A research into the effect of agency costs on the level of forward looking disclosures : empirical evidence from U.S. multinationals on foreign operations

(1)

Amsterdam Business School

MSc Accountancy & Control

Master Thesis

A research into the effect of agency costs on the level of forward

looking disclosures: empirical evidence from U.S. multinationals

on foreign operations.

Name: Ms. S.M. Terwindt Student number: 10858989

Thesis supervisor: Dr. W. Janssen & Dr. S.W. Bissessur Date: June 20th₂₀₁₆

Word count: 14155

(2)

Statement of Originality

This document is written by student Sophie Terwindt who declares to take full responsibility for the contents of this document.

I declare that the text and the work presented in this document is original and that no sources other than those mentioned in the text and its references have been used in creating it.

The Faculty of Economics and Business is responsible solely for the supervision of completion of the work, not for the contents.

(3)

Abstract

The purpose of this study is to examine the effect of agency costs on the level voluntarily forward looking disclosures for foreign sales in multinationals of the United States (U.S.).

As proxy for agency costs the expense ratio and the asset utilization ratio is used. For the quantity of forward looking disclosures the forward looking sentences are extracted from the Management's Discussion and Analysis (MD&A) sections of 10-K filings, using

automated language processing techniques. The sample consists of 2090 U.S. multinational organizations from the years 1996 to 2000. A linear regression is used to measure the effect of agency costs on the quantity of forward looking disclosures for foreign sales.

There is a positive significant effect for expense ratio on the quantity of forward looking disclosures for foreign sales. This result states that managers withhold more forward looking disclosures when the agency costs are high for foreign sales. Managers withhold more forward looking bad news than good news when the agency costs are high for foreign sales. The management makes the decision to withhold forward looking bad news for their own interest and objectives. Managers tend to withhold forward looking disclosures with relatively low abnormal profits and/or high expenses, this is equal to a high expense ratio. In addition, managers avoided forward looking disclosures about the segment operations. Managers exploited the discretion afforded under SFAS No. 14 to opportunistically conceal bad news forward looking disclosures.

In difference with much of the prior researches, the focus in this study is on the quantity of the forward looking disclosures in the MD&A section of the annual reports. There is a gap in the literature about the effect of agency costs on level forward looking disclosures for international organizations on foreign sales. The contribution of this research is to investigate this effect in a unique setting of U.S. multinationals.

There are a few limitations in this research. The focus is only on foreign sales. Investigation on this topic might be extended by analysing diverse context. Finally, the evidence does not explain specifically why a low asset utilization ratio not lead to a higher level of withholding of forward looking disclosures.

Key words: forward looking disclosures, agency costs, expense ratio, asset utilization ratio,

(4)

1 Introduction

The discussion and background information about forward looking disclosures is described. Based on this paragraph the research question is defined. Finally, the contribution and the structure of this study is given.

1.1 Background

This research examines the effect of agency costs on voluntarily forward looking disclosures in multinationals of the United States (U.S.). These forward looking disclosures are based on the disclosures in the Management's Discussion and Analysis (MD&A) section of 10-K filings. Voluntary forward looking disclosures are still susceptible to manipulation which could lead to agency costs. The MD&A section is an important part of the annual report that includes voluntarily managerial commentary about a firm’s current state and future prospects. The MD&A section provides managers with an opportunity to convey their future expectations and strategic plans directly to the public. However, regulators and users have been critical of the quantity and quality of forward-looking disclosures in the MD&A section (Tavcar, 1998; Securities and Exchange Commission, 2003). The MD&A section of the annual report is a quasi-mandatory or a quasi-voluntary disclosure setting, it remains largely voluntary (Beyer, Cohen, Lys, & Walther, 2010).

The intention of the Securities and Exchange Commission (SEC) with the MD&A section is to provide useful and transparent information about the organization for investors. The section has to give periodically guidance about the disclosures in the MD&A section of the annual report (Securities and Exchange Commission, 1989). Investors have a greater need for forward looking disclosures than for disclosures about actions and choices in the past. The MD&A sections have guided organizations to be informative about trends, events, commitments, plans, and uncertainties that are likely to materially affect company liquidity, capital resources, or future operations (Securities and Exchange Commission, 1989; 2003). MD&A disclosures provide unique information to the capital market (Mayew & Venkatachalam, 2015). However, the SEC finds the disclosures of the MD&A sections deficient (Cole & Jones, 2005). The disclosures typically include Standard statements and immaterial details but not enough information about the matter itself (Securities and Exchange Commission, 2003).

(6)

Opponents argue that the management of organizations already voluntarily disclose sufficient information in the MD&A section of the 10-K. The instructive power of the MD&A content itself is almost the same magnitude as financial information. The language tone alone already provides a quarter part of the explanatory power of the forward looking disclosures in the MD&A sections. Not only the financial information but also the market information include more and probably different aspects of the information environment including financial variables, auditor reports, and nonfinancial information contained within a 10-K (Mayew & Venkatachalam, 2015).

The most recent research of forward looking disclosures in the MD&A section is performed by Muslu et al (2015). They suggest that organizations with more forward looking disclosures have poor information environments before the 10-K filing. Forward looking disclosures are related to improved information environments. The disclosures mitigate poor information environments. The poor information environments could be a cause of the agency theory (Muslu, Radhakrishnan, Subramanyam, & Lim, 2015). The manager reveals his information when it is either sufficiently high or sufficiently low, regardless of their own assessments of their valuation implications. Managers do not base their disclosure decision on their own interpretation of their private information, but rather on their prediction of how the market analyst will interpret it. Given that the analyst’s prior valuation of the firm was favourable, the manager is willing to disclose information he personally finds unfavourable for the organization, as the manager expects the analyst to interpret it in a favourable manner (Ball & Shivakumar, 2012).

The reasons by which managers can commit to truthful disclosure of private information are not well understood (Ball & Shivakumar, 2012). Research on forward looking disclosures has established such a role for management earnings predictions issued by US organizations (Baik & Jiang, 2006). Empirical evidence so far suggests that a higher frequency of forward looking disclosures does indeed correlate with stock prices that are more informative about future earnings. Therefore, as forward looking disclosures are not verifiable or auditable immediately, managers could also use them when they have incentives to be misleading or untruthful at the time of the disclosure. To guard against misleading performance disclosures, investors may look for information quality in the audited financial statements (Athanasakou & Hussainey, 2014).

According to Berger and Hahn (2007) there are two main reasons for why managers withhold segment related news, agency and proprietary costs.

(7)

This study examines the effect of agency costs on forward looking disclosures in U.S. multinationals on foreign sales. Managers face agency costs of segment disclosure if the revelation of a segment that earns low abnormal profits reveals unresolved agency problems and, hence, leads to heightened external monitoring. Managers may want to withhold bad news forward looking disclosures if this news reveals negative consequences of managers’ investment strategy and actions. Therefore, the managers may want to withhold good news forward looking disclosures if others, such as competitors, can use the information to extract profits from the disclosing organization (Berger & Hahn, 2007). In the recent financial crisis, a lot of organizations were not able to raise new funds from investors. To raise significant new funds for organizations, investors depend on the organizations financial conditions (Song & Yang, 2016).

If the forward looking disclosures about the financial conditions of organizations are not well disclosed, investors doubt about their investment decisions. Managers could disclose the organizations financial conditions in order to raise new funds (Song & Yang, 2016). The agency problem is that the interests of the management and the users of the financial statement differ from each other. This problem results in costs and inadequacies who are payable by the society. These agency costs are difficult to measure precisely but they are important. The estimates of agency costs for large manufacturing organizations and small organizations are positioned at costs at 0.2 and 5.0 percent of the revenue, respectively (Ang, Cole, & Lin, 2000; Dobson, 1992). The agency theory states that management will influence the investors and financial statement uses in order to economize on these costs (Bosse & Phillips, 2016).

In the studies of Boehmer and Ljungqvists (2004) and Leuz et al (2008) it is found that management has an influence on reporting decisions for personal objectives that are not aligned with the objectives of the organization. Prior research about segment reporting do not attempt to directly test whether managerial self-interest plays a large role in segment aggregation decisions (Boehmer & Ljungqvist, 2004; Leuz, Triantis, & Wang, 2008). Even in the broader empirical disclosure literature, little evidence exists on the agency costs motive for voluntarily forward looking disclosure decisions (Berger & Hahn, 2007). Managers would withhold forward looking disclosures when the knowledge of investors about the investment is incomplete (Dye, 1985). According to Dutta and Trueman (2002), managers could make the forward looking disclosures decisions based on their prediction of the interpretation of the market (Dutta & Trueman, 2002).

(8)

Other studies also examine forward looking disclosures provided in annual reports. These studies are often limited to a small sample size (Clatworthy & Jones, 2003; O'Sullivan, Percy, & Stewart, 2008).

Segment reporting could have an influence on the level of impact of agency costs on forward looking disclosure decisions. Prior research provides evidence that organizations who operate in different segments trade at a discount relative to stand-alone organizations and internal capital markets in organizations with diversified line of business transfer funds across segments in a suboptimal manner. This difference could be related to agency costs in segment reporting. In this study the focus is on foreign sales in U.S. multinationals. Segment data provide information about the diversification strategy of an organization and the transfers of resources across the divisions (Berger & Hahn, 2007).

1.2 Research Question

Based on the discussion in the literature and the background information this research examines how agency costs are associated to the level of forward looking disclosures on foreign sales in U.S. multinationals.

1.3 Contribution of this Research

In this research the focus of the main reasons for why managers withhold segment related news is on agency costs (Berger & Hahn, 2007). In the prior literature of segment reporting there is not tested whether managerial agency costs have a large role in segment aggregation decisions. Even in the wider empirical disclosure researches, little evidence occurs on the agency costs motive for withholding disclosures (Berger & Hahn, 2007). Investors and financial statement users use the information that is audited in the financial statements (Athanasakou & Hussainey, 2014). Prior studies who examine forward looking disclosures that are provided in annual reports are often limited to a small sample size (Clatworthy & Jones, 2003; O'Sullivan, Percy, & Stewart, 2008). In this study computer intensive techniques will be used. Several more recent studies have used computer intensive techniques to examine large samples of MD&A disclosures in the U.S. (Li, 2008; 2010; Brown & Tucker, 2011; Davis & Tama-Sweet, 2011) Most of these researches focused on the qualitative characteristics of the information that is provided in the MD&A section (Muslu, Radhakrishnan, Subramanyam, & Lim, 2015). In difference with much of the prior researches, the focus in this study is on the quantity of the forward looking disclosures in the MD&A section of the annual reports.

(9)

There is a gap in the literature about the effect of agency costs on forward looking disclosures for international organizations. The contribution of this research is to investigate this effect in a unique setting of U.S. multinationals. The results of this research could be interested for the SEC based on their concern about the forward looking disclosures in de MD&A section of annual reports.

1.4 Structure of this Research

Following the introduction, the next chapter provides an overview of the prior literature about the variables used in this research, forward looking disclosures, agency costs and foreign sales. In the third chapter the hypothesis development is described. The research methodology is explained in chapter four including the descriptive statistics. The results are presented in chapter five with a detailed discussion. Finally, the conclusion of this research is given in chapter six.

(10)

2 Literature review

The variables that are used in this research are forward looking disclosures, agency costs and foreign sales. The main accounting theory and information from prior literature of the variables is described. The relationship between the variables is explained in the last paragraph.

2.1 Forward Looking Disclosures

The definitions of forward looking disclosures is explained. Forward looking disclosures are filed in the Management's Discussion and Analysis (MD&A) section of the annual reports. In this study the forward looking disclosures are based on the sentences in the MD&A section of annual reports. The reporting requirements for the MD&A section in the annual reports are described.

2.1.1 Definition forward looking disclosures

Forward looking disclosures are predictions of the business of an organization about the future. These disclosures provide investors with information about an organizations future forecast that could be used for decision making (Alkhatib, 2014). They can be an essential factor to help investors and financial statement users to better understand what an organizations previous and present performance might be in the future (Liu, 2015). Forward looking disclosures includes information about financial predictions of an organization. This information contains revenue prediction, sales and cash flows. It also includes information about non-financial predictions about factors that could affect the future performance of an organization. Examples of non-financial predictions are risks, uncertainties, agency relationships, operations general relevant information of an organization, analysis and evaluation (Uyar & Kilic, 2012; Aljifri & Hussainey, 2007). Forward looking disclosures could be found in the annual report. There are several terms that are included in the organizations chairman’s report in the annual report that represent forward looking disclosures. For example, terms as predicting, expecting, estimating, forecasting and hoping. For forward looking disclosures, a distinction could be made between good, bad and neutral news disclosures (Alkhatib, 2014).

Prior literature shows that there are financial, economic and social benefits in forward looking disclosures (Alkhatib, 2014).Some organizations are trying to avoid the advantage of an agency relationship by voluntarily disclosing good and bad news equally (Clarkson, Kao, & Richardson, 1994; Johnson, Kasznik, & Nelson, 2001).

(11)

Therefore, other organizations disclose information about future performances voluntarily without any reasons that could be to attract investors to construct decision making (Uyar & Kilic, 2012; Lim, Matolcsy, & Chow, 2007; Dutta & Trueman, 2002). The literature suggests that there are differences in forward looking disclosures for different organizations. There is a relation between the prediction of future earnings and performance with firm-specific characteristics and the size of an organization. Large sized firms likely disclose more additional and high quality information about future earnings and performance than smaller organizations (Kent & Ung, 2003; O'Sullivan, Percy, & Stewart, 2008). There are even differences between countries. In some developed countries organizations are likely to disclose forward looking information which is considered to be expensive and less revealing. Therefore, in other developed countries forward looking disclosures have been considered beneficial and important (Vanstraelen, Zarzeski, & Robb, 2003). Shareholders of organizations question the management about the future forecasting frequently. The management of a company could not predict specifically or give a concrete answer about the future performance of a company. However, they are in a position to observe most recent market trends. The management could present this information to the investors with explanations about what the organization is suggesting or planning to do (Alkhatib, 2014).

2.1.2 Financial reporting forward looking disclosures

Despite of the understanding that specific statements of an organization’s annual statement are considered speculative. The Securities Exchange Commission (SEC) states that listed organizations are obligated to include a disclaimer on all of the management debates, disputes and discussions with investors with the intention to emphasize this point (Healy & Palepu, 2001; Kent & Ung, 2003). Forward looking disclosures in most cases reduces information asymmetry between the management of an organization and the investors. These disclosures are a contribution for investors to make better investment decisions (Alkhatib, 2014; Uyar & Kilic, 2012). The annual reports which contains the forward looking disclosures are considered as the best possible way of communicating the information that investors use for investment decisions. The SEC (2016) requires that the wealth of information for a United States (U.S.) public company is in the company’s annual report on Form K. Among other things, the 10-K describes detailed an organizations business, the risks, and the operating and financial results for the fiscal year. The management of the organization also discusses its perspective on the business results and what the drivers are (Securities and Exchange Commission, 2016).

(12)

Most U.S. public companies are required to produce a 10-K each year and file it with the U.S. SEC. The rules of the SEC require that 10-Ks follow a set order of topics (Securities and Exchange Commission, 2016).

The rules also require companies to send an annual report to their shareholders when they are holding annual meetings to elect members of their boards of directors. There is a lot of overlay in the requirements for the 10-K and the annual report to shareholders, nonetheless there are also important differences. The 10-K includes more detailed information than the annual report to shareholders. A number of companies take their 10-K and send it as their annual report to shareholders. In those cases, the 10-K filed with the SEC and the annual report to shareholders are the same document. The company writes the 10-K and files it with the SEC. Laws and regulations prohibit companies from making materially false or misleading statements in their 10-Ks. Likewise, companies are prohibited from omitting material information that is needed to make the disclosure not misleading. In addition, as noted above, the Sarbanes-Oxley Act requires a company’s CFO and CEO to certify the accuracy of the 10-K. The SEC neither writes the 10-K nor vouches for its accuracy. The SEC sets the disclosure requirements and the SEC staff reviews 10-Ks to monitor and enhance companies’ compliance with the requirements. Both the SEC and the staff also provide interpretive advice about the disclosure requirements. All 10-Ks filed with SEC are available to the public by the SEC (Securities and Exchange Commission, 2016). Forward looking disclosures has established an important role for future performance issued by U.S. multinationals (Baik & Jiang, 2006).

The lack of firm-specific knowledge in the market forces the external need for and reliance on forward looking disclosures (Aharony, Lin, & Loeb, 1993). The disclosures are not without costs because they could reveal proprietary information about an organization and they are not protected from litigations. The effectiveness of self-serving causal disclosures tends to be context-sensitive. This might result in a setting were organizations avoid disclosure precedents by an extensive use of such disclosure mechanisms (Graham, Harvey, & Rajgopal, 2005). According to Gibbins, Richardson and Waterhouse (1990) this could increase the incentives of an organization to adopt an opportunistic disclosure position. The resources of organizations who commit to financial statement verification by independent auditors are an improving function of the level of their management predictions. The market response to disclosures increases in the resources committed to the audit. This implies that organizations commit to greater auditor financial statement authentication when their managers put more resources into forward looking disclosures (Gibbins, Richardson, & Waterhouse, 1990).

(13)

The management carry this out by making more frequent and more informative forward looking disclosures (Gibbins, Richardson, & Waterhouse, 1990). The investors then perceive that the forward looking disclosures are more credible. As a result, the financial statement verification enhances the information value of the forward looking disclosures (Ball & Shivakumar, 2012).

The mechanisms by which managers can commit to truthful disclosure of private information are not well understood. Credible disclosure must acquire some costs (Spence, 1973). The costs of committing to a particular level of independent audit includes audit firm fees, related internal accounting and control expenses, internal audit expenses, management time and the decreased usefulness form managers restricting their ability to manipulate the financials. The level of independent audit has an influence on the measurement of accuracy and independency from managerial manipulation of the reported financial outcomes. The expectation of more accurate and independent ex post accounting restrains the managers to be more truthful ex ante in their private information disclosures. The mechanism works in this way. Reporting ex post financial outcomes that are both accurate and independent of management manipulation permits outsiders with a means of evaluating the truthfulness of past management disclosures. Examples of ex post financial outcomes are revenues, total assets and earnings. The expectation that actual outcomes will be accurately and independently reported assists to restrain managers to make more truthful and hence more informative voluntary disclosure ex ante. The financial reporting that is based on independent verifiable information will incentivize managers to implicitly commit to truthfully disclose unverifiable information. The credibility of the disclosures by management will improve the disclosures of private information. This complementarity in roles suggests that financial reporting is an input to the organizations total information environment. It can only be designed and evaluated by its contribution to the efficiency of the complete system of communication with investors and other users of the financial statements. (Ball & Shivakumar, 2012).

Nevertheless, it is not certain that forward looking disclosures in the MD&A section are informative. The MD&A section of the annual reports is not a timely way of communication. There are other ways of communication that are timelier and would anticipate information provided through the MD&A section. Examples of these channels of communication are analyst hosted investor conferences, conference calls and press releases (Muslu, Radhakrishnan, Subramanyam, & Lim, 2015).

(14)

A second reason is that investors might rely less on the MD&A sections than other parts of the annual reports. Only for the consistency with the other parts of the annual report are the MD&A sections reviewed by auditors (Muslu, Radhakrishnan, Subramanyam, & Lim, 2015). Previous researches lack of evidence of the usefulness of the MD&A sections. The criticism of the SEC suggests a possibility that the forward looking disclosures in the MD&A sections are not informative. Besides, the theories of voluntarily disclosures are contradictory whether the management disclose useful information. The perspective of informative forecasts that the management disclose value-relevant information. Early signalling models claim that organizations disclose all value-relevant information to reduce agency problems (Grossman, 1981; Milgrom, 1981). Following models impose costs and derive selective strategies of disclosures where the management generally disclose good news (Verecchia, 1990; Dye, 1985). The differences with the forecasts of the signalling models is that the litigation risk hypothesis claims that organizations voluntarily disclose bad news (Skinner, 1994). At last, prior studies discuss that management disclose either good or bad news to meet the information needs of investors (McNichols, 1989; Lang & Lundholm, Corporate Disclosure Policy and Analyst Behavior, 1996). It is also possible that management aligns their expectations with the expectations of the investors (Ajinkya & Gift, 1984). The perspective of informative forecasts argues that voluntarily forward looking disclosures are informative, even when they are selective. The perspective of opportunism forecasts that the forward looking disclosures are formed by the motivations of the management. The management will disclose good news to raise new capital or extracting rents (Aboody & Kasznik, 2000; Lang & Lundholm, 2000). Managers will disclose good news timely but withhold bad news based on their career motives. The perspective of opportunism overall argues that misleading of investors and the market in the short term is caused by voluntarily forward looking disclosures. Concluding, it remains an empirical question if the forward looking disclosures in de MD&A section of the annual reports communicate useful information to investors and the market (Kothari, Shu, & Wysocki, 2009).

2.2 Agency Costs

Agency costs originate from the agency theory. The definition of the agency theory and the costs are described. The way agency costs operate and the risks that come along with agency costs are further explained.

(15)

2.2.1 Definition agency costs

The agency theory describes the relationship between the stakeholders, the shareholder (principal) and director (agent). This relationship has emerged because of the separation between the supply of capital and operations. The conflict of interests could arise between two parties because of this separation (Bonner & Sprinkle, 2002). Managers are in the possession of private information that is not known by shareholders and stakeholders. Managers could make decisions that are in line with their own interests and this may not be in line with the best interest of the organization and the shareholders. This could happen when managers are not monitored by shareholders. Managers could withhold information when this results in negative consequences for the organization and reducing of shareholder wealth. Managers could also be incentivized to report more forward looking disclosures when this results in positive consequences for the organization (Berger & Hahn, 2007).

When the interests of the managers of an organization are not aligned with the owners of the organization agency costs will arise (Ang, Cole, & Lin, 2000). Agency costs arise in any situation when a cooperative effort is involved by two or more people, even when there is not a clear cut principal-agent relationship (Jensen & Meckling, 1976). These costs could have the form of preferences for on-the-job perks, avoidance and making decisions in their own self-interest that reduce shareholder wealth (Ang, Cole, & Lin, 2000). The problem of inducing an agent to behave as if he were maximizing the principal’s welfare is quite general. This problem exists in all kind of organizations and in all cooperative efforts at every level of management. For example, universities, mutual organizations, cooperatives, authorities and bureaus of the government, unions and in relationships who are normally classified as agency relationships such as are in common in the performing arts and the real estate market. The literature explains the form of which agency costs take place in each of these situations and types of organizations and cooperative efforts (Jensen & Meckling, 1976).

2.2.2 Performance of agency costs

To explain the mechanism of agency costs and disclosures the agency costs hypothesis is explained. The agency cost hypothesis implies that when the agency costs motive dominates, managers tend to withhold forward looking disclosures with relatively low abnormal profits and/or high expenses. Disclosures are withheld as a result of interest conflicts between managers and shareholders (Berger & Hahn, 2007). Managers face agency cots of segment disclosure. These costs arise when segment data provide information that is indicative of

(16)

Prior papers indicate that diversified organizations trade at a discount relative to stand-alone organizations. The diversification discount is related to the measurements of agency costs (Berger & Ofek, 1995).

In the study of Ang et al (2000) they have examined the effect of agency costs on managerial equity ownership. Their research demonstrates that the expense ratio and asset utilization ratio are a measurement for agency costs (Ang, Cole, & Lin, 2000). In the prior literature there is no clarity about what the best measurement is for agency costs because agency costs are very difficult to determine. The studies of Singh & Davidson (2003), Florackis (2008), Jelinek & Stuerke (2009) and Ibrahim & Samad (2011) already used the measurement of Ang et al (2000). The proxies for agency costs for this research are used from Ang et al (2000) and measured as firm efficiency with the expense ratio and the asset utilization ratio. The agency costs can be measured with the expense ratio and the asset utilization ratio from the research of Ang et al (2000). The expense ratio is the operating expense scaled by annual sales. The ratio indicates the effectiveness of the organization’s management control operating costs. The expense ratio is a measure to show how effectively the management of the organization controls the operating costs. This costs include the extreme perquisite consumption and other direct agency costs. A high expense ratio is also related to the agency cost hypothesis of Berger and Hahn (2007). This hypothesis stated that managers withhold more forward looking disclosures when the profits are abnormal low and the expenses are high. The asset utilization ratio is calculated by annual sales divided by total assets. The ratio measures the effectiveness of how the management of the organization uses or organizes the assets. It is a proxy for the loss in revenues allocated to inefficient asset utilization, which could result from bad investment decisions. In distinction with the expense ratio, the agency costs are negatively related to the asset utilization ratio. These costs increase because the management acts in a way that is not aligned with the goals of the organization. For example, exerts insufficient effort, making bad investment decisions which result in lower revenues. With lower revenues the management could consume executive perquisites which could lead to the purchase of unproductive assets. Examples of unproductive assets are expensive cars, fancy furniture in their offices, resort properties and expensive business excursions (Ang, Cole, & Lin, 2000).

(17)

2.3 Foreign Operations

This research focusses on the foreign sales of U.S. multinationals with foreign operations. With foreign operations segment reporting is applicable. Geographical segment reporting is described and explained to understand the mechanism of foreign sales.

2.3.1 Geographical segment reporting

Managers had more discretion about the reporting profitability even when the organizations were required by the SFAS No. 131 to disclose additional information about the segment operations. The agency costs theory explains that managers are inclined to stop reporting segment profitability or costs in different geographic areas when the information will result in negative consequences for the organization (Berger & Hahn, 2007). To improve the quality of segment information the United States Financial Accounting Standards Board (FASB) issued Accounting Standards (FASB, 1997). According to the FASB, the Association for Investment Management and Research (AIMR) declared that organizations should report information in the way that it reflects how the business of the organizations is organized and managed. There are two standards for the disclosures of information in financial statements, SFAS No. 14 and SFAS No. 131. The standard SFAS No. 14.9 state that many organizations have broadened the scope of their operations into different industries, foreign countries and markets. The SFAS No. 14 published in 1996 requires that the financial statements of an organization include information about the organizations operations in different industries, its foreign operations and export sales, and major customers. The statement also requires that an organization operating predominantly or exclusively in a single industry should identify that industry (FASB, 1976). The standard SFAS No. 131 describes the standards for the way that public business organizations report information about their operating segments in annual financial statements. The standard requires that those organizations report selected information about operating segments in interim financial reports issued to shareholders. The standards also establish standards for related disclosures about products, services, geographic areas and major customers (FASB, 1997).

Before to issue of SFAS No. 131, all multinational organizations were required to disclose their profits, sales and total assets by geographic area. They also had to reveal their line-of-business information (Berger & Hahn, 2007). After the adoption of SFAS No. 131 it was required that all organizations have to disclose their sales and total assets. The disclosure

(18)

A main concern with SFAS No. 14 was the discretion with regard to a segment’s industry definition allowed managers to report more aggregated segment information to external users of the financial statements than what was reported internally (Ernst & Young, 1998). Prior studies show a significant increase in the number of reported segments at the time of the adoption of SFAS No. 131 (Berger & Hahn, 2003).

The new standard offers relatively less discretion for segment aggregation. The increase in the reported segments under SFAS No. 131 suggests that managers used the discretion afforded under SFAS No. 14 to cover segments that could have been separately reported. With the emphasis on the research of Berger and Hahn (2003), they find that the more disaggregated reporting under SFAS No. 131 creates a greater diversification discount. These findings are suggestive of managers hiding information about agency problems under SFAS No. 14. Through, they are also consistent with organizations being forced by SFAS No. 131 to reveal additional proprietary information (Berger & Hahn, 2003).

2.4 Review prior literature

Agency costs problems can be expected to be primary dimensions of opportunistic disclosure behaviour (Gibbins, Richardson, & Waterhouse, 1990). For investors and other users of the financial statements of the organization the forward looking disclosures could be more useful for decision making (Berger & Hahn, 2007). The agency theory motivates voluntary forward looking disclosures as a mechanism to reduce information asymmetry (Athanasakou & Hussainey, 2014). The manager will disclose information when it is either sufficiently high or sufficiently low, regardless of their own assessments of their valuation implications. Managers do not base their disclosure decision on their own interpretation of their private information, but rather on their prediction of how the market analyst and investors will interpret it. Disclosure strategies depend on these valuations. The disclosure strategies are driven by the assessment of the manager. These assessments are about the analysts’ valuation of the disclosed facts. If the analysts have an unfavourable assessment about the valuation of the organization, then the forward looking disclosures of the manager will be withheld. These disclosure decisions could be favourable or unfavourable information but are thus driven based on the agency theory (Dutta & Trueman, 2002).

(19)

Organizations who report under SFAS No. 131 have a lower discretion over the amount of segments they report. This lower discretion over the segment disclosures leads to disaggregation, which could result in consequences for agency costs. New segments under SFAS No. 131 are related with lower abnormal profits than the old segment under SFAS No. 14. This suggests that managers avoided forward looking disclosures about the segment operations. This is consistent with the agency cost hypothesis (Berger & Hahn, 2007). The level of agency costs is limited by the extent of how well owners and delegated third parties monitor the actions and decisions of the outside managers (Ang, Cole, & Lin, 2000). Some organizations stopped reporting the profitability and their expenses of their geographical segments after the adoption of SFAS No. 131 (Herrmann & Thomas, 2000). This increased the information asymmetry and thus the agency costs (Berger & Hahn, 2007).

Based on the prior literature there is limited evidence on the impact of forward looking disclosures on agency costs in a large sample size. Berger and Hahn (2007) find that managers have incentives to withhold bad news about the segments when the news will reveal negative consequences of the managers’ investment strategy and decisions. If a greater disclosure will reveal more of the extent of value-devastation at an underperforming organization, the potential for corporate governance and the control mechanisms to discipline the underperforming manager might increase (Berger & Hahn, 2007). The results of Berger and Hahn (2007) are consistent with the agency costs hypothesis. New segments tend to have lower abnormal profits than the old segments when the agency costs motive dominates. They suggest that managers exploited the discretion afforded under SFAS No. 14 to opportunistically conceal negative information. Based on the prior literature there is limited evidence on the impact of forward looking disclosures on agency costs. Berger and Hahn (2007) find that managers have incentives to withhold bad news about the segments when the news will reveal negative consequences of the managers’ investment strategy and decisions. To fill this gap in the literature this research examines the effect of agency costs on forward looking disclosures in foreign sales.

(20)

3 Hypotheses

Berger and Hahn (2007) concluded that a better understanding about forward looking disclosures can help to resolve agency problems. Forward looking disclosures are withheld as a result of conflicts of interest between shareholders and managers (Berger & Hahn, 2007). Most of the literature and theories are often limited to a small sample size. This is the reason to measure the forward looking disclosures on foreign sales in multinationals of the United States (U.S.) for agency costs. The quantity of forward looking disclosures will be used as the dependent variable. This variable could be separated in two classifications, good and bad news. The independent variable is agency costs. The measurement for agency costs is used from Ang et al (2000). They used the expense ratio and the asset utilization ratio. The expense ratio is the operating expense scaled by annual sales. The ratio indicates the effectiveness of the organization’s management control operating costs. The expense ratio is a measure to show how effectively the management of the organization controls the operating costs. When the expense ratio is high, this indicates that the organization has high agency costs. The asset utilization ratio is calculated by annual sales divided by total assets. The ratio measures the effectiveness of how the management of the organization uses or organizes the assets. It is a proxy for the loss in revenues allocated to inefficient asset utilization, which could result from bad investment decisions. In distinction with the expense ratio, the agency costs are negatively related to the asset utilization ratio. These costs increase because the management acts in a way that is not aligned with the goals of the organization. When the organization has a low asset utilization ratio, this indicates high agency costs. (Ang, Cole, & Lin, 2000). The effect of agency costs on the level of forward looking disclosures is hypothesised for foreign sales of the U.S. Multinationals.

Managers might want to withhold forward looking disclosures if this news reveals negative consequences of managers’ investment strategy and actions (agency costs). Therefore, high agency costs leads to higher level of withholding forward looking disclosures. The disclosures include good, bad and neutral news disclosures. Based on the prior literature and theories, this leads to the following main hypothesis:

Main hypothesis 1: U.S. multinationals with high agency costs withhold more forward looking disclosures on foreign sales.

(21)

This study predicts that there is a positive relation between agency costs and the level of forward looking disclosures. As proxy for agency costs this research uses the expense ratio and the asset utilization ratio. The predictions for the proxies are that the expense ratio has a positive relation with the level of forward looking disclosures and the asset utilization ratio a negative relation. Derived from the main hypothesis the following hypotheses are formulated:

Main hypotheses 1A: U.S. multinationals with a high expense ratio withhold more forward looking disclosures on foreign sales.

Main hypotheses 1B: U.S. multinationals with a low asset utilization ratio withhold more forward looking disclosures on foreign sales.

(22)

4 Research methodology

The research methodology is described which consists of the sample, data, variables and the research method. Finally, the descriptive statistics of the data are given.

4.1 Overview

This research aims to understand the effect of agency costs on the level of withholding of forward looking disclosures. The significant role of the conceptual specification of constructs of a theory-based study could be showed in the position of the predictive validity framework as presented in Figure 1.

Figure 1 Predictive Validity Framework

This research examines if the level of the expense and asset utilization ratio has an influence on the quantity of forward looking disclosures. This could be relevant in the context of the agency theory, forward looking disclosures and segment reporting in an international market. The research question will be answered through a quantitative research combining United States (U.S.) data which are collected from the University of Amsterdam. I will only examine the effect of agency costs on forward looking disclosures on U.S. multinationals on foreign sales. Prior literature often limited to a small sample size which makes this setting unique.

Conceptual Independent variable Dependent variable

→ Agency Theory →

↓ ↓

Operational Proxy Proxy

-Foreign Sales → Linear regression model → Total FLD MD&A section:

-Expense ratio -FLD: Good news

-Asset Utilization ratio -FLD: Bad news ↑

Control Variables

(23)

4.2 Sample and data

This research examines how agency costs are associated to the level of withholding of forward looking disclosures on foreign sales in U.S. multinationals. The dataset is obtained from the University of Amsterdam. The dataset consists of forward looking sentences, which are the classification of forward looking disclosures on foreign operations. These forward looking sentences are often qualitative in nature. The data has been extracted from the Management's Discussion and Analysis (MD&A) sections of 10-K filings, using automated language processing techniques. The sample consists of 2090 U.S. multinational organizations from the years 1996 to 2000. This dataset is unique because a lot of studies are often limited to a small sample size (Clatworthy & Jones, 2003; O'Sullivan, Percy, & Stewart, 2008). The organizations that engage in geographical segment reporting have at least two geographical non-U.S. segments for each of the five years. The information is about the following subjects. The number of forward looking sentences on foreign segment operations per firm-year. The classifications on forward looking sentences will be good news, bad news or neutral news sentences to investors and financial statement users. At last, the classification will be made into the number of forward looking sentences on foreign segment sales, costs, profits and investments. The additional information in the dataset that is extracted from the annual reports are total assets, total assets lagged, total foreign assets, long-term debt, pre-tax income, pre-tax income domestic, pre-tax income foreign, standard industrial classification, total sales, total foreign sales, total foreign segments and total business segments.

4.3 Variables

The variables that are used in the regressions are distinguished in dependent and independent variables. The dependent variable is the variable forward looking disclosures. The independent variables are the agency costs with as proxy the expense ratio and asset utilization ratio. The quantity of forward looking disclosures (TOTAL_FLD) is based on the sentences of organizations in the annual report. These sentences have been extracted from MD&A sections of 10-K filings for US multinationals from the years 1996 to 2000, using automated language processing techniques. The automated language processing technique classifies a sentence as a forward looking sentence on foreign disaggregated operations if it meets the following requirements. The first requirement measures if the sentences are forward looking.

(24)

The sentences are forward looking if the following words occur in these sentences: “will”, “should”, “can”, “could”, “may”, “might”, “expect”, “anticipate”, “believe”, “plan”, “hope”, “intend”, “seek”, “forecast”, “objective”, “goal”, “project”, “seek”. This list of words is originating from the research of Li (2010). The second requirement is the measurement of foreign country or region. Sentences are classified as foreign sentence if the company is on the country list provided by GATE, region list provided by GATE, country adjusted list provided by gate (e.g. French, German, Dutch, Latin America etc.) and region adjusted list provided by GATE (e.g. Asian, European, Latin American etc.). List items who are related to the U.S. are deleted from the initial GATE lists. The last requirement is to identify if the sentences are related to foreign operations. The four categories are sales, costs, profits and investments. For sales the stem is “revenue”, the string is “sales”. For costs the stems are “cost” and “expense”, the strings are “SG&A”, and “COGS”. For profits the stems are “profit”, “loss”, “earnings”, “income”, “margin”, “EBIT”, “EBITDA”, “EBITA” and “eps”. The sequence of strings is “operating results”, “results of operations” and “operating performance”. Finally, for investments the stems are “expenditure”, “invest”, “investment”, “acquire” and “acquisition”. Although the data coding procedure extracts relevant sentences, there are a lot of unsuccessful matches. To filter out these unsuccessful matches there are classifications made in successful vs. unsuccessful matches (Q1) and further classification of successful matches (Q2 & Q3). The first classification identifies if the forward looking sentence is on disaggregated foreign operations (Q1). If no: FL_GEO = 0, If yes: FL_GEO = 1. The next classification is the main classification (Q2). Q2a: Code sentence on whether it is about: future disaggregated sales (FL_SALES = 1), future disaggregated profits (FL_PROFITS = 1), future disaggregated costs (FL_COSTS = 1) and future disaggregated (des)investment (FL_INVEST = 1). Q2b: Code sentence on whether it is about: future disaggregated operations outside North America (FL_FOREIGN = 1), future disaggregated operations inside North America which is the U.S. and Canada (FL_NA = 1), future disaggregated operations in US (FL_US = 1) and future disaggregated operations in Canada (FL_CA = 1). Q2c: Code sentence on the news sign: positive news about future disaggregated operations (FL_GOOD = 1), negative news or caution about future disaggregated operations (FL_BAD = 1) and no or neutral news about future disaggregated operations (FL_NO = 1). Positive news deals with increases in firm value (increase in sales, increase in profits, and decrease in costs). Negative news deals with decreases in firm value (decrease in sales, decrease in profits, and increase in costs).

(25)

For investments positive news deals with investments/expansions of the firm, whereas negative news deals with disinvestments. Q3: Specify whether sentence informs about currency risks: currency risks: (CUR=1).

In the literature there is no clarity about what the best measurement is for agency costs because agency costs are very difficult to determine. Some studies use the expense ratio and asset utilization ratio as proxy. The studies of Singh & Davidson (2003), Jelinek & Stuerke (2009) and Ibrahim & Samad (2011) already used the measurement of Ang et al (2000). The proxies for agency costs for this research are used from Ang et al (2000) and measured as firm efficiency with the expense ratio (EXPRATIO) and the asset utilization ratio (UTIRATIO). The expense ratio is the operating expense scaled by annual sales. The operating expenses are calculated by total sales minus the pre-tax income. A benchmark for the expense ratio is between 55% and 80%. The range of values is so wide due to the differences in sizes and types of operations, as well as the capital structure and individual efficiencies of a business. According to Jelinek and Stuerke (2009) a ratio over 80% may indicate profitability problems, while a ratio under 55% may indicate great efficiency. Ang et al (2000) explains when the expense ratio is high, this indicates that the organization has high agency costs. The asset utilization ratio is calculated by annual sales divided by total assets. A firm whose asset utilization ratio is lower than the base case firm have positive agency costs. When the organization has a low asset utilization ratio, this indicates high agency costs (Ang, Cole, & Lin, 2000). The ratios are not measured without error. The sources of measurement error include the differences in the accounting methods that are implemented with respect to the recognition and timing of revenues and costs.

As explained in the literature review managers will withhold forward looking disclosures if this information could have negative consequences for an organization. Investors and financial statement users are the most focused and interested in the sales and profit of organization to make a decision about for example an investment. It is expected that forward looking disclosures have a positive relation with the expense ratio and a negative relation with asset utilization ratio (Rashid, 2013).

4.3.1 Control variables

A number of control variables are used in this research consistent with prior researches. When these control variables are not included in the regression analysis this could affect the results

(26)

The control variables used in this study are foreign sales (FORSALES), firm size (SIZE), debt of the firm (DEBT), segment diversity (SEGDIVERISTY) and leverage (LEVERAGE).

The first control variable is foreign sales (FORSALES) where a ratio is used for the measurement. This variable is calculated by dividing the total foreign sales by total assets. A few organizations stopped with the reporting of their profitability and expenses of their geographical segments (Herrmann & Thomas, 2000). This led to an increase of agency costs (Berger & Hahn, 2007). The second control variable is firm size (SIZE) which indicates the size of the firm. The size of the firms could influence the level of forward looking disclosures. The larger a firm, the higher the probability of withholding forward looking disclosures. The size of the organization is determined by the amount of the total assets (Uyar & Kilic, 2012; Ang, Cole, & Lin, 2000; Ibrahim & Samad, 2011; Jelinek & Stuerke, 2009). This variable indicates an amount, therefore the logarithms of total assets are used for the determination of the control variable (SIZE=ln (at)). This way, the wide distribution of the amounts does not influence the linear regression later. (Field, 2013). The debt of the firm (DEBT) indicates the position of long-term debt. If a firm has a lot of long-term debts this could increase the probability to withhold forward looking disclosures. The management might want to hide information about the sales and profit because this could have negative consequences for the organization and creditors. If a firm has a lot of long-term debts and high sales and profit their investors and financial statement users could question the firm about their operations. (Ibrahim & Samad, 2011). This variable also indicates an amount, therefore the logarithms of long-term debt are used for the control variable (DEBT=ln (dltt)) (Field, 2013). Segment diversity (SEGDIVERISTY) indicates that organizations who have a higher segment diversity are operating in more different segments and are more likely to aggregate their segment information to hide sales and profit. The more different segments that organizations operate in, the more likely that these organizations hide their sales and profit information. The segment diversity is the amount of the business segments (busseg) organizations have (Berger & Hahn, 2007). Leverage (LEVERAGE) indicates if an organization is capable to meet its financial obligations. The debt ratio is used as calculation for leverage and is calculated by dividing long-term debt by total assets. If an organization has a high leverage ratio, this could increase the probability to withhold forward looking disclosures. The management could hide information about the sales and profit because this could have negative consequences for the organization and creditors (Uyar & Kilic, 2012; Ibrahim & Samad, 2011; Jelinek & Stuerke, 2009).

(27)

4.4 Research Method

This research tests whether agency costs have an impact on the quantity of forward looking disclosures using a linear regression model. First the descriptive of the statistics are given. Second the correlation between the variables are tested. Finally, a linear regression is performed.

The purpose of regression analysis is to evaluate the relative impact of a predictor variable on a particular outcome. This is different from a correlation analysis, where the purpose is to examine the strength and direction of the relationship between two random variables. A linear regression model contains only one independent (explanatory) variable and is linear with respect to both the regression parameters and the dependent variable. The corresponding dependent (outcome) variable is labelled. Typical steps for regression model analysis are the following: (a) determine if the assumptions underlying a normal relationship are met in the data, (b) obtain the equation that best fits the data, (c) evaluate the equation to determine the strength of the relationship for prediction and estimation, and (d) assess whether the data fit these criteria before the equation is applied for prediction and estimation (Zou, Tuncali, & Silverman, 2003). To test the main hypotheses, linear regressions in the program Stata are tested. For hypothesis 1A the following regressions are performed:

TOTAL_FLD= α + β1EXPRATIO+ β2FORSALES + β3SIZE + β4DEBT + β5SEGDIVERISTY + β6LEVERAGE

GOOD_FLD=α + β1EXPRATIO+ β2FORSALES + β3SIZE + β4DEBT + β5SEGDIVERISTY + β6LEVERAGE

BAD_FLD=α + β1EXPRATIO+ β2FORSALES + β3SIZE + β4DEBT + β5SEGDIVERISTY + β6LEVERAGE

For hypothesis 1B the following regressions are performed:

TOTAL_FLD = α + β1UTIRATIO + β2FORSALES + β3SIZE + β4DEBT + β5SEGDIVERISTY + β6LEVERAGE

GOOD_FLD = α + β1UTIRATIO + β2FORSALES + β3SIZE + β4DEBT + β5SEGDIVERISTY + β6LEVERAGE

(28)

BAD_FLD = α + β1UTIRATIO + β2FORSALES + β3SIZE + β4DEBT + β5SEGDIVERISTY + β6LEVERAGE

TOTAL_FLD= The quantity of good, bad and neutral forward looking disclosures GOOD_FLD= The quantity of good forward looking disclosures

BAD_FLD= The quantity of bad forward looking disclosures β1 EXPRATIO = Operating expenses divided by annual sales β1 UTIRATIO = Annual sales divided by total assets

β2FORSALES= Foreign sales divided by total assets β3SIZE = Logarithm of total assets of the firm

β4DEBT = Logarithm of long-term debt of the firm

β5SEGDIVERISTY = The amount of the business segments β6LEVERAGE = Ratio long-term debt total assets

This type of model is referred to as a linear regression model. The β values are constants and are called regression coefficients or regression weights. This regression tests whether the independent variables have a significant (p<0,05) effect on the dependent variable forward looking disclosures. For the variable forward looking disclosures there are three distinctions made, bad news, good news and neutral news. These three types are all tested in this

regression as a total amount called TOTAL_FLD. If there are significant results that the independent variables have an impact on the total of the forward looking disclosures, the regression will be tested separately per type of news disclosure to see which type has the most significant impact. The expense ratio is the operating expense scaled by annual sales. The ratio indicates the effectiveness of the organization’s management control operating costs. When the expense ratio is between 55% and 80%, this indicates that the organization has high agency costs. The asset utilization ratio is calculated by annual sales divided by total assets. The ratio measures the effectiveness of how the management of the organization uses or organizes the assets. In distinction with the expense ratio, the agency costs are negatively related to the asset utilization ratio. When the organization has a low asset utilization ratio, this indicates high agency costs (Ang, Cole, & Lin, 2000). In the regressions for hypothesis 1A and 1B the control variables are included because they could influence the outcome if they are not included. The control variables are foreign sales (FORSALES), firm size (SIZE), debt of the firm (DEBT), segment diversity (SEGDIVERISTY) and leverage (LEVERAGE).

(29)

4.5 Data and descriptive statistics

The sample consists of 2090 U.S. multinationals from years 1996 to 2000.The data of forward looking disclosures are obtained from the MD&A section in the annual report. The proxies for agency costs are the expense ratio and the asset utilization ratio. Table 1 presents the descriptive statistics for our sample size of 2090 U.S. multinational firms.

Notes: Table 1 Panel A presents descriptive statistics for the variables used in the regression. The sample consists of 2090 U.S. multinational organizations with forward looking sentences in the MD&A sections from years 1996 to 2000. TOTAL_FLD indicates the quantity of good, bad and neutral forward looking disclosures. GOOD_FLD indicates the quantity of good forward looking disclosures. BAD_FLD indicates the quantity of bad forward looking disclosures. EXPRATIO is the proxy for agency costs and is the ratio of operating expenses annual sales. UTIRATIO is the proxy for agency costs and is the ratio of annual sales and total assets. FORSALES is the ratio of foreign sales and total assets. SIZE is the logarithm of total assets of the firm. DEBT indicates the logarithm long-term debt of the firm. SEGDIVERISTY indicates segment diversity and is the amount of the business segments that the organization operates in (busseg). LEVERAGE is the ratio of long-term debt and total assets. Table 1 Panel B presents the descriptive statistics of the independent and control variables. There is a difference in the sample for long-term debt. Not all of the organizations have a long-term debt. From the sample of 2090

Variable N Mean SD p25 p50 p75 TOTAL_FLD 2090 0,392 0,790 0,000 0,000 1,000 GOOD_FLD 2090 0,170 0,493 0,000 0,000 0,000 BAD_FLD 2090 0,168 0,496 0,000 0,000 0,000 EXPRATIO 2090 0,952 0,226 0,875 0,932 0,985 UTIRATIO 2090 1,225 0,640 0,832 1,102 1,458 FORSALES 2090 0,493 0,406 0,251 0,417 0,616 SIZE 2090 6,571 1,950 5,257 6,562 7,777 DEBT 1822 4,683 2,636 3,029 5,138 6,570 SEGDIVERSITY 2090 2,548 2,108 1,000 1,000 4,000 LEVERAGE 2090 0,191 0,197 0,028 0,154 0,286 Variable Mean EXPRATIO 0,952 UTIRATIO 1,225 FORSALES 0,493 SIZE 6,571 DEBT 4,683 SEGDIVERSITY 2,548 LEVERAGE 0,191

TABLE 1 Sample of Annual Report Disclosures

Panel A: General descriptive statistics

(30)

Table 1 Panel A reports the general descriptive statistics for the quantity of forward looking disclosures, expense ratio, asset utilization ratio and the control variables. TOTAL_FLD has a mean (average) of 0.392. Comparing to the average of GOOD_FLD with 0.170 and BAD_FLD 0.168 the average is lower than TOTAL_FLD. The forward looking disclosures are classified as 1 for a forward looking disclosure and 0 for none. The average of the total forward looking disclosures means that almost half of the forward looking sentences are forward looking disclosures. There are a bit more good news disclosures but this is a small difference of 0.002. The standard deviation of these variables has a range of 0.790 to 0.493. The numbers lie in a range close around the average. Jelinek and Stuerke (2009) showed that a ratio over 80% may indicate profitability problems, while a ratio under 55% may indicate great efficiency. The mean of the expense ratio (EXPRATIO) with 0.952 indicates that the agency costs are high. For the asset utilization ratio (UTIRATIO) there is not a clear range indicated in the literature. Based on these descriptive statistics with a mean of 1.225 this presents that for most U.S. multinationals the sales are higher than their total assets and indicates agency costs are existing. The mean of foreign sales (FORSALES) is 0.493. This indicates that the foreign sales of the U.S. multinationals are almost half the amount of their total assets. Only the long-term debt (DEBT) shows that of the 2090 U.S. multinationals, 1822 of them have a long-term debt. The diversity of the segments that the organization operates in (SEGDIVERSITY) shows a mean of 2.548. This indicates that each firm operates in 2.6 segments, on average. The mean of 0.191 for leverage (LEVERAGE) indicates that the amount of long-term debt has almost 20% of the value of the total assets. Table 1 Panel B presents the mean of the independent variables expense ratio and asset utilization ratio and control variables used in this research.

(31)

5 Results

The correlation between the variables used in this research are presented and explained. In the second paragraph the linear regression is performed and the results are showed with an explanation. At last, based on the results, the effect of the agency costs on the level of withholding forward looking disclosures is discussed.

5.1 Correlations variables

A correlation matrix is used to investigate the dependency between variables at the same time. The correlation says something about the linear relationship between two variables. It checks how the scores on one variable are associated with scores on the other variable. If high scores on one variable are associated with high scores on the other variable, there is a positive correlation between the two variables. If high scores on one variable associated with low scores on the other variable, there is a negative correlation between the two variables. After the test is performed, there are three steps made for the determination of the correlation. First, the determination of a significant relation between the variables. A relation is significant at the level of 5% (p<0.05). The next step is to determine the direction of the correlation. Finally, the determination of the strength of the relation. The measure the strength and the direction of the correlation the correlation coefficient is used. The coefficient can take values from -1 to +1. The sign indicates the direction of the correlation. The absolute value of the correlation coefficient indicates the strength of the correlation. The value for a perfect positive correlation is +1, and the value for a perfect negative correlation is -1. If there is no relationship between the two variables is the correlation coefficient of 0. Guidelines for interpreting the magnitude of the correlation coefficient are little or no correlation (0,00 <r <0,30), low correlation (0,30 <r <0,50:) , moderate correlation (0,50 <r <0,70) , high correlation, (0,70 <r <0,90: ) very high correlation (0,90 <r <1,00:). These guidelines are used in this study for the correlation (Field, 2013).

Table 2 illustrate the pairwise correlations between the forward looking disclosures, proxies for agency costs and the control variables. Overall the correlations between the variables are not quite significant but there are a few correlations with a significance of p<0.05. Not all variables correlate positively with each other, a few correlate negatively.

(32)

The total forward looking disclosures (TOTAL_FLD) is significant correlated with good forward looking disclosures (GOOD_FLD), forward looking disclosures (BAD_FLD), expense ratio (EXPRATIO) and foreign sales (FORSALES) with a significance of p<0.05. GOOD_FLD correlates significant with BAD_FLD, firm size (SIZE) and segment diversity (SEGDIVERSITY). BAD_FLD correlates significant with EXPRATIO. EXPRATIO correlates significant with SIZE, DEBT, SEGDIVERSITY and the leverage ratio (LEVERAGE). UTIRATIO correlates significant with FORSALES, SIZE, DEBT, SEGDIVERSITY and LEVERAGE. The control variables correlate significant with each other at a level of significance p<0.05. Overall, the variables correlate with each other and not to strong. This indicates of a good data set, every variable measures something different. If the variables had a strong correlation that was close to 1, the variables would measure the same purpose.