Wealth Creation

Consider a firm’s track record, as illustrated in Chapter 4, as representing a company’s up-to-date history. A forecast of a company’s long-term, future life cycle is an intuitive way to generate a future NCR stream—that is, NCRs implied by the forecast economic returns and reinvestment rates applied to today’s asset base. Figure 5.1 packages this process in terms of a warranted value that is calculated as the present value warranted by a particular forecast of a firm’s future life cycle and by the assigned discount rate. Although not necessary for the discussion in this chapter, a more detailed version of Figure 5.1 would specify NCRs from existing assets and, separately, NCRs from future investments.

As a total system, all the variables in Figure 5.1 are interrelated. How one specifies operating assets influences the calculations for economic returns and reinvestment rates. Consequently, the observed historical fade rates for economic returns and asset growth rates (proxy for reinvestment rates) also depend on the specification of operating assets. For example, the life-cycle track record for a typical pharmaceutical company is significantly different if R&D expenditures are, or are not, capitalized and included in operating assets. Also true, but less obvious, is that the assignment of a company-specific discount rate should be logically consistent with the NCR forecasting procedures used.

FIGURE 5.1 Life-Cycle Valuation Model

Figure 5.1 shows economic returns as one component of the process that generates NCRs. It is a gross understatement to say that economic returns are important to valuation. A fundamental wealth-creation principle is that investments that yield economic returns above (below) the investors’ discount rate, or cost of capital, create (destroy) wealth.

Let’s define more carefully what we mean by an economic return. Consider a project with cash outflows followed by after-tax cash inflows, inclusive of recovery of the value of nondepreciating assets at the end of the project’s life. The achieved internal rate of return for this project is its economic return. If the outflows and inflows have not been adjusted for changes in the purchasing power of the monetary unit, it is a nominal economic return. If all outflows and inflows are expressed in monetary units of the same purchasing power (e.g., dollars of purchasing power of a specified year), it is a real economic return.

Note that a nominal return of 8 percent with 0 percent inflation provides a real 8 percent return, whereas the same nominal return of 8 percent coupled to 8 percent inflation provides a 0 percent real return. This is quite significant. With an 8 percent real return, wealth doubles in nine years compared to no change for a 0 percent real return.

It is reasonable to expect that equity investors set stock prices with expectations of achieving a specified real return after anticipated payments for any personal taxes on dividends and capital gains (Madden, 1999, pp. 86-87; Sialm, 2006). With this line of reasoning, when investors experience an increase in their real tax burden, they should demand a higher cost of capital from corporations (stock prices drop) as compensation in order to maintain their real, net-of-personal-tax return goal.

Real economic returns for projects were at the heart of my model corporation work at CMA. Input to the model included period-by-period economic returns for specific projects and reinvestment rates. Output included balance sheets and income statements. The model corporation software was written to represent the firm as a portfolio of projects. Those projects had specified characteristics: economic life for depreciable assets, proportion of nondepreciating assets released at the end of the project, plus a real economic return that determined period-by-period, after-tax cash inflows. New investment in the form of capital expenditures and additional net working capital in each period represented the start of a new project to replace a project completed in that period. The amounts for new investment conformed to the specified real, reinvestment rate, which also involved targeted proportions for debt in the capital structure and dividend payouts from earnings.

This view of the firm as a portfolio of ongoing projects is depicted in Figure 5.2. A project consisted of an initial investment outlay (down arrow) followed by cash inflows (up arrows) over the life of the project, including a final release of any nondepreciating assets. The income statement at a point in time, such as 2008, represented cash inflows from prior projects that were still productive in 2008. The 2008 gross plant account consisted of past capital expenditures for not-yet-completed projects. Also on the balance sheet were nondepreciating assets, such as net working capital and land, from past projects.

The above perspective illustrates the commonsense intuition that the value of existing assets, at year-end 2008, depends on the wind-down pattern of anticipated cash inflows in years 2009 to 2011 (right-hand side of Figure 5.2). This figure also provides an intuition about the origin of the CFROI metric. The CFROI is calculated as a project ROI using aggregate data from the financial statements. The initial down arrow is gross plant plus nondepreciating assets, followed by equal cash inflows, or up arrows, over the average economic life of the assets, and a final up arrow for release of nondepreciating assets. It made sense to keep these variables in plain sight because of the multitude of accounting issues that can give a significantly distorted picture of economic reality (e.g., accelerated depreciation). For an updated technical discussion of CFROI returns, see Larsen and Holland (2008).²

The CFROI framework has proven very useful for identifying and solving issues in which accounting treatments differed from business economics (i.e., economic reality). A small sample of such issues include: capitalization of R&D expenses, operating lease capitalization, acquisition intangibles, financial subsidiaries, off-balance-sheet liabilities, special items, stock option expenses, franchise rights, asset impairments, and the list goes on.

Consider an economic perspective for the straight-line accounting treatment for depreciation charges. The productive capacity of plant and equipment does not decline nearly as rapidly as implied by straight-line depreciation (Thomas and Atra, 2009). This depreciation assumption operates behind the scenes in the calculation of a conventional RONA (return on net assets). Plant and equipment that is substantially depreciated yet fully operational is carried at full historical cost in the gross plant account but has a minuscule value in the net plant account. Based on gross plant value, the CFROI return is not distorted in such situations, whereas a conventional RONA is too high. Those who prefer the use of a RONA metric should calculate depreciation charges in a manner that overcomes the inadequacy of straight-line depreciation.

Much care is needed in analyzing how a firm’s business economics translate to the mathematics of a model’s present value calculations. Consider the handling of a firm’s existing assets. Recall that the warranted value of a firm is the present value of NCRs from both existing assets and future investments.

Quantifying the value of future investments implied in current stock prices is important and is calculated as the total market value of debt and equity less the estimated present value of existing assets. The proportion of a firm’s total market value due to future investments can be interpreted as an indicator of competitive advantage. The closer the estimated value of existing assets is to economic reality, the more accurate will be the implied value of future investments.

A relevant example would be the analysis of a potential acquisition of an oil and gas exploration company. Would it not make sense to first estimate the present value of the wind-down of NCRs from existing (proven) reserves? Next, separately assess the value of future investments from drilling new wells on owned or leased properties with more uncertain prospects and from new discoveries due to the exploration skill of the firm.

Mainstream finance has popularized two ways to treat existing assets that ignore the cash flow wind-down approach, yet are mathematically correct within the context of a particular valuation model. One approach is to estimate a normalized level of earnings and treat it as a perpetual annuity. This necessitates the assumption that depreciation charges are automatically reinvested every year in the future regardless of the level of economic return achieved. This also puts a considerable amount of future investments into the present value calculation of today’s existing assets. The other approach is to use a book value for today’s assets regardless of the level of economic return being achieved on those assets.

We should not lose sight of the role of a valuation model as a thinking template for strategic options. A focus on explicit cash flows from existing assets raises the relevant question as to whether certain assets might have a higher value to a different owner better able to generate higher cash flows from these assets in the future. For management in particular, the present value computations for existing assets should focus on the wind-down pattern of cash flows.³ The annuity approach and book value approach simplify the mathematics of a valuation model at the cost of obscuring economic reality for those who let their thinking follow the logic of their valuation model.

As a practical matter, is it worth the effort to strive for consistency in measurement units? An application of the model corporation addressed this question by inputting a time series of repetitive 6 percent real project ROIs and 2 percent real reinvestment rates, project characteristics similar to the average S&P 500 industrial firm, and similar financial leverage. One example illustrated in Figure 5.3 used data from 1875 to 1995 for nominal interest rates and for the GDP Deflator series to reflect levels of inflation/ deflation.⁴

Figure 5.3 plots the calculated CFROI returns from modeled annual financial statements, and these CFROI returns match the repetitive 6 percent real economic returns being achieved on all projects. For comparison purposes, the popular Earnings/Common Equity was calculated. As a levered ROI metric, one would expect its values to be a bit higher than the 6 percent real project ROIs. But the actual plot of the simulated Earnings/Common Equity is a wildly gyrating line going from a low of 3 percent to a high of 20 percent, essentially due to the variation in a single variable—the purchasing power of the dollar. Lessons learned from studies of long-term competitive fade, covering the time period of Figure 5.3, would be a bit misleading, to say the least, using a rubber ruler of Earnings/Common Equity as a measuring stick.

A few more points about this model corporation work relevant for security analysis in general and, in particular, for how finance students learn about valuation deserve attention. Those involved with discounted cash flow valuation models should gain absolute clarity on the calculation of net cash receipts (Madden, 1999, p. 68). NCRs can be calculated from the firm’s perspective (cash flows from operations less reinvestment). The identical NCRs can be calculated from the perspective of the capital owners (dividends, share repurchases, interest payments, and debt repayments less stock sales and new debt issuance). The above is important because “free cash flows” are often used as a substitute for NCRs and there are numerous definitions in use for free cash flow.

When the model corporation software is properly programmed, an accurate warranted value is calculated each period based on future NCRs. This can be verified in that the achieved equity investor return (dividends plus capital gains), year-by-year, equals the equity cost of capital. This is a useful exercise, for finance students especially, to see how the entire process checks out: forecasting life-cycle variables, generating NCRs, calculating warranted values, and proving that the present value computations are accurate. A particularly revealing model corporation exercise for students would be to calculate a CFROI return or adjusted RONA from as-reported financial statements that mirrors the repetitive real project ROIs. This requires one to understand how to make adjustments for inflation and deflation.

Simulation, along the lines of the model corporation work, deserves more academic attention in order to address the most difficult and important measurement challenges (e.g., intangibles) in connecting business economics to accounting data, and then to valuation. A good example is the simulation work of Healy, Myers, and Howe (2002) that addressed the complex issues of R&D capitalization.

This systems mindset leads one to the conclusion that the assignment of a discount rate is dependent on the procedures used to forecast NCRs. This is particularly relevant to models incorporating standard ways of forecasting future fade rates based on company characteristics.⁵ In contrast, mainstream finance relies on either a capital asset pricing model (CAPM)-based calculation for the discount rate or some other twist to the CAPM theory. These discount rates are then parachuted into valuation models without regard to how users make NCR forecasts.

A helpful approach to the topic of forward-looking discount rates is to observe how discount rates are handled in the bond market. For bonds, the anticipated NCR streams are composed of interest and principal payments. Knowing today’s price for a bond enables one to calculate a yield-to-maturity (YTM)—that is, the implied discount rate.

Consider a group of bonds with known credit quality ratings that are about to be sold to investors. Our objective is to estimate the discount rates that will be assigned by investors as implied in the soon-to-be-traded market prices for these bonds. These estimates could be obtained from a regression equation developed from a large universe of publicly traded bonds. For this universe of bonds, we would record YTM observations as the dependent variable and credit quality ratings as the independent variable.

For the group of soon-to-be-traded bonds, we could then use this regression equation to estimate forward-looking, market-derived discount rates. We can apply the exact same methodology to stocks, even though future NCRs are substantially more difficult to forecast for stocks than for bonds. Similar to credit quality ratings being the dominant variable determining demanded discount rates in the bond market, there are two variables for stocks that are logical choices and have proven to consistently have a dominant influence on the investors’ discount rates (Madden, 1998), which are weighted averages of firms ’ equity and debt discount rates.

The first variable is financial leverage. Note that CFROI returns are calculated using after-tax cash flows, which reflect the tax-deductible benefit of interest payments. But there is an offset to this benefit. As financial leverage increases, equity investors should require a higher return to compensate for a higher risk of financial difficulties. It is clear that high leverage played a significant role in crushing the common stock values of numerous firms during the 2008-2009 housing and credit crisis.

The second variable is company size, based on the plausible assumption that investors demand a higher return from smaller, less liquid companies’ stocks. This is due to both higher transactions costs in buying and selling positions in smaller companies and elevated business risk that cannot be diversified away.

A recent experience with Taiwanese companies serves as an excellent example of the benefit of a systems mindset for market-derived, forward-looking discount rates. In a systems approach, learning is a function of identifying problems and developing solutions by paying attention to interactions among variables. Along these lines, a (2006) Credit Suisse HOLT report, authored by Ng, Jhaveri, and Graziano, described a major improvement for Taiwanese companies.

Let’s begin with problem recognition. The aggregate market-derived discount rate for Taiwanese companies seemed implausibly high. Also, Taiwanese companies with low financial leverage had higher discount rates than the high-leverage companies—a result that did not make economic sense.

The root cause of these problems was identified as excessively high CFROI returns for the many companies that generously dispensed shares for employee stock bonuses. From the shareholders’ perspective, this outlay was clearly an economic expense, although it was ignored in computing accounting net income.⁶ This artificially boosted market-derived discount rates.

Figure 5.1 is helpful in understanding this point. Substitute a firm’s known market value for the warranted value. The market value can be matched by either one of the following:

• Discounting higher NCRs (boosted by ignoring employee stock bonuses) at a higher rate

• Discounting lower NCRs (this is more accurate) at a lower rate

The solution was to incorporate an appropriate charge for employee stock bonuses. This lowered cash flow used in calculating CFROI returns. With the new, lower CFROI returns (better reflecting business economics), calculated market-derived discount rates declined. Interestingly, technology companies were the biggest users of employee stock bonuses and these companies also tend to have low financial leverage. Thus, the CFROI fix also resolved the problem of a too -high discount rate for low-leverage companies. Finally, there was an across-the-board improvement in the tracking of warranted values with actual stock prices for all Taiwanese companies.

The accusation of illogic reveals an inability to think outside the CAPM framework. The life-cycle valuation model’s standard fade forecast for a typical technology company is much less favorable compared to that of a typical food company. A technology company with above-cost-of-capital, but highly variable, economic returns and/or high reinvestment rates would be assigned a faster downward fade compared to a food company, which typically has more stable economic returns and slower reinvestment rates.⁷ The life-cycle approach handles the “risk” difference in the numerator.

To sum up, there are three reasons for preferring some form of a market-derived discount rate instead of a CAPM/Beta discount rate. First, to repeat, a discount rate that is estimated consistent with the procedures for forecasting NCRs should be preferred over a CAPM/Beta discount rate that is essentially parachuted into any and all valuation models.

Second, a forward-looking discount rate should be preferred over a discount rate, such as CAPM/Beta, that is based on historical data and incapable of adjusting for near-term changes in the environment (inflation expectations, new tax legislation, etc.).

Third, application of the CAPM equation requires two inputs that are notoriously difficult to judge—Beta and the equity market risk premium over the risk-free rate. These are applied as forward-looking variables but they are necessarily estimated from historical data.

Depending on the past time periods selected, a stock’s Beta could easily range from say 1.2 to 1.5 and the market premium could easily range from say 4 to 7 percent. Users of CAPM have little to guide them in the selection of these two critical inputs. Combining a risk-free rate of 3 percent with a Beta of 1.2 and a 4 percent market premium yields a 7.8 percent equity cost of capital. In contrast, substitution of a Beta of 1.5 and a market premium of 7 percent yields a 13.5 percent equity cost of capital.

The valuation impact of using a 7.8 or 13.5 percent equity cost of capital is enormous. A similar big impact on an economic value added, or EVA^® (trademark of Stern Stewart & Co.), calculation occurs when the equity cost of capital is estimated with the CAPM equation or alternative procedures, such as arbitrage pricing theory or the Fama-French three-factor model that is increasingly being used by quantitative portfolio managers (Fabozzi, Focardi, and Jones, 2008). In practice, market -derived discount rates for a sample of companies have a much smaller range than CAPM/ Beta discount rates.

This process of improvement is a good example of the PAK Loop in operation. Critical to effective cycles of perceiving, acting, and knowing is the use of long-term charts that plot annual, warranted equity values alongside firms’ actual stock prices over time. These warranted value charts plus the valuation model (Figure 5.1) and the life-cycle track records (Chapter 4) are the basic tools for researching the causes of levels and changes in stock prices over time.

These three tools are used in an unending cycle of problem identification and resolution. Typically, a potential problem is observed as a systematic under- or over-tracking of warranted versus actual stock prices for a firm. The source of the problem is tracked down using the valuation components of Figure 5.1. For example, a firm’s per-share warranted equity values might be substantially below its actual stock price, year after year. Perhaps the accounting life (calculated as gross plant divided by depreciation charges) used as a proxy for economic life is clearly too short because of accelerated depreciation charges. Using a longer life that more closely fits economic reality would increase the value of existing assets. And it also would increase CFROI returns and lead to a higher value for future investments. This fix would move warranted values closer to actual stock prices.

Confidence in implementing a fix increases when there is a compelling economic reason for adjusting the accounting data. Also, confidence increases when other firms with the same issue show substantial tracking improvement after the fix is applied to their data.

The calculation, at a point in time, of a warranted value requires a company-specific discount rate and a long-term NCR forecast. The assignment of a discount rate has already been discussed. As to the NCR forecast, the more critical issue is the long-term fade pattern of CFROI returns. Think of a five-year window of future CFROI returns where the +1-year CFROI return is tied to security analysts’ EPS forecasts and the +5-year CFROI return is calculated from an assigned fade rate according to observed company characteristics. After the fifth year, CFROI returns are forecasted to regress toward the long-term cost of capital level.⁸

Life-cycle empirical research (Madden, 1996) on company characteristics and fade rates was based on the underlying assumptions about the interplay between managerial skill and competition described in Chapter 4.⁹ Here are some highlights of observed patterns and reasons based on life-cycle interpretations.

The higher the CFROI level, the faster will be the fade. This is due to competitors seeking to obtain a share of this wealth-creation opportunity. High CFROI returns coupled with high reinvestment rates lead to fast fade rates because (1) big reinvestment rates signal a big product market opportunity, which is especially attractive to competitors, and (2) the degree of difficulty in managing a business that is experiencing rapid growth necessarily increases. All else equal, above-average CFROI return firms with low year-to-year variability in their CFROI returns fade more slowly. This makes sense since the more controlled the firm’s operations, the more likely that higher managerial skill is involved. Finally, firms with above-average CFROI returns tend to fade down, those with average (cost of capital) CFROI returns tend to stick at that level, and below-average CFROI firms tend to fade up due to pressure on management to improve, shrink the business, or both.

Over many years of research and feedback from knowledgeable users, the procedures for forecasting NCRs have improved, leading to closer tracking of warranted versus actual stock prices. This process produces NCR forecasts that, on average, more and more closely mirror investor expectations. Consequently, as the universe of companies expands and contains better proxies for investor NCR expectations, the market-derived discount rates also become more accurate.

Keep in mind that these discount rates are attuned to the fade forecasting procedures used—a total system approach. Users of the valuation model are familiar with the standard fade forecasts based on company characteristics. When making their own judgments about a company’s future performance, they pay particular attention to the standard fade forecast versus the fade forecast implied in today ’s stock price as a gauge of the market ’s current degree of pessimism or optimism.

The learning tasks described above are rooted in fast and effective cycles through the PAK Loop. As described in the opening chapter, these cycles do not have a “start point.” One’s knowledge base affects how the world is perceived and what constitutes a problem that interferes with achieving a purpose or an anomaly showing a deficiency in an existing assumption or theory. Dealing with problems and testing hypotheses provides critical feedback that in turn affects one ’s knowledge base. This loop perspective avoids having to prefer an inductive or deductive methodology since both are at work in cycles through the PAK Loop.

In sharp contrast to the above, Robert Haugen’s quote at the beginning of this chapter accurately portrays mainstream finance’s reverence for elegant theory that sets the direction for subsequent empirical work. For example, many mainstream finance researchers rely heavily on top-down (deductive) theory such as market efficiency and the CAPM. In the past, the strong pull of this dominant view impeded experimentation with variables that might jeopardize the market efficiency and the CAPM constructs. Note how long it took for behavioral finance articles to be published in top-tier journals.

Modern finance researchers by and large have used the CAPM to guide much of their work. An elegant explanation of a mathematically logical relationship between expected returns on stocks and risk, the CAPM provides a blueprint, given its assumptions, for investors to optimize their portfolios to the highest expected return for a given level of risk. Notwithstanding the CAPM’s poor empirical record of predictability (Fama and French, 2004) and its challengeable assumptions, it continues to exert a strong hold on mainstream finance.¹⁰

Increasingly, behavioral finance researchers have presented serious challenges to the premises and empirical underpinnings of mainstream finance theory (Thaler, 2005). But proponents of the status quo seem little concerned about dealing with the weaknesses of their theory. Rather, they take the offensive, asking, Where is the better theory? Believing none has been offered as yet, the core body of knowledge presented in finance textbooks and taught to finance students has not been significantly changed. Thus the dominant theory remains intact.

Early on, the accepted goal of the life-cycle research program became to better understand levels and changes in company stock prices on a global basis so portfolio managers could make better investment decisions. In contrast, mainstream finance was focused on a logically consistent equilibrium model that related risk to expected return—and the CAPM became the answer. It was not designed to explain the level of firms’ market prices, but rather to explain the change in prices that drive investor returns.

Strong beliefs in market efficiency and the CAPM were never a part of the life-cycle work. Nevertheless, those involved with the research gained a great appreciation for the market’s ability to see through complex accounting issues and, most of the time, to astutely anticipate firms’ future profitability. An integral part of the life-cycle research was a knowledge-building process focused on anomalies. The research was designed to identify anomalies, gain an understanding of them, communicate the new knowledge to clients, and incorporate the findings into practical tools.

As previously noted, the research tools help to continually pinpoint situations where there is a patterned mismatch between actual stock prices and the warranted prices calculated with the existing algorithms for implementing the life-cycle valuation model. This often leads to learning how to better adjust accounting data to approximate economic returns. This in turn leads to improved discount rates and a sharper lens by which to identify new anomalies, perhaps concerning investor expectations. With this intensive, looped working with data within the context of a specified valuation model, every so often, a fundamental breakthrough would occur, such as a superior way to forecast long-term fade rates. Knowledge building (Gilbert and Christensen, 2005) through a systematic identification and study of anomalies (large differences between warranted versus actual stock prices) holds much promise for finance researchers.¹¹

Anomalies offer a fruitful path to improve the valuation model itself, or more likely, the calculation of input variables to the model. In terms of the PAK Loop, one begins with a purpose of evaluating and improving a particular valuation model. Anomalies are perceptions that cause problems for the existing knowledge base or model. This leads to a more penetrating analysis of cause and effect that, if successful, yields a change (action) that improves the tracking of warranted versus actual stock prices (consequence). In general, the testing and evaluation of alternative hypotheses provides critically important feedback that would not have occurred if the anomalies were treated as regression equation outliers and ignored.

Although the Wal-Mart success story is well known, the magnitude of Wal-Mart’s wealth-creation achievement is striking when displayed in life-cycle terms as seen in Figure 5.4. CFROI returns rose from 12 to about 15 percent from 1970 to 1990, coupled with enormous real asset growth rates. That remarkable performance was continually underestimated by investors and the stock outperformed the S&P 500 by 100-fold over that 20-year span.

In 1991, Wal-Mart’s stock price implied no downward competitive fade in both CFROI returns and real asset growth rates for the next five years. While possible, our experience suggested that at its much bigger size relative to the 1970s and 1980s, Wal-Mart was unlikely to meet those extremely optimistic investor expectations. The stock subsequently underperformed the market substantially from 1991 to 1996 (see bottom panel of Figure 5.4) as CFROI returns declined and asset growth sharply fell off.

At the time of the 1996 Forbes article, investor expectations were for Wal-Mart’s CFROI returns to rapidly fade downward over the next five years to a level close to the long-term corporate average of 6 percent CFROI returns. We felt comfortable in betting against an expectation that Wal-Mart was on the verge of becoming an average firm. This time, the stock subsequently rose sharply more than the S&P 500 during the next three years as Wal-Mart handily beat the 1996 expectations.

Although it is convenient to distill investor expectations into a single, best-estimate forecast, more rigorous analysis deals with warranted value as the expected value of a probability-weighted distribution of scenarios for future fade of economic returns and reinvestment rates (Alessandri, Ford, Lander, Leggio, and Taylor, 2004). Real options analysis is relevant for dealing with alternative scenarios, although application at the firm level is substantially more difficult compared to the project level.

To illustrate the concept of fade distribution, let’s return to Figure 5.4 and reflect on the process that produced such extraordinary excess shareholder returns during the 1970s and 1980s. At various times during this period, I analyzed Wal-Mart and decided not to buy it because I viewed the probability as low for a scenario in which Wal-Mart would maintain high CFROI returns while sustaining an extraordinarily high 25-percent-per-year organic asset growth rate. I was wrong. My mistake was in not sufficiently understanding Wal-Mart’s business model and exceptional managerial skill, which enabled the firm to perform so spectacularly as to drive its chief competitor, Kmart, into bankruptcy on its way to becoming the dominant retail company in the United States.

As for intangibles, it is widely agreed that this is an especially difficult and important challenge (Corrado, Haltiwanger, and Sichel, 2005; Hand and Lev, 2003). The conceptual accounting issue is invariably framed around the definition of an asset. Outlays for R&D, employee training, advertising, organizational changes to improve processes, and the like can certainly generate benefits well beyond the current accounting period, which argues for capitalization as assets.¹² But the issue is not so simple.

Accounting rule makers view the intangibles issue through a conceptual lens that serves up the following difficulties:

• Decide which outlays clearly will bring benefits in future years and should be treated as “investments” and recorded as assets on the balance sheet so that future revenues will be matched with appropriate expenses.

• Quantify the amortization schedules that reflect how the intangible assets will depreciate in the future.

Let’s frame the problem differently and focus on the valuation needs of investors and the notion that the more valuable the intangible outlay, the harder to quantify for accounting purposes. For example, outlays that enhance a firm’s knowledge-creation capability are especially instrumental in a firm gaining competitive advantage:

Unlike physical assets, knowledge assets are process rather than substance, and therefore in continuous change. They are an indispensable, internal resource for creating values that cannot be readily bought and sold. Much of a firm’s economic value is measured in explicit knowledge assets, such as know-how, patents, copyrights, and brand image, because they are easier to measure. But, in fact, these are the results of past knowledge-creation endeavors. The more valuable asset is the underlying tacit knowledge that was needed to create them because that knowledge and its methodology are the source of knowledge-creation capability at the firm and therefore the gauge of future value.

(Nonaka, Toyama, and Hirata, 2008, p. 42)

One approach to intangibles is for accounting information to include relevant details about intangible outlays so that investors could use this information in whatever ways are most workable for their valuation models. At any point in time, some investors might choose not to attempt to capitalize and amortize a particular intangible item, and instead adjust their long-term fade forecast for economic returns as a way to capture the impact of this intangible investment. Perhaps many investors would be comfortable with the capitalization and amortization of R&D expenditures as required by new accounting rules. Perhaps some investors with especially deep knowledge about certain industries would want to handle R&D differently.

This approach puts a premium on flexibility and learning and is conducive to a new accounting system evolving over time. Investors, provided with detailed information, would have choices in how to handle intangibles. Consumer choice and competition can work even for accounting standards.

The above is not an abstract idea; rather it is eminently practical due to the emergence of XBRL—Extensible Business Reporting Language (www.xbrl.org). With XBRL, accounting items are tagged with precise definitions enabling fine-grained analysis as opposed to aggregate accounting data (e.g., a net plant figure). With electronic access to company financial statement data, investors would be able to manipulate XBRL information about intangibles according to the needs of their own valuation models.

EFFICIENT MARKETS VERSUS BEHAVIORAL FINANCE

VALUATION MODEL PRINCIPLES

MEASUREMENT UNITS

FORWARD-LOOKING, MARKET-DERIVED DISCOUNT RATES

PROBLEMS WITH CAPM COST OF CAPITAL

IMPROVING THE VALUATION PROCESS

INVESTOR EXPECTATIONS: THE WAL-MART EXAMPLE

CRITICAL ACCOUNTING ISSUES

REPLY TO CRITICS