9

Baby Steps: Young and Start-Up Companies

Valuing companies early in the life cycle is difficult, partly because of the absence of operating history and partly because most young firms do not make it through these early stages to success. This chapter looks at the challenges we face when valuing young companies and the shortcuts employed by many who have to estimate the value of these businesses to arrive at value. Whereas some of the rules that have developed over time for valuing young businesses make sense, other rules inevitably lead to erroneous and biased estimates of value.

Because young companies usually are small, they represent only a small part of the overall economy. However, they tend to have a disproportionately large impact on the economy for several reasons.

• Employment: Although few studies focus just on start-ups, evidence exists that small businesses account for a disproportionate share of new jobs created in the economy. The National Federation of Independent Businesses estimates that about two-thirds of the new jobs created in recent years have been in small businesses and that start-ups account for a large share of these new jobs.¹

• Innovation: In the early 1990s, Clayton Christensen, a strategy guru from the Harvard Business School, argued that radical innovation—innovation that disrupted traditional economic mechanisms—was unlikely to come from established firms. They have too much to lose from the innovation. Radical innovation was more likely to come from start-up companies that have little to lose. Thus, online retailing was pioneered by a young upstart, Amazon.com, rather than by traditional retailers.

• Economic growth: The economies that have grown the fastest in the last few decades have generally been those that have a high rate of new business formation. Thus, the U.S. generated much more rapid economic growth than Western Europe during the 1990s, primarily as a consequence of the growth of small new technology companies. Similarly, much of the growth in India has come from smaller companies rather than from established companies.

• No history: At the risk of stating the obvious, young companies have very limited histories. Many of them have only one or two years of data available on operations and financing. Some have financials for only a portion of a year, for instance. Some of the youngest do not even have a product to sell, just an idea for a product and a market that it might serve.

• Small or no revenues, operating losses: The limited history that is available for young companies is rendered even less useful by the fact that they contain little operating detail. Revenues are small or nonexistent for idea companies, and the expenses often are associated with getting the business established, rather than generating revenues. In combination, they result in significant operating losses.

• Dependent on private equity: With a few exceptions, young businesses are dependent on equity from private sources, rather than public markets. At the earlier stages, the equity is provided almost entirely by the founder (and friends and family). As the promise of future success increases, and with it the need for more capital, venture capitalists become a source of equity capital, in return for a share of ownership in the firm.

• Many don’t survive: Most young companies don’t survive the test of commercial success. Several studies back up this statement, although they vary in the failure rates they find. A study of 5,196 start-ups in Australia found that the annual failure rate was in excess of 9% and that 64% of the businesses failed in a ten-year period.² Knaup and Piazza (2005, 2008) used data from the Bureau of Labor Statistics Quarterly Census of Employment and Wages (QCEW) to compute survival statistics across firms.³ This census contains information on more than 8.9 million U.S. businesses in both the public and private sector. Using a seven-year database from 1998 to 2005, the authors concluded that only 44% of all businesses that were founded in 1998 survived at least four years and that only 31% made it through all seven years. In addition, they categorized firms into ten sectors and estimated survival rates for each one. Table 9.1 presents their findings on the proportion of firms that made it through each year for each sector and for the entire sample.

Table 9.1 Survival of New Companies Founded in 1998

	Proportion of Firms That Were Started in 1998 That Survived Through…
	Year 1	Year 2	Year 3	Year 4	Year 5	Year 6	Year 7
Natural resources	82.33%	69.54%	59.41%	49.56%	43.43%	39.96%	36.68%
Construction	80.69%	65.73%	53.56%	42.59%	36.96%	33.36%	29.96%
Manufacturing	84.19%	68.67%	56.98%	47.41%	40.88%	37.03%	33.91%
Transportation	82.58%	66.82%	54.70%	44.68%	38.21%	34.12%	31.02%
Information	80.75%	62.85%	49.49%	37.70%	31.24%	28.29%	24.78%
Financial activities	84.09%	69.57%	58.56%	49.24%	43.93%	40.34%	36.90%
Business services	82.32%	66.82%	55.13%	44.28%	38.11%	34.46%	31.08%
Health services	85.59%	72.83%	63.73%	55.37%	50.09%	46.47%	43.71%
Leisure	81.15%	64.99%	53.61%	43.76%	38.11%	34.54%	31.40%
Other services	80.72%	64.81%	53.32%	43.88%	37.05%	32.33%	28.77%
All firms	81.24%	65.77%	54.29%	44.36%	38.29%	34.44%	31.18%

• Note that survival rates vary across sectors. Only 25% of firms in the information sector (which includes technology) survive seven years. On the other hand, almost 44% of health service businesses make it through that period. Table 9.2 provides updated values for survival statistics, still from the Bureau of Labor Statistics, for the aggregate sample through 2015.

Table 9.2 Survival Rate of New Companies Founded, by year: 2005–2015

		Year Founded
		2005	2006	2007	2008	2009	2010	2011	2012	2013	2014	2015
Survival (in years)	0	100%	100%	100%	100%	100%	100%	100%	100%	100%	100%	100%
	1	80%	78%	77%	75%	77%	79%	79%	79%	80%	80%	NA
	2	69%	66%	64%	63%	66%	69%	69%	69%	69%	NA	NA
	3	60%	57%	56%	57%	60%	62%	62%	62%	NA	NA	NA
	4	53%	50%	50%	52%	55%	56%	56%	NA	NA	NA	NA
	5	47%	45%	46%	48%	50%	51%	NA	NA	NA	NA	NA
	6	43%	42%	43%	44%	46%	NA	NA	NA	NA	NA	NA
	7	41%	40%	40%	41%	NA	NA	NA	NA	NA	NA	NA
	8	38%	37%	38%	NA	NA	NA	NA	NA	NA	NA	NA
	9	36%	35%	NA	NA	NA	NA	NA	NA	NA	NA	NA
	10	34%	NA	NA	NA	NA	NA	NA	NA	NA	NA	NA

• Based on this table, 34% of all companies that were started in 2005 survived ten years and were still in existence in 2015 and 35% of companies started in 2006 made it nine years, to survive in 2015. Survival rates have remained low over time and are affected by the overall economy, as evidenced by the slightly higher failure rates in 2008 and 2009, in the midst of the banking crisis. Note also that the failure rate is highest in the early years and if a company makes it through year 3 or 4, the failure rate drops off significantly.

• Multiple claims on equity: The repeated forays made by young companies to raise equity do expose equity investors, who invested earlier in the process, to the possibility that their value can be reduced by deals offered to subsequent equity investors. To protect their interests, equity investors in young companies often demand and get protection from this eventuality. This can take the form of first claims on cash flows from operations and in liquidation. Investors also might have control or veto rights, allowing them to have a say in the firm’s actions. As a result, different equity claims in a young company can vary on many dimensions that can affect their value.

On each of these measures, young firms pose estimation challenges that can be traced back to their common characteristics.

• The absence of historical data makes it difficult to assess how well the revenues from existing assets will hold up if macroeconomic conditions become less favorable. In other words, if all you have is one year of financial data, deciding whether the revenues represent a flash in the pan or are sustainable is difficult. The lack of data from prior years also makes analyzing how revenues would change if the company changed its pricing policy or faced new competition more difficult.

• The expenses that young companies incur to generate future growth are often mixed in with the expenses associated with generating current revenues. For instance, it is not unusual to see the selling, general, and administrative (S, G, and A) expenses at some young companies be three or four times larger than revenues, largely because they include the expenses associated with lining up future customers. To value existing assets, we have to be able to separate these expenses from genuine operating expenses, and that is not easy to do.

In general, even if there are financial statements with historical data that can be used to value existing assets, these assets represent only a sliver of the overall value of a young business. That would suggest that if valuing this business, you should be spending less of your time on this component of value and more of your time on valuing growth assets.

• The absence of revenues in some cases, and the lack of history on revenues in others, means that we cannot use past revenue growth as an input into the estimation of future revenues. As a result, we are often dependent on the firm’s own estimates of future revenues, with all the biases associated with these numbers.

• Even if we could estimate revenues in future years, we also have to estimate how earnings will evolve in future years as revenues change. Again, the fact that young companies tend to report losses and have no history on operating income makes assessing what future profit margins will be more difficult.

• Chapter 2 noted that revenue or even earnings growth per se is not what determines value, but the quality of that growth. To assess the quality of growth, we looked at how much the firm reinvested to generate its expected growth. We noted that value-creating growth arises only when a firm generates a return on capital greater than its cost of capital on its growth investments. This intuitive concept is put to the test with young companies, because there is little basis for the expected return on capital on new investments. Past data doesn’t provide much guidance, because the company has made so few investments in the past, and these investments have been in existence for short periods. The current return on capital, which is often used as a starting point for estimating future returns, is generally a negative number for young companies.

In summary, we have a tough time estimating future growth in revenues and operating margins for young companies. The estimation problems are accentuated by the difficulties we face in coming up with reinvestment assumptions that are consistent with our growth estimates.

With young companies, these assumptions are open to challenge. First, most young companies are not publicly traded and have no publicly traded bonds outstanding. Consequently, there is no way we can run a regression of past returns to get an equity beta, or use a market interest rate on debt. To add to the problem, the equity in a young company is often held by investors who are either completely invested in the company (founders) or only partially diversified (venture capitalists). As a result, these investors are unlikely to accept the notion that the only risk that matters is the risk that cannot be diversified away. Instead, they will demand compensation for at least some of the firm-specific risk. Finally, we noted that equity in young companies can come from multiple sources at different times and with very different terms attached to it. It is conceivable that the differences across equity claims can lead to different costs of equity for each one. Thus, the cost of equity for an equity claim that has first claim on the cash flows might be lower than the cost of equity for an equity claim that has a residual cash flow claim.

One final and very important risk is the risk of failure. As we saw in Tables 9.1 and 9.2, many young companies don’t make it through the winnowing process to become healthy, money-making entities. That risk, large though it might be, does not easily fit into discount rates, specifically, and into a discounted cash flow valuation (DCF), more generally. A DCF is a going concern value and the chance of failure is a discrete, truncation risk. Trying to push discount rates up to reflect failure risk is incompatible with the very notion of a going concern (and DCF).

• Will the firm make it to stable growth? As noted in the last section, many young firms will never make it to stable growth, and the terminal value will not provide the large windfall to value that it does for a going concern. Estimating a firm’s probability of survival early in the life cycle therefore is a critical component of value, but it’s not necessarily an easy input to estimate.

• When will the firm become a stable growth firm? Even if we assume that a firm will make it to stable growth in the future, estimating when that will occur is a difficult exercise. After all, some firms reach steady state in a couple years, whereas others have a much longer stretch of high growth before settling into mature growth. The judgment of when a firm will become stable is complicated by the fact that the actions of competitors can play an important role in how growth evolves over time.

• What will the firm look like in stable growth? Chapter 2 noted that not just the stable growth rate determines the magnitude of terminal value. Also important are the concurrent assumptions we make about risk and excess returns during the stable phase. In effect, assuming that a firm will continue to generate excess returns forever will lead to a higher terminal value than assuming that excess returns will converge on zero or be negative. Although we must make this judgment for any firm, the absence of any historical data on excess returns at young firms complicates estimation.

The net effect of these diverse equity claims is that allocating the value of equity across different claims requires us to value both the preferential cash flow and control claims and the protective rights built into some equity claims but not others.

• What do you scale value to? All valuation multiples have to be scaled to some common measure. Conventional scaling measures include earnings, book value, and revenues. With young companies, each of these measures can pose problems. Because most of them report losses early in the life cycle, multiples such as price earnings ratios and EBITDA multiples cannot be computed. Because the firm has been in operation only a short while, the book value is likely to be a very small number and won’t reflect the true capital invested in the company. Even revenues can be problematic, because they can be nonexistent for idea companies and miniscule for companies that have just transitioned into commercial production.

• What are your comparable companies? When relative valuation is used to value a publicly traded company, the comparable firms are usually publicly traded counterparts in the same sector. With young companies, the comparison would logically be to other young companies in the same business, but these companies are usually not publicly traded and have no market prices (or multiples that can be computed). We could look at the multiples at which publicly traded firms in the same sector trade at, but these firms are likely to have very different risk, cash flow, and growth characteristics than the young firm being valued.

• What is the best proxy for risk? Many of the proxies used for risk, in relative valuation, are market-based. Thus, beta or standard deviation of equity returns are often used as measures of equity risk, but these measures cannot be computed for young companies that are privately held. In some cases, the standard deviation in accounting numbers (earnings and revenues) is used as a measure of risk, but this, too, cannot be computed for a firm that has been in existence for a short time.

• How do you control for survival? In the context of discounted cash flow valuation, we looked at the problems created by the high failure rate of young companies. This is also an issue with using relative valuation. Intuitively, we would expect the relative value of a young company (the multiple of revenues or earnings we assign it) to increase with its likelihood of survival. However, putting this intuitive principle into practice is not easy.

In conclusion, the use of relative valuation might seem like an easy solution when faced with the estimation challenges posed in intrinsic valuation. But all the problems we face with the latter remain problems when we do the former.

• Top line and bottom line, no detail: Estimating the details of cash flow and reinvestment for young companies is difficult. Consequently, many valuations of young companies focus on the top line (revenues) and the bottom line (earnings, usually equity earnings). Little or no attention is paid to either the intermediate items (that separate earnings from revenues) or the reinvestment requirements (that separate earnings from cash flows).

• Focus on the short term, rather than the long term: The uncertainty we feel about the estimates we make for young companies become greater as we go further out in time. Many analysts use this as a rationale for cutting short the estimation period, using only three to five years of forecasts in the valuation. “Forecasting beyond that point in time is too difficult” is the justification they offer for this short time horizon.

• Mixing relative with intrinsic valuation: To deal with the inability to estimate cash flows beyond short time periods, analysts who value young companies use relative valuation as a crutch. Thus, the value at the end of the forecast period (three to five years) is often estimated by applying an exit multiple to the expected revenues or earnings in that year. The value of that multiple is itself estimated by looking at what publicly traded companies in the business trade at right now.

• Discount rate as the vehicle for all uncertainty: The risks associated with investing in a young company include not only the traditional factors—earnings volatility and sensitivity to macroeconomic conditions, for example—but also the likelihood that the firm will not survive to make a run at commercial success. When valuing private businesses, analysts often hike up discount rates to reflect all their concerns about the firm, including the likelihood that the firm will not make it.

• Share count adjustments for dilution: Young growth firms are often dependent on new equity issuances in future years to cover cash flow needs arising both from operating shortfalls and reinvestment needs. That resulting dilution in future years has to be factored into the value of existing equity shares, because existing equity investors will own smaller and smaller proportions of the business in future years. Many analysts, faced with this reality, try to adjust for future dilution by estimating future equity issues and adjusting the share count today for those issues. That creates a circularity problem because you need to estimate share prices in future years to estimate future share issuances and you need future share issuances to estimate current share price.

• Ad hoc and arbitrary adjustments for differences in equity claims: As we noted in the preceding section, equity claims in young businesses can have different rights when it comes to cash flow and control. When asked to make judgments on the value of prior claims on cash flows or superior control rights, many analysts use rules of thumb that are either arbitrary or based on dubious statistical samples.

All of these practices come into play in the most common approach used to value young firms—the venture capital approach. This approach, which we described briefly in Chapter 3, has four steps:

1. We begin by estimating the expected earnings or revenues in a future year, but not too far into the future: two to five years is the typical range. In most cases, the forecast period is set to match the point in time at which the venture capitalist plans to sell the business or take it public.

2. The value at the end of the forecast period is assessed by multiplying the expected earnings in the future year by the multiple of earnings (PE ratio) that publicly traded firms in the sector trade at. In some cases, the multiple is based on other companies in the sector that have been sold or have gone public recently.

   Equity Value at End of Forecast Horizon = Expected Earnings_{Year n} × Forecast PE

   Alternatively, the revenues at the end of the forecast period can be multiplied by the revenue multiple at which publicly traded firms trade at to arrive at an estimate of the value of the entire business (as opposed to just equity):

   Enterprise Value at End of Forecast Period = Expected Revenues_{Year n} × Forecast EV/Sales

   This approach is used for companies that might not become profitable until later in the life cycle.

3. The estimated value at the end of the forecast period is discounted back at a target rate of return. This is generally set high enough to capture both the perceived risk in the business and the likelihood that the firm will not survive. Because the latter is a high, venture capital required rates of return tend to be much higher than the discount rates we see used with publicly traded companies.

   Equity Value Today=Equity Value at End of Forecast Horizon(n)(1+Target Rate of Return)n

   Table 9.3 summarizes the target rates of return demanded by venture capitalists, categorized by how far along a firm is in the life cycle.

   Table 9.3 Venture Capital Target Rates of Return: Stage in the Life Cycle

   Stage of Development	Typical Target Rates of Return
   Start-up	50% to 70%
   First stage	40% to 60%
   Second stage	35% to 50%
   Bridge/IPO	25% to 35%

   How do we know that these rates of return have survival risks built into them? In addition to the intuitive rationale that they decrease as firms move through the life cycle and the chance of failure drops off, the actual returns earned by venture capitalists at every stage of the process are much more modest. Table 9.4 summarizes the actual returns earned by venture capitalists (VC) in the aggregate for investments across the life cycle, through 2014.

   Table 9.4 Returns Earned by Venture Capitalists in 2014

   	Three-Year	Five-Year	Ten-Year	25-Year
   Early/seed VC	19.25%	16.83%	9.81%	29.41%
   Late and Expansion VC	14.21%	16.57%	12.80%	13.15%
   Multistage VC	17.42%	14.81%	10.33%	13.33%
   All VC	18.04%	16.07%	10.28%	22.11%
   NASDAQ	22.05%	15.85%	8.09%	9.83%
   S&P	20.41%	15.45%	7.67%	9.62%

   Over the longest time period (25 years), the returns earned by venture capitalists, especially on early-stage investments, are significantly higher than the returns earned by investors in equity in public markets over the same period. The high returns, though, reflect significant survivor bias, because the least successful VC investors go out of business well before 25 years.⁴ But even these skewed returns are nowhere near the target returns listed in Table 9.2. For instance, early-stage VC investors earned an annual return of 29.4% over the last 25 years, well below the 50 to 70% target returns. In effect, the high target rates of return that are used in analysis are not delivered by most investments (usually the ones that fail to make it to the exit valuation). Over shorter time periods, the returns on VCs are not significantly different from returns on public markets, suggesting that if venture capital investing is a winning game, it requires a long time horizon to deliver results.

4. Venture capitalists receive a proportion of the business in return for the capital they bring to the firm. In making this judgment, it is worth drawing a contrast between two valuations, a pre-money valuation, which is the value that the business would have without the VC capital infusion and a post-money valuation, which is the value with the VC capital infusion. The difference between the two numbers, at least in the VC approach, is the new capital infusion:

   Post-Money Valuation = Pre-Money Valuation + New Capital Infusion

   The question then becomes whether the value estimated in step 3 is a pre-money or post-money valuation and the answer depends on who you ask. Using the argument that their capital infusion is integral to delivering future earnings in step 2 and the resulting value in step 3, venture capitalists contend that the value from step 3 is a post-money valuation and net out their capital infusion to estimate what percent of the equity that they are entitled to.

   Proportion of equity to new VC=New Capital Provided/(Equity Value(from VC perspective)Today from Step 3−New CapitalProvided)

   The pushback from the founders and existing equity investors in the company could be that they could have been able to access the capital elsewhere and that the value in step 3 is therefore a pre-money valuation. The post-money value, with that assumption, would be the value in step 3 augmented by the cash balance, and the proportion that the VC is entitled to be will be much smaller.

   Proportion of equity to new VC=New Capital Provided/(Equity Value(from owner perspective)Today from Step 3+New CapitalProvided)

   Given the opacity of the VC approach, it is tough to argue that either side is right, or has the high ground and the final value will reflect their relative bargaining power.

As we see it, the venture capital approach has several problems. Many of them are rooted in the practices we listed before:

• By focusing on revenues and earnings, and ignoring both the intermediate items and those that come after, venture capital valuations encourage game playing. Because value increases as the projected earnings (revenues) increase, the existing owners of the business try to push up these values, without having to flesh out the consequences in terms of future capital investment. On the other side of the bargaining table, venture capitalists will argue for using lower numbers for earnings and revenues, because this pushes down the estimated value (and gives them a greater share of equity for the same capital investment). Consequently, the projected value becomes a bargaining point between the two sides rather than the subject of serious estimation.

• Venture capital valuations try to avoid the serious challenges of estimating operating details for the long term by cutting off the estimates prematurely (with a short forecast period) and using a multiple that is usually based on what comparable companies are trading at currently. However, the multiple of earnings or revenues that a business will trade at three years from now will be a function of the cash flows after that point. Not estimating those cash flows or dealing with the uncertainty in the cash flows does not mean that the uncertainty has gone away.

• A degree of sloppiness is associated with the use of a target rate to discount the firm’s future value. This target rate is the rate demanded by venture capitalists, who are equity investors in the firm. It incorporates the likelihood that the business will fail. Using this number as the discount rate on the future value of the business creates two problems. The first is that the future value discounted has to be an equity value. Of course, this is the case when we use expected equity earnings and a PE ratio, but it is not so if we use revenues and enterprise value multiples. In the latter case, we should consider the cost of capital as the discount rate and not the rate demanded by just equity investors. The second problem is that building in a probability that the business will not survive into the discount rate also implies that this rate will not change over time, as a firm moves through the life cycle.

• The distinction between pre- and post-money values in the VC approach is largely in the eyes of the beholder, partly because the approach does not explicitly make assumptions about how the VC capital is used and its link to the estimated future earnings. For instance, if some or all of the new capital infusion is used by existing equity investors to cash out of their ownership in the firm, the portion that is removed from the firm should not be considered to get to the post-money value.

Ultimately, it is worth keeping in mind that venture capitalist success has less to do with valuation skills and more to do with the capacity to pick the right companies to invest in, based upon product offerings and management quality, pricing these companies (rather than valuing them) well, and exiting from these investments at the right time.

To get from a story about a company to a valuation does require a few steps, which we have outlined in Figure 9.2.

To tell a story about a business, in step 1, you need to understand the business it is in and the competition. To make sure that the story that you have told is not a fairy tale, you have to check the story in step 2 for possibilities, plausibility, and probability. Each test becomes more stringent than the previous one; for every hundred possible stories, perhaps a dozen are plausible and a few are probable. In step 3, you convert each part of your story to a valuation input, with revenue growth, margins, reinvestment, and risk all tied to your story. In step 4, the valuation inputs are fed into a valuation model to arrive at a value estimate for the company. Finally, in step 5, you open your valuation up for feedback and criticism, with the intent of improving your story.

1. Potential market for the product/service: The first step in deriving the firm’s revenues is estimating the total potential market for its products and services. At this juncture we face two challenges:

   Defining the product/service offered by the firm: If the product or service offered by the firm is defined narrowly, the potential market will be circumscribed by that definition and will be smaller. If we use a broader definition, the market will expand to fit that definition. For example, defining Amazon.com as a book retailer, which is what it was in 1998, would have yielded a total market of less than $10 billion in that year, representing total book retailing sales in 1998. Categorizing Amazon.com as a general retailer would have yielded a much larger potential market. While that might have been difficult to defend in 1998, it became more plausible as Amazon expanded its offerings in 1999 and 2000.

   Estimating the market size: Having defined the market, we face the challenge of estimating the size of that market. For a product or service that is entering an established market, the best sources of data tend to be trade publications and professional forecasting services. Almost every business has a trade group that tracks the operating details of that business. The U.S. alone has almost 7,600 trade groups, tracking everything from aerospace to telecommunications.⁶ In many businesses, firms specialize in collecting information about the businesses for commercial and consulting purposes. For instance, the Gartner Group collects and provides data on different types of information technology businesses, including software.

   Evolution in total market over time: Because we have to forecast revenues into the future, getting a sense of how the total market is expected to change or grow over time would be useful. This information usually is available from the same sources that provide the numbers for the current market size.

2. Market share: After we have a sense of the overall market size and how it will change over time, we have to estimate the share of that market that will be captured by the firm being analyzed, both in the long term and in the time periods leading up to steady state. Clearly, these estimates will depend on both the quality of the product or service that is being offered and how well it measures up against the competition. A useful exercise in estimation is to list the largest players in the targeted market currently and to visualize where the firm being valued will end up as soon as it has an established market. However, two other variables must be considered concurrently. One is the capacity of the management of the young company to deliver on its promises. Many entrepreneurs have brilliant ideas but do not have the management and business skills to take them to commercial fruition. That is part of the reason that venture capitalists look for entrepreneurs who have a track record of past success. The other variable is the resources that the young company can draw on to get its product/service to the desired market share. Optimistic forecasts for market share have to be coupled with large investments in both capacity and marketing; products don’t produce and sell themselves.

3. Operating expenses/margins: Revenues might be the top line, but for investors, a firm can have value only if it ultimately delivers earnings. Consequently, the next step is estimating the operating expenses associated with the estimated revenues. We are stymied in this process with young companies, both by the absence of history and the fact that these firms usually have very large operating losses at the time of the estimate. Again, we would separate the estimation process into two parts. In the first part, we would focus on estimating the operating margin in steady state, primarily by looking at more established companies in the business. After we have the target margin, we can look at how we expect the margin to evolve over time. This “pathway to profitability” can be rockier for some firms than others, with fixed costs and competition playing significant roles in the estimation. One final issue that has to be confronted at this stage is the level of detail that we want to build into our forecasts. In other words, should we just estimate the operating margin and profit, or should we try to forecast individual operating expense items such as labor, materials, selling, and advertising expenses? As a general rule, the level of detail should decrease as we become more uncertain about a firm’s future. While this might seem counterintuitive, detail in forecasts leads to better estimates of value, if and only if we bring some information into that detail that otherwise would be missed. An analyst who has a tough time forecasting revenues in year 1 really is in no position to estimate labor or advertising costs in year 5 and should not even try. In valuing young companies, less (detail) is often more (precision).

4. Investments for growth: When owners are asked for forecasts of revenues and earnings (steps 2 and 3), it is natural that they go for optimistic values. Revenues increase at exponential rates, and margins quickly move toward target values. In any competitive business, though, neither revenue growth nor margin improvement is delivered for free. Consequently, estimating how much the firm is reinvesting to generate the forecast growth is critical. With a manufacturing firm, this takes the form of investments in additional production capacity. With a technology firm, it includes not only investments in R&D and new patents, but also human capital (hiring software programmers and researchers). There are two reasons to pay attention to this step of the process. The first is that these investments require cash outflows and thus affect the final bottom line, which is the cash flow that can be delivered to investors. The second reason (and this is especially so with young firms) is that this reinvestment often results in negative cash flows, which then have to be covered with new capital infusions. Thus, existing equity investors either see their share of the ownership reduced (when new equity investors come in) or are called on to make fresh investments to keep the business going.

5. Compute the tax effect: With healthy firms, computing the tax effect is usually a simple exercise of multiplying the expected pre-tax operating income by the tax rate. The only real estimation question we face is what tax rate to use—marginal or effective. Young firms that are losing money have two estimation challenges. The first is that these firms have generally never paid taxes in the past (because they have never generated earnings) and thus have no effective tax rates. The second challenge is that the losses that have been made in the past and that you often expect in the near future create net operating losses that can be carried forward and used to shelter positive earnings in future years. The most direct way of dealing with these losses is to cumulate them as they are expected to occur over time and to keep track of the net operating loss (NOL) carryforward. In the first few years of positive earnings, we can draw on this NOL and essentially not pay taxes. When the NOL is exhausted, we should move to a marginal tax rate based on the statutory tax codes. This is a conservative solution. The alternative is to use the average effective tax rate paid by healthy firms in the sector.

6. Check for internal consistency: One of the perils of the top-down approach is that operating income and reinvestment are estimated separately, so these numbers might not be internally consistent. In other words, we might be reinvesting too little, given our forecasts of expected revenue growth, or too much. One simple test that can be used to check for consistency is to compute an imputed return on capital, based on the earnings and reinvestment forecasts:

   Imputed Return on Capital=Expected Operating Income after TaxtCapital Invested in Firmt−1

   The numerator is the forecast operating income, and the denominator is computed as the cumulated total of all reinvestment (net capital expenditures and change in noncash working capital) over time, through period t −1, added to the initial capital invested (at the time of the valuation):

   Capital Investedt−1=Capital Investedo+∑n=1n=t−1Reinvestmentn

   The imputed return on capital, as you approach steady state, can then be compared to both the industry average return on capital (to ensure that you are not making your company an outlier) and to the company’s own steady state cost of capital. An imputed return on capital well above the industry average and the cost of capital is an indication that the reinvestment forecast for the firm over the forecast period is insufficient, given the expected earnings. Conversely, an imputed return on capital below the cost of capital would indicate that the reinvestment numbers are too high, given the revenue and earnings forecasts.

Note that earnings become positive well before cash flows do; the latter are weighed down by the reinvestment made to sustain future growth. In fact, the cash flows remain negative for the next ten years, representing what is referred to as “cash burn” at young companies. Secure Mail will have to raise fresh capital, from new equity issues or debt, to cover these cash shortfalls. Later in this chapter, we will come back and address the resulting dilution and how it affects value per share today.

Earlier in this section, we argued that to value a company, you need to come up with a story first and then tie to value inputs and a value. If you are wondering where the story about Secure Mail is in the valuation, it is worth noting that every input in this valuation was tied to a piece of the Secure Mail story that we are crafting. Figure 9.3 shows the links between story and numbers for Secure Mail.

The value that we will derive for Secure Mail will therefore reflect our expectation of a company that has a strong chance to grow to gain a significant share of the antivirus software business, with strong competitive advantages allowing it to earn high margins and returns on capital. Changing that story to make it more expansive, where Secure Mail expands into new markets, or more restrictive, where technology brings fresh competition into the antivirus market, will change the value.

1. Capacity size/investment: The process begins with an estimate of what we will need to invest to get the business off the ground; this also determines the production capacity. A trade-off is inherent in this step. Investing in more capacity allows us to produce and sell more in the future, but the capital (both financial and human) needed to sustain this capacity will also be greater. To the degree that either human or financial capital is limited, we might have to settle for less capacity over more.

2. Unit sales/revenues: Once we have chosen a capacity constraint, we have to estimate how many units we can sell each period, for the forecast period, and the price that will be charged per unit. At this stage, we have to consider not only the potential market for the product or service we offer, but also the competition (both current and potential) in this market. The choices we make on pricing can determine the number of units sold. Lower prices generally translate into more sales, but not necessarily higher profits.

3. Operating costs: With the number of units sold each period as an input, we can estimate the costs of production in each period. These estimates should include not only the costs of inputs that go into the product, but also selling, administrative, and other costs. The latter must be consistent with the unit sales assumptions in step 2.

4. Taxes: The revenue and expense estimates are used to estimate the taxable income that the firm will generate each period and the resulting taxes. At this stage, we also have to separate capital from operating expenses, and estimate depreciation and amortization on the former. We also must separate operating expenses from financial expenses (interest expenses) to determine cash flows to the firm and cash flows to equity. The former is before financial expenses, whereas the latter is after.

5. Additional reinvestment: Although we estimated the initial investment in step 1, we might have to make additional investments over time to augment or preserve the earning capacity of the business. We need to determine what the business will have to reinvest to preserve its income-generating capacity. If the business requires working capital, growth in revenues might also lead to investments in working capital (inventory and accounts receivable) that have to be considered as reinvestment.

As a general rule, bottom-up approaches of cash flows yield lower expected cash flows and earnings, because we work with capacity constraints. Consequently, bottom-up is more suited for businesses that either face significant restrictions on raising additional capital (too small and/or in the wrong type of business) or are dependent on a key person or people for their success. As a general rule, personal-service businesses (medical practices, a plumbing business, restaurants) are better valued using this approach rather than the top-down approach, unless the service can be franchised or replicated easily.

• Beta and cost of equity: Young companies are often held by either undiversified owners or partially diversified venture capitalists. Consequently, assuming that the only risk that should be priced in is the market risk does not make sense. The cost of equity must incorporate some (in the case of venture capitalists) or maybe even all (for completely undiversified owners) of the firm-specific risk. The standard practice of estimating betas from stock prices will not work, because young firms are generally not publicly traded.

• Cost of debt: Young firms almost never have bonds outstanding; instead, they are dependent on bank loans for debt. Consequently, no bond rating will measure default risk. Even though we might be able to use the process described in Chapter 2 to estimate a synthetic rating, the resulting cost of debt might not appropriately capture the interest rates actually paid by these small and risky businesses, because banks might charge them a premium.

• Debt ratio: Because the equity and debt in young companies are not traded, no market values can be used to weight the debt and equity to arrive at the cost of capital.

The confluence of these problems is used to justify the use of arbitrary “target rates” by venture capitalists. We suggest an alternative process built around the following steps:

1. Sector averages: While the company being valued might not be traded, there are generally other companies in the same business that have made it through the early stage in the life cycle and are publicly traded. We would use the betas of these firms to arrive at an estimate of the market risk associated with being in this business. Generally, this requires taking an average of the regression betas across the publicly traded firms and unlevering the beta to arrive at the beta of the business:

   Unlevered Beta for Sector=Average Regression Beta for Publicly Traded Firms(1+(1−Tax Rate) Average Market D/E Ratio for Publicly Traded Firms)

2. Adjust for diversification or its absence. As noted earlier, the owners of young businesses tend not to be diversified. In fact, the entire firm might be held by the founder, who, in turn, has all of his or her wealth tied up in that investment. To account for this absence of diversification, we will again draw on the publicly traded firm sample. The same regressions that yielded the market betas for these firms also provide an estimate of how much of the risk in these firms comes from the market (through the R-squared and correlation coefficients in the regressions). Dividing the market beta by the correlation of the publicly traded firms with the market gives us a scaled-up version of beta (that we will call the total beta) that captures all the risk of being in a specific business, rather than just the market risk:

   Total Beta=Market BetaPublicly Traded Firms in BusinessCorrelation with MarketPublicly Traded Firms in Business

   This total beta will be much higher than the market beta. The resulting cost of equity will reflect the cost of equity to an investor who is completely invested only in this business. As the firm expands and taps into venture capital, it attracts investors who have some diversification. Venture capitalists tend to hold investments in multiple companies, but often in the same sector or a few sectors. The portfolio of investments held by a venture capitalist will be more highly correlated with the market than an individual company is, and the resulting total beta to a venture capitalist will be lower:

   Total BetaVC=Market BetaPublicly Traded Firms in BusinessCorrelation with MarketVC Portfolio

   Thus, as firms move through the life cycle and attract larger and more diversified venture capitalists into the fold, they should see lower costs of equity. Ultimately, the cost of equity will converge on the market beta measure if the firm goes public or is sold to a publicly traded entity.

3. Consider the use of debt and its cost. The absence of a rating should not be used as an excuse to use book interest rates or arbitrary costs of debt. As described in Chapter 6, “A Shaky Base: A ‘Risky’ Risk-Free Rate,” synthetic bond ratings can be estimated for any firm based on financial ratios that are available even for private businesses. Thus, an interest coverage ratio can be computed for a small business and used to come up with a synthetic rating and a pre-tax cost of debt (by adding the default spread based on the rating to the risk-free rate). The one additional adjustment we would consider making to this cost of debt is to add a spread to capture the small size of these businesses. It is likely that a bank would charge more for a BBB-rated firm with revenues of $1 million than for a BBB-rated firm with revenue of a billion.

4. Look at management proclivities and industry averages. There are some young businesses where the owners come in with strong views on using (or, more commonly, never using) debt. In these cases (and they are unusual), we can use the target debt ratio specified by management to compute the cost of capital. In the more common scenario, where the owners are unclear about how much they will use debt, especially as they grow, it is best to revert to the publicly traded firms in the business and use their average market debt ratio as the debt ratio for the firm being analyzed.

5. Build in expected changes in all of these inputs over time. As firms move through the life cycle, we should expect their risk and cash flow characteristics to change. In fact, we build in these expected changes in the earnings and cash flows we forecast. To preserve consistency, we should allow the cost of equity, debt, and capital to change over time. Thus, a firm that is all equity-funded and owned entirely by its founder, with a cost of equity of 30%, as a start-up, should not only see its cost of equity decline over time as it attracts more diversified investors into the mix. It also should be more open to the use of debt as earnings become larger and more stable.

With some young companies, you might find yourself unable to gain traction with the approach described previously, perhaps because they have no products or services yet or because their business model is in flux. Rather than spend your time trying to fine-tune your discount rate, we suggest that you adopt a much simpler approach to estimating discount rates that draws on the range of costs of capital across companies. Figure 9.4 summarizes the distribution of cost of capital for both U.S. and global companies, at the start of 2017, in U.S. dollar terms. (To switch currencies, just add the differential inflation rate in that currency, relative to the U.S. dollar, to each number.)

Thus, if you have to value a young start up, all you have to do is make a judgment (to go with the many others in valuation) on where you would put your firm in the distribution. Thus, if you are valuing a very risky young company with diversified investors and a global market presence in 2017, you would use a 10.68% U.S. dollar cost of capital, the 90^th percentile of the distribution for global companies. As with the earlier approach, you would adjust this cost of capital down over time toward the median of 8.03% for global firms as your firm becomes larger, more profitable, and more mature,

• We can value the firm as a going concern, making reasonable assumptions about cash flows growing in perpetuity. Chapter 2 noted that the terminal value could then be written as a function of the perpetual growth rate and the excess returns accompanying the growth rate (with excess returns defined as the difference between returns on invested capital and the cost of capital).

• If the assumption of cash flows continuing in perpetuity is too radical for the firm being valued, either because the firm is dependent on a key person or persons for survival, or because it is a small business, we can make an assumption about how long we expect cash flows to continue beyond the forecast horizon and estimate the present value of these cash flows as the terminal value.

• The most conservative assumption that we can make about terminal value is that the firm will be liquidated at the end of the forecast period and that the salvage value of any assets that the firm might have accumulated over its life is the terminal value.

Note that using relative valuation (multiples) to estimate terminal value, as is often the practice, is inconsistent with the notion of intrinsic value. Of the three approaches described, the right approach for estimating terminal value depends on the characteristics of the firm being valued. When valuing firms, where success translates into an initial public offering or sale to a publicly traded firm, the perpetual growth model makes the most sense. For smaller, less ambitious firms, where success is defined as surviving the forecast period and delivering cash flows beyond, assuming a finite life for the cash flows will yield the most reasonable value. Finally, liquidation value is best suited for businesses that come with time limits on their operating lives, such as an operating license that will end in five years.

• Sector averages: Earlier in the chapter we noted a study by Knaup and Piazza (2007). It used data from the Bureau of Labor Statistics to estimate the probability of survival for firms in different sectors from 1998 to 2005. We could use the sector averages from this study as the probability of survival for individual firms in the sector. For a software firm that has been in existence for one year, for instance, the likelihood of failure (from Table 9.1) over a five-year period would be assessed at 40.07% (the difference between the probability of surviving two years [64.85%] and the probability of surviving seven years [24.78%]). We are painting with a broad brush, in this case, and generalizing findings from a very specific time period (1998−2005) to all firms.

• Probits: A more sophisticated way to estimate the probability of failure is to look at firms that have succeeded and failed over a time period (say, the last ten years). Then we can try to build a model that can predict the probability of a firm’s failing as a function of firm-specific characteristics—the firm’s cash holdings, the age and history of its founders, the business it is in, and the debt it owes.

• Simulations: Chapter 3 noted that simulations can be put to good use when confronted with uncertainty. If we can specify probability distributions (rather than just expected values) for revenues, margins, and costs, we might be able to specify the conditions under which the firm will face failure (costs exceed revenues by more than 30% and debt payments are coming due, for example) and estimate the probability of failure.

After the probability of failure has been assessed, the firm’s value can be written as an expected value of the two scenarios—the intrinsic value (from the discounted cash flows) under the going-concern scenario and the distress value under the failure scenario:

No simple formula will help determine how many cash flows will be lost as a result of the loss of key personnel, because this varies not only across businesses but also across the personnel involved. One way to assess it is to survey existing customers to see how they will respond if the key personnel leave and then build this impact into operating forecasts.

To deal with the former, we suggest caution. Rather than add the cash balance from the most recent financial statements to operating asset value, we recommend obtaining an updated value (reflecting the cash balance today). To deal with capital infusions, we would revert to the concept of pre-money and post-money valuations that we introduced in the section on venture capital valuation. When we discount free cash flows to the firm, where reinvestment needs are treated as cash outflows, we are in effect computing a value of the operating assets, with no consideration for the cash that we might have on hand to make these investments. Adding the company’s prevailing cash balance yields a pre-money valuation of the firm:

If the firm raises additional capital in the form of either debt or equity, the portion of that capital infusion that stays in the firm (as opposed to being used by owners who want to cash out their ownership) augments value to yield a post-money valuation:

If you buy into the notion that pre-money value represents the value of the firm, with no capital infusion, and the post-money value, the value with the capital infusion, we would argue that rather than just using the capital infusion to separate the two values, we look at a richer way of measuring the value difference. For the pre-money valuation, you could estimate cash flows, growth, and risk, reflecting the capital access that the firm would have without the capital infusion. At one extreme, it is conceivable that this is the only source of capital for the firm and without it, it faces failure; the pre-money valuation here could be zero. At the other, it is possible that the private firm has access to multiple venture capitalists and that its cash flows, growth, and risk would be unaffected, if the specific VC capital infusion is not made. That would yield a pre-money value that is separated from the post-money value by the capital infusion. In the middle, though, are more interesting cases, where the company, without the capital infusion, might have to scale down its growth ambitions and settle for lower value and perhaps a different risk profile.

After we estimate the equity value, we can apportion the value between the option holders (in the convertible debt or elsewhere) and standard equity investors.

To value first claim on the cash flows from operations (preferred dividends), the simplest mechanism to use is to discount these dividends back at a lower rate than other cash flows to equity. This should lead to a premium for those owning these claims. The practical issue is coming up with an appropriate discount rate. If we accept the premise that preferred equity is similar to a debt issue, we can approach this question in much the same way that we estimate the pre-tax cost of debt. In effect, the risk-adjusted rate for fixed preferred dividend-paying equity is as follows:

The default spread can be estimated using an approach that is often used to estimate the cost of debt for nonrated companies. We estimate a synthetic rating for a company based on its financial ratios and then use that rating to come up with a default spread. In fact, one ratio that is widely used for synthetic bond ratings is the interest coverage ratio:

This ratio can be adapted to incorporate preferred dividends in the denominator (treated like interest expenses):

The resulting number should yield a synthetic rating for preferred stock. In turn, this can be used to estimate the default spread and the risk-adjusted cost of preferred stock. The resulting number should be higher than the pre-tax cost of debt, because preferred dividends are paid after interest expenses. But it should be lower than the cost of equity, because preferred stockholders get their dividends before common stockholders.

The question of whether the dividend is cumulative or noncumulative can be examined in this context as well, with the rate on cumulative preferred stock being lower than the rate on noncumulative preferred stock.

In some cases, preferred stockholders also get first claim on the firm’s cash flows on liquidation. Unlike dividends, which represent an ongoing claim, liquidation is a one-time event, and the valuation approaches we use reflect the difference. One approach to bring in liquidation cash preferences is to try to incorporate the likelihood of and expected cash flows from liquidation into a discounted cash flow model and arrive at a value today. The simplest way to do this is to create two scenarios. In the first, you value the equity claims assuming that the firm is a going concern. In effect, you assume that the cash flows (dividends or free cash flows) continue forever and compute the present value. In the second scenario, you assume that the firm will be liquidated at a specific point in time (say, five years from now). Then you compute the value of the equity claims on the firm based on the cash flows each period during the time period and the cash flows in liquidation. After the claims have been valued under both scenarios, you estimate the probability of each scenario (going concern and liquidation) and compute an expected value. This approach is predicated on the assumption that liquidation will occur only at the specified point in time and that the probability of its occurrence can be estimated with reasonable ease.

Veto power—the power to keep an event from occurring—does protect the rights of the equity claim endowed with the power, but it does so at the expense of overall firm value. By reducing the probability of a specific event (acquisition, initial public offering) that might increase overall firm value at the expense of a specific claim on equity, it reduces the expected value of the business and thus the value of all claims on the business. For instance, assume that a firm’s value, run by existing managers, is $10 million, the value to an acquirer is $15 million, and the probability of an acquisition is 40%. If the firm has only one class of shares outstanding, and there are ten million shares, the value per share can be estimated as follows:

Now assume that there are two classes of equity—5 million class A shares with no special rights, and 5 million class B shares with veto rights over acquisitions. As a consequence, the probability of an acquisition drops to 20%. The estimated value of equity per share reflects this change:

Note that the class B shareholders are costing the firm a million dollars in value. It is possible that they could negotiate to give up their veto rights for approximately that amount. Consequently, the value per class B share can be computed as follows:

The veto power that the class B shareholders have gives them a higher value than the class A shareholders, but they can monetize this value only if they are willing to give up their veto power.

Protective rights can be more complicated to value, because the right extends beyond the power to say no. In effect, the equity claim holders who have the right receive cash flows to compensate for the loss of value from the event. It is more akin to an option, providing protection against negative consequences, and it can be valued as such.

To see how DCF incorporates dilution, let us start with the recognition that the need for fresh capital is precipitated by having negative free cash flows. When valuing a young company, if you project out negative free cash flows to the firm for the next few years, you are also assuming that the firm will have to raise fresh capital during those years, with the proportions of debt and equity that you have used to compute your cost of capital. Thus, if you have assumed a 100% equity mix for your cost of capital, you are assuming that you will be raising fresh equity, by issuing new shares, for the negative cash flow years. That will increase your share count, but you don’t have to factor that increased share count in your value per share calculations today, for a simple reason. The value that you get for the operating assets today is the present value of future free cash flows, and that value will be reduced by the presence of negative cash flows in future years. That, in effect, is already a dilution adjustment and increasing the share count today for future share issues will be double counting.

To make the dilution effect more explicit, let us go back to the Secure Mail valuation in Table 9.19. Note that the expected free cash flows to the firm are negative for the next ten years, before turning positive and generating a terminal value. The value of Secure Mail’s operating assets, obtained by discounting all free cash flows and the terminal value back to today, is $177.56 million, but that value is depressed because of the expected negative cash flows during the forecast period. In fact, if we had not counted the negative free cash flows in years one through ten, the value of Secure Mail would have been $456.72 million. The cash burn is reducing the value of Secure Mail’s operating assets by 61.12%, which is our dilution adjustment. Dividing the estimated value ($177.56 million) by an augmented share count, to reflect expected dilution, will be double counting.

• Arm’s-length transactions: One of the perils of using prices from private transactions is that some of them are not arm’s-length transactions, where the price reflects just the business being sold. In effect, the price includes other services and side factors that might be specific to the transaction. Thus, a doctor selling a medical practice might get a higher price because he agrees to stay on for a period of time after the transaction to ease the transition.

• Timing differences: Private business transactions are infrequent and reflect the fact that the same private business will not be bought and sold dozens of time during a particular period. Unlike public firms, where the current price can be used to compute the multiples for all firms at the same point in time, private transactions are often staggered across time. A database of private transactions can therefore include transactions spread from June 2008 and December 2008, a period when the public markets lost almost 45% of their value.

• Scaling variable: To compare firms of different scale, we generally divide the market price by a standardizing variable. With publicly traded firms, this can take the form of revenues (price/sales, EV/sales), earnings (PE, EV/EBITDA), or book value. While we could technically do the same with private transactions, there are two potential roadblocks. The first is that some young firms have little to show in terms of current revenues and earnings, and what they do show might not be a good indication of their ultimate potential. The second is that there are broad differences in accounting standards across private businesses, and these differences can result in bottom lines that are not quite equivalent. The third is that in countries where private business owners are focused on minimizing taxes paid, it should come as no surprise that the numbers in the reporting statements (for taxes) bear little resemblance to the real earnings of a company.

• Nonstandardized equity: As we noted in the preceding section, equity claims in young, private businesses can vary widely in terms of cash flow and control claims. The transaction price for equity in a private business reflects the claims that are embedded in the equity in that business. It might not easily generalize to equity in another firm with different characteristics.

• Non-U.S. firms: Most of the transaction databases that are available and accessible today are databases of transactions of private businesses in the U.S. As we are called on increasingly to value young businesses in other markets, some of which are riskier, emerging markets, it is not clear how or even whether this data can be used in that context.

If we decide to employ private company transactions to value a young business, some general practices can help deliver more dependable valuations:

• Scale to variables that are less affected by discretionary choices: As a counter to the problem of wide differences in accounting and operating standards across private companies, we can focus on variables where discretionary choice matters less. For instance, multiples of revenues (which are more difficult to fudge or manipulate) should be preferred to multiples of earnings. We could even scale value to units specific to the business being valued, such as the number of patients for a general medical practice or the number of customers for a plumbing business.

• Value businesses, not equity: Chapter 4 classified multiples into equity multiples (where equity value is scaled to equity earnings or book value) and enterprise value multiples (where the value of the business is scaled to operating earnings, cash flows, or the book value of capital). Given the wide differences in equity claims and the use of debt across private businesses, focusing on enterprise value multiples is better than on equity multiples. In other words, valuing the entire business and then working out the value of equity is better than valuing equity directly.

• Start with a large dataset: Because transactions with private businesses are infrequent, starting with a large dataset of companies and collecting all transaction data is best. This allows us to screen the data for transactions that look suspicious (and thus are likely to fail the arm’s-length test).

• Adjust for timing differences: Even with large datasets of private transactions, timing differences will occur across transactions. While this is not an issue in a period where markets are stable, we should make adjustments to the value (even if they are crude) to account for the timing differences. For instance, using June 2008 and December 2008 as the transaction dates, we would reduce the transaction prices from June 2008 by the drop in the public market (a small cap index like the Russell 5000 dropped by about 40% over that period) to make the prices comparable.

• Focus on differences in fundamentals: The notion that the value of a business depends on its fundamentals—growth, cash flows, and risk—cannot be abandoned just because we are doing relative valuation. The estimated value is likely to be more reliable if we can collect other measures of the transacted private businesses that reflect these fundamentals. For instance, obtaining not only the transaction prices of private businesses but also the growth in revenues recorded in these businesses in the period prior to the transactions and the age of the business (to reflect maturity and risk) would be useful. We can explore the data to see whether a relationship exists between transaction value and these variables. If there is one, we can build it into the valuation.

• Life cycle affects fundamentals: If we accept the premise that only young firms that make it through the early phase of the life cycle and succeed are likely to go public, we also have to accept the reality that public firms will have different fundamentals than private firms. Generally, public firms are larger, often have less potential for growth, and have more established markets than private businesses. These differences manifest themselves in the multiples that investors pay for public companies.

• Survival: A related point is that young firms have a high probability of failure. However, this probability of failure should decrease as firms establish their product offerings. Firms that go public should have a greater chance of surviving than younger private firms. The former should therefore trade at higher market values, for any given variable such as revenues, earnings, or book value, holding all else (growth and risk) constant.

• Diversified versus undiversified investors: Earlier we discussed estimating risk and discount rates for young, private businesses. We noted the different perspectives on risk that diversified investors in public companies have, relative to equity investors in private businesses. That difference can manifest itself as higher costs of equity for the latter. When we use multiples of earnings or revenues, obtained from a sample of publicly traded firms with diversified investors, to value a private business with undiversified investors, we overvalue the latter.

• Scaling variable: Assuming that we can obtain a reasonable multiple of revenues or earnings from our public company dataset, we face a final problem. Young firms often have very little revenues to show in the current year, and many will be losing money, so the book value is usually meaningless. Applying a multiple to any of these measures will result in strange valuations.

Some simple practices can prevent egregious valuation errors and lead to better valuations:

• Use forward revenues/earnings: One of the problems we noted with using multiples on young companies is that the company’s current operations do not provide much in terms of tangible results: revenues are very small, and earnings are negative. One solution is to forecast the firm’s operating results later in the life cycle and to use these forward revenues and earnings as the basis for valuation. In effect, we will estimate the value of the business in five years, using revenues or earnings from that point in time.

• Adjust the multiple for your firm’s characteristics at time of valuation: If we are valuing the firm five years down the road, we have to estimate a multiple that is appropriate for the firm at that point in time, rather than today. Consider a simple illustration. Assume that you have a company that is expected to generate a compounded revenue growth of 50% a year for the next five years, as it scales from being a very small firm to a more established enterprise. Assume that revenue growth after year 5 will drop to a more moderate compounded annual rate of 10%. The multiple we apply to revenues or earnings in year 5 should reflect an expected growth rate of 10%, not 50%.

• Adjust for survival: When we estimated the intrinsic value for young firms, we allowed for the possibility of failure by adjusting the value for the probability that the firm would not make it. We should stick with that principle, because the value based on future revenues/earnings is implicitly based on the assumption that the firm survives and succeeds.

• Adjust for nondiversification: The value estimated for the firm or equity, based on future earnings and revenues, must be discounted back to the present to arrive at the value today. We can use the techniques we developed for adjusting the beta and cost of equity for private businesses in the intrinsic value section. We can discount for the forecast future value of the business by a high-enough rate to reflect the nondiversification of equity investors today. In effect, we are assuming that the firm will go public in the future year (where the multiple is applied) and that the nondiversification issue will dissipate.

  Valuing Secure Mail Software: Relative Valuation

  We will use publicly traded firms as comparable firms in the relative valuation of Secure Mail for two reasons:

  • It aspires to become a much larger firm and eventually go public.

  • Very few transactions involve young, private software companies.

  In coming up with a sample of comparable firms, we initially looked at only the three antivirus firms that we used in the venture capital approach—Symantec, McAfee, and Trend Micro—but we decided that we could not base a valuation on a sample this small. Consequently, we expanded our sample to include publicly traded software companies with a market capitalization less than $100 million. We regressed the ratio of enterprise value to sales at these companies against three variables: the beta (as a measure of risk), the expected growth in revenues over the next five years (to capture growth differences), and the return on capital (as a measure of the quality of growth):

  EV/Sales = 0.33 − 0.6 (Beta) + 7.6 (Revenue Growth) + 5.3 (Return on Capital)

  We then applied this regression to get a predicted EV/sales ratio for Secure Mail in year 5, using the following inputs as independent variables: the firm’s total beta in year 5 (1.60), the expected growth rate in revenues from years 6 to 10 (which is computed to be 21.2%, based on the forecasts in Table 9.8), and the return on capital in year 5 (estimated to be 13.31% in year 5, from Table 9.11):

  Predicted EV/Sales_{Secure Mail, Year 5} = 0.33 − 0.6 (1.6) + 7.6 (0.212) + 5.3 (0.1331) = 1.6866

  Applying this multiple to revenues of $686 million in year 5, we obtain a value for the firm of $1,157 million in year 5. We make two adjustments to get to value today:

  • We adjust for the probability that the firm will fail before the fifth year (40%) and arrive at an expected value:

    Expected Value=Estimated Value in Year 5(1−Probability of Failure)=1,157(1−.40)=$694.22 million

  • We discount the estimated value in year 5 back to today using the higher costs of equity that we estimated in the intrinsic valuation for years 1 to 5:

    Value Today=694.22(1.19)2(1.16)21.12=$325.29 million

  The final value that we obtain for the operating assets at Secure Mail, using this approach, is $339 million. It is higher than the intrinsic value, partly because we are ignoring the possibility of negative cash flows from years 1 through 5. Consequently, the number of shares we divide this value by should be modified to reflect expectations that more equity will be issued to cover capital requirements.

• New products: Success with an existing product or service can sometimes provide an opening for a firm to introduce a new product. A classic example is Microsoft building off the operating systems (MS-DOS and Windows) it developed for the PC to produce Microsoft Office, an immensely profitable addition to its product line. Another example is Apple’s introduction of the iPhone, which took advantage of the customer base Apple developed with the iPod. While neither new product (Microsoft Office and the iPhone) could have been predicted at the time of the original product’s introduction, the success of the initial product was clearly the launching pad for these offerings.

• New markets: In some cases, companies that succeed with a product in one market might be able to expand into other markets, with similar success. The most obvious example of this is expanding into foreign markets to build on domestic market success, a pathway adopted by companies like Coca-Cola, McDonalds, and many retail companies. The more subtle examples are products that are directed at one market but that serendipitously find new markets. An ulcer drug that reduces cholesterol is a good example.

Why can’t we build expectations about new products and new markets into our cash flows and value? We can try, but there are two problems. The first is that our forecasts about these potential product and market extensions will be very hazy at the time of the initial valuation, and the cash flows will reflect this uncertainty. In other words, neither Microsoft nor Apple would have been able to visualize the potential markets for Microsoft Office or the iPhone at the time they were introducing MS-DOS and the iPod. The second problem is that the information gleaned and lessons learned during the initial product launch and subsequent development allow firms to take full advantage of the follow-up offerings. This learning and adaptive behavior give rise to the option value.

1. Estimate the expected value and cost of going ahead with the expansion option today. The process of valuing real options begins with a fairly counterintuitive first step. We determine what the present value of the expected cash flows would be if we expanded into the new product today, and the cost of that expansion. In other words, this would have required Apple to consider the possibility and the potential cash flows of introducing the iPhone at the time it introduced the iPod. Many analysts resist making these estimates, arguing that they know too little about the potential product and market. But that is precisely where the option value is derived.

2. Assess the uncertainty in the estimated value of the expansion option. In the second step in the process, we not only confront the inherent uncertainty in the process, but we also try to measure this uncertainty in the form of a standard deviation in the value of the cash flows. There are two ways in which we can do this. The first is to fall back on a market-based measure. The standard deviation of publicly traded firms in the business could be used as a proxy. The other way is to run simulations on the expansion investment and derive a standard deviation in the value of the expected cash flow across simulations.

3. Determine the point in time when the firm will have to make the expansion choice. The option to expand into new markets and products cannot be open-ended. Practically speaking, there must be a date by which the firm must decide to either expand or abandon that option. In some cases, this time period might be a function of specified factors, such as a patent expiring or a license renewal, and in others it might be self-imposed.

4. Value the option to expand. The inputs to value the option are now in place, with the following pieces going into value. The present value of the expected cash flows from expansion, assuming we expand now, becomes the value of the underlying asset, and the cost of expansion today becomes the strike price. The standard deviation in value is the volatility in the underlying assets, and the life of the option is the point in time by which the expansion decision must be made. In theory, binomial option pricing models should work better at pricing real options, because they allow for early exercise, but the traditional Black-Scholes model provides reasonable approximations for most real options.

Consider, for instance, the two examples we used to illustrate the real-options argument in the first part of this section. Microsoft’s exclusivity in developing Office arose from its control of the operating system. Thus, it had a significant advantage over the competition (Lotus, WordPerfect) when developing its software. Apple’s exclusivity came from a reputation for innovation and coolness it developed with the iPod. Both were critical components in the adoption of the iPhone.

The allure of the real-options argument is the premium you can add to traditional discounted cash flow valuation. Some people push the use of this argument to its logical limit and beyond. Thus, some analysts argue that discounted cash flow valuations undervalue all young companies and that we should add option premiums to all of them. Other analysts mistake opportunities for options, using the real-options argument to add premiums to any company that has high growth potential. These range from technology companies in growing markets (software and alternative energy, for example) to small companies in large, emerging markets (such as India and China). In the process, they often double-count the value of growth—once through the expected cash flows in discounted cash flow valuation, and again when they add the premium. While real options are a powerful and effective tool for assessing value, they have to be used selectively. They can be used only when the expected expansion opportunities cannot be adequately captured in the expected cash flows, and when the company in question has significant competitive advantages over the competition.

This chapter has described processes that can be used to apply conventional valuation models to young companies. While these approaches require us to estimate inputs that are often difficult to nail down, they are still useful insofar as they force us to confront the sources of uncertainty, learn more about them, and make our best estimates. While we might be tempted to add premiums to these values for potential opportunities we see in the future, the use of real-options premiums should be limited to companies that have some degree of exclusivity in exploiting these opportunities.

___________________________

1. NFIB Small Business Policy Guide, Small Business Contributions in Small Business Policy Guide.

2. John Watson and Jim Everett, 1996, “Do Small Businesses Have High Failure Rates?” Journal of Small Business Management, v34, pp. 45–63.

3. Knaup, Amy E., May 2005, “Survival and longevity in the business employment dynamics data,” Monthly Labor Review, pp. 50–56; Knaup, Amy E. and M. C. Piazza, September 2007, “Business Employment Dynamics Data: Survival and Longevity,” Monthly Labor Review, pp. 3–10.

4. These returns are from a venture capital index maintained by Cambridge Associates, which tries to control for survivor bias by including liquidated partnerships. That correction only partly takes care of the survivorship bias problem.

5. As the venture capitalist, you would probably argue for an even lower number (Symantec’s multiple). To counter, the founder of Secure Mail would probably argue that his company will be priced more like McAfee.

6. Wikipedia has an excellent listing of industry trade groups, with links to each one. See http://en.wikipedia.org/wiki/List_of_industry_trade_groups_in_the_United_States.

7. The alternative is to use the average capital invested over the period. In keeping with the fact that we are using end-of-the year cash flows (rather than mid-year cash flows), we chose the capital invested at the start of each year.

8. Mr. Black has lived in the town for a long time and is a local celebrity.

9. Mr. Black has enough taxable income this year to claim the tax deduction immediately.

10. Although the kitchen can produce 60 meals a day, it is unrealistic to expect it to produce and sell this many meals every day of the year.

11. We use a sample of 12 companies involved in the security software business, rather than stick with the stricter sample of firms that just produce antivirus software. We assume a marginal tax rate of 40% applied to all of these firms.

12. The sample had 22 publicly traded restaurants, and we assumed a marginal tax rate of 40% applied to them.

13. Because this is a fresh bank loan, we assume that the bank is charging a fair interest rate, given perceived default risk.

14. This equation is a shortcut. You can obtain the same answer by estimating the cash flows each year for ten years and discounting back at the cost of capital.

15. Needless to say, the buyer will want Mr. Black to sign an agreement that he will not compete with the existing owner for the customer base.

16. This chapter offers a compressed version of how best to value cash flow and control claims. A more comprehensive paper on this topic is Damodaran, A., 2008, “Claims on Equity: Voting and liquidity differences, cash flow preferences and financing rights.”

17. It is unlikely that the venture capitalist will accept the higher valuation, unless he gets full credit for the increase in value, because it is his capital infusion that creates the increase.

18. The values we derive for d1 and d2 are as follows: