Decline in Financial Reporting for Joint Ventures? Canadian Evidence on Removal of Financial Reporting Choice

Abstract

In 1995, the Canadian Institute of Chartered Accountants (CICA) changed its Generally Accepted Accounting Principles on accounting for joint ventures from permitting a choice between the equity method (EM) and proportionate consolidation (PC) to requiring only PC. More recently, the International Accounting Standards Board (IASB) has decided to issue a new standard that will eliminate choice between EM and PC and require only EM; but as of October 2010 a new standard was still to be issued. The past Canadian and proposed IASB changes are similar in that choice between the same two reporting methods is removed but differ in the required alternative, PC for Canada and EM for the IASB. In this paper we use a sample of Canadian companies over the period 1985–2003 to study financial reporting for joint ventures. To our knowledge, our Canadian sample is the only one reflecting a reduction of choice in financial reporting methods for joint ventures. Therefore, our results have particular relevance for evaluating the IASB's proposed change. Specifically, we investigate whether firms that use EM between 1985 and 1994 experience a decline in value relevance of key balance sheet amounts such as total assets and liabilities when forced to use PC from 1995 onwards. Since 1995 firms are also required to provide footnote disclosures on their share of joint venture assets and liabilities in addition to revenues, expenses and cash flows. Using these disclosures, we investigate whether disaggregate joint venture assets and liabilities are incrementally and overall value relevant. We find that firms that are forced to switch from EM to PC experience a decline in value relevance of reported assets and liabilities. The firms that use PC for the entire sample period experience no such decline. We also find that joint venture assets and liabilities are incrementally and overall value relevant when disclosures are mandatory from 1995 onwards. Our results show that the removal of choice of financial reporting method does have value-relevance implications, something that is of importance to users. We also find that the requirement of additional disclosure of joint venture assets and liabilities is value relevant, which may offset, to some extent, the costs of the reduction in choice. Our inferences may have implications for a number of jurisdictions across Europe and beyond that are affected by a similar reduction of accounting choice proposed by the IASB.

Acknowledgement

We thank the Society of Management Accountants of Ontario for financial support of this project. We are grateful to David Moore, CA of the Canadian Institute of Chartered Accountants, who provided us with the names of companies listed as reporting joint ventures in 1994 in Financial Reporting in Canada, and Dr C. Morrill of the University of Manitoba, who shared the list of companies used in his research study. We thank the editor Salvador Carmona, an anonymous referee, Vasiliki Athanasakou, Mary Barth, Al Bhimani, Stefano Cascino, Peter Miller, John O'Hanlon and Martin Walker for helpful comments.

Notes

All EU listed firms were required to adopt IFRS in 2005, so IAS 31 automatically became the standard for listed UK firms from 2005 onwards.

Under the gross equity method disaggregate shares of joint venture revenues, earnings, assets and liabilities are disclosed on the face of the financial statement in addition to equity method recognition.

For example, Canada is in the process of adopting IFRS in full by 2011 (CICA, 2007), and in 2007, the Australian Accounting Standards Board modified their accounting standards so that they are identical to IFRS as issued by the IASB.

Joint ventures are jointly controlled by co-venturer firms where joint control is the sharing of power between venturer firms. No single firm unilaterally controls the joint ventures, and joint consent of all venturers is required for major operating and financing policy decisions.

The cost method, which reports only dividends received as income, was recommended in cases where the equity from the joint venture was not likely to accrue to the co-venturer or there had been a loss of joint control.

Revised CICA 3055 defines a joint venture as ‘an economic activity resulting from a contractual arrangement whereby two or more venturers jointly control the economic activity’ (CICA 3055.03).

Comment letter 86 from the AcSB to the IASB, 11 January 2008.

The independent variables are deflated by number of shares outstanding.

Only one firm changed between EM and PC. This firm changed from EM to PC one year and then back the following year. Our results are substantively the same whether this observation is included or excluded.

It is beyond the scope of this paper to study explicitly an accounting choice or its determinants, or to investigate signalling implications of an accounting choice.

David Moore, CA of the Canadian Institute of Chartered Accountants supplied the names of these companies to us. We checked the completeness of our sample by comparing it to those used by Graham et al. (2003) and Kothavala (2003), who kindly provided lists of companies in their samples.

Rather than trying to investigate how value relevance for firms in general changes over time our study examines how value relevance changes for the same firms over time.

Because the original Section 3055 allowed a choice between EM and PC, we need not be concerned with the issue of early adopters. This conclusion is reinforced by the fact that there was a relatively short time between the release of the exposure draft (June 1993) and finalisation of the revised Section 3055 (October 1994).

The proportion of firm-years in each time period is close to the proportion of number of years in each 10-year time period: 50.3% for 1985–94 and 49.7% for 1995–2003.

These pre-1995 amounts are not used in the regression analysis and are hence untabulated. We believe the pre-1995 set of firms disclosing share of joint venture assets and liabilities is too small to perform any rigorous and conclusive analysis.

Untabulated statistics show that for each year in 1985–94, when choice was permitted, PC was by far the method of choice. On average PC was chosen 68.7% of the time compared to 31% for EM. Untabulated statistical tests show that the PC to EM distribution rejects the assumption of a 50:50 distribution at the 1% level for each of the 10 years in the sample.

For example, suppose a company's primary business is low risk but it has large joint venture investments in high risk industries, and the net joint venture investment is small. Under EM the assets and liabilities would be primarily those of its core business whereas under PC these assets and liabilities would be a mix of low and high risk, core and joint investments. In this case the PC coefficients would likely be lower than EM coefficients because of the differences in risk. For companies that used PC in both periods there would be no expectation of decline in coefficients. In this sort of scenario the decline in coefficients for EM firms would not be due to inappropriateness of accounting method. We do not have detailed information on the type and nature of joint venture investments because when disclosing share of joint venture assets and liabilities, or providing any information about joint ventures, companies tend to refer to all their joint venture investments collectively. However, anecdotally, we see that joint ventures tend to be in similar or related industries. For example, a mining company would have an extractive joint venture, and a property company might have joint ventures in a chain of shopping complexes.

Investigating each of these possibilities is beyond the scope of this study. However, they stand as reasonable explanations for the decline in value relevance of accounting amounts experienced by firms that chose EM in the former period.

In further specifications, O'Hanlon and Taylor (2007) distinguish between associate and joint venture liability disclosures incremental to assets and liabilities reported under EM. These additional aspects are beyond the scope of our study.

Bauman (2003) includes indicator variables for investor guaranteed and non-guaranteed debt. However, this is beyond the scope of our data and analysis.

Additional information

Notes on contributors

A. William Richardson

Paper accepted by Salvador Carmona.

Raafat R. Roubi

Paper accepted by Salvador Carmona.

Kazbi Soonawalla

Paper accepted by Salvador Carmona.