What is the impact of fair value adjustments on dividend policy?

1 MSc Accountancy & Control, track Accountancy Faculty of Economic and business,

University of Amsterdam

Master thesis:

What is the impact of fair value adjustments on dividend policy?

Student: Abther Bar Student number: 10593861 Date: 17th august 2015 Words: 9501

Faculty of Economics and Business, University of Amsterdam Supervisor Dr. Alexandros Sikalidis

2 Table of Contents 1 Introduction ... 4 2 Literature review ... 7 3 Institutional framework ... 9 4 Hypothesis development ... 12 5 Sample selection ... 14 5.1 Research design ... 15 6 Findings ... 17 7 Conclusion ... 20 8 References ... 21

3

Abstract

This research will examine the impact of the fair value measures on the dividend policy. This research is done in Germany, because Germany is the biggest economy in Europe (CNN Money 2015). The corporate governance in Germany is very strict and does not allow the distribution of unrealized income (IESE 2008). The German GAAP does not allow the distribution of unrealized dividend (KPMG 2008). The distribution of dividend is linked to the Framework of Lintner (1956). The framework of Lintner (1956) states that dividend should only be distributed if income is persistent over time. This study explores if the fair value adjustments are persistent in Germany and if the fair value adjustments are distributed. The findings of this research show that the fair value adjustments are not persistent in Germany. The results of this research show that there is no distribution of fair value measures. The fair value measures do not influence the dividend policy in Germany. This is in line with the framework of Lintner (1956). This paper is a replication of the research of Igor Goncharov and Sander Van Triest (2011).

4

1 Introduction

The ambiguity about the advantages and cons of the fair value accounting method has caused an ongoing debate (e.g. Barth, 2007; Christensen and Nikolaev, 2013; Wallison, 2008). This paper is a replication of the research of Goncharov and van Triest (2011). The research of Goncharov and van Triest (2011) has examined the impact of fair value accounting in Russia. This thesis focuses on the impact of fair value accounting in Germany. There are two accounting methods available for measuring the value of a corporation (IFRS 2012). These methods are the historical cost method and the fair value method, respectively. This study will explore the fair value method and the impact of this method on the dividend policy. Before IFRS was implemented most counties had their own General Accepted Accounting Principles (GAAP). Each country could value their investment property according to their GAAP rules (KPMG 2008). Corporations are in some countries required to choose a certain method, for example in the Netherlands it is required to use the fair value method (KPMG 2008). Other counties have the freedom to choose between these methods like in the US (KPMG 2008). Once a method is chosen then a corporation cannot change to the other method (PWC 2014). Since the adoption of IFRS by the European Union (EU) it has become a voluntary act for all of the listed firms of European Union counties to value their investment property in either the fair value method or the historical cost method (PWC 2014). There are 283 counties worldwide who have adopted the IFRS (PWC 2014). How to value the properties in an organization is, according to IFRS, an important booking issue (PWC 2014). As mentioned before, IFRS allows the valuation of investment property of listed firms in both ways i.e. fair value and historical cost method. The fair value method values the property of a firm each year; this is done by real estate agents (KPMG 2008) and the valuation is then based on the market expectations. The value of the property could increase or decrease depending on the market. The fair value method provides a future expectation that is not yet realized (Barth and Landsman 1995, Barth 2007). Therefore, it is interesting to explore how organizations handle these unrealized gains or losses. The reason why fair value method has been introduced was to provide more relevant and up to date information to shareholder and stakeholders (Barth and Landsman 1995, Barth 2007). This, in turn could improve the decision making process of investors. With historical cost the price of the assets remains the same as the purchase price (Barth and Landsman 1995, Barth 2007). The changes in market prices do not influence the value of the property in a firm. However, this will not provide up to date information for stakeholder and shareholders (IFRS 2012). The historical cost method will not have an impact on the distribution of dividend, because the value of the property will

5 remain the same. This changes when using the fair value method. With this method the value could increase, decrease or remain the same during the lifetime of the property. The fair value delivers information about the expected future value and cash flows that could be achieved with the investment property (e.g. FASB 2000, para. 36). This makes the fair value accounting method more relevant than the historical cost method (e.g. FASB 2000, para. 36). The information which the fair value method provides will improve the decision making process of shareholders and stakeholders (e.g. FASB 2000, para. 36). This information is relevant for investors as it allows them to make the right decisions at the right time (Barth and Landsman 2001, Hitz 2007).

There are researches that oppose the fair value method. The opposition argues that the fair value method is expensive and that it is not easily adopted, they state that the costs outweigh the benefits (Ball, 2006). Ball (2006) states that the costs of estimating the fair value of investments property outweighs the benefits caused by having more relevant information. Furthermore, the fair value method increases managerial verdict. Managers have to judge the credibility of the information that is attained through fair value (Hail, Leuz & Wysocki, 2010). The fair value method can cause earnings management Hail et al. (2010). Managers can manipulate fair value measures in order to protect their own interests. These interests could be for example, reaching certain targets or earning bonuses (Hail, Leuz & Wysocki, 2010). It can also occur that managers are not sure about the value of the property, due to the fair value method. The study of Hail et al. (2008) states that fair value accounting does not improve the relevance of the income statement, because it is not certain that the delivered information is credible. This is because the information is based on estimates. These estimates could be wrong which could then lead to incorrect figures being issued. The use of Fair value accounting could also lead to momentary changes in the net income (Penman 2007, Plantin et al. 2008). The net income could suddenly change and could cause noise in the decision-making process (Petroni and Wahlen 1995, Hung and Subramanyam 2007). According to Petroni and Wahlen (1995) the fair value method could lead to a reversed impact and harm the decision making process, because of the sudden change in net income. The corporate effect will also increase because of this (Enria et al., 2004). This is because firms want to meet the expectations of the stakeholders which could lead them to wanting to manipulate this information. This will lead to earnings management and will harm the firm and the decision making process of the stakeholders.

This paper will examine what the impact of fair value adjustments is in Germany. Germany is a county that has adopted IFRS in 2005 and is allowed to use both the historical

6 cost and the fair value method (PWC 2014). IFRS does not have regulations about the distribution of unrealized income that is based on the fair value method. However, in Germany it is not allowed to distribute unrealized income (KPMG 2008). It will be interesting to explore whether in Germany unrealized income affects dividend distribution. It is difficult to decide if unrealized income should impact the distribution of dividend. There are researchers that have investigated how to handle the unrealized income based on the fair value method. Lintner (1956) has investigated the conditions which allow for the distribution of fair value income. The framework of Lintner, which has been also mentioned in the research of Jagannathan et al. (2000) and Brav et al. (2005), makes the connection between dividends and earnings. The Linther (1956) framework describes a condition for when it is allowed to distribute dividend based on fair value income. The framework of Lintner (1956) describes that firms want to see a constant relation between dividend development and their earnings. Lintner (1956) developed a theory that approves of the distribution of the dividend that is based on unrealized income, which is measured using the fair value method. Unrealized income should be persistent over time in order to be distributed (Lintner 1956). If there is no continual relation between fair value adjustments and earnings then the dividend should not be distributed. The opposition of Lintner (1956) and his model claim that the distribution of non-realized income will make organizations inconsistent, because the dividend developments could rapidly change due to the constant fair value changes (Boyer 2007, Caruana and Pazarbasioglu 2008). These adjustments will create swings in the economy (Plantin et al. 2008, Laux and Leuz 2009). According to Plantin et al. (2008) the distribution of unrealized income will have a reversed impact on the use of accounting numbers. Stake- and shareholders will not be able to rely on this information because of the continual changes as it will be difficult to make decisions since the information could change within a short time span. According to Plantin et al. (2008) the reliability of the information is reduced. Plantin et al. state that the method of Lintner (1956) will make it very hard for stakeholders to use the accounting numbers. This will lead to large transitory changes in net income (Plantin et al. 2008). Because of these reasons it is a relevant and important research question to know how the fair value adjustments impact the dividend policy.

This thesis will explore the impact of fair value adjustments on dividend policy in Germany from the period of 2009 till 2012. These firms all have adopted IFRS in 2005. These firms are well aware of the use of the fair value accounting method. It will be researched whether there is persistence in fair value measures over these four years and if these firms distributed this unrealized income during the crisis period. It will be important to

7 explore if the dividend policy changes when the damage of the financial crisis is still present. Most organizations did not have growth opportunities during the financial crisis (Carmassi, J., Gros, D.,Stefano, M., 2009). The crisis lasted from 2007 till 2014 (Carmassi, J., Gros, D., Stefano, M., 2009). This thesis will contribute to the literature by examining the consequences of fair value accounting on the dividend policy in Germany.

The paper is structured as follows: literature review, institutional framework, hypothesis development, sample selection findings and the conclusion. In the next section, there will be a theoretical background for the relationship between dividends and earnings components, and the possible impact of fair value accounting. In the third section the instructional framework and the requirements of the relevant German accounting standards will be discussed. The fourth section will discuss the research question and the hypothesis. The fifth section discusses the research design. Section six will sums up the findings. The paper will end with a conclusion and the limitations of this research.

2 Literature review

Investment property of firms is sensitive to changes in value through time (IFRS 2012). This change is not always measured by firms. The measurement depends on the chosen or required accounting method (IFRS 2012). The value of the investment should either be measured by the historical costs or by the fair value method (IFRS 2012). The change in value can have influence on the dividend distribution of the organization. This influence will only occur if the fair value method is used (IFRS 2012). The issue with fair value is that the value of the property changes as it depends on the market value. It can either increase or decrease (IFRS 2012). For example, value of buildings could increase over time. The investment property is according to PWC (2014) all the possessions of an organization. This could be both land and buildings that are held for a lease or that are owned by the corporation (PWC 2014). The possessions of firms are sensitive to changes. These changes depend on the prices on the market (IFRS 2012). These changes van be measured with the fair value method (IFRS 2012). Although the change in value is not realized yet, it is available (PWC 2014). This change in value could have impact on the dividend policy. The important research question regarding this matter is if this unrealized income should impact the dividend policy of a firm. Whether an entity decides to distribute the unrealized income is based on the regulations in that country and the policy of the organization (KPMG 2008). The policy of an entity is among other things a result of the financial situation of the organization (KPMG 2008). An example is the growth opportunities of a firm. According to Gul (1999), an organization with

8 growth opportunities will decide to keep the non-realized income and invest in projects that could realize the growth of this firm. However, an organization with enough borrowing capacity will not need the unrealized income, because they can borrow an amount of money to fulfill their financial needs Cotter and Zimmer (1995). In this case the distribution of the unrealized income will be easier Siokis (2014). The distribution of unrealized income could also be determined by the corporate governance of organization. The corporate governance decides how strictly a firm follows the regulations of the country regarding the distribution of the unrealized income based on fair value measures Fama and French (2001). The distribution of dividend is decided by the corporate governance Fama and French (2001).

The connection between earnings and unrealized investment property has been researched before. Earlier I mentioned that Lintner (1956) researched the connection between earnings and their influence on the dividend policy. Lintner (1956) came up with a theory that determines the distribution of dividend. The framework of Lintner (1956) does allow the distribution of dividend based on unrealized income. Lintner’s model (1956) states that unrealized income based on fair value measures could be distributed if the unrealized income is persistent over time. The distribution of unrealized income is linked to the stability of the fair value measures. The research of Kormendi and Zarowin (1996)) also mention that distribution of non-realized income should be linked with the persistence of the fair value measures. According to them it is then assumed that the persistence is a good indicator that the firm is doing well, so it is allowed to distribute the unrealized income. The research of Jagannathan, Stephens & Weisbach (2000) also agrees with the framework of Lintner (1956). Jagannathan et al. also confirm that persistent unrealized income is allowed for distribution. They state that firms are operating well when their unrealized income keeps increasing. Jagannathan et al. state that not permanent components should not be considered distributed. According to Watts (1977) information from accounting is important to decide about the distribution of the dividend. Future expectations about the value of an organization are important to determine the dividend policy. Watts (1977) also makes the link between earnings persistence and the distribution of dividend. The same argument is made by Jensen and Meckling (1976). They also agree that persistent fair value measures should impact the distribution of dividend. The research of DeAngelo (2006) is in line with this theory. The dividend policy should be based on the net income in previous years (DeAngelo 2006). The connection between dividend and earnings is also described in the research of Brav et al. (2005). They also mention that dividend change is related to the continual change in earnings. When earnings keeps increasing then the dividend should also increase. The same findings

9 are confirmed by Goergen et al. (2005). Earnings persistence is, according to Goergen et al. (2005), important to determine the dividend. This is in alignment with the theory of Lintner. This theory is also confirmed by Edwards and Mayer (1985) who state that temporary aspects of earnings are not part of the determination of the earnings. Earnings should be consistent in order to influence the distribution of dividends. These parts should not impact the dividend. Grullon et al. (2002) also confirm the relation between earnings and persistence. They state that persistence is the key factor for determining dividend distribution. Grullon et al. state that the dividend policy must be based on the earnings persistence. The research of Renneboog and Szilagyi (2007) which is done in the Netherlands also makes the link between earnings and dividend. According to them dividend should be based on earnings which are persistent. The research of Fama and Baiak (1968) also confirm this theory. The researchers Fama and Babiak (1968) have conducted a research in the US. They show the relation between the dividend and the persistence of the earnings. Fama and Babiak use a regression model to explain the link between past and current earnings and the dividend distribution. This regression model is also used in this paper. They state that dividends should be depended on constant earnings. The research that has been done by Kasanen, Kinnunen & Niskanen (1996) is also in line with the theory of Lintner; they also make the connection between the dividend policy and the continuity of earnings. The link between earnings and dividend is also stated by Correia da Silva, Marc & Renneboog (2007). They proved this theory by proving that the dividend is not financed by the reserves of an organization. They confirmed that dividend will reduce when the earnings reduce decrease. The reduction in earnings won’t be filled by the reserves to pay the dividend.

The impact of fair value adjustments on the dividend policy is investigated before by Goncharov and Van Triest (2011). The research of Goncharov and Van Triest (2011) was done in Russia from the period of 2003 till 2006. In that time there was a growth in the economy. The choose Russia because the GAAP of Russia required the use of fair value accounting. The results of the research of Goncharov and Van Triest (2011) concluded that positive fair value accounting lead to lower dividend payouts.

3 Institutional framework

IFRS is adopted by the EU in 2005 (IFRS 2012). IFRS provides the option for all listed companies to value their property according to the fair value method (IFRS 2012). The fair value method is introduced to create more relevancy and comparability in accounting numbers among countries worldwide (Barth and Clinch 1998, Barth et al. 2001, Hitz 2007).

10 The international comparison of accounting numbers has increased because of the international adoption of IFRS by 283 countries (PWC 2014). The fair value accounting method provides users of financial information relevant information (Barth and Landsman 1995, Barth 2007). This information gives a clear view of the current state of the firm and its expectations of the future (KPMG 2008). The fair value method achieves these statements by measuring the current value of the firm and by giving predictions about future value and cash flows (IFRS 2012). These increases or decreases are not realized yet. This information could be used by investors. This information increases relevance, but it could decrease the reliability. The reliability could be decreased due to the consistent change of the market (Plantin et al. 2008, Laux and Leuz 2009). The fair values measures are sensitive to changes of the market. So the provided information could rapidly change and loose it reliability (Plantin et al. 2008, Laux and Leuz 2009).

IFRS do not have rules about the distribution of the unrealized income that are measured according to the fair value method (IFRS 2012). IFRS leaves this up to the regulations of the involved countries. The rules about the distribution of the unrealized dividend are determined by the jurisdiction of each country (PWC 2014). The regulations regarding the distribution of the unrealized fair value income are different in each country. For example in The Netherlands it is allowed to distribute dividend that is based on the measures according to fair value method (KPMG 2008). In other countries it is not allowed to distribute dividend that is based on the unrealized fair value measures, like for example the United Kingdom and Germany (KPMG 2008). These countries cannot distribute dividend that is based on unrealized income measured by fair value.

This paper will investigate the influence of fair value adjustments on the dividend policy in Germany. Germany has its own German GAAP, but they adopted IFRS in 2005 (PWC 2014). There are a lot of similarities between the German GAAP and the IFRS (PWC 2014) this is due the adoption of rules by both regulators. The application of the IFRS is in Germany mandatory for companies that are listed in any of the EU states. Companies who are not listed in an EU state have according to the German GAAP the option to voluntarily adopt IFRS in their consolidated financial statements (PWC 2014). The difference between IFRS and German GAAP is the valuation method of property (IFRS 2012). IFRS gives organizations the option to value the property both in the fair value method as historical cost method (IFRS 2012). German GAAP requires corporations to value their property in the historical cost method (PWC 2014). The German GAAP makes two exceptions. The first exception is for banks and financial intuitions (PWC 2014). The financial instruments of

11 banks and financial intuitions which are held for trading must be valued at fair value (PWC 2014). The second exception is for assets who are deprived of all other creditor’s access and executively relate to long term liabilities (PWC 2014). This paper will investigate the German firms that use IFRS. The motivation to conduct this research in Germany is because according to the Gross Domestic Product (GDP) Germany is the fourth biggest economy in the world and the biggest economy in Europe (CNN Money 2015). It is in Germany not allowed to distribute dividend that is based on fair value adjustments (KPMG 2008). It will be interesting to find out how German firms deal with this. Furthermore the corporate governance in Germany is stakeholder oriented (IESE 2008). This means that the corporate governance look out for the interests of the whole society (Deegan and Unerman 2006). A stakeholder is a person, group or organization that has interest or concern in an organization (Kormendi, R. and Zarowin 1996). The board in a stakeholder oriented county is also more concerned about the long-term growth and stability (Kormendi, R. and Zarowin 1996). It is interesting to research if Germany distributes dividend that is based on unrealized income, because it is not allowed in Germany to distribute dividend based on unrealized income. Germany is also a code law county. In a code law county rules are systematic list of laws that have been codified and are enforceable by law. In a common law county rules are more based on court rulings (The Economist 2013). It is interesting to find out how the corporate governance of Germany deals with this. The distribution of unrealized income will be monitored by the corporate governance. Germany is a code law country. The corporate governance works better in a common law country (Christopher M. Bruner 2014). It will be interesting to research how the corporate governance in a code law county deals with the distribution of unrealized income when it is not allowed to distribute this income. Another reason why I choose to conduct this research in Germany is because Germany has a unique accounting system (Andres et al. 2009). The distribution of dividend is in Germany different than in other countries (Andres et al. 2009). The dividend distribution in Germany is based on the cash flows (Andres et al. 2009). This is because the German accounting regulations are conservative (Andres et al. 2009). The dividend policies are however very flexible in reducing the dividend. German firms can easily cut the distributed dividend (Andres et al. 2009). The distribution of dividend in Germany is regulates by German Stock Corporation Act (Andres et al. 2009). The amount of dividend that is distributed is decided by the management (Andres et al. 2009). Firms can retain a maximum of 50% of the profits and distribute the other 50 percent (Andres et al. 2009). The management of the parent firm decides the amount of dividend (Andres et al. 2009). It will be interesting to research if fair

12 value adjustments influence the dividend distribution in Germany, which is a country that is very strict in their accounting rules and in the distribution of dividend.

This research contributes to academic literature by investigating the distribution of unrealized income in Germany, a county who does not allow the distribution of unrealized income, and who is more stakeholder oriented (Kormendi, R. and Zarowin 1996).

4 Hypothesis development

Like I mentioned before the IFRS does not have rules about the distribution of unrealized income that is based on fair value measures. These rules are decided by the regulations of the counties (KPMG 2008). According to Lintner’s model (1956) dividend based on unrealized income should only be distributed if the changes in fair value measures are persistence over time. In Germany it is not allowed to distribute dividend based on the unrealized income (KPMG 2008). Unrealized income could however influence the distribution of dividend. Organization can hide the distribution of the unrealized income based on fair value measures. They can give this distribution other reason (DeAngelo 1990). It is difficult to find out if the distribution of dividend is based on the unrealized income (DeAngelo 1990). I will test if the fair value adjustments in Germany have an influence on the dividend policy. This leads to the following research question:

Does unrealized income measured by the fair value method impact the distribution of dividend in Germany?

Fair value measures the expectations of the present and future value of an organization. It also gives a clear view about the economic conditions of an organization (Barth and Clinch 1998, Barth et al. 2001, Hitz 2007). It serves to enhance the transparency and decision making relevance of the financial information (Barth and Clinch 1998, Barth et al. 2001, Hitz 2007). The fair value method also has a downside. Investor needs both relevant and reliable information to make the right decisions (Landsman 2007). The information that is provided by the fair value adjustments is relevant, because it give the stake- and shareholder time to make the right decisions (Landsman 2007). The fair value method includes additional transitory components in the income statement. So the income statements could change in a short timeframe. This could increase the volatility of aggregate income, which could reduce the usefulness of predicting the long-run performance (Landsman 2007). This will decrease the reliability of the provided information (Petroni and Wahlen 1995, Cornett et al. 1996,

13 Hung and Subramanyam 2007). Previous researches have exposed three reasons of increased volatility while using the fair value accounting method. The first reason is not permanent adjustment in the underlying economics (Penman 2007, Plantin et al. 2008). The second reason is that the fair value method is not able to match adjustments in the fair value of assets of assets recognized with the fair method with negatively correlated changes in the fair value of liabilities not recognized with the fair value method (Penman 2007, Plantin et al. 2008). The last reason is the addition of bubble prices into financial statements (Penman 2003). The fair value adjustments could produce information that makes it not possible for stakeholders to judge the future earnings (Sloan 1996, Xie 2001). This will make the use of fair value adjustments even noisier than other forms of financial information (CAS Task force 2002, p. 9). This is because investors get unrealized values without knowing if these values will impact the future earnings. The research of Ball (2006) states that fair value accounting could give more noise to decision-making. This will increases the risks for the users of financial information that is based on fair value measures.

The papers of Lintner (1956) and Jagannathan et al. (2000) predict that only persistent fair value measures should impact the dividend policy. If is fair value adjustments are not persistent and have no effect on the earnings (Ohlson 1999), I predict no link between positive fair value adjustments and dividends. This research is linked to the framework of Lintner (1956).The persistence of fair value measures is critical in this research. This way I can link the distribution of dividends to the persistence. Then I can find out if the fair value measures have an influence on the distribution of dividend. This leads to my hypothesis:

H1: Positive fair value adjustments have no distribution consequences.

There are two assumptions linked to this hypothesis. The first one is that fair value adjustments are not permanent and that the stakeholders estimate the persistence of the fair value adjustments in a right way. Prior research has shown that managers are very optimistic and that they often overestimate the future earnings (Jensen 1993, DeAngelo et al. 1996). There is also research that proves that relevant stakeholders can’t assess the persistence of earnings (Sloan 1996, Hirshleifer et al. 2004). If the situation occurs when the persistence of fair value adjustments is overestimated, positive adjustments could be distributed as part of bottom-line income. This point of critic is also used by Enria et al. (2004) and the European Central Bank (ECB). Enria et al. (2004) state (2004) state that in times of economic growth

14

unrealized fair value adjustments are unfairly distributed as dividends. These organizations will suffer when the fair value adjustments are not realized Enria et al. (2004).

A positive impact of fair value changes on dividends can lead to the rejection of H1. In this situation the fair value accounting could have consequences and lead to more risk. The fair value adjustments during times of economic growth are often positive. This distribution of these fair value adjustments will increase the leverage, because unrealized income is distributed. This situation does occur a lot in banks (IMF 2008). Banks are sensitive to continually change in their unrealized income (Enria et al. 2004). Both the International Monetary Fund (IMF) and the ECB insist that it is important to link the distribution of dividend with core earnings and not with unrealized income (IMF 2008, p. 129). This is important because with the fair value method the organization is distributing money that isn’t realized yet. The distribution of unrealized income will reduce the ability of the banks to react to economic changes. Enria et al. (2004) are in line with this statement. They state that the dividend that is based on fair value adjustments reduces the bank ability to react to changes in the economy. The Banco de Espana is also in line with this statement (Vinals 2008, p. 124). They state that fair value accounting could corrupt the behavior of bank managers. Managers will look out for their own interest and are more sensitive to manipulate the fair value measures.

5 Sample selection

This paper explores a sample of all the public listed firms in Germany. The data is collected from the data collection program DataStream Thomson Reuters. DataStream provides financial information for listed firms in Germany. The information of DataStream is detailed. It is possible to gather information about the fair value, dividend, the net income etc. It provides financial information directly from the issuer. My sample selects data from the time span from 2009 till 2012. 2007 is the first year after the economic crisis (Carmassi, J., Gros, D., Stefano, M., 2009). I have chosen for this time span because it will show if firms are distributing dividends based on fair value adjustments during the financial crisis. The financial crisis lasted from 2007 till 2014 (Carmassi et al. 2009). This time span is also quite recent. The research will give a representation of the current distribution of dividend by German firms. My final collected data consists of 1733 unique firms, which leads to 6921 observations.

The first hypothesis predicts that positive changes in fair value adjustments have no influence on the distribution of dividend. I will first run the descriptive statistics. The use of

15

the descriptive statistics gives a clear view of the distribution of the variables. This is because it controls for unobservable characteristics (Francis and Lennox 2008). The descriptive statistics show the limitations in the data.

Note: The Data consists of 6921 observations of German corporations. NIit (NIit21) is net income in year t (t − 1). REVit means positive fair value adjustments due to revaluations of short-term and long-term financial assets in year t. While DIVit-1 and ΔDIVit are lagged dividends and change in dividends. All variables except indicators are scaled by average total assets.

5.1 Research design

The persistence of fair value measures is important for my hypothesis. This is because the distribution of dividend is linked to the framework of Lintner (1956). The earnings persistence regression of Sloan (1996) is used to measure the increasing fair value measures of future earnings. In this model I regress the net income on lagged net income before revaluations and lagged revaluation income. The regression model is as follows:

NIit = y0 +y1NIBREVit−1 +y2REVit−1 +

ε

In this model: NIit stands for net income in year t, NIBREVit-1 is net income before fair value adjustments in year t-1, and REVit-1 stands for the fair value increase on (revaluation of) short-term and long-short-term financial assets in year t-1. The variables are all scaled by average total assets, which are described in the research of Sloan (1996). The sensitivity of the results will be controlled by rescaling all the variables by lagged total assets and average total assets adjusted for the effect of a revaluation. This rescaling of variables will reduce the risk that revaluation impacts the deflator via the current year’s assets. These extra tests will not impact the interpretation of my analysis.

The effect and sign of y2 is built upon the persistence of the positive fair value changes. When the fair value changes are fully transitory then the coefficient of REVit-1 is predicted to be zero. When fair value changes provides information about future earnings, then y2 must be lager then zero for good news and smaller than zero is there is a decline in future performance.

Table 1 Descriptive statistics

Variable Obs Mean Median Std. Dev. Min Max

NIit 6921 -.0010356 0,060550126 .8793516 -16 47.18269

NIit-1 6921 -.0276788 0,078339616 .8484419 -39 20.42874

REVit 6921 -0.00077 -0,000530866 .0075249 -.4529632 .1430197

DIVit-1 6921 .0001373 0,0035804 .0016425 -.0003591 .0569859

16 The heteroscedasticity-corrected standard errors adjusted for clustering at the firm level is used (Petersen 2009). The results of table 2 show that the fair values don’t predict future income. The fair values are not persistent. So according to the model of Lintner (1956) fair values should only be distributed if they are persistent. In this research fair values are not persistent and should not influence the distribution of dividend. According to the results the German firms cannot distribute dividends based on fair value accounting.

In the next step I will evaluate the effect of the revaluation and historical cost on the dividend. It is difficult to isolate the impact of accounting standards. A well-specified model should be used in order to achieve this (Beatty 2007). It is a challenge to empirical research the accounting choice. The partial adjustment model of Lintner (1956) is used to estimate the connection between earnings components and the dividend policy. The independent variables in this regression are based on the research of Fama and Babiak (1968). The research of Fama and Babiak (1968) use lagged dividend and current and lagged earnings as independent variables. In this research I use total assets as deflator which is based on the Fama and French (2002) paper. This leads to the following regression:

ΔDIVit = a0 +a1NIit +a2NIit−1 +a3DIVit−1 +

ε

In this regression ΔDIVit and DIVit−1 are the adjustments in dividend from year t-1 and lagged dividends. The net income in year t and t-1 is indicated by NIit and NIit−1. The research of Sloan (1996) is used to scale all the variables by average total assets.

In the following step I will separate net income. I will separate the net income into net income before revaluation of financial investments and fair value adjustments. I will use the following formula:

ΔDIVit = β0+ β1NIBREVit+ β2NIBREVit−1 + β3REVit + β4 DIVit−1 + ζit

All the variables have the same meaning as before. I will investigate if H1 holds by looking at

the sign and the statistical significance of the coefficient on the fair value adjustments (REVit). Hypothesis 1 will be evaluated with this formula. When fair value changes are not permanent and stakeholders/investors can estimate their implications for future earnings, then according to hypothesis ,1 fair value adjustments are not distributed and β3 = 0. The pervious results of the persistence of fair value adjustments can influence the expectations. It can occur that β3 is more than 0. However if investors can’t estimate the persistence of fair value adjustments then I expect that β3>0 and that β1 and β3 are equal. Then firms pay the same

17

amount of fair value changes and earnings. The results of the models two and three are shown in table 3 and in table 4.

6 Findings

I will begin with analyzing the first regression. The results of the first regression are displayed in table 2. All the tables in this research are based on 6921 observations. The adjusted R-squared of table 2 is 5.7 %. So 5.7 % of the net income (NIit) is explained by unrealized gain and losses (REVit) and by net Income before fair value adjustments

(NIBREVit-1). This percentage is low because there were not many observations of the fair

value variable. In first regression (table 2) I have tested if the fair values adjustments are persistent in Germany. This information is important to this research, because the research is based on the framework of Lintner (1956). The framework of Lintner (1956) allows the distribution of unrealized income if the fair value adjustments are persistent. The research results show that the fair value adjustments are not persistent. The REVit is not significant

(p=0.914). Like I said before if fair value adjustments have some information about future earnings, y2 (REV it) must be bigger than zero for good news and smaller than zero if fair

value adjustments signal a decrease in future performance. The coefficient of y2 is bigger

than zero. The coefficient is .2463537. In other words y2 (REV it) provides good news about

future earnings. The p-value of REV it is 0,914. This p-value is not significant. According to

the framework of Lintner (1956) firms can only distribute dividend based on unrealized income is the fair value adjustments are persistent. The results of table 2 show that the German firms cannot distribute dividends based on fair value accounting, because the fair value adjustments are not persistent. The Prob > F in table 2 is 0,0000. Unrealized gains and losses due fair value measures (REV it) are not significant. This is could be because there

were not many observations about the fair value measures in Germany. This is a limitation of this research. However according to this research it is clear that the fair value measures are not persistent and should not impact the distribution of dividend in Germany. The relation between the tables in this research and the Lintner (1956) framework is discussed in the next tables. It is also investigated if the fair value adjustments are distributed.

18 Table 2 Earnings persistence analysis

Model 1 NIit = y0 +y1NIBREVit−1 +y2REVit−1 +

ε

Variabels P>(t) Coef. Constant 0.921 .0012046 NIBREVit-1 0.000 .0806277 REVit 0,914 .2463537 N 6921 Adj.R² 0,0057 Prob > F 0.0000

This regression consists of 6920 observations of German corporations. The dependent variable in this regression is NIit net income in year t. The independent variables in this regression are: NIBREV it-1 which means net income before fair value adjustments in year t − 1. REV it-1 means positive fair value adjustments because of revaluations of short-term and long-term financial assets in year t − 1. The variables are all scaled by average total assets. The Adj.R² is the percentage of the regression that is explained by the variables. Prob > F is if the regression is significant. All the t-statistics are based on heteroscedasticity corrected standard errors adjusted for clustering at the firm level.

I will now discus the results of tables 3 and 4. These tables provide results about the influence of the fair value adjustments on the dividend policy in Germany. The Adjusted R-squared of table 3 and table 4 are 58 percent. This means that 58 percent of the results are explained by the variables. This is a high percentage. This means that the results in these tables are well explained by the variables. The prob > F in all tables is 0.000. That means that the result of the regressions is significant. Not all the findings of the tables are consistent with the framework of Lintner (1956) though. Whether H1 holds is dependent on the coefficient and statistical significance of the fair value (REVit). I will compare the results of my research

with the framework of Lintner (1956). REVit is described as β3 in model 3. I will test the

distribution of with the following, β3=0 if fair value adjustments are not distributed. The

results of table 2 indicated that the fair value adjustments are not persistent, because of these results I also expect that β3>0. In contrast with the framework of Lintner (1956) I find that the

coefficient of lagged dividend (DIVit-1) is positive and significant. DIVit-1 has a value of

0.9709789 and the significance is 0.0000. This means that the dividend has an impact on the distribution of dividend. Similar to the framework of Lintner (1956) is the contemporaneous net income (NIit). This variable is positive (0.0000521). Also similar to the Lintner (1956)

framework is the coefficient of income before revaluations (NIBREVit). This variable is also

positive (0.00004350) in my research. Consistent to the Lintner (1956) model is the positive fair value adjustments (REVit). This variable is also negative (-.0021326). According to the

19 research. My results show that fair value adjustments are not distributed. This is because REVit is not equal to zero.

The first test in table 2 pointed out that the fair value adjustments are not persistent in Germany. The second and third test in respectively table 3 and 4 pointed out that the REVit is less than zero and that the fair value adjustments are not distributed in Germany.

The results are in line with the framework of Lintner (1956). The fair value adjustments are not persistent in Germany and are also not distributed.

Table 3 Earnings persistence analysis

Model 2 ΔDIVit = a0 +a1NIit +a2NIit−1 +a3DIV it−1 + εit

Variabels P>(t) Coef. Constant 0.012 .0000391 NIit 0.815 .0000412 NIit−1 0.776 .0000521 DIV it−1 0.000 .8037757 N 6921 Adj.R² 0,5841 Prob > F 0.0000

This regression consists of 6920 observations of German corporations. The dependent variable in this regression is NIit net income in year t. The dependent variable in this regression is ΔDIVit is defined as change in dividends.The Independent variables are: NIit (NIit-1) means net income in year t (t – 1). The variable DIVit-1 stands for lagged dividends. The Adj.R² is the percentage of the regression that is explained by the variables. Prob > F is if the regression is significant. All the t-statistics are based on heteroscedasticity corrected standard errors adjusted for clustering at the firm level.

Table 4 Earnings persistence analysis

Model 2 ΔDIVit = β0+ β1NIBREVit+ β2NIBREVit−1 + β3REVit + β4DIVit−1 + ζit

Variabels P>(t) Coef. Constant 0,012 .0000391 DIVit−1 0.000 .8035718 REVit 0,464 -.0021326 NIBREVit 0.805 .0000435 NIBREVit−1 0.780 -0,0000511 N 6921 Adj.R² 0,5841 Prob > F 0.0000

This regression consists of 6920 observations of German corporations. NIBREVit (NIBREVit-1) stands for net income before fair value adjustments in year t (t – 1). REVit means is positive fair value adjustments due to revaluations of short-term and long-term financial assets in year t. The variable DIVit-1 stands for lagged dividends. The Adj.R² is the percentage of the regression that is explained by the variables. Prob > F is if the regression is significant. All the t-statistics are based on heteroscedasticity corrected standard errors adjusted for clustering at the firm level.

20

7 Conclusion

This research has contributed to literature by examining the connection between fair value adjustments and the dividend policy. I have used the framework of Lintner (1956) to analyze the link between fair value adjustments and the dividend policy. The hypothesis of my research is that if fair value adjustments are not persistent then they should not be distributed. This is in line with the framework of Lintner (1956). This framework is useful because it enables the controlling of core earnings and past dividends. This research is done in Germany, because Germany is the biggest economy in Europe (CNN Money). Germany also has unique accounting regulations (Andres et al. 2009). Germany does not allow the distribution of dividend that is based on fair value adjustments (KPMG 2008). It is also a stakeholder oriented country (IESE 2008). It is interesting to see how such a county deals with the distribution of dividend. The distribution of fair value depends on the accurate estimation of the persistence of the fair value adjustments and their influences on future earnings. I have analyzed 1733 unique firms in a time span of 4 years from 2009 till 2012. These have lead to 6921 observations. The first model (table 2) tested if the fair value adjustments are persistent in Germany. It was clear that the fair value measures were not persistent. According to the framework of Lintner (1956) the fair value adjustments should not be distributed. Models two and three (tables 3 and 4) tested if the fair value measures has an impact on the distribution of dividend. My research shows that fair value adjustments are not distributed in Germany. This is in line with the model of Lintner (1956). The framework of Lintner (1956) only allows the distribution unrealized income if the unrealized income is persistent over time. It is also in line with the dividend distribution policy of the German GAAP (PWC 2014). The research question in this paper is: Does unrealized income measured by the fair value method impact the distribution of dividend in Germany? This research shows that the fair value accounting method does not have an impact on the dividend policy in Germany. This is because the results conclude that the German corporations don’t distribute dividend that is based on fair value measures.

This research does have limitations. The fair value data that was collected in this research consisted of NA (Not available) measures. The information about fair value in these firms was not available. So the fair value measures of these firms were not included in this research. This reduces the generalization of this research. Nevertheless the findings of this research are in line with the paper of Goncharov and Van Triest ( 2011) and with the framework of Lintner (1956). This makes the results more suited for generalization.

21

8 References

Andres, C., Betzer, A., Goergen, M., & Renneboog, L. (2009). Dividend policy of German firms: A panel data analysis of partial adjustment models. Journal of Empirical Finance, 2009, Vol.16(2), pp.175-187 Ball, R. (2006). International financial reporting standards (IFRS): pros and cons for investors. Accounting and

Business Research, International Accounting Policy Forum, 5-27.

Barth, M.E. (2007). Standard-setting measurement issues and the relevance of research. Accounting and

Business Research, 37(1), 7-15.

Barth, M.E. and Clinch, G. (1998). Revalued financial, tangible, and intangible assets: associations with share prices and non-market-based value estimates. Journal of accounting research, 36 (3), 199–233.

Barth, M.E., Beaver, W.H., & Landsman, W.R. (2001). The relevance of the value relevance literature for financial accounting standard setting: another view. Journal of Accounting and Economics, 31(1), 77-104.

Barth, M.E., & Landsman, W.R. (1995). Fundamental issues related to using fair value accounting for financial reporting. Accounting Horizons, 9(4), 97-107.

BBC.com. (2015). Research: Germany country profile. Retrieved 05-08-2015, from

http:/www.bbc.com/news/world-europe-17299607

Beatty, A. ( 2007). How does changing measurement change management behaviour? A review of the evidence. Accounting and business research, 37 (Special issue), 63–71.

Bøhren, Ø., Josefsen, M. G., & Steen, P.E. (2012). Stakeholder conflicts and dividend policy, Journal of Banking & Finance 36, 2852-2864.

Boyer, R. (2007). Assessing the impact of fair value upon financial crises. Socio-economic review, 5 (4), 779–807.

Brav, A., Graham, J.R., Harvey, C.R., & Michaely, R. (2005). Payout policy in the 21st century. Journal of

Financial Economics, 77(3), 483-527.

Carmassi, J., Gros, D.,Stefano, M. (2009). The Global Financial Crisis: Causes and Cures. Journal of Common Market Studies, Volume 47, Issue 5, pages 977–996.

CAS Task force, (2002). White paper on fair valuing property/casualty insurance liabilities. Arlington, VA: Casualty Actuarial Society.

Christopher M. Bruner., (2014). Corporate Governance in the Common-Law World. 77-84

CNN.com (2015). Research. World’s largest economies. Retrieved 05-08-2015, from

http://money.cnn.com/news/economy/world_economies_gdp/

Cornett, M.M., Rezaee, Z., and Terhranian, H. (1996). An investigation of capital market reactions to pronouncements on fair value accounting. Journal of accounting and economics, 22 (1–3), 119–154. Correia da Silva, L., Marc, G., & Renneboog, L. (2004). Dividend policy and corporate governance. Oxford:

Oxford University Press.

Cotter, J.,Zimmer, I. (1995). Asset Revaluations and Assessment of Borrowing Capacity. Abacus, 1995, Vol.31(2), pp.136-151.

Caruana, J. and Pazarbasioglu, C. (2008). Revisiting valuation practices throughout the business cycle: some symmetry is needed [Special issue: Valuation and financial stability]. Banque de France. Financial stability review, 12, pp. 15–22.

DeAngelo, H., & DeAngelo, L. (1990). Dividend policy and financial Distress: an empirical investigation of troubled NYSE firms. Journal of Finance, 45(5), 1415-1431.

DeAngelo, H., DeAngelo, L., and Stulz, R.M. (2006). Dividend policy and the earned/contributed capital mix: a test of the life-cycle theory. Journal of financial economics, 81 (2), 227–254.

Deegan, C. and Unerman, J. (2006), Financial Accounting Theory: European edition.

Do Quynh- Nhu. (2009). Leverage, Growth Opportunities and Stock Price Response to New Financing. International Journal of Business and Management, Vol 4, Iss 9.

Edwards, J. and Mayer, C. (1985). An investigation into the dividend and new equity issue practices of firms: evidence from survey information. London: Institute for Fiscal Studies.

22

Enria, A., Cappiello, L., Dierick, F., Grittini, S., Haralambous, A., Maddaloni, A., Molitor, P., Pires, F., & Poloni, P. (2004). Fair value accounting and financial stability (European Central Bank occasional paper series, 13). Frankfurt: European Central Bank.

Fama, E.F., & Babiak, H. (1968). Dividend policy: an empirical analysis. Journal of the American Statistical

Association, 63, 1132-1161.

Fama, E., and French, K. (2001). Disappearing dividends: changing firm characteristics or lower propensity to pay? Journal of Financial Economics 60, 3–43.

Fama, E., and French, K. (2002). Testing trade-off and pecking order predictions about dividends and debt. Review of financial studies, 14 (1), 1–33.

FASB. (2000). Statement of Financial Accounting Concepts No. 7: using cash flow information and present value in accounting measurements. Norwalk, CT: Financial Accounting Standards Board.

Francis, J.R. and Lennox, C.S. (2008). Selection models in accounting research. Working Paper, University of Missouri–Columbia.

Goergen, M., Renneboog, L., and Correia da Silva, L. (2005). When do German firms change their dividends? Journal of corporate finance, 11 (1–2), 375–399.

Goncharov, I., & van Triest, S. (2011). Do fair value adjustments influence dividend policy? Accounting and

Business Research, 41(1), 51-68.

Grullon, G., Michaely, R., and Swaminathan, B. (2002). Are dividend changes a sign of firm maturity? Journal of business, 75 (3), 387–424.

Gul, F. (1999) Growth opportunities, capital structure and dividend policies in Japan. A journal of Corporate Finance, Vol.5(2), pp.141-168

Gugler, K. (2003). Corporate Governance and Investment. International Journal of the Economics and Business, Vol 10, (3) pp.261-289.

Hail, L., Leuz, C., & Wysocki, P.D. (2010). Political factors and future scenarios for U.S. accounting standards.

Accounting Horizons, 24(4), 567-588.

Hilton, A.S. and O’Brien, P.C. (2009). Inco Ltd.: market value, fair value, and management discretion. Journal of accounting research, 47 (1), 179–211.

Hirshleifer, D., Hou, K., Teoh, S.H., and Zhang, Y. (2004). Do investors overvalue firms with bloated balance sheets? Journal of accounting and economics, 38 (Conference issue), 297–331.

Hitz, J.M. (2007). The decision usefulness of fair value accounting - a theoretical perspective. European

Accounting Review, 16(2), 323-362.

Hung, M., & Subramanyam, K.R. (2007). Financial statement effects of adopting international accounting standards: the case of Germany. Review of Accounting Studies, 12(4), 623-657.

IESE (2008). Research. Shareholder vs. Stakeholder: Two approaches to corporate governance. Retrieved f rom http://www.ieseinsight.com/doc.aspx?id=866

IFRS.org (2012). Technical summary: IAS 40 Investment Property. Retrieved 02-01-2014, from http://www.ifrs.org/IFRSs/Documents/English%20IAS%20and%20IFRS%20PDFs%202012

/IAS%2040.pdf

IMF, (2008). Financial stress and deleveraging: macrofinancial implications and policy. Global financial stability report. Washington, DC: International Monetary Fund, October.

Jagannathan, M., Stephens, C.P., & Weisbach, M.S. (2000). Financial flexibility and the choice between dividends and stock repurchases. Journal of Financial Economics, 57(3), 355-384.

Jensen, M. (1986). Agency costs of free cash flow, corporate finance, and takeovers. American Economic

Review, 76(2), 323– 329. 29.

Jensen, M., & Meckling, W. (1976). Theory of the firm: managerial behavior, agency costs and ownership structure. Journal of Financial Economics, 3(4), 305–360.

Kalbers, L.P. (2009). Fraudulent financial reporting, corporate governance and ethics: 1987-2007, Review of Accounting and Finance, Vol. 8 No. 2 , pp 187-209

Kasanen, E., Kinnunen, J., & Niskanen, J. (1996). Dividend-based earnings management: empirical evidence from Finland. Journal of Accounting and Economics, 22(1-3), 283-312.

Kormendi, R., & Zarowin, P. (1996). Dividend policy and permanence of earnings. Review of Accounting

23

KPMG (January 2008) Feasibility study on alternative to capital maintenance regime established by the Second Company Law Directive 77/91/EEC of 13 December 1976 and an examination of the impact on profit distribution of the new EU accounting regime. Retrieved from the EU single market website:http://ec.europa.eu/internal_market/company/capital/ index_en.htm

Landsman,W.R. (2007). Is fair value accounting information relevant and reliable? Evidence from capital market research. Accounting and Business Research, p.19(12).

Laux, C., & Leuz, C. (2009). The crisis of fair-value accounting: making sense of the recent debate. Accounting,

Organizations and Society, 34(6-7), 826-834.

Lintner, J. (1956). Distribution of incomes of corporations among dividends, retained earnings and taxes.

American Economic Review, 46(2), 97–113.

McCabe, G. (1979). The empirical relationship between investment and financing. Journal of Financial and Quantitative Analaysis 14 (March): 119-35.

Ohlson, J.A. (1999). On transitory earnings. Review of accounting studies, 4 (3–4), 145–162.

Penman, S. (2007). Financial reporting quality: is fair value a plus or a minus? Accounting and Business

Research, 37(1), 33–44.

Petersen, M.A. (2009). Estimating standard errors in finance panel data sets: comparing approaches. Review of financial studies, 22 (1), 435–480.

Petroni, K.R. and Wahlen, J.M. (1995). Fair values of equity and debt securities and share prices of propertyliability insurers. Journal of risk and insurance, 62 (4), 719–737.

Plantin, G., Sapra, H., and Shin, H.S. (2008). Marking-to-market: panacea or Pandora’s box? Journal of accounting research, 46 (2), 435–460

Puleo, A., Smith, S.F., Casey, K.M. (2009). Insurance company dividend policy decisions; Evidence on the role of corporate governance and regulation, Journal of Managerial Finance, Vol 35 No. 6, 2009, pp 493-500.

PWC (October 2014) Research on the adoption and impact of IFRS. Retrieved from the PWC http://www.pwc.com/us/en/issues/ifrs-reporting/publications/ifrs-status-country.jhtml

Renneboog, L., & Szilagyi, P.G. (2007). How relevant is dividend policy under low shareholder protection? Discussion Paper, Tilburg University.

Rozeff, M. (1986). How companies set their dividend payout ratios. In J. Stern and D. Chew (eds.), The Revolution in Corporate Finance, pp. 320-26. New York: Basil Blackwell.

Sloan, R.G. (1996). Do stock prices fully reflect information in accruals and cash flows about future earnings? The accounting review, 71 (3), 289–315.

The Economist.com (2013). Research. What is the difference between common and civil law? From http://www.economist.com/blogs/economist-explains/2013/07/economist-explains-10

Vinals, J. ( 2008). Improving fair value accounting [Special issue: Valuation and financial stability]. Banque de France. Financial stability review, 12, pp. 121–130.

Wallison, P. (2008) Fair value accounting: a critique. American Enterprise Institute for Public Policy Research outlook series. Retrieved from http://www.aei.org/article/27917

Watts, R.L. (1977). Corporate financial statements: a product of the market and political processes. Australian journal of management, 2 (1), 52–75.

Woolridge, J., and Ghosh, C. (1986). Dividend cuts: Do they always signal bad news? In J. Stern and D. Chew (eds.), The Revolution in Corporate Finance, pp. 327-41.