Are foreign earnings disclosures value-relevant?

Accepted Manuscript Title: Are Foreign Earnings Disclosures Value-Relevant? Disaggregation Solves the Puzzle Author: R. Joliet A. Muller PII: DOI: Reference:

S0275-5319(15)30011-8 http://dx.doi.org/doi:10.1016/j.ribaf.2015.09.009 RIBAF 390

To appear in:

Research in International Business and Finance

Received date: Revised date: Accepted date:

24-1-2014 13-6-2015 8-9-2015

Please cite this article as: Joliet, R., Muller, A.,Are Foreign Earnings Disclosures ValueRelevant? Disaggregation Solves the Puzzle, Research in International Business and Finance (2015), http://dx.doi.org/10.1016/j.ribaf.2015.09.009 This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

Are Foreign Earnings Disclosures Value-Relevant?Disaggregation Solves the Puzzle R. [email protected], A. Muller2*[email protected] 1

IESEG School of Management, Rue de la Digue, 3, F-59000 Lille, France. HEC Management School of the University of Liège, Ethias Chair in Asset and Risk Management, Liège – Belgium

ip t

2

cr

HEC Management School of the University of Liège, Finance Department, Rue Louvrex, 14, B4000 Liège, Belgium. Phone: +32.4.232.7435; Fax: +32.4.232.7376. Abstract

M

an

us

This study examines whether foreign region- and country-specific performance disclosures relate differentially to shareholder wealth compared to domestic performance disclosures. Bearing in mind the fact that, in previous literature, empirical findings regarding the value-relevance of geographic performance disparities have been surprisingly conflicting and puzzling, we question in this paper prior sample selection procedures and data treatments. Using an extensive international company-level database, we distinguish between foreign ‘region-’ and ‘country-’specific performance disclosures and find that (1) the use of disaggregated foreign earnings information increases significantly the precision and significance of the value-relevance of foreign performance; (2) the disclosure of positive earnings movements generated in Latin American countries has a consistently negative value impact and (3) the value-relevance of foreign segment disclosures is particularly perceptible for Asian Pacific countries and the United Kingdom, confirming the attractiveness of these areas for foreign direct investment.

Ac ce p

te

JEL classification: F23, M41, G14

d

Keywords Firm valuation; accounting disclosures; foreign earnings; multinational firm; disaggregation

Introduction

With the ongoing internationalization of trade and capital flows, the pricing impact of disclosed foreign operations performance has grown to one of the most interesting and exciting debates in accounting, business and international finance. The last three decades have indeed witnessed an impressive surge in companies’ foreign operations and assets all over the world. Corporate rationales for geographic diversification include economies of scale, access to larger customer bases, proximity of natural resources, operational flexibility and the ability to arbitrage institutional as well as financial restrictions and regulations. Value-based studies seem to suggest that investors are willing to pay a premium for shares of internationally diversified companies, as they represent a claim on a collection of diversified profit streams generated in various areas of the world. Empirically, however, the link between firms’ foreign activities and shareholder value remains ambiguous and puzzling. While in earlier years, empirical studies fairly unanimously supported the theoretical concept that geographic diversification enhances firm value (Caves, 1971; Kogut, 1983; Errunza and Senbet, 1981; Hines and Rice, 1994) by showing that disclosed foreign operations resulted in positive valuation effects (Errunza and Senbet, 1984; Kim and Lyn, 1986), later evidence is

Page 1 of 25

Ac ce p

te

d

M

an

us

cr

ip t

mixed and conflicting. While Denis et al. (2002) report evidence on the negative value impact of international diversification, Morck and Yeung (1991), Garrod and Rees (1998), Bodnar et al. (2003b) and Errunza et al. (2008) suggest that geographic diversification is positively related to firm value. On the other hand, focusing on disclosed foreign activities, Boatsman et al. (1993), Christophe (1997) and Christophe and Pfeiffer (2002) show that investors do not value foreign operations as highly as they value domestic operations, while Bodnar and Weintrop (1997) and Bodnar et al. (2003a) report that firm value is more strongly affected by changes in foreign earnings than by changes in domestic earnings. The purpose of this study is to address this puzzle and the key issues related to the impact of foreign operations performance disclosures on shareholder wealth. It also aims to fill the gap left by previous research in performing a comprehensive and detailed analysis of the impact of geographic earnings disparities. The basic idea is simple: given the increasing number of companies providing country-specific disclosures of their foreign operations, it is surprising that no existing study has yet investigated the value-relevance of detailed foreign performance disclosures. As country-specific disclosures have the most easily identifiable economic characteristics relative to domestic operations disclosures, they are the most informative in terms of differential value-relevance – they are the most appropriate means of telling us whether and why investors conjecture either a stronger or weaker value impact of foreign operations relative to the value impact of domestic activities. We investigate, using a sample of 390 European, Canadian, Australian, New Zealand and U.S companies, how foreign versus domestic operations are reflected in shareholder wealth, how the disclosures of foreign earnings in specific regions constitute incremental value-relevant news for investors and how the disaggregation of this region-specific performance information into country-specific performance information provides shareholders with additional value-relevant information above and beyond the information contained in more commonly investigated aggregated disclosures. We conduct four major tests of the relationship between foreign performance disclosures and firm value. The following key points have emerged from our empirical findings: Both domestic and foreign performance disclosures have a significant positive impact on firm value. While valuation effects seem to depend on the country where the company is located, evidence on the differential impact of domestic versus foreign performance remains inconclusive – region, respectively country-specific information is needed. Disaggregated foreign country-specific earnings are statistically significantly reflected in firm value. The comparison between the value-relevance of country-specific versus region-specific profit disclosures confirms, moreover, that previous empirical designs based on the aggregation of foreign country-level performance disclosures into foreign region-specific disclosures lead to an important loss of value-relevant information for investors and are hence misspecified. In contrast with the positive value impact of domestic performance, most authors (Rask et al., 1998; Wurgler, 2000; Durnev et al., 2004; Wet, 2004) anticipate that when the foreign performance disclosures relate to a vulnerable or less transparent market, they will have a relatively weaker – or even negative1 – value impact on firm value. Our study finds support for a

1

Given that a firm’s foreign region- and country- specific performance is usually publicly announced after the disclosure of its overall consolidated earnings, foreign earnings disclosures constitute ‘additional’ relevant signals for investors. This is consistent with the functional fixation theory, which states that investors respond to bottom line accounting numbers without fully reflecting on the valuation implications of the different components of earnings.

2

Page 2 of 25

ip t

negative link between firms’ disclosed earnings in specific Latin American markets and shareholder value. Results show that investors perceive positive changes in foreign earnings related to Asia Pacific and the United Kingdom as outstandingly good news. They associate relative higher responses to these foreign performance disclosures –hence confirming the attractiveness of these countries for foreign direct investment.

us

cr

The outline of the remainder of this paper is as follows. In the next section, we discuss the background and motivation of this paper. Section 2 describes the selection procedure used to form the sample of multinationals analyzed. It details carefully various sample characteristics. In section 3, we describe the empirical design used to measure the value-relevance of geographic earnings disclosures. The estimates obtained from alternative specifications are presented and the results analyzed. Section 4 concludes.

an

Background and motivation

Ac ce p

te

d

M

The purpose of this study is to investigate the relative value-relevance of foreign earnings disclosures with respect to domestic earnings disclosures. Following Francis and Schipper (1999), Core et al. (2003) and Kothari and Shanken (2003), we regard the value-relevance of this accounting information as the degree to which domestic versus foreign earnings disclosures are reflected in shareholder wealth. Recent evidence describes extensively how firms manage financial reporting practices to provide support for superior managerial ability (Iatridis and Valahi, 2010) to influence their stock price (Graham et al., 2005). They show that earnings matter more to managers than cash flows for financial reporting purposes. Investors are hence expected to assess the value of a firm based on its earnings information, basing their assessment less on its cash flow information. The underlying question is whether and how investors use geographical segmented earnings disclosures when establishing the value of a company. First of all, it should be emphasized that information regarding a firm’s foreign region- and country-specific performance is usually publicly reported ‘after’ the disclosure of its overall consolidated earnings. Moreover, according to the functional fixation hypothesis, investors tend to be influenced by “functional fixation” on consolidated net income and do not fully interpret the information on earnings components. They subsequently correct this mispricing. They react to segmented information in earnings components with a time delay (Hand, 1990; Tiniç, 1990; Sloan, 1996; Harris and Nissim, 2006). As a consequence, segmented foreign earnings disclosures constitute relevant news for investors. Authors unanimously agree that the geographical disaggregation of a firm’s performance conveys additional informative news to investors and analysts (Hussain and Skerratt, 1992; Herrmann, 1996; Nichols et al., 1996; Doupnik and Seese, 2001; Kou and Hussain, 2007). However, given the complex set of issues associated with this task, there is no doubt that there are numerous reasons why systematic mispricing errors may arise when investors assess the relationship between a firm’s foreign activities and its shareholder value. The central controversy over whether investors conjecture either a stronger or weaker value impact of foreign operations relative to the value impact of

Subsequent price adjustments occur when the information contained in the financial statements is unscrambled (Hand, 1990; Tiniç, 1990; Sloan, 1996; Harris and Nissim, 2004).

3

Page 3 of 25

Ac ce p

te

d

M

an

us

cr

ip t

domestic activities remains empirically unanswered and has grown to represent a wide debate in the literature. On the one hand, in line with international diversification and foreign direct investment theories (Caves, 1971; Kogut, 1983; Errunza and Senbet, 1981; Hines and Rice, 1994), Bodnar and Weintrop (1997) argue that U.S. multinational foreign operations have a stronger value impact than domestic earnings, as they are related to foreign markets identified and selected by the firm for its underexploited operational flexibility, diversification and growth opportunities. Their findings are supported in a subsequent study by Bodnar et al. (2003a) for firms listed in Australia, Canada and the United Kingdom.2 On the other hand, several authors argue that there are multiple reasons why foreign operations should have weaker wealth effects. In comparison with domestic earnings, foreign earnings are in fact characterized by an additional source of risk – exchange rate risk – as they are affected by future exchange rate movements (Boatsman et al., 1993). Exchange rate volatility thus increases the variability and risk of the domestic currency denominated value of foreign earnings. It is for this reason that one can presume that investors are more careful when incorporating information related to foreign activities in their valuation, as they may not be able to gather, analyze and accurately interpret all the information disclosed in foreign performance reports. Moreover, as mentioned in Eaker et al. (1996), foreign activities may be riskier and more costly than domestic ones.3 Foreign regulatory and cultural environments as well as barriers to entry have to be discerned and actively managed by the firm. All these reasons may lead investors to suspect that foreign investment opportunities are potentially sub-optimal, that they have been primarily undertaken for the benefit of managers4 and that they are hence less value-relevant for the firm and for investors’ investment portfolios (Anderson et al., 2011). Duru and Reeb (2002) suggest that international diversification increases, moreover, the difficulty of forecasting earnings accurately because of the greater volatility of foreign earnings and the additional opportunities for managerial discretion – which in turn increases the information asymmetry between analysts and management. The difficulty of reflecting foreign performance in investors’ pricing is, furthermore, enhanced by an additional information bias, as it has been empirically shown that investors tend to systematically underestimate the time persistence of reported international performance information (Thomas, 2000a; Khurana et al., 2003; Callen et al., 2005). Finally, Duru and Reeb (2002) document the fact that greater international diversification is associated with more optimistically biased analysts’ forecasts of earnings. This is notably due to analysts’ unfamiliarity with the international information environment and the necessity to maintain favorable relations with managers who allow analysts to access private information (Lim 2001). The preceding discussion reveals that the relative value-relevance of domestic versus foreign earnings disclosures is intimately and inextricably associated with the characteristics of the country where the foreign earnings have been generated. The characteristics of the host country play a major role in determining the way investors perceive and interpret the accounting information related to this country.

2

It should be mentioned that, according to Hope and Kang (2005), foreign earnings are no longer incrementally significant in explaining stock returns when controlling for other value-relevant information. 3 Empirical support is provided by Reeb et al. (1998), who find that investors use higher discount rates when valuing international operations than when valuing domestic operations. 4 See, for instance, Christophe (2002) and Jensen (1986) for a detailed discussion on potential agency problems associated with firms’ foreign activities.

4

Page 4 of 25

Ac ce p

te

d

M

an

us

cr

ip t

While it is a generally held view that emerging markets are weakly correlated with developed markets and offer promising diversification benefits at the corporate level, investors regard emerging economies as a separate group of countries. They are commonly considered as a basket of countries with one important characteristic in common: being risky, or at least being more risky than developed markets. In times of volatile world markets, the emerging markets are usually more affected by volatility. Macroeconomic instability and exchange rate risk also play a major role when trading or operating with emerging economies. As investors tend to be riskaverse, they are inclined to underestimate the value impact of foreign performance originating from developing economies. Collins (1990), as well as Kwok and Reeb (2000), finds empirical support for these expectations. In spite of high rates of return on direct investment in developing countries, Collins (1990) finds that market model performance measures, ordering models, and efficient frontier analysis unanimously indicate that multinationals in developed countries perform better, while multinationals in developing countries have the worst performance overall. Kwok and Reeb (2000) similarly suggest that a multinational’s home and target market conditions play an important role in predicting how its internationalization influences its riskiness. From a more general perspective, Stiroh and Rumble (2006) describe the so-called dark side of diversification – when the cost of increased exposure to volatile activities is higher than the expected gains of diversification. When a firm headquartered in a more developed economy expands its direct investments into less stable markets, the overall systematic firm risk will increase – and the ultimate impact on firm value may indeed be negative. It should, furthermore, be stressed that investors have a preference for accounting information that can be easily researched, where all relevant information is available and informed return expectations can be made. Previous studies reveal that information and news from culturally as well as institutionally distant countries or regions are differentially perceived by investors. The asset allocation process towards these countries turns out to be negatively affected (Anderson et al., 2011; Beugelsdijk and Frijns, 2010). So-called lower-class or reduced information harms as a matter of fact investors’ decision-making, decreases the confidence in the firm and destroys value. (Wurgler, 2000; Durnev et al., 2004) As emerging economies suffer to a greater extent from asymmetric information problems (Wet, 2004) and are often perceived as culturally distant, accounting information deriving from operations in emerging markets is often considered noisy or unreliable (Rask et al., 1998). Less developed markets may hence suffer from an information as well as culture barrier causing accounting information related to emerging market operations to be insignificantly related to shareholder wealth (Anderson et al 2011). The financial and macroeconomic stability, the prevalent growth perspectives and the cultural and geographical proximity of the foreign country represent some of the many factors influencing and determining the way investors assess the relative valuation impact of foreign region and/or country specific earnings with respect to domestic earnings disclosures. Surprisingly, however, little attention has been given to the exact and detailed origin of foreign performance disclosures. While most studies consider the decomposition of earnings into domestic versus foreign components (Bodnar and Weintrop, 1997; Thomas, 2000a; Christophe, 2002; Khurana et al., 2003; Bodnar et al., 2003a; Thomas, 2004; Callen et al., 2005; Hope and Kang, 2005), only a few authors have investigated how far investors incorporate the pricing effects of disclosed foreign operations performance into specific geographic regions (Boatsman et al., 1993; Thomas, 2000b; Christophe and Pfeiffer, 2002). However, the use of too aggregated foreign performance measurements mixes foreign earnings disclosures together, which originate from a range of distinctive markets whose stability, growth opportunities as well as managerial

5

Page 5 of 25

Ac ce p

te

d

M

an

us

cr

ip t

characteristics may be extremely heterogeneous and differentially perceived by investors (Hussain and Skerratt, 1992; Herrmann, 1996; Nichols et al., 1996; Doupnik and Seese, 2001; Kou and Hussain, 2007, Antonczyk and Salzmann, 2013). The aggregation of these disparate effects leads inevitably to a difficult interpretation of the information – and hence to a downward bias in the estimation of the value-relevance of these aggregated performance measures. Consequently, problems in the design of previously reported research causing statistically weak and inconclusive results require further research on country-specific earnings disclosures in order to shed new light on the empirical question of whether foreign performance disclosures are as value-relevant as domestic ones. This study adds to the existing literature by testing and comparing how markets assess the valuerelevance of aggregated foreign versus domestic earnings and, more importantly, how investors value region-specific versus country-specific foreign performance disclosures. Both regionspecific and country-specific foreign performance disclosures originate from a large variety of regions and countries characterized by as many economically and institutionally distinctive environments. The discussion is based on an international database consisting of 390 European, Canadian, Australian, New Zealand and U.S companies reporting foreign operations, spread all around the world, at a region- and/or country- specific level.5 We estimate the value-relevance of foreign performance disclosures by regressing total security returns on ‘scaled’ measures of geographic region-specific and/or country-specific earnings – foreign earnings being defined as firms’ disclosed earnings within the foreign region and/or country deflated by firms’ market value. As no existing study has yet incorporated the huge amount of information contained in these country-specific earnings disclosures to estimate the value-relevance of foreign operations, this paper enables us to make four major contributions: (i) First, it confronts the traditional geographic segment earnings coefficients with an empirical specification that measures the valuation impact of foreign ‘scaled’ earnings – allowing hence legitimately the comparison between domestic versus foreign as well as the comparison between foreign region- versus foreign country-specific pricing effects. (ii) Second, we compare, moreover, the valuation impact of disaggregated region- and/or country-specific performance disclosures with the pricing effect of aggregated measures of foreign earnings. We verify hence whether the use of too aggregated foreign performance measures may explain the inconclusive results previously reported in the existing literature. (iii) Third, the contrast between the pricing effects of country-specific earnings disclosures is shown to be extremely informative regarding the way investors incorporate foreign performance news originating from a large and heterogeneous variety of economic and institutional environments. (iv) Finally, a series of robustness checks, including the distinction between home countries and the distinction between dividend and no dividend paying companies enables us to investigate whether the characteristics of the firm’s home market matters or whether investors associate foreign performance disclosures with a signal having a stronger or weaker value impact compared to dividend announcements – or whether signaling through dividend policy predominates.

Sample selection and data description

5

It should be noted that our sample is not restricted to firms with a December 31 year-end – avoiding thus the selection bias that has been systematically observed in many previous studies.

6

Page 6 of 25

Ac ce p

te

d

M

an

us

cr

ip t

Disclosed earnings components are of particularly crucial interest to investors because of the potential for investors to understand more precisely earnings, to better forecast their evolution (Moumen et al., 2015) and hence to better estimate firm value (Khurana et al., 2003). Consequently, most accounting practices invite companies to provide a separate financial reporting on the performance of their international activities. Following the introduction of the Statement of Financial Accounting Standards (SFAS) 131 in the United States, firms are required to report financial information on the basis of geographic "operating segments".6 SFAS 131 stipulates that "operating segments of a company shall be segregated to provide a better understanding of performance and a better assessment of its future cash flows, allowing users of financial statements to make a more informed judgment about a company as a whole”7. From the manager's point of view, SFAS 131 states that the segmented operating results "are regularly reviewed by the enterprise's chief operating decision maker to make decisions about resources to be allocated to the segment and assess its performance" (management approach). Critics of SFAS 131 argue that such room to maneuver could harm the comparability and reliability of segment reporting across U.S. companies. Berger and Hann (2003), for instance, deplore the lack of a precise segment profit definition or the incentive for managers to manipulate internal segment information used in performance evaluation. Like its U.S. counterpart, the Canadian Chartered Accountants of Canada (CICA) Handbook section 1701 mandates geographic segment disclosures. Similarly, the Australian Accounting Standard Board’s AASB 1005 "Segment Reporting" guidelines aim to "facilitate the development of accounting standards that (...) allows users to make and evaluate decisions about allocating scarce resources (...), [and to be] relevant to assessing performance, financial position, financing and investment (...)". Furthermore, the Australian standard specifies that such segment reporting should "facilitate the Australian economy by reducing the cost of capital, enabling Australian entities to compete effectively overseas and having accounting standards that are clearly stated and easy to understand".8 International Standard IAS 14 – initially issued in August 1981 by the IASB and predominantly used by European firms – imposed similar rules for reporting financial information by geographical segment. For IAS 14, however, the definition of geographical segments was required to be based on a “risk and reward approach” (IAS 14, paragraph 19). A geographical segment was considered as a segment providing products and services within this particular economic environment were subject to risks and returns that were different from those affecting other operating units in different economic environments. According to IAS 14, companies had to identify both “primary” and “secondary” segment disclosures (IAS 14 paragraph 26). To determine whether geographical segments were to be used for their primary segment reporting format, companies had to check whether the source and nature of risks and the rates of return of the entity were affected predominately by the fact that it operated in a different geographical area. Again the purpose was to help investors to better understand past performance as well as the nature of both risks and returns involved. Despite the fact that objectives were identically formulated and segments were based on the same quantitative criteria, it was not clear whether 6

SFAS No. 131 became effective for fiscal years beginning after December 15, 1997. We observe in our sample that, since 1997, the proportion of country-specific disclosures has increased while the proportion of broader geographic area segment disclosures has decreased. It should be stressed that these findings are consistent with the observations made by Nichols et al. (2000) and Herrmann and Thomas (2000). 7 Securities and Exchange Commission (SEC) RIN 3235-AH43 Segment Reporting. SFAS N°131 superseded SFAS N°14 in 1998 and the established standards for reporting information about "operating segments" rather than following the "industry segments" standards that were in place previously. 8 ASIC Act Part 12, Section 224.

7

Page 7 of 25

Ac ce p

te

d

M

an

us

cr

ip t

the same geographical disclosures could satisfy both IAS 14 and SFAS 131. IAS 14 was criticized due to the flexibility introduced by the “risks and returns” qualification, which could potentially allow managers to report segments in an inconsistent manner with the internal reporting structure of the business. This could happen if the internal reporting structure did not reflect the nature and source of the business “risks and returns” (Bugeja et al. 2015). In spite of obvious problems such as geographical groupings, common cost allocations, intragroup transfers, management discretion, inflation and exchange rate movements, geographic segment reporting has been shown to be informative to investors (Berger and Hann 2003, Hossain 2008, Hope et al., 2009). The informative power of segment disclosure has, moreover, been particularly strong since the Financial Accounting Standards Board (FASB) and the International Accounting Standards Board (IASB) committees required segments to be identified in line with managers’ own view of them9. The way earnings are segmented is thus coherent with the way managers guide operating/financial decisions and the way they assess strategic performance (Radebaugh et al., 2006). The management approach of geographical segment disclosure leads to a more disaggregated geographical segment reporting (Street et al. 2000, Berger and Hann 2003, Nichols et al. 2013, Bugeja et al. 2015). Such disaggregated disclosures are value-relevant, since they reduce information asymmetry, noise in the foreign earnings signal and information acquisition costs for investors (Hussain 2008, Hope et al. 2009). Graham et al. (2005) argue that earnings provide more information content about firm value than do cash flows. They report that accounting earnings matter more to managers than cash flows for financial reporting purposes and find that most earnings management convey reliable information on real decision-making – as opposed to accounting manipulations. This is consistent with the IASB Conceptual Framework for financial reporting (2010 version) according to which relevance and faithful representation are the fundamental qualitative characteristics of useful financial information. Relevant information makes a difference to users’ decision-making process if it has predictive value, confirmatory value or both. To be useful, relevant information needs as well to depict faithfully economic phenomena by exhibiting completeness, neutrality and freedom from error. Still, evidence shows that while the number of disclosed segments increased with the adoption of the management approach, the number of line items disclosed per operating segment decreased (Nichols et al. 2013, Bugeja et al. 2015). This reduction results from a corporate choice that may be due to proprietary cost reasons. Firms decreasing line item information have significantly less loss making segments and operate in more concentrated industries (Bugeja et al. 2015). In this study, changes in geographic segment earnings are measured by annual changes in foreign pre-tax operating earnings deflated by the firm’s market value at the beginning of the yearly interval.10 We gather accounting segmental data from the Thomson Worldscope database.11 Data is collected from 1993 to 2006. Given that economic crises affect the reporting quality of earnings, the scope for earnings manipulation, the overall corporate management of earnings as well as firms’ profitability (Campello et al., 2010, Kousenidis et al., 2013) we exclude the crisis

9

IFRS 8 superseded IAS 14 in 2009. By introducing the management approach, IFRS 8 is aimed to reduce differences between US GAAP and IFRS. 10 Non-operating data are not available in geographic segment disclosures. Following Bodnar et al. (2003a), we hence concentrate our study on the comparison between the value-relevance of domestic versus foreign performance at the operating performance level. 11 We consider disclosures as originally reported, not as subsequently restated.

8

Page 8 of 25

Ac ce p

te

d

M

an

us

cr

ip t

period starting in 2007.12 We compute for each individual company and each specific year geographic segmented annual earnings measures for each disclosed country/region of operation. Performance is calculated as the ratio between the foreign pre-tax operating income generated within the country/region as reported in the financial accounts and the market value of the firm. All values are denominated in the firm’s domestic currency. As changes in earnings have to be estimated, we only include firm-year observations for which we observe at least two consecutive years of region- respectively country-specific accounting data. Our selection procedure consists of several steps. The first selection step investigates all firms being incorporated and traded on the most important developed markets. Specifically, we verify the data availability for all S&P 500, S&P 400, and TSX companies in North America, for all SBF 120, DAX 30, MDAX 50, AEX and AMX companies in Western Europe and for all ASX 200 and NZSX 40 companies in Australasia. Financial services companies are excluded from our sample.13 Subsequently, companies that are not internationalized (i.e. that do not report foreign activities in at least one foreign area) are similarly eliminated. We are interested in firms disclosing sufficiently detailed geographical segmental information (i.e. foreign segmental profit and foreign segmental assets). Consistent with the current literature (Boatsman et al., 1993; Thomas, 2000b; Christophe and Pfeiffer, 2002), companies that aggregate foreign operations into a single disclosure labeled “international” or “foreign” are hence excluded. Firms from some developed countries (for instance Spain, Italy or Switzerland) do not disclose precise geographic segmented data and are thus similarly excluded. The entire procedure yields a total sample of 417 multinational firms. According to the details provided in the Thomson Worldscope database, we observe that the degree of accuracy of geographic segmented performance disclosures differs between our sample firms. A company does not systematically convey specific accounting information for each country in which it operates. Some companies disclose data for large (and often heterogeneous) geographic regions (e.g. North America, Latin America or Asia Pacific), while others provide more accurate information in their disclosure (distinguishing between the United States, Canada, Brazil, Chile, Korea, Japan, etc.). There is no doubt that the degree of precision depends on the importance of the company’s activities within the foreign country/region relative to its total economic activity. However, the deliberate choice of a company to smooth or dissimulate performance in specific countries/regions may also play a role. Moreover, it should be emphasized that other factors such as limited incentives to provide segmental information because of proprietary, preparation or competitive concerns also influence firms’ disclosure policies. We distinguish between three quality levels of disclosure. First, we add up the disclosed series of countries’ foreign performance into a single aggregate measure of foreign performance. To do this, we sum for each company all operating incomes generated in foreign countries/regions. Second, we focus on region-specific disclosures as they are reported by the firm in its financial statements – namely Europe, Eastern Europe, North America, Latin (or South) America, Africa, 12

Kousenidis et al. (2013) and Iatridis et al.(2013) observe that during crisis periods earnings quality seems to deteriorate for firms that have an incentive for earnings management as these firms tend to engage in earnings management to mask their lower profitability. In contrast firms that have less incentive for earnings management tend to improve actively the quality of their earnings leading to a higher value-relevance of their earnings during crisis periods. While these issues are of interest the analysis from an international operation perspective of these differential behaviors lies beyond the scope of this study. 13 Following Bodnar et al. (2003a), financial services companies are not included in our sample, as their foreign revenue disclosures are not comparable to the disclosures of non-financial firms.

9

Page 9 of 25

te

d

M

an

us

cr

ip t

Asia(-Pacific) and Middle East. It should be stressed that investors do not necessarily know the accurate country composition of these region-specific disclosures. Finally, we investigate the country-specific disclosures – selecting twenty-two countries that meet the above-described sample selection requirements.14 All these data requirements result in a sample of 2,018 firm-year observations for 390 distinct firms sorted into eight different industry sectors.15 The 390 firms of our final sample have their registered offices in Europe (43), the United States (225), Canada (62) and Australia/New Zealand (60). More than 83% of these firms (324 companies) usually disclose geographic segmental data across specific regions. Among these companies almost one in two (148) also discloses country-specific performance.16 Of all the companies disclosing country-specific data – 54% of the total sample (214 companies) – it should be emphasized that 31% (66 firms) report exclusively disclosures at a country level, with no aggregation for any geographical segment. The dependent variable of our analysis is based on firms’ market value. The stock returns employed are based on daily adjusted prices denominated in firms’ domestic currencies. They are obtained from Datastream International.17 Following Boatsman et al. (1993), Bodnar and Weintrop (1997), Thomas (2000a, 2000b, 2004), Christophe (2002), Bodnar et al. (2003a), Khurana et al. (2003), and Callen et al. (2005), we compute annual stock returns over an interval beginning and ending 3 months after the firm's fiscal year-end. A lag of three months is chosen to ensure that the geographically segmented earnings have been fully disclosed to investors through annual reports and/or other publications – while preceding the disclosing of any new quarterly report. It should be noted that this data collection procedure is carried out in spite of the fact that many firms' fiscal years end on different dates – and not necessarily on December 31st. The measurement window to calculate returns is hence based on each individual company’s accounting data release. This is in sharp contrast with previous studies (Boatsman et al., 1993; Thomas, 2000b; Christophe 2002), which limit their sample to firms with December fiscal year ends.

Ac ce p

The breakdown of the sample across country of origin, industry sector and geographic location of foreign operations is reported in Table 1. Obviously, Europe constitutes the major geographic region for which foreign activities are disclosed in the financial statements. Europe is followed by Asia Pacific, and then by North and Latin America. Industrial companies are most represented in the sample. It should be mentioned that, in most sectors, companies have well-diversified international activities across Europe, Asia Pacific, and North and Latin America. As expected, foreign operations in the Middle East are predominantly focused on the energy (oil) and industrial sectors. Similarly, companies operating in Africa are mostly active in consumer non cyclical, energy (oil) and materials (commodities).

14

Countries included are: France, Germany, Ireland, Spain and the UK in Europe, Canada and the USA in Northern America, Argentina, Brazil, Chile, Ecuador and Mexico in Latin America, South Africa in Africa, and China, HongKong, Japan, Korea, Yemen, Australia, Indonesia, New Zealand and Papua New Guinea in Asia Pacific. 15 These industry sectors are: consumer cyclical, consumer non cyclical, energy, healthcare, industrials, materials, technology, and utilities. 16 The increasing percentage of companies providing country-specific disclosures of their foreign operations enables us to investigate the value-relevance of these detailed foreign performance disclosures. 17 We consider exclusively the listing information in the firm’s country of incorporation. If a company is also listed on foreign exchanges, these listings are ignored in our study.

10

Page 10 of 25

Methodology and empirical findings

d

M

an

us

cr

ip t

Table 2 describes, for each specific region of activities, cross-sectional mean changes in foreign profitability – measured as the operating earnings per year deflated by the total foreign assets held by the company within the same country/region. It should be emphasized that, like Herrmann (1996), who studied the predictive ability of geographic segment information in forecasting operating results, we study the disclosed earnings movements at three distinctive levels: at the consolidated, region and country level. Focusing on domestic versus foreign operations, we observe that on average both domestic and foreign changes in profitability are positive for the entire sample period. Interestingly, aggregated foreign profitability changes are systematically higher than domestic profitability movements from 1994 to 2000. From 2001 onwards, the reverse is observed. Profitability changes in firms’ domestic markets have been consistently stronger than foreign performance movements. When concentrating on foreign region-specific disclosures, we observe a variety of negative cross-sectional means. In Europe and Asia Pacific, profitability has increased over almost the entire sample – only 2005 and 2006 were characterized by an average downward trend. Country-specific disclosures, in contrast, clearly reveal the influences and implications of major macroeconomic events. For instance, Asian countries such as China, Korea, Indonesia and Papua New Guinea exhibit large profitability declines around the Asian financial crisis in 1997. We discern, too, the negative impact of other well-known crises: for instance, Mexico in 1994-1995, Argentina in 2001-2002 and the terrorist attacks in the United States in 2001. On the other hand, very prosperous years are witnessed in China (including Hong Kong) from 2003-2005.18

Ac ce p

te

Following the extensive literature on the value-relevance of accounting performance measures – originally documented by Ball and Brown (1968) and subsequently adapted to geographically segmented performance measures (see e.g. Boatsman et al., 1993; Bodnar and Weintrop, 1997; Francis and Schipper, 1999) – we estimate a firm’s performance response using the following model:

∆Vi ,t = η0 + η0 DEi ,t + η1' ∆FEi ,t + η2' ∆FEi ,t + ε i ,t where ∆Vi ,t designates the change in firm i’s shareholder value, DE i,t firm i’s performance generated in its home country ∆DE i,t the change in firm i’s performance generated in its home , country, FE i,t firm i’s performance achieved through firm i’s foreign activities , and ∆ FE i ,t the change in firm i’s performance achieved through firm i’s foreign activities. The original methodological design has been improved to explore new research questions and to overcome previously reported weaknesses: We explore the impact of firm i’s geographical disclosures as originally reported and publicly announced by the firm, i.e. we keep the quality of the disclosures unaltered. 54% of our sample firms provide country-specific disclosures of their foreign operations. While country-specific data doubtlessly provide a crucial informational content to explore the value-relevance of 18

Important oil discoveries were made in China and Hong Kong from 2003-2005.

11

Page 11 of 25

d

M

an

us

cr

ip t

specific country-related performance disclosures, it is surprising that existing studies have ignored the value-relevance of these detailed foreign performance disclosures.19 Following an extensive literature (Boatsman et al., 1993; Bodnar and Weintrop, 1997; Thomas, 2000b; Christophe, 2002; Bodnar et al., 2003a; Hope and Kang, 2005), movements in geographically segmented performance are defined as country- or region-specific operating earnings scaled by the market value of the firm, respectively changes in country- or regionspecific operating earnings scaled by the market value of the firm. This intuitively sound specification is more appropriate than alternative performance measures focusing either on raw operating earnings (Hughes and Ricks, 1987) or on the ratio between a firm’s foreign earnings and total consolidated accounting data, such as its total assets or its equity book value (Thomas, 2000a; Khurana et al., 2003; Thomas, 2004; Callen et al., 2005). This study controls for intertemporal macroeconomic influences by including year dummies. As firms in our sample are active in a large range of industry sectors, being hence quite heterogeneous, industry effects are controlled for by the addition of sector dummies. Finally, we incorporate country dummies to verify whether the country of origin of firm i significantly affects our estimation. A large variety of robustness checks are performed. The model is consecutively estimated using total stock returns and market risk-adjusted cumulative returns (Boatsman et al., 1993; Bodnar and Weintrop, 1997). The independent variable is first defined as the change in disclosed performance. Subsequently it is defined as the level of disclosed performance (Bodnar et al., 2003a). Finally our empirical specifications are verified for different lag structures (e.g. Christophe, 2002) and we distinguish between dividend-paying-stocks and non-dividend-paying stocks to investigate the potentially competitive signal conveyed by dividends.

Ac ce p

te

The results from estimating the value-relevance of country- versus region-specific geographic disclosures are presented in Table 3. First, total stock returns are regressed on domestic and aggregate foreign earnings movements. Results are reported in the upper panel of table 3 (Panel A). 20 It investigates for the total sample the value impact of domestic versus foreign operations in the aggregate. Significant positive association coefficients are indicated for changes in both domestic and foreign earnings. It should be noted that, in the total firm sample, we do not reject the null hypothesis of equality between these response coefficients. The impact of disclosed earnings levels are remarkably less significant. Given that the disclosed earnings level variables enter the model to assess the impact of the transitory component of earnings (Ali and Zarowin, 1992), the weak significance of the association coefficients between returns and these earnings level variables suggests that the incremental explanatory power of the transitory component of earnings remains low. A potential criticism of the above findings is that total stock returns may be driven by external factors that are not specifically related to the performance of firms’ domestic or foreign activities. Other forces such as local market movements may be dominating and failure to 19

If firms provide simultaneously country- and region-specific performance disclosures, we test the effect of their disclosed region-specific performance when they report aggregate data and also verify the impact of their disaggregated country-specific performance when they disclose specific data. 20 Since the objective of this study is to investigate how far domestic versus foreign profitability disclosures affect firm value, we do not display year, industry and country dummy coefficients. These are available from the authors upon request. It should be noted that while year and industry effects are statistically significant, no significant country effects could be detected. As a consequence, country dummies are not included in subsequent model specifications.

12

Page 12 of 25

Ac ce p

te

d

M

an

us

cr

ip t

include them in our model could hence result in exaggerated estimates of our response coefficients. To consider this concern, we calculate for each individual company risk-adjusted cumulative abnormal returns and estimate the sensitivity of these cumulative abnormal returns to changes in domestic and aggregate foreign earnings. Overall the value-impact of domestic versus foreign earnings movements remains unchanged. However, interestingly, the impact of domestic earnings levels on cumulative abnormal returns is significantly positive. The significance of both domestic association coefficients reveals hence that both the permanent and the transitory components of domestic earnings matter when we try to explain firms’ cumulative abnormal returns levels. 21 With respect to the variables that are of primary interest for this study22, we consider in Panel B the relevance of region-specific disclosures in terms of shareholder wealth. In line with previous empirical findings (Boatsman et al., 1993; Thomas, 2000b; Christophe and Pfeiffer, 2002), results are mostly inconclusive and mixed. First, it should be stressed that the adjusted R² declines to 13.04%, compared to 13.64% when using aggregate foreign performance measures. The response coefficient to earnings levels generated by foreign activities in Latin America (resp Asia Pacific) is significantly positive at the 5% (resp. 10%) level. Surprisingly, the response coefficient to earnings levels generated by foreign activities in Europe is significantly negative at the 5% level On the other hand it is puzzling to observe that the link between earnings level changes in Asia Pacific and Europe (resp. between earnings level changes in Latin America) and firm returns is significantly positive (resp. negative). It should be stressed, moreover, that both the magnitude and sign of these response coefficients vary notably when studying the impact of foreign earnings disclosures on cumulative abnormal returns. Panel C of Table 3 reports the results of our regressions when country-specific disclosures are investigated. As expected, the use of disaggregated country-specific performance movements increases the precision and significance of the response coefficients. Several observations are noteworthy: First of all the adjusted R2 of both the regression estimation using returns and the regression estimation using cumulative abnormal returns as dependent variable increase when compared to results observed in Panel A and Panel B. Moreover, Western continental European and North American countries, such as Germany and the United States, exhibit relatively small but strong positive response coefficients to disclosed earnings changes. We cannot find significant differences in these response coefficients. This homogeneity leads us to gather these countries together in Table 3.23 It should, furthermore, be stressed that the response coefficient of this cluster is not statistically different from the domestic market response (Chi-squared=0.78, p-value=0.38). News about performance changes from developed Anglo-Saxon European markets, such as the United Kingdom and Ireland, also have a positive and strongly significant impact. However, this influence is statistically larger compared to other European or Northern American countries. The United Kingdom and Ireland are therefore presented as a separate cluster. One possible explanation for these results may be the fact that the United Kingdom remains the top destination for inward foreign R&D investment and projects.24 The fact that, 21

Please note that the impact of both domestic earnings levels and domestic eanings changes on firms’ cumulative abnormal returns remains significantly positive in Panel B and Panel C. 22 Year and industry dummy coefficients are not displayed in Table 4. These are available from the authors upon request. 23 For the null hypothesis of equality between continental European countries vs. USA and Canada: Chisquared=1.48, p-value=0.22. Wald tests for each country do not show any evidence of statistical differences in the coefficients. These results are available upon request. 24 UK Trade and Investment (UKTI).

13

Page 13 of 25

25

Ac ce p

te

d

M

an

us

cr

ip t

among European countries, two separate – heterogeneous – clusters are observed explains why Europe as an aggregate region has a puzzling impact on firms’ market value in Panel B. Regarding Latin American countries, positive earnings change disclosures from Argentina, Brazil, Ecuador and Mexico are shown to be significantly negatively affecting firm value while good performance news from Chile is positively perceived – though not significantly so.25 Wald tests (Chi-squared=6.14, p-value=0.19) suggest that we can cluster these countries together. These findings, at the country level, support results reported in Boatsman et al. (1993) and Thomas (2000b) on regional disclosures. Indeed, these authors find a negative response coefficient for the South American segment.26 The puzzling negative value-impact of positive performance news emanating from firms’ permanent component of earnings in South America seems hence to be persistent over time. Thomas (2000b) argues that the negative response coefficient may be attributable to the high operating risk in this area. The negative sign of these coefficients may be interpreted as being related to an indicator that leads investors to revise their pricing downwards: higher earnings in these emerging markets may be associated by investors with a higher than expected portion of riskier and lower-valued cash-flows and thus a lower than expected portion of cash-flows originating from less risky countries (Eaker et al., 1996; Reeb et al., 1998). Moreover, investors may question the accounting information related to operations in Latin American countries (Rask et al., 1998) given greater asymmetric information problems associated with emerging markets (Wet, 2004) and particularly with Latin American markets (Chollete et al., 2012a, 2012b). It should be noted that the negative value-impact of Mexican news may also explain why the value-impact of North America, as an aggregate region, is not significant (Panel B). The response coefficient of Mexico being strongly negative in contrast with the positive coefficient of the USA, it is obvious that investors are not able to discern the informational content and the value impact of performance changes in the aggregate North American region.27 It should be stressed that, overall, our results are in line with the functional fixation hypothesis, which argues that investors “fixate” on reported (consolidated) net income and react to specific information in earnings components with a delay (Hand, 1990; Tiniç, 1990; Sloan, 1996; Harris and Nissim, 2006). The response coefficients to earnings changes for Asian Pacific countries such as China, Papua New Guinea, Japan, and Australia are all positive, high and strongly significant at the 1% level. For these countries, we cannot reject the null hypothesis of equality between the estimated coefficients (Chi-squared=8.24, p-value=0.41).28 It seems that news originating from each country of Asia Pacific is relatively highly priced.29 The performance disclosures form hence The positive impact of a firm’s performance in Chile is interesting. Chile is in fact one of the few Latin American countries that has reported solid growth for decades and has achieved a solid investment-grade credit rating while still suffering from many of the political weaknesses that are often held up to explain Latin America’s economic underperformance. 26 Their coefficient was nevertheless insignificant at the regional level. 27 The expression “North America” may be interpreted differently by different people according to the context. As a region, North America may exclusively include the USA and Canada. In a different context, North America may include Mexico and other countries. Of course, this heterogeneity in the definition of the North American region may strengthen investors’ muddle. 28 It is also interesting to note that the coefficients associated with emerging Asian countries are found to be statistically identical to those related to developed Asian countries. 29 The sharply contrasting results found for Latin America and Asia deserve further discussion. Hillebrand (2003) argues that the development of a unified system of world trade – together with the rise of the East Asian countries (China in particular) – has intensified competition between countries of the South. Asian countries (led by China) capture larger and larger shares of foreign investments, export markets and even enter Latin American national

14

Page 14 of 25

Ac ce p

te

d

M

an

us

cr

ip t

quite a homogeneous cluster, whose response (3.3137 significantly different from zero at the 5% significance level) is similar to the results related to the Asian Pacific region-specific disclosure (2.8229 significantly different from zero at the 5% significance level) reported in Panel B. This outcome is not surprising, since a predominant part of all foreign direct investment to developing countries goes to China (including Hong Kong) during our sample period.30 It is, moreover, natural to observe that Australasian countries (e.g. Australia and New Zealand) are quite heavily enmeshed in economic relations with South-East Asia.31 Our findings suggest that country-specific disclosures increase the precision and significance of the response coefficients. The adjusted R² climbs to 15.01%. If foreign earnings disclosures that originate from a range of distinctive markets whose stability and growth opportunities are extremely disparate and differentially perceived by investors are added up in regional performance measures (as in Panel B), the aggregate disclosures may be too coarse and uninformative. For investors, the aggregation of heterogeneous effects leads inevitably to a difficult interpretation of the information and to insignificant response coefficients. For Europe, North America and Latin America, this study solves the puzzle raised in previous literature (Boatsman et al., 1993; Bodnar and Weintrop, 1997; Thomas, 2000b; Christophe and Pfeiffer, 2002; Bodnar et al., 2003a) and demonstrates that these regions are too heterogeneous to be aggregated in pre-defined geographic clusters.32 Overall our results suggest that the association between market value and foreign accounting earnings documented in previous literature should be qualified and could be mainly driven by specific regions or countries. The two last columns of Table 3 present additional tests of the information content of regionand country-specific disclosures when risk-adjusted abnormal returns are used rather than total returns. With respect to the results discussed above, the magnitude of the impact of earnings change disclosures emanating from activities in the Asian Pacific region appears to be larger (Panel B) when compared to the impact of the same variable on raw returns. On the other hand findings concerning the valuation impact of earnings change disclosures related to operations in developed countries, such as continental Europe and North America, support the results reported in the first two columns. While overall, relative to our findings displayed in the first two columns, significance levels tend to be notably weaker, it should be stressed that the comparison between Panel B and Panel C reveals that findings based on country-specific earnings disclosures (Panel C) tend to be more consistent and robust (in terms of magnitudes and signs) across both model specifications than results reported in Panel B. Several robustness checks are added. Results are disaggregated according to firms’ country of origin. It appears that the relative value-relevance of foreign versus domestic performance varies markets. Hillebrand shows, moreover, that type and orientation of foreign investments differ between these two regions. While in China a large share of foreign investments is directed towards Greenfield investments in the manufacturing sector, in Latin America a major share of foreign investment goes to company acquisitions and privatizations. According to Hillebrand, in Latin America, foreign investments have tended to crowd out domestic investors. 30 See Dicken, 2003. Global Shift: Reshaping the Global Economic Map in the 21st Century, London (Sage), 4th Edition, p.61. 31 Australia has one of the highest rates of economic growth in OECD countries. See Australian Government Department of Foreign Affairs and Trade. 32 We find that Wald tests fail to reject the equality of the coefficients for all Asian countries and UK/Ireland (Chisquared=0.94, p-value=0.33). However, for obvious geographical reasons, we decide to keep these two clusters distinct when presenting Panel C.

15

Page 15 of 25

cr

ip t

across countries of origin. For companies located in the United States and Europe, the response coefficient of foreign earnings is less significant than for the rest of our sample. By contrast, the value-impact of domestic earnings is not significant for Canadian firms. Further robustness checks are performed. We run both sets of regressions by specifying the dependent variable as annual stock returns coinciding with the firm’s fiscal year (without lags). Cumulative abnormal returns are similarly re-specified. We also test the robustness of our findings by focusing exclusively on the impact of disclosed earnings changes on firm value (without including earning levels). Finally, in order to investigate the potentially competitive signals conveyed by dividends concerning the performance of foreign operations, we form twosub samples: we distinguish between no-dividend-paying stock and dividend-paying companies. Overall the findings are consistent and our conclusions remain unaffected.

us

Concluding remarks

Ac ce p

te

d

M

an

This paper is devoted to investigating the relative value-relevance of foreign earnings disclosures. While the central issue of whether and why foreign performance disclosures are differentially associated with shareholder wealth has remained unsolved for decades, the present study sheds new light on the value-relevance of foreign geographically segmented performance disclosures by investigating the huge amount of information contained in firms’ country-specific disclosures. On an extensive dataset consisting of 2,018 firm-year observations from 1993 to 2006, we perform a thorough examination of the value-impact of segmented accounting disclosures related to foreign operations in the aggregate, related to foreign region-specific operations and to foreign country-specific operations. Overall the sign, size and statistical significance of reported foreign earnings response coefficients are unambiguous and solve the long-lasting debate on the relative value-relevance of foreign performance disclosures: (i) Performance disclosures of both domestic and foreign operations in the aggregate are shown to be positively linked to firm value. The relative importance of these relationships remains however inconclusive. Empirical evidence does not enable us to conclude whether firm value is more or less sensitive to foreign versus domestic performance. (ii) The value-impact of too aggregated foreign performance news – aggregating earnings movements from a large and heterogeneous variety of economic and institutional environments – turns out to be statistically weaker and less consistent across distinctive model specifications. It is hence reasonable to conclude that if the market is confronted with performance news disclosed at a heterogeneous region-level, the market fails to understand the information and does not properly assess its impact on firm value. (iii) By explicitly testing the impact of country-specific performance disclosures on firm value, we observe that findings are significantly more consistent. Detailed country-specific response coefficients enable us to explain why previous research generated mixed evidence. When foreign earnings disclosures emanating from activities in Europe, whose stability and growth opportunities are extremely disparate, are for instance added up in a European regional performance measure, results demonstrate that aggregate disclosures turn out to be too coarse and uninformative. In contrast, the country-specific performance disclosures help us to identify and disentangle the differential valuation impacts – and hence to solve the puzzles debated in previous studies (Boatsman et al., 1993; Bodnar and Weintrop, 1997; Thomas, 2000b; Christophe and Pfeiffer, 2002; Bodnar et al., 2003a). (iv) Specifically, we observe that an increase in earnings related to operations in Latin American economies is associated with a

16

Page 16 of 25

an

us

cr

ip t

negative impact on shareholder wealth, suggesting that investors interpret this increase in earnings as a higher than expected proportion of low-valued income. Investors seem hence to be consistently doubtful towards operations in more vulnerable, less stable and less regulated markets. The impact of performance disclosures emanating from activities in Asian Pacific countries and in the United Kingdom appears to be the highest compared to any other geographic area, suggesting the attractiveness of Asia Pacific and the United Kingdom as top foreign direct investment destinations. The exploration of both foreign region-specific and foreign country-specific segment data enable us to confront previously reported findings with new empirical results that are demonstrated to be robust to several alternative methodological designs and that consistently reveal the crucial importance of disaggregation. Empirical findings using disaggregated data do not only provide an explanation for the weak and conflicting evidence reported in previous empirical work but help us as well as to understand how investors interpret foreign performance news originating from a large and heterogeneous variety of economic and institutional environments and how they incorporate these news into firm value.

Ac ce p

te

d

M

References Ali, A., Zarowin, P., 1992. The role of earnings levels in annual earnings-returns studies. Journal of Accounting Research 30(2), 286-296. Anderson, C.W., Fedenia, M., Hirschey, M., Skiba, H., 2011. Cultural influences on home bias and international diversification by institutional investors. Journal of Banking and Finance 35, 916-934. Antonczyk, R.C., Salzmann, A.J., 2013. Overconfidence and optimism: The effect of national culture on capital structure. Research in International Business and Finance, in press. Ball, R., Brown, P., 1968. An empirical evaluation of accounting income numbers. Journal of Accounting Research 6, 159-178. Berger, P. G., Hann, R., 2003. The impact of SFAS No. 131 on information and monitoring. Journal of Accounting Research 41, 163-223. Boatsman, J., Behn, B., Patz, D., 1993. A test of the use of geographical segment disclosures. Journal of Accounting Research 31, 46-64. Bodnar, G.M., Hwang, L., Weintrop, J., 2003a. The value relevance of foreign income: an Australian, Canadian, and British comparison. Journal of International Financial Management & Accounting 14, 171-193. Bodnar, G. M., Tang, C., Weintrop, J., 2003b. The value of corporate international diversification. NBER Working Paper No. W6224. Bodnar, G., Weintrop, J., 1997. The valuation of foreign income of U.S. multinational firms: A growth opportunities perspective. Journal of Accounting and Economics 24, 69-98. Bugeja, M., Czernkowski, R., Moran D., 2015. The impact of the management approach on segment reporting. Journal of Business Finance and Accounting, forthcoming. Callen, J., Hope, O., Segal, D., 2005. Domestic and foreign earnings, stock return variability, and the impact of investor sophistication. Journal of Accounting Research 43, 377-412. Campello, M;, Graham, J.R., Campbell R.H., 2010. The real effects of financial constraints: Evidence from a financial crisis. Journal of Financial Economics 97(3), 470-487. Caves, R., 1971. International corporations: the industrial economics of foreign investment. Economica 38, 1-27.

17

Page 17 of 25

Ac ce p

te

d

M

an

us

cr

ip t

Chollete, L., De la Pena, V., Lu, C., 2012a. International diversification: An extreme value approach. Journal of Banking and Finance 36, 871-885. Chollete, L., De la Pena, V., Lu, C., 2012b. International diversification: A copula approach. Journal of Banking and Finance 35, 403-417. Christophe, S., 1997. Hysteresis and the value of the U.S. multinational corporation. Journal of Business 70, 435-462. Christophe, S., 2002. The value of U.S. MNC earnings changes from foreign and domestic operations, Journal of Business 75, 67-93. Christophe, S., Pfeiffer, R., 2002. The valuation of MNC international operations during the 1990s. Review of Quantitative Finance and Accounting 18, 119-138. Collins, J., 1990. A market performance comparison of U.S. firms active in domestic, developed and developing countries. Journal of International Business Studies 21, 271-287. Core, J. E., Guay, W. R., Buskirk, A. V., 2003. Market valuations in the new economy: An investigation of what has changed. Journal of Accounting and Economics 34, 43-67. Denis, D., Denis, D., Yost, K., 2002. Global Diversification, Industrial Diversification, and Firm Value. Journal of Finance 57, 1951-1979. Dicken, P., 2003. Global Shift: Reshaping the Global Economic Map in the 21st Century, Sage, 4th Edition. Doupnik, T., Seese, 2001. Geographic area disclosures under SFAS 131: materiality and fineness. Journal of International Accounting, Auditing, and Taxation 10, 117-138. Durnev, A., Morck, R., Yeung, B., 2004. Value-enhancing capital budgeting and firm-specific stock return variation. Journal of Finance 59(1), 65-105. Duru, A., Reeb, D., 2002. International diversification and analysts’ forecast accuracy and bias. The Accounting Review 77, 415-433. Eaker, M., Fabozzi, F., Grant, D., 1996. International Corporate Finance. Harcourt Brace College Publishers. Errunza, V., Senbet, L., 1984. International corporate diversification, market valuation, and sizeadjusted evidence. Journal of Finance, 39, 727-745. Errunza, V., Senbet, L., 1981. The effects of international operations on the market value of the firm: theory and evidence. Journal of Finance 36, 401-417. Francis, J., Schipper, K., 1999. Have financial statements lost their relevance? Journal of Accounting Research 37, 319-353. Garrod, N., Rees, W., 1998. International diversification and firm value. Journal of Business Finance and Accounting 25(9), 1255-1281. Gelb, D. S., 2000. Corporate signaling with dividends, stock repurchases and accounting disclosures: An empirical study. Journal of Accounting Auditing and Finance 15, 99-120. Graham, J., Harvey, C., Rajgopal, S., 2005. The economic implications of corporate financial reporting. Journal of Accounting and Economics 40, 3-73. Hand, J., 1990. A test of the extended functional fixation hypothesis. The Accounting Review, 65, 740-763. Harris, T., Nissim, D., 2006. The differential value implications of the profitability and investment components of earnings. Working paper. Herrmann, D., 1996. The predictive ability of geographic segment information at the country, continent, and consolidated levels. Journal of International Financial Management and Accounting 7, 50-73.

18

Page 18 of 25

Ac ce p

te

d

M

an

us

cr

ip t

Herrmann, D., Thomas, W., 2000. An analysis of segment disclosures under SFAS no 131 and SFAS no 14. Accounting Horizons, 287-302. Hillebrand, E., 2003. South-South competition: Asia versus Latin America?, Friedrich-EbertStiftung, Dialogue on Globalization: conference reports. Hines, J., Rice, E., 1994. Fiscal paradise: foreign tax havens and American business. Quarterly Journal of Economics 109, 149-182. Hope, O., Kang, T., 2005. The role of “other information” in the valuation of foreign income for U.S. multinationals. Journal of Accounting, Auditing & Finance 20, 355-377. Hope, O., Kang, T., Thomas, A. B., Vasvari, F., 2009. The effects of SFAS 131 Geographic Segments disclosures on the valuation of foreign earnings. Journal of International Business Studies,40(3), 421-443 . Hossain, M., 2008. Change in value relevance of quarterly foreign sales data of U.S. multinational corporations after adopting SFAS 131. Review of Quantitative Finance and Accounting 30, 1-23. Hughes, J., Ricks, W., 1987. Associations between forecast errors and excess returns near to earnings announcements. Accounting Review 62, 158-175. Hussain, S., Skerratt, L., 1992. Gains from disaggregation and the definition of a segment: a note on SSAP 25. Accounting and Business Research 22, 370-376. Iatridis, G., Dimitras, A.I., 2013. Financial crisis and accounting quality: Evidence from five European countries, Advances in Accounting 29(1), 154-160. Iatridis, G. Valahi, S., 2010. Voluntary IAS 1 accounting disclosures prior to official IAS adoption: An empirical investigation of UK firms. Research in International Business and Finance 24, 1–14. Jensen, M., 1986. Agency costs of free cash flow, corporate finance, and takeovers. American Economic Review 76, 323-329. Kousenidis, D.V., Ladas, A.C., Negakis, C.I., 2013. The effects of the European debt crisis on earnings quality. International Review of Financial Analysis 30, 351-362. Khurana, I. K., Pereira, R., Raman, K. K., 2003. Does analyst behavior explain market mispricing of foreign earnings for U.S. multinational firms? Journal of Accounting, Auditing and Finance 18(4), 453-478. Kim, W., Lyn, E., 1986. Excess market value, the multinational corporation, and Tobin’s q-ratio. Journal of International Business Studies 17, 110-125. Kogut, B., 1983. Foreign direct investment as a sequential process. In: The Multinational Corporations in the 1980s, edited by C. Kindelberger. MIT Press. Kothari, S., Shanken, J., 2003. Time-series coefficient variation in value-relevance regressions: A discussion of Core, Guay and Van Buskirk and new evidence. Journal of Accounting and Economics 34, 69-87. Kou, W., Hussain, S., 2007. Predictive gains to segmental disclosure matrices, geographic information and industry sector comparability. The British Accounting Review 39, 183-195. Kwok, C., Reeb, D., 2000. Internationalization and firm risk: an upstream-downstream hypothesis. Journal of International Business Studies 31, 611-629. Lim, T., 2001. Rationality and analysts’ bias. Journal of Finance 56, 369-385. Morck, R., Yeung, B., 1991. Why investors value multinationality. Journal of Business 64, 165187.

19

Page 19 of 25

Ac ce p

te

d

M

an

us

cr

ip t

Moumen, N., Ben Othman, H., Hussainey, K., 2015. The value relevance of risk disclosure in annual reports: Evidence from MENA emerging markets. Research in International Business and Finance 34, 177–204. Nichols, D., Street, D., Tarca, A., 2013. The impact of segment reporting under the IFRS 8 and SFAS 131 Management Approach: a research review. Journal of International Financial Management and Accounting 24, 261-312. Nichols, D., Street, D., Gray, S., 2000. Geographic segment disclosures in the United States: reporting practices enter a new area. Journal of International Accounting, Auditing, and Taxation 9, 59-82. Nichols, D., Tunnell, L., Waldrup, B., 1996. An exploratory study of the effect on forecast accuracy of using different geographic segment data sources. Journal of International Financial Management and Accounting 7, 125-136. Radebaugh, L., Gray, S., Black, E., 2006. International Accounting and Multinational Enterprises, 6th Edition, John Wiley & Sons. Rask, R., Chu, D., Gottschang, T., 1998. Institutional change in transitional economics: The case of accounting in China. Comparative Economic Studies 40(4), 76-100. Reeb, D., Kwok, C., Baek, H., 1998. Systematic risk of the multinational corporation. Journal of International Business Studies 29, 263-279. Sloan, R., 1996. Do stock prices fully reflect information in accruals and cash flows about future earnings?. The Accounting Review, 71, 289-315. Street, D., Nichols, N., Gray, S., 2000. Segment disclosures under SFAS No131: has business segment reporting improved. Accounting Horizons 14, 259-285. Thomas, W., 2004. Firm characteristics and the pricing of foreign earnings of US multinational firms. Journal of International Financial Management and Accounting 15, 145-173. Thomas, W., 2000a. A test of the market's mispricing of domestic and foreign earnings. Journal of Accounting and Economics 28, 243-267. Thomas, W., 2000b. The value-relevance of geographic segment earnings disclosures under SFAS 14. Journal of International Financial Management and Accounting 11, 133-155. Tiniç, S., 1990. A perspective on the stock market’s fixation on accounting numbers. The Accounting Review 65, 781-796. Wet, W. A., 2004. The role of asymmetric information on investments in emerging markets. Economic Modelling 21, 621-630. White, H., 1980. A heteroscedasticity-consistent co-variance matrix estimator and a direct test for heteroscedasticity. Econometrica 48, 817-838. Wurgler, J., 2000. Financial markets and the allocation of capital. Journal of Financial Economics 58(1-2), 187-214.

20

Page 20 of 25

Table 1: Sample structure

0 0 21 0 11 0 0 0

10 4 0 2 7 2 0 0

926

1,332

32

25

161 249 207 154

31 60 243 71

17 6 18 7

332 51 422 121

128 126 874 204

TOTAL

2,018

771

405

48

926

347 263 187 166 503 271 263 18

131 90 84 56 224 119 61 6

68 88 56 26 64 94 5 4

4 10 13 0 6 15 0 0

2,018

771

405

48

TOTAL

us

an

M

Consumer cyclical Consumer non cyclical Energy Healthcare Industrials Materials Technology Utilities

cr

370 328 1,085 235

TOTAL

Eastern Europe

219 169 71 122 371 142 231 7

Australia/New Zealand Canada USA Europe

ip t

Middle East

201 100 43 42 238 134 158 10

Europe

25

Asia (Pacific)

32

Africa

1,332

Latin America

0 1 6 18

North America

5 10 17 0

Domestic Industry of origin

Country of origin

Regions of destination

1,044 831 2,872 810

980 724 475 414 1,424 777 718 45

Ac ce p

te

d

Note: – This table reports across countries of origin, types of industry and areas of operations the number of firmyears satisfying the following conditions: (1) The firm reports foreign activities at the region and/or country-level in its financial statements. (2) The firm discloses detailed geographical segmental information, i.e. it does not aggregate foreign operations into a single disclosure labeled "international" or "foreign". (3) The firm is not active in the financial industry sector. (4) Geographic segmented data (i.e. operating earnings) are disclosed. (5) At least two consecutive years of region- respectively country-specific accounting data must be observed, as changes in earnings have to be estimated.

32

Page 21 of 25

Table 2: Mean changes in domestic, foreign, foreign region-specific and foreign countryspecific profitability from 1994 to 2006 1994

1995

1996

1997

1998

1999

2000

2001

2002

2003

2004

2005

2006

0.32 0.28

0.36 0.13

0.63 0.61

0.22 0.14

0.57 0.03

-3.13 -1.14 1.19 0.34 0.07 0.48 0.26

2.11 0.76 0.17 0.28 0.04 -0.09 1.01

0.32 0.85 0.25 -0.06 0.60 -0.13 -0.15

-0.28 -0.10 0.67 -0.27 -0.62 0.18 2.01

-0.40 -0.26 -0.21 0.05 -0.42 0.10 1.45 0.98 1.78 -0.05 0.25 -0.96 -1.08 -0.13 2.53 -0.84 3.00 1.53 0.21 0.93 0.56 -0.06 23.35 -0.49 0.73 -4.48 0.78 3.56 -2.72 0.16 -0.18 -0.11 0.08 2.89 -0.08 -0.39 -0.41 -0.11 -0.66 -4.44 1.23 0.02 0.53 28.93 3.33 39.60 -0.05 -0.36 -0.06 0.37 -0.14 -0.23 1.60 -0.68 -0.29 -1.55 -0.08 0.07 0.20 0.14 0.82 0.51 -0.06 0.10 -0.24 0.02 0.65 0.69 0.06 0.14 0.38 0.09 -

-0.29 -0.45 -3.14 0.92 0.01 -0.38 1.16 -0.56 3.77 0.39 0.21 2.07 -0.19 -

Domestic Foreign

0.88 0.89

0.34 0.44

0.25 0.26

0.00 0.18

0.26 0.38

0.21 0.62

0.49 1.01

0.10 -0.20

PANEL B: Foreign region-specific disclosures 0.19 -0.01 0.00 0.30 3.86 -0.41 1.76

-0.16 -1.86 2.51 0.72 1.62 -1.73 0.04

0.07 -0.07 -0.27 1.91 2.12 2.05 -

0.06 -0.45 0.30 3.12 0.57 4.86 -

0.32 0.63 0.55 1.71 -0.05 13.38 4.51 1.76 0.55 0.98 1.86 0.92 1.62 -0.48

-0.76 -0.67 0.17 1.65 0.34 0.27 2.19

2.57 1.34 -0.26 0.21 0.41 -2.38 4.65

cr

0.50 0.60 0.63 2.55 0.74 -6.75 -

us

North America Latin America Africa Asia (Pacific) Europe Middle East Eastern Europe

ip t

PANEL A: Domestic versus foreign disclosures

PANEL C: Foreign country-specific disclosures

an

-0.35 0.30 0.22 0.46 -0.73 2.53 1.58 0.81 1.48 -0.16 -0.01 0.79 0.15 -19.10 -0.11 -4.73 0.89 0.24 2.80 1.17 -0.44 -2.20 0.21 1.18 1.23 0.08 1.08 2.25 -0.28 1.11 -0.24 -0.66 0.88 13.78 -0.21 5.03 -0.33 -0.93 -0.08 0.87 -0.79 0.26 -0.03 1.88 -0.23 -0.26 0.69 2.27 0.63 -0.57 1.65 -0.65 0.54 -0.31 0.47 0.06 0.31 -0.10 0.63 17.06 1.16 -6.81 -0.25 0.07 0.40 0.40 4.03 -1.14 -2.84 -

M

0.69 0.03 0.23 1.58 -11.33 0.04 0.09 -1.05 -0.07 0.48 1.40 0.81 1.62 -0.01 -0.02 -0.19 -0.74 0.25 -17.42 2.51 8.31 1.64 0.95 -7.33 1.02 0.75 0.68 -0.76 -0.52 0.44 0.35 0.16 -0.45 0.36 -0.71 -0.04 0.12 -0.69 -2.55

d

0.05 -7.06 -0.13 -0.46 1.43 -0.21 -0.26 0.07 0.47 -0.56 0.85 0.98 0.07 4.31 0.00 -0.26

te

0.37 0.39 2.17 0.39 0.58 1.88 0.21 0.00 1.15 -0.26 -0.30 8.57 0.56 0.34 0.70

Ac ce p

France Germany Ireland Spain UK USA Canada Argentina Brazil Chile Ecuador Mexico South Africa China Hong Kong Japan (South) Korea Yemen Australia Indonesia New Zealand Papua N. Guinea

Note: – Panel A of this table reports over time the cross-sectional mean statistics for the changes in domestic versus foreign profitability for all sample companies (390 companies). Panel B of this table reports over time the crosssectional mean statistics for the changes in foreign region-specific profitability for the 324 companies disclosing region-specific profitability information. Panel C of this table reports over time the cross-sectional mean statistics for the changes in foreign country-specific profitability for the 214 companies disclosing country-specific profitability information. Annual profitability is calculated as the ratio between the pre-tax operating income within the country/region and the total foreign assets held by the company within the same country/region.

33

Page 22 of 25

Ac ce p

te

d

M

an

us

cr

ip t

Table 3: Value relevance of changes in domestic versus foreign earnings disclosures, domestic versus foreign region-specific earnings disclosures and domestic versus foreign country-specific earnings disclosures

34

Page 23 of 25

Ac ce p

te

d

M

an

us

cr

ip t

Table 3: cont’d

Note: – This table reports the valuation impact of disclosed changes in domestic, foreign, foreign region-specific and foreign country-specific earnings. The value-relevance is measured in terms of total returns – respectively cumulative abnormal returns – using the following model: 7

13

DEi ,t

p

t

MVi ,t 1

Vi ,t    ip Industry p   t Yeart  1

  '1

DEi ,t MVi ,t 1

s

FSEij ,t

j 1

MVi ,t 1

   j 1

s

FSEij ,t

j 1

MVi ,t 1

   ' j 1

  i ,t

where Vi ,t is either the percentage change in the firm i‟s adjusted stock price on an annual interval beginning at the end of the first quarter of the firm‟s fiscal year t and ending at the end of the first quarter of the next fiscal year t+1 (first two results columns), or the cumulative abnormal return on an annual interval coinciding with the firm‟s fiscal year t (last two results columns). Cumulative abnormal returns are computed as CARi ,t   1  Ri ,k   i   i RWM ,k   1 12

k 1

where Ri , k is firm i‟s monthly stock return in month k of fiscal year t, RW M,k is the MSCI World index monthly return in month k of fiscal year t, and

i

and

i

are the coefficients from the OLS estimation over the 36 months

prior to the beginning of fiscal year t.

35

Page 24 of 25

Ac ce p

te

d

M

an

us

cr

ip t

Industry refers to sector dummies equal to one for the proper industry of firm i, zero otherwise, Year refers to the year dummies equal to one for the corresponding year t and zero otherwise. DEi,t is the level of domestic earnings, defined as the operating earnings generated in firm i‟s home country, from fiscal year t-1 to fiscal year t. DEi,t is the change in domestic earnings, defined as the operating earnings generated in firm i‟s home country, from fiscal year t-1 to fiscal year t. FSEij,t is firm i‟s performance achieved through firm i‟s foreign activities from region (country) j, defined as the as-disclosed operating earnings generated by firm i in the region (country) j, from fiscal year t-1 to fiscal year t. FSEij,t is the change in specific foreign earnings from region (country) j, defined as the as-disclosed operating earnings generated by firm i in the region (country) j, from fiscal year t-1 to fiscal year t. MVi,t-1 is firm i's market value at the beginning of the computed annual interval. The F-stat is a test of the null hypothesis that the earnings coefficients are equal to zero. P-values are provided in parentheses. The null hypothesis of the Wald test is the equality of the estimated coefficients for the regions/countries. P-values are provided in parentheses. In Panel C, geographic clusters are formed for countries from North America, continental Europe, Latin America and Asia Pacific exhibiting similar effects on equity valuation, i.e. those for which we do not find statistically different coefficients at the 10% level. *, **, *** indicate that the estimated coefficients are statistically different from zero at the 10%, 5% and 1% level respectively. White‟s (1980) heteroskedasticity-consistent standard errors are also exhibited.

36

Page 25 of 25