Please help me to write a summary for attached article (about 900 words)

Accounting Horizons Vol. 22, No. 4 2008 pp. 453-470 American Accounting Association DOI: 10.2308/acch.2008.22.4.453 On the Balance Sheet-Based Model of Financial Reporting Ilia D. Dichev SYNOPSIS: FASB adopted a balance sheet-based model of nancial reporting about 30 years ago, and this model has been gradually expanded and solidied to become the required norm around the world. This article argues that the balance sheet orientation of accounting standard-setting is awed for the following reasons. First, accounting is supposed to reect business reality, and thus the essential features of the nancial reporting model need to reect the essential features of the underlying business model. However, the balance sheet orientation of nancial reporting is at odds with the economic process of advancing expenses to earn revenues, which governs how most businesses create value, and which represents how managers and investors view most rms. Second, the adoption of the balance sheet approach was driven by conceptual considerations; standard-setters argued that the concept of assets is more fundamental and logically prior to the concept of income. However, this article argues that the concept of income is clearer and practically more useful than the concept of assets, especially with the recent proliferation of intangible assets. Third, earnings is the single most important output of the accounting system. Thus, intuitively, improved nancial reporting should lead to improved usefulness of earnings. However, the continual expansion of the balance sheet approach is gradually destroying the forwardlooking usefulness of earnings, mainly through the effect of various asset revaluations, which manifest as noise in the process of generating normal operating earnings. During the last 40 years, the volatility of reported earnings has doubled and the persistence of earnings is down by a third, while little has changed in the properties of the underlying business fundamentals. INTRODUCTION urrently, the U.S. Financial Accounting Standards Board FASB and the International Accounting Standards Board IASB are reconsidering their conceptual frameworks. This reassessment is a key part of the two Boards' strategic plan for long-term convergence of the two dominant systems of nancial accounting since agreement on specic issues is difcult C Ilia D. Dichev is a Professor at the University of Michigan. Special thanks go to Steve Penman, who provided detailed comments and guidance on this project. Thanks to Bob Colson, Dan Givoly, David Ziebart the editor, and two anonymous referees for specic comments on the paper. The themes developed in this study have beneted from numerous discussions over the years with Greg Waymire, Steve Zeff, Shiva Rajgopal, Jim Ohlson, Sudipta Basu, Richard Sloan, Patty Dechow, and many other colleagues in the profession. The author gratefully acknowledges funding from the Center for Excellence in Accounting and Security Analysis CEASA at Columbia Business School and from the Ross School of Business. Submitted: August 2007 Accepted: July 2008 Published Online: November 2008 Corresponding author: Ilia D. Dichev Email: dichev@umich.edu 453 454 Dichev without shared conceptual foundations. The project is still in its early stages, and the ultimate form of the shared framework that is likely to be reached is unclear. On the table are foundational concepts such as \"assets,\" \"liabilities,\" and \"revenue recognition,\" which implies that the effect of the decisions about to be reached will impact the state of nancial reporting for decades.1 This new development is necessary and welcome, since uniformity and comparability of accounting standards around the world is a laudable goal. However, uniformity in itself is not enough, and in fact may turn out to be harmful unless it embraces the \"right\" principles for nancial accounting. The point is that the Boards' reconsideration of the conceptual framework does not include a reassessment of its balance sheet orientation, which has become the dominant feature of nancial reporting during the last 30 years. This paper argues that this omission is a grave shortcoming, and calls on the Boards to expand their effort to a more thorough reassessment of their conceptual framework. Specically, the main argument is that the income statement approach to accounting is the natural foundation for nancial reporting for most rms, and a disregard for this approach is bound to result in faulty accounting, no matter what desirable characteristics the rest of the nancial reporting model might have. In the rest of the paper, these arguments are developed and presented in more detail, concluding with some suggestions about what a good model of nancial reporting might look like. The broad goal in the paper is to spark a renewed interest and debate about these issues in the accounting profession at large. A more specic goal is to aid the current work of several bodies that aim to improve the nancial reporting model, such as the Pozen Committee's initiative to reduce complexity and otherwise address the diminished utility of nancial reporting.2 HISTORY AND BACKGROUND FOR THE CURRENT DEVELOPMENTS Some sense of history and a review of the balance sheet versus the income statement approach to accounting help provide a proper context. Accounting has a long-standing debate about two alternative and competing approaches to doing nancial reporting. The essence of the so-called balance sheet-based approach is that it views the proper valuation of assets and liabilities as the primary goal of nancial reporting, with the determination of other accounting variables considered secondary and derivative. The principal implication from this perspective is that the determination of income statement amounts and especially earnings is governed by balance sheet considerations. The balance sheet approach, taken to its logical conclusion and extreme, prescribes that the correct determination of assets and liabilities completely determines earnings, where earnings for a given period is simply the change in net assets over that period adjusted for distributions and contributions from equity holders. This view of earnings has strong underpinnings in economics, where it is known as Hicksian income.3 1 2 3 See Bullen and Crook 2005 and McGregor and Street 2007 for a more complete account of the scope of this project. The proposed new conceptual framework is discussed in more detail later in the paper. The formal name of the Pozen Committee is Advisory Committee on Improvements to Financial Reporting, known informally by the name of its chairman, Robert Pozen. This committee was formed by the SEC in June 2007. It is currently working on its agenda and plans to release its recommendations in August 2008. Proponents of the balance sheet model and fair value-type accounting frequently use the Hicksian idea of income as one of their key constructs and arguments. It is little known that John Hicks himself advocated historical cost accounting and recommended against including fair value-type estimates in nancial statements see Brief 1982 for a review of Hicks's views on accounting. In addition, note that the Hicksian denition of income requires that wealth be measured well at both the beginning and end of the requisite period. When there are difculties in measuring wealth e.g., when using fair value accounting for assets where no reliable market-based prices are available, the resulting change in wealth is not really \"Hicksian income,\" so invoking the Hicksian income concept becomes strained. Accounting Horizons American Accounting Association December 2008 On the Balance Sheet-Based Model of Financial Reporting 455 In contrast, the income statement approach views the determination of revenues, expenses, and especially earnings as the primary goal of nancial reporting. The emphasis here is on the proper determination of the timing and magnitude of the revenue and expense amounts, whereas balance sheet accounts and amounts are secondary and derivative. The two major guiding principles in this process are the principles of revenue recognition and matching of expenses to revenues. The goal of accounting is to record accruals, which properly record the timing of economic achievements revenue and the alignment of associated expenses matching. Balance sheet accounts and amounts are mostly the residual of this process, and assets and liabilities are in essence the cumulative effect of periodic accruals. For example, accounts receivable arise because revenue is recorded before actual collections, and the PPE account represents unexpired costs of acquired equipment. The income statement approach has always had strong support among accounting practitioners, and especially in the investment community, where to this day investment managers and nancial analysts primarily think of stock value as arising from the rm's ability to generate a stream of earnings, and therefore nancial reporting's goal should be the correct determination of earnings. Although an inherent conceptual tension exists between these two approaches, in practice nancial accounting has always been a pragmatic compromise between them. Historically, the income statement view of nancial reporting was dominant in accounting. By the rst half of the twentieth century, this view was rmly embedded in the theory and practice of accounting, reaching its epitome in Paton and Littleton 1940, an enormously inuential work that was later proclaimed \"the accounting book of the century.\" Looking back on this book today is fascinating and instructive because it is entirely built around income statement concepts and problems, and balance sheet considerations are relegated to a peripheral status. It should be emphasized that there have always been proponents of balance sheet and even fair-value approaches to accounting. For example, MacNeal 1939/1970 and Chambers 1966 can be viewed as early proponents of fair-value accounting for assets.4 On balance, though, it is fair to say that the income statement approach to nancial accounting dominated theory, practice, standard-setting, and pedagogy until the mid-1970s. An important new stage in the development of accounting was set in 1973 with the start of FASB's reign as the ofcial standard-setter in the United States. The predecessor of FASB, the Accounting Principles Board APB, was a part-time organization, with limited staff and resources. Thus, the emphasis in the APB's work was mostly on codifying and clarifying existing principles of accounting, with little ability or attempt to inuence nancial reporting in a major way. However, widespread dissatisfaction with the APB's work led to the creation of FASB as a full-time professional unit, which was much better funded and staffed, and was ready and able to pursue a more ambitious and activist approach to accounting standard-setting. For the interested reader, Storey and Storey 1998 provide an exhaustive account of these developments. Soon after FASB started its work, it became apparent that a piecemeal approach to standardsetting was not going to work, because absent shared conceptual foundations, internal inconsistencies, and even contradictions were inevitable across the standards that govern specic accounting areas because of changing Board composition, political pressures, and the effect of a myriad of other idiosyncratic factors. Thus, at the very dawn of its existence FASB embarked on an extensive project to provide a conceptually sound and workable foundation for nancial reporting and standard-setting. FASB received several rounds of input from its constituencies and was engaged 4 See Barth 2006 for a modern view on fair value accounting. Note also that the disagreement about the balance sheet versus the income statement orientation is only one aspect of the larger debates about the proper model of nancial reporting. For example, Watts 2003 suggests that conservatism is a valuable and necessary feature of nancial reporting, while standard-setters have largely eschewed it. Accounting Horizons December 2008 American Accounting Association 456 Dichev in vigorous internal and external debates, as it was clear that the nature of the conceptual foundation would have a major inuence on all future standard-setting activities, on the economy, and on the world at large. The Board quickly reached two conclusions: rst, that the income statement and the balance sheet approach to accounting are the two major alternatives for a nancial reporting model; and second, that if one wants to ensure conceptual clarity and internal consistency, one has to choose an alternative and avoid some sort of a muddled compromise between them. Against this background, in the late 1970s FASB reached a major decision. FASB concluded that the balance sheet approach is the only logical and conceptually sound basis of accounting and that therefore the balance sheet approach should become the cornerstone of standard-setting and nancial accounting in general. Storey and Storey 1998, Bullen and Crook 2005, and other accounts of this decision clearly indicate that the main reason for this conclusion was the perceived conceptual supremacy of the balance sheet approach. FASB's reasoning can be summarized as follows: Earnings is a \"change in value\" concept, and it is impossible to dene a change in value concept before one denes \"value.\" Thus, the determination of assets and liabilities logically precedes and supersedes the determination of earnings, which implies that the balance sheet approach is the natural basis of accounting. In contrast, the income statement approach is conceptually suspect because it relies on vague concepts such as matching. In addition, the implementation of the income statement approach results in deferrals and accruals, which create assets and liabilities of questionable substance. See Sprouse 1966 for an early and forceful exposition of this view. During the years that have followed, accounting standard-setters have expanded and solidied the balance sheet approach on several dimensions. First, older rules have gradually transitioned to conform to the new conceptual framework. For example, APB Opinion No. 11 from 1967, which prescribed an income statement approach for income tax reporting, was superseded by a balance sheet orientation in SFAS No. 96 in 1987 and SFAS No. 109 in 1992, and the change was specically motivated by a desire to conform to FASB's conceptual framework. Second, FASB has been increasingly adopting more pure and extreme forms of the balance sheet approach, particularly with the recent broad initiative for moving to \"fair-value\" accounting. As the name suggests, fair-value accounting afrms the primacy of market and market-type prices as the benchmark for value for company accounting. Specic examples of fair-value accounting include SFAS No. 133 in 1999 hedging, SFAS No. 141 in 2001 acquisitions and goodwill accounting, SFAS No. 156 in 2006 securitization, and the recent sweeping SFAS No. 159, which allows fair-value accounting for a broad class of assets and liabilities. As a subject, fair-value accounting is signicant enough to merit separate and deeper study, and the interested reader is referred to more focused efforts in this direction e.g., Barth et al. 2001; Barth 2006; and especially Nissim and Penman 2007. In this article a broader perspective is taken; fair-value accounting is viewed as just the latest and more extreme form of a longer evolution that has been gaining momentum for the last 30 years. The balance sheet approach has also expanded geographically, moving from its U.S. roots to international standard-setting, and in the process becoming the dominant world-wide accounting doctrine. Since it began, FASB has been a model for international standard-setting, where various foreign bodies have sought to emulate the success and power that FASB exerts in the United States. Specically, the International Accounting Standards Committee IASC was founded in 1973, and the conceptual framework it issued in 1989 was heavily based on FASB's, adopting the balance sheet model of reporting Camfferman and Zeff 2007, 259-264. The IASC was eventually replaced in 2001 by the International Accounting Standards Board IASB, largely modeled Accounting Horizons American Accounting Association December 2008 On the Balance Sheet-Based Model of Financial Reporting 457 after FASB's organization and process, and cross-sharing with it a number of key personnel.5 Recently, the two Boards have actively sought to coordinate their philosophy and activities, adopting in 2002 a formal memorandum known as the The Norwalk Agreement, which details their joint commitment to convergence of U.S. and international accounting standards. Since such convergence is possible only under shared conceptual foundations, the two Boards share a rm commitment to the balance sheet approach. Thus, while this paper focuses on the U.S. and FASB experience, the following analysis has much applicability to the international domain as well. On July 6, 2006, FASB issued a document called Preliminary Views see Financial Accounting Series 1260-001, FASB 2006, which summarizes the Boards' current thinking about the conceptual framework and asks for feedback. Preliminary Views represents an initial step before issuing a more formal exposure draft, but it already contains a number of important indications about the shape of the future conceptual framework. For example, the conceptual framework is envisioned as a single unied document, along the lines of the IASB's existing framework rather than being like FASB's collection of seven concept statements. Also, the conceptual framework will have a high standing in the hierarchy of nancial reporting rules, and specic standards will be expected to comply with and embody the framework. For the purposes of this paper, the most important aspect of the Preliminary Views document is that the Boards have decided to concentrate on ironing out existing inconsistencies within and across the two systems of nancial reporting rather than on a deep review and rethinking of the entire existing conceptual foundation paragraph P7. In fact, Preliminary Views presents a strong endorsement of the balance sheet model of nancial reporting; the authors envision a further deterioration in the status of the income statement and earnings in particular. The magnitude and even the language of the proposed changes are startling, especially considering what steps would likely be taken to esh out the proposed conceptual framework in specic nancial accounting standards. For one thing, the Preliminary Views largely avoids using terms such as \"revenues,\" \"expenses,\" \"earnings,\" and \"income,\" mentioning them only as a nod to existing conventions paragraph BC1.30.. The preferred terms in Preliminary Views are variations on \"changes in economic resources and claims to them\" paragraphs OB22 and BC1.28., a language that embodies pure-grade balance sheet accounting.6 A more substantive and indeed critical change in the proposed framework concerns the importance of earnings in nancial reporting. Note that FASB's existing conceptual framework has always had a somewhat ambivalent attitude to earnings because on the one hand it endorses a balance sheet model of accounting, but on the other hand it also has a formal statement about the primary importance of earnings in nancial reporting. Specically, FASB's Concepts Statement No. 1 paragraph 43 says: The primary focus of nancial reporting is information about an enterprise's performance provided by measures of comprehensive income and its components. Investors, creditors and others, who are concerned with assessing the prospects for enterprise cash ows are especially interested in this information. 5 6 For example, James Leisenring has held a number of positions at FASB since 1982, including being a member from 1987 to 2000 and serving as the Director of International Activities, when he left in 2001 to join as a member of the IASB. Anthony Cope was a member of FASB from 1993 until 2001, when he left to become a member of the IASB. Thomas Jones, the current vice chairman of the IASB, was previously a trustee for the Financial Accounting Foundation which oversees FASB, and was a member of FASB's Emerging Issues Task Force. Mary Barth, a professor from Stanford University, served on FASB Advisory Council until 2001 when she became a part-time member of the IASB. The current chairman of FASB, Robert Herz, was a part-time member of the IASB before joining FASB. In addition, the term matching is never used, even in passing. Accounting Horizons December 2008 American Accounting Association 458 Dichev Preliminary Views, however, is clear that: to designate one type of information as the primary focus of nancial reporting would be inappropriate. paragraph BC1.29. Displays of ... changes in economic resources and claims, and displays of the list of economic resources are equally important. paragraph BC1.30. Of course, the implication is that now standard-setters view earnings as much less important than before.7 Summarizing, Preliminary Views crowns and further develops the balance sheet model of nancial reporting. Some changes are critical, and they prompt the more structured critique that follows. A CRITIQUE OF THE BALANCE SHEET-BASED MODEL OF FINANCIAL REPORTING The critique is built around the following four main themes: The Balance Sheet Approach is Problematic Because it is at Odds with How Most Businesses Operate, Create Value, and are Managed Most rms are essentially sophisticated devices for continually advancing expenses, hoping to earn revenue and earnings. In relation to this fundamental purpose, most assets are just supplementary and temporary devices; one could say that they are props that serve the continual stream of company operations. Notice that once acquired, most assets have little independent existence and value. In other words, the balance sheet approach would make sense if rms were \"asset greenhouses,\" where the primary mission of the rm is to earn money by acquiring, storing, and growing assets, and earnings represents the realized or unrealized growth in these assets. But most rms are not asset greenhouses; they are more like \"asset furnaces,\" where acquired or internally created assets are continually sacriced or transformed for the larger goal of producing revenue and earnings. The balance sheet makes it look as if there is a permanent store of assets and asset values, but this impression is illusory because the stock of assets exists only because of the continuous process of assets renewal and sacrice. Relaxing the continuity of this process clearly reveals the temporary and subservient nature of most assets. Using a historical example, in the late Middle Ages the Italian trading guilds' business was much more fragmented and piecemeal. Partners would get together to nance the purchase or rental of a ship, appoint the crew, and advance other expenses. On completing the trading mission, the partners would split the prots, dissolve the partnership, and possibly start considering other ventures. In this example, it is clear that what really matters is the expenses advanced, and the revenue and prots ultimately earned. The assets ship and inventory are just temporary implements to carry out the trading business. This situation is far from an isolated historical curiosity, and is commonly found in business today. For example, many consulting rms operate on a project-by-project basis, and the key to creating value in this business is making sure that the revenues on the piecemeal projects exceed the requisite costs by a reasonable margin. In addition, even if practiced, assigning \"assets\" to each project has a temporary and provisional nature. Similar observations apply to, for example, the construction business, shipbuilding, and businesses related to military contracts. Note that in practice no real dichotomy exists between continual and piecemeal types of business, where most rms embody some varying mixture of both, and this mixture also varies 7 Over the years FASB members and afliated parties have repeatedly denied that the adoption of the balance sheet model leads to decreased emphasis on the primary importance of earnings e.g., Storey and Storey 1998, 82. Accounting Horizons American Accounting Association December 2008 On the Balance Sheet-Based Model of Financial Reporting 459 with the time horizon, where more and more of a rm's business represents one-shot projects over longer horizons. As the horizon lengthens, whole product lines and divisions represent one-shot deals, and the rm itself is a one-shot project over its complete life. Once the rm is nished, all assets are dissolved, and the only lasting impact is the cumulative amount of revenue earned, costs incurred, and resulting and distributed earnings. The point is that if rms operate as a process of advancing expenses to earn revenues, and assets have a secondary and supporting role in this process, then proper accounting must reect and follow this reality, and that implies a natural and logical supremacy for the income statement view of nancial reporting. Accounting can be dened as a system of tracking wealth and the creation of wealth in an economic unit. Intuitively, the essential features of the economic process of wealth creation must be reected in the essential features of the accounting system that tracks this process. It truly is like putting the cart before the horse if rm operations follow an income statement logic, while the nancial reporting process follows a balance sheet logic.8 Since most businesses follow an income statement mode of operations, it is not surprising that managers manage their businesses following an income statement approach. When managers prepare budgets, they produce a forecast of revenues rst and then predict the costs needed to support these revenues. Only after the budgeted income statement is produced will managers think about the asset base necessary to support these projections and the nancing needed to make it all happen.9 The same process applies to managerial decisions such as opening a new product line or starting a new division. Managers will think about the total operating and investment costs needed to get this project going, the revenue that is likely to be earned, and whether the resulting return on investment is acceptable. Managers will not consider the buildup of assets in a new division as the source of added value; if anything, excessive buildup of assets will be seen as a drag on rm performance. Investors, which FASB recognizes as the most important users of nancial reporting information, also base their decisions on income statement considerations. For example, the typical valuation of a business starts with a projection of revenues and costs rst, followed by balance sheet amounts and the resulting free cash ows. Consider also that when nancial analysts go beyond the prediction of earnings, they typically produce some sort of an income statement. Analysts' projections of balance sheet amounts are rare, and one never sees a projection of balance sheet amounts as a means to compute changes in net assets, which will comprise earnings. Similar observations and conclusions apply to the activities of most other parties interested in the success of rm activities, including participants in mergers and acquisitions and credit lending.10 To put the preceding ideas differently, the main problem with the balance sheet approach is that it is largely silent about the notions of business model and business performance that are central to a rm's success and value creation. The balance sheet approach takes asset values as given, as stores of value that are divorced from what the rm is doing, and diverts attention from 8 9 10 The balance sheet model, and especially its more extreme form of fair value accounting, is also at odds with the going concern assumption in accounting and auditing. The going concern assumption is that accounting views the rm as a continuous and ongoing stream of operations. In contrast, the fair value model views the rm as a collection of resources valued at exit market values. Of course, in practice there are many variations in the order presented above; for example, the managers of a rapidly growing retail chain may rst think about the maximum possible number of new stores they can open assets, which will drive the prediction of sales. Thus, the asset base can be a real constraint and inuence on operations and income. However, the point is that even in the retail chain situation managers care about assets rst because they hope that the asset expansion will drive sales, which is where value is created in this business. Managers are not opening stores hoping that stores will go up in value, creating earnings; even if this happens, store value appreciation is incidental to the business. Of course, asset values are also an important and often independent consideration in such decisions, for example, in terms of potential collateral or in the form of underperforming assets producing low earnings. Accounting Horizons December 2008 American Accounting Association 460 Dichev operations, which are the key to rm success and value. In contrast, the income statement model by its nature focuses the attention on rm operations, on the fact that rm value arises not from a static pile of resources but from continually using and putting these resources at risk in executing a business model. The income statement model clearly reveals that business success is determined in real operations, where the rm must go out and engage customers and markets for its goods or services, and where it is the customers who provide the ultimate business model verication by buying and paying for what the company offers. The fact that sales to a customer is the critical verication of a company's business model is recognized in accounting as the order in the process of recognition of income, where revenue recognition always comes rst, and then in turn triggering the recognition of expenses and income. In contrast, the balance sheet model makes it look as if rm value comes from the value of a store of resources, for example, as valued at exit prices in fair-value accounting. In short, for most rms the value of their resources comes from value-in-use and not from value-in-exchange. The rm is an ongoing process of business operations and not a collection of \"things,\" implying that the income statement model is the natural foundation for nancial reporting. To illustrate and bolster the arguments above, Table 1 provides empirical evidence about the TABLE 1 Stocks and Flows of Aggregate U.S. PPE during 1990 to 2005 (in $ million) Year n Capex Depreciation Sales of PPE Level of PPE 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 3,619 3,728 4,025 4,403 4,710 5,163 5,560 5,622 5,779 5,967 5,919 5,445 5,315 5,220 5,146 4,473 588,533 591,587 593,316 619,158 717,832 814,474 937,737 1,027,705 1,106,928 1,169,959 1,316,866 1,284,313 1,159,903 1,108,077 1,215,723 1,133,675 292,801 309,991 330,693 376,048 508,367 579,783 624,720 674,603 749,954 819,191 912,476 969,085 940,677 1,017,712 1,032,240 821,247 35,754 43,434 44,310 43,663 48,361 47,319 74,077 87,549 140,869 181,581 170,846 173,581 162,151 140,846 120,858 100,062 3,776,710 3,939,801 4,097,584 4,249,339 4,729,200 5,053,372 5,509,646 5,712,572 6,182,074 6,768,889 7,144,928 7,355,361 7,566,578 8,048,863 8,366,860 7,042,813 All data are from Compustat, for rms with assets exceeding $50 million. n is number of available rms for the respective year. Variable Denitions: Capex capital expenditures Compustat annual item #128; Depreciation depreciation and amortization item #14 amortization of intangibles item #65; Sales of PPE sale of PPE item #107; and Level of PPE PPE net item #8 as of year-end. Accounting Horizons American Accounting Association December 2008 On the Balance Sheet-Based Model of Financial Reporting 461 relative roles of internal use versus market-based considerations for the most typical and largest group of operating assets, PPE property, plant, and equipment. Table 1 lists the aggregate stocks and ows of U.S. rms' PPE over the last 15 years. All data are from Compustat, for rms with total assets exceeding $50 million. N is number of available rms for the respective year, Capex is Capital Expenditures Compustat annual item 128, Depreciation is dened as Depreciation and Amortization item 14 Amortization of Intangibles item 65, Sales of PPE is Sale of PPE item 107, Level of PPE is PPE Net item 8 as of year-end.11 The idea behind the table is that capital expenditures increase the Level of PPE, while depreciation and Sales of PPE reduce the Level of PPE, so by assessing the relative magnitudes of these amounts, one can judge the relative importance of PPE uses for internal and external purposes, and thus draw some conclusions about which model of nancial reporting is more appropriate for such assets. Figure 1 provides a graphical view of some of the relations in Table 1, plotting the ratio of Sales of PPE to Depreciation in Panel A, and the ratio of Sales of PPE to Level of PPE in Panel B. An examination of Table 1 and Figure 1 reveals two main results relevant for our considerations. First, the amount of Sales of PPE is small compared with the amount of Depreciation. Panel A of Figure 1 reveals that the ratio of these two amounts varies over the years but for all practical purposes it looks as if it is bounded between 10 and 20 percent. Thus, the use of PPE for internal purposes exceeds the use of PPE for external purposes on the magnitude of ve to 10 times, which suggests that the primary use of PPE is by far internal to the rm and its operations. Second, the amount of Sales of PPE is tiny compared with Level of PPE. Panel B of Figure 1 reveals that the ratio of these two aggregate amounts hovers between 1 and 2.5 percent, suggesting that only a small amount of PPE is relevant for current considerations of PPE's market value. Such magnitudes also question the wisdom of fair value-type accounting for PPE and PPE-type assets. If the motivation is that one needs to include the fair values of PPE on the balance sheet to reect what such assets would fetch on the market, this motivation is strained considering that the fair values even if they can be determined are irrelevant for 98 to 99 percent of the assets. This motivation is further strained when one considers that for the sake of properly reecting the value of a small percentage of assets, one revises the value of all assets, and this revaluation ows through the income statement, injecting great volatility in earnings from operations. The tenor of the empirical results above is consistent with business intuition and with arguments that rms primarily invest in PPE assets to use them in their production process, and that sales and market values of PPE are of second-order consideration and largely incidental to the business. The implication is that the proper accounting for such assets needs to reect this business reality, and thus PPE accounting has to be primarily concerned with the internal use of PPE, and much less with the uctuations of outside PPE values. Thus, the income statement model of nancial reporting is the natural basis for the accounting for such assets. Note that a large minority of business activities and whole businesses do follow a process of value creation, which has a balance sheet orientation and where balance sheet-based accounting is sensible. A stark example is a rm whose only assets are marketable securities. Since the assets are freely divisible and separable, and have value independent of the existence of the rm that holds them, keeping track of their changing values is all that one needs to know about this rm, and one would readily agree that \"earnings\" for this rm is just the changes in the value of the underlying assets during the current period. The logic of this example can be extended to other situations; for 11 Note that the amounts in Table 1 do not provide a full reconciliation of the PPE account; that is, Capex - Depreciation - Sales does not equal Level of PPE. The lack of reconciliation is because of the effects of M&As and write-downs of PPE, for which Compustat provides insufcient detail. Nevertheless, it is clear that Table 1 captures the rst-order effects in the PPE account. Accounting Horizons December 2008 American Accounting Association 462 Dichev FIGURE 1 The Relative Importance of Sales of PPE example, rms whose assets are undeveloped real estate holdings or rms that hold a collection of patents. To varying degrees, most nancial rms also justify some sort of balance sheet-based accounting because most nancial assets have values and lives that are largely independent of the existence and operations of the company that holds them. Accounting Horizons American Accounting Association December 2008 On the Balance Sheet-Based Model of Financial Reporting 463 Roughly speaking, the dividing line between income statement-oriented operations and balance sheet-oriented operations seems to be the distinction between operating and nancing activities. For nancing types of assets and liabilities, and rms that hold them, balance sheet-oriented accounting makes sense. However, the balance sheet orientation seems inappropriate for the operating activities of most companies that continually engage in the productive destruction of their assets and where assets have little independent existence and therefore little independent and separable value. As Paton and Littleton 1940 and many others have put it, most assets are \"unexpired costs\" rather than free-standing economic units with independent and separable value. Thus, for most types of rms, and for most types of activities, the income statement approach seems the natural t for their process of value creation, while the balance sheet approach is the preferred option in more limited and specic circumstances. The Alleged Conceptual Superiority of the Balance Sheet Approach is Unclear. If Anything, One Can Argue that the Concept of Income Provides a Clearer and Stronger Foundation for Financial Reporting FASB has consistently expressed a belief that the balance sheet approach allows establishing a solid conceptual foundation, which naturally and logically leads to building the theoretical structure of nancial accounting. Specically, FASB considers the concept of \"asset\" as the most important and fundamental in accounting, and other concepts as derivative and secondary. For example, \"liabilities\" are essentially the converse of \"assets;\" \"equity\" is the residual of assets and liabilities; \"revenues\" are increases in assets or decreases in liabilities, and \"expenses\" are decreases in assets or increases in liabilities see Storey and Storey 1998; Bullen and Crook 2005. The structure of the accounting that results from this ordering indeed seems appealing, but this article suggests that its foundations are far from solid, and from there the whole structure looks shaky. Recall that FASB considers the balance sheet approach superior because earnings is a change in value, and one cannot dene a change in value before establishing what value is, and that leads to asset and balance sheet-based accounting. In light of this argument, consider FASB's formal denition of assets: \"Assets are probable future economic benets obtained or controlled by a particular entity as a result of past transactions or events\" Statement of Financial Accounting Concepts No. 6. This denition states that an asset is something that brings future \"benets,\" and probably the implied meaning here is net benets; for example, a rental property is an asset because it produces rental income after rental expenses. However, for most assets, \"net benets\" sounds very much like \"earnings,\" which leads to a circularity in FASB's argument.12 FASB is arguing that asset-oriented accounting is superior to income-oriented accounting because one needs to dene assets before one can dene earnings, and then proceeds to dene assets in terms of expected earnings! The point is that economically the concept of asset and income are inextricably connected. If there is income, one can infer that some type of economic asset is producing it. If tangible assets are not visible, then one can infer that some intangible assets or what accountants call goodwill the synergy of other assets must be behind it. And the converse applies as well; if one has an \"asset,\" that means some kind of an income stream is attached to it; otherwise, it will not be an asset. FASB seems to suggest that the two concepts can be divorced, and one can be made primary and superior to the other. Perhaps what the Board had in mind was that sometimes it seems that no 12 Of course, assets such as cash and gold do not need a relation to future earnings to be recognized as assets. As discussed earlier, most such assets are separable, have value independent of the rm's operations, and are meant to be realized on some external markets. The argument here is mostly about operating assets. Accounting Horizons December 2008 American Accounting Association 464 Dichev clear stream of income is attached to an asset, but we can value it, perhaps based on an appraisal or some type of exchange or market value. However, implicit in the appraisal or the market value is again some stream of future benets or distribution of benets, be it rental income, higher margins because of a successful brand, or dividends on stock. Thus, the conceptual superiority of the balance sheet approach is unclear. If anything, one can argue that the concept of income is more fundamental and clear, especially in light of the increasing prominence of intangible assets. As discussed above, the concept of income is crystal-clear in one-shot deals; for example, selling for $10 something that cost $7 produces income of $3. For more realistic situations and rms, the concept of income is also reasonably clear over long horizons because by design income and net cash ows converge over such horizons; for example, we can get a reasonably clear idea about \"how much money\" Microsoft made over a 10-year period. The challenges of applying the income concept stem from the fact that most rms have continuous, overlapping, and interwoven activities, and that the horizon of interest is short; for example, it is a lot harder to say what \"income\" is for Microsoft over any given quarter. However, these are not conceptual but operational and technical difculties, and after all, it is precisely in solving such difculties where the value-added of accounting is. Otherwise, note that the clarity of the income concept is universal; it applies in a comparable manner from the most mundane to the most sophisticated business. No matter whether the business is a coffee cart, a manufacturing company, a lm studio, or a biotech company, it is easy to see whether the company is \"making money,\" especially in the long run. In contrast, it is far from clear what the \"assets\" are of a company such as Microsoft, and how to account for them. The balance sheet for Microsoft lists some assets, but these are not much help in accounting for Microsoft's extraordinary protability. Based on the abnormal prots, we can infer that there must be some missing assets, and we could call them intangibles, goodwill, human capital, monopoly position, or captive customer base, but these will be just names, and the whole exercise is not very illuminating or helpful in practice. The point is that today many, if not most, assets are elusive conceptually and difcult to operationalize in any helpful way. The steadily growing market-to-book ratio of most rms is a succinct illustration and testament to this problem. Ironically, those who try to derive estimates of intangible assets typically use projections and discounting of some sort of income to accomplish the task; for example, one can derive the value of the Coca-Cola brand by estimating the future price and volume premium because of the brand name, and discounting it to derive a present value. Such derivations clearly reveal that the existence and valuation of intangible assets is derivative and conditional rather than fundamental. This presents a tremendous practical problem considering that today intangible assets apparently account for a great part, and perhaps even the majority, of rm assets; for example, the average market-to-book ratio is much higher than one see also Lev 2003. Balance Sheet Accounting is Likely a Major Contributor to the Substantial Temporal Decline in the Forward-Looking Usefulness of Earnings Investors use earnings as the primary metric to evaluate prospective and existing investments. The classics of investment theory advise us to \"buy earnings,\" and surveys reveal that investors and analysts consider earnings to be by far the single most important number about rms e.g., Graham et al. 2005. Note that the notion of earnings that investors nd useful is not \"changes in assets\" but \"recurring earnings,\" essentially the current earnings that are the best predictor of the future stream of earnings that the rm will produce. Thus, to investors, \"good earnings\" means a metric that is highly persistent and predictive of future earnings. In contrast, the balance sheet approach views assets as the store of value and earnings as \"changes in net assets,\" which implies low persistence and predictability of earnings. In the Accounting Horizons American Accounting Association December 2008 On the Balance Sheet-Based Model of Financial Reporting 465 extreme, balance sheet accounting is pure mark-to-market accounting, where every asset and liability is updated to market or fair value each period. Since it is well known that market values changes are unpredictable Samuelson 1965, pure market value accounting implies earnings that are pure noise, with high volatility and zero persistence and predictability. Thus, the balance sheet approach creates earnings that are at odds with what investors consider \"good earnings.\" Unfortunately, the empirical behavior of earnings suggests that this is not just a theoretical concern, and indeed a lot of damage has already been done. Dichev and Tang 2008 examine the empirical properties of earnings of the 1,000 largest U.S. rms during the last 40 years, and nd that comparable earnings volatility has more than doubled during this period, while earnings persistence has fallen from 0.91 to 0.65, a substantial deterioration in the basic properties of accounting earnings. This evidence is especially troubling because the study nds little change in the properties of the underlying revenues, expenses, and cash ows over the same period, and, more generally, the evidence suggests that the bulk of the changes in the properties of earnings are because of changes in the accounting rather than changes in the real economy. On the practical level, these changes in the properties of earnings likely happen because the balance sheet approach mandates various asset revaluations that result in an increasing number and magnitude of writeoffs, \"one-time\" charges, and other nonrecurring items. Givoly and Hayn 2000 also nd that the volatility of earnings has greatly increased over time, while the volatility of cash ows has largely remained the same. Given these temporal changes in the properties of earnings, it is not surprising that Collins et al. 1997 nd that the relation between stock prices and earnings has steadily weakened over time. Thus, existing research suggests that the balance sheet-based model of nancial reporting has already produced a marked deterioration in the forward-looking informativeness of earnings. Earnings today are much more volatile and less persistent, which implies that current earnings tells less and less about future earnings. Of course, these changes can be due not only to deciencies in the model itself but to faulty implementation of the model. In any case, the critical consideration here is that if these trends continue unabated for another 30 to 50 years, the danger is very real that earnings will become a meaningless number for forward-looking applications, a situation that will have far-reaching repercussions in many directions. One possible consequence is that useless earnings threaten the very utility of the accounting system, and that may lead to an erosion of the accounting function and profession, at least as far as nancial accounting is practiced today. A related implication is that investors and their various proxies will increasingly turn to non-GAAP metrics of value. The experience with pro forma earnings during the last 10 to 15 years offers a foretaste of what is to come, and the fact that it was a chaotic and confusing experience suggests that very real costs exist in producing inferior earnings. The deterioration in the informativeness of earnings also suggests the possibility of a further stratication between sophisticated and unsophisticated investors in the security markets. One could argue that the major beneciaries from good earnings are small and unsophisticated investors, who tend to value investments by using heuristics such as price-to-earnings ratios. As long as current earnings is a good guide to future earnings, and thus provides a solid link to rm value, such heuristics work fairly well. However, unsophisticated investors who continue to uncritically rely on earnings will increasingly feel as if they are standing on shifting sand as the predictive power of earnings continues to erode. In contrast, sophisticated investors will have an increasingly keener edge because of their differential abilities to understand and work through the mounting problems of earnings as a guide to investment value. Given that a mandate of security regulators is to level the playing eld in nancial markets, it is ironic that FASB, being the proxy for the SEC in accounting standard-setting, has implemented a course of action that threatens to produce the opposite result. Accounting Horizons December 2008 American Accounting Association 466 Dichev Substantial Problems Exist with Applying the Balance Sheet-Based Model of Accounting in Practice Some of these problems have already been discussed in other sources; for example, \"markto-market\"-type accounting rules are difcult to apply when there are no reliable estimates of market values, and thus in practice rms have to resort to \"mark-to-model\" accounting instead. Of course, the weakness of the \"mark-to-model\" approach is that it involves considerable managerial discretion with respect to inputs, and from there the potential for large estimation errors and outright manipulation. For example, after Enron fell, \"mark-to-model\" accounting became notorious because it was extensively used to manage earnings see Palepu and Healy 2003. Although Enron's fall is an extreme case, it illustrates the more general dangers of introducing much and avoidable subjectivity in estimating nancial results. Because the subjectivity theme is already prominent in other sources, here the attention is on another point that is less well recognized. Balance sheet-based accounting and especially its more extreme forms of mark-to-market and fair-value accounting create a feedback loop between nancial markets and the real economy, and can possibly lead to or exacerbate market bubbles. Generally speaking, fair-value accounting puts a lot of faith in market prices and elevates them to an unfailing standard of correctness. This premise is dangerous when market prices can deviate from fundamental values, which recent research increasingly recognizes and documents e.g., see reviews in Hirshleifer 2001; Shleifer 2000.13 Accounting must recognize the difference between the real economy where real economic value is created and the nancial markets world that makes educated guesses about the values of claims to real economy wealth and trades them. For most rms, accounting needs to reect their real economic activities and provide inputs and independent checks on the valuation and trading process in nancial markets. Failure to do so confuses what is being measured and collapses accounting's independent check function. In the extreme of pure mark-to-market accounting, accounting and nancial markets functions are fused, where rm performance and rm valuation become an empty and self-fullling tautology. For example, markets go up because rms have earnings, which they have because markets are going up and their assets are revalued up see also Plantin et al. 2005 for an elaboration on this point. And the converse is happening when prices are heading down, in a self-propagating spiral. Of course, to some extent this description is oversimplifying the matters to make the pointbut we are moving in this direction. SUGGESTIONS ABOUT WHAT A \"BETTER\" CONCEPTUAL FRAMEWORK MIGHT LOOK LIKE As an extension and conclusion, this article offers some observations on what can potentially serve as a \"better\" model of nancial reporting. One caveat here is that, as claried in Gonedes and Dopuch 1974, any discussion of \"better\" needs to be understood as taken from the perspective of a given user or constituency of nancial reporting, especially if different constituencies have conicting preferences. Consistent with prevalent beliefs and FASB's own statements including the proposed new conceptual framework, this article takes the outside investor perspective. 13 Note that here market prices are viewed as potentially faulty measures of value, while the earlier discussion of marketto-book ratios assumes that markets produce proper measures of value. Both assumptions are true depending on the context; because of market frictions and behavioral biases market values can become distorted measures of value sometimes severelybut eventually the market mechanism ensures that there is a long-term convergence with fundamentals. Accounting Horizons American Accounting Association December 2008 On the Balance Sheet-Based Model of Financial Reporting 467 The rst major feature of an alternative model of nancial reporting is a clear theoretical and practical distinction between operating and nancing activities. Operating activities include all activities that are related to the regular business of the company and that are related to the transformation of purchased and internally produced inputs into goods and services to be sold on the open market. Thus, operating activities here encompass most of what is labeled today as \"operating\" and \"investing\" activities. A dening feature of operating activities is that resources used in them operating assets have a primary purpose of supporting and enhancing these within-the-rm activities, and have only limited and peripheral value as independent, freestanding, and marketable stores of value; that is, even if available, fair values for these assets do not reect their primary value to the company. Thus, operating assets are primarily noncash assets such as PPE. In contrast, nancing activities revolve around cash and cash-equivalent assets and liabilities such as marketable securities and perhaps accounts receivable, which are separable from the rm and have value that is largely independent from the rm's fortunes. In practice, some controversy and debate will arise about the precise operational divide between operating and nancing assets and liabilities, and the demarcation between these two categories may shift depending on the nature of the business. However, such controversies and judgments are nothing new for accounting, and in fact they are an integral part of its essence and utility to users of nancial information. The important distinction between operating and nancing assets and activities should be reected in all nancial statements. The income statement must clearly identify the difference between earnings from regular operating activities, which have much persistence and forwardlooking informativeness, and earnings from value uctuations in nancial assets, which have little persistence and predictive power. Thus, accounting will have to move away from the hallowed notion of a single \"bottom-line,\" a number that neatly summarizes the entire performance of the rm. Of course, the abandonment of a bottom-line number leads to a regrettable loss of parsimony and could place at risk the preeminence of earnings in investor decisions. However, the accumulation of changes in the business world and accounting has led to a situation where continuing to mix two very different sources of income also has substantial and probably higher costs, and thus the evolution toward two and maybe more groupings of income seems necessary. This projected evolution is also in line with what is already happening. In March 2007 FASB revealed that it is considering sweeping changes in the way income is reported, including reporting separate income subtotals for a company's operating, investing, nancing, and tax activities see Reilly 2007 and Cain 2008 for more complete accounts. The balance sheet also needs to separate assets from operating and nancing activities because of their different nature and implications for valuation. Operating assets are not really independent assets but essentially just a listing of unexpired costs. One can rightly view them as a listing of shorter and longer-term operating bets or as commitments on streams of future operating costs. Most of these costs will be realized internally, and there is little reason to pursue fair-value accounting for them. In other words, the value of these assets is value-in-use, and there is no sense in using fair value or some other benchmark of outside utility to value them. In contrast, nancial assets are freely divisible and separable from the rm, and are almost by denition going to be realized on some market, and therefore some type of mark-to-market or fair-value accounting makes sense for them. The value of such assets is value-in-exchange, so exchange-based benchmarks of value are entirely appropriate. Finally, the cash ow statement also should distinguish between cash generated from the company's operations and from nancing activities. The second major feature of an alternative model of nancial reporting is renewed emphasis on the matching principle and, to a lesser extent, the revenue recognition principle as cornerstones of the accounting for operating activities. Note that, as argued above, the accounting for cash-like Accounting Horizons December 2008 American Accounting Association 468 Dichev value-in-exchange assets is already well provided for under the implemented and proposed rules of the fair-value initiative. However, there should be a realignment of the accounting for operating, value-in-use assets along the lines of the income statement model of nancial reporting. Most rms create value and are managed in an income statement mode, and thus accounting should follow and reect this economic reality. The two overriding principles in income statement accounting are revenue recognition and matching. Thus, accounting must be clear about the denition and content of these two principles, and then the more specic provisions must align with and follow them. Of these two principles, revenue recognition seems comparatively more straightforward, at least in theory; basically, revenue is earned when goods or services have been exchanged for cash or cash equivalents.14 The matching principle, which relates relevant expenses to associated revenues, is thornier and produces considerable interpretation and implementation problems because many expenses are difcult to trace to specic revenues or periods. However, repairing and reafrming the primacy of these two principles shows a clear general direction for making progress. If one is thinking about measuring performance for a given period, one must derive a measure of the economic achievements revenue and the sacrices that were made to accomplish these achievements expenses, and performance is the difference between these two earnings. The practical upshot from these considerations is that income statement-oriented accounting must concentrate its efforts on mapping the economic link between rm expenditures and receipts and translating this link into revenues and expenses. For example, if a reliable link exists between research and development R&D expenditures today and revenues for three years ahead, R&D expenditures should be capitalized and expensed over the next three years. The goal is to make the accounting reect and follow the economic logic of the business as much as possible. If following the economic logic of the business is embraced as a guidepost for accounting, the argument for renewed emphasis on the matching principle is straightforward. Most businesses are run on the explicit or implicit logic of matching costs and benets. Almost all managerial decisions contain an element of weighing the benets of some action against the costs. For example, in considering whether to open a new product line or a new division, managers project the investment and operating expenditures, and will only green-light the project if the corresponding cash inows or other benets exceed the costs by a requisite margin. Thus, using matching in accounting is logical and necessary because it reects the inescapable reality of cost-benet considerations and results that pervade every business. In other words, matching considerations are at the very core of how businesses are run and create value, and may be even considered to dene what a \"business\" is i.e., an economic unit that produces a surplus of cash receipts over associated expenditures. Thus, the real question is whether accounting will choose to incorporate this core feature of business and to what extent. If accounting aims to faithfully reect the business, its core principles should reect the core drivers of the business, and this intuition suggests that matching considerations should be a pivot for the nancial reporting system. Of course, accounting can choose to ignore this intuition but that may prove to be perilous considering that matching is hard-wired in the essence of what a business is and how it runs. 14 There are signicant practical difculties in revenue recognition, however, when the contractual arrangements are complex e.g., multiple deliverables, signicant rights-of-return, and contingencies. In addition, the fact that about half of the SEC's enforcement actions are related to revenue recognition illustrates the commonality of practical problems with revenue recognition. In attempting to address these difculties, FASB and the IASB have embarked on a multiyear project aimed to align revenue recognition with the balance sheet-based model of nancial reporting. At this point, it is unclear what the specics of the nal product will be, but the chance is high that these new rules will be another substantial deviation from the income statement model. Accounting Horizons American Accounting Association December 2008 On the Balance Sheet-Based Model of Financial Reporting 469 Unfortunately, the current thinking at FASB and the IASB completely ignores the concept of matching. An inspection of the Preliminary Views for a new conceptual framework reveals the consideration of a number of useful concepts such as relevance, faithful representation, and comparability, but there is not a single reference to matching. In contrast, this article argues that, together with revenue recognition, matching should be the cornerstone of nancial reporting, and failing that, all these other concepts will be decient in content and utility. In conclusion, the development of a full alternative model of nancial reporting is a formidable task; thus these observations are offered more with the intent to spark a debate rather than as a solution. The accounting standard-setters are currently reconsidering the foundations of nancial reporting, and the repercussions of this process will be felt for decades. The time is right to join these deliberations because the costs of having a decient reporting model are substantial and far reaching. The converse is also true; accounting touches many levels and functions of society, and having the right model of reporting contributes to the welfare of us all. REFERENCES Barth, M. E., W. H. Beaver, and W. R. Landsman. 2001. The relevance of the value relevance literature for nancial accounting standard setting: Another view. Journal of Accounting and Economics 31 1-3: 77-104. -. 2006. Including estimates of the future in today's nancial statements. Accounting Horizons 20 3: 271-285. Brief, R. P. 1982. Hicks on accounting. The Accounting Historians Journal 9 1: 101-111. Bullen, H. G., and K. Crook. 2005. Revisiting the Concepts: A New Conceptual Framework Project. Norwalk, CT: Financial Acc