https://orcid.org/0009-0001-8739-4596
![Sample our Area Studies journals, sign in here to start your access, latest two full volumes FREE to you for 14 days]({"type":"image" , "src" : "/sda/5922/TOC-Subject-Sample-2021-Area-Studies.jpg"})
ABSTRACT
This article studies the effects of IFRS 11 on the value relevance of accounting numbers (VRAN) in France. In 2014, IFRS 11 made the equity method (EM) mandatory to account for joint ventures (JVs) and disallowed proportionate consolidation, the method previously preferred by French groups. Panel method regressions are used to examine the evolution of value relevance in listed groups’ financial statements over a long period (2007–2020). Generalization of the EM reallocates the VRAN, and post-IFRS 11 EM-related numbers are significantly and negatively linked to market value, raising questions about their faithfulness. These results concern all groups using the EM, whatever method they previously used for JVs. This study also looks at the standard-setters’ proposed integral/non-integral classification of net income from JVs and associates, which is found to be non-value relevant. These results have implications at standard-setting level for improving the quality of financial reporting, and for investors.
Acknowledgements
We thank the editors, Andrei Filip and Anne Jeny, as well as two anonymous reviewers for their insightful comments and suggestions. We also thank Ann Gallon for language editing and assistance. We are grateful to the ANC for its grant, to its members and staff for valuable comments, and to the practitioners we have met. Any views and opinions expressed herein are those of the authors and do not necessarily reflect the official policy or position of the ANC. We also thank Pascal Barneto, Serge Evraert, Michel Legain, Charles Signorini, Sabrina Texandier, and other colleagues for their reviews, comments and suggestions.
Disclosure Statement
No potential conflict of interest was reported by the author(s).
Notes
1 IAS 31, Interests in joint ventures.
2 Recommendation n° 2013–01 issued by the French Accounting Standards Authority (ANC), and later, a proposal by the IASB in its Exposure Draft ED 2019/7.
3 54.5% in our panel and 87% in Sarquis et al.’s (Citation2022) sample of 105 groups.
4 For example, the Safran, JCDecaux, and Véolia groups have experienced major modifications in their consolidated financial statements.
5 This was confirmed by Lopes and Lopes (Citation2019).
6 Which defines two main characteristics of useful accounting information, relevance and faithful representation, and four secondary characteristics: understandability, timeliness, verifiability, and comparability.
7 Conversely, in French standards applicable to companies that are not listed on a regulated market (ANC regulation 2020-01), the EM is mandatory for associates but remains optional for joint ventures.
8 This proportion remains steady from 2014 to 2020.
9 In the sample in 2020: Vivendi, Bollore, Holcim, Saint Gobain, Assystem, Sodexo.
10 https://www.ifrs.org/news-and-events/updates/iasb/2021/iasb-update-october-2021/ and later in the Project Report and Feedback Statement (IASB, Citation2022a).
12 However, the IASB is still working on EM through its project to reform IAS 28[12], and the French standard-setter, too, is still interested in the subject.
13 In France, listed groups have two independent auditors.
14 i.e. accounts for 2013 established under 2014 standards, including the new standard IFRS 11, for publication in 2014 annual reports.
15 Barth et al. (Citation2023) illustrate this topic with the example of Lev and Sougiannis (Citation1996) for R&D: under US GAAP, R&D outlays are charged to expenses, and yet, like capital expenditure for intangible assets, they are positively associated with future operating earnings. This means that ‘R&D expense is not a faithful representation of the value of R&D activities, which implies its VR derives from its relevance’.
16 The Vuong test is based on the likelihood ratio and uses the Kullback-Leibler information criterion (Kullback & Leibler, Citation1951).
17 This is the case between models with and without a breakdown, but not between models with and without interaction variables.
18 Keeping all the SIC codes instead of the dummy business sector variable was tested, and gave similar results.
19 Running regressions with or without these interactions does not modify the results.
20 EXPC in interaction with other variables is used in the sub-period regression. See Section 5.3.
21 0.56 for the whole period (), rising to 0.7 for the sub-period 2014–2020 ().
22 In the models used, the ‘Debt’ variable (D) is the ‘total net financial Debt’ rather than the ‘total debt and liabilities’ as used by Richardson et al. (Citation2012) and Gavana et al. (Citation2020), because of the strong correlations observed between the ‘total Assets’ (A) and ‘total financial debt’ and between ‘A’ and ‘total debt and liabilities’ (respective coefficients of 0.80*** and 0.95***). In the models presented, total net financial debt and the asset variables (A or ALEM) showed moderate (see and ) but highly significant correlation coefficients.
23 There is no collinearity between ‘A’ and ‘LnA’ because they capture two different dimensions: ‘A’ measures total assets per share. It is a relative measure. LnA, as a control variable, is a proxy for size and is not a relative measure. and show weak correlation coefficients (0.35 and 0.37). We run the model without LnA: global results do not change significantly.
24 Adding all the products between EXPC and all the variables in model 1 does not change these results, as all the corresponding coefficients are non-significant.
25 As Gavana et al. (Citation2020, p. 10) points out, the rule change has modified the contribution of each accounting variable to the VR.
26 Comparisons cannot be tested because the Vuong test cannot be applied to a regression performed on different periods.