Presentation on theme: "Capital Investment in Information Robert M. Hayes 2002."— Presentation transcript:
Capital Investment in Information Robert M. Hayes 2002
Overview §Capital Investment in InformationCapital Investment in Information §APB 17 of the AICPAAPB 17 of the AICPA §Application to Database CompaniesApplication to Database Companies §Database as a Machine ToolDatabase as a Machine Tool §Determination of "Goodwill"Determination of "Goodwill" §Analogies to Evaluation of a DatabaseAnalogies to Evaluation of a Database §Application Of Standard No. 53Application Of Standard No. 53 §The Combinatorial Effect of Database SizeThe Combinatorial Effect of Database Size §ConclusionConclusion
Capital Investment in Information §This presentation considers a question that can be stated quite simply: How does one assess the value of the capital investment in information. §This question was an issue of concern during the deliberations of the Public Sector/Private Sector Task Force of the NCLIS when some members of the panel were adamantly opposed to the suggestion that a database should be viewed as a capital resource. Their view was that to do so would give it undue importance since, as they saw it, a database really has value only as a tool in support of other activities. The counter- argument was that, for an information company at the very least, its database was as much a capital investment as the manufacturing plant of an auto company. But that argument fell on deaf ears.
Increasing Importance of Information §In the years since then, efforts to assess the impact of information upon national economies at both a macro level and a micro level repeatedly have shown evidence that increasingly important expenditures are incurred in developing databases and creating other forms of information. Those expenditures simply were not being recognized. The result has been serious mis- measurement of phenomena of fundamental economic importance. It appears even in such areas as the balance of international trade.
Tangible vs. Intangible Assets §In principle, of course, the value of any company is a combination of tangible net assets and intangible ones, the latter usually being what euphemistically is called goodwill. The tangible net assets are directly derived from the books of the company as the difference between tangible assets and liabilities. Tangible assets include such items as accounts receivable, inventory, and fixed assets such as land, buildings, and machinery. The liabilities include accounts payable, loans and other financial obligations, contingencies for possible losses.
Accounting for Tangible Assets §Of course, there are subtleties in accounting for both tangible assets and liabilities, and in many cases they result in significant failures in properly representing the "true" value of a company. An example that comes to mind is of a company with substantial investments in real estate that are carried on the books at the costs of purchase but for which the current market value is far greater. Such a company is clearly the target for a hostile takeover with the intent of converting that real estate into immediate profit. But, recognizing those potentials for erroneous assessment of value, there are no fundamental difficulties with respect to the tangible assets and liabilities.
Accounting for Intangible Assets §In contrast, though, assessment of the intangible assets is fraught with difficulties. They arise as a result of accounting practices that reflect consideration of tax implications rather than operational value. First, let's look at the definition of intangible assets: they are "assets having a life of one year or more but which lack physical substance". Classical examples include patents and copyrights and they clearly reveal the role of investment in information resources.
APB 17 of the AICPA §APB 17, of the AICPA (American Institute of Certified Public Accountants) covers accounting practices for intangible assets, whether purchased or internally developed. The cost of intangibles acquired from others are to be reported as tangible assets and are to reflect the cash or fair market value of the consideration given for them. That seems straight- forward and surely will represent at least the value of such intangible assets as perceived at the time of acquisition. The book value of such assets will then be treated like that of tangible assets, typically with amortization over an expected period of benefit.
Internally Developed Intangible Assets §But the situation for intangible assets that are internally developed is significantly different. Their costs, for both developing them and maintaining them, are to be charged against earnings "if the assets are not specifically identifiable, or have indeterminate lives, or are inherent in the continuing business". Now, the reasons for this quite different treatment presumably reflect tax implications. The more that costs can be legitimately charged against earnings, the lower those earnings and hence the lower the taxes. It must be said, though, that the distinction between acquired intangibles and internally developed ones seems to be arbitrary, especially in the context of "out-sourcing" in which increasing amounts of things are being acquired rather than internally developed.
Application to Database Companies §Leaving that aside, though, let's now consider a database. While it will always exist in some physical form and thus might be treated as a physical asset, the physical form is literally immaterial. The costs incurred in producing it have virtually nothing to do with production in that physical form but reflect instead the production of information content. It is almost certain that the database will be regarded as an intangible asset. Now, it clearly is "specifically identifiable", but it also is likely to have an "indeterminant life" and, for the database company, is "inherent in the continuing business". Hence, the costs in developing and in maintaining the database are to be charged against earnings. Voila, so be it, and amen!
Database will NOT be among Assets! §In other words, given the accounting practices, in general any such intangible asset as a database will not be reflected on the books, so the problem in assessing the value of the company is how to treat such assets. Let's explore the issues involved in dealing with this problem. Aside from doing so with companies like ISI and OCLC, let's explore them as they relate to academic libraries. And there is a final context to do so as well – Internet companies for which their databases are the means for providing online services.
Intangibles in Software §Aside from the database itself, there are also intangibles in the form of the software developed to process it for production of products and services from it. Again the costs in creation of that software typically are also expensed and do not appear as assets on the books (although there is a new potential change to which in that respect to which reference will be made shortly).
Traditional Elements of Goodwill §Of course, aside from these issues, which are specific to companies that depend upon a database, there are the more traditional elements of goodwill, including such things as the ability to penetrate a given market, or having significant skills in marketing and management, or owning a patent on a process or a copyright on a publication. Each of those is not in itself a tool for production, so they will not be of concern here. There would seem to no reason to expect there to be any difference in treatment of them for any particular type of company.
Database as a Machine Tool §The facts are that the database, with related software, is literally the machine tool from which products and services are produced in information-based companies. §Let’s look at some examples: l Institute for Scientific Information (ISI) l OCLC l An Academic Research Library l An Internet Company
ISI §At the time of first sale of the Institute for Scientific Information, in 1971, the accumulated cost of its database was probably about half of the sale price. The important point is not that amount per se but the fact that the costs of the database were not represented on the books, and from an accounting standpoint, it had no value in the assessment of the company for purchase, despite the magnitude of the expenditures.
OCLC §In the same vein, let's consider OCLC, for which the database (WorldCat) is not only the machine tool but the very heart of its business, indeed the very reason for its existence. This is a fascinating example since the costs involved in creating WorldCat are a complex combination of expenditures by OCLC, by the Library of Congress, and by the great array of libraries that cooperatively have contributed cataloging data to it. The file today contains over 30 million records. Any proper accounting would assign a cost of at least $30 to the work in creating just one of those records.
Book value of the OCLC Database? §Beyond the catalog records as such, WorldCat includes data for each record identifying the libraries holding that item; the result is a database of value not only for the cataloging data but as the tool for access. There are several hundred million entries for holdings, each of which entailed a cost by the library entering it. But looking only at the cost of the cataloging, that alone says that the cost of the WorldCat database is nearly $1 billion. The books of OCLC show a value of nothing for it. In fact, they do not show even the small part of the costs that OCLC itself incurred in creating the database.
An Academic Research Library §As another example, for an academic library, its database (as represented in its collection and its catalog) is the very reason for its existence. Of course, there is no real counterpart of a corporate balance sheet for an academic library, but leaving that aside, there is also no representation of the costs represented by the collection or by the catalog in any reporting about any library. At most there may be a value of the collection established for insurance purposes, but it will not usually reflect the costs in selecting the books, in acquiring and processing them, in cataloging them, in maintaining them. In a very real sense those are all costs that could be regarded as creating the capital resource from which library services are derived.
An Internet Company. §Finally, if we consider any company engaged in providing Internet products and services, a key portion of its means for doing so is its database and, for many, it is the only means. The fact that its investment in creating the database in not recorded on its books makes it almost impossible to assess the value of that company.
Determination of "Goodwill" §Let's start with the representation of the determination the value of the database as part of "goodwill". One general and effective rule is that the total value of a company should reflect the present value of the expected future income stream to be derived from its operations. Within that, there are two means for determining goodwill: capitalization of earnings and capitalization of excess earnings, which are equivalent if the underlying estimations (for industry average rate of return, values of tangible assets, and average expected annual earnings) are equal. To illustrate, consider the following hypothetical data:
Differences between the Means §There is a variation that can result in a difference between the two calculations. Specifically, the approach using excess earnings permits one to demand a rate of return from goodwill different from that for tangible assets (typically, substantially greater so as to reflect relative uncertainty in the investment in goodwill). For example, if the rate of return from goodwill were chosen at 40% (instead of the 20% used for tangible assets), the goodwill estimated from the excess earnings would be only $25M. But for the purposes of this exploration, they will be taken as equivalent.
Ratio of “Goodwill” to Tangible Assets §One reasonable means for assessing the significance of goodwill is by the ratio of it to tangible assets. In the example above, it would be one-half of the tangible asset value, one-third of the total value. On the surface, all of this is reasonable, even unquestionable. But consider a database company which, quite typically, might have NO tangible assets. The ratio of goodwill to tangible assets would be infinite, 100% of the total value. The problem, of course, is that the database is not like what one would normally think of in considering "goodwill". It is clearly a means for production and, as such, is far more tangible than such things as the ability to penetrate a given market. or having skills in marketing and management, or even of owning a patent on a process or a copyright on a publication.
Analogies to Evaluation of a Database §The issue at hand, therefore, is how to value a database as a means for production? §As a starting point, let's consider some contexts analogous to that of investment in databases: l investment in research and development, leading to patents and copyrights, l investment in computer software, leading to publishable products, and l investment in motion pictures, leading to distribution through a variety of markets. §Each reflects substantial expenditures for information resources that serve as the means for producing products and services that generate income in a variety of ways from a variety of markets.
Research & Development. §Unfortunately, the first analogy leads into the same dead end. Indeed, in its Statement of Accounting Standards No. 2, October 1974, the FASB ruled that all research and development costs must be charged to expense when incurred. (There are two exceptions. One is when the costs are incurred for others, under contractual arrangements; the second is for extractive industries, such as oil exploration.) The argument in favor of this policy (aside from the evident tax implications) is that to capitalize investments in research and development would fail to recognize that future returns from them are uncertain.
Lack of Recognition of Importance §The problem, though, is that one of the most important elements of future growth in every industry is totally unrecognized in accounting, both for individual companies and for the nation as a whole. Beyond that, a patent or a copyright is not in itself a means for production, though it may give the right to engage in production.
Computer Software. §The second analogy is much more promising, though only recently having become so. In a Los Angeles Times article of Friday, October 29, 1999, it was reported that, "As part of its periodic update of its methods, the Commerce Department redefined software as an investment, something of value in its own right and thus counts (sic) toward economic output." It went on to say, "For last year alone, the rejiggering added about $250 billion to estimates of total economic output. Fully two-thirds of that was due to the redefinition of software." Now, that is getting somewhere!
Restructuring of National Accounts §Incidentally, the current "update of methods" has much wider significance, since it involves re- structuring of the national economic accounts, replacing the hoary SIC codes with new ones called NAISC (North American Industrial Classification System). This long overdue overhaul begins to recognize the role of information industries in the national economy.
What has been the effect? §It must be said, though, there this step of the federal government by no means implies that there will be immediate changes in accounting policies or practices in individual companies. Nor it is clear what aspects of the investment in software are to be capitalized, those of the producers of the software or those of the purchasers of them. For the former, the software is indeed the basis for production of a variety of products; for the latter, with some exceptions, software is more of a supporting tool than a means for production.
Is software a means for production? §Incidentally, the latter point reflects exactly the position of those members of the Public Sector/Private Sector Task Force who objected to treating a database (or, analogously, software) as a capital resource, and one must essentially agree with them in that respect, except for the exceptional cases. There are those exceptions, though, for which software really is a means for production by the company acquiring it. An example that come immediately to mind is imaging software which is vital in the production of television advertising, in a variety of kinds of motion pictures, in companies that depend on CAD-CAM software for engineering and architecture. For them, amortizing the imaging software makes eminent sense.
Potential implications? §It is possible, at least, that the new federal policy concerning amortization of software may eventually provide a basis for analogous treatment of expenditures for databases. For the moment, however, there are no accounting policies that provide doing so.
Motion Picture Production. §The third analogy, though, is quite relevant and there appears to be some basis for application of it to database companies. FASB Statement of Financial Accounting Standards No. 53 relates to "Financial Reporting by Producers and Distributors of Motion Picture Films". The copy at hand is dated December 1981 and, while it may have changed in the years since then, it probably reflects current policy and practice. Paragraph 10 states, "Costs to produce a film (production costs) shall be capitalized as film cost inventory."
Based on Ratio of Revenues §It goes on to state that the costs shall be amortized "in the same ratio that current gross revenues bear to anticipated total gross revenues." It is evident that the film, once produced, provides the means for producing further products and even services, and the accounting should recognize that by capitalization of those costs.
Primary vs. Secondary Exploitation §In an earlier, 1979, report, the AICPA discussed the background that led to FASB Standard No. 53 and, among other things, considered "allocation of production costs to primary and secondary exploitation". It defines the primary market as that for which the film was principally produced, with the secondary market as all residual exploitation. It comments, "When a film is expected to be exploited in secondary markets, it is appropriate to allocate a portion of production costs to revenues to be realized from secondary exploitation."
Application of Standard No. 53 §Let’s see the effect of applying Standard No. 53 to the same database contexts: l Institute for Scientific Information (ISI) l OCLC l An Academic Research Library l An Internet Company
Institute for Scientific Information l This would appear to be highly relevant to a database company. For ISI, as an example, production costs are incurred in creating the database entries from which successive issues of their various citation indexes are published. The primary market for each index presumably is in the sale of those current issues to the libraries subscribing to it.
Secondary Markets for ISI l But there are substantial secondary markets for those indexes. Historically most evident were the five-year cumulations. Indeed, it seems that the operations of the ISI index products first became profitable with issuance of the first five-year cumulation of Science Citation Index. Today, though, there are innumerable secondary markets in the sale of online services, CD-ROMs, and a variety of other potential products and services.
OCLC l In the case of OCLC, the production costs are incurred as a combination of efforts at contributing libraries and OCLC itself. The primary market presumably is in the subsequent sale of records for copy-cataloging by other libraries. Historically, that was the purpose for both OCLC and RLIN and it is still the primary source of income for them. l But there are substantial secondary markets for those catalog records. Most evident are use of them as means for adding holdings data which then serve as the basis for inter-library loan services. Others are in the sale of subsidiary products (such as CD- ROMs) and other kinds of online services, such as reference services.
An Academic Research Library l Turning to the academic research library, the situation presumably is quite different. These are entities that do not function as though they were profit making companies; they do not have balance sheets; they do not generally "market" their services; there is no reason to set a "value" on them, as there may be for a company being acquired. In other words, on the surface, there appears to be little basis for considering them in the context of the question of evaluating "database companies". Despite that, let's do so, with the following rationale.
Budget as equivalent to Income l Treat the budget of the academic research library as the equivalent of income from sales of services as represented and measured by circulation. (Measurement of library services by circulation alone is obviously simplistic but it will suffice for the purposes of this discussio and may even be rational if we assume that all other services are roughly proportional to it.) Thus, for the median ARL library, with a budget of $15M and circulation of 500K, the fee (i.e., cost) per circulation is $30. Considered as a company, the income and expenditures are in balance, so there is neither profit nor loss and taxes will be zero.
The Database in the Library l An academic research library spends a significant portion of its income in creating and maintaining the database, the collection of books and the catalog that provides access to it. That database is its very reason for existence. The yearly costs for creating that database include those for purchase of the materials, of course, but they will also include the costs in technical processing (selection, acquisition, cataloging, physical processing, and preservation). For the median library, the costs for purchase of materials will be taken at $5.5M and for technical processing staff, at $2.5M for a total of $8M. The remaining $7M are in direct services (circulation, reference, etc.) to the using public.
How the value the library’s database? l It is the set of costs ($8M) for purchase and technical processing that is at issue. l How are they to be treated? To answer that, we first must deal with an interesting question: What are the primary and secondary markets for delivery of services from the database? l There are a variety of potential answers to that question. The markets could be segmented by type of customers (e.g., faculty, students, other libraries, etc.) or by type of usage (circulation use, in-house use, reference use, etc.). l Either of those, however, becomes quite complex and probably difficult to measure. Beyond that, they do not provide ready generalization to other contexts.
Market Segmentation by Time of Use l Let's consider the time period of use as a relatively simple but provocative basis for segmentation. The primary market is defined as uses occurring within, say, the five year period after acquisition of material, and the secondary market, as all subsequent uses. l These are relatively easy to measure, at least for circulation uses. It is a fact of typical library operation that the great frequency of use of a given item occurs within that first five year time period so, in that sense, it is the primary market. It is also the case, though, that the very purpose of the academic research library is to assure that uses of an item subsequent from that period of maximum frequency will also be met, and those uses can properly be considered as the secondary market.
Differences in Time Periods l Underlying these definitions of primary and secondary markets is the view that uses during the two times periods are fundamentally different. In particular, during the early five year period, the uses are in the nature of environmental scanning, maintaining currency with what is going on today. Subsequent usage has more the nature of reference use, retrospective in nature.
Illustrative Numbers l For the median ARL library, the holdings are 3.5M volumes and yearly additions are 86K volumes. Thus the five year period would represent a total of 430K volumes or about 12% of the collection. Let's conjecture that about 25% of the usage of the library in a given year will be to that 12% of the collection, the remaining 75% being scattered across the rest of the collection. Then, the primary market would represent 25% of the income and the secondary market, 75%.
Amortization Period, Primary Market l The simplest rule for doing so would be linear, with the resulting amortization each year equaling the $2M so the books would show a steady state of about $10M for the capitalization account for the primary market. However, if the rules of Standard No. 53 were applied there would be a much more rapid amortization, since the general pattern is that use of materials drops exponentially during that initial five year period.
Amortization Period, Secondary Market l But how should the amortization of the capitalization for the secondary market be treated? Aha! A very difficult question. It could be argued though that there should be no amortization of that capitalization and that therefore the capitalization for the secondary market, the long-term market, will steadily grow. l The rationale is that the market for use beyond the first five-year period is essentially infinite so the income during any given year has a ratio of zero to the total market.
Long-term Amortization §Recognize, of course, that no library will have an infinite life, but the one at Harvard has been around for several hundred years and surely will be so for centuries to come. If one is uncomfortable with an infinite lifetime for a collection of books, we could amortize over three or four hundred years instead, with a result not dramatically different.
Division between Markets l Now, in this description there are some conjectured values (five years as the period for the primary market, 25%/75% as the split between primary and secondary, and infinity as the time period for amortization of the secondary market, in particular). Those clearly can be whatever they need to be to reflect reality and policy. The point is that, whatever those values may be, there is a basis for reflecting the investment in the database that has some rationale.
Online Database Services. l The same approach seems eminently applicable to segmenting the primary and secondary markets for online database services (although the infinite lifetime clearly is not at all appropriate). Typically, the primary market for such services represents immediacy of data and very current awareness. Stock market data are perhaps the characterizing example, with the value and therefore the fee that can be charged decaying literally within minutes if not seconds. But there is a secondary market for which immediacy is no longer significant but for which, instead, comprehensiveness is the crucial need.
Secondary Market l That secondary market partakes of much of the characteristics of the secondary market for the academic research library. It is likely to be at a virtual steady-state indefinitely into the future. However, the analogy with the academic research library fails because there is an important marketing strategy currently operating. It reflects the fact that immediacy is worth money, so the online service wants to attract customers that will use and pay well for access to the most current data. To attract those customers, the secondary market services are likely to be made available at no cost. Thus, while there is a secondary market it is not likely to generate enough income to warrant the separate accounting.
Combinatorial Effect of Database Size §Any rational rule for assigning value to a database will almost certainly be represented by a linear function of its size, and each of the examples presented above (ISI, OCLC, academic research libraries, and online database services) illustrates that fact. §But, in all of those examples, the operational value of the database is actually not a linear function of its size. As a database grows, the number of combinations that can be made from the data within in it increases. §Indeed, the effect is combinatorial and, as a result, exponential. If the size of the database doubles, its value more than doubles and may even be as much as four times as great.
Combinatorial Effect of Database Size §Just to illustrate, a real value of a major academic research library is evident when one must pursue a "chain of references", leading one after another through a succession of items from the collection. Only in a very large collection can such a chain of references be continued for very long. Clearly there is value in being able to relate, in this way as well as others, a number of items from the collection. As the size of the database grows, the ability to make such connections grows combinatorially.
Combinatorial Effect of Database Size §There appears to be no rational means for representing this combinatorial effect in a simple accounting process, and it is doubtful that there will be any formal recognition of it in assessments of the value of database companies. However, despite that, it is important to recognize the phenomenon, if only informally.
Conclusion §In summary, the purpose of this exploration has been to identify a problem and, by analogy with the motion picture industry, a potential answer to it. The discussion has been essentially qualitative and hypothetical, since the objective is not to solve the problem for specific contexts but rather to raise the issue and suggest that there may be appropriate means to do so.