Value relevance is defined in the existing accounting literature as an accounting number that has a predicted association with equity market values or share prices (Barth, Beaver, & Landsman, 2001). The literature examining this association dates back to the study of Miller and Modigliani (1966), and the first study to use the term of value relevance is Amir et al. (1993). As Barth, Beaver, & Landsman (2001) state, the purpose of value relevance studies is to provide evidence regarding the relevance and reliability of accounting numbers, which are reflected in equity market values and share prices. In the academic literature value relevance is defined as the empirical operationalization of relevance and reliability. An accounting number will be value relevant when this number reflects information which is relevant to investors in order to value the entity, thus implying an accounting number is value relevant when it can make a difference in the decisions taken by a user (Barth, Beaver, & Landsman, 2001). The different studies done on value relevance use different models to structure their test, which most often used as their benchmark, equity market value. Equity market value is perceived to reflect information used by investors. For my thesis, I examine the association between net income, comprehensive income and other comprehensive income with both equity market values and share prices. The first step to examine this association in my thesis is selecting a valuation model. The most common used model is that based on Ohlson (1995) and the adjustments made to the original model (e.g. Feltham and Ohlson, 1995; Ohlson 1999). Firm value is presented by the Ohlson (1995) model as a linear function of book value of equity and the expected future abnormal earnings. Ohlson (1995) presents the following view: equity valuation can be performed accurately if book value of equity can be measured accurately and abnormal earnings can be predicted based on ‘clean surplus’ accounting. The model of Ohlson (1995) assumes that capital markets are perfect, however this model allows for imperfect product markets for a finite number of periods. The models is not dependent on permanent earnings or assets values or liability values, which makes it possible to implement this model without requiring specifying a link between accounting numbers and economic constructs. This model received critics from academics, whom argue that it does not include the possibility of economic rents (i.e. returns in excess of the cost of capital for a finite number of periods). However, this claim is waived by Barth, Beaver, & Landsman (2001) stating that economic rents are viewed by the Ohlson model as being reflected in the persistence of abnormal earnings. Research on the value relevance of accounting numbers only needs one assumption, share prices reflect investors’ consensus beliefs. The literature on investors’ consensus beliefs began with Ball and Brown (1968), who found that share prices are affected by the valuation implications of the information which is publicly available. Thus, the model presented by Ohlson (1995), with the assumption that investors’ consensus beliefs is reflected in share prices to research, the usefulness of net income, comprehensive income and other comprehensive income to investors can be determined. Finally, value relevance studies do not require market efficiency (Barth, Beaver, & Landsman, 2001). As they explain, research on value relevance does not assume true or unbiased measures of equity market values or true economic values of the firms’ assets and liabilities or income.
The research by Ohlson (1999) presents three attributes of transitory (non-recurring) performance measures, of which two must meet the requirement to be indeed transitory. These three attributes of Ohlson (1999) are used for the research of my thesis to answer one of the most common arguments against comprehensive income and other comprehensive income, and the research question of this thesis, which is that these performance measures are not transitory, thus not useful to investors. The following are the three attributes of Ohlson (1999): first, the accounting number is not value relevant (price and returns relevance), second, the accounting number does not have the ability to predict itself (persistence) and third, the accounting number is not relevant in forecasting abnormal income in the following period (forecasting ability).
2.2.1. Price relevance
Black (2016) defines price relevance as the relation between an accounting number (i.e. net income, comprehensive income and other comprehensive income) and a firms’ equity market price. Value relevance studies examine the association between accounting numbers and equity market values, which can be done by examining share prices or changes in share prices. The distinction between using changes in share prices or share prices is that studies examining price changes are interested in changes in firms’ value over a specific period of time, while the association with share prices determines what is actually reflected in a firms’ value. This distinction is important to make, as change in prices is only used for timeliness studies, while the association with share prices is not. Thus, for my research the association with share prices will be used, as this association is not merely to measure the timeliness of accounting numbers.
2.2.2. Returns relevance
The definition of returns relevance is the relation between a firms’ accounting number (i.e. net income, comprehensive income and other comprehensive income) and that firms’ equity market return (Black, 2016). Returns relevance is used to investigate which performance measure of a firm is reflected in its stock returns (Dhaliwal, Subramanyam, & Trezevant, 1999). Research examining value relevance can also make a distinction here between the association with returns or the changes in returns. Like the association between share prices or changes in share prices, I use returns instead of changes in returns, as I am not only interested in the timeliness of the different performance measures. However, the returns models are suitable for the timeliness of accounting numbers as examined by Easton (1999).
2.2.3. Persistence
In this thesis, I refer to predictability- the ability of an accounting number to predict itself- as persistence. Different performance measures have different levels of persistence. Persistence captures the different performance measurers’ sustainability over time. Performance measures with higher persistence are more likely to be viewed as desirable by investors, as these performance measures are recurring. Consistent with this view different researches have showed that the information content of earnings components increased, when the persistence increased. Earnings components which tend to be more persistent (e.g. core earnings), than earnings components which are either transitory or have zero persistence (e.g. special items) are more value relevant (Brown and Shivakumar, 2003).
2.2.4. Forecasting ability
Other than having different levels of persistence, performance measures also differ in forecasting ability. Ohlson (1999) shows that current gains from a forward contract might not predict future gains from this forward contract. However, this same forward contract might still be able to predict future earnings. Thus, the relation with comprehensive income and other comprehensive income is that, gains and losses from these earning components might be able to predict future cash flows, as these gains and losses accumulate on the balance sheet for several years, prior to being realised (i.e. assets sold, liability settled or a pension plan funded). The FASB (2002) states that performance measures that predict future cash flows are more useful to investors and investors view these performance measures as more desirable.
2.3. Institutional setting
In my master thesis, the International Financial Reporting Standards adopted by the European Union (EU), are the main standards the performance measures (i.e. income, comprehensive income and other comprehensive income) are presented. IFRS are issued by the International Accounting Standards Board (IASB) and were adopted by the EU in 2005, whom required member states to apply these standards, in order to enhance comparability and understandability within financial reporting in its member states. In this section I discuss the IFRS, IFRS in Europe and other IFRS requirements related to the different performance measures.
2.3.1. International Financial Reporting Standards
In my thesis IFRS, endorsed by the EU, are the main financial reporting standards. IFRS are issued by the International Accounting Standards Board, which assumed responsibilities from its predecessor in April 2001, the International Accounting Standards Committee (IASC). The IASC issued International Accounting Standards (IASs). Some of these ISAs are still in use, however in my thesis all standards issued by the IASB are referred to as IFRS. All entities listed on the stock exchanges of European Member States were required to adopt IFRS for financial statements beginning on or after January 1, 2005. Whilst the majority of entities were required to apply IFRS did so when it became mandatory, some European countries permitted the adoption of IFRS before 2005 (Barth, Beaver, & Landsman, 2001). IFRS grew and advanced compared to its adoption in 2005 in the EU. Nowadays, IFRS is a global financial reporting framework, that provides principle-based standards on how to prepare the financial statements. The principle-based approach of IFRS have been both endorsed and opposed by academics and users of financial statements. On the one hand, those who agree with the principle-based approach argue that these standards allow managers more flexibility and greater discretion in preparing the financial statements, thus enhancing the transparency, accountability and efficiency of the financial statements and the preparation of these financial statements. Those who oppose the principle-based standards, argue that these standards allow too much discretion to managers, thus enhancing earnings management and accounting distortions in financial statements. Overall, IFRS as principle-based standards, are said to enhance the comparability and understandability of financial statements, and earnings management and accounting distortions also occur under a rules-based financial reporting standards (e.g. one of the most famous accounting scandal in the history of accounting, Enron, whom prepared its financial statements under U.S. GAAP).
2.3.2. IFRS adoption in the EU
IFRS presented a major shift in financial reporting for many European firms, because these standards differentiate from the domestic standards which were required in member states. To become the main financial reporting standards of member states IFRS had to be endorsed by the EU. The EU regulation that made IFRS a requirement for member states is EU regulation 1606/2002. This regulation requires all listed companies to prepare their consolidated financial statements, which will be subsequently published to the public, to be prepared in accordance with IFRS, as endorsed by the EU, and explicitly state that this has been prepared in accordance with said standards.
The adoption of IFRS by European entities introduced a substantial shift in financial reporting, especially the use of fair value measurement. The fair value measurement resulted in different accounting numbers, compared to the pre-adoption of IFRS. These differences were also reflected in the performance measurement statements. In this thesis, I discuss IAS 39 Financial Instruments and IAS 19 Employee Benefits, as the standards which affected financial reporting and subsequently the financial statement performance measures. IAS 39 requires many financial instruments (e.g. derivatives) had to be recognized at fair value and changes in the fair value should be recorded in net income. The requirements of IAS 39 differed significantly from the previous (domestic) financial reporting standards, which were affective in each member state. This resulted in concerns from European financial institutions. They expressed their concern regarding the fact that the adoption of IFRS 39 would result in the reduction of the usefulness of an entity its financial statements. According to Armstrong et al. (2010), critics of IFRS argued that the adoption of IAS 39 would result in financial instability of Europe, as this would affect many financial institutions. IAS 19 required that all the expenses made for providing employee benefits to be recognised when said benefits were earned. For defined benefit plans IAS 19 offered two methods, first, the actuarial gains and losses had to be recognised immediately in the income statement (i.e. net income or other comprehensive income). Second, the actuarial gains could not be recognised immediately, but amortised (also referred to as the corridor approach). Whilst IAS 19 provided standards for employee benefits, domestic financial reporting standards did not offer specific rules or guidelines to account for employee benefits.
2.3.3. Summary
IFRS, as principle-based standards, which grant managers greater discretion and flexibility to prepare financial statements, was adopted by the EU in 2005 to enhance a financial statements its transparency, comparability and capital markets’ benefits (ICAEW, 2015). The discretion granted to managers can either be used to reflect the true underlying economic reality of an entity or to distort the financial statements. Overall, IFRS as a principle-based standard, are said to enhance the comparability and understandability of financial statements. The adoption of IFRS was not only met with praise, but also critical comments from European financial institutions. One comment, which is important to this thesis is that the application of fair value measurement would decrease a financial statement its usefulness. This comment also refers to the presentation of gains and losses in net income or other comprehensive income. Thus, this makes my research interesting by researching the usefulness of the different performance measures (i.e. net income, comprehensive income and other comprehensive income).
3. Literature review and hypothesis development
The literature review of prior empirical literature related to the usefulness of net income, comprehensive income and the components of other comprehensive income. I review the existing literature on said subjects, and present a summary of these studies and their conclusions (a summary of all main literature is presented in table 1). Finally, at the end of this chapter the hypothesis of my thesis are developed.
3.1. Literature review
The reporting of comprehensive income as a separate item in the financial statements was mandatory in the USA for fiscal years beginning after 15 December 1997 (FASB, 1997). While this has been well implemented by the FASB, under the IASB this is a more recent phenomenon. The IASB standards required entities to issue a statement of comprehensive income for annual periods beginning on or after 1 January 2009. Ever since the FASB has required entities to report comprehensive income, this had been a subject of value relevance and other research. The research on value relevance of accounting numbers is relevant in order to assess their usefulness, which highlights whether these accounting numbers reflect the information used by investors to reach their economic conclusions, which are based on said accounting numbers reported in the financial statements. Thus, value relevance is a typical method used in the (accounting) literature to measure accounting quality and the usefulness of financial statement information to investors (Barth et al., 2008; Chen et al., 2010).
Most of the academic research on comprehensive income and other comprehensive income related questions can be classified into four categories. The first category, academic research which examines the relation between standard-setting issues (i.e. FASB and IFRS) related to comprehensive income. The second category is related to the presentation of comprehensive income and other comprehensive income. The third category is related to the convergence of IFRS and U.S. GAAP, which focuses on the identification of other comprehensive components, the recycling of other comprehensive income components and other differences and issues in the convergence of the financial reporting standards. The fourth category, which is related to my thesis, is the usefulness of other comprehensive income and its components to investors.
Usefulness of earnings and comprehensive income
One of the first studies that examined the usefulness of other comprehensive and its components was by Dhaliwal, Subramanyam, & Trezevant (1999). The first question Dhaliwal et al. (1999) answered is whether earning (i.e. net income) or comprehensive income has greater explanatory power for annual stock returns and stock prices. The second question, which performance measure (i.e. earnings or comprehensive income has a greater predictive ability for future earnings and future cash flows. The study by Dhaliwal et al. (1999) was right after the implementation of SFAS 130, thus the authors used existing financial data prior to the implementation and constructed ‘as-if reported’ numbers for comprehensive income. Findings indicate that net income is a better measure of firm performance than comprehensive income. Net income has a higher association with stock prices and returns and has greater explanatory power of future earnings and cash flows, compared to comprehensive income. The conclusions of Dhaliwal et al. (1999) have been confirmed by more recent studies by Barton et al. (2010) and Pronobis and Zulch (2010). Chambers et al. (2007) examines the value relevance of other comprehensive income and its components using actual as-reported numbers and as-if calculated numbers. The main idea behind this reasoning is that the as-if reported numbers might have introduced measurement errors in the regression models used in previous research (e.g. Dhaliwal et al. (1998)). The results of Chambers et al. (2007) indicate that other comprehensive income is not value relevant in both the pre-and post-SFAS 130 period when as-if reported numbers are used, while its positively priced when actual reported numbers are used. This result can be attributed to the increased transparency of other comprehensive income disclosures since its implementation in 1997. Thus, proving that the research design of Dhaliwal et al. (1999) and other papers using as-if numbers might have caused the different conclusions. Barton et al. (2010) uses data from 46 countries to measure value relevance of eight performance measures (including net income and comprehensive income). Barton et al. (2010) find that performance measures which are in the middle of the financial statements (i.e. operating income, earnings before interest, taxes depreciation and depreciation) have higher value relevance than performance measures lower in the financial statements (i.e. comprehensive income). Thus, the findings indicate that comprehensive income is one of the least value relevant performance measure. Pronobis and Zulch (2010) examine the predictive ability of comprehensive income compared to earnings in a homogeneous institutional setting of German IFRS firms and find that net income has greater predictive power of future firm operating performance compared to comprehensive income. Goncharov et al. (2011) research whether aggregated realised or unrealised income (comprehensive income) influences the decision usefulness by assessing the information, valuation and prediction content of investors. The findings indicate that net income and comprehensive income are positively associated with price changes and that aggregated comprehensive income does not improve the measurement association of net income.
Usefulness of other comprehensive income its components
Most of prior research has studied the individual components of other comprehensive income, under U.S. GAAP, which consisted of the following components: first, foreign currency translation adjustments, second, gains/losses on derivative instruments that are designated as cash flow hedges, third, unrealized holding gains from available-for-sale securities and fourth, certain components of pensions and post-employment benefits (i.e., gains or losses not recognized in earnings, prior service costs, and transition costs not recognized in earnings). A study of individual components of other comprehensive income, before the implementation of SFAS 130, was by Soo and Soo (1994). They examined the valuation effect of foreign currency reported in equity (after SFAS 130, this had to be included in other comprehensive income) and found that the market uses foreign currency translation adjustments in reaching their economic decisions, however the valuation effect of foreign currency translation adjustments is much smaller than other components reported in earnings. Another study that finds that the foreign currency translation adjustments are value relevant is by Bartov (1997). However, a study by Louis (2013) indicates that a negative change in a firms’ value is associated with a positive change in foreign currency translation adjustments. Louis (2013) uses a sample of manufacturing firms, to illustrate that a negative foreign currency translation adjustment is caused when the local or foreign currency depreciates in value. Thus, this indicates that an entity its operating environment is an important factor for the value relevance of foreign currency translations. Furthermore, these results indicate the importance of disaggregation of the other comprehensive income components.
A study by Barth et al. (1996) examined the value relevance of one of the components of other comprehensive income (i.e. fair value of financial instruments) for a sample of U.S. banks, prior to the implementation of SFAS 130. The findings by Barth et al. (196) indicate that the bank share prices are associated with the disclosure of fair values of securities, loans and other long-term debt. The findings of this researh indicate that unrealized holding gains from available-for-sale securities (as referred to in SFAS 130) and fair values of financial instruments are valued by the market. Another study that examined the value relevance of individual comprehensive income components is by Dhaliwal et al. (1999). The findings of Dhaliwal et al. (1999) support those of Barth et al. (1996), gains/losses on marketable securities are value relevant, however, this component of other comprehensive income is only value relevant for the financial services industry. The other two components of other comprehensive income, foreign currency translation adjustments and minimum pension liability in excess of unrecognized proir service costs, examined by Dhaliwal et al. (1999) were not value relevant. O’Hanlon and Pope (1999) use a sample of firms in U.K. reporting under U.K. GAAP, and examine the value relevance of different earnings components (i.e. extraordinary items, foreign currency translation adjustments and revaluations). The findings indicate that only extra ordinary items are value relevant, if measures over multiple years, and the other earnings components are not considered value relevant.
Price relevance
Value relevance studies examine the usefulness of accounting information by assessing the association between accounting numbers and stock returns and prices. The higher the explanatory of accounting information for stock prices and returns, indicates more usefulness of that accounting information to investors. Almost all papers mentioned in this literature review assed the price and returns relevance of earnings and comprehensive income. Dhaliwal et al. (1999) find that net income has a higher association with stock prices than does comprehensive income. Cahen et al. (2000) assessed the price relevance of individual other comprehensive components in New Zeealand and finds that only fixed asset revaluation adjustments are associated with stock returns. Kanagaretnam et al. (2009) assess the association of other comprehensive income components and find that unrealised gains and losses on available-for-sale securities are significantly associated with the entity its stock prices.
Goncharov and Hodgson (2011) also assess the association of individual other comprehensive income components and find that both comprehensive income and other comprehensive income are value relevant, however net income is found to be more significant. The components of other comprehensive income (i.e. revaluation reserve adjustments, foreign currency translation adjustments, and unrealised gains and losses on available-for-sale securities) are value relevant, but when the authors controlled for net income, the components were significantly less value relevant. Goncharov and Hodgson (2011) also find that revaluation reserve adjustments, foreign currency translation adjustments are not only value relevant for investors, but also for analysts, when assessing their stock price targets.
Returns relevance
The same value relevance studies examining the usefulness of accounting information by assessing the association between accounting numbers and stock prices also examine the association with stock returns. Dhaliwal et al. (1999) find that the association of stock returns and comprehensive income is stronger than the association between net income and stock returns. One component of other comprehensive income, in particular unrealised gains and losses on available-for-sale securities, explains the stronger association of comprehensive income and stock returns. This association is driven by financial institutions; however, the authors conclude that the association between comprehensive income and stock returns is also stronger for non-financial firms, and that this could be due to the magnitude of other comprehensive income. Thus, the results of Dhaliwal et al. (1999) indicate that the environment and the business model of the entity should be considered as an important determinant when assessing which performance measure is value relevant. O’Hanlon and Pope (1999), who assessed the association of other comprehensive income and stock returns for a sample of U.K. firms find that other comprehensive income component are significantly associated with stock returns. Kanagaretnam et al. (2009), find that comprehensive income has a stronger association with stock returns compared to the association of net income and stock prices. Kanagaretnam et al. (2009) who also assess the association of other comprehensive income components find that unrealised gains and losses on available-for-sale securities are significantly associated with stock returns. Barton et al. (2010) find that comprehensive has the strongest association with stock returns, out of eight different performance measures. Jones and Smith (2011) their findings indicate that special items have a stronger association with stock returns compared to comprehensive income components. And finally, Goncharov and Hodgson (2011) also tested the association of other comprehensive income components and the findings indicate that only unrealised gains and losses on available-for-sale securities are returns relevant.
Persistence
Persistence of performance measures (i.e. comprehensive income and other comprehensive components) has significant implications for an entity’s firm value (Friedman 1957; Kormendi and Lipe 1987; Collins and Kothari 1989). The valuation model of Ohlson (1995, 1999) suggest that a financial reporting statement that emphasizes the importance of persistence is useful to investors. Jones and Smith (2011) indicate that special items are not persistence (i.e. zero persistence), while other comprehensive income is negatively persistent.
Forecasting ability
Dhaliwal et al. (1999) find empirical evidence that net income is a better predictor of next years’ cash flow from operations and income, compared to comprehensive income. This conclusion is supported by the findings in the study of Kanagaretnam et al. (2009), who find that net income is a significantly better predictor of future income, compared to comprehensive income. However, Kanagaretnam et al. (2009), find that comprehensive income is a better predictor of future cash flow from operations, thus this is the opposite of Dhaliwal et al. (1999). Kanagaretnam et al. (2009) find that unrealized holding gains from available-for-sale securities is the most significant component of other comprehensive income in forecasting future cash flow from operations. Barton et al. (2010), who empirically tested the forecasting ability of eight performance measures, finds that comprehensive income has the least predictive power to forecast future cash flow from operations. The findings of Barton et al. (2010) are consistent with those of Dhaliwal et al. (1999). Jones and Smith (2011) examine the forecasting ability of special items and other comprehensive income and find that special items have a greater forecasting ability than other comprehensive income. Goncharov and Hodgson (2011) test the forecasting ability of comprehensive income and net income, and find that comprehensive income has a lower forecasting ability for cash flow from operations.