The impact and valuation of off-balance-sheet activities concealed by equity method accounting.

SYNOPSIS: This paper reports the results of a study of the financial reporting effects of off-balance-sheet activities concealed by the equity method of accounting. The study examines footnote disclosures relating to equity method investees, offers suggestions for improving the usefulness of

INTRODUCTION

Financial analysts have long been concerned with the effect of off-balance-sheet activities on a firm's financial condition. The financial statement analysis literature emphasizes the importance of examining footnote disclosures for the existence of economic assets and liabilities that are not recognized under generally accepted accounting principles (GAAP); Lasman and Weil (1978) provide an early example. Further, off-balance-sheet activities receive increased attention in the post-Enron world (Chang 2002).

This paper examines financial analysis and valuation issues caused by off-balance-sheet activities that are not fully reported under equity method accounting. Described in Accounting Principles Board (APB) Opinion No. 18 (APB No. 18, APB 1971), investors must use the equity method for investments in common stock that provide "significant influence over operating and financial policies of an investee even though the investor holds 50% or less of the voting stock" (APB No. 18, [paragraph] 17). There is a presumption that significant influence exists with ownership interests of 20 percent.

Given that GAAP requires the equity method, in analysis a key issue is how to best reflect the investment in the investor's financial statements. In most cases, the equity method summarizes the investor's share of the investee's net assets and net income in single lines in the balance sheet and income statement, respectively. An alternative approach consolidates the components of the investee's assets, liabilities, and income with those of the investor. Another alternative approach consolidates only the investor's proportionate share of the components of the investee's assets, liabilities, and income. These alternative approaches can produce large differences in the investor's reported performance and leverage. (1)

The equity method of accounting is controversial. Davis and Largay (1999, 281) perform a critical analysis of the accounting for significant influence equity investments, defined as "situations where one entity possesses more than a passive investment in another entity but does not control that entity." They develop two conclusions based on financial analysis considerations, consolidation theory, and aggregation issues. First, they find "no substantive justification for continued use of the equity method ... due to the method's intrinsically limited informational characteristics" (Davis and Largay 1999, 281). Second, they recommend alternative accounting methods--proportionate consolidation and the expanded equity method. Proportionate consolidation and the expanded equity method are substantially similar, differing only in presentation format. For purposes of this paper, I consider these methods equivalent.

Proportionate consolidation is common outside the United States. International Accounting Standard No. 31 (International Accounting Standards Board [IASB] 1990) recommends proportionate consolidation for jointly controlled entities and Canadian accounting principles require proportionate consolidation of joint ventures (CICA Handbook, Section 3055, Canadian Institute of Chartered Accountants (2000)). Bierman (1992) advocates the use of proportionate consolidation even for majority-owned subsidiaries.

Some analysts regard the equity method as enabling firms to avoid balance sheet recognition of the assets and liabilities of investees (e.g., Ciesielski 2002). From a financial analysis perspective, the lack of detail about investee activities is a weakness of the equity method. The balance sheet does not report the investor's interest in the individual assets and liabilities of an investee, and the income statement lacks detail on investee revenues and expenses. Under GAAP, investing firms must present summarized information about the assets, liabilities, and operating results of material unconsolidated investees. Thus, unless the separate financial statements of investees can be obtained, these footnote disclosures are the primary source of financial data for these entities.

This paper reports the results of two analyses utilizing recent financial statement footnote disclosures of equity method investments for a sample of U.S. manufacturing firms. First, I review these disclosures and find that an aggregated presentation of investee information frequently hinders accurate estimation of off-balance-sheet assets and liabilities. This review enables me to offer suggestions for improving the usefulness of equity investment footnote disclosures. The review also includes examining the financial statement impact of equity method investees on companies that provide investee financial data. This analysis indicates that under pro forma proportionate consolidation, 55, 73, and 73 percent of investors report 5 percent or larger increases in assets, liabilities, and sales, respectively.

The second analysis examines the stock market's valuation of equity method investees by comparing the association between firm equity value, reported financial statement data, and pro forma financial data reflecting the off-balance-sheet activities concealed by the equity method. The results indicate that market participants place greater weight on off-balance-sheet liabilities than assets for firms that explicitly guarantee investee obligations.

Overall, the study has two main implications. First, although financial statement users can use footnote disclosures to consolidate investees on a pro forma basis, several limitations impair the accuracy and usefulness of such analyses. Second, based on the association between firm value and off-balance-sheet activities concealed by the equity method, market participants find the disclosure of these off-balance-sheet activities useful in some cases. The finding pertaining to guaranteed off-balance-sheet liabilities provides empirical support for recent rulemaking. In particular, FASB Interpretation No. 45 on guarantees (FIN No. 45, FASB 2002) calls for recognizing a liability based on the estimated fair value of guarantee obligations issued after December 31, 2002.

BACKGROUND

Equity Method and Proportionate Consolidation

Under the equity method, a firm reports its investment in a nonconsolidated entity as a single line item in the balance sheet. The amount initially reported in the investment account is either (1) the investor's capital contribution to form the entity, or (2) the price paid to acquire an interest in an ongoing enterprise. In each accounting period the investor's income statement contains its proportionate share of the investee's net income (loss) as a single line item. Any difference between the acquisition cost and the investor's proportionate share of the net assets of an acquired investee is allocated to identifiable assets of the investee and/or to goodwill. When appropriate, amortization of these fair value adjustments reduces both the investment account and the investor's income from the investee. (2) The investment account is reduced by any dividends received from the investee and increased if the investor makes additional capital contributions. For these reasons, an investor's periodic net income and stockholders' equity does not differ between the equity method and consolidation. (3)

Proponents of proportionate consolidation believe the equity method provides a distorted picture of firm profitability and risk by relegating details about investees' balance sheet and income statement composition to the financial statement footnotes. For example, including its share of investee earnings in net income and understating the asset base used to generate these earnings overstates the investor's return on assets. Similarly, netting investee assets and liabilities in the investment account impacts leverage ratios, and shielding investee interest expense in the one-line presentation of investee income overstates interest coverage.

A key issue in the debate involves the relation between the investee and the investor. If the products and customers of the investee are unrelated to those of the investor, then including the results of the investment in single lines on the income statement and balance sheet may be reasonable. However, when the respective operations are closely related and the investor has significant influence over the operating and financial policies of the investee, analysts consider the investee as an integral part of the investor. In these cases, texts such as White et al. (2003) recommend consolidating the investor's pro rata share of investees' assets, liabilities, revenues, expenses, and cash flows into its accounts. Graham et al. (2003, Figure 1) provide a detailed example of the differences in a hypothetical investor's financial statements using proportionate consolidation versus the equity method. (4)

Proportionate consolidation is not without controversy. Some object to the investor recognizing its share of the investee's individual assets because the investor does not control the benefits from using those assets, and the asset criteria in FASB Concepts Statement No. 6 (FASB 1985) are not satisfied. Others object to recording investee liabilities on the investor's books because, in the absence of explicit guarantees, the investor has no legal obligation to pay the debts of an investee. Because the same claim can be made when majority- or wholly owned subsidiaries are fully consolidated, though, this represents an argument against consolidation in general. Finally, some financial analysts dislike proportionate consolidation as the integration of investor and investee accounts hinders separate evaluation of significant investees (AICPA 1994). While some parties do not favor proportionate consolidation in the primary financial statements, most users want detailed supplemental disclosures. But as shown below, the information required under GAAP is often not sufficient for analysis purposes.

Research Related to Off-Balance-Sheet Activities

The potential financial statement effects of off-balance-sheet activities examined in the accounting literature include those related to research and development limited partnerships (Shevlin 1987), pensions (Dhaliwal 1986; Landsman 1986), and operating leases (Imhoff et al. 1993; Ely 1995). Existing research also addresses certain financial analysis considerations peculiar to the equity method. Graham et al. (2003) examine whether financial statements of Canadian firms prepared under proportionate consolidation provide better predictions of future profitability than pro forma statements prepared under the equity method. Using a sample of 78 firms, they find that financial statements under proportionate consolidation better predict one-year-ahead return on common equity. The present study extends this line of research by focusing on U.S. firms and current U.S. footnote disclosures.

Existing research also addresses valuation-related issues bearing on equity method investments made by U.S. firms in the shares of publicly traded investees. Graham and Lefanowicz (1996) find a positive association between investor and investee stock returns around investee earnings and dividend announcements, and conclude that the market value method may be appropriate for equity investments. (5) Graham et al. (1998) examine the value-relevance of fair value disclosures of equity method investees with quoted market prices. They find that investor stock prices are positively associated with the difference between book values recognized on the balance sheet and disclosed fair values. Because the current sample includes only five firms with publicly traded investees, this paper extends valuation research to firms with equity method investees that are not publicly traded.

SAMPLE AND DATA

The sample is selected from the 3,391 manufacturing firms (SIC code between 2000 and 3999) appearing in the 2000 Compustat active files. Of this total, 431 publicly traded firms have available stock prices and non-missing, non-zero amounts for "Equity in Earnings" (Compustat item #55). Although 359 of these firms are incorporated in the United States, only 228 meet the calendar yearend requirement of this study. I examined these firms' footnote disclosures to identify those that report enough data to permit proportionate consolidation of equity method investees. There are 90 such firms in the year 2000. Of these firms, 76 also filed 2001 financial statements on EDGAR prior to May 1, 2002. Using Belsley et al.'s (1980) regression diagnostics, I identified one firm as an outlier in the valuation analysis and excluded it from all of the analyses below. The final sample consists of 150 firm-year observations--75 firms for the years 2000-2001. I obtained financial statement data from Compustat and Form 10-Ks, and stock price data for April 30, 2001 (2002) from The Center for Research in Security Prices files (Yahoo! Finance).

FINDINGS

This section reports the results of a review of sample firms' footnote disclosures and an empirical examination of the valuation of off-balance-sheet activities arising under the equity method of accounting.

Disclosure Review

From paragraph 20(d) of APB No. 18:

   When investments in common stock of corporate joint ventures or
   other investments of 50% or less accounted for under the equity
   method are, in the aggregate, material in relation to the financial
   position or results of operations of an investor, it may be
   necessary for summarized information as to assets, liabilities, and
   results of operations of the investees to be presented in the notes
   or in separate statements, either individually or in groups, as
   appropriate. (emphasis added)

Although the wording of this disclosure requirement leaves the door open for nondisclosure, SEC registrants are subject to formula-based requirements. Accounting Series Release No. 302 (SEC 1981) requires detailed footnote disclosure when one of three specific materiality thresholds is met. In general, satisfying these materiality thresholds occurs when, individually by any investee or on an aggregated basis, (1) the investment account exceeds 10 percent of consolidated assets, (2) the proportionate share of total assets exceeds I 0 percent of consolidated assets, or (3) equity in income from continuing operations before income taxes, extraordinary items, and cumulative effect of accounting changes exceeds 10 percent of such consolidated income.6 SEC rules require investors to disclose data such as current assets, noncurrent assets, current liabilities, noncurrent liabilities, redeemable preferred stock, minority interests, net sales or gross revenues, gross profit, income from continuing operations, and net income.

Surprisingly, one-fifth of the sample firms reported a majority-owned investee via the equity method. SFAS No. 94 (FASB 1987) consolidation requirements exclude majority-owned subsidiaries when control does not rest with the majority owner. For example, an investor may own a majority of equity shares, but hold less than a majority of the positions on the board of directors.

Using Footnote Disclosures in Financial Analysis

The above discussion suggests that proportionate consolidation is most appropriate when investees' operations are integrated with those of the investor. My review of footnote disclosures indicates that the sample consists of operationally integrated investors and investees, as virtually all investees are directly involved in related production and/or distribution activities. Sample firms AGCO Corp. and Polaris Industries account for their retail finance subsidiaries under the equity method. Moreover, the disclosures indicate that most of the investors in the sample partnered with third parties to create a new entity, rather than acquiring a stake in an existing operating business. Thus, the sample consists of firms for which proportionate consolidation is a meaningful alternative.

To prepare pro forma financial statements based on proportionate consolidation, analysts multiply investees' balance sheet and income statement items by the relevant ownership percentage. However, data are often aggregated across investees, making accurate proportionate consolidation adjustments difficult to compute. Inaccuracies occur unless (1) the investor's ownership percentage is the same for all of its investees, or (2) profitability and capital structure are identical across affiliates.

For example, Chevron discloses a 50 percent interest in Caltex Group and a 26.5 percent interest in Dynegy Inc. Chevron could have aggregated these investees' 2000 financial data as follows:

($ million)                         Caltex     Dynegy     Aggregated

Investee liabilities $5,928         $16,444    $22,372     $22,372
Reported equity in income of
  investees                             259        127         386
Reported net income of investees        519        501       1,020

To determine the investor's share of investee liabilities, Stickney and Brown (1999) recommend dividing the equity in the income of investees ($386) by the investees' total reported income ($1,020), which yields a composite ownership percentage of 37.8 percent. (7) Applying this percentage to the aggregated liabilities of $22,372 results in $8,457 of estimated additional liabilities. In contrast, multiplying the reported ownership interests by the separate liability amounts listed above results in additional liabilities of $7,322. Thus, the estimate overstates the actual by 15.5 percent [($8,457 - $7,322) / $7,322] in this case.

This error results from two offsetting effects. First, using the composite ownership percentage of 37.8 percent underweights Caltex and overweights Dynegy. The fact that Dynegy's liabilities are greater overstates Chevron's estimated off-balance-sheet liability. Second, the estimated 25.3 percent (= $127 / $501) ownership percentage for Dynegy understates the reported 26.5 percent. Such a discrepancy may be due to amortization of fair value adjustments from Chevron's acquisition of its stake in Dynegy and/or elimination of Chevron's share of unrealized intercompany profits. To the extent these amounts are not disclosed, estimated ownership percentages may contain errors. (8) Without these items, Chevron would report equity in the incomes of Caltex and Dynegy of $392.3 [(.5)($519) + (.265)($501)], yielding a composite ownership of 38.5 percent. Thus, the second error causes Chevron's share of its investees' liabilities to be understated. (9) In summary, analysts' estimates of proportionate consolidation amounts can be misstated due to errors from using footnote data aggregated across multiple investees.

Another significant impediment to pro forma proportionate consolidation is the lack of required disclosure of income statement line items. An important element of solvency analysis is the measurement of interest coverage, generally computed as earnings before interest and taxes divided by interest expense. However, only 13 percent of sample firms disclose interest expense of their investees. Although interest expense can be estimated by applying an interest rate to off-balance-sheet debt, less than one-third of sample firms disclose the amount of investees' interest-bearing debt. Further, while the current market rate of interest can be used, the weighted-average interest rate on outstanding debt is the appropriate measure.

Overall, the above-cited deficiencies in footnote disclosure requirements potentially hinder effective financial statement analysis. Accordingly, I suggest that existing requirements be modified to include disclosure of (1) the investor's share of investee amounts, including interest-bearing debt, and (2) investee income statement details sufficient to compute pro forma interest coverage and margins.

Magnitude of Investee Activities

Table 1 presents descriptive statistics for several measures of the magnitude of equity method investments. The first two measures are based on amounts reported by investors. The investment account relative to total assets has a mean of 9.4 percent and a median of 4.9 percent. The larger mean is due to several large values, including two greater than 85 percent for a holding company (Tremont Corporation). Two sample companies report credit balances in their investment accounts due to distributions in excess of earnings (Pennzoil-Quaker State Co.) or allocated losses in excess of investment (Orbital Sciences Corp.). (10) Table 1 also compares the equity in income of investees relative to the investor's reported net income. Given the existence of losses, I use absolute values of these numbers. The large difference between the mean (65.2 percent) and median (15.6 percent) values is due to several observations having small denominators--reported net income near zero.

The next two measures in Table 1 reflect the incremental amount of total assets/liabilities to be reported under proportionate consolidation. Footnote seven indicates that composite ownership estimates are required for 37 percent of the sample firms; accordingly, the empirical analyses in this paper assume that market participants use an estimation method similar to that described above. As total stockholders' equity is the same under the equity method and proportionate consolidation, the increments to assets and liabilities are identical by definition (Graham et al. 2003, Figure 3). On average, incremental assets comprise 11.2 percent of reported assets, with a median value of 6.1 percent. With respect to the materiality of these amounts, off-balance-sheet assets exceed 5 percent of reported assets for 54.7 percent of the sample observations. The next row reports incremental liabilities relative to reported total liabilities. As most sample firms report positive shareholders' equity, these measures exceed those in the previous row. These off-balance-sheet liabilities comprise 5 percent or more of reported liabilities for nearly 73 percent of the sample observations. Moreover, the median (10.0 percent) and third quartile (23.7 percent) values indicate that these liabilities can be substantial. Finally, the last row of Table 1 measures the pro rata share of investee sales relative to reported sales; on average these sales comprise 17.5 percent of reported sales, with a median value of 10.0 percent. Thus, the off-balance-sheet activities concealed by the equity method are often significant relative to reported amounts.

Valuation of Off-Balance-Sheet Activities

The empirical model employed in this study uses a general equity valuation model, which expresses the market value of common equity for firm i, MV[E.sub.i], as a function of the firm' s net income from continuing operations, N[I.sub.i], and book value of common equity, B[V.sub.i] (Collins et al. 1999; Graham et al. 1998). To allow for public release of the annual financial statements, I measure MVE based on the stock price four months after fiscal year end.

(1) MV[E.sub.i] = [alpha]o + [[alpha].sub.1]N[I.sub.i] + [[alpha].sub.2]B[V.sub.i] + [[epsilon].sub.i].

I make several modifications to Equation (1). First, to allow the relation between MVE and profits to differ from the relation with losses, I interact NI with a dummy variable (LOSS) that is (1) equal to 1 if NI is less than zero and (2) equal to 0 otherwise. Second, I decompose book value into reported total assets (ASSETS) and reported total liabilities (LIABS). This allows the coefficients on these variables to differ. Third, I add a variable representing the magnitude of off-balance-sheet activities arising under the equity method (OBS), equal to the incremental amount of assets/liabilities reported under pro forma proportionate consolidation. I anticipate that market participants could place a different value on OBS, depending on whether the investor guarantees the obligations of the investee, and allow for this possibility by interacting OBS with two dummy variables, GUAR and NOGUAR. The former (latter) takes on a value of 1 if the investor provides (does not provide) an explicit guarantee. These modifications result in the following empirical model: (11)

(2) MV[E.sub.i] = [[beta].sub.0] + [[beta].sub.1] N[I.sub.i] + [[beta].sub.2] N[I.sub.i] * [LOSS.sub.i] + [[beta].sub.3][ASSETS.sub.i] + [[beta].sub.4][LIABS.sub.i] + [[beta].sub.5][OBS.sub.i] * [NOGUAR.sub.i] + [[beta].sub.6][OBs.sub.i] * [GUAR.sub.i] + [[gamma].sub.i].

With regard to the coefficients for Equation (2), I expect NI to be positively related to MVE ([[beta].sub.1] > 0). Since shareholders can liquidate a firm rather than suffer from persistent losses, investors perceive losses as temporary (Hayn 1995). The coefficient on NI*LOSS reflects how the market's valuation of losses differs from its valuation of profits. To the extent losses are more weakly associated with firm value than profits, I expect that [[beta].sub.2] < 0. Since assets represent probable future economic benefits, while liabilities represent probable future sacrifices of economic benefits, I expect the coefficient on ASSETS to be positive ([[beta].sub.3] > 0), and the coefficient on LIABS to be negative ([[beta].sub.4] < 0).

The remaining terms in Equation (2) pertain to the valuation of the off-balance-sheet activities concealed by the equity method. As reported shareholders' equity does not differ between the equity method and proportionate consolidation, off-balance-sheet assets are by definition equal to off-balance-sheet liabilities. Thus, both amounts cannot be included in the regression model. (12) To address this issue, I include OBS, and interpret the coefficient estimate for OBS*NOGUAR as the impact of off-balance-sheet activities on MVE, incremental to reported assets and liabilities. Specifically, a positive coefficient implies that the market values investee assets in excess of non-guaranteed obligations. Conversely, a negative coefficient implies that the market values non-guaranteed investee liabilities in excess of the related assets. If the market values each dollar of resources and obligations equally, then the coefficient will be zero. Thus, there is no clear expectation for the sign of [[beta].sub.5].

As indicated above, investing firms sometimes provide explicit guarantees of investee liabilities, thereby creating a legal obligation. Explicit guarantees include formal debt guarantees, deficiency agreements, take-or-pay contracts, and the like. Although the precise dollar amount of some guarantees can be identified, others cannot be quantified. For example, Cummins Inc. discloses that it is "required to purchase product at transfer prices sufficient to allow recovery of 50 percent of (the investee's) cost, including interest and financing expenses." Because the guarantee creates a legal obligation, I expect the valuation multiple on guaranteed off-balance-sheet liabilities to exceed that on the associated off-balance-sheet assets. Thus, I expect that [[beta].sub.6] < 0.

Table 2 presents information concerning the variables in Equation (2). Panel A includes descriptive statistics expressed in millions of dollars. The sample includes many large firms, as evidenced by the median market capitalization of over $1 billion. However, the raw data are highly skewed as the mean for each of the continuous variables exceeds the 75th percentile. To address the problem of interfirm scale differences, all variables are scaled by ASSETS (note that scaling ASSETS by itself results in a value of 1 for all observations). As seen in Panel B of Table 2, scaling results in "better behaved" variables. Only the means for NI (due to several large loss observations) and OBS*NOGUAR lie outside the interquartile range. Based on statistics designed to assess the impact of individual observations on the magnitude of coefficient estimates (Belsley et al. 1980), these observations do not significantly affect the reported regression results.

Table 3 reports the results from estimating the valuation model; coefficients' expected signs appear in the second column. Considering first the coefficient estimates based on the year 2000 data, the coefficient for NI (10.50, t = 4.38) is significantly positive, while the coefficient estimate for NI*LOSS (-15.02, t = -4.88) is significantly negative. The coefficient estimates on ASSETS (2.16, t = 4.98) and LIABS (-2.37, t = 4.31) are also consistent with economic intuition.

Note that both of the coefficient estimates reflecting the valuation of off-balance-sheet activities--OBS*NOGUAR (-0.99, t = -2.62) and OBS*GUAR (-1.72, t = -3.10)--are significantly less than zero. This is consistent with market participants placing more weight on investee liabilities than assets.

The last column of Table 3 reports the coefficient estimates based on 2001 data. Similar to 2000, the coefficient estimates for NI (15.67, t = 5.86) and ASSETS (1.53, t = 4.36) are significantly positive, while the coefficients for NI*LOSS (-17.46, t = -6.47) and LIABS (-1.51, t = -3.17) are significantly negative.

The results pertaining to off-balance-sheet activities differ from those of the prior year. Although the coefficient estimate for OBS*NOGUAR (-0.17, t = -1.10) is negative, it is not significantly different from zero. Thus, contrary to the results from 2000, there is only weak evidence in 2001 that the market values non-guaranteed investee obligations in excess of the associated assets. However, the results for guaranteed investee obligations are consistent with prior results, with the coefficient on OBS*GUAR (-1.35, t = -1.76) being significantly less than zero.

The result that market participants place greater weight on off-balance-sheet liabilities than assets for firms providing explicit guarantees of investee obligations provides empirical support for FIN No. 45. In addition to expanded disclosure about the nature of any guarantees, FIN No. 45 requires liability recognition for the fair value of the obligation undertaken when issuing new guarantees or modifying existing guarantees after December 31, 2002.

SUMMARY

This study examines the financial reporting effects of off-balance-sheet activities whose details are concealed by the equity method of accounting. First, based on a review of sample-firm footnote disclosures, I offer suggestions for improving the usefulness of these disclosures for financial analysis. Second, I estimate the market's valuation of off-balance-sheet activities. The results indicate that market participants place more weight on off-balance-sheet liabilities than assets for firms that provide explicit guarantees of investee obligations.

The study has two main implications. First, although financial statement users can utilize footnote disclosures to consolidate investees on a pro forma basis, limitations of current disclosure requirements affect the accuracy of such analyses. Second, based on the association between firm value and off-balance-sheet activities of equity method investees, market participants generally find the disclosure of these off-balance-sheet activities useful. The finding that market participants value off-balance-sheet liabilities supported by guarantees similar to reported liabilities provides empirical support for recent rulemaking.

TABLE 1

Magnitude of Investee Activities

                                                   n (a)       Mean

Investment account ($ million)                    150          542.4
Investment account as % of investor's
  reported assets                                                9.4
Equity income ($ million)                         150           67.1
Equity income as % of investor's reported
  net income (b)                                                65.2
Share of investee assets ($ million) (c)          150          820.1
Share of investee assets as % of investor's
  reported assets                                               11.2
Share of investee liabilities ($ million) (c)     150          820.1
Share of investee liabilities as % of
  investor's reported liabilities                               22.6
Share of investee sales ($ million) (c)           141 (d)    1,916.5
Share of investee sales as % of investor's
  reported sales                                                17.5

                                                 Std. Dev.       25%

Investment account ($ million)                   1,589          18.31
Investment account as % of investor's
  reported assets                                   14.9         2.6
Equity income ($ million)                          293          -1.0
Equity income as % of investor's reported
  net income (b)                                   182.1         6.1
Share of investee assets ($ million) (c)         2,591          29.2
Share of investee assets as % of investor's
  reported assets                                   18.9         3.2
Share of investee liabilities ($ million) (c)    2,591          29.2
Share of investee liabilities as % of
  investor's reported liabilities                   51.5         4.6
Share of investee sales ($ million) (c)          5,933          67.0
Share of investee sales as % of investor's
  reported sales                                    29.1         4.6

                                                    50%         75%

Investment account ($ million)                     70.55       311.6
Investment account as % of investor's
  reported assets                                   4.9          9.8
Equity income ($ million)                           3.0         19.6
Equity income as % of investor's reported
  net income (b)                                   15.6         35.9
Share of investee assets ($ million) (c)           86.7        328.5
Share of investee assets as % of investor's
  reported assets                                   6.1         11.9
Share of investee liabilities ($ million) (c)      86.7        328.5
Share of investee liabilities as % of
  investor's reported liabilities                  10.0         23.7
Share of investee sales ($ million) (c)           283.7        985.2
Share of investee sales as % of investor's
  reported sales                                   10.0         15.9

(a) Data include 150 firm-year observations (two annual observations
for each of 75 sample firms).

(b) This ratio is based on absolute values due to observations
with losses.

(c) Investee assets, liabilities, and sales represent the incremental
amounts to be reported under pro forma proportionate consolidation.

(d) The number of observations does not equal 150 due to unreported
investee sales (4 observations) and zero investor sales
(5 observations for holding companies).

TABLE 2

Descriptive Statistics for Valuation Model Data (a)

Variable (b)          Mean     Std. Dev.    25%       50%        75%

Panel A: Unscaled ($ millions)
MVE             9,899.8     35,714.7     222.8     1,050.4    3,816.7
NI                433.2      1,992.7     -10.5        16.6      136.5
LOSS                0.33         0.47      0           0          1
ASSETS          7,764.2     19,323.8     446.7     1,729.5    6,676.2
LIABS           4,847.8     10,531.8     220.6     1,285.4    3,615.0
OBS               820.1      2,591.3      29.2        86.7      328.5
NOGUAR              0.65         0.48      0           1          1
GUAR                0.35         0.48      0           0          1

Panel B: Scaled by ASSETS
MVE                 1.080        1.058     0.335       0.734      1.327
NI                 -0.025        0.181    -0.017       0.015      0.059
LIABS               0.624        0.230     0.489       0.627      0.763
OBS                 0.112        0.189     0.032       0.061      0.119
OBS*NOGUAR          0.070        0.187     0           0.020      0.061
OBS*GUAR            0.042        0.085     0           0          0.050

(a) Data include 150 firm-year observations (two annual observations
for each of 75 sample firms).

(b) Variable definitions:

MVE = market value of common equity at April 30;
NI = net income from continuing operations;
LOSS = 1 if NI < 0, 0 otherwise;
ASSETS = reported total assets;
LIABS = reported total liabilities;
OBS = additional assets/liabilities resulting from
proportionate consolidation;
NOGUAR = 1 if investee debt is not guaranteed, 0 otherwise; and
GUAR = 1 if an explicit guarantee of investee debt is disclosed,
0 otherwise.

TABLE 3

Estimation of Valuation Model

(all variables scaled by ASSETS, t-statistics in parentheses)

MV[E.sub.i] = [[beta].sub.0] + [[beta].sub.1] N[I.sub.i]
* [[beta].sub.2] N[I.sub.i] * LOS[S.sub.i] + [[beta].sub.3]
ASSET[S.sub.i] + [[beta].sub.4] LIAB[S.sub.i] + [[beta].sub.5]
OB[S.sub.i] * NOGUA[R.sub.i] + [[beta].sub.6] OB[S.sub.i] *
GUA[R.sub.i] + [[gamma].sub.i].

Variable (a)              Sign         Year 2000       Year 2001

# observations                           75              75
Constant                   ?            -26.01 *         -1.72
                                        (-2.56)         (-0.35)
NI                         +             10.50 **        15.67 **
                                         (4.38)          (5.86)
NI*LOSS                    -            -15.02 **       -17.46 **
                                        (-4.88)         (-6.47)
ASSETS                     +              2.16 **         1.53 **
                                         -4.98           (4.36)
LIABS                      -             -2.37 **        -1.51 **
                                        (-4.31)         (-3.17)
OBS*NOGUAR                 ?             -0.99 *         -0.17
                                        (-2.62)         (-1.10)
OBS*GUAR                   -             -1.72 **        -1.35 *
                                        (-3.10)         (-1.76)
Adjusted [R.sup.2]                        0.523           0.508

*,** Denotes significance at the 0.05 and 0.01 levels or better,
respectively, one-tailed test (except for Constant and OBS*NOGUAR).
Significance is based on White (1980) heteroscedasticity-adjusted
t-statisctis.

(a) Variable definitions:

MVE = market value of common equity at April 30;

NI = net income from continuing operations;

LOSS = 1 if NI < 0, 0 otherwise;

ASSETS = reported total assets;

LIABS = reported total liabilities;

OBS = additional assets/liabilities resulting from
proportionate consolidation;

NOGUAR = 1 if investee debt is not guaranteed, 0 otherwise; and

GUAR = 1 if an explicit guarantee of investee debt is
disclosed, 0 otherwise.

(1) These options are currently not available, except as voluntary disclosures.

(2) Goodwill is no longer amortized to expense in accordance with FASB Statement No. 142 (SFAS No. 142, FASB 2001). However, the investor's income from the investee could be reduced if the goodwill implicit in the investment becomes impaired. Note also that the investor's share of an extraordinary item must be reported on a separate line in the investor's income statement.

(3) An exception occurs when an investor's share of investee losses exceeds the carrying amount of the investment (APB No. 18, [paragraph] 19).

(4) A related alternative, the expanded equity method, represents a compromise between the equity method and proportionate consolidation in that investee accounts are presented separate from the investor's accounts (see Dieter and Wyatt 1978). Davis and Largay (1999) assert that the expanded equity method may be preferable to proportionate consolidation when investees are not operationally related to the investee.

(5) Under the market value method prescribed in FASB Statement No. t 15 (FASB 1993), securities with a public market are reported on the balance sheet at their current market value, while the income statement includes dividends, realized changes in market value, and unrealized value changes in trading portfolio securities.

(6) Regulation S-X, Rule 1-02(w) specifies the guidelines for determining the significance of an investee. Guidance regarding test calculations, including how to handle loss figures and decreases in income from historical levels, appears in the computational notes to Rule 1-02(w) and SEC staff interpretations.

(7) The need to make estimates of overall ownership interest is not uncommon. In the current sample' estimates of composite ownership percentages are required for 37 percent of the firms.

(8) The elimination of mandatory annual goodwill amortization under SFAS No. 142 will mitigate this problem. However, because SFAS No. 142 requires firms to test annually for goodwill impairments, financial statement users must determine whether equity income includes any loss from such impairments.

(9) Four sample firms in the oil and gas industry, including Chevron, present both aggregated data for all investees and separate disclosure of the investor's proportionate share of assets, liabilities, etc. For Chevron, the overall error in estimated liabilities increases to +34 percent when its unspecified interest in other investees is considered. The error rates in 2000 for the other firms are -2 percent (Conoco), -3 percent (duPont), and +3 percent (ExxonMobil). While die estimates axe accurate in three of the four oil and gas cases, large errors may result.

(10) Under GAAP, losses in excess of investment are reported as a liability when the investor has economic exposure, such as a loan guarantee.

(11) Including LOSS and GUAR dummies as additional variables in the model has no effect on the sign or magnitude of the coefficients of interest.

(12) When two variables are perfectly collinear, ordinary least squares estimates cannot be computed.

REFERENCES

Accounting Principles Board (APB). 1971. The Equity Method of Accounting for Investments in Common Stock. Opinion No. 18. New York, NY: AICPA.

AICPA Special Committee on Financial Reporting. 1994. Improving Business Reporting: A Customer Focus. New York, NY: AICPA.

Belsley, D. A., E. Kuh, and R. E. Welsch. 1980. Regression Diagnostics: Identifying Influential Data and Sources of Collinearity. New York, NY: John Wiley & Sons.

Bierman, H. 1992. Proportionate consolidation and financial analysis. Accounting Horizons 6: 5-17.

Canadian Institute of Chartered Accountants (CICA). 2000. CICA Handbook. Toronto, Ontario: CICA.

Chang, J. 2002. Wall Street sharpens focus on off-balance-sheet items. Chemical Market Reporter (April 8).

Ciesielski, J. T. 2002. Reading the 2001 annual reports: A user's guide. The Analysts' Accounting Observer 11.

Collins, D. W., M. Pincus, and H. Xie. 1999. Equity valuation and negative earnings: The role of book value of equity. The Accounting Review 74: 29--61.

Dhaliwal, D. 1986. Measurement of financial leverage in the presence of unfunded pension obligations. The Accounting Review 61:651-661.

Davis, M. L., and J. A. Largay III. 1999. Financial reporting of "significant-influence" equity investments: Analysis and managerial issues. Journal of Managerial Issues 11: 280-298.

Dieter, R., and A. R. Wyatt. 1978. The expanded equity method: An alternative in accounting for investments in joint ventures. Journal of Accountancy 145: 89-93.

Ely, K. M. 1995. Operating lease accounting and the market's assessment of equity risk. Journal of Accounting Research 33: 397-415.

Financial Accounting Standards Board (FASB). 1985. Elements of Financial Statements. Statement of Financial Accounting Concepts No. 6. Norwalk, CT: FASB.

--. 1987. Consolidation of All Majority-Owned Subsidiaries. Statement of Financial Accounting Standards No. 94. Norwalk, CT: FASB.

--. 1993. Accounting for Certain Investments in Debt and Equity Securities. Statement of Financial Accounting Standards No. 115. Norwalk, CT: FASB.

--. 2001. Goodwill and Other Intangible Assets. Statement of Financial Accounting Standards No. 142. Norwalk, CT: FASB.

--. 2002. Guarantor's Accounting and Disclosure Requirements for Guarantees, Including Indirect Guarantees of Indebtedness of Others. Interpretation No. 45. Norwalk, CT: FASB.

Graham, R. C., and C. Lefanowicz. 1996. Evidence of the relation between accounting for equity investments and equity valuation. The Journal of Accounting, Auditing, and Finance 11: 587-605.

--, --, and K. Petroni. 1998. The use of the equity method and the value-relevance of fair value disclosures. Working paper, Oregon State University, Michigan State University.

--, R. D. King, and C. K. J. Morrill. 2003. Decision usefulness of alternative joint venture reporting methods. Accounting Horizons 17 (2): 123-138.

Hayn, C. 1995. The information content of losses. Journal of Accounting and Economics 26: 125-153.

Imhoff, E. A., Jr., R. C. Lipe, and D. W. Wright. 1993. The effects of recognition versus disclosure on shareholder risk and executive compensation. The Journal of Accounting, Auditing, and Finance 8: 335-368.

International Accounting Standards Board (IASB). 1990. Financial Reporting of Interests in Joint Ventures. IAS No. 31. London, U.K.: IASB.

Landsman, W. 1986. An empirical investigation of pension fund property rights. The Accounting Review 61: 662-691.

Lasman, D. A., and R. L. Weil. 1978. Adjusting the debt-equity ratio. Financial Analysts Journal 34: 49-58. Securities and Exchange Commission (SEC). 1981. Separate Financial Statements Required by Regulation SX. Accounting Series Release No. 302. Washington, D.C.: SEC.

Shevlin, T. 1987. Taxes and off-balance-sheet financing: Research and development limited partnerships. The Accounting Review 62: 480-509.

Stickney, C. P., and P. R. Brown. 1999. Financial Reporting and Statement Analysis: A Strategic Perspective. Fourth edition. Fort Worth, TX: The Dryden Press.

White, G. I., A. C. Sondhi, and D. Fried. 2003. The Analysis and Use of Financial Statements. Third edition. New York, NY: John Wiley & Sons, Inc.

White, H. 1980. A heteroscedasticity-consistent covariance matrix estimator and a direct test for heteroscedasticity. Econometrica 48: 817-838.

Mark P. Bauman is an Assistant Professor at the University of Wisconsin-Milwaukee.

I thank Chris Bauman, Paul Kimmel, Robert Lipe, and two anonymous referees for helpful comments and suggestions. I also acknowledge the capable research assistance of Jessica Marschall.

Submitted: August 2002

Accepted: May 2003

Corresponding author: Mark P. Bauman

Email: mpbauman @uwm.edu