Project topics and materials
Get Complete Project Material File(s) Now! »
Research related to the recognition and measurement of investments in associates
An early paper investigating equity accounting is that of Comiskey and Mulford (1986:523). They find that entities deliberately avoided acquiring an interest greater than 20 per cent if an investee appeared likely to report losses. They suggest that the findings imply that equity accounting requirements have real economic consequences (1986:525). The findings of Comiskey and Mulford (1986) are not entirely surprising when considering the history behind the threshold in IAS 28 (IASB, 2003). Nobes (2002:28) notes that the 20 per cent threshold for equity accounting appears to have as its basis a United Kingdom tax rule, whereby tax losses could be shared between the members of the investing consortium if the company had five or fewer shareholders. He suggests that the inclusion of the 20 per cent threshold in accounting standards at the time represented the start of a tradition of standard-setters accommodating companies’ wishes (2002:29). The main implication of Nobes’ conclusion is that the 20 per cent threshold has no theoretical basis to suggest that a shareholding of 20 per cent results in a significant influence over either legal requirements relating to the decision-making processes of firms or economic relationships.
However, the results of Comiskey and Mulford (1986) should also be considered within the relevant historical context. The accounting measurement requirement for financial assets during the sample years of Comiskey and Mulford (1986:520) was to carry these investments at cost. With changes to accounting standards, financial assets are now generally required to be carried at fair value (Power, 2010:197). As a result, the incremental cost of applying equity accounting may have decreased in the ensuing years and so too the benefits of avoiding its application, as losses of associates are likely to be reflected in decreased fair values for such investments. Indeed Nobes (2002:41) suggests fair value accounting investments in associates as an alternative to the equity method, as fair value accounting does not rely on an arbitrary threshold and is a “more honest” valuation approach. This comment by Nobes (2002) reflects a perception of the equity method as an alternative valuation method, as opposed to a simplified form of consolidation. In fact, the equity method has indeed been viewed as a valuation method in some countries, such as The Netherlands (Nobes, 2002:21).
Perhaps because the equity method is sometimes viewed as a valuation of the investment, prior research suggests that the equity method provides value-relevant information to equity investors. Soonawalla (2006:411) finds, for example, that equity accounted carrying amounts of both associates and joint ventures are value-relevant in Canada and the United Kingdom. The study shows specifically that disaggregation of investments in associates and joint ventures provides value-relevant information to equity investors. In contrast Soonawalla (2006:409) finds that the equity accounted income of joint ventures is value-relevant, while that of associates is not. One implication of these findings is therefore that changes in the equity accounted carrying amount of joint ventures appear to be a timelier reflection of changes in the value of the investment than those of associates.
The findings of Soonawalla (2006) were in contrast to prior research by Graham, King and Morrill (2003a:135), who found that proportionate consolidation of joint ventures forecast accounting return on equity more accurately than equity accounted results of joint ventures do. Although these findings relate to joint ventures, they do suggest that the equity method may not be the optimal accounting treatment for significant investments. Interestingly, Richardson, Roubi and Soonawalla (2012:390) subsequently find that when Canada decided to remove the equity method as an alternative for measuring investments in joint ventures in 1995, the firms forced to switch over to proportionate consolidation suffered a decline in value-relevance in certain balance sheet amounts (such as total assets). This would suggest that equity accounting results in information with greater decision-usefulness than that provided in terms of proportional consolidation. To summarise, prior research findings thus suggest that great uncertainty remains around the appropriateness of the equity method as an accounting treatment.
For this reason some prior researchers have empirically considered the use of fair value measurements for investments in associates as an alternative to the equity method. Fair values as a measurement base for investments is grounded in finance theory, whereby the value of any asset is the present value of its expected future cash flows (Ilmanen, 2011:66). Unlike equity accounted carrying amounts, which focus on the net asset value of the associate and historical information, fair values incorporate expectations around future cash flows relating to the investment in the associate. Expected future cash flows are implicitly incorporated into the fair values of listed associates. As the fair values of listed associates are disclosed based on market prices, should market participants take into consideration the future expected cash flows of the associate, these expectations are carried forward to the financial report of the investor. Similarly, fair values of unlisted associates are based on directors’ valuations. Directors’ valuations represent management’s assessment of the value of the investment in the associate. Such a value assessment, finance theory suggests, will represent the present value of expected future cash flows that relate to the investment in the associate. When the reporting entity (i.e. the investor) is valued by market participants, such future cash flow expectations should then be incorporated into the value that they place on the equity of the reporting firm.
In this branch of research, an early study that considers the use of fair values for investments in associates is that of Barth and Clinch (1998). They find that the disclosed fair values of investments in associates were value-relevant only for mining firms. Their sample consisted of Australian firms observed from 1991 to 1995 (1998:217). Barth and Clinch (1998:217) also find that the recognised carrying amounts of investments in associates in their sample are value-relevant, but only for mining and financial firms. However, as equity accounted carrying amounts were only utilised in Australian financial statements from 1998 onwards (Nobes, 2002:26), the Barth and Clinch (1998) study compares fair value measurements with the cost of these investments. A consideration in this respect is that Barth and Clinch (1998:217) utilise the total fair value and carrying amount of investments in associates in their regressions. Therefore, as the study does not utilise a measure of the difference between the fair values and the carrying amounts (cost in this case) of investments in associates, inferences drawn from the total fair value measurements are still relevant, despite changes in accounting requirements. More importantly, because of the changes in accounting requirements to carry investments in associates under the equity method as opposed to cost, it is uncertain how much of the value-relevance of fair values may have been captured by the equity accounted carrying amounts.
A study that gives some insight into this, is that of Graham et al. (2003b:1076) who find that the excess of disclosed fair values over the equity accounted carrying amounts of listed associates is value-relevant for a sample of listed United States firms reporting from 1993 to 1997. Similarly, the study finds that the equity accounted carrying amounts of associates are value-relevant. However, because of differences in sampling methods, sample periods, model specifications and accounting requirements, the findings of the Graham et al. (2003b) and Barth and Clinch (1998) papers discussed above are not directly comparable. This leads to some unanswered questions, discussed in further detail below.
Apart from the fact that the studies were performed in different countries with sample periods that do not directly overlap, the Graham et al. (2003b:1065) paper is focused on investigating investments in associates. To this end, their sample only includes firms where investments in associates comprise more than one per cent of total assets (ibid. 1070). As Barth and Clinch (1998:199) investigate several fair value disclosures, their sample is significantly larger and not targeted to investments in associates. Furthermore, Graham et al.
Research related to the disclosures around equity accounted investments
As discussed earlier, the version of IAS 28 (IASB, 2003) applicable during the sample period sets out several disclosure requirements for investments in associates. One of the most important requirements for the purpose of this study is that fair values must be disclosed for associates where quoted prices are available. The research implications of this requirement were discussed in the preceding subsection. However, other relevant disclosures required by IAS 28 (IASB, 2003) include summarised financial information of the associate, including total assets, total liabilities, revenues and profit or loss. The accounting standards relating to joint ventures required similar disclosures for such investments.
Prior research about the value-relevance of disclosures for equity accounted investments has tended to investigate joint ventures and, in particular, the disclosures related to the total liabilities of the equity accounted investment. Baumann (2003:313) finds, for example, that investor-guaranteed obligations of equity accounted investments are significantly negatively associated with the investor’s market value. In a later study O’Hanlon and Taylor (2007:284) find that the disclosed liabilities of equity accounted investments in the United Kingdom are negatively associated with the investor’s market value and that the relationship is stronger for joint ventures than associates. Richardson et al. (2012:390) confirm the value-relevance of liability disclosures for the equity accounted Canadian joint ventures in their sample. However, IAS 28 (IASB, 2003) also requires disclosures around the total assets, revenues and profit or loss of associates. Prior research has not investigated the value-relevance of these disclosures. Based on the Ohlson (1995:666–667) model, the fair value of an investment in an associate should be derived in some measure from its net assets and profit or loss. This suggests that the disclosures around equity accounted investments in associates may be used as inputs into a valuation model to determine an alternative fair value measurement (i.e. intrinsic value) for the associate. The question that therefore arises is whether or not the disclosure of summarised financial information is necessary where the associate is listed and its fair value is already disclosed. If the summarised financial information is indeed incorporated by investors in an intrinsic value of the investment associate for decision-making purposes, then such summarised financial information would be value-relevant. However, if investors accept that the market value of the associate already incorporates all necessary value-relevant information, the additional summarised financial information disclosures should have no value-relevance of their own. Importantly, prior research investigating the value-relevance of liability disclosures of equity accounted investments does not control for the fair value of the investment.
Therefore, the fair value of equity accounted investments may subsume the disclosed summarised financial information as well as the equity accounted carrying amount. However, the degree to which this occurs may be significantly affected by the nature of the fair value information available in the financial statements. The fair values of listed associates are determined by market forces and are directly verifiable by users of the financial statements. In contrast, the fair values of unlisted associates depend on the discretion of management. As a result, they may have a comparably muted effect on the value-relevance of disclosed summarised financial information, where investors may use this information to rather develop an independent fair value measurement. However, prior research (Song et al., 2010:1376–1377) suggests that investors will still take management’s fair value measurement into account in developing a fair value of their own. Accordingly, this study also investigates the value-relevance of the disclosures of summarised financial information of listed and unlisted associates when the fair value of the associate is controlled for.
1. Introduction
1.1. Introduction
1.2. Research question
1.3. Contribution of the study
1.4. Delimitations
1.5. Summary of the main findings
1.6. Chapter outline
1.7. Summary and conclusion
2. Literature review and hypotheses development
2.1. Introduction
2.2. Current accounting requirements for associates
2.3. Value-relevance research
2.4. General valuation research
2.5. General valuation research and value-relevance research
2.6. Associate accounting research
2.7. Hypotheses
2.8. Summary and conclusion
3. Research methodology
3.1. Introduction
3.2. Research model and instruments
3.3. Summary and conclusion
4. Sample selection
4.1. Introduction
4.2. Sample firms and sample period selected
4.3. Sources of the sample data obtained
4.4. Sample numbers
4.5. Summary and conclusion
5. Detailed findings: fair values of listed associates
5.1. Introduction
5.2. Descriptive statistics
5.3. Univariate investigations
5.4. Detailed multivariate regression findings
5.5. Results of robustness tests
5.6. Summary and conclusion
6. Detailed findings: fair values of unlisted associates
6.1. Introduction
6.2. Descriptive statistics
6.3. Univariate investigations
6.4. Detailed multivariate regression findings
6.5. Results of robustness tests
6.6. Summary and conclusion
7. Detailed findings: value-relevance of listed associates across time
7.1. Introduction
7.2. Descriptive statistics
7.3. Univariate investigations
7.4. Detailed multivariate regression findings
7.5. Results of robustness tests
7.6. Summary and conclusion
8. Detailed findings: value-relevance of unlisted associates across time
8.1. Introduction
8.2. Descriptive statistics
8.3. Univariate investigations
8.4. Detailed multivariate regression findings
8.5. Results of robustness tests
8.6. Summary and conclusion
9. Detailed findings: value-relevance of listed associates between countries
9.1. Introduction
9.2. Descriptive statistics
9.3. Univariate investigations
9.4. Detailed multivariate regression findings
9.5. Results of robustness tests
9.6. Summary and conclusion
10. Detailed findings: value-relevance of unlisted associates between countries
10.1. Introduction
10.2. Descriptive statistics
10.3. Univariate investigations
10.4. Detailed multivariate regression findings
10.5. Results of robustness tests
10.6. Summary and conclusion
11. Detailed findings: other disclosures of listed associates
11.1. Introduction
11.2. Descriptive statistics
11.3. Univariate investigations
11.4. Detailed multivariate regression findings
11.5. Results of robustness tests
11.6. Summary and conclusion
12. Detailed findings: other disclosures of unlisted associates
12.1. Introduction
12.2. Descriptive statistics
12.3. Univariate investigations
12.4. Detailed multivariate regression findings
12.5. Results of robustness tests
12.6. Results of additional analyses
12.7. Summary and conclusion
13. Summary and conclusion
13.1. Introduction
13.2. Summary of the main findings
13.3. Summary and conclusion
14. References
