex to apply to corporate strategic investments. This article introduces a new real options approach developed at Boeing in which these challenges have been resolved through the availability of new spreadsheet add-in software.

Boeing’s real option method provides technologists with investment and risk modeling tools and methods that can be incorporated alongside standard systems engi- neering design modeling techniques. The real option method and its extensions provide a mathematical foun- dation for a scientifi c and “engineering-like” approach to identifying risk and quantifying impact. In turn, they enhance the technology managers’ level of confi dence in reducing risk through targeted allocation of mitigation funds, and help to shape project value outcomes that in- crease the likelihood of achieving strategic objectives.

Business Case for an Air Freighter Project To illustrate a real options valuation, let us fi rst examine a project using a simple business case analyzed from an

Scott Mathews is a Boeing Associate Technical Fellow and technical lead for business engineering within the Boeing research and development division, in Seattle, Washington. He provides technical consulting across Boeing for investment and risk models for new prod- ucts and strategically signifi cant projects. For the past 15 years he has been engaged in stochastic modeling, capital markets investment and fi nancial analysis, and international strategic analysis. He has approximately 20 patents and patents pending in the fi eld. Previously, he worked in the United States, Europe and Asia as an engineer in robotic control systems, artifi cial intelli- gence, and systems and software development. He has a B.S. from Cornell University in electrical engineering and control systems and an M.S. in fi nance from Seattle University. This article is adapted from Chapter 43 of the forthcoming book: The Handbook of Technology Management, 3 Volume Set, Hossein Bidgoli, Editor- in-Chief, John Wiley & Sons, Inc., 2009, Hoboken, N.J. [email protected]

September—October 2009 33

NPV approach. Imagine Boeing has the opportunity to design and build a small aircraft specialized for air cargo transport. This case closely parallels one Boeing asked to be evaluated by a class of students in the Global Inte- grated System Engineering (GISE) graduate certifi cate program at the University of Washington.

There is a rapidly expanding market for high-value goods that are shipped effi ciently from airports near a manufacturer and directly to airports near consumer market outlet locations. Historically, freighter planes were converted from older passenger planes. Ineffi ciency in design and weight of converted freighters makes the cost of transported cargo, as measured by ton-mile, uncom- petitive with cargo transported by truck or by ship. A new specialized freighter might be competitive for transport- ing luxury goods. However, the specialty freighter re- mains a risky investment proposition because the air cargo market is diffi cult to forecast and therefore unit sales and price forecasts can vary greatly depending upon assumptions about the market.

An actual Boeing business case for an air freighter is complex, involving many factors, such as fuel price vol- atility, competition and the global economic environ- ment. However, the valuation concepts presented in this paper can be suffi ciently illustrated with the following simplifi ed business case.

Table 1 sets forward example projections of revenues and costs using a most-likely NPV scenario based on assumptions about product strategy and market recep- tion. The engineers and marketing analysts are re- questing authorization to spend $75M (million) in R&D expenses over the next three years. The objec- tive is to reduce uncertainty in order to be better in- formed about whether to launch the product. The funding will enable the engineers to develop a prelim- inary design and better determine recurring and non- recurring costs. Meanwhile, the marketing analysts will obtain a better estimate on the unit quantity and price of the air freighter. After three years, contingent

on the success of the engineering efforts and a promis- ing market forecast, Boeing would commit a one-time expenditure of $2.0B (billion) to launch the air freight- er. This signifi cant launch cost provides funds for fi nal design, production facilities, FAA certifi cation and marketing outlays.

The strategic investment question facing senior manage- ment in this example is whether to authorize the $75M R&D expenditure. Over the last half-century or so, cor- porations have answered that question by applying NPV analysis to determine whether project net profi ts exceed the initial investment. NPV of the net profi ts is common- ly based on the value estimates for the business case variables refl ecting the most likely outcome of assump- tions about the product strategy and market reception (1). NPV analysis typically discounts all cashfl ows at a common rate: the Project Risk rate, a rate established by management to meet a required rate of return for project investments. This rate is sometimes referred to as a hur- dle rate. Applying the Excel NPV function with the Proj- ect Risk discount rate of rp = 15 percent, the project net present value at Year 0, today, is estimated to be a nega- tive $187M (see Table 2).

Under standard NPV decision-making, senior manage- ment would deny the request for initial funding of $75M because the freighter project is forecasted to lose money. Following that guidance, Boeing most likely would ter- minate the project, and engineering and marketing re- sources would be directed to other projects with positive NPV values. The termination of the R&D efforts would essentially preclude any near-term participation in the freighter market. Without consideration for strategic in- vestments, the company is left unprepared should a plausible but lower probability upside scenario actually materialize, potentially leaving an opening for a com- petitor. Some managers might be tempted to declare the freighter project “strategic” and invest anyway in order to preserve the opportunity to explore the market potential. Of course, doing so would circumvent the discipline of fi nancial decision-making.

Table 1.—Most Likely Business Case

($M)

Year

0 1 2 3 4 5 6 7 8 9 10

Most-Likely

Target Unit Price $34 Unit Cost $38 $30 $25 $22 $21 $19 $19 Unit Quantity 15 30 45 60 60 60 60 Revenues $510 $1,020 $1,530 $2,040 $2,040 $2,040 $2,040 Recurring Costs $575 $888 $1,133 $1,340 $1,238 $1,167 $1,114 NPV Op Profi ts $0 $0 $0 ($65) ( $132) $397 $700 $802 $873 $926 NPV Launch Cost $0 $0 ($2,000)

Research . Technology Management34

Given the uncertainty of the market forecast for unit sales and price, there are good reasons for the project manager to be skeptical about a decision restricted to an analysis of a single most likely scenario, such as this NPV business case. This limiting approach is further re- inforced by spreadsheet formatting that constrains each cell to a single value. Consequently, a business case spreadsheet typically refl ects a single scenario, while eliminating other less probable though still plausible scenarios. The result is that a standard NPV analysis does not provide insight into the business case opportu- nities and risks at the margin, nor does it take into con- sideration management’s ability to respond and take advantage of contingencies.

NPV mathematics, having originated in the banking in- dustry for use in calculating interest on savings accounts, unfortunately are misapplied, especially when used to analyze projects that have high uncertainty. The result is that NPV analysis tends to justify only projects that are more conservative (bank-like), where uncertainty is in- consequential, the investment amount and timing is es- tablished, and the near-term outcome is more certain.

Multi-Scenario Approach To extend the NPV analysis beyond a single most-likely scenario, a common practice is to assess several scenarios. Typically, two additional scenarios are considered: an op- timistic and a pessimistic one, which refl ect different out- comes, more or less favorable, of the assumptions applied in the most likely case. These two scenarios are plausible, although lower probability than the most likely case.

To reduce the complexity in creating viable scenarios, the emphasis should be on identifying those half-dozen or so assumptions that are potentially consequential or impact 10 percent or more of the total value of the project. Fac- tors impacting less than 10 percent of the project value are assumed to be manageable by a good project team, analo- gous to an insurance deductible, where minor scenario excursions do not pose a perilous risk. This is in line with standard engineering practices for a fi rst- or second-draft design effort where the focus is on those components that contribute a substantial portion of the targeted function.

Under standard NPV decision-making,

senior management would deny the

request for initial funding.

Begin the multi-scenario modeling process by envision- ing three scenarios: optimistic, most likely, and pessimis- tic, such as Table 3. The most likely scenario estimates are premised on high-likelihood outcomes of assump- tions, or contingencies, that a majority of experts (tech- nology, engineering, marketing, fi nance, management) anticipate materializing, including engineering technical challenges and marketing response, all of which have quantifi able impact on cash fl ow estimates. The optimis- tic and pessimistic scenarios respectively are derived by positively and negatively challenging the assumptions.

Each scenario—optimistic, most likely and pessimis- tic—results in an operating profi t forecast that corre- sponds to a plausible outcome within the market. Figure 1 shows a graph of the three operating profi t scenarios as a cone of uncertainty, representing the range of variation of future events. There are also different scenarios for the launch cost. Table 4 provides the complete optimis- tic and pessimistic business case scenarios.

Calculations for optimistic and pessimistic project NPV values are $2,099M and ($1,568M) respectively. Given

Table 3.—Variables for Various Scenarios

Experts Variable

($M) Optimistic Most

Likely Pessimistic

Engineering Unit Cost $18 $20 $23 Engineering Launch Cost $1,500 $2,000 $2,500 Engineering Production

Ramp 15–30/

Year 15/Year 15/Year

Marketing Unit Price $40 $34 $30 Marketing Unit Quantity 455 330 270 Finance Discount Rate 15% Project; 5% Investment

Range Management Outlook 10% 10% Technology R&D Costs $75 $75 $75

Table 2.—NPV Most Likely Results

Discount Rate Assumptions

Project Risk Rate, rp 15.0%

NPV Calculations ($M)

PV0 NPV Operating Profi ts $1,203 PV0 NPV Launch Costs ($1,315) R&D Expenses ($75)

Total Project NPV Value ($187)

September—October 2009 35

the wide disparity between the outcomes of the two sce- narios with their respective contingencies, we under- stand there is substantial uncertainty in this business case. Clearly, the optimistic scenario has a signifi cant profi t opportunity. On the other hand, the pessimistic scenario forecasts market and engineering situations that result in substantial losses. Although there is only a 10 percent chance of either sce- nario occuring, this additional information doesn’t pro- vide any more clarity on whether senior management ought to commit the $75M R&D investment. What does become apparent, however, is that in the optimistic sce- nario, it is worthwhile in Year 3 for management to com- mit the one-time launch cost of approximately $2B. On the other hand, in the pessimistic (and the most likely) scenario, management should consider abandoning the project at Year 3, or at least delay committing the launch cost, to avoid substantial losses. Which scenario will manifest itself will only become ap- parent by committing the upfront R&D investment and initiating the preliminary engineering and marketing ac- tivities. By Year 3 management will have obtained the information necessary to make a better decision. We will see how option analysis by Monte Carlo simulation does blend these three scenarios to provide the investment in- formation management seeks.

“What-If ” Multi-Scenario Modeling Just as Monte Carlo simulation technology at Boeing ex- tends the ability to investigate “what if ” scenarios on matters of engineering concern where there is large un- certainty (such as performance, quality and operating sta- bility), this same technology can be applied to scenarios

for price, unit sales and cost for the air freighter business case. In fact, the term “business engineering” can be ap- plied to the more advanced models, which combine both the business and engineering aspects of a project. Monte Carlo simulation offers the ability to incorporate into the analysis hundreds of scenarios, including those that are plausible albeit lower probability, but potentially conse- quential to the outcome of the project such as the opti- mistic and pessimistic scenarios (see Table 5).

We can develop a valuation approach that effectively uses this technology and provides a more useful estimate for the project than the NPV multi-scenario approach. The estimates of the annual operating profi ts of the three sce- narios can be interpreted as representing the corners of a triangular distribution (2). Using Monte Carlo software, it is relatively straightforward to construct a range fore- cast triangular distribution for each year of the operating profi t forecast. The annual values for the minimum (pes- simistic), most likely, and maximum (optimistic) param- eters form a triangular distribution range forecast for each year as shown in Figure 2. Similarly, the Year 3 launch cost also can be presented as a triangular distribution.

When the model is simulated, the Monte Carlo makes successive random draws of values, or “what-ifs,” from the annual operating profi t and launch cost distributions. The simulation of successive random values is an emu- lation of the range of values of scenario forecasts, each with variations of the project assumptions. The Monte Carlo simulation generates a succession of “what-if ” net profi t scenarios (a minimum of 500 draws or “trials” is recommended), each of which is valued using NPV mathematics, and is treated as a plausible cash fl ow forecast.

Figure 1.—Three operating profi t scenarios are envisioned.

Research . Technology Management36

Datar–Mathews Real Option Valuation Model The relatively recent (2001) Datar–Mathews option pricing model attempts to provide a more transparent and intuitive approach to valuing project options. The D–M Method (U.S. Patent 6862579, © Boeing), as it is called, can be understood as an extension of the NPV multi-scenario Monte Carlo model with an adjustment for risk-aversion and economic decision-making.

The D–M Method uses two discount rates: 1. rp = 15%, the hurdle rate, for the project operating profi ts, those cashfl ows that are at market risk. 2. rf = 5%, for investments that are relatively secure and over which the corporation has fairly extensive con- trol, such as the launch cost. In advanced fi nancial analyses, it is fairly common to discount various cash fl ows in accordance with the risk

to those cash fl ows. The differential discount rate implic- itly allows the D–M Method to correctly ‘risk-adjust’ the project value, accounting for the differing risks (3) with- in the project. The operating profi ts and the launch cost distribution ranges are both simulated with Monte Carlo, and discounted to Year 0. The net profi t is the difference between the two discounted cashfl ows (OperatingProfi ts − LaunchCost). The result is the Year 0 net profi t present value distribution (histogram) for the hundreds of cash fl ow scenario trials shown in Figure 3. The project manager must be fi nancially rational for the investments in the air freighter project (see Figure 4). The Terminated Outcomes section on the left tail of the present value distribution represents those scenarios in which the discounted launch cost is anticipated to ex- ceed the operating profi ts. In these instances of negative net profi t, the rational choice is to avoid potential sub- stantial losses by terminating the project. The effect is as

Table 4.—Optimistic and Pessimistic Business Case

($M)

Year

0 1 2 3 4 5 6 7 8 9 10

Optimistic Target Unit Price $40 Unit Cost $34 $27 $22 $19 $17 $16 $15 Unit Quantity 15 30 60 80 90 90 90 Revenues $600 $1,200 $2,400 $3,200 $3,600 $3,600 $3,600 Recurring Costs $517 $800 $1,311 $1,531 $1,569 $1,468 $1,394 Optimistic Op Profi ts

$0 $0 $0 $83 $400 $1,089 $1,669 $2,031 $2,132 $2,206

Launch Cost $0 $0 ($1,500)

Pessimistic Target Unit Price $30 Unit Cost $44 $34 $29 $26 $25 $23 $22 Unit Quantity 15 30 45 45 45 45 45 Revenues $450 $900 $1,350 $1,350 $1,350 $1,350 $1,350 Recurring Costs $661 $1,022 $1,303 $1,184 $1,107 $1,051 $1,007 Pessimistic Op Profi ts

$0 $0 $0 ($211) ($122) $47 $166 $243 $299 $343

Launch Cost $0 $0 ($2,500)

Table 5.—Three-Scenario Monte Carlo Setup

($M)

Year

0 1 2 3 4 5 6 7 8 9 10

3 Scenarios Optimistic ($1,500) $83 $400 $1,089 $1,669 $2,031 $2,132 $2,206 Most Likely ($2,000) ($65) $132 $397 $700 $802 $873 $926 Pessimistic ($2,500) ($211) ($122) $47 $166 $243 $299 $343 Cost-Monte Carlo $0 $0 ($2,000)

September—October 2009 37

Figure 2.—Monte Carlo range forecasts are fi tted to the optimistic, most likely and pessimistic scenarios: top, middle and lower curves, respectively.

if operating profi t and launch cost cash fl ows were ze- roed out for these scenarios. On the other hand, the solid section on the right tail of the present value distribution indicates a successful project forecast. The right tail cor- responds to the 32 percent of scenarios with a positive NPV outcome where the discounted operating profi ts exceed the launch cost. The real option value can be understood as the average of the appropriately discounted Year 0 net profi t, contin- gent on terminating the project if a loss is forecast. The payoff distribution illustrates that 68 percent of the sce- narios are terminated with zero cash fl ow, while the re- maining scenarios yield a range of positive net profi ts (Figure 5). The real option value, which is easily gener- ated by the Monte Carlo software, is the average value of this payoff distribution, approximately $113M in this example. This value is the best estimate today of the dis- counted future expected net profi t, conditional on ratio- nal decision making at the time of launch.

The $113M option value is the maximum amount the company would be willing to pay today for the oppor- tunity to participate in the air freighter project. The company could choose to invest the total amount or a portion into its R&D effort. The engineers and market- ing analysts are requesting authorization to spend $75M in R&D expenses over the next three years. Based on the real option value of the business pros- pects, management has the justifi cation to invest in the freighter project. The real options approach delays the “go forward” or “terminate” decision until Year 3, when there is more information and a better decision can be made. Furthermore, it appears that management can purchase the freighter option ($75M) for less than its calculated value ($113M), an immediate boon to the shareholders. The formal calculation of the real option value is done using the Boeing Datar–Mathews Method (4). The spreadsheet D–M Method formula is as follows:

Figure 3.—Year 0 net profi t present value distribution. Figure 4.—Year 0 risk-averse rational decision distribution.

Research . Technology Management38

Re

operating profits launchcost ,0)

al option value

Mean[MAX( ]

=

−

The formula, which is a combination of Excel and Mon- te Carlo functionality, captures the intuition described above. The overscore bar in the equation represents a distribution—formally a random variable—of the dis- counted cash fl ows at time 0. The “operating profi ts” and “launch cost” are the present value distributions. The payoff distribution is created by simulating several hun- dred scenario trials, and calculating the MAX value, with a zero threshold for terminated projects represent- ing no cash fl ow. The option value is the mean value of this payoff distribution. An intuitive understanding of real options is useful dur- ing strategy discussions. An estimator for the real option value can be expressed as a function of positive NPV project outcomes in the following formula:

Real option value = +NPV Risk Adjusted Probability × (Operating Profi ts − Launch Costs)

For example, in Figure 4 the risk-adjusted probability of positive NPV forecast is 32%, and the appropriately discounted mean net profi ts value (operating profi ts − launch costs) of the successful outcomes is $0.35B. Using these values in the above formula produces a real option value of the project given its contingencies:

Real option value = 32% × ($0.35B) ≈ $113M A real options approach gives management the justifi ca- tion to commit a contingent strategic investment in tech- nology, engineering and marketing R&D prior to the launch. These funds will enable the engineers and mar- keting analysts to advance the air freighter project to a state of readiness in preparation for the launch decision, while effectively reducing the uncertainty of that deci- sion. Contrast this result with that of the NPV (a nega- tive $187M) approach that would terminate the project even before initiating the R&D effort (5).

Real options methods work for strategic decisions be- cause of their ability to simplify and manage complex investment problems, such as those at Boeing. The D–M Method has been used to get a better sense of value on some of Boeing’s largest projects to date. A simplifi ed version (see “Range Options,” next page) is the basis of value assessment for a portfolio of early-stage innova- tive business opportunities in one division of the com- pany. Generally, it is not possible to know all of the potential factors that might affect the outcome of such in- vestment. But it is suffi cient in an uncertain environment to bound the problem, yet remain confi dent in the deci- sion-making process. By acquiring the initial resources and information necessary for informed decisions, real options allows us to make better decisions at a future date while concentrating scarce investment resources on those truly promising opportunities. Real options think- ing is an approach to project planning that extends the investment and risk modeling tools and methods to pro- vide engineers with a solid fi nancial economic construct to incorporate into strategic thinking and contingency planning.

The Datar-Mathews option pricing model

has provided a better sense of value on some of Boeing’s largest projects.

Figure 5.—Year 0 payoff distribution.

September—October 2009 39

Range Options: Option Values without Simulation In very new business ideas we still would like to calculate an approximate real option value even though there has not been suffi cient time or resources to gather the necessary quantitative information required for a complete cash fl ow simulation. An approximate option value can be estimated simply using range estimates of the present values of operating profi ts and launch costs.

Assume we are able to estimate the present value (PV0) of the operating profi ts and launch cost ranges (pessimistic, most-likely, and optimistic: Min, ML, Max, respectively), correctly observing the differential discount rates. In Figure 6A, below, two triangular distributions have been constructed from the present value ranges, one for the operating profi ts and the other for the launch cost. I shall illustrate this example using the approximate values from the preceding air freighter case.

The real option value is the expected value of the mean of the tail of a net profi t distribution. The net profi t distribu- tion is the difference between the operating profi t and the launch cost distributions. Assuming that the operating profi ts and the launch costs are independent (as are most business cases), we can approximate the magnitude of the launch cost by “collapsing” the triangular distribution to its mean value, a determined or “scalar” value. The imputed value for the net profi t distribution is then calculated as the operating profi t distribution minus the launch cost mean value shown in Figure 6B.

The real option value is calculated as the mean of the right tail (i.e., those values greater than zero) of the net profi t triangle, times the probability of the right tail. The Table below provides the formulas for the mean and prob- ability of a simple right triangle tail. (Other types of triangular sections, such as left tail quadrilateral, require somewhat more calculation effort.)

The range option value is a good approximation of true value of the business opportunity. The range option value will never quite equal that of a correctly calculated real option owing to the asymmetric nature of option calcula- tions that are only fully captured with a simulation. However, the result is a real option value derived from simple range estimates of early-stage project operating profi ts and launch cost.—S.M.

Figure 6A.—Present Value Distributions. Figure 6B.—Range Option Concept.

Research . Technology Management40

Once the path or scenario is identifi ed, the second factor is a contingent course of action associated with the sce- nario that preserves the original intent of the option val- uation (see Table 6). If the pessimistic scenario is actualized, the manager can conserve the substantial launch cost investment, terminate the project immedi- ately, and perhaps sell off any derived patented assets. If it is the optimistic scenario, the manager can invest the launch cost and garner the expected operating profi ts. If the actuality is the most-likely scenario, the manager may determine that it is worthwhile to delay the launch, and perhaps invest additional R&D funds to preserve the opportunity while also attempting to reduce the uncer- tainty further in order to make a clear decision later.

References and Notes 1. In most corporations, business cases applying NPV analysis typically will use, given a list of assumptions, the most-likely (technically the mode of the distribution) values estimates for cost, market price and volume, etc. However, NPV analysis correctly should use the mean value estimates. The diffi culty is that most corporations do not have easy access to a suffi cient quantity of historical actuals from which to derive a mean value. Additionally, by defi nition a new technology has no historical actuals. 2. Most risk distributions are skewed, including the triangular distributions used in the case. A skewed distribution captures the risky project concept of a low-probability but high-consequence phenomenon. Distributions other than triangular can be used. A lognormal distribution, used in formal options theory, is a type of skewed distribution, but its defi ning parameters, such as mean and standard deviation, are more diffi cult to determine in the context of standard engineering and business practices. The easily comprehensible parameters Max—Most Likely—Min that defi ne a skewed triangular distribution can more or less approximate the more formal lognormal distribution without material impact on decision outcomes. 3. Note that the risk adjustment is the result of differential discounting, effectively shifting the relative values of the operating profi t and the launch cost distributions discounted to Year 0. The launch cost cash fl ow is more highly valued, and therefore discounted at a lower rate, because it represents “cash on hand,” versus the operating profi ts that are anticipated but not guaranteed. This natural risk-averse perception of our world is captured in the adage, “A bird in the hand is worth two in the bush.” 4. The Datar-Mathews and the well-known Black-Scholes option methods are mechanically different representations of the same

Strategic Thinking and Contingency Planning Much of the worth of real options resides not in the ac- tual calculation of the option value, but rather in what is termed “real options thinking.” That is the application of real options logic without necessarily carrying out the detailed calculations. Savvy Boeing managers conduct their project management using processes and planning remarkably similar to real options thinking. In addition, when they articulate the challenges of the project using the language of real options, they provide structure for scenario and strategy discussions. The real options planning approach of a phased series of incremental risk-averse investments linked to the probability of a successful outcome contrasts with that of NPV-driven planning, which tends to commit large dollar amounts up front to a single course of action.

In the technology world, here are a few rules of thumb when we ought to apply real options thinking:

Uncertainty is large enough that it is sensible to wait • for more information.

Uncertainty is large enough to make fl exibility a • consideration.

The investment decision is contingent on material • assumptions or events.

The technology or product profi tability is not in the • current offering, but a future extension or derivative product where there are possibilities for future growth.

Project updates and mid-course strategy corrections • are anticipated. The real option-driven approach to project planning is tied to two key factors: an initial investment directed to reducing uncertainty followed by a contingency-based course of action. The fi rst factor involves targeting of small investments toward engineering and marketing risk abatement initiatives prior to the launch commit- ment. As a result, at the downstream decision point of the irreversible launch investment, the project manager will be able to better determine which project scenario (optimistic, most likely, or pessimistic) is being borne out in reality.

The language of real options provides

structure for scenario and

strategy discussions.

Table 6.—Course of Action, Contingent on Scenario Actualized at Launch Date

Actualized Scenario

Year 3 Course of Action

Optimistic Launch Immediately, Receive Operating Profi ts

Most Likely Delay Launch, Additional R&D Investment

Pessimistic Terminate Program

September—October 2009 41

underlying economics, and, given the same constraints, calculate the same option value. The D-M method provides a better estimate of option value when the strict theoretical assumptions of Black-Scholes are compromised in real life. For example, the D-M method can easily deal with non-lognormal (such as triangular) cash fl ow distributions and a launch cost that is a range rather than a scalar value. More detail is provided in Chapter 43 of the forthcoming Handbook of Technology Management. 5. Distinguishing between strategic and tactical investment decisions provides a further rationale for using real options. Strategic decisions are risky because the investment resources are committed up front while the outcome benefi ts are far from certain. Having the option to cancel the project, if warranted, signifi cantly reduces the cor porate exposure to the tactical launch investment decision risk. This risk- lowering option practice enables companies to take on smaller, higher-risk but potentially higher-return projects while maintaining fi scal respon sibility. In comparison, tactical decisions are made by fully commit ting whatever resources and information are on hand at the decision moment. For example, the launch commitment at Year 3 is a tactical decision, where the substantial launch investment is irreversible. It is worthwhile to point out a subtlety about the fi nancial risk of the project—there is no guarantee of positive net profi ts. The purchase of an option is a strategic decision, such as investing in R&D, which permits preliminary participation in a business venture that can be reevaluated later. A real option valuation does not preclude conditions at launch time from changing, necessitating a re-valuation of the prospective project profi tability, or that the launch decision will be fi nancially risk-free. Exercising a real option on a project nearly always exposes management to the subsequent tactical decision of whether or not to invest the signifi cant launch costs in the risky underlying project/product asset. In contrast to a real option, when exercising a fi nancial call option, the owner can eliminate the tactical risk of possessing an asset that might decline in value by using a so- called “cash settlement” on the exercise date (by simultaneously selling the equivalent shares of stock). It is a tactical decision whether or not to “exercise” the project option and commit the substantial launch investment. An NPV analysis of the business case at the launch decision date can evaluate the proper conditions to exercise the option, i.e., the prospect for posi- tive net profi ts with an appropriate discount rate to account for the risk. However, the evaluation does not preclude the possibility for

conditions to change at a future date (perhaps an unanticipated souring of the air freighter market), and eventually result in an overall profi t loss.

To Learn More

Mathews, Scott H., Datar, Vinay T. and Johnson, Blake. 2007. A Practical Method for Valuing Real Options. Journal of Applied Cor- porate Finance, Spring (19), No. 2, pp. 95-104.

Mathews, Scott H. and Salmon, Jim. 2007. Business Engineering: A Practical Approach to Valuing High-Risk, High-Return Projects Us- ing Real Options. Tutorials in Operations Research. INFORMS.

Websites

Annual International Conference on Real Options: Theory Meets Practice. www.realoptions.org/ The most important annual event on real options, organized by Real Options Group. A great collection of important papers to download from current and past conferences makes this website one of the main references for real options re- searchers and practitioners.

Investment Science. The purpose of this site is to promote and discuss modern investment analysis, ideas and techniques as applied to the pricing and management of real assets, or what has come to be known as Real Options. By Prof. David G. Luenberger of Stanford Univer- sity’s Department of Management Science and Engineering. www. investmentscience.com. See especially this site which discusses The Two-Rate Method of Discounting: http://investmentscience. com/Content/newsArticles/news3.html.

Luehrman, Timothy A. His series of articles in Harvard Business Re- view presents a framework that can bridge the gap between the prac- ticalities of real-world capital projects and real options for the general business audience. What’s It Worth?: A General Manager’s Guide to Valuation, May 1997; Investment Opportunities as Real Options: Getting Started on the Numbers, July 1998; Strategy as a Port- folio of Real Options, September 1998.

Real Options: Managing Strategic Investment in an Uncertain World. A website that supports the book of the same name. Extensive refer- ences. www.real-options.com/

Reprints

FINANCIAL MANAGEMENT OF R&D Forty-two RESEARCH • TECHNOLOGY MANAGEMENT articles on this subject are now available in paper- back. To order, see inside back cover.

The Role of Core Competencies in the Corporation Financial Management of R&D 2002 3000 Raw Ideas = 1 Commercial Success! Linking R&D to Growth and Shareholder Value Traps, Pitfalls and Snares in the Valuation of Technology Risk-Adjusted Valuation of R&D Projects Assessing the Value of Your Technology Evaluating R&D Performance Using the Cost Savings Metric Evaluating R&D Performance Using the New Sales Ratio Applying ‘Cost of Innovation’ to Technology Planning

R&D In An EVA World When Choosing R&D Projects, Go With Long Shots Cost Modeling as a Technical Management Tool Financial Analysis Extends Management of R&D Allocating R&D Resources: A Quantitative Aid to Management Insight The Role of Risk in Business Decision-Making, or How To Stop Worrying and Love the Bombs Resolving Uncertainty in R&D Portfolios How To Manage Risk Better AND MORE

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.