Mention intellectual capital to most businesspersons and they think of traditional forms of intellectual property: patents, trademarks, copyrights, and trade secrets. However, companies must not only begin to recognize, as implied above, that intellectual property includes some very fuzzily defined assets—but that these assets may be among the most valuable in a company’s portfolio. The old adage "Knowledge is power" may be restated for the nineties as "Intellectual Capital is power," and in fact, it has, in an article by historian James Burke chronicling 1.5 million years of knowledge-based competitive advantage (Burke, 1997).

So powerful is the new intellectual currency—and so different from traditional assets—that it has created a whole new genre for business writers and academics: the digital economy. First introduced as the name of Don Tapscott’s 1996 book, this genre has become the latest rage in business writing, with books like Killer App appearing on a seemingly monthly basis.

While perhaps faddish, these books underscore an important point: that information technology is fundamentally changing the way companies do business; in fact, IT is causing us to rethink the very laws of traditional economics that govern business. In our lifetimes it’s likely that we will see a shift from the tangible-resource-based wealth that has dominated economics since men first traded stone ax heads for wild nuts and berries, to an information-resource-based wealth centered around intellectual capital (Stewart, 1991).

Defining intellectual capital and how to value it, then, is a fundamental step toward understanding the digital economy. Since traditional forms like patents, trademarks, service marks, trade dress, copyrights and trade secrets are already well defined elsewhere, I’ll focus on defining what I’ll call nontraditional intellectual assets. After defining these, however, I’ll broaden the scope of the paper to include both traditional and nontraditional intellectual assets, which I’ll alternately refer to as information assets (IA), intellectual property (IP) and intellectual capital (IC). While subtle differences between these terms will become obvious in context, they are all closely related to the broad notion of intellectual assets.

My intent here is not to provide definitive answers, but rather to bring together some ideas from the recent IC literature, in combination with some of my own, to give an admittedly limited view of how intellectual capital should be defined and valued—and what the proper purpose of the resulting valuations should be. I’ll begin by attempting to define information assets, then discuss how to inventory and value them, and conclude with some thoughts about the role of these valuations.

According to the Financial Accounting Standards Board (FASB), assets are "probable future economic benefits obtained or controlled by a particular entity as a result of past transactions or events" (King, 1997). Note that any mention of the form of those benefits is conspicuously absent from this definition. Despite Downes and Mui’s jab, even the accountants, it seems, don’t specifically exclude information as a potential asset!

But I’m not an accountant, so let me offer a working definition for information assets that builds on the FASB’s:

Probable future economic benefits obtained or controlled by a particular entity as a result of information or knowledge acquired through past transactions or events.

Thinking of intellectual capital in terms of economic potential is the most basic step towards our goal of putting a value on it, and one to which we shall return when we actually start looking at valuing intellectual assets.

Expanding on the economic benefits assumption, Arthur Andersen presents four prerequisite features for intellectual property (Bertolotti, 1995). According to them, it must be:

1. Readily identifiable and capable of being separated from the other assets employed in the business;
2. Non-physical in nature;
3. Capable of producing future economic benefits; and
4. Protected legally or through a de facto right.

Arthur Andersen is, of course, primarily an accounting consultancy, so it follows that they would use accounting-centric prerequisites to identify intellectual capital; the referenced article goes on to discuss valuing what I’ve termed traditional information assets, with hardly a mention of their nontraditional counterparts. I list these rules to illustrate two points: A. The arguably limited perspective with which information assets are currently viewed by the group (accountants) best positioned to recognize and account for them; and B. The ease with which the current perspective can be broadened to encompass information assets.

Regarding A., one of this paper’s primary motivations is to address whether accountants really need to worry about valuing information assets; we’ll tackle this after we address how to value them. To point B, however, notice that the only issues that might prevent Andersen’s definition from applying to nontraditional information assets are:

• 1. Identification and separation—because identifying information assets is often as much an art as it is a science; and
• 4. Rights of protection—because whether all the potential information assets we’ll explore here are protectable may be questionable.

Before we move on, I’d like to more rigorously define economic benefit. In the context of intellectual capital, I consider economic benefit to be economic value over and above that provided by fungible resources. If, for example, suppose I have a patent on drug A, but I’ve recently patented drug B, which is a more effective treatment than A for the same ailment and costs the same. After the introduction of drug B, though I might still be selling drug A, the existence of its patent doesn’t add any economic value since if I ceased production of A, drug B would simply absorb whatever market share drug A had managed to maintain. Similarly, if I employ unskilled minimum wage workers to mop floors, and the amount of time it takes to train a new worker to mop floors is trivial, I’ve lost no material economic value by losing an employee. On the other hand, if I hire unskilled workers who, due to substantial training and/or cultural indoctrination become more productive working for me than they would be working for my competitor, if one of my them quits, I have lost economic value—especially if the quitter takes his productivity and applies it working for my competitor.

Attempting to come up with a comprehensive list of information assets is a sisyphian task. In Intellectual Capital, Edvinsson and Malone list no less than 164 measures of a company’s intellectual capital, including such questionable items as "number of women directors" and "share of employees less than 40 years old" (Edvinsson, 1997).

I think this list is rather extreme, and certainly beyond the scope of this paper. However, I think their macro breakdown of IC, which originates with the Swedish insurance and financial services company Skandia, provides a useful framework:

Intellectual Capital = Human Capital + Structural Capital

Here, Human Capital (HC) breaks down into employee and corporate components: the employee side includes knowledge, talent, morale, creativity, skill, and foresight; on the corporate side are values, traditions, expertise, culture and philosophy.

HC is not exactly owned by the company; rather, you can think of the human elements as being rented, while the corporate elements are built up over time via a combination of human elements, structural elements, market conditions, and a certain amount of chaos and luck.

Structural Capital (SC), which includes "everything left at the office when the employees go home" breaks down into internal and external elements. Internal elements include hardware, software, databases, organizational structure, traditional IC (patents, etc), and everything else inside the company that supports the employees’ productivity. External elements include relationships with customers, suppliers, partners and other entities outside the company that provide it with some information-based advantage. Probably the fundamental differentiator between HC and SC is that SC can be owned and more straightforwardly traded (for example, it shows up as goodwill on the balance sheet of an acquirer), thereby making it more quantifiable.

Notice the inherent blurriness of the lines separating HC, SC and traditional tangible assets. One of the issues that makes measuring IC so frustrating is that it invariably relies upon and is often intermingled with tangible assets; the contents of a database, for example, are arguably information assets, but the hardware and software necessary to access those assets have a very quantifiable, depreciable accounting book value.

Even when strictly sticking to the Human Capital side of the equation, we often find overlap between the corporate and human components. For example, corporate philosophy is considered a strong intellectual asset by many companies; at Hewlett Packard, where I used to work, the "HP Way" provides a guide to how HP does business, creating consistent approaches that lead to valuable efficiencies across the entire huge organization. However, such an asset can only be formed by a complex combination of leadership, employee buy-in, and proof over time that it works. And while the HP Way has been proven over many years, we can’t even jump to the conclusion that long timeframes are required for corporate philosophies to accumulate value; relatively new companies like Diamond Technology Partners and Viant Corporation (both digital economy consulting firms) consider their relatively young corporate cultures to be prime assets, especially in the attraction of talent.

We find similar overlap within the individual components of HC. For example, strong leadership is often cited as Intellectual Capital (Malone, 1997), but think about the complexities of how this asset forms: not only do you need a talented, visionary group of leaders, but they cannot be effective without 1. cohesion amongst themselves and 2. the respect and allegiance of the rest of the workforce.

Now that we have some idea of what IC looks like, let’s examine how we might inventory it.

Before you can assign a value to intellectual capital, you must first determine if it’s present in your organization; assuming it is, you must then identify it.

The first task is relatively straightforward. From the perspective of an analytical finance purist, a company’s market valuation equals the "capitalized value of average earnings under a no-growth policy, plus PVGO, the present value of growth opportunities…" (Brealey, 1996). At first glance, this doesn’t seem to leave much room for the value of intangible assets. However, what allows a company to generate earnings and potentially grow in the first place? Certainly not just the brick and mortar assets it owns. Rather, it’s the synergies between those tangible assets and all the IC assets we described above: knowledge, creativity, culture, relationships, etc. Financial analysts use historic performance, market conditions, and future investment plans (along with a lot of guessing) to value PVGO; valuing IC is really just attacking the same problem from another angle.

So we are not in conflict with the finance folks when we say that if a company’s market value exceeds the book value of its tangible assets it must possess some intangible assets. Since market to book multiples for U.S. companies typically range from two to nine (Edvinsson, 1997), it’s safe to assume that virtually all U.S. companies have intangible assets.

We’ll now briefly reconsider traditional IC, since its value can sometimes be overlooked. Then we’ll move on to nontraditional IC.

Many companies already report intellectual property on their balance sheets; in fact, many have to, since it’s all they own. For example, upon its emergence from bankruptcy, by far the largest asset on the balance sheet of TriplEdge Products, Inc., was the patent for the windshield wiper that the company sold on TV. This patent was valued at $112,000, or 25% of total assets—even though the product based on it couldn’t provide enough cash flow to service the company’s debts.

Other companies on whose balance sheets one might expect to find astronomically valuable IC items often downplay their significance or omit them entirely. For example, Motorola, a company that depends heavily on patents and trade secrets, doesn’t specifically list them on its balance sheet. Even more counterintuitive, yet unsurprising given current accounting rules, Diamond Technology Partners, which generates revenues based only on the intellectual capital of its consultants, doesn’t make a single mention of IC on its balance sheet.

We will return to the issue of whether such assets should be included on the balance sheet later. Here, I’d like to concentrate on locating potential traditional IC items in the course of performing an IC inventory; I’ll assume for the sake of brevity that finding existing trade secrets, trademarks, service marks, copyrights and patents is a straightforward task.

Intuitively, one would expect a company to be most familiar with its patentable IC, since patenting inventions has become a more-or-less routine part of the R&D process for most companies that invent. A more interesting question arises when an unpatented item or process is found: should it be patented, thereby publishing its design for all to see immediately and anyone to copy in 20 years, or kept a trade secret, potentially preventing copies indefinitely (provided the secret can be kept, of course)? An analysis of this decision is beyond the scope of this paper; the important thing here is that potentially patentable subjects are identified.

Intuitively, we expect to find such subjects in R&D departments. However, they can exist in any part of the organization. In the case of trade secrets, for example, the 100+ year old "secret formula" for Coke is more of a marketing secret than an R&D secret. As another example, in E.I. duPont deNemours & Co. v. Rolfe, the trade secret covered the configuration of a plant. Both of these could have conceivably been patented instead of kept as secrets, but it’s interesting to note the significant impact that such a decision would have had on the value of Coke’s formula; had it become public in 1903 (17 years after its invention), it’s unlikely that Coke would have developed into as valuable a brand as it has.

The lesson here is that when searching for potential trade secrets and patents, it pays to look in "nonobvious" places besides R&D, including (Lott, 1997):

• Operations—are there any inherent competitive advantages in your manufacturing processes or in other ways that your business is run?
• Marketing—since virtually any information that is "capable of adding economic value to the plaintiff" is eligible for trade secret protection (Jorde, 1997), marketing data such as customer lists, competitive intelligence, future product plans, service agreements, pricing, and even the organization and compensation structures of the sales force should, if practical, be considered for trade secret protection.
• Accounting/Finance—even though public companies must periodically report financial information to the SEC, all companies have financial data that may benefit from trade secret protection. For example, product cost structure data may not hold an intrinsic value, but if allowed into the hands of a competitor could turn into a competitive weapon to be used against the owner.
• MIS—computer software, networking and support products and processes, even if they are developed for internal use only, should be considered for patent and/or trade secret protection if they give your business a competitive advantage.

Copyrights, along with trademarks and related Lanham Act designations, are different because the materials they protect are inherently exposed to the public; furthermore, in most circumstances there is partially or completely automatic protection provided under Lanham Act marks.

The automatic nature of the copyright makes it unlikely that any uncopyrighted material will turn up during an IC audit, but there is a possibility that internal documents may be productized and thereby made economically valuable. For example, technical assistance centers for high-end computers routinely write up fixes to common problems and send them out to customers; this eliminates repetitive calls, freeing up technicians for more productive work and saving money. These types of documents should be inventoried as IC.

Trademarks and service marks, on the other hand, are similar to patents in the way they are typically acquired: companies generally know when they exist, and usually design and register them on purpose. While section 43a of the Lanham act may provide "federal common law" protection for these marks along with trade dress (Jorde, 1997), companies may consider registering them to avoid potential future disputes over issues like priority. More apropos to the topic at hand, companies should ensure that unregistered marks are inventoried for proper valuation.

We now return to the fuzzy world of nontraditional IC. How can this stuff be inventoried? It’s certainly a lot more subjective than inventorying traditional IC; concrete guidelines are impossible to define. In this section, we’ll look at some general guidelines for identifying nontraditional IC; later, we’ll examine how to value both types.

When auditing nontraditional IC, we must always keep in mind our original definition and look for economic potential. Even if an information asset’s value is difficult to quantify, if it has value we should add it to our IC inventory. Beyond that, it will help to organize our attack with a framework; we’ll use Edvinsson’s IC=HC+SC.

Starting with the employee component of HC, some guidelines for identifying IC in your labor force follow. Keep in mind that while the presence of these attributes can increase your IC bottom line, their absence can decrease it; when taking inventory, note the opportunities for improvement as well as the presence.

• Knowledge—is the most common area of economic potential in human resources. In the context of IC, knowledge refers to a fluid mix of experience, information and insight that provides a framework for incorporating new experiences and information (Davenport, 1998). However, we’re not looking for ubiquitous or easily transferable knowledge; going back to our floor mopper example, in order for IC to add economic value, it must be nonfungible.
• Skill—is related to knowledge, but more task-specific and focused. Skill can provide valuable intellectual capital because it implies speed and confidence in performance of a job.
• Talent—is like an innate skill. Talented workers provide high IC value because they are quicker to learn, adapt, improve or transfer their talents.
• Creativity—implies an ability to approach a problem from several different perspectives when a solution isn’t obvious, or to find unique new solutions that are superior to the standard way of doing things. Creative workers provide high IC value by solving problems more efficiently than they otherwise would be solved.
• Foresight—also known as vision, foresight implies an ability to predict the future; usually it complements the other attributes covered so far. Foresight is a highly valuable intellectual asset, as it provides strategic competitive advantages which enable other forms of IC to continue generating economic benefits into the future.
• Morale—is one of those crossover attributes that often depends on group consensus, leadership, and/or structural capital. Since happy workers are productive workers, morale is economically valuable and can therefore be considered an intellectual asset.

Turning now to the corporate side of HC, with the same caveats we enumerated above:

• Values—are principles that provide ethical guidance, ensuring that all employees behave in a consistent, above-board manner. This provides valuable IC through positive PR, reputation building, and improved employee morale.
• Expertise—can be thought of as corporation-wide knowledge and skill. Expertise is acquired with experience in an industry—not only internal production and organizational experience, but experience with the industry’s customers, suppliers, and competitors. In an organization that makes efficient use of IC, both internal and external experience is assimilated and communicated throughout the entire workforce, creating highly valuable IC in the form of expertise.
• Culture—can be thought of as the personality of an organization. It also forms over time, as employees form friendships and identify common traits and goals. Culture serves two IC roles: 1. It boosts morale, since employees feel like they’re part of a family; and 2. It can be used to attract valuable employees away from competitors, assuming they find your culture more attractive than your competitors’.
• Philosophy—combines and drives all of the above. A corporate philosophy is a way of doing things that develops as a result of corporate values, expertise, and culture, and also influences the development of each. A good corporate philosophy can provide value by boosting all of the above attributes, but beware: a misguided philosophy—e.g. resulting from getting too comfortable with a certain expertise and therefore becoming unresponsive to the market, or letting a small group dictate the organization’s values—can strip away IC value faster than any other attribute.

Since internal Structural Capital is "what’s left when the humans go home," inventorying it is straightforward. It is useful to identify exactly which tangible assets facilitate the existence or use of exactly which intangible assets, so that you can find potential areas for improvement of the intangibles based upon their supporting tangibles.

External SC is also fairly straightforward: it’s all about external relationships. When attempting to inventory external SC, you must identify all external relationships that can potentially add economic value (or take it away). This is a task that is highly organization-specific, but some general guidelines follow.

• Customers—customer relationships can be an enormous information asset. This is obvious across a spectrum of high-touch businesses ranging from consulting or high-dollar fixed asset sales, where the relationship can make or break deals worth millions of dollars, to the doctor/patient relationship in a physician’s practice. However, even in commodity industries, improved information flow has some potential for economic benefit—if not in delivering information to the customer, in extracting information from him. For example, in the highly competitive gasoline industry, Mobil Oil’s Speed Pass alleviates the customer’s need to pay inside or even swipe a credit card at the pump; it’s economically free to the customer, but the application requires the customer to supply personal information that Mobil can then use for marketing purposes (along with the customer’s buying habits, which the pass makes easy to track). If the potential for such economic benefit exists, you can bet that if you don’t seize it, one of your competitors will. Therefore, the presence of such information assets should be inventoried, and the unaddressed potential noted.
• Suppliers—as with customers, supplier relationships can vary widely depending on the industry; however, in virtually all industries, they include opportunities to create or leverage IC. Some examples include supplier contracts that favor you over your competitors, and electronic information sharing (such as EDI) that gives you a cost advantage.
• Competitors—provide two opportunities for value creation through IC: inbound and outbound. On the inbound channel, an active competitive intelligence program can provide information assets that provide high value through a competitive edge. Take the mirror image of this and you can see how to generate value in the outbound direction: do everything practical to keep your competitors from getting valuable information on you.
• Others—include market entrants, substitutes, the community, special interest groups, and the government. All of these external structures can benefit from managing information flow and/or relationships in manners analogous to those outlined for the structures above.

Having defined our information assets and inventoried them, we’re now ready to tackle the thorny issue of determining their value.

Their valuation process consists of the following steps:

1. Determining from where the asset’s costs and benefits are derived. For example, is it a single patent? If so, did we invent it, purchase it, or are we licensing it? What exactly are we using it for? Or is it employee knowledge that can be attributed to training or on-the-job-learning? Or employee creativity that can be partially attributed to corporate culture? Where did the employee get this asset, and to what is he applying it?
2. Quantifying the costs and benefits. Once we have a handle on where the costs and benefits come from, we can quantify them, which will give us a base for our choice of the appropriate valuation method or methods.
3. Value the costs and benefits of the IC. There are several ways to estimate both costs and benefits depending on the IC in question and the circumstances.

To demonstrate this, let’s briefly digress to look at an extremely simplified application of the process.

Imagine you’re a very simple consultant, specializing in intellectual property law. You own a laptop worth $1000 and you just bought a suit on credit for $500 (you’re a rare, thrifty consultant). You’re also public; there is one share of stock in you that has been trading on the New York Stock Exchange for an average of $3000. Your balance sheet looks like this:

Your market to book ratio is therefore 3000/1500 or 2; the extra $1500 is the gross economic benefit of an information asset. In your case, its source is obvious: it’s the value of your impressive consulting brain power; your personal intellectual capital owing most likely to your consulting knowledge. If we further assume that you obtained this knowledge at a seminar presented by the world-renowned, universally acclaimed, and very tolerant intellectual property law Professor John Allison, and the seminar cost $500, this would be the cost and quantification of the information asset. You can then say that you’ve earned a net economic benefit (profit) on your intellectual asset of $1000.

I’ll call this a simple market to book valuation method. Unfortunately, valuing IC in real world companies is infinitely more complicated. The sections that follow go through the steps in more detail for the various types of IC we’ve discussed so far.

Determining the costs and benefits of traditional IC instruments is relatively straightforward.

For example, say they hypothetical Concave Computer Corporation sets out to design a new server. It can reasonably predict that the design will require several trade secrets and patents, the software and documentation will require copyright protection, and perhaps even the name will be trademarkable. Cost accounting and project management software can be used to track the development efforts allocated to each component over the course of the project; with sufficient preliminary foresight and diligence during development, this should provide reasonably accurate cost figures for each resulting piece of intellectual capital.

This method has the advantage that it can theoretically come closest to the actual cost of developing the asset. Of course, the disadvantage is that this is an idealization of an already idealized situation; in the real world, such perfect planning and tracking—even if the tools are available to do it—seldom takes place, and the boundaries between various components of a complex development project like a new computer architecture are seldom easily identifiable. However, if we want to accurately cost intellectual property, this kind of system must at least be our goal.

When development cycle tracking is impossible, the next best solution is probably the comparables method, in which the value of the intellectual property in question is estimated based on the value of similar technology being used in a similar application in a similar industry (Stiroh, 1998).

For example, let’s say Concave didn’t have the foresight to track the development costs on a patented memory subsystem used in their new server. To value it according to a comparable, they could find a case in which a similar patent was purchased or licensed for a similar application in the past and use its purchase or license price, adjusted for the provider’s profit, and for market and economic fluctuations that occurred in the interim. Alternately, they could determine what it would cost to outsource production of the component, or to license a comparable substitute today, adjust the cost as described, and use the result as an estimate.

The comparables method can provide reasonable estimates if an appropriate comparable exists for the intellectual asset you’re trying to cost. The drawbacks are that, for lack of complete data, you will almost invariably have to make assumptions about the profit margins—and even then, there’s no guarantee that the price you’re looking at was an accurate reflection of cost. If several decent comparables are available, such problems can be somewhat mitigated by averaging, but for most intellectual capital of any value, there will typically be fewer comparables rather than more.

When no other data is available, estimating the cost to produce the item may be your only alternative.

This method is similar to development cycle tracking, only instead of measuring the actual time and expense, it requires you to predict the expected time and expense using assumptions that are as close as possible to the actual situation when the IC was developed. Depending on the IC in question and your skill at predicting things, this could give a decent estimate, or it could be about as good as a wild guess.

Going back to Concave, if they had no alternative but to estimate production costs for the memory subsystem, they would do so by envisioning the state of their own expertise, as well as R&D costs (labor and materials as well as the fixed costs of the facilities and equipment), at the time the memory subsystem project began. They would then construct a project plan based on these assumptions which ignored any memory of how the actual project went. Once the plan was generated, it could be modified to include deviations that were agreed upon by developers who worked on the actual project.

This method, of course, has the disadvantages of relying on assumptions that under many circumstances may be difficult to make, and also of relying on developers’ memories of how the project went. However, for IC whose development cycle is consistent and quantifiable, this method can be fairly accurate; for example, if Concave wanted to cost the copyright for one of the server’s manuals, a fairly accurate estimate might be obtainable, since manual development cycles tend to be predictable.

Benefits are in some ways more straightforward and in some ways more difficult to determine than costs for traditional IC.

They are straightforward when either most of or very little of a product’s value to the customer is due to the IC component. When the IC component is the only reason the product sells—for example, TriplEdge wipers sell mostly due to their patented design, and we can assume a major reason for Coke’s sales is its brand—then any surplus economic benefit (after the fixed and variable costs of producing the product are covered) should be allocated to the IC component.

Likewise, even in a product with many IC components, benefit allocation can be fairly straightforward if one is dominant. For example, if the CPU, network adapter, I/O subsystem and memory subsystem on our Concave computer were all patented, but the primary selling feature was the memory subsystem due to its high speed, it would clearly be allocated a proportionate share of the economic benefit. To determine the exact proportion, Concave could estimate how many servers they’d sell with an off-the-shelf memory subsystem vs. how many they’d sell with an off-the-shelf CPU.

The problem is when more than one IC component in a product is a strong selling feature. In this case, vendor estimates of what might sell with various pieces of IC removed are not sufficient; some market research is needed. Such research could take one of two forms:

• Customer surveys—to force rank product features;
• Analysis of comparables—comparables can sometimes be used on the benefit side too. If another product exists that only possesses a subset of the features provided by your product, its acceptance, price and market share can be used to determine the benefit provided by the IC subset that it emulates.

We’ll now look at methods for determining the costs and benefits of nontraditional IC, using Edvinsson’s framework of IC=HC+SC.

Intrinsic HC includes talent, and often includes creativity, skill, knowledge and perhaps even some morale. Intrinsic here does not necessarily mean "born with;" it simply means that when Sally became your employee, she possessed the information asset in question. Depending on the employee’s past experience, learning ability and time with your company, then, all of her HC may be intrinsic.

Intrinsic HC can be costed by looking at the employee’s compensation, and also by what your company may have given up to hire her. If the choice was between a tangible asset and the employee, we expect the NPV of the employee’s intrinsic HC to be worth at least as much as the NPV of the foregone asset.

Costing by compensation requires analysis at two levels. First, the entire class of employees that does comparable work is likely paid a premium over "blank slate" employees who possess no HC and must be trained to do anything productive. Second, an individual employee may provide HC in excess of her peer group, and may be compensated accordingly.

Acquired HC can include everything but talent, but generally varies based on the employee’s amount of time at the company. Most employees (we might ignore high school kids in their first jobs) come into a position with some relevant (intrinsic) HC; they acquire more knowledge, and often creativity, skill and foresight as a result of learning from coworkers or taking training classes. Morale is sometimes intrinsic, but more often instilled by the work environment.

Company-sponsored classes are straightforward to cost, as they almost always involve a quantifiable expense consisting of the price of the course, incidentals if necessary, and the employee’s time away. Courses taken by the employee on his own outside of work and without the company’s sponsorship may also contribute to his acquired HC; for example, a recreational painting course might make an engineer more creative. Some companies, such as Ford Motor Company, provide employee reimbursement for such training. However, given the variability of the return on such courses (relative to the return on job-related training), this is probably better viewed as a benefit than an investment in HC.

Costing the on-the-job training, practice, and experience that produce knowledge as well as skill, creativity, foresight and morale is more difficult. To be rigorous, observation of employee work habits and their rate of ascension of the learning curve would be required.

Alan Benjamin, a retired British consultant, has actually developed a very robust framework for allocating HC costs—both training and compensation—between expense and asset accounts (Stewart, 1996). For example, under Benjamin’s system, all the salary of a clerk is expensed, because companies don’t invest in clerks for the future; on the other hand, some portion of a marketing staffer’s salary would be expensed—to cover the current value he brings to the company—and the other half capitalized, in anticipation of future value. Benjamin capitalizes all of the R&D staff’s salaries.

Expertise, values, culture and philosophy are difficult enough to define, let alone cost. However, we can make some reasonable attempts.

While they almost by definition must develop organically as an organization grows, the cost of values, culture, and philosophy may be quantifiable when programs such as team building and mission/value statement generation are conducted.

Another way to measure the cost is by looking at what choices they require you to make; if part of your corporate culture is to provide a laid-back work environment, and that means you must have a pool table in the office, then the cost of the table and the cost of the space it occupies, plus any quantifiable cost of the time employees spend playing pool instead of working, equals some part of the cost of the culture.

Expertise is a special case; since it’s manifested as an increase in efficiency as the company gains experience, it is typically zero-cost. If you really want to complicate things, you can look at the cost of not acquiring expertise; however, I’ll leave that as an exercise for the reader.

You may be able to measure intrinsic HC benefits by comparing a new employee possessing them to a peer who doesn’t, or by the overall increase in productivity you experience when you bring in a new employee possessing intrinsic HC.

Acquired HC can be easier to quantify if you actively measure for it. This would involve, for example, tracking an employee’s progress as she progresses up the learning curve, comparing performance at milestones along the way with the starting baseline. Company-sponsored training, especially when attended by groups of employees in the same area at the same time, provides an excellent opportunity to measure acquired HC: you know when to take the baseline measurement and how much the HC acquisition costs, and if a group participates, you can compare members as a test of intrinsic HC, or look at the group average to get a better estimate of future improvements.

As mentioned above, expertise is basically a free by-product of corporate experience, and it is fairly easy to measure by comparing current performance (preferably in a single constrained area) to historic performance. Expertise is not guaranteed to improve performance noticeably in all industries, especially in mature companies that have been producing the same product in the same way for a long time. However, even if expertise isn’t expected to have a noticeable positive affect, it should be monitored as an indicator of deeper problems in the case of negative affects.

Measuring the benefits of values, culture and philosophy is probably easier than measuring their costs. Improvement of these intangibles will often lead to improvements in efficiency and morale. We just discussed efficiency improvements, and morale can be measured to some degree through surveys, focus groups and interviews. These benefits may also show up in reduced attrition and improved communication.

Note that all of these measurement suggestions are subject to measurement noise and faulty assumptions; as hard as we might try to be quantitative and analytical about it, valuing the HC component of the IC equation is as much art as it is science.

Recall that internal structural capital is "what’s left when the humans go home;" virtually all of it is already reported under assets on the balance sheet, so neither its costs nor benefits should be calculated along with intangible assets. To keep IC strictly separated from tangible assets, if a certain information asset’s value is greatly enhanced by its supporting tangible asset, the extra value should be allocated to the intangible, leaving the tangible at book.

External relationships may seem like the most awkward form of IC to value; however, as with all the other forms, a little creativity and tolerance for ambiguity will allow us to at least get close.

The cost of building relationships that is common to all external SC components is time, which is, luckily, quantifiable. Here are some guidelines to help you quantify the time and other costs involved:

• Customer relationships—are, in most industries, the most important of the external structural capital components, because your customers are your direct source of revenue. Customer acquisition and retention costs vary widely from industry to industry, but in general, they can be quantified by 1. the amount of time contributed by sales, marketing, and/or service personnel and 2. the amount of money spent on marketing and advertising resources, including third-party resources like market research.
• Supplier relationships—can vary widely in importance depending on the ratio of suppliers to buyers, which is a function of both the economy and the industry. Costing supplier relationships, however, is similar to costing customer relationships: how much time is your purchasing agent spending on the relationship, and how much are you spending on peripheral expenses like delivery or payment terms?
• Competitor, substitute and new entrant relationships—going back to the inbound and outbound competitor relationship model, how much are you spending on a competitive intelligence function? How much on security measures to keep your information out of your competitors’ hands?
• Government and special interest group relationships—for most large companies, paying a lobbyist is an accepted part of doing business. Public relations staff and lawyers must also be paid to handle these kinds of relationships. And don’t forget the costs of PR event sponsorship.

As with most of the IC components we’ve looked at, measuring the benefits of external structural capital is far more haphazard than measuring the costs.

In most cases, history-based tracking is the best you can do. In the customer area, this should be familiar to marketing professionals, who often track sales over time to look for trends. Correlations between these trends and the introduction of new customer-relationship-impacting efforts can then hopefully be drawn. Similarly, limited marketing tests can also be conducted to gauge their benefits before moving to a full-scale launch of the effort.

Such metrics are useful for judging the benefits of programs aimed at customers, but other groups are less intuitive to measure. Initiatives to improve supplier relationships might be measured based on an increase in logistical efficiency, or a lowering of costs.

PR benefits can be quantified by public opinion and name recognition studies. As for the benefits of government relationships, a lobbyist’s job usually has some clear goals which can be measured.

Competitor, substitute and new entrant relationships are the most difficult to measure. Inbound competitive intelligence efforts can be measured by the degree to which they enable the salesforce to compete and/or the marketing group to improve your product’s competitive position. Unfortunately, the benefits of outbound competitive efforts are virtually impossible to quantify.

Assuming we’ve actually managed to use the framework provided in this paper to come up with some IC valuations, what do we do with them now? Put them on a balance sheet?

Hopefully, the last 18 pages or so have demonstrated that valuing information assets—especially the nontraditional kind—is anything but a simple task; but hopefully I’ve also demonstrated that it’s a task that’s doable, given appropriate attention. In this closing section, I’d like to examine some of the arguments for and against officially requiring that information assets appear in financial statements—and recommend a course of action.

Proponents of revising accounting rules to allow broader IC capitalization make some compelling arguments. Among them:

• As the advance of information technology continues to accelerate at breakneck speed and its presence in business nears ubiquity, more and more tangible assets will be replaced by information assets, and, in the words of Downes and Mui, "Everything you move from the world of atoms to the world of bits will [behave] the way brands and trademarks do today" (Downes, 1998). Proponents of IC capitalization argue that as more and more atoms are transformed into bits, businesses’ asset values will become less and less relevant—unless they include information assets.
• This trend can already be seen: 30 years ago, 25% of stock price changes could be correlated to changes in earnings; now that figure is less than 10% (Lev, 1997). Implication: traditional financial value predictors are becoming less and less useful—due to the increasing influence of IC.
• Where current accounting rules do recognize IC, they sometimes require that it be expensed immediately. For example, R&D costs must be expensed immediately; but what business activity is more about future payback than research and development? (Lev, 1997)
• Research shows a correlation between stock price changes and expenses like R&D and customer acquisition costs, indicating that investors are already valuing them as assets rather than expenses
  (Lev, 1997).
• There is a high correlation between acquisition prices (in mergers) and the cumulative R&D expenditures of the acquired company, again showing that investors already value R&D as an asset (Lev, 1997). In fact, in the acquisition of Macy’s department stores by Federated department stores, disputes over the value of Macy’s information assets led to the hiring of Trademark & Licensing Associates, a consulting firm specializing in the valuation of intellectual capital (Rutberg, 1995). Wouldn’t it have been easier if these assets were listed in the balance sheet according to some standardized principles?
• Software already exists to help companies value their intellectual property based on an audit of everything from copyrights to staff skill (Matthews, 1997).
• Companies like Swedish insurance company Skandia are proving that information assets can be measured—be measuring them. Skandia includes a 20-page IC balance sheet with its annual report, covering over a hundred measures of IC that they’ve developed (Matthews, 1997).

On the other side, proponents of the status quo provide their own list of compelling arguments:

• Despite the existence of IC-valuing software, and the evolution of valuation techniques, valuing anything beyond the currently-allowed traditional IC is just too subjective of a task to pass accounting’s conservancy principle.
• There may be a correlation between R&D expenditures and market value, but that doesn’t mean the market is smart: Harvard’s Michael Jensen claims that despite the market response, most companies get a negative return on R&D (Karlgaard, 1997).
• There is broad disagreement about current techniques for valuing traditional IC, and the techniques that exist are highly complex (Bertolotti, 1995). Take the Macy’s case cited above, for example; before bringing in the consultants, estimates of the value of Macy’s information assets ranged from less than $100 million to over $1 billion (Rutberg, 1995).
• Allowing such subjectively defined assets onto financial statements allows unscrupulous accountants one more variable with which to manipulate paper earnings. A company may be very smart, but that doesn’t necessarily mean it will be well managed or profitable (Rutledge, 1997).
• Current financial statements are the product of 500 years of evolution. Not only does this imply some beneficial natural selection in their contents, but also 500 years of evolution of techniques to interpret the statements; the rise of the importance of information assets is only about 20 years old. Valuing them may be a valuable exercise—but rushing to put them on the balance sheet goes against 500 years of evolution (Rutledge, 1997).
• The market has always comprehended the value of intangibles; that’s why there’s a disparity between a company’s book and market values. Why meddle with this age-old mechanism?
• By allowing IC into financial statements, companies would open up yet another avenue for government interference in business; perhaps this is why the "SEC loves IC" (Karlgaard, 1997).

I think both IC-reporting supporters and opponents would agree to a few points:

• IC is become very relevant to the value of a company.
• There is an unmet demand for IC valuation information.
• Current research is figuring out ways to make IC easier to quantify and value.

And I think these points of agreement suggest a potential compromise: voluntary reporting of IC valuation based on standardized valuation techniques.

Standard techniques need not be universally standard. In fact, I’d advocate a several-year trial period in which any techniques, as long as they were sufficiently documented, would be allowed; the market could then choose the best method for a given industry to standardize around. And it should be clear that even these "optimal" techniques should not be written in stone; they would be expected to evolve like every accounting practice introduced over the last 500 years. However, we do have to acknowledge that the rate of change in all sciences is now accelerating like never before, and accounting practices—especially those as technologically-correlated as IC valuation techniques—must adapt to the new rate of change.

Voluntary reporting will allow the market to show its opinion on the ultimate value to investors of IC value reporting: companies that do it should realize some market benefits, which will encourage other companies to start. And if the market decides that IC value reporting is worthless, it’s opponents will have a strong argument against pursuing it further.

Accounting, despite its decimal places and commas, is a fundamentally inexact science; if you need evidence of this, look at the number of companies with seemingly strong financials that suddenly find themselves in financial trouble.

However, given the amount of variability they have to deal with in industry, current accounting standards probably do as reasonable job as possible of providing comparability between companies, and this is really all we can ask it to do. After carefully examining both the mechanics of valuing IC and the pros and cons of reporting its value, I’m convinced that IC valuation is possible, valuable, and in need of improvement. I think my proposed voluntary reporting scheme will leverage the current possibilities while providing value for the companies that choose to do it and improving the state of the art in the process.