I



An Alternative Perspective to Internet Valuations

April 2000

Richard Solano

Duke University

Fuqua School of Business

404-886-9425

E-mail: richardsolano@

Introduction

The recent rise in stock prices for Internet companies has sparked debate over whether the apparent unbridled optimism is warranted or whether this is simply another example of investor’s “irrational exuberance”. Although the most spectacular returns have been observed in the NASDAQ composite index, a market site that trades mostly technology-related issues (Exhibit 1), this effect has bled over to what are known as “old economy” stocks, which make up the Dow Jones Industrial Average (Exhibit 2).

This performance has sent analysts scrambling to find reasons that justify the valuations being observed. Everything from revenue growth to subscribers to “eyeballs” are being used to value these stocks, without regard to future profitability. Because Internet stocks are experiencing growth as a new medium is adopted, naturally most of these parameters appear to have significant correlations to the stock prices. In some cases, this becomes a self-fulfilling prophecy. Investors begin to seek out these stocks, further driving up the price.

Clearly, however, traditional valuation methods do not suffice. The future value of GM is not driven by the same factors that drive the future value of Yahoo!. GM is in a mature industry where future growth can be forecasted relatively accurately. Yahoo!, on the other hand, is in the “Internet Portal Industry”. What is the future market capitalization of that industry? How do you measure its future sources of revenue? The future potential of this new technology is beyond our current level of understanding. The applications for this value proposition are as unknown to us as the applications of the automobile were to this unidentified analyst that wrote in 1910 “The world market for automobiles is at most one million units.” His projection was based on the assumption that only about one million families could afford chauffeurs.

This paper seeks to provide a framework for understanding value drivers in an industry that is experiencing a structural shift and present a trading strategy that exploits these drivers to deliver returns that beat an appropriate benchmark measure.

The adoption of a new technology

There is no doubt that a structural shift is taking place. The Internet is impacting existing industries and creating new markets for services. The example of Yahoo! mentioned above is a powerful example of the “market-creating” power of this new medium. As this new technology is adopted by the general population, some familiar trends emerge. The first is that new technology is adopted by the population in varying degrees over time. Initially, the adoption rate increases over time and the slope of the adoption curve is increasing. As the technology reaches adoption by half of the eligible population (defined as that segment of the population that can be defined as wither the early or late majority of adopters – to the exclusion of technology laggards – or those people that will never adopt the technology) there is a point of inflection where the adoption rate continues to increase, but the slope of adoption begins to decrease. By viewing adoption curves of previous technologies (Exhibit 3), that point of inflection occurs when about 40-50% of the population adopts the technology. In regard to Internet technology, numbers vary. The number of people online as of February 2000 stands at 123.6 million people or 45.3% of the population. Another measure called the “Active Internet Universe” captures the number of unique people that actually use the Internet on a daily basis. As of April 3, 2000, that measure stands at approximately 59.1 million people or 21.7%. In order to be consistent with previous measures of adoption that address the relative availability of the technology as opposed to actual usage patterns, we will use the former in this discussion.

It is clear that the number of people online has reached a stage at which growth can be expected to continue for the next few years, but that growth will decrease in rate. In addition, the adoption of the technology can be expected to reach a steady-state after a few years. At that point, growth will follow a more traditional pattern. Opportunity will arise from alternate applications of the technology, and not so much the initial adoption by new users. The adoption of Internet technology creates an opportunity for firms that specialize in this medium to realize abnormal earnings, once a competitive position has been staked out. Old economy stock do not have the same opportunity as any sales that are carried out online can be considered a net zero sum game. Overall consumer demand will not change as a result of the adoption of the new medium, only the method by which the consumer carries out the purchasing transaction. The real opportunity is for firms that specialize in selling over the Internet and those that offer unique services that cannot be obtained over traditional channels.

Structural disruption creates an opportunity to capture wealth

Similar to the wealth that was created through the adoption of television technology through the creation of television production companies, television networks and advertising agencies that specialized in television promotion, Internet technology will create the same types of opportunities. Each of these opportunities represents a particular series of cash flows that results in a particular market capitalization. In the case of Internet portals, the major competitors will lay claim to a set of cash flows, most likely as a result of selling advertising, which can be estimated to be a certain Net Present Value. That NPV can be thought of as the expected “prize” or “purse” for the competitors. Because the adoption of Internet technology is currently in flux, the level of that purse is unknown and difficult to estimate. We do know that it will be significantly larger than it is today as more people go online. This high probability of future payoff creates a situation where the market no longer values firms competing for that purse as a function of existing and immediate future cash flows, but rather as a probability of reaching that ultimate end-state purse. The higher the probability the firm can lay claim to a larger piece of the pie, the higher the estimated value of the firm. This can be thought of as a similar situation one finds in a horse race where the odds placed on a horse represent bettors’ best estimate as to the probability of success for a particular horse. In the case of Internet stocks, the market is placing odds on the firms success reflected in stock prices. Because of the dynamic nature of the industry, traditional methods will not work and we must view this pricing behavior as a variant of option pricing behavior where common stock is a call option on the value of the firm, and the exercise price is the par value of the firms expected debt position (Brav, Gray, Harvey and Maug, 1998).

Recall that the pricing of options can be represented using the Black-Scholes formula given by:

c = S * N(x1) – PV(k) * N(x2)

where

[pic]

and

[pic]

where

S = current stock (underlying asset) price

k = exercise price of theoption

T = time to maturity

B = price of a zero coupon riskless bond

c = call option value

[pic] = annual standard deviation of rate of return of stock

N(x) = cumulative normal probability of value less than x

The value of the call option is a function of the volatility of the return of the stock. The more volatile the stock, the higher the probability that the underlying asset will exceed the exercise price. The call option value is simply the difference between the current stock price and the present value of the exercise price. In other words, the call option represents the expected future payoff. By changing one assumption, we can use this same framework to understand how stock in an industry that is experiencing a structural shift take on this type of option pricing behavior.

In order to change this assumption, we must first discuss the concept of network effects. Network effects describe the disproportionate increase in value a service or product represents due to the higher rate of usage by the population. The telephone is a good example. The value of the telephone increases as more people put on in their home. The Internet experiences similar effects. The more people that use the Internet, the more valuable it is to other users. For example, the Internet has been around for about 20 years. Its true value as a medium for interaction increased when the browser was introduced, allowing more people to “surf the web” and providing an incentive for more users to place information online. The value to the user increased as more people created an online presence and placed more information (and now services) on the Internet. For now, we represent these network effects as a parameter NE where NE is directly proportional to the level of usage by consumers.

Returning to the option pricing formula, if we substitute the total value of the firm (TV) for the “underlying asset” and the “strike price” is the face value of outstanding debt for the firm at the time the debt is due (LTD), then the call option price is simply the value of all the outstanding shares of common stock (MCAP). In addition, we can assume that the future value of the firm is a function of network effects (NE), instead of volatility. We can make this assumption because we know we are in the middle of a structural shift where a new technology is being adopted. This gives us an alternate representation of:

MCAP = TV * N(x1) – PV(LTD) * N(x2)

where

[pic]

[pic]

Clearly, the quantitative usefulness of this relationship is constrained by the qualitative nature of the network effects term. In addition, the present value of long term debt is dynamic and the capital structure of a firm cannot be easily represented with one number. However, the qualitative aspects of option pricing behavior help us understand the dramatic changes in stock price as the market’s assessment of the network effects enjoyed by a firm change. In this case, the value MCAP dramatically increases if network effects increase. Conversely, if network effects decrease, MCAP decreases as well, to a point at which it become worthless once the adoption of the new technology reaches higher levels.

While we cannot prove empirically that this relationship holds due to the nature of long term debt, we can show that, all else held constant, network effects dramatically affect stock prices. Schwartz and Moon (1999) report, however, that based on their simulation results, “variance of the distribution of future growth rates is important in the valuation because it determines the option value of the Internet firm… higher growth rates lead to larger cash flows, which imply a more valuable firm…in contrast, if growth rates are sufficiently low, the firm may go bankrupt…if the firm goes bankrupt, it will be worth zero if growth rates are just low enough for the firm to go bankrupt or even if growth rates are far lower than that critical level”. Their research demonstrates the “option pricing” behavior that is exhibited by Internet firms.

Understanding the new value drivers

In order to understand the role of network effects, and its role on the value of an Internet firm, its necessary to identify the critical value drivers. Considerable research has been done on this subject. What follows is a discussion of this research and a conclusion on the use of REACH (a variable that describes a particular Internet site’s increase in active user as a share of the total Internet universe) as a proxy for network effects.

Schwartz and Moon (1999) apply real options theory and modern capital budgeting to the problem of valuing an Internet stock. They report that “depending on the parameters chosen, the value of an Internet stock may be rational, given high enough growth rate in revenues. Even with a very real chance that a firm may go bankrupt, if the initial growth rates are sufficiently high and if there is enough volatility in this growth over time, then valuations can be what would otherwise appear to be dramatically high”. While a high growth rate in revenues certainly explains a dramatic increase in stock price, it does not answer the question of why firms that report negligible revenues are still acquired for apparently exorbitant prices (e.g. ICQ, and Internet chat software site that gave away its software for free was purchased by America On Line for over $200 million). In addition, they use to illustrate their methodology. Amazon’s core business depends on the sale of products and necessarily must show revenues in order to cover cost of goods sold. This methodology was not applied to firms such as Yahoo! where the marginal cost of delivering their service (before overhead allocation) is negligible, and therefore revenues are not dependent on costs of sale, but rather on true market acceptance of the value proposition.

Hand (2000) reports that sales and marketing and research and development costs are valued by the market as assets, as opposed to expenses which is the traditional view of accrual-based accounting. He finds that “consistent with the argument that very large marketing costs are intangible assets, not period expenses, Net firms’ market values are reliably positive and concave in selling and marketing expenses when net income is negative…R&D expenditures are priced in a similarly concave manner, although more durably beyond the IPO than are marketing costs”. For Internet firms, the market clearly believes that by engaging in activities that will enhance a firm’s adoption by a population, the value of the firm increases (in this case exponentially), at least during a period of growth for the new technology. By developing useful technology and doing whatever it takes to put that new technology in the hands of consumers, the firm is enhancing the probability of success by staking out a market position that will increase the durability of its abnormal earnings going into the future.

Perhaps the most compelling and useful research is reported by Rajgopal, Kotha and Venkatachalam (2000). They report that “[web] traffic creates future growth potential through network effects and customer relationships that do not necessarily result in current realized revenues”. They go on to demonstrate the explanatory nature of variables that proxy for network effects, including unique visitors, reach, alliance with AOL, media visibility, affiliate programs, marketing expenditures and cash constraints. They limit their research to the identification of explanatory variables that, through linear regression analysis, may explain Internet stock returns. No attempt is made at using this information to develop a trading strategy that beats benchmark measures of return. They do provide a good indication of the relative explanatory strength of each of the variables, useful in developing a trading strategy, and help to fill in the picture of what the value drivers are for an Internet firm in the during a structural shift phase.

Based on their research, they find that REACH has the strongest association to MVE or market value of equity. REACH is defined as the number of unique users divided by the total estimated population viewing the web during the reported time period. They run several regressions and REACH consistently demonstrates strong explanatory power as demonstrated by the high t-statistics (in some cases as high as 13.95). Thus, using REACH as a proxy for the level of adoption of a specific technology or value proposition by the population is appropriate.

This research provides the necessary pieces of the puzzle to understand the value drivers that manifest themselves in the stock price of the Internet firm. Exhibit 4 depicts the relationship of the value drivers to each other. Fundamentally, the decisions the firm makes in allocating its resources are reflected in R&D and sales and marketing expenditures. These expenditures are necessarily constrained by the cash position of the firm. The combination of these three factors represents the truest measure of the firms strategy to increase its adoption by consumers. These data are reported quarterly and thus are not useful in creating a large enough sample size to yield meaningful results or to prove that above-benchmark returns can be obtained by creating a trading strategy based on these information. The resource decisions the firm makes manifest themselves in the level at which the firm’s technology is adopted by consumers. We might call this “customer acquisition” and use REACH as a proxy for this level of adoption. This data is available weekly from PC Data Online, an Internet traffic measurement service. Obviously, the effectiveness with which the firm’s R&D and Sales and Marketing resources are allocated is critical, however, it is impossible to quantify the skill of the management team in making these decisions. For now we will show this relationship as direct and proportional.

This performance in customer acquisition drives network effects which in turn drives competitive position of the firm as the market for the firm’s products or services grows. The improved competitive position increases the probability that the firm will enjoy a larger share of the market capitalization “purse” when the market leaves the “tornado” phase and enters the mature stage where traditional market and industry forces begin to take over.

Capitalizing on the value drivers

Taking into consideration the factors surrounding the Internet phenomenon such as the s-curve nature of adoption, the option pricing behavior driven by network effects and the identification of a strong proxy for network effects in the form of REACH, we can develop a trading strategy that capitalizes on these factors to beat the appropriate benchmark, which in this case is represented by a buy and hold strategy of the stock sample used in this methodology.

The methodology uses a sort strategy using weekly REACH growth percentage as the primary sort key. Various scenarios were tested and the results are summarized in Exhibit 5. In order to construct a sample for testing the methodology, I started with a base group of stocks that comprise the ISDEX® Internet index. Some of these stocks clearly are “infrastructure” firms (e.g. Cisco, JDS Uniphase) that benefit from the Internet through providing equipment, services and/or software to other Internet firms or traditional firms that are developing a web presence. Network effects are important in adoption of their technologies, but their website’s REACH percentage is not related to their success in increasing their competitive position. I limited the sample of stocks to those that could benefit from increased traffic to their site. This include firms that are selling traditional products and services over the web (e.g. Amazon, CD Now) and firms that provide new Internet-specific services over the web (e.g. Yahoo!, Real Networks) (Exhibit 6). I limited the time period to begin February 1999, as that is when PC Data Online first began to compile web traffic performance metrics. As a matter of record, they track web usage metrics and purchasing patterns from their 100,000 contracted web users. Every week, PC Data Online publishes the previous’ weeks web traffic reports. These reports cover web usage from Sunday through Saturday. The data is dated as of the last day of measurement for the week (e.g. data for the week of April 16 to April 22 would be dated April 22. PC Data Online also publishes a statistic called UNIVIS or the number of web-active individuals who visited a particular site or web company within a given time frame. However, REACH alone is used for this analysis because the correlation between REACH and UNIVIS was 0.97 (Rajgopal, Kotha and Venkatachalam, 2000).

Each week where data was available, I calculated the growth in REACH percentage over the previous week for each stock in the sample. For that same week, I calculated a one week return assuming that one could buy the stock at the closing price on the Monday after the report was available and would sell the stock at the closing price on the following Friday. For each week, I then calculated what the average return would have been from buying equal amounts of those stocks that had positive REACH percentage growth over the previous week. I also tested other trading strategies such as buying only the top 1 through 10 stock each week. All results are summarized in Exhibit 5.

In addition, commission fees for trading activity are estimated to be $8 per trade. This is most likely an overestimation of the actual cost an institutional investment firm would incur for making trades. Because at most 10 stocks are being traded per week, and we assume a large enough portfolio value over which to spread transaction costs, the commission is negligible in this analysis.

The buy-and-hold strategy whereby all stock are purchased and held throughout the sample period delivers an average weekly return of 1.15% and total return on one dollar invested of $1.58. The strategy of buying positive REACH percentage growth stocks only yields an average weekly return of 2.04% - almost double that of the buy and hold strategy – and dollar return of $2.34. Comparing to the ISDEX® index, a similar buy and hold strategy of all Internet stocks (the ISDEX® index contains non-REACH dependent stocks) yields a average weekly return of 1.5% and dollar return of $1.95.

Exhibit 7 shows what would have been the total return over the sample period from investing one dollar in each of the trading strategies. The highest total return is obtained through the positive REACH growth trading strategy. It is interesting to note that simply buying the “top 3” or “ top 5” stocks after the sort yields inferior results. This is because is some cases, by doing this arbitrary selection of N stocks, negative REACH growth stocks are selected in weeks where overall usage growth slows down. By selecting only those stocks that are positive growth, especially robust stocks are identified even in weeks where usage growth is weak.

Conclusion

In this paper I explore the importance of taking into context the structural shift taking place in the economy due to the “s-curve” adoption of the Internet by consumers, the option-pricing behavior of stocks the market engages in during these periods, and the importance of network effects in identifying firms that are especially well-positioned to create sustainable competitive advantage as the Internet industry transforms to a mature market. I also developed a trading strategy that significantly outperforms a buy-and-hold strategy over the past 13 months. Its worth noting that due to lack of data for earlier periods, I am unable to determine the soundness of the strategy over the early periods of technology adoption and, further, am unable to determine how this strategy holds up as adoption moves into the late majority of users of the Internet. It is my suspicion that volatility will increase significantly during this “late majority adoption period” as the market attempts to decide whether it should be valuing firms based on future potential or more traditional measures of value such as revenue – and perhaps even discounted cash flows or abnormal earnings growth. Further, it is hypothesized that as money flows into Internet firms, it is possible that a condition may arise where the market capitalizations of all firms in a given industry (e.g. online auction industry), can exceed the expected “purse” or prize that can be expected in the steady-state. In this case, the market for that industry is clearly overvalued as there is no financial justification for the price of the stock under that scenario.

Bibliography

Rajgopal, S., Kotha, S. and Venkatachalam, M. 2000. The relevance of Web Traffic for Internet Stock Prices. University of Washington and Stanford University.

Brav, A., Gray, S., Harvey, C. and Maug, E. 1997. Option Contracts. Global Financial Management course. Duke University – Fuqua School of Business.

Hand, J. 2000. Profits, losses and the non-linear pricing of Internet stocks. UNC Chapel Hill.

Schwartz, E. and Moon, M. 2000. Rational Pricing of Internet Companies. UCLA and Fuller & Thaler Asset Management.

PC Data Online 2000. Top weekly web sites. Internet Audience Measurement Reports Reston, VA.

................
................

In order to avoid copyright disputes, this page is only a partial summary.

Google Online Preview   Download