ACCOUNTING FOR JOINT VENTURES AND ASSOCIATES IN ...

ACCOUNTING FOR JOINT VENTURES AND ASSOCIATES IN CANADA, UK, AND US: DO US RULES HIDE INFORMATION?

Kazbi Soonawalla∗ London School of Economics and Political Science May 2005

∗

I am indebted to my thesis chair, Mary Barth, for invaluable support and guidance. My sincere thanks my other committee members Bill Beaver and George Foster for help and guidance. I am very grateful to an anonymous referee for valuable suggestions. I thank workshop participants at Columbia University, Emory University, London Business School, London School of Economics and Political Science, Oxford University, Stanford University, University of California at Berkeley, and University of Washington for helpful suggestions. I gratefully acknowledge financial assistance from the Olin Foundation. All mistakes are my own.

Abstract: Using Canadian, UK, and US data, this study investigates the potential loss of forecasting and valuation relevant information from aggregating joint venture and associate accounting amounts. Unlike US GAAP, accounting principles in Canada and the UK require disclosure of disaggregated components of joint ventures and associates. Findings show that aggregating joint venture and associate earnings, and aggregating joint venture revenues and expenses, each leads to loss of forecasting and valuation relevant information. Thus, current US accounting principles mask information that financial statement users could use to predict future earnings and explain share prices.

1

1.

INTRODUCTION

This study investigates whether income statement information relating to joint ventures and associates is relevant for earnings forecasting and equity valuation.1 The study bases inferences on the association between recognized amounts for various income statement components, and share prices and future earnings, to address the following questions: (1) Is there loss of information for earnings forecasting and equity valuation when joint venture earnings are aggregated with earnings from associates? (2) Is there loss of information for forecasting and valuation purposes when joint venture revenues and expenses are aggregated together as joint venture earnings? An issue currently debated internationally is the appropriate method of accounting for joint ventures and associates. In most countries either the equity method or proportional consolidation is required for joint venture investments, and the equity method is required for associates.2 Some countries, such as Canada and the United Kingdom (UK), require joint venture and associate amounts to be reported separately, as well as additional disclosures relating to these investments. Other countries, such as the United States (US), require no additional disclosures and do not even require that joint venture and associate amounts be reported separately. Due to the considerable diversity in practice, the accounting treatment for these investments is of interest to standard setters. At a meeting in London in September 2002, the International Accounting Standards Board (IASB) and representatives from national standard setters concluded that accounting for interests in joint ventures should be the subject of an 1

“Associates” is the term used to describe “equity investments in associated companies” or “other equity investments.” They are known as long-term investments in Canada. Throughout this paper the UK term “associates” is used for these types of investments, and associate earnings refers to earnings from other equity investments or associated companies. 2 This study is not concerned with the differences between the equity method and proportional consolidation. Instead, it focuses on the disclosure of disaggregated accounting amounts relating to joint ventures and associates.

2

international convergence project. In March 2003, the Canadian Accounting Standards Board (AcSB) asked financial statement preparers whether they would favour a change in reporting practice from proportional consolidation to the equity method. As yet, no decision has been made regarding a global accounting standard. Accounting treatments for joint ventures and associates vary considerably, as demonstrated by the differences in US, Canadian, and UK GAAP. The equity method is required in the US, whereby the sum of net earnings from joint ventures and associates (also called equity earnings), and the sum of net investments in joint ventures and associates (also called equity investments), are presented as single line items in the income statement and balance sheet, respectively (APB 18, 1971). Operating results are aggregated in two ways. First, joint venture earnings are aggregated with earnings from other equity investments in associated companies. Second, joint venture revenues and expenses are aggregated together as joint venture earnings. Although the aggregate of joint venture and associate amounts are recognized under US GAAP, disaggregated earnings and equity book value components of these investments are not reported separately. US firms do not disclose information sufficient to conduct empirical analyses on joint venture and associate accounting amounts. Canadian and UK GAAP require that investor firms disclose their pro rata share of joint venture revenues, operating income, assets, and liabilities.3 Canadian GAAP requires proportional consolidation for reporting joint venture interests along with disclosures of disaggregated revenues, earnings, assets, liabilities, and cash flows. Under proportional consolidation, the investor’s pro rata share of each of the assets, liabilities, revenues, and expenses that are subject to joint control is combined on a line-by-line basis with the investor’s 3

In this paper the term “investor” or “investor firms” refers to the firm that has entered into the joint venture arrangement and/or owns an interest in the associate. Investor firms are often referred to as “co-venturers” or “joint venturers” or “parents” in the academic literature and popular press.

3

other assets, liabilities, revenues, and expenses. UK GAAP requires the gross equity method for joint ventures, whereby disaggregated shares of joint venture revenues, earnings, assets, and liabilities are disclosed in addition to equity method recognition. Although the accounting treatments and disclosure requirements for joint ventures are different under Canadian and UK GAAP, firms in both countries are required to provide disaggregated information sufficient to address the research questions in this study. Both countries use equity method reporting for associates, and associate earnings and investment amounts are reported separately from joint venture earnings and investments.4 The availability of this data creates a rich environment for investigating the information lost under US GAAP. This study investigates the potential loss of information from failing to provide detailed disaggregated income statement and balance sheet information, using samples of Canadian, UK, and US firms. A loss of usefulness of aggregated accounting amounts for predicting future earnings implies loss of forecasting relevance. Similarly, aggregated accounting amounts that diminish the information used by shareholders in valuing the firm reduce value relevance of financial statements (Barth et al., 2001). The evidence provided bears directly on reporting practices under Canadian and UK GAAP. The evidence also bears indirectly on current issues facing the US Financial Accounting Standards Board regarding consolidations and off-balance sheet entities. The study predicts that presenting aggregated data for joint ventures and associates results in loss of information that could otherwise be used to forecast future earnings and explain share prices. It tests these predictions by estimating relations between future earnings or share prices, and investor earning components, joint venture earning components, associate earnings, and their corresponding equity book value components. The investigation is performed on a 4

Disaggregated amounts on associate revenues and expenses are not disclosed in either country.

4

sample of 68 Canadian firms and 91 UK firms that have joint ventures and associates in the late nineties.5 Where possible, analogous analyses are conducted on a sample of US firms. This establishes the relation between investor earnings and earnings from equity investments. It permits inferences about the potential benefits of additional disclosures in the US, from Canadian and UK data. Initial findings show that equity earnings are significantly different from investor operating earnings for equity valuation, across the Canadian, UK, and US samples. This establishes that the valuation relation is similar in all three accounting jurisdictions. As predicted, aggregating joint venture and associate earnings suppresses forecasting and valuation relevant information. Also as predicted, aggregating joint venture revenues and expenses results in loss of forecasting and valuation relevant information. Overall, the results indicate that failure to report disaggregated information on joint ventures and associates suppresses information that is potentially useful to financial statement users. The rest of this paper proceeds as follows. Section 2 describes relevant accounting standards under Canadian, UK and US GAAP. Section 3 reviews related background literature. Section 4 presents the research design. Section 5 presents descriptive statistics and findings. Section 6 concludes.

2. RELEVANT ACCOUNTING METHODS FOR JOINT VENTURES AND ASSOCIATES Currently, no single definition or accounting method for joint venture investments exists across accounting jurisdictions or within the United States (AICPA, 1979; Nobes and Parker, 2002;

5

The relevant accounting standards requiring joint venture and associate amounts to be separately disclosed have been in effect since 1996 in Canada and 1998 in the UK.

5

Roberts et al., 1998; Milburn and Chant, 1999). The key feature that sets joint ventures apart from other investments and operations is that joint ventures are jointly controlled by the investor firms. Joint control is the sharing of power between investors: no single investor unilaterally controls joint ventures, and the joint consent of all co-investors is required for major operating and financing policy decisions. In contrast to joint control, investor firms have participating interests in associates and exercise significant influence over the associates’ activities. The joint consent of all investors is not required for operating and financing decisions of associates and, in general, investors have less control over associates than over joint ventures (Bierman, 1992). While significant influence gives the power to participate in financial and operating policy decisions of the investee (IAS 28, 1988), it gives the investor less power and decision-making ability over associates’ financial and operating activities than joint control or full control (FRS 9, paragraph 6, 1997). One of the contentious issues surrounding accounting treatments for joint ventures is that not all accounting jurisdictions agree on the definition of “joint control,” which is the key defining feature of joint ventures.6 The G4+1 proposed a definition that is different from the definition used in the relevant international accounting standard (IAS 31, 1990). The Canadian and UK definitions of joint ventures also vary. Under Canadian GAAP, jointly controlled assets, operations, and enterprises are considered joint ventures. These jointly controlled assets and operations need not be entities distinct from the investor. Under UK GAAP, joint ventures must

6

Under Canadian GAAP, “a joint venture is an economic activity resulting from a contractual arrangement whereby two or more venturers jointly control the economic activity” (CICA 3055.03). Under UK GAAP a joint venture is “an entity in which the reporting entity holds an interest on a long-term basis and is jointly controlled by the reporting entity and one or more other venturers under a contractual arrangement” (FRS 9, paragraph 4). US GAAP only talks about incorporated joint ventures and these would be corporations “owned and operated by a small group of businesses (the joint venturers) as a separate and specific business or project for the mutual benefit of the members of the group” (APB Opinion No. 18).

6

be distinct entities, and thus jointly controlled assets and operations are not considered joint ventures.7 Ownership interests in both joint ventures and associates usually range from 20 to 50 percent, so ownership amounts alone do not help distinguish between joint ventures and associates. The distinguishing feature is the amount of control the investor has over the investment. Control over joint ventures and associates is shared but by differing amounts, making them distinctive from each other and from other operations and activities that are fully controlled by the investor. Canadian GAAP has required proportional consolidation of joint ventures with additional disclosures, since 1996 (CICA, 1996). Proportional consolidation accounts for the investor's interest by consolidating in the investor's financial statements the investor's share of joint venture assets, liabilities, revenues, and expenses. In Canada, the investor is also required to disclose the following amounts related to its interests in joint ventures: (a) current and long-term assets, (b) current and long-term liabilities, (c) revenues, expenses, and income, and (d) cash flows from operating, financing, and investing activities.8 The treatment for associates is different from that for joint ventures. Canadian GAAP requires that associate investments be accounted for using the equity method, whereby net operating earnings from and net investments in associates are represented by single line items in the income statement and balance sheet, respectively (CICA, 1994). UK GAAP has required the gross equity method for joint venture investments since 1998 (FRS 9, 1997). This method requires equity method recognition in the income statement and balance sheet, with the investor's share of the gross assets and liabilities underlying the 7

This may hinder cross-country comparison to some extent. However, the key feature of joint control is required for Canadian and UK joint ventures. 8 Note that investors disclose the aggregate for these amounts across all their joint venture investments.

7

investment shown, in aggregate, on the face of the balance sheet. Investor’s share of joint venture earnings is presented as a separate line item in the income statement. Similarly, joint venture revenues are disclosed as an additional line item in the income statement before equity method revenues. Similar to Canadian GAAP, under UK GAAP, associates are accounted for using the equity method. In the US, the equity method is required for joint ventures and associate investments whereby the sum of joint venture and associate earnings is presented in aggregate in the income statement and the aggregate net investment in joint ventures and associate investments is presented in the balance sheet (APB 18, 1971; AICPA, 1979; Milburn and Chant, 1999). The varying control levels for joint ventures and associates lends itself to investigate whether joint venture and associate accounting amounts should be disclosed separately, and whether joint venture revenues and expenses should be disclosed separately. Investigating these issues requires data from countries in which both joint venture and associate earnings are disclosed separately from other income. In some accounting jurisdictions such as the US the only available accounting amounts on these investments are the aggregate of joint venture and associate earnings, and the aggregate of joint venture and associate investments. Because the necessary disaggregate accounting amounts are available in Canada and the UK, this study uses this data to investigate the research questions. The study obtains data on joint venture revenues, earnings, and net investment amounts for Canadian firms from footnote disclosures, and calculates joint venture expenses as joint venture revenues minus earnings. The study also obtains associate earnings and investment amounts from the relevant line items in the income statement and balance sheet. For UK firms,

8

the study obtains joint venture revenues, earnings, and net investment amounts, and associate earnings and net investment amounts from the income statement and balance sheet, respectively. Overall, the rules under Canadian and UK GAAP provide that level of disclosure reflects the degree of the investor’s control over the investee; i.e. disclosure is high for subsidiaries, intermediate for joint ventures, and low for associates (Flower and Ebbers, 2002). Therefore, Canadian and UK GAAP provide information that US GAAP fails to. US GAAP requires aggregation across investments that are inherently different in nature from each other and subject to different degrees of control. From a forecasting and value relevance perspective, permitting reporting of aggregated accounting amounts of joint venture and associate earnings, or joint venture revenues and expenses implies that the disaggregated components are not viewed to be relevant to financial statement users for predicting future earnings and estimating current share prices. If knowing the disaggregated accounting amounts is relevant to a financial statement user it can also be capable of making a difference to that user’s decisions. Because of differences in the investments and differences between full control, joint control, and significant influence, lack of reporting requirements for the disaggregated components potentially limits the usefulness of financial statements and hinders the user’s decision-making ability. However, based on information provided and on institutional features of these investments, it is difficult to predict whether any single component is more significantly associated with future earnings or equity value than any other component or combination of components.

3. REVIEW OF RELATED RESEARCH Prior research documents an association between joint venture investments and share prices (Koh and Venkatraman, 1991; Madhavan and Prescott, 1995; Park and Kim, 1997). However, it does

9

not document whether joint venture accounting amounts are associated with share prices. The extant literature also does not document whether there is a differential association between investor accounting amounts and joint venture and associate accounting amounts, and share prices. Three factors affect whether joint venture and associate earnings have different associations with prices. First, investor firms have different degrees of control over fully controlled operations, joint ventures, and associate investments. Investors control majority owned investments, jointly control joint ventures, and have only significant influence over associates. Some suggest that the power of joint control is substantively greater than significant influence and accordingly jointly controlled investments should be distinguished from other equity investments (Flower and Ebbers, 2002). Park and Kim (1997) find that the market reacts positively to joint venture announcements if the investor has more control over the joint venture, suggesting that degree of control affects equity valuation. Second, joint ventures tend to be of finite duration (Kogut, 1988; Park and Russo, 1996; Williamson, 2000). Subramanyam and Wild (1996) show that earnings persistence is inversely related to probability of termination of the firm. This suggests that joint ventures have lower earnings persistence, which Ohlson (1999) shows would be translated into lower association with market value of equity.9 Third, joint ventures and associates may be in different lines of business or industry from the investor and, accordingly, have different industry-specific properties such as earnings persistence, share returns, and financial risks and rewards (Dieter and Wyatt, 1978; Harrigan, 1988). Joint ventures may be formed to invest and expand in new markets that could result in 9

Joint venture terminations often result in reallocation of ownership between the existing investor firms, where the investor firm effectively uses the joint venture as an option to expand (Kogut 1991). This study does not explore reasons for joint venture termination.

10

unpredictable earnings and differing associations with market value of equity from other controlled and significant influence investments (Kogut, 1991). There is very little extant literature on accounting for joint venture investments. Graham et al. (2001) compare proportional consolidation and the equity method and find that proportionally consolidated financial statements are more useful in predicting future returns on shareholders’ equity than equity method financial statements. Kothavala (2003) investigates the risk relevance of joint venture accounting amounts and ratios for a sample of Canadian firms and finds that disaggregated information on joint venture accounting amounts helps explain variation in market risk. However, Graham et al. (2001) and Kothavala (2003) focus only on Canadian firms with joint ventures. This study differs in that it focuses on the information lost from various levels of aggregation and uses data from Canadian, UK, and US firms, providing a comprehensive view of the issues relating to accounting for joint ventures and associates. Maines et al. (2000) investigate the role of analyst knowledge and supplemental disclosures on joint venture investments. They examine whether analysts assign different equity values depending on whether a firm uses the equity method or proportional consolidation to report joint venture interests. They find that analysts with low familiarity in joint venture accounting rules assigned higher equity values to firms with equity method financial statements than to firms with proportionally consolidated financial statements. However, disclosure of separate joint venture financial information eliminated assigned equity value differences. Their study supports the idea that aggregating joint venture accounting amounts leads to loss of value relevant information. It also suggests that analyst perception of economically identical joint venture investments varies depending on the accounting treatment used and that these differences can be undone with additional disclosures.

11

This study also contributes to the literature on the information content of earnings and its components. Prior research generally finds that disaggregating earnings into its components, such as cash flows, revenues, and segment earnings, increases the predictive ability and valuation relevance of earnings (Lipe, 1986; Wilson, 1986; Wilson, 1987; Bernard and Stober, 1989; Balakrishnan et al., 1990; Livnat and Zarowin, 1990; Swaminathan and Weintrop, 1991; Dechow, 1994; Barth et al., 1999). This study extends prior research by partitioning joint venture and associate earnings, and equity book value components from investor earnings and equity book value to investigate loss of information from aggregation.

4. RESEARCH DESIGN (i)

Disaggregating Joint Venture and Associate Earnings

To investigate whether there is loss of information for earnings forecasting or equity valuation when joint venture earnings are aggregated with associate earnings, the study builds on the following set of equations:10 NIt+1 = ω0 + ω1234 NIt + ω567 BVt + ε1t+1

(1)

Pt = α0 + α1234 NIt + α567 BVt + ε2t

(2)

where P is share price at time t, NI is net income of the investor including joint venture and associate earnings, and BV is equity book value of the investor including joint venture and associate investments. NI and BV are included in equations (1) and (2) as summary measures of financial statement information (Ohlson, 1995). The first research question investigates the potential loss of information from aggregating joint venture and associate earnings given operating and non-operating earnings. Before testing

10

The “ω” coefficients presently have more than one subscript as the corresponding variables are subsequently disaggregated into their relevant earnings components and each component will retain a single subscript.

12

this, the study establishes the general relation between operating earnings and equity earnings. Given that the sum of joint venture and associate earnings are presented in the US as equity earnings, this investigation is carried out for Canadian, UK, and US firms. The estimating equations are: NIt+1 = ω0 + ω1 OIBefEQt + ω2 OINOt + ω34 OIEQt + ω5 BVBefEQt + ω67 BVEQt + ε3t+1 Pt = α0 + α1 OIBefEQt + α2 OINOt + α34 OIEQt + α5 BVBefEQt + α67 BVEQt + ε4t

(3) (4)

where OIBefEQ is operating earnings of the investor minus the sum of joint venture and associate earnings, OINO is non-operating earnings of the investor, OIEQ is equity earnings which is the sum of joint venture and associate earnings, BVBefEQ is equity book value of the investor minus the sum of joint venture and associate investments, and BVEQ is net equity investments which is the sum of net investments in joint ventures and associates. All variables, except price, are deflated by number of shares outstanding at fiscal year end.11 The study predicts that the coefficient on OIEQ is significantly different from the coefficient on OIBefEQ. After establishing the general relation above, the first research question tests whether the coefficients on joint venture operating earnings and associate earnings are statistically similar, and for this OIEQ is disaggregated into OIJV and OIAS. Simultaneously, the study investigates whether the corresponding balance sheet components are statistically similar for forecasting and valuation purposes. Disaggregating the equity book value coefficients is consistent with disaggregating the equity earnings variables, as presumably the joint venture and associate earnings are being generated by the joint venture and associate equity book value components, respectively. To this end, BVEQ is disaggregated into BVJV and BVAS. To address these research questions, the study estimates the following equations (5) and (6) and tests whether ω3 = ω4, ω1

11

The variables are deflated by number of shares outstanding to control for heteroscedasticity. Price is not deflated because it already exists in the deflated form, i.e. market capitalisation divided by number of shares outstanding.

13

= ω3, ω1 = ω4, α3 = α4, α1 = α3, and α1 = α4. OIJV is joint venture earnings, OIAS is associate earnings, BVJV is net investment in joint ventures, and BVAS is net investment in associates. NIt+1 = ω0+ω1OIBefEQt +ω2OINOt +ω3OIJVt+ω4OIASt+ω5BVBefEQt+ω6BVJVt+ω7BVASt+ ε5t+1

(5)

Pt = α0 + α1OIBefEQt+ α2OINOt + α3OIJVt +α4OIASt +α5BVBefEQt + α6BVJVt + α7BVASt + ε6t

(6)

The study predicts that the forecasting and equity valuation multiples on OIJV and OIAS are different, implying loss of information for financial statement users when joint venture and associate earnings are aggregated (Harrigan, 1988; Bierman, 1992; Subramanyam and Wild, 1996; Park and Kim, 1997). The study also predicts that joint venture and associate earnings are significantly different from operating earnings. Based on prior research, the study predicts that OIBefEQ, OIJV, OIAS, BVBefEQ, BVJV, and BVAS, have positive coefficients and OINO has a negative coefficient (Barth et al., 1999).

(ii)

Disaggregating Joint Venture Earnings into Revenues and Expenses

Building on the prior results, the next set of tests determine whether there is loss of information for earnings forecasting and equity valuation when joint venture revenues and expenses are aggregated into joint venture earnings. To do so, OIJV is disaggregated into REVJV and EXPJV, and the equalities µ4 = ─µ5 and δ4 = ─δ5 are tested in equations (7) and (8).12 Retaining a parallel disaggregation for the investor firm, operating earnings are disaggregated into revenues and operating expenses. Equations (3) through (8) are consistent with Ohlson (1999), where the

12

Readers are reminded that for associate investments the income statement has only a single line item showing associate earnings, and neither associate revenues nor expenses are disclosed.

14

various earnings components together represent net income.13 Associate earnings, OIAS, are retained as a disaggregate earnings component, resulting in the following equations: NIt+1=µ0+µ1REVt+ µ2EXPt+ µ3OINOt+µ4REVJVt+µ5EXPJVt+ µ6OIAst + µ7BVBefEQt+µ8BVJVt+µ9BVASt+ε7t+1

(7)

Pt = δ0+ δ1REVt+ δ2EXPt+ δ3OINOt+ δ4REVJVt+ δ5EXPJVt+ δ6OIAst + δ7BVBefEQt+ δ8BVJVt+ δ9BVASt +ε8t

(8)

where REV is investor revenues, EXP is investor expenses, REVJV is joint venture revenues, and EXPJV is joint venture expenses. Rejecting the magnitude equalities of REVJV and EXPJV coefficients would imply that presenting only the aggregate joint venture earnings amount suppresses information that is useful to financial statement users for earnings prediction and equity valuation purposes. The study also predicts that the coefficients on joint venture revenues and expenses are different from the coefficients on investor revenues and expenses.

5. DATA AND RESULTS (i)

Data

The study uses data from Canadian and UK firms that have joint venture and associate investments for years 1995-2000 and 1997-2000, respectively. The data were identified using keywords like “joint ventures,” “jointly controlled entities,” “joint arrangements,” and “FRS 9” for exhaustive country-specific searches on Global Researcher, Global Vantage, and Worldscope, to identify samples of Canadian and UK firms with interests in joint ventures.14

13

The income component in an Ohlson framework is generally abnormal earnings, which are calculated using an assumed cost of capital and previous period book value. Econometrically, using either earnings or abnormal earnings should not make a difference. 14 FRS 9 makes a distinction between joint ventures and joint arrangements. Joint ventures are accounted for under the gross equity method, whereas joint arrangements are accounted for under proportional consolidated. The search terms merely identify a list of firms that potentially have joint ventures. Therefore, the search often identified firms

15

The searches identified 110 Canadian firms and 133 UK firms that have joint venture investments. Of these, 68 Canadian firms and 91 UK firms also have associate investments, and provide financial statement information on joint ventures and associates for at least one year.15 Data on associates and joint ventures are hand collected from company financial statements. Specifically, joint venture earnings and revenues, joint venture net investments, associate earnings, and associate net investments are obtained from investor financial statements. Other financial statement information and price data are obtained from Compustat’s Canadian Industrial Annual, Global Vantage’s Global Commercial and Financial Services, and Datastream. Where possible, the study runs analogous investigations for a sample of US firms that report equity investments and earnings for the years 1995-2000. These firms are identified from Compustat’s US Industrial Annual files where equity investment amounts are non-zero or nonmissing. 10,268 such firm-year observations were identified. Of these, only observations having non-missing equity earnings and positive equity book value were retained, bringing the sample to 6,176 observations.

(ii)

Descriptive Statistics

Table 1, Panel A presents the descriptive statistics of the variables used in the study.16 Median joint venture earnings, OIJV, constitutes 12% of median net income, NI, for the Canadian firms and 10% of NI for the UK firms. In comparison, associate earnings, OIAS, contribute a smaller portion of net income in both countries, so that total equity income comprises mainly of joint that did not in fact have joint ventures as defined by FRS 9 and after careful examination of the annual reports these firms were excluded. 15 Recall that the relevant accounting standards have been in effect since 1996 for Canadian firms and 1998 for UK firms. Using restated observations results in six years of data for Canadian firms and four years for UK firms. 16 To mitigate the undue influence of extreme observations the extreme 1 percent of all observations are winsorised.

16

venture income. Whereas total equity income, OIEQ, for US firms is higher on average than that for Canadian and UK firms, US median equity income is quite a bit smaller than that of Canadian firms and similar to that of UK firms. The mean net equity investment, BVEQ, is considerably larger for US firms than for Canadian and UK firms. However, the median value is smaller than that of Canadian firms and larger than that of UK firms. It is worth noting that the variables OIEQ and BVEQ are slightly different for the US sample than for the Canadian and UK samples. The Canadian and UK samples consist of those firms that have both joint venture and associate investments. Given that US firms do not provide this disaggregate information, the US sample consists of firms that have either joint venture or associate or both types of investments. This likely explains the large variance in OIEQ and BVEQ for the US firms. Table 1, Panel B presents Pearson correlations for the Canadian and UK samples and Table 1, Panel C presents Pearson correlations for the US sample.17 For both the Canadian and UK samples, joint venture and associate earnings have low correlation with each other and with net income and price. For the US sample, OIEQ has a low correlation with price, P, but has a higher correlation with net income, NI, and net equity investment, BVEQ.

(iii)

Findings

As a precursor to the equations addressing the primary research question, equations (3) and (4) are estimated where joint venture and associate earnings, and joint venture and associate investments, are aggregated into single line items called equity earnings, OIEQ, and equity investments, BVEQ, respectively.18 In addition to Canadian and UK data, a sample of US data are also tested using this specification. The results in Table 2 show that, as predicted, forecasting 17

Untabulated Spearman correlations are similar in magnitude and significance to Pearson correlations. The results are robust to using total assets as an alternative deflator. t-statistics are based on White (1980) heteroscedasticity-consistent standard errors.

18

17

and valuation coefficients for OIEQ are generally significantly positive for all three countries.19 The t-statistics are 1.62, 2.13 and 19.10 for earnings forecasting and 1.85, 1.69, and 11.44 for equity valuation for the Canadian, UK, and US samples, respectively.20 These findings indicate that the sum of joint venture and associate earnings are forecasting and valuation relevant.21 Tests of equality of income statement coefficients of OIBefEQ and OIEQ, ω1 = ω34 and α1 = α34, are rejected for the Canadian, UK, and US samples for equity valuation, and for the US sample for earnings forecasting as well. As predicted, the coefficients on OIBefEQ and OINO are significantly positive and negative, respectively. On the equity book value side, the valuation coefficient on the equity investment book value component, BVEQ, is consistently significantly positive for all three samples. However, BVBefEQ has no positive association with future earnings, conditional on BVEQ, for the Canadian and UK samples (t = –1.18 for Canada and = –0.71 for UK), and has a significantly negative association with future earnings for the US sample (t = –3.25). This is inconsistent with expectations but similar to results obtained by the Barth et al. (1999). The findings also show that tests of equality for equity book value coefficients BVBefEQ and BVEQ, ω5 = ω67 and α5 = α67, are rejected for earnings forecasting and equity valuation for all three samples. The results from this analysis indicate that equity earnings and investments have significant associations with future earnings and share prices, and that knowing equity earnings and investments separately from investor earnings and equity book value is valuation relevant for financial statement users.

19

The exceptions are the earnings forecasting coefficient on OIEQ for the Canadian sample and the equity valuation coefficient on the UK sample where marginal significance is obtained. 20 The t-statistics are much higher for the US sample than for the Canadian and UK samples. This may be due to the much larger and high variance nature of the US sample. 21 The difference in sample size between the forecasting and valuation specification is due to the forecasting specification needing one additional years data compares to the valuation specification.

18

Next the study looks at the principle research question investigating the potential loss of information from aggregating joint venture and associate earnings. Summary statistics from estimating equations (5) and (6) are presented in Table 3. For these specifications, equity earnings, OIEQ, is disaggregated into OIJV and OIAS, and equity investment BVEQ is disaggregated into BVJV and BVAS. Given the unavailability of disaggregate US data, these tests are restricted to the Canadian and UK samples. To investigate the potential information loss, the equality of coefficients for OIJV and OIAS, ω3 = ω4 and α3 = α4, are tested. The findings indicate loss of valuation-relevant information, as indicated by statistically different coefficients for OIJV and OIAS, for both samples when joint venture and associate earnings are aggregated (p-value = 0.00 for Canada and = 0.05 for UK). Findings also indicate loss of earnings forecasting information from this aggregation for the UK sample, and marginally significant loss of information for the Canadian sample (t = 0.10 for Canada and = 0.02 for UK). The earnings forecasting and equity valuation coefficients for OIJV are significantly positive for both samples, whereas only the earnings forecasting coefficient on OIAS is significantly positive for both samples. For equity valuation the coefficient on OIAS is significantly negative for the Canadian sample and marginally negative for the UK sample. The tests of equality of coefficients for BVJV and BVAS, ω6 = ω7 and α6 = α7, are generally rejected for earnings forecasting and equity valuation for both the Canadian and UK samples, with p-values of 0.00 and 0.05 (0.07 and 0.04) for earnings forecasting and equity valuation, respectively, for the Canadian (UK) sample. Together the results indicate that aggregating joint venture and associate line items in the income statement and balance sheet masks information that financial statement users find useful for predicting earnings and estimating share prices. The tests also show that joint venture and associate earnings and investment components are

19

significantly different from investor operating earnings and equity book value for equity valuation. The study next investigates the loss of information when joint venture revenues are aggregated with joint venture expenses to give joint venture earnings. The results from estimating equations (7) and (8) are presented in Table 4. For this specification, joint venture earnings, OIJV, are disaggregated into REVJV and EXPJV. To maintain consistency across the investor and joint venture variables, operating earnings, OIBefEQ, are disaggregated into revenues, REV, and operating expenses, EXP.22 The remaining variables are carried forward as before. All variables are constructed so that the sum of income statement components gives total net income of the investor, and sum of balance sheet components gives total book value of equity of the investor. Tests of equality on revenue and expense coefficients for REVJV and EXPJV, µ4 = –µ5 and δ4 = –δ5, are rejected for earnings forecasting and equity valuation for both the Canadian and UK samples (p-values = 0.02 and 0.05 for Canada and = 0.00 and 0.00 for UK). As predicted, individual coefficients for REV and REVJV are significantly positive, the t-statistics for the coefficients on REVJV being 2.10 and 1.94 (2.33 and 1.71) for earnings forecasting and equity valuation, respectively, for the Canadian (UK) sample. Also as predicted, the individual coefficients on EXP and EXPJV are significantly negative, with t-statistics for the coefficients on EXPJV being –2.13 and –1.99 (–2.31 and –1.69), for earnings forecasting and equity valuation, respectively, for the Canadian (UK) sample. Additional tests show that investor revenues and joint venture revenues, and investor expenses and joint venture expenses are different from each other for equity valuation for both samples and for earnings forecasting for the UK sample. The

22

It is not possible to disaggregate associate earnings into revenues and expenses as this level of detailed information is not provided.

20

results from this analysis indicate that aggregating joint venture revenues and expenses into joint venture earnings results in loss of forecasting and valuation relevant information, as evidenced by statistically different coefficients on REVJV and –EXPJV. Interestingly, this is the case even though the revenue and expense coefficients appear to be of similar magnitudes. Taken together the findings show that there are two levels of potential information loss when reporting interests in joint ventures and associates. The first occurs when joint venture and associate earnings are aggregated in equity earnings. These results suggest that in reporting regimes where joint venture and associate earnings are disclosed only in aggregate, shareholders’ incorrectly place the same forecasting and valuation multiples on joint venture and associate earnings. The second type of loss occurs when joint venture revenues and expenses are aggregated to give joint venture earnings. The results suggest that failure to clearly delineate joint venture revenues and expenses masks information that is forecasting and valuation relevant.23

6. SUMMARY AND CONCLUDING REMARKS This study provides evidence that aggregating joint venture and associate earning components results in loss of forecasting and valuation relevant information. This is consistent with prior research that shows that disaggregating earnings into its components increases its information content. Because the study suggests that the nature and characteristics of joint ventures and

23

For all the regression analyses the estimation pools available firm-year observations over time. This potentially overstates significance levels. One solution is to estimate regressions separately for each year. However, the number of available observations in any given year is too small to effectively estimate the equations. Instead, the regression analysis used permits specifying that observations are independent across groups, but not necessarily independent within groups, the groups here being sample firms. The procedure estimates the model by Ordinary Least Squares but uses the linearization/Huber/White/sandwich (robust) estimates of variance (and thus standard errors). These variance estimates are robust in the sense of providing correct coverage rates to much more than panel-level heteroscedasticity. In particular, they are robust to any type of correlation within the observations of each group.

21

associates are different, it predicts that the coefficients on joint venture and associate earnings differ. It also predicts that joint venture revenues and expenses are significantly different from each other, and that aggregating these components results in information loss. The tests performed on the Canadian, UK, and US samples show that equity earnings differ from investor operating earnings for equity valuation, and that equity investment amounts differ from investor equity book value amounts for forecasting and valuation. The evidence indicates that information is lost when data is aggregated. Specifically, the results indicate that joint venture and associate earnings are significantly different from each other for forecasting and valuation for the UK sample and for valuation for the Canadian sample. It also shows that joint venture revenues and expenses are significantly differently from each other for both samples, and aggregating them into joint venture earnings results in loss of information. The study concludes that failure to separately report joint venture and associate earnings and investment components, and joint venture revenues and expenses, results in loss of information that could be used to predict future earnings and estimate share prices. The evidence suggests that accounting regimes like the US, that do not require more detailed accounting information on joint venture and associate investments, mask information that is potentially forecasting and valuation relevant. This could result in shareholders incorrectly placing the same forecasting and valuation multiples on joint venture and associate accounting amounts. In contrast, in regimes like Canada and UK which require disclosure of disaggregate information, shareholders place different forecasting and valuation multiples on the different investment components. By investigating its research questions, this comprehensive study uses data from

22

three countries to provide evidence that is timely and of interest to accounting standard setters, academics, and practitioners around the world.

23

REFERENCES American Institute of Certified Public Accountants (AICPA), (1979), Joint Venture Accounting, Issues Paper. New York, NY: AICPA. Accounting Principles Board U.S. (APB), (1971), The Equity Method of Accounting for Investments in Common Stock, Accounting Principles Board Opinion No. 18, New York, NY: AICPA. Accounting Standards Board UK (ASB), (1997), Associates and Joint Ventures, Financial Reporting Standard 9, London, UK: ASB. Balakrishnan, R., T. S. Harris and P. K. Sen (1990), ‘The Predictive Ability of Geographic Segment Disclosures’, Journal of Accounting Research, Vol. 28, No. 2 (Autumn), pp. 305-325. Barth, M. E., W. H. Beaver, J. Hand and W. R. Landsman (1999), ‘Accruals, Cash Flows, and Equity Values’, Review of Accounting Studies, Vol. 3, No. 3/4, pp. 205-229. Barth, M. E., W. H. Beaver and W. R. Landsman (2001), ‘The relevance of value relevance literature for financial accounting standard setting: another view’, Journal of Accounting and Economics, Vol. 31, No. 1-3 (September), pp. 77-104. Bernard, V. and T. L. Stober (1989), ‘The Nature and Amount of Information Reflected in Cash Flows and Accruals’, The Accounting Review, Vol. 64, No. 4 (October), pp. 723-747. Bierman Jr., H. (1992), ‘Proportional consolidation and financial analysis’, Accounting Horizons, Vol. 6, No. 4 (December), pp. 5-17. Canadian Institute of Chartered Accountants (CICA), (1994), Long-Term Investments, Canadian Institute of Chartered Accountants 3050, Toronto, ON. Canadian Institute of Chartered Accountants (CICA), (1996), Interests in Joint Ventures, Canadian Institute of Chartered Accountants 3035, Toronto, ON. Dechow, P. M. (1994), ‘Accounting Earnings and Cash Flows as Measures of Firm Performance: The Role of Accounting Accruals’, Journal of Accounting and Economics, Vol. 18, No. 1 (July), pp. 3-42. Dieter, R., and A. R. Wyatt (1978), ‘The Expanded Equity Method - An Alternative in Accounting for Investments in Joint Ventures’, The Journal of Accountancy (June), pp. 89-94. Flower, J., and G. Ebbers (2002), Global Financial Reporting (Palgrave Macmillan, New York, NY).

24

Graham, R. C., R. D. King and C. K. J. Morrill (2003), ‘Proportionate Consolidation vs. the Equity Method: A Decision Usefulness Perspective on Reporting Interests in Joint Ventures’, Accounting Horizons, Vol. 17, No. 2 (June), pp. 123-137. Harrigan, K. R. (1988), ‘Joint Ventures and Competitive Strategy’, Strategic Management Journal Vol. 9 (February), pp. 141-158. International Accounting Standards Board (IASB), (1988), Investments in Associates, International Accounting Standard No. 28, London, UK. International Accounting Standards Board (IASB), (1990), Financial Reporting of Interests in Joint Ventures, International Accounting Standard No. 31, London, UK. Kogut, B. (1988), ‘Joint ventures: theoretical and empirical perspectives’, Strategic Management Journal, Vol. 9, No. 4, pp. 319-332. Kogut, B. (1991), ‘Joint Ventures and the Option to Expand and Acquire’, Management Science, Vol. 37 (January), pp. 19-33. Koh, J. and J. Venkatraman (1991), ‘Joint Venture Formations and Stock Market Reactions: An Assessment in the Information Technology Sector’, Academy of Management Journal, Vol. 34 No. 4, pp. 869-893. Kothavala, K. (2003), ‘Proportional consolidation versus the equity method: A risk measurement perspective on reporting interests in joint ventures’, Journal of Accounting & Public Policy, Vol. 22, No. 6 (November), pp. 517-539. Lipe, R. (1986), ‘The Information Contained in the Components of Earnings’, Journal of Accounting Research, Vol. 24 (Supplement), pp. 37-64. Livnat, J. and P. Zarowin (1990), ‘The incremental information content of cash-flow components’, Journal of Accounting and Economics, Vol. 13, No. 1 (May), pp. 25-46. Madhavan, R. and J. E. Prescott (1995), ‘Market value impact of joint ventures: The effect of industry information-processing load’, Academy of Management Journal, Vol. 38, No. 3, pp. 900-915. Maines, L. A., R. D. Mautz Jr., G. B. Wright, J. A. Yardley, A. J. Rosman and L. E. Graham (2000), Implications of International Differences in Joint Venture Financial Reporting Standards for Financial Analysts’ Stock Values. Working Paper. Milburn, A. and P. Chant (1999), ‘Reporting Interests in Joint Ventures and Similar Arrangements’, Special Report 201-E. Financial Accounting Standards Board (FASB), Norwalk, CT.

25

Nobes, C. and R. Parker (2002), Comparative International Accounting (7th edition. Financial Times, Prentice Hall). Ohlson, J. (1995), ‘Earnings, book values, and dividends in equity valuation’, Contemporary Accounting Research, Vol. 11, No. 2 (Spring), pp. 661-687. Ohlson, J. (1999), ‘On Transitory Earnings’, Review of Accounting Studies, Vol. 4, No. 3/4, pp. 145-162. Park, S. and D. Kim (1997), ‘Market Valuation of Joint Ventures: Characteristics and Wealth Gains’, Journal of Business Venturing, Vol. 12, pp. 18-108. Park, S. and M. V. Russo (1996), ‘When Competition Eclipses Cooperation: An Event History Analysis of Joint Venture Failure’, Management Science, Vol. 42, No. 6, pp. 875-891. Roberts, C., P. Weetman and P. Gordon (1998), International Financial Accounting: A Comparative Approach (Financial Times, Pitman Publishing). Subramanyam, K. R. and J. J. Wild (1996), ‘Going-Concern Status, Earnings Persistence and Informativeness of Earnings’, Contemporary Accounting Research, Vol. 13, No. 1 (Spring), pp. 251-273. Swaminathan, S., and J. Weintrop (1991), ‘The Information Content of Earnings, Revenues, and Expenses’, Journal of Accounting Research, Vol. 29, No. 2 (Autumn), pp. 418-427. Williamson, C. (2000), ‘Many “marriages” fail’, Pensions and Investments, Vol. 28, No. 8, p. 8. Wilson, G. P. (1986), ‘The Relative Information Content of Accruals and Cash Flows: Combined Evidence at the Earnings Announcement and Annual Report Date’, Journal of Accounting Research, Vol. 24 (Supplement), pp. 165-200. Wilson, G. P. (1987), ‘The Incremental Information Content of the Accruals and Funds Components of Earnings after Controlling for Earnings’, The Accounting Review, Vol. 62, No. 2 (April), pp. 293-322.

26

TABLE 1 Descriptive Statistics and Correlation Coefficients Panel A: Descriptive Statistics Canada (n = 121) Variable P

Mean 15.88

UK (n = 246)

Median Std. Dev. 11.33 16.92

US (n=6,176)

Mean 4.43

Median 3.44

Std. Dev. 3.88

Mean 24.94

Median 18.25

Std. Dev. 24.74

NI

1.33

0.74

2.59

0.25

0.20

0.36

4.67

0.87

29.96

OIBefEQ

2.33

1.24

3.19

0.43

0.35

0.37

27.39

1.60

230.11

OINO

–0.96

–0.65

1.42

–0.17

–0.14

0.30

–1.15

–0.72

2.13

REV

26.32

20.25

28.18

4.91

2.99

5.83

EXP

–24.15

–16.86

25.95

–4.49

–2.46

5.71

OIJV

0.21

0.09

0.36

0.04

0.02

0.05

OIAS

0.04

0.02

0.14

0.02

0.01

0.04

OIEQ

0.24

0.12

0.41

0.05

0.03

0.09

0.50

0.03

3.44

12.09

7.93

17.11

1.90

1.44

1.69

162.04

7.97

1,415.00

BVJV

1.77

1.44

1.56

0.14

0.07

0.20

BVAS

1.08

0.44

1.55

0.10

0.04

0.17

BVEQ

2.92

2.15

2.96

0.24

0.13

0.33

15.75

0.70

123.62

REVJV

3.74

2.23

4.59

0.31

0.09

0.59

EXPJV

–3.52

–2.01

4.40

–0.28

–0.08

0.55

BVBefEQ

27

TABLE 1 (Continued) Descriptive Statistics and Correlation Coefficients Panel B: Pearson Correlation Coefficients for the Canadian Sample (UK Sample) in Upper (Lower) Triangle P P

NI

OIBefEQ

OINO

REV

EXP

OIJV

OIAS

OIEQ

BVBefEQ

BVJV

BVAS

BVEQ

REVJV

EXPJV

0.69

0.67

-0.27

0.59

-0.57

0.22

0.32

0.29

0.72

0.34

0.22

0.31

0.24

-0.23

0.87

-0.17

0.70

-0.67

0.06

0.30

0.14

0.81

0.17

0.34

0.33

0.16

-0.17

-0.60

0.81

-0.78 -0.01

0.25

0.07

0.86

0.25

0.35

0.35

0.14

-0.14

-0.51

0.48

0.09

-0.06

0.05

-0.42

-0.25

-0.17

-0.20

-0.06

0.06

-0.99

0.13

0.31

0.23

0.73

0.26

0.27

0.29

0.38

-0.39

-0.14 -0.32

-0.25

-0.70

-0.25

-0.25

-0.27

-0.41

0.41

0.35

0.93

0.15

0.13

-0.10

-0.05

0.69

-0.63

0.60

0.33

0.02

0.10

0.01

0.42

-0.41

0.26

0.09

-0.11

-0.09

0.69

-0.64

0.18

0.13

0.16

0.23

-0.22

0.36

0.73

0.24

-0.25

0.88

0.17

-0.19

0.20

-0.22

NI

0.55

OIBefEQ

0.63

0.54

OINO

-0.11

0.38

-0.52

REV

0.22

0.23

0.42

-0.22

EXP

-0.19

-0.20

-0.36

0.19

-0.99

OIJV

0.35

0.32

0.32

0.02

0.33

-0.32

OIAS

0.11

0.10

0.10

-0.01

0.30

-0.30

0.33

OIEQ

0.32

0.31

0.25

0.07

0.37

-0.36

0.87

0.73

BVBefEQ

0.51

0.51

0.73

-0.26

0.30

-0.25

0.29

0.19

0.31

BVJV

0.24

0.24

0.27

-0.03

0.25

-0.23

0.59

0.47

0.64

0.32

BVAS

0.18

0.11

0.14

-0.02

0.32

-0.31

0.35

0.80

0.65

0.19

0.49

BVEQ

0.24

0.22

0.24

-0.03

0.31

-0.31

0.55

0.70

0.73

0.30

0.89

0.81

REVJV

0.14

0.13

0.17

-0.04

0.48

-0.48

0.56

0.27

0.51

0.14

0.57

0.29

0.53

EXPJV

-0.11

-0.10

-0.15

0.06

-0.47

0.48

-0.48 -0.25

-0.44

-0.11

-0.54

-0.28

-0.51

28

-0.99 -0.99

TABLE 1 (Continued) Descriptive Statistics and Correlation Coefficients Panel C: Pearson Correlation Coefficients for the US Sample P P NI

NI

OIBefEQ

OINonOp

OIEQ

BVBefEQ

BVEQ

0.35

0.11

-0.24

0.19

0.09

0.14

0.87

-0.16

0.79

0.81

0.84

-0.23

0.78

0.93

0.93

-0.07

-0.20

-0.19

0.82

0.84

OIBefEQ OINO OIEQ

0.94

BVBefEQ BVEQ

Notes: P is price per share at fiscal year end. NI is net income of the investor firm. OIBefEQ is operating earnings of the investor firm minus joint venture earnings and associate earnings. OINO is non-operating earnings of the investor firm. REV is revenues of the investor firm. EXP is operating expenses of the investor firm. OIJV is joint venture earnings. OIAS is associate earnings. OIEQ is the sum of joint venture and associate earnings. BVBefEQ is book value of equity at fiscal year end minus net investments in joint ventures and associates. BVJV is net investment in joint ventures. BVAS is net investment in associates. BVEQ is sum of net investments in joint ventures and associates. REVJV is joint venture revenues. EXPJV is joint venture expenses. n is number of firm-year observations. NI, OIBefEQ, OINO, REV, EXP, OIJV, OIAS, OIEQ, BVBefEQ, BVJV, BVAS, BVEQ, REVJV, and EXPJV are deflated by number of shares outstanding at fiscal year end. Correlation coefficients significantly different from zero at p-values less than 5% are in boldface type.

29

TABLE 2 Summary Statistics from Regression of Future Earnings and Share Prices on Operating Earnings, Non-Operating Earnings, Total Equity Earnings, and their Corresponding Equity Book Value Components NIt+1 = ω0+ ω1 OIBefEQt + ω2 OINOt + ω34 OIEQt + ω5 BVBefEQt + ω67 BVEQt + ε3t+1 Pt = α0 + α1 OIBefEQt + α2 OINOt + α34 OIEQt + α5 BVBefEQt + α67 BVEQt + ε4t Canada

US

UK

NIt+1

P

NI t+1

P

NI t+1

P

Predicted Sign +

Coefficient (t-statistic) 0.79 (13.78)






OINO

─

─0.35 (─2.69)

─1.41 (─1.65)

─0.26 (─2.50)

─3.57 ─ (2.56)

─0.82 (─6.90)

─2.65 (─15.88)

OIEQ

+

0.58 (1.62)

3.25 (1.85)

0.90 (2.13)

4.38 (1.69)

0.60 (19.10)

4.03 (11.44)

BVBefEQ

+

─0.01 (─1.18)

0.62 (2.79)

─0.01 (─0.71)

0.15 (3.68)

─0.01 (─3.25)

0.25 (14.13)

BVEQ

+

0.17 (2.68)

1.26 (3.59)

0.06 (1.89)

0.05 (3.10)

0.06 (9.49)

0.36 (5.13)

N Adj R2

91 0.75

124 0.64

162 0.39

248 0.51

4,514 0.68

6,176 0.33

Tests

p-value

p-value

p-value

p-value

p-value

p-value

ω1 = ω34, α1 = α34

0.14

0.04

0.15

0.01

0.01

0.01

ω5 = ω67, α5 = α67

0.01

0.05

0.05

0.01

0.01

0.01

Variable OIBefEQ

30

TABLE 2 (Continued) Summary Statistics from Regression of Future Earnings and Share Prices on Operating Earnings, Non-Operating Earnings, Total Equity Earnings, and their Corresponding Equity Book Value Components Notes: P is price per share at fiscal year end. NI is net income of the investor firm. OIBefEQ is operating earnings of the investor firm minus joint venture earnings and associate earnings. OINO is non-operating earnings of the investor firm. OIEQ is the sum of joint venture and associate earnings. BVBefEQ is book value of equity at fiscal year end minus net investments in joint ventures and associates. BVEQ is sum of net investments in joint ventures and associates. NI, OIBefEQ, OINO, OIEQ, BVBefEQ, and BVEQ are deflated by number of shares outstanding at fiscal year end. N is number of firm-year observations. t-statistics are based on White (1980) heteroscedasticity-consistent standard errors.

31

TABLE 3 Summary Statistics from Regression of Future Earnings and Share Prices on Operating Earnings, Non-Operating Earnings, Joint Venture Earnings, Associate Earnings, and their Corresponding Equity Book Value Components NIt+1 = ω0+ω1OIBefEQt +ω2OINOt +ω3OIJVt+ω4OIASt+ω5BVBefEQt+ω6BVJVt+ω7BVASt+ ε5t+1 Pt = α0 + α1OIBefEQt+ α2OINOt + α3OIJVt +α4OIASt +α5BVBefEQt + α6BVJVt + α7BVASt + ε6t Canada

UK

NIt+1

Pt

NI t+1

Pt

Predicted Sign +





OINO

─

–0.21 (–2.61)

–2.64 (–1.73)

–0.19 (–3.42)

–3.45 (–3.78)

OIJV

+

0.46 (2.19)

2.91 (1.71)

0.59 (3.31)

7.99 (1.70)

OIAS

+

1.13 (2.32)

–7.45 (–1.82)

1.32 (4.39)

–4.00 (–1.62)

BVBefEQ

+

–0.01 (–0.92)

1.42 (13.41)

–0.01 (–0.17)

1.15 (1.69)

BVJV

+

0.26 (1.38)

1.70 (3.55)

0.11 (2.09)

1.16 (2.53)

BVAS

+

0.33 (3.05)

–0.56 (–0.93)

0.02 (0.19)

1.20 (1.98)

89

124

161

248

Adj R2

0.75

0.79

0.50

0.55

Tests

p-value

p-value

p-value

p-value

ω3 = ω4, α3 = α4

0.10

0.00

0.02

0.05

ω1 = ω3, α1 = α3

0.24

0.05

0.36

0.03

ω1 = ω4, α1 = α4

0.33

0.00

0.01

0.01

ω6 = ω7, α6 = α7

0.00

0.05

0.07

0.04

ω5 = ω6, α5 = α6

0.10

0.01

0.03

0.01

ω5 = ω7, α5 = α7

0.01

0.10

0.12

0.01

Variable OIBefEQ

N

Income

Book value of equity

32

TABLE 3 (Continued) Summary Statistics from Regression of Future Earnings and Share Prices on Operating Earnings, Non-Operating Earnings, Joint Venture Earnings, Associate Earnings, and their Corresponding Equity Book Value Components Notes: P is price per share at fiscal year end. NI is net income of the investor firm. OIBefEQ is operating earnings of the investor firm minus joint venture earnings and associate earnings. OINO is non-operating earnings of the investor firm. OIJV is joint venture earnings. OIAS is associate earnings. BVBefEQ is book value of equity at fiscal year end minus net investments in joint ventures and associates. BVJV is net investment in joint ventures. BVAS is net investment in associates. NI, OIBefEQ, OINO, OIJV, OIAS, BVBefEQ, BVJV, and BVAS, are deflated by number of shares outstanding at fiscal year end. N is number of firm-year observations. t-statistics are based on White (1980) heteroscedasticity-consistent standard errors.

33

TABLE 4 Summary Statistics from Regression of Future Earnings and Share Prices on Investor Revenues, Operating Expenses, Non-Operating Earnings, Joint Venture Revenues, Joint Venture Expenses, Associate Earnings, and their Equity Book Value Components NIt+1=µ0+µ1REVt+ µ2EXPt+ µ3OINOt+µ4REVJVt+µ5EXPJVt+ µ6OIASt+µ7BVBefEQt+µ8BVJVt+µ9BVASt+ε7t+1 Pt = δ0+ δ1REVt+ δ2EXPt+ δ3OINOt+ δ4REVJVt+ δ5EXPJVt+ δ6OIASt+ δ7BVBefEQt+ δ8BVJVt+ δ9BVASt+ε8t

UK

Canada NIt+1

Pt

NI t+1

Pt

Predicted Sign +





EXP

─

─0.87 (─14.72)

─2.87 (─1.72)

─0.61 (─6.06)

─5.07 (─4.00)

OINonOp

─

─0.24 (─2.88)

─1.48 (─1.70)

─0.27 (─2.82)

─2.31 (─2.28)

REVJV

+

0.74 (2.10)

4.25 (1.94)

0.93 (2.33)

7.55 (1.71)

EXPJV

─

─0.78 (─2.13)

─4.75 (─1.99)

─0.97 (─2.31)

─7.41 (─1.69)

OIAS

+

1.01 (2.53)

7.17 (1.50)

1.14 (3.45)

─6.14 (─0.72)

BVBefEQ

+

0.01 (0.26)

0.71 (4.12)

─0.01 (─0.80)

0.36 (1.69)

BVJV

+

─0.10 (─1.25)

2.02 (4.19)

─0.12 (─0.75)

2.75 (5.48)

BVAS

+

0.51 (4.05)

0.91 (1.34)

0.12 (0.05)

2.79 (2.81)

89

124

161

248

Adj R2

0.71

0.63

0.35

0.49

Tests

p-value

p-value

p-value

p-value

µ4 = ─µ5, δ4 = ─δ5

0.02

0.05

0.00

0.00

µ1 = µ4, δ 1 = δ 4

0.36

0.04

0.05

0.03

µ2 = µ5, δ 2 = δ 5

0.25

0.04

0.05

0.03

Variable REV

N

34

TABLE 4 (Continued) Summary Statistics from Regression of Future Earnings and Share Prices on Investor Revenues, Operating Expenses, Non-Operating Earnings, Joint Venture Revenues, Joint Venture Expenses, Associate Earnings, and their Equity Book Value Components Notes: P is price per share at fiscal year end. NI is net income of the investor firm. REV is revenues of the investor firm minus joint venture revenues. EXP is operating expenses of the investor firm minus joint venture expenses. OINO is non-operating earnings of the investor firm. REVJV is joint venture revenues. EXPJV is joint venture expenses. n is number of firm-year observations. OIAS is associate earnings. BVBefEQ is book value of equity at fiscal year end minus net investments in joint ventures and associates. BVJV is net investment in joint ventures. BVAS is net investment in associates. NI, REV, EXP, OINO, REVJV, EXPJV, OIAS, BVBefEQ, BVJV, and BVAS are deflated by number of shares outstanding at fiscal year end. N is number of firm-year observations. t-statistics are based on White (1980) heteroscedasticity-consistent standard errors.

35