Is Comprehensive Income Value Relevant and Does Location Matter? A European Study

Abstract

This paper investigates the relative and the incremental value relevance of Comprehensive Income (CI) and Other Comprehensive Income (OCI) across European countries after the mandatory adoption of the International Accounting Standards Board (IASB) standards. This topic, which has already been analysed in other countries, drew the attention of academics and practitioners in Europe after the issuance of International Accounting Standard 1-revised, which requires entities to prepare a statement of comprehensive income (SOCI) in which both CI and OCI components are disclosed. The investigation of the value relevance of accounting amounts is important in order to evaluate their usefulness, because it highlights whether they reflect information investors use in making economic decisions. This study analyses a sample made of all the European listed entities of countries belonging to the EU at the date of issuance of EU Regulation 1606/2002. Our analysis involves a period from 2006 to 2011 and includes 16,511 firm-year observations. Our results show that Net Income (NI) is more value-relevant than CI, even though the total OCI of the period adds relevant information to those already disclosed in other accounting items such as NI and Book Value. In this regard, we found that the coefficient of the total OCI of the period is lower than that of NI, a result to be expected because of its transitory nature. Our findings also suggest that the requirement to issue an SOCI has not produced a significant change in the value relevance of both CI and the total OCI of the period leading to the conclusion that its location does not affect the value relevance of these items. Finally, we found significant differences in the incremental value relevance of the total OCI of the period across European countries, differences that seem to be caused by the countries' characteristics, such as the source of funds (credit/equity and insider/outsider) and the legal systems.

Notes

1 Note that the IAS 1-revised allows for two options when preparing the SOCI: (a) a single statement that shows both the NI and the CI or (b) two separate statements, one that shows NI and the one that, starting from the NI, recognizes the OCI in order to show CI.

2 In this case, we cannot test for the absence of a structural break between two different regression coefficients (e.g. NI and CI) because, while the Chow (1960) test allows us to verify whether a relation remains stable over two periods of time or whether the same relationship holds for two different groups of economic units (Chow, 1960, p. 591), it does not allow us to verify whether the difference between coefficients of different regressors included in different models are statistically significant.

3 The total stock returns have been calculated as follows: RET_it = (P₁+D−P₀)/P₀, where P₁ is the price per-share at the end of the period; P₀ is the price per-share at the beginning of the period and D refers to the dividends paid by the entities during the period.

4 Since a high correlation between regressors could bias results, we perform a multicollinearity test assessing the variance inflation factor (VIF) and the condition number. In our case, as the mean VIF is 6.81 and 5.59 for the price and the return models and the condition numbers 7.78 and 7.72 (both below 10), a serious multicollinearity problem is unlikely.

5 Following the mainstream of the literature, we try to limit noises due to the scale effect by using a suitable deflator, that is the market value at the beginning of the period in the main analysis and the number of shares in the sensitivity analysis described in the text. It is noteworthy that some scholars suggested to face the ‘scale effect’ not by deflating all the variables, but by using models different from the classic OLS, such as the Generalized Linear Model (Wu and Xu, 2008), the fixed effects model (Devalle, Onali and Magarini, 2010) and the addition of scale proxy as independent variable (Barth and Kallapur, 1996).