For the past decade, a debate has been raging among finance practitioners involved in investment decision making. A few have been advocating something called real options as an alternative to the more conventional and well-known discounted cash flow techniques, both for valuing investment opportunities and for choosing among such opportunities when faced with budget constraints. Over the past few years, I have had the good fortune to work not only in large companies where conventional techniques are well established, but also with leading thinkers in the area of real options analysis. And after having been exposed to both sides of the real options debate, I have ended up becoming a proponent of real options—of both its way of thinking about corporate strategy and, where appropriate, its technique for quantifying the value of corporate assets and strategies.
In this article, I discuss why the usual objections to real options may be misguided. But by way of background, I will start with a brief detour highlighting certain experiences that have shaped my thinking.
In the early 1990s, I was an intern at the Inkjet Business Unit at Hewlett-Packard. There I was introduced by a creative HP manager to a concept called “postponement.” The idea behind postponement is that it is better not to fully customize a final product until the last moment—that is, until the consumer is ready to pick it up. The managers responsible for the decision did not call it real options, but they clearly understood the value of postponing certain aspects of the manufacturing process. In real options parlance, they decided to extend the maturity of the option.
The product in this case was an inkjet printer in a box, including manuals, power cords, and all the other accessories. Once a printer had been put in a Japanese box with a Japanese manual and a 100V power supply, it could be sold only to customers in the Far East—at that point HP had lost the value of any “postponement” options. The alternative approach, which HP began to put in place in early ’90s, was to standardize parts and assemble them in stages, reaching higher and higher levels of customization only as they approached the eventual sale. Such progressive customization had the benefit of keeping the postponement option alive (until it was optimal to exercise) while also increasing standardization upstream, which provided scale advantages in manufacturing. This simple idea of postponement also led to a complete rethinking of HP’s supply chains, including the location of manufacturing and assembly plants as well as warehouses.
My second formative experience with real options came while I was at Deloitte Consulting in the mid ’90s. One of my clients was a software company that was developing software to help customize products in the computer hardware industry. By making the software specific to computer hardware companies, the client could develop the product faster, provide “hard-wired” custom features, and go to market sooner with a product tailor-made for the targeted segment. Alternatively, they could develop a more generic software as a “platform” product that could be used across a variety of industries—from furniture manufacturers to airlines. This platform product could then be customized as appropriate to the needs of the ultimate customer.
After considering the choices, the company settled on investing in the platform technology and preserving the “option” to move into a variety of industries in the future. This decision, although somewhat delaying the market entry of the new software, significantly broadened its market potential. Again, the term “real options” never surfaced during the debate and analysis that led up to this decision; but the telltale signs of real options thinking were there.
My third experience came while I was at a large pharmaceutical company in the latter part of the ’90s. The company was considering acquiring a prototype molecule from a biotechnology company. The molecule was in the very early stages of R&D, and the question was how much it was worth. Since these transactions are quite common in the biotech industry, one would assume that similar questions had been asked and answered numerous times in the past and that formal valuation techniques had been developed as a result. But this was not the case. Of course, companies engaging in these types of transactions have developed some “rules of thumb” based on trial and error and past experience. They go something like this:
A discounted cash flow analysis of the costs, revenues, and expected in-licensing terms (at the pharmaceutical company’s weighted average cost of capital) indicated that the molecule had a negative NPV. But this result contradicted what our intuition was clearly telling us. How could a product that showed such promise and had no legal issues have a negative value with very reasonable licensing terms? In fact, we could not get a reasonably positive NPV even when assuming no milestone payments or royalties to the inventor. Baffled by this result, we looked for answers.
The literature was already rich in real option theory and techniques at that time. So we formulated a partial differential equation and solved it using finite difference techniques and established a fair value for the transaction. Unfortunately, we were not able to complete the analysis in time to make the decision. But this experience persuaded me to take a more serious look at real options analysis.
Over the past several years, I have spent a lot of time talking with academics, consultants, and corporate executives about how best to provide insights into making better decisions. There are debates among academics as well as practitioners about the usefulness of real options techniques in investment analysis. In the past couple of years, we have seen no fewer than 20 major conferences focusing on real options. And while it appears that the interest in real options techniques is increasing, the debate goes on and there continues to be a great deal of skepticism. I have heard many arguments against real options and will now respond to the ones I encounter most often.
This statement brings to mind the various debates about the “efficient markets” hypothesis since the theory was first put forward in the 1960s. The question of whether markets are efficient cannot be answered by a simple “yes” or “no.” Clearly, it is the degree of efficiency that is at issue. Anyone who believes that large equity markets are not at least “weak-form” efficient is effectively assuming that the market-pricing process regularly provides opportunities for investors to earn above-market returns on the basis of publicly available information—a proposition that, given the intensity of competition among money managers and other investors, seems hard to accept. In fact, there is an abundance of academic evidence—including the well-documented failure of the vast majority of fund managers to outperform market averages—that suggests markets are reasonably efficient.
Similarly, the question of whether real options techniques work or not does not have a simple answer. The best way to think about this is that the real options framework provides a generalized asset pricing methodology, of which the more conventional techniques like DCF are special, simplified cases. In situations with little variability in expected outcomes and no flexibility in future decision choices, conventional techniques like DCF are adequate. But once some of the assumptions that underlie the traditional techniques are no longer applicable, it makes sense to move toward the more generalized real options framework.
Again, this is not a black or white issue. There are classes of decisions in every industry that require a generalized framework like real options analysis. But it would be a mistake to move into a generalized framework when the assumptions underlying traditional techniques are valid. In other words, complicating the analysis unnecessarily does not add value.
There also appears to be a lot of confusion about how to evaluate “private” or technical risks (for example, the possibility that toxic effects are revealed in a Phase I experiment) as distinguished from “market” risks (such as the possibility that a recession reduces the demand for “lifestyle” drugs). One common argument is that real options techniques do not work when technical risks dominate market risks. It is true that when the value of an investment is heavily dominated by technical risks, the managers of the investing company have much less flexibility. And it is also true that in cases where there is little flexibility, traditional techniques will provide an adequate way of estimating value. But the point to keep in mind here is that technical risks are treated the same whether you use real options or the plain old DCF approach. It is only the market risks that are treated differently. But one has to be careful in concluding that market risks do not matter before analyzing the problem. It is important to remember that the product has to be sold in the market to make any money!
Let me provide at least two types of problems in the pharmaceutical and biotechnology industries where a generalized real options framework is necessary:
(a) Pricing a molecule, technology, or intellectual property (IP) in the course of transactions among companies.
(b) Value-based management of a portfolio of investment opportunities, including discovery, development, marketing, and infrastructure investments ranging from information technology to manufacturing.
The real options framework has already been successfully applied to both types of problems, and so we are beyond the feasibility testing stage here. Further refinement and customization of the real options approach can add significant value in these industries, which are currently struggling to cope with changing business environments. Real options can transform industry thinking and lead to more favorable overall economics in drug making.
This is a slippery slope. One can also question whether we have proof that markets are pricing assets by discounting cash flows according to the Capital Asset Pricing Model. Many studies show that only a small percentage of the total value of certain types of companies (such as so-called “information” and “knowledge” firms) can be explained by DCF values. Much of the rest of the value comes from so-called “growth options,” or investment opportunities. For pharmaceutical companies, this growth option value represents as much as 70% of market cap. So, there is no doubt that the market is capable of appreciating value beyond what can be assessed using traditional DCF techniques. But the extent to which market values reflect real options value is, of course, an open question—one that is likely to be difficult to resolve.
Even so, let’s consider what happened in the case of financial option pricing models. Many of the early users of the Black-Scholes formula for stock options, eager to make money with the new technique, came to the painful conclusion that markets were already pricing options in essentially this way. What the Black-Scholes formula did was to model an existing market pricing process and thereby institutionalize it. In similar fashion, real options analysis should be viewed as providing a framework—one that is lacking in traditional corporate finance—that replicates and quantifies sound management intuition about the kinds of investments that are most likely to create value over the longer term.
Some companies fear that if they adopt real options approaches in decision making, their stock prices will suffer. Managers of these companies tend to believe that markets are pricing assets using a highly simplistic DCF model—one that effectively capitalizes just the next quarter’s earnings. But, to the extent we can judge from the high P/E ratios of many growth firms, and from the willingness of investors to assign large values even to companies without earnings, this view clearly seems mistaken. What’s more, given the current concerns about corporate governance, those managers who take shortsighted actions to boost quarterly EPS are most likely to find themselves being punished by the market. Next quarter’s earnings are important, to be sure, but not as important as the company’s ability to demonstrate that it has a coherent strategy based on building and maintaining its long-run competitive advantages. To the extent conventional capital budgeting techniques are preventing managers from thinking strategically and encouraging them to use tactics for meeting near-term earnings targets (originating from a simplistic NPV model), the long-term impact on the company will certainly be negative. Managers need to adopt real options thinking and, equally important, find effective ways to communicate their thinking to help investors understand the value of the company’s portfolio of growth options and opportunities.
This depends on what one means by “complicated.” It is certainly true that Texas Instruments does not market a real options calculator, nor can one put a real options formula into an EXCEL worksheet (yet). But there are complicated calculations in engineering, too, where detailed modeling of aircraft parts relies on techniques such as finite element analysis and computational fluid dynamics that are certainly not available on a calculator or in EXCEL. Yet large aircraft manufacturers do not shy away from these techniques.
Corporate investment decisions should be evaluated using the best available techniques. An accurate assessment of value can make a difference, and a technique that provides a more accurate quantification should not be ignored. Of course, an analysis doesn’t have to be complex to be correct; but perceived complexity in a technique is not sufficient grounds for dismissing it. There are many ways to make real options models more user-friendly. But such models will never become as generic as an EXCEL formula because real options techniques require careful framing of the problem, not just a forecast of expected cash flows followed by rote discounting.
I don’t know of any good senior executive who would approach an investment problem as a now-or-never proposition with no future flexibility or variability (the assumptions one is implicitly making in a DCF analysis). It would be hard to find a senior executive who, faced with a five-year investment program, did not think about delaying, abandoning, or expanding aspects of the program over time, investing on a small scale initially as a means of learning about cash flow potential, or forming contingency plans given a very uncertain future. This is, and has long been, a natural way of thinking for managers. Until real options, however, corporate finance had not provided a way to structure such thinking and quantify the value of strategic alternatives.
Try explaining discounted cash flow to a good senior executive. It seems unnatural to forecast single point estimates of all cash flows and then discount those back at one rate (typically the weighted average cost of capital) to get an NPV. More sophisticated finance departments run scenario analyses to create optimistic, normal, and pessimistic projections. But it’s also important to keep in mind that neither scenario analyses nor Monte Carlo simulations of NPV are good substitutes for real options since neither incorporates management flexibility— and there still is the problem of how to determine an appropriate discount rate. To assume that there is no flexibility in the future and that one needs to make the investment now or never is simply unrealistic. In more ways than one, good senior executives have internalized real options thinking, which makes it more difficult for them to be satisfied with standard DCF analysis. And in many cases, they make decisions “in spite of” the financial analysis, which can be a frustrating experience for financial professionals starting out in the corporate world, newly armed with DCF and NPV.
“The data requirements for real options analysis are extensive and the analysis itself is time consuming.” In responding to this objection, it’s once again useful to make a comparison here to traditional techniques. The amount of data used in an analysis is more a function of the analyst than the method employed. A CEO of a Fortune 100 company once said that he was more comfortable with one data point than with two—because he could draw an infinite number of lines through one but with two data points he felt constrained.
In general, real options analyses actually tend to be less data intensive. Since the mean underlying value is used together with an estimate of volatility (variability), real options users tend to be comfortable with less precision in that mean value. But in a traditional DCF analysis, all the information is wrapped into the mean cash flow, so there is a stronger focus on getting the average “right.” As a result, extensive market research is conducted and scenario analyses are performed to fully define the average. Curiously, after exerting significant effort on collecting the data and performing these analyses, most of the information is thrown away and the expected cash flows are then discounted at the weighted average cost of capital to create a generally meaningless NPV.
The immediate response that comes to mind relates to the discount rate in a DCF analysis. Discount rates for projects that have time-varying risk are not observable in the market. There is a number—the weighted average cost of capital—that is always available and is sometimes used without regard even to the Capital Asset Pricing Model. But just because the weighted average cost of capital is easy to calculate does not mean that it is appropriate either for every project, or throughout the entire life of a single project.
Of course, volatility is an important driver of value in a real options analysis. In many cases, volatility is observable in the market, perhaps in the volatility of a biotechnology index or the volatility of Merck’s stock price. In other cases, it can be calculated by applying expert opinion to historical data. It is not perfect, but it has a lot more real-life information wrapped into it than an apparently easy-to-use discount rate. In fact, it may be a lot harder to accurately determine the appropriate discount rate for a specific project or a specific decision than it is to estimate volatility—and a project’s value is generally much more sensitive to the discount rate than to the volatility estimate.
This is true to a certain extent. On the other hand, how many decision makers really understand the underlying assumptions of the Capital Asset Pricing Model when they use DCF? It is not necessary that all users of a technique understand all the underlying assumptions and mathematical derivations. In fact, decision makers routinely take NPV at its face value without fully appreciating the math underlying it. Moreover, the prescribed mechanics and assumptions of the CAPM are often violated in practice, but nobody seems to care much.
Consulting firms who promote decision tree approaches have long maintained that they are doing real options analysis. The problem here is obvious: decision trees are merely pictorial representations of DCF calculations. Technical risk is represented in the branches. The appropriate probabilities are applied to the cash 46 JOURNAL OF APPLIED CORPORATE FINANCE flows past that decision point, just as we adjust the cash flows for risk before discounting them. In performing DCF, we adjust (or should be adjusting) cash flows in exactly the same way. Decision trees are a good way to frame the decision problem and are good communication vehicles. But they do not provide the benefits of a real options analysis.
It is true that a correctly formulated DCF analysis will typically give a lower value than its real options cousin. However, it’s important to keep in mind the role of the discount rate as a key driver of value in DCF analysis (much as volatility is a critical value driver in real options). As mentioned earlier, discount rates are not generally observable in the market for a specific project or decision. For this reason, managers who are predisposed to taking a certain project (regardless of its value) are able to start with a WACC and then adjust it downward until they get a “believable” number for the value of the project. Although people like to say that valuation is more an art than a science, I contend that there is (or could be) a lot more science in valuation, and that assumptions that are not market-based make the whole process useless. In this context—as well as in a very practical sense—the DCF versus real options debate does not serve much purpose.
I once had an experience with an investment bank that valued a biotechnology company using traditional techniques. They took the company’s products in the R&D stage and did a bottoms-up valuation. Somehow, they forgot to adjust the cash flows for technical risk in R&D and came up with a number that was three times the current stock price of the company. A real options analysis resulted in a price much closer to the market price. I suspect that the ease of use of the DCF technique in EXCEL and other such analytical tools may have let these types of mistakes creep in.
“Real options caused the technology bubble and recent crash.” This strikes me as casting about for a scapegoat. In fact, if real options analysis were more mainstream today, it might have prevented the type of tech bubble that we have recently experienced. Market participants may have had difficulty assessing value in times of high uncertainty without sophisticated tools such as real options. A reliance on rules of thumb and precedent may have actually contributed to the overvaluation of the technology stocks, which tend to show extremely high volatility. The technology bubble is a strong argument for, rather than against, the institutionalization of real options analysis in valuation.
The bottom line is that the real options framework is just a generalized asset pricing model. Every industry encounters situations that could benefit from such a generalized framework. Real options techniques can be useful in transactions involving sales or transfers of real assets among pharmaceuticals, biotechnology firms, and even insurance companies. DCF is nothing more than a special case that is applicable when there is little variability in outcomes and no management flexibility in decisions. But, in the absence of these simplifying assumptions, real options techniques are likely to provide a more complete financial analysis for the decision maker.
In fact, good managers have always thought and acted in a real options way—in large part, because markets appear to value companies and products in the same way. It is important for companies, particularly those for which “growth options” constitute a significant portion of value, to institutionalize real options-based analyses in their decision processes. Those companies that use traditional techniques and focus excessive attention on near-term earnings in their investment decisions are likely to be punished by the market in the long run.
I have discussed a number of misconceptions about real options, some of which are based on unexamined biases and others perhaps on an understandable fear of the unknown. Objective analysis and more experience with real options will eliminate these roadblocks and accelerate the process of incorporating this kind of analysis into corporate decision-making. We will then be able to fully link strategy and finance, making financial analysis a truly relevant tool for strategic decision making.
CEO – Decision Options, LLC