Subcommittee on Securities


Hearing on Adapting a 1930's Financial Reporting Model to the 21st Century


Prepared Testimony of Mr. Peter J. Wallison
Resident Fellow
American Enterprise Institute


10:00 a.m., Wednesday, July 19, 2000 - Dirksen 538


Mr. Chairman and members of the Subcommittee:

As we all know, the U.S. economy is constantly changing, and change seems to be coming more and more rapidly. Of course, we encourage these changes, because by and large they are the product of competition--as companies seek to function more efficiently, to use technology more effectively, and to serve their markets better than their competitors.

But large-scale change is a challenge for investors--individual and institutional--who have to assess whether these companies are successfully positioning themselves for growth and profitability in the future. To do this, investors need reliable information. As this Subcommittee knows so well, much of the regulation of the securities markets is devoted to assuring that investors are getting the best possible information with which to make their judgments.

Good information is vitally important, of course, because it reduces the uncertainty associated with making investments, and thus reduces one element of investment risk. Reduced investment risk in turn reduces the cost of capital. If capital costs are low, more capital will be available for companies that need it, capital will be allocated more efficiently, we will have faster and broader-based economic growth, and the welfare of all Americans will be enhanced.

Thus, there is little that policymakers can do in addressing the securities markets that is more important than assuring that investors have access to high quality, informative data that will enable them to evaluate the current positions and future prospects of public companies.

It may come as a surprise, then, to learn that--because of changes in economy--it has become quite difficult for investors to get all the information they need to evaluate companies. This is not because companies are withholding anything they are otherwise required to disclose, or that existing laws and regulations are not being vigorously enforced. It is because the nature of the assets on which most public companies now rely have changed so radically in the last quarter century that we literally do not have the skills or means to describe their value.

According to some estimates, about 80 percent of the value of companies listed in the S&P 500 is attributable to their intangible assets. These are familiar items such as patents, trademarks, and software, and less familiar items such as employee skills, customer satisfaction, and efficiency of product innovation.

However, conventional accounting has no effective means for recording intangible assets on balance sheets, and thus corporate balance sheets--prepared in accordance with Generally Accepted Accounting Principles (GAAP)--may simply not contain most of the assets a company holds. Worse still, since the purpose of accounting is to match costs with revenues--and a major aspect of cost is the depreciation or amortization of assets--earnings may be overstated or understated because the assets that are producing the earnings are not on the balance sheet and are thus not being depreciated or amortized.

I want to emphasize that this problem is not the result of a deficiency in the skills or imagination of the accounting profession. There is simply no reliable way to place a value on internally generated intangible assets. There is no market for intangibles which could be a reference point, and appraisals are subject to error and manipulation. Traditionally, accounting has relied on cost in order to place a reliable value on assets. But the cost of a patent or a trademark--let alone the cost of employee skills or customer satisfaction--bears no relationship to their value.

In the largest sense, this creates a discrepancy between the balance sheet net worth of companies and how the market might value them, because the market--knowing that balance sheets and income statements may not be accurate--is attempting to estimate what these companies are really worth, not what their financial statements show. So, we now find that the ratio of market capitalization to book net worth--which was 1-to-1 in the 1970s--is today about 6-to-1.

As an aside, this raises the interesting possibility that all the media talk about a "bubble"--or "crazy" price/earnings ratios--is misplaced. Price/earnings ratios obviously assume that the earnings are properly stated, but for the reasons outlined above they might not be at all.

In a narrower sense, however, this is not a healthy situation for the economy. Although companies report a good deal of information about themselves in text form, it is their financial statements which should be the best indicators of where they stand. If the financial statements are not fully relevant to answer their questions, investors are simply left to guess. Guessing of course is risky; it raises the cost of capital, promotes volatility, and ultimately distorts the allocation of capital.

An example of the problem I have been pointing to is furnished by AOL's recent settlement of an SEC charge that, in capitalizing its customer acquisition costs during the years 1994, 1995 and 1996, it adopted a misleading accounting treatment. When a company capitalizes a cost like this, it means that it is treating the cost as though it were an investment rather than an expense. In subsequent years, it amortizes or depreciates this investment, and this process reduces its earnings in the those later years. In capitalizing its customer acquisition costs, AOL claimed in effect that these costs were producing an intangible asset--customers who would use its services in the future--and that the proper accounting treatment would be to capitalize them when they occurred, and to write them off against the earnings from these customers in future years.

The SEC disagreed. It argued that by capitalizing these very large costs AOL showed earnings in 1994 through 1996, when--if it had written off these costs as incurred--it would have shown losses for those years.

In 1997, under pressure from the SEC, AOL changed its accounting treatment, so that it expensed its customer acquisition costs. Was this the right treatment? The answer--in light of later developments--is clearly no. AOL, as we all know, turned out to be a great success. In other words, its customer acquisition costs really were investments, since they produced a very large and profitable customer base. Because these costs were--at the SEC's insistence--written off as they were incurred, AOL's current earnings have been reduced since 1997 but future earnings were relieved of a burden that would have come from the amortization of these capitalized costs.

So if you were an investor in AOL in 1997, and you thought it was going to be a success, how would you arrive at a value for the company? You would believe that current earnings were unnecessarily suppressed by writing off customer acquisition costs as incurred, but that future earnings would be much higher as a result. In other words, if you guessed right about the value of the customer acquisition costs--a classic intangible asset--you would now be very happy indeed.

But notice that investors had very little information with which to decide whether AOL's customer acquisition costs were a solid investment for the company--that would pay big dividends in the future--or a waste of money. If they guessed right, and bought or held AOL, they would be sitting pretty today. If they guessed wrong, and sold AOL when it began to expense its customer acquisition costs in 1997, they are probably kicking themselves. It is the absence of any way of making this decision--other than as a guess--that is the significant point here, and I will return to it later in this statement.

Meanwhile, one question that immediately arises is: What does a price/earnings ratio mean in this context? If investors believed that current earnings were unnecessarily suppressed by writing off customer acquisition costs, they would conclude that future earnings would be higher than simply a projection of current earnings. This would result in a high price/earnings ratio in relation to current earnings. Thus, whether this classic intangible--customer acquisition costs--is an asset or an expense determines whether the company's earnings are overstated or understated, and what does that say about the focus on price/earnings ratios as an indicator of whether there is excessive speculation in the market?

And we can go a step further: what if AOL had turned out to be a failure. In that case, it might be argued that its customer acquisition costs should have been written off in the year incurred, because they were no longer properly classified as investments--they didn't produce the expected revenues. In that case, writing off the customer acquisition costs would have produced losses in the years 1994-1996, perhaps signaling to investors that this was a risky investment.

Unfortunately, this means that how an intangible should be treated depends on what happens in the future--clearly an untenable situation.

In reality, there are two questions here: what is the value of the internally generated intangible--in this case the customer base AOL was creating? And what is the future of the company itself?

When intangibles were a small part of company assets--when most of what companies used to earn profits were tangible assets like plant and equipment that had ascertainable values--the difficulty in valuing intangibles was not of much consequence. However, when substantially all of a company's profitability depends on intangible assets, the accounting problems associated with intangibles become quite serious. It becomes very difficult if not impossible for an investor to answer either of these questions with respect to a company that produces its revenues through the use of intangible assets. Again--keeping our eye on the larger question--it is important to recall that where there is uncertainty about value, capital costs are higher and the efficiency of capital allocation is reduced.

In other words, the almost 60 year effort to create a more efficient market--in large part by making sure that investors have the information they need--has been at least partially defeated by the advent of changes in our economy. These changes have made intangible assets the real drivers of company value, but this has happened before we have had an opportunity to develop the means to assess and communicate what these assets are actually worth. The result is an increasing discrepancy between what financial statements are telling us and what the market sees--and hence more uncertainty, more volatility, higher than necessary capital costs, and less efficiency in the allocation of capital in our economy.

There are also other inherent problems with financial statements that are worth noting at this point.

For one thing, they are inherently backward-looking. They tell us what happened to the company over the last quarter, or the last year, but not much about what will happen in the future. In fact, under some circumstances financial statement can be actively misleading about the future.

Take the case of Xerox Corporation. While Xerox was exploiting its patent--from the mid-50s through roughly the mid-70s--it was a very profitable company. Unfortunately, however, the copiers the company was producing, while in high demand because of their unique features, were highly unreliable and frequently needed repair. The company found that by selling the copiers instead of renting them it could make even more money--first on the sale, and then by charging for repair services. So until its patent expired, Xerox showed increasing earnings.

However, as soon as competitors were able to use its technology, Xerox was nearly driven out of business. The poor quality of its copiers had infuriated customers, and as soon as they had a choice they changed brands. In other words, even though its financial statements were showing healthy and profitable growth, the company was hollowing out. Investors who relied on Xerox's financial statements, and did not know the degree of its customer dissatisfaction, were in for a shock.

Here again, we encounter an intangible asset--customer satisfaction--that does not appear on a balance sheet and yet can be more important for predicting the future than what does.

In fact, if one audits conference calls between company managements and financial analysts, it turns out that relatively few questions are addressed to the company's financial statements. Most are about what is going to happen in the future, not what happened in the past. Thus, in a recent Pepsico conference call, there were 30 questions, of which 5 were about the financials. And in a recent Exxon Mobil conference call, there were 40 questions, about which 6 were on the financial statements. The other questions were about the future: "How do you see your North American growth as opposed to world-wide growth?" "Will canned beverage sales increase relative to other forms?"

In other words, financial statements, because they are backward-looking, are inherently deficient in the information that investors want most to know about--the company's future. What is needed is information that supplements the financial statements--that provides some indication of where the company is likely to go in the future.

Finally, financial statements suffer from one other inherent deficiency. They are issued periodically--quarterly or annually. Between these reports, the market is full of speculation about what they will contain. This speculation adds to uncertainty and volatility, and therefore to risk. The way financial statements are prepared--involving decisions by management and the aggregation of many different items into a relatively few lines--this delay is probably unavoidable. But that should not necessarily mean that all data on a company's operations has to be issued at periodic intervals.

In the past, when it took a while to assemble financial and other data, periodic releases were unavoidable. Today, however, when a good deal of information is available to management in real time, there is a question whether some of it could not be released more quickly.

In fact, the delayed release of financial data may be leading to earnings management--where companies coach analysts to reach conclusions concerning the company's earnings, and then companies come forth with earnings that slightly beat these estimates. If this is happening, it further suggests that investors' lack of confidence in financial statements is well founded.

These inherent deficiencies in financial statements have drawn the attention of the accounting profession. As early as 1991, the FASB issued a report--now known as the Jenkins report--concluding that information furnished by companies should be forward-looking and user-driven. As the problems associated with intangible assets became more pronounced, accounting firms themselves began developing ideas for ways in which companies might communicate the value of their intangible assets and how they were meeting their goals.

Two particularly detailed reports were prepared by PriceWaterhouseCoopers and Arthur Anderson. Both focused on the development of nonfinancial indicators that would communicate the degree to which a company is able to innovate, to satisfy and retain its customers, invest in and retain its employees, and increase the efficiency with which it is accomplishing its strategic objectives. All these categories are considered the "drivers" of future success, and information about them would be highly useful to investors.

On an official level, the Organization for Economic Cooperation and Development (OECD), which is a program supported by many of the major industrialized nations, has begun an intensive effort to find and develop nonfinancial indicators or measures that would permit investors to assess the prospects of companies or the value of their intangible assets. The indicators are nonfinancial in the sense that they involve information that does not appear in financial statements.

To take the AOL and Xerox examples, the hope would be that nonfinancial indicators or measures could be developed that would have allowed investors and analysts to get a better picture of whether AOL's customer acquisition costs were likely to pay off in the future, or that Xerox was incurring customer enmity rather than fostering customer satisfaction. In both cases, a great deal of uncertainty would have been eliminated, investors would have a better sense of how to value the intangible assets of both companies, and the securities markets would have functioned more efficiently.

In other words, many groups see the need for making useful nonfinancial information available to investors. But up to now progress has been slow.

The obstacles to the development of any set of nonfinancial indicators are formidable. First of all, these indicators are derived information, not simply raw data. Where they are numbers, they are presented as ratios, averages, or other kinds of compilations. For example, some have suggested an indicator for the quality of a company's patents that would report the number of times its patents are cited in later applications; the assumption is that the more citations there are the more original and higher quality the patent. Without arguing whether this is true, it is clear that disclosing an indicator of this kind requires some judgment by management. This raises the possibility of legal liability, which is discussed below.

Second, there is really no comprehensive theoretical framework for the development of this information. There hasn't been any significant testing of the efficacy of various indicators. There are no universal indicators that could be applied to all businesses, and very few for specific industries or activities.

Third, there is little interest among companies in proceeding down this road. Companies say that they are concerned that the information they will have to disclose will be helpful to their competitors, or that disclosures will result in legal liability. There is some merit in these concerns, but they may be somewhat exaggerated.

In reality, there are several areas where businesses are already cooperating in activities that are closely related to the development of indicators that would be useful for investors. First, many industries participate in a process called "benchmarking," in which they seek to develop the best practices by exchanging information about the way they conduct certain kinds of operations, such as employee recruitment and training. Testing whether these practices are effective involves statistical comparison of indicators.

Second, a number of industries are currently developing supply chain standardization, so that they can save procurement costs by creating accepted definitions for commonly used parts and services. This would enable a manufacturer, for example, to solicit bids for a particular part from a world-wide group of potential suppliers--all of whom would understand the specifications the manufacturer desires without having to meet and discuss them. The definitional problems associated with this activity is not far removed from what would be required to develop common measures or indicators.

Finally, for a number of years businesses have been developing, for internal use, indicators which tell management whether and how well a company is achieving its goals. These indicators are not shared outside the company--and there is no common set of indicators even for companies in the same industry--but they could be. In fact, a few companies do make these indicators public. Skandia International Insurance Company, a large Swedish insurer with a world-wide financial services business, has for almost 10 years been making public, in time series, the indicators it uses to measure the success of some of its subsidiaries.

To be sure, even if these internal indicators were made public they would still not permit the comparisons across competing companies that would be most useful to investors. Nevertheless, it would be a good start. It's important to remember that Generally Accepted Accounting Principles--on which we rely today to make comparisons among companies--hardly existed as late as the 1950s. Financial accounting is hundreds of years old, and for most of that time first individual firms, and then whole industries, had unique ways of accounting and reporting their results. Although we don't have hundreds of years--or even half a century--to develop the indicators that are necessary to supplement financial statements, a good place to start might be with companies making public the indicators they use themselves.

The fundamental reason for the slow movement in this direction is probably that companies do not yet believe that all the effort and cost involved in disclosure of new data will benefit them in any appreciable way.

However, there is data indicating that increased disclosure can have the effect of lowering capital costs. This stands to reason--since more information reduces uncertainty and hence volatility and risk. If this can be demonstrated to the satisfaction of companies--either through analytical work or by observing experience of others--a virtuous circle could result, in which successful companies disclose extensive amounts of nonfinancial information in order to achieve lower capital costs, and others must follow suit in order to remain competitive.

The issue for policymakers is how to stimulate development in this direction. The SEC has thus far exhibited no serious interest in promoting alternative or supplemental forms of disclosure. In a sense, this is good, because one of the fears of public companies is that SEC will come in with mandated disclosure requirements that will be costly to implement, misleading and possibly the source of legal liability.

Indeed, action of this kind by the SEC would be exactly the wrong way to get this process started. As we have seen in the past, mandated SEC requirements quickly stifle innovation. There are sufficient potential benefits for companies in the form of lower capital costs to believe that once the best companies start disclosing additional information--and seeing these benefits--others will feel compelled to follow suit.

On the other hand, the SEC could perform a valuable role without issuing mandates. It could encourage voluntary action--convening groups of companies, dealing with objections, seeking solutions that attract support, emphasizing that investors need information in order to make rational choices.

Having said this, there is clearly a serious question of potential legal liability when companies disclose information of their own devising--especially when investors may not have a clear understanding of what the information actually means. In this case, management is using its judgment about what to disclose and the content of the disclosure. This is exactly the kind of situation in which management's decisions can be second-guessed in the light of later developments.

There are at least two ways to deal with this. One of course is to create a safe harbor--similar to the safe harbor for forward-looking information. This would be quite difficult as a political and legal matter. Defining what is permissible, without weakening investor protections, would not be easy to do.

Another way--perhaps more promising--is to encourage the development of commonly understood indicators, based on clearly defined underlying financial or operating data, and accompanied by standardized explanatory information which outlines what it may or may not mean. Publication of information of this kind--assuming it is accurate--is highly unlikely to be regarded as deliberately deceptive or reckless, which is the statutory standard after the recent Securities Litigation Reform Act. Again, the SEC could be quite helpful in encouraging the cooperation of public companies in developing these commonly accepted indicators.

Thus far, I have discussed indicators that are derived or interpreted from operating or other data of public companies. But there is also a class of information that might be called data elements. These are raw financial or nonfinancial facts that generally do not require any interpretation or compilation. As I indicated above, the number of patent citations might be an example of an indicator; it requires the distillation of information from a number of sources. On the other hand, a data element might be a company's daily sales. This is the basic information a company uses to prepare its financial statements. Data elements can also be nonfinancial information, such as the number of employees.

This data can also be disclosed, some or all of it in real time. The development of the Internet makes instantaneous communication, at virtually no communication cost, entirely feasible. In part, this would address the problem of periodicity in financial statements. If the market were to have access to significant information in real time it would put to rest a lot of speculation between quarters.

But developments involving the Internet enable us to go a step further with this kind of quantitative raw data. A new Internet language known as eXtensible Markup Language, or XML, is now coming into common use. Up to now, information on the Internet has been stated in a language known as Hyper Text Markup Language, or HTML. HTML is basically a set of instructions to a display mechanism--a monitor or a printer--on how to display the document as a whole. It does not generally permit individual items of data to be identified and extracted from a document. XML, on the other hand, permits the tagging of individual items of data with definitions and context, so that they can be extracted from the document in which they are imbedded.

For example, let's assume that a company lists, in HTML, all twelve of the laptops it manufactures, together with their prices. If you wanted to know the median price of these laptops, you would have to print out or otherwise display the document, take off the price information by hand, and input it into another application that would compute the median price. However, if the same data were listed in XML format, your application could extract the price information directly. This is because in XML each data item has a tag that identifies what it is. In other words, next to the numbers in the list of laptops (but of course hidden from human view) would be a symbol that says, essentially, "this is a price." So when your application goes into the document looking for prices, it can find and extract them.

Clearly, this kind of disclosure is far different from the use of indicators. For one thing, companies would simply be disclosing factual information. Assuming it is accurate, there would be little if any legal liability associated with disclosures of this kind. Moreover, it is far less costly to develop this information than to develop the information used in indicators, since by and large it is information that companies maintain anyway--either to prepare their financial statements or for other business purposes.

There is still the question of providing useful information to competitors, but the question would be whether any of this data would provide as much useful information to competitors as the many things companies have to do "in the clear" -such as building plants, making acquisitions, or hiring skilled personnel.

The practical consequence of this for our purposes is that if a company were to make raw data available on the Internet--let us say in a data base--it would be possible for analysts to build models that extract the data they want and use it to forecast the company's prospects. This suggests that a whole new industry would develop--analysts who sell their services directly to the investing public. Right now there are certainly financial analysts, but generally of only two kinds--sell side analysts who work for the securities firms, and whose forecasts are suspect for that reason--and buy side analysts who work for institutional investors. There is generally no group of analysts who work solely for, and in the interests of, the general investing public.

At first, of course, the models that used various data elements would be fairly primitive, but over time the successful ones would survive and the unsuccessful would fall by the wayside. In the end, the public would have much greater, and more reliable, sources of information, and of course for this reason capital would be allocated more efficiently.

A good deal of work is already being done with XML. Most notably, the AICPA has been developing what it calls XBRL--eXtensible Business Reporting Language--which would theoretically enable applications to drill down into financial statements to the underlying transactional level, and to analyze the data then made available in real time. The initial users of this capability will be company managements, who will be able to get more data much faster, but at some point in the future some of this data could be made available to investors and analysts.

In summary, then, Mr. Chairman, there is a strong perceived need for information that supplements financial statements. This need is growing as the growth of intangible assets continues to reduce the relevance of conventional financial reports. Through the Internet there are now inexpensive ways that this information could be made available to investors, and a strong basis to believe that both companies and investors will be benefitted by such disclosure. What is needed, however, is the will among policy makers and businesses to proceed.


Home | Menu | Links | Info | Chairman's Page