Dividends in the UK: did the adoption of the IFRS
have an impact on the decision to pay dividends for
multinational and domestic companies differently?
Master Thesis
Rick Nijenhuis
University of Groningen: S1798456 Uppsala University: 890531-P752
Email: ricknijenhuis@live.com Supervisor: Dr. J. H. von Eije
1
Abstract
The decision of paying dividends or not by UK firms is analyzed to see if the change from UK GAAP to IFRS has had an effect on it. The decision to pay or not pay dividends has been tested by logit and probit panel regressions over the period from 2000 till 2012. The results indicate that changing to IFRS change the decision to pay dividends more positively for multinational companies compared to domestic companies.1
JEL Classification: G28, G30
Key words: Decision to pay dividends, Multinational and Domestic companies, accounting standards, IFRS
1 I wish to thank my supervisor Dr. Henk von EIje, for providing valued support and thorough
2
1. Introduction
In 2002 the European Union (EU) announced that all the publicly traded firms within the European Union were required to use the International Financial Reporting Standards (IFRS) for preparing their consolidated financial statements. The companies were required to do so for the first time in the fiscal year end of 2005. This change in financial reporting standards was received as large in the accounting world (Larson and Street, 2004; Schipper, 2005; Whittington, 2005; Aharony, Barniv and Falk, 2010). The IFRS were introduced with two main goals, increase in transparency and comparability. These goals were meant to improve the functioning of the capital market. There has been done much research on whether the mandatory adoption of the IFRS in the European Union has reached its intended goals or not. However, the literature on the unintended consequences of the adoption of IFRS is still in its infancy (Brüggemann, Hitz and Sellhorn, 2012). Researching if the adoption of IFRS has reached its intended goals is valuable but researching its unexpected consequences might be just as important. Therefore it is essential to explore this gap in the literature and give firms, managers and policy makers a more complete view of the consequences, intended and unintended, of a change in accounting standards.
3 The link between accounting standards and dividend policies has also been touched upon by KPMG in their research on profit distribution in the United Kingdom (UK) (KPMG, 2006). When KPMG looked at the results of their CFO questionnaire (UK) they noticed how accounting standards could influence the profit distribution, the dividends paid by a firm and how this differs between large (multinational) and small (domestic) companies (Aggarwal and Kyaw, 2010). The questionnaire shows that smaller firms were less likely to differentiate between accounting earnings (transitory earnings included) and real earnings (core earings) because of the costs associated of calculating both types of earnings. Goncharov and van Triest (2011) say that this could lead to a situation in which managers are less likely to pay dividends and use the transitory income as an excuse.
This thesis distinguishes itself in three ways, first by focusing on the decision to pay dividends as the dependent variable. This is an important factor because by measuring the decision to pay dividends with a dummy variable it does not involve dividend amounts. For dividend amounts it might be difficult to find an appropriate denominator2, but the decision to (not) pay does not have this disadvantage. Second I focus on the effect that adopting the IFRS has on the decision to pay dividends of UK firms and if this effect domestic and multinational companies differently. Analyzing the decision to pay (or not pay) dividends in the UK is the third reason why this paper distinguishes itself. The decision to (not) pay dividends has not been researched up until 2012 before. Using the UK market as a sample has several advantages. Firstly, sufficient data about the UK market is available and therefore no problems with the sample size are expected. Secondly, externalities from previous IFRS disclosures are eliminated because firms listed in the UK were not allowed to switch to IFRS before 2005 (Horton and Serafeim, 2010). Thirdly, Von Eije and Megginson (2008) researched the decision to pay dividends for the period 1989 till 2005. I will also investigate if the decision to pay dividends is different for the UK compared to the rest of Europe because of its different law system and see how the decision to pay dividends have further evolved in recent years.
2 To scale the dividend amounts, earnings or total assets are usually used. These become the
4 The thesis shows companies, managers and investors that a change in accounting regulations does not only reach its intended goals but also has unintended consequences. Even more important is that these unintended consequences do not always affect every firm in the same way.
2. Literature review and hypotheses development
The literature review contains literature that links dividends, IFRS and the effect it might have on domestic and multinational companies. Furthermore, out of the literature the hypotheses will be formed. This section will start with a broad introduction of IFRS and its purpose.
2.1. International Financial Reporting Standards
In 2002 the European Union announced the International Accounting Standards (IAS) Regulation. This regulation required all listed firms in the European Union to adopt IFRS from fiscal year 2005 and onwards. This regulation had three goals. The first two goals were to ensure a high degree of transparency and comparability of financial statements. These first two goals would then lead towards the third goal: the improvement of the efficient functioning of the EU capital markets (EC 2002). Besides the two financial reporting objectives and the capital market objective the adoption of IFRS also has two macroeconomic goals. When the capital market would work efficiently the economy would grow and unemployment would decrease (McCreevy, 2005).
2.1.1. Intended consequences of adopting the IFRS
5 achieve its bigger goal. All in all, these findings are rather inconsistent because it would be expected that the regulation achieved better transparency and comparability which would lead to more efficient capital markets or that the regulation would fail in its primary objective which would prevent the capital market from getting more efficient. It is this inconsistency that makes researching the effects of the adoption of IFRS relevant because although it did not work (no increase in transparency and comparability), it did work (increase in efficiency of capital markets, growth and employment).
2.1.2. Unintended consequences
Brüggemann et al. (2012) define unintended consequences as those absent from the IAS Regulation’s explicitly stated objectives and relate them to the contracting uses of IFRS financial statements. The first unintended consequence is the compensation schemes. Compensation schemes often depend on accounting numbers in order to mitigate agency costs. Chen and Tang (2009) provided evidence that gains from adopting IFRS are positively associated with the cash compensation of executives, while losses are not.
A second unintended consequence is lending agreements. Lending agreements contain covenants that are based on the current reporting standards. These covenants have to be changed when the company changes to IFRS. When this change occurs there is a possible wealth transfers between lenders and shareholders (Holthausen and Leftwich, 1983). Christensen, Lee and Walker (2009) find indirect evidence that wealth was redistributed from lenders to shareholders because of the mandatory adoption of IFRS. Other unintended consequences are taxation, banking and rate regulations. However this paper will focus on the unintended consequence dividend payouts.
2.1.3. Unintended consequence: dividend payouts
6 part of the dividend policy of a firm and therefore I will use the literature on the dividend policies as ground to build my hypotheses on.
The effect of the adoption of IFRS on dividend payouts is most likely an indirect effect that is caused by the change in accounting earnings. The EU’s second Company Law Directive’s balance sheet test, states that the maximum amount of distributable profit of EU companies is limited to accumulated accounting earnings. Under the domestic general accepted accounting principles (GAAP), those accounting earnings would have been calculated in the companies’ unconsolidated, legal entity financial statements. When the owners of the a company perceive group earnings as a more precise indicator of performance than the parent firm’s unconsolidated earnings, it is likely that the owners claim their dividends as a portion of the group earnings and not the unconsolidated earnings. Although the adoption of IFRS does not influence the unconsolidated earnings it does change the group earnings and thus it might have an effect on the dividend policy of a firm (Brüggemann et al., 2012).
There is not much empirical research that studies the link between the adoption of IFRS and dividend policies. The closest empirical sources available are two research papers of Goncharov and van Triest (2011, 2013). Their first paper (2013) is a case study about United Energy System (UES), a Russian energy company. This case study is interesting because the Russian Accounting Standards (RAS) were changed in an IFRS-style (mandatory use of fair value) and the companies were forced to disclose the gains or losses that were caused by the change in the accounting standards. UES saw its accounting earnings increase significantly because of the IFRS-style change in RAS and this led to the highest profit in UES’s history. Russian firms have to pay at least 2% of their net profit to the preferred shareholder by law. The only way to avoid this from happening was to omit all dividends for all the shareholders. This case illustrates how a change in accounting standards can strongly affect the dividend policy of a firm.
7 Goncharov and van Triest (2011) provide two reasons for this finding. Firstly, it is possible that managers used the increase in transitory earnings caused by the change in accounting standards to justify a lower level of dividends. Managers prefer lower levels of dividends because this leaves more funds in the firm, which they can at their own discretion (Pinkowitz, Stulz and Williamson, 2006). However, the managers are aware that this might lead to negative stock market reactions (Ghosh and Woolridge, 1988) and damages the company’s reputation on the financial markets (Gomes, 1996). Managers may then explain the relatively lower dividends to the shareholders by referring to unrealized and transitory fair value components. Their second explanation suggests that an unobservable managerial response to high growth might correlate with the fair value gains. Managers perceive the fair value gains as an unsustainable growth. The authors argue that if this is the case, maintaining a dividend payout based on the level of income before the accounting standards changed might increase the firm’s riskiness, which could explain a relatively lower dividend payout. Although, Goncharov and van Triest (2011) did not empirically test the two explanations, their results do indicate that the use of IFRS-style accounting (in this case the increase of usage of fair value accounting) can lead to an increase in transitory earnings and consequently upset the longstanding relations between accounting earnings and dividends. Hence, a switch from domestic accounting standards to IFRS can lead to changes in dividend policies of companies (Brüggemann et al., 2012).
2.1.4. The difference between UK GAAP and IFRS
8 The first major difference is caused by classifying leases differently under IFRS (IAS 17) compared to how they were classified under UK GAAP. Some leases that were classified as operational leases under UK GAAP are classified as finance leases under IFRS. Because of this, finance income is differently recognized under IFRS and thus the reported leverage will differ as well. The second difference comes from IAS 19 that causes for most UK firms to transfer the pension surplus or deficit from their notes to the balance sheet and report an extra charge to earnings. The change in share-based payments is the third major difference between IFRS and UK GAAP. IFRS 2 increased the expenses for compensation and therefore reduced the firms’ earnings. For deferred tax, the fourth difference, IFRS differs from the UK GAAP mostly in a conceptual way. Under IFRS deferred tax is more likely to be recognized because it uses a wider scope and removes discounting. Because of this IFRS increases deferred tax liability in general and this can lead to a decrease in shareholders’ equity. Under IFRS (IAS 38) automatic amortization of goodwill is replaced for a test for impairment of goodwill. Normally goodwill would be amortized by a certain percentage per year but under IFRS companies have to value their intangible assets by performing an impairment test. This impairment test values the goodwill according to the market value instead of the historical value. Because UK firms are forced to not only reconcile after 2005 but also the year previous to the adoption year of IFRS the biggest adjustment will be to reverse the previously reported amortization of goodwill and replace it with an impairment test. According to Horton and Serafeim (2010) this will lead to higher but more volatile earnings. The last major difference between UK GAAP and IFRS is their way of reporting financial instruments. Under IFRS it is expected that the carrying values of assets and liabilities in the balance sheet will change because they must be valuated according to market prices. Besides these six major differences there are countless other differences between the UK GAAP and IFRS (Ernst & Young, 2011).
2.2. The link between IFRS and the decision to pay dividends
9 case only the independent explanatory variables (and in particular earnings) may then still be affected by the IFRS but the dependent variable is not. This is because the measure is not reported in amounts and therefore not influenced by the IFRS. Common determinants of dividend policies are size, growth opportunities and profitability (Fama and French, 2001; Denis and Osobov, 2008). Since the decision to pay (or not pay) dividends is part of the dividend policies of companies I will use the literature on dividend policies, including dividend payouts, to build the hypotheses on.
The link between accounting information and dividends originates from the work of Lintner (1956). Lintner’s partial adjustment model explains how firms try to pay out a certain percentage of their core earnings as dividends and that those companies only slowly change their dividends over time. Many authors have tested this model over time, by modeling dividends as a function of core earnings and lagged dividends, and find evidence to support it in the USA (DeAngelo, DeAngelo and Skinner,1992; Brav, Graham, Harvey and Michaely, 2005) and also in Germany (Goergen, Renneboog and Correia da Silva, 2005). Thus, the Lintner model (1956) shows that core earnings influence dividends but the transitory part of the earnings do not (Jagannathan, Stephens and Weisbach, 2000). In line with this reasoning, Kormendi and Zarowin (1996) find evidence that companies with higher core earnings paid out more dividends. Edwards and Mayer (1985) find similar results; UK firms only reduce their dividends when their core earnings decline. The link between core earnings, IFRS and dividends is implies that there should not be a change in the decision to pay dividends by a change in accounting standards because the change in accounting standards does not affect core earnings. The first set of hypotheses is as followed:
H10: The change from UK GAAP to IFRS did not affect the decision to pay (or not
pay) dividends of listed firms in the UK.
H1a: The change from UK GAAP to IFRS did affect the decision to pay (or not pay)
dividends of listed firms in the UK.
10 only influence transitory earnings and not the core earnings. I reject the null hypothesis and accept the alternative hypothesis when the results are in line with the evidence that Goncharov and van Triest (2011) find in Russia.
2.3. Multinational companies versus domestic companies
The second part of this thesis will look at whether the effect of IFRS on the decision to pay dividends of UK firms is different for multinational firms compared to domestic firms. Since there is little direct literature on the effect of IFRS on the decision to pay dividends of firms, a logical consequence is that there is no literature that directly researches the different effect IFRS might have on the decision to pay dividends of multinational and domestic companies. Evidence does suggest that there is a clear advantage for multinational firms to change to IFRS compared to domestic companies (Gazzar, Finn and Jacob, 1999). Gazzar et al. (1999) point out that multinational companies profit by using IFRS as their accounting standards because they enhance their exposure to foreign markets, improve their customer recognition, are able to secure more foreign capital and reduce their political costs of doing foreign business. Hence, it is reasonable to assume that if IFRS actually did improve the core earnings for multinational companies and that they would profit more from the change to IFRS than domestic companies.
Another reason why there might be a difference between domestic and multinational companies is that companies only have to apply the IFRS to their consolidated accounts and to not the individual accounts of their subsidiaries. Since it is very costly to both have UK GAAP and IFRS figures smaller companies often do not put that much effort in providing both figures compared to bigger (multinational) companies (KPMG, 2006). It is more likely for a multinational to disclose both the real earnings (core earnings) and accounting earnings (earnings with a transitory part). This could result, as Goncharov and van Triest (2011) already used to explain their findings, in a situation in which managers explain the rise in income as transitory and this would lead to fewer domestic companies paying dividends.
H2o: The effect on the decision to pay (or not pay) dividends from changing from UK
11 H2a: The effect on the decision to pay (or not pay) dividends from changing from UK
GAAP to IFRS does not affect multinational companies and domestic companies equally.
The null hypothesis will be accepted when neither multinational nor domestic companies gain (lose) from the change in accounting standards or when the change affects both multinationals and domestic companies in equal ways. The null hypothesis will be rejected when multinationals make the decision to pay dividends more frequently because of the change to IFRS. Alternatively, the null hypothesis will also be rejected when domestic companies decide to pay dividends more regularly than multinational companies.
2.4. The decision to pay dividends of firms listed in the UK
The decision to pay (or not pay) dividends of UK firms will be analyzed by analyzing the data and controlling for profitability, investment opportunities, size (Fama and French, 2001), repurchases of shares (Skinner, 2008) and dividend premium (Baker and Wurgler, 2004).
2.4.2. Dividends in the UK
The USA has usually been the most researched market over the last couple of decades in finance and dividend policies are no exception. Fama and French (2001) discovered that the proportion of dividend paying companies declined heavily between 1978 and 1999 (from 66.5% to 20.8%) in the USA. A part of this change was caused by the changing characteristics of the companies in the USA but after controlling for this change in characteristics Fama and French (2001) still find significant evidence of lower propensity to pay dividends.
12 to the rest of Europe, according to the agency cost model (LaPorta, Lopez-de-Silanes, Shleifer, and Vishny, 2000). This is in line with what Ferris, Sen and Yui (2006) find. They also find a decline in the propensity to pay dividends of UK firms, however, they find it to be less strong than in the USA.
3. Methodology
The methodology is ordered as follows. The first part, data and sample, will describe the data this thesis utilizes. The research design section describes how the hypotheses will be tested and provides descriptive statistics.
3.1. Data and sample
All the data are collected from Datastream. The sample consists out of all the companies listed on the London Stock Exchange from 1999 to 2012 except financial companies (sic code 6000-6999) and utility companies (sic code 4900-4999). Financial and utility companies have been excluded from the sample because of regulatory issues (Ferris et al., 2006). Additionally, all firms without an International Securities Identifying Number (ISIN), without an industry code or do not use Pound Sterling as their reporting currency were excluded. The data does include “live” and “dead” companies in order avoid survivor bias. In Appendix 1 the Worldscope codes and the corresponding Datastream data types can be found.
13 The effect of outliers has been reduced by winsorizing extreme values at the 1st and 99th percentiles (Graham, Michaely and Roberts, 2003). To tackle endogeneity problems all independent variables are lagged once (except if they are time invariant or represent time itself). The data that is used is called panel data. The advantage of panel data is that it makes it possible to control for variables that cannot be observed or measured, like cultural factors or difference business practices, and thus accounts for individual heterogeneity. Appendix 2 shows that the coefficients of correlation between the independent variables are below 0.7, so there is not much evidence of a problem of multicollinearity.
3.2.1. Research design
Von Eije and Megginson (2008) shows that the amount of firms paying dividends was decreasing over a period from 1989 to 2005. The analysis will start by analyzing the further evolution of this trend. After that the decision to pay dividends will be analyzed by exploring the relationship of them with an IFRS dummy variable, a dummy that indicates if a company is domestic or multinational, and interaction variable between the IFRS and multinational dummy.
The decision to (not) pay dividends will be measured with a dummy variable that takes the value of 1 if a firm pays dividends and 0 if the firm does not pay dividends. Because the dependent variable is a dummy variable a binary estimation method is used. According to Brooks (2008, page 518) using a logit or probit model will give similar characterizations of the data because the densities are very similar. In the end the using a logit or probit model will lead to the same results except when the dummy variable is very unbalanced. The dummy variable that measures the decision to pay dividends is balanced sufficiently (see table 1.) and therefore both the probit and logit models will be tested. The results of the probit model are used in the analysis and the results of the logit model serve as robustness test and can be found in Appendix 3.
14 accounting standard. Originally, the expectation was that there should be only one switching year, namely 2005, because all the firms within the EU were required to switch for the first time in the fiscal year end of 2005. UK firms were not even allowed to switch to IFRS prior to 2005, however it is not mandatory for UK firms to have 31 December year-ends and therefore it might occur that their first IFRS reports relate to the following year (Nobes, 2009). This would still only lead to two possible switching years but the London Stock Exchange (LSE) has a sub-market called the Alternative Investment Market (AIM) that was not obliged to switch to IFRS until 2007. Because of the different year-ends for the companies listed at the AIM there are four possible switching years for companies listed in the UK (See Table 1).
The amount of foreign sales to total sales will indicate whether a company is domestic or multinational. When this percentage will be higher than 20% the firm will be regarded as a multinational company (Michel and Shaked, 1986). The 20% point is not only obtained from the paper of Michel and Shaked (1986) but also happens to be about the median of the international sales to total sales ratio. According to this percentage a dummy variable will be created where 1 represents a multinational company and 0 a domestic company.
15 Table1.
Number of observations for the dependent variable, the decision to pay dividends, and the two main variables, variable that indicates if a company is multinational or domestic and a variable that indicates when companies switch from UK GAAP to IFRS by listed companies in the UK from 2000 to 2012.
Year #Dividend Payers #Not Dividend Payers #Multinational Companies #Domestic Companies
Number of companies that have switched from UK GAAP to IFRS
2000 835 772 440 520 0 2001 765 889 481 536 0 2002 710 984 505 526 0 2003 686 1057 517 546 0 2004 681 1136 525 582 0 2005 682 1188 543 578 265 2006 647 1218 555 580 355 2007 631 1159 587 517 313 2008 607 1063 585 458 341 2009 526 1030 572 409 0 2010 497 975 569 401 0 2011 502 916 563 350 0 2012 489 862 526 265 0 Total 8258 13249 6968 6268 1274
3.2.2. Control variables and regression
16 between the decision to pay dividends and investment opportunities because firms generally prioritize investing over paying out dividends. A firm with a lot of investment opportunities is therefore less likely to pay dividends.
17 DDPit = α0 + α1 IFRSit + α2 DMULTit + α3 IFRSit*MULTit + α4 SIZEit-1 + α5 PROFit-1 +
α6 MTBVit-1 + α7 ∆Tai-1t + α8 DREPit + α9 REPUit-1 + α10 CRISt + α11 DPREt + εit
(1)
Where
DDPit = Dummy variable: When Divit >0 then 1, Divit =0 then 0
IFRSit = Dummy variable: When 1 after SY and 0 before
DMULTit = Dummy variable: When ISTSit>20% then 1, ISTSit<20% then 0
SIZEit = Log (TAit)
PROFit = OIit /TAit
MTBVit = Vit /TAit
Vit = TAit + MEit – BEit
∆TAit = (TAit –TAit-1)/TAit
DREPit = Dummy variable: When Repit >0 then 1, Repit =0 then 0
REPUit= Rep divided by TAit
DPREt= Log(VWMDPt)-Log(VWMNDPt)
CRISt= Dummy variable: When t=2000 till 2007 then 0, t=2008 till 2012 then 1
And
TAit=Total Assets for the firm i at time t
Repit= (gross) Repurchases of stocks for the firm i at time t
Divit= Cash Dividends for the firm i at time t
OIit= Operational Income for the firm i at time t
MEit= Market Value Equity for the firm i at time t
BEit=Book Value Equity for the firm i at time t
VWMDPt= Value weighted mean MTBV dividend payers at time t
VWMNDPt= Value weighted mean MTBV non-dividend payers at time t
18 Table 2.
Descriptive Statistics from companies listed in the UK from 2000 to 2012 (Winsorized).
DDP is a dummy variable that takes the value of 1 if a firm pays dividends and 0 if a firm does not pay dividends. IFRS is a dummy variable that takes the value of 1 if a firm uses IFRS as its accounting standard and 0 if it takes UK GAAP as its accounting standard. DMULT is a dummy variable that takes the value of 1 if 20% of the firm’s sales consist out of international sales and 0 if the firm’s international sales is less than 20% of its total sales. IFRS*MULT is an interaction variable that indicates how much affect IFRS has on the decision of multinational companies. PROF is calculated as operating income divided by total assets. SIZE is calculated as the natural log of total assets. MTBR is the market to book value of the firm. ∆TA is calculated as the relative change in assets. DREP is a dummy variable that takes the value on 1 if a firm repurchases dividends and 0 if a firm does not repurchase dividends. REPU is calculated as repurchases of shares divided by total assets. DCRIS is a dummy variable that has a value of 1 from 2008-2012 and 0 from 2000-2007. DPRE is the dividend premium. Std. Dev. indicates the standard deviation. Obs. indicates the total observations.
Mean Median Maximum Minimum Std. Dev. Obs.
DDP 0.454 0 1 0 0.498 18192 IFRS 0.370 0 1 0 0.483 16900 DMULT 0.527 1 1 0 0.499 13231 IFRS*DMULT 0.247 0 1 0 0.431 12537 PROF -0.091 0.037 0.377 -2.809 0.445 18484 SIZE 10.448 10.317 16.663 4.694 2.414 18520 MTBV 2.335 1.42 23.179 0.484 3.107 16892 ∆TA 0.006 0.05 0.927 -3.161 0.536 17147 DREP 0.143 0 1 0 0.35 17959 REPU 0.007 0 21.694 0 0.18 17957 CRIS 0.308 0 1 0 0.462 39351 DPRE 0.27 0.281 0.706 -0.45 0.25 39351
19
4. Analysis
4.1. The decision to pay dividends in the UK
Von Eije and Megginson (2008) show a decrease in firms paying dividends in their period (1989-2005) in Europe. Figure 1 and 2 show us that this decline has stopped after 2005 and has since then stabilized. Table 4 shows the results of the probit panel regression that tests the two main sets of hypotheses. Table 4 shows that variables influencing the decision to pay dividends in the UK are similar to the EU. Size and profitability of a firm are positively related to the decision to pay dividends and investment opportunities are negatively related. These findings are in line with the results of Von Eije and Megginson (2008) and indicate that the UK firms represents the firms in the EU well regarding the decision to (not) pay dividends.
Fig 1. Absolute amount of firms paying and not paying dividends between 2000 and 2012 in the UK.
Dividend Payers indicates the total amount of firms that pay dividends, NonDividend Payers indicates the total amount of firms that do not pay dividends and Total Observations indicates the total amount of observations there are on dividend and not dividend paying firms.
20
4.2. The effect of switching to IFRS
The first set of hypotheses have been tested by a probit and logit panel regression and the results of the probit regression can be observed in table 4 (Panel A) while the results from the logit panel regression will be used as an robustness test and are in Appendix 3. The null hypothesis of the first set of hypotheses can be rejected according to table 4, panel A. This means that the switch from UK GAAP to IFRS did have a significant effect on the decision to pay dividends. This finding is in line with Goncharov and van Triest (2011) because the coefficient is negative. This means that the switch to IFRS made it less likely for firms to pay dividends which fits with their results that the change in accounting standards in Russia was negatively related to dividends changes.
The results, however, change when I add the dummy variable, that indicates if a company is multinational or domestic, to the model (panel B). Apparently when the model controls for being multinational or domestic the effect of changing to IFRS is not significant anymore. Goncharov and van Triest (2011) did not control for multinationality in their paper and this could be a finding that would explain this phenomena in a more complete manner. Remarkably, moreover, is that multinational firms from the UK are less likely to pay dividends. To test the robustness of this surprising finding I will measure the degree of multinationality in three different Fig 2. Relative amount of firms paying and not paying dividends between 2000 and 2012 in the UK. %DP
indicates the relative amount of companies that pay dividends and %NDP indicates the relative amount of firms that do not pay dividends.
21 ways. The first two different measures and the associated results are in Appendix 4. In panel A multinationality is measure by the ratio of international sales to total sales and in panel E it is measured by a dummy variable that takes the value of 1 if the international sales to total sales ratio is more than 0% and 0 if the ratio is 0%. Appendix 4 shows similar results, indicating that the negative relationship between the decision to pay dividends and the multinationality of a firm is negative. The results of the third regression are in panel D of table 4. This robustness test looks if the relationship between international sales to total sales ratio and decision to pay dividends is quadratic. The marginal significance of the quadratic variable suggests that the relationship is possibly quadratic. Moreover, the coefficient becomes positive which means that multinationality has a positive relationship with the decision to pay dividends. These findings are in line with previous research of Aggarwal and Kyaw (2010) who only find marginal evidence on the positive relationship between dividend policies and multinationality.
The interaction variable between IFRS and MULT in panel C of table 4 tests the second set of hypotheses; does the change to IFRS has different consequences for the decision to pay dividends between domestic and multinational companies? At a significance level of 5% the null hypothesis of the second set of hypotheses is rejected and the alternative hypothesis is accepted which means that the change from UK GAAP to IFRS does not affect the decision to pay dividends of domestic and multinational firms equally. Because of the marginal quadratic nature of the relationship between multinationality and decision to pay dividends I also checked if the interaction variable was quadratic of nature (panel D). IFRS*MULT2 is the variable that measures the quadratic relationship and is significant at a 1% level which strengthens the assumption that the nature of the relationship between multinationality and decision to pay dividends is quadratic.
22 Table 4.
Probit panel regression coefficients for the decision to pay dividends for 2000-2012 in the UK.
The dependent variable is a dummy variable that takes the value of 1 if a firm pays dividends and 0 if it does not pay dividends. Co indicates the coefficient of the variable. * indicates a significance level at 10%, ** at 5% and *** at 1%. An L before the variable indicates that the variable is lagged by one year. IFRS is a dummy variable that takes the value of 1 if a firm uses IFRS as its accounting standard and 0 if it takes UK GAAP as its accounting standard. DMULT is a dummy variable that takes the value of 1 if 20% or more of the firm’s sales consist out of international sales and 0 if they are less than 20%. IFRS*DMULT is an interaction variable between IFRS and DMULT. MULT is the international sales to total sales ratio. MULT2 is the international sales to total sales ratio in quadratic form. IFRS*MULT is an interaction variable between IFRS and MULT. IFRS*MULT2 is the quadratic form of the interaction variable between IFRS and MULT. PROF is calculated as operating income divided by total assets. SIZE is calculated as the natural log of total assets. MTBR is the market to book value of the firm. ∆TA is calculated as the relative change in assets. DREP is a dummy variable with the value of 1 if a firm repurchases shares and 0 if the firm does not. REPU is calculated as repurchases of shares divided by total assets. DCRIS is a dummy variable that has a value of 1 from 2008-2012 and 0 from 2000-2007. DPRE is the dividend premium. Cons is the intercept of the regression equation. Total Obs is the total amount of observations for which all information was available. Prob indicates the significance of the equation and McF R2 is the McFadden R-squared. Panel A B C D Co Co Co Co IFRS -0.081** -0.042 -0.110** IFRS -0.206*** DMULT -0.188*** -0.245*** MULT -0.007*** MULT2 0.000* IFRS*DMULT 0.125** IFRS*MULT 0.016*** IFRS*MULT2 0.000*** LPROF 6.746*** 7.673*** 7.676*** LPROF 7.662*** LSIZE 0.250*** 0.247*** 0.248*** LSIZE 0.264*** LMTBV -0.120*** -0.141*** -0.141*** LMTBV -0.135*** L∆TA -0.210*** -0.098** -0.101** L∆TA -0.069 DREP 0.462*** 0.411*** 0.408*** DREP 0.388*** LREPU -0.052 -0.237 -0.233 LREPU -0.188 DCRIS -0.492*** -0.472*** -0.470*** DCRIS -0.433*** DPRE -0.170*** -0.180*** -0.179*** DPRE -0.170** Cons -2.456*** -2.344*** -2.329*** Cons -2.446***
Total Obs 14351 11184 11184 Total Obs 11184
Prob(LR statistic) 0.443 0.427 0.427 Prob(LR statistic) 0.000***
McF R2 0.000 0.000 0.000 McF R2 0.437
23 All the control variables have the expected coefficients and significance levels with two exceptions. First, the REPU variable is not significant, however the main reason to include the REPU variable was for correcting the transitory part of earnings so any significance was not expected. Second, the negative coefficient of the dividend premium (DPRE) gives us reason to believe that the catering theory does not hold in the UK, which is not unexpected because Von Eije and Megginson (2008) came to the same conclusion when analyzing the EU. The decision to repurchase shares (DREP) has a positive coefficient which indicates that repurchase of shares are not replacing dividends but rather supplement them, which is in line with the literature of Floyd, Li and Skinner (2013).
The other control variables reconfirm that companies that pay dividends are big, profitable companies with a relatively low market to book ratio and fewer investment opportunities. Finally, the crisis dummy does not only control for the crisis years but also indicates that the crisis has a negative effect on the decision to pay dividends for firms listed in the UK. The results from the logic regression (see Appendix 3), that serve as a robustness check, indicate similar coefficients and significance levels.
5. Conclusions
This paper shows that the decline in companies paying dividends have stopped around 2005 and has been stable since then. The UK also turns out to be a good reflection of the EU with regard to the decision to pay dividends. Out of the first set of hypotheses the null hypothesis was rejected and the alternative hypothesis accepted, which means that the change to IFRS had a negative impact on the decisions of firms to pay dividends listed in the UK. This finding contribute to the literature on unintended consequences of the implementation of IFRS and are in line with the findings of Goncharov and van Triest (2011).
24 decision to pay dividends of multinational companies was affected more positively by the change to IFRS than that of the domestic companies.
The limitation of this thesis is that I cannot prove why the decision to pay dividends of multinational companies are affected more positively by the change to IFRS compared to domestic companies. This limitation is at the same time an opportunity for future research. The empirical evidence on the unintended consequences of the adoption of IFRS is still in its infancy, especially compared to the empirical evidence on the intended goals of the IFRS.
The practical implication of this paper for firms, managers, policy makers and investors is that it shows that a change in accounting standards do not only have an impact on their intended goals but can also have unintended consequences that do not affect everyone equally.
References
Aggarwal, R., Kyaw, N. A., 2010. Capital structure, dividend policy, and multinationality: Theory versus empirical evidence. International Review of Financial Analysis 19(2), 140-150.
Aharony, J., Barniv, R., Falk, H., 2010. The impact of mandatory IFRS adoption on equity valuation of accounting numbers for security investors in the EU. European Accounting Review 19(03), 535-578.
Aisbitt, S., 2006. Assessing the effect of the transition to IFRS on equity: the case of the FTSE 100. Accounting in Europe 3(1), 117-133.
Baker, M., Wurgler, J., 2004. A catering theory of dividends. The Journal of Finance 59(3), 1125-1165.
25 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.
Brooks, C., 2008. Introductory econometrics for finance. Cambridge university press 518.
Brüggemann, U., Hitz, JM., Sellhorn, T., 2012. Intended and unintended consequences of mandatory IFRS adoption: A review of extant evidence and suggestions for future research. SFB 649 Discussion Paper, 2012-011.
Callao, S., Jarne, J. I., 2010. Have IFRS affected earnings management in the European Union? Accounting in Europe 7(2), 159-89.
Chen, K. C. W., Tang, F., 2009. Do firms use the unrealized gains mandated by IFRS to increase executive cash compensation? Evidence from family-owned property companies in Hong Kong. Working paper. Hong Kong University of Science and Technology.
Christensen, H. B., Lee, E., Walker, M., 2009. Do IFRS reconciliations convey information? The effect of debt contracting. Journal of Accounting Research 47(5), 1167-99.
DeAngelo, H., DeAngelo, L., Skinner, D. J., 1992. Dividends and losses. The Journal of Finance 47(5), 1837-1863.
Denis, D. J., Osobov, I., 2008. Why do firms pay dividends? International evidence on the determinants of dividend policy. Journal of Financial Economics 89(1), 62-82.
26 Edwards, J., Mayer, C., 1985. An investigation into the dividend and new equity issue practices of firms: evidence from survey information. Institute for Fiscal Studies (IFS).
El-Gazzar, S. M., Finn, P. M., Jacob, R., 1999. An empirical investigation of multinational firms' compliance with international accounting standards. The International Journal of Accounting 34(2), 239-248.
Ernst and Young (2011) “UK GAAP vs. IFRS: The basics”
Fama, E. F., French, K. R., 2001. Disappearing dividends: changing firm characteristics or lower propensity to pay? Journal of Financial economics 60(1), 3-43.
Ferris, S. P., Sen, N., Yui, H. P., 2006. God Save the Queen and Her Dividends: Corporate Payouts in the United Kingdom*. The Journal of Business 79(3), 1149-1173.
Floyd, E., Li, N., Skinner, D., 2011. Payout policy through the financial crisis: The growth of repurchases and the resilience of dividends. Chicago Booth Research Paper 12 (01).
Ghosh, C., Woolridge, J. R.,1988. An analysis of shareholder reaction to dividend cuts and omissions. Journal of Financial Research 11(4), 281-294.
Goergen, M., Renneboog, L., Correia da Silva, L., 2005. When do German firms change their dividends? Journal of Corporate Finance 11(1), 375-399.
Gomes, A., 1996. Dynamics of stock prices, manager ownership, and private benefits of control. Manuscript. Harvard University.
27 Goncharov, I., van Triest, S., 2013. Unintended consequences of changing accounting standards: the case of fair value accounting and mandatory dividends. Working Paper. WHU – Otto Beisheim School of Management and Amsterdam Business School.
Graham, J. R., Michaely, R., Roberts, M. R., 2003. Do price discreteness and transactions costs affect stock returns? Comparing ex-dividend pricing before and after decimalization. The Journal of Finance 58(6), 2611-2636.
Grullon, G., Michaely, R., 2002. Dividends, share repurchases, and the substitution hypothesis. The Journal of Finance 57(4), 1649-1684.
Holthausen, R. W., Leftwich, R.W., 1983. The economic consequences of accounting choice: implications of costly contracting and monitoring. Journal of Accounting and Economics 5, 77-117.
Horton, J., Serafeim, G., 2010. Market reaction to and valuation of IFRS reconciliation adjustments: first evidence from the UK. Review of Accounting Studies 15(4), 725-751.
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.
Kormendi, R., Zarowin, P., 1996. Dividend policy and permanence of earnings. Review of accounting studies 1(2), 141-160.
KPMG, 2006. Contract ETD, feasibility study on capital maintenance, Main Report.
Lang, M., Maffett, M., Owens, E., 2010. Earnings comovement and accounting comparability: The effects of mandatory IFRS adoption. Working paper. UNC-Chapel Hill and University of Rochester.
28 Larson, R. K., Street, D. L., 2004. Convergence with IFRS in an expanding Europe: progress and obstacles identified by large accounting firms’ survey. Journal of International Accounting, Auditing and Taxation 13(2), 89-119.
Lee, B. S., Rui, O. M., 2007. Time-series behavior of share repurchases and dividends. Journal of Financial and Quantitative Analysis 42(01), 119-142.
Li, S., 2010. Does mandatory adoption of International Financial Reporting Standards in the European Union reduce the cost of equity capital? The Accounting Review 85(2), 607-36.
Lintner, J., 1956. Distribution of incomes of corporations among dividends, retained earnings, and taxes. The American Economic Review 46(2), 97-113.
Márquez-Ramos, L., 2011. European accounting harmonization: consequences of IFRS adoption on trade in goods and foreign direct investments. Emerging Markets Finance and Trade 47(5), 42-57.
McCreevy, C., 2005. IFRS: No pain, no gain? European Commissioner for Internal Market and Services Speech 05-621. Brussels, 18 October.
Michel, A., Shaked, I., 1986. Multinational corporations vs. domestic corporations: Financial performance and characteristics. Journal of International Business Studies 89-100.
Nobes, C. W., 2009. Observations on measuring the differences between domestic accounting standards and IAS. Journal of Accounting and Public Policy 28(2), 148-153.
29 Schipper, K., 2005. The introduction of International Accounting Standards in Europe: implications for international convergence. European Accounting Review 14(1), 101–126.
Schleicher, T., Tahoun, A., Walker, M., 2010. IFRS adoption in Europe and investment cash flow sensitivity: outsider versus insider economics. The International Journal of Accounting 45(2), 143-68.
Skinner, D. J., 2008. The evolving relation between earnings, dividends, and stock repurchases. Journal of Financial Economics 87(3), 582-609.
30
Appendix 1.
Description of variables gathered from Datasteam
Name Variable Name in Datasteam Worldscope Code
Cash Dividends Cash Dividends Paid #04551
Total Assets Total Assets #02999
Operating Income Operating Income #01250
Accounting Standards Accounting Standards Followed #07536
International Sales International Sales #07101
Net Sales Net Sales or Revenue #01001
Market Value Equity Market Capitalization #08001
Book Value Equity Common Equity #03501
Gross Repurchases Common/Preferred Redeemed, Retired, Converted #04751
Appendix 2.
Correlation matrix independent variables (Winsorized)
IFRS is a dummy variable that takes the value of 1 if a firm uses IFRS as its accounting standard and 0 if it takes UK GAAP as its accounting standard. DMULT is a dummy variable that takes the value of 1 if 20% or more of the firm’s sales consist out of international sales and 0 if they are less than 20%. IFRS*DMULT is an interaction variable between IFRS and DMULT. PROF is calculated as operating income divided by total assets. SIZE is calculated as the natural log of total assets. MTBR is the market to book value of the firm. ∆TA is calculated as the relative change in assets. REPU is calculated as repurchases of shares divided by total assets. DREP is a dummy variable with the value of 1 if a firm repurchases shares and 0 if the firm does not. DCRIS is a dummy variable that has a value of 1 from 2008-2012 and 0 from 2000-2007. DPRE is the dividend premium.
IFRS DMULT IFRS*DMULT PROF SIZE MTBV ∆TA REPU DREP CRIS DPRE
31 Appendix 3.
Logit panel regression coefficients for the decision to pay dividends for 2000-2012 in the UK.
The dependent variable is a dummy variable that takes the value of 1 if a firm pays dividends and 0 if it does not pay dividends. Co indicates the coefficient of the variable. * indicates a significance level at 10%, ** at 5% and *** at 1%. An L before the variable indicates that the variable is lagged by one year. IFRS is a dummy variable that takes the value of 1 if a firm uses IFRS as its accounting standard and 0 if it takes UK GAAP as its accounting standard. DMULT is a dummy variable that takes the value of 1 if 20% or more of the firm’s sales consist out of international sales and 0 if they are less than 20%. IFRS*DMULT is an interaction variable between IFRS and DMULT. MULT is the international sales to total sales ratio. MULT2 is the international sales to total sales ratio in quadratic form. IFRS*MULT is an interaction variable between IFRS and MULT. IFRS*MULT2 is the quadratic form of the interaction variable between IFRS and MULT. PROF is calculated as operating income divided by total assets. SIZE is calculated as the natural log of total assets. MTBR is the market to book value of the firm. ∆TA is calculated as the relative change in assets. DREP is a dummy variable with the value of 1 if a firm repurchases shares and 0 if the firm does not. REPU is calculated as repurchases of shares divided by total assets. DCRIS is a dummy variable that has a value of 1 from 2008-2012 and 0 from 2000-2007. DPRE is the dividend premium. Cons is the intercept of the regression equation. Total Obs is the total amount of observations for which all information was available. Prob indicates the significance of the equation and McF R2 is the McFadden R-squared. Panel A B C D Co Co Co Co IFRS -0.134** -0.060 -0.169* IFRS -0.319*** DMULT -0.339*** -0.430*** MULT -0.012*** MULT2 0.000** IFRS*DMULT 0.202* IFRS*MULT 0.026*** IFRS*MULT2 0.000*** LPROF 14.101*** 14.391*** 14.394*** LPROF 14.274*** LSIZE 0.436*** 0.431*** 0.433*** LSIZE 0.463*** LMTBV -0.261*** -0.265*** -0.265*** LMTBV -0.254*** L∆TA -0.279*** -0.174** -0.178** L∆TA -0.125 DREP 0.786*** 0.720*** 0.715*** DREP 0.682*** LREPU 0.050 -0.188 -0.181 LREPU -0.124 DCRIS -0.861*** -0.803*** -0.801*** DCRIS -0.736*** DPRE -0.273*** -0.296** -0.294** DPRE -0.284** Cons -4.345*** -4.148*** -4.122*** Cons -4.352***
Total Obs 14351 11184 11184 Total Obs 11184
Prob(LR statistic) 0.000*** 0.000*** 0.000*** Prob(LR statistic) 0.000***
32 Appendix 4.
Probit and logit panel regression coefficients for the decision to pay dividends for 2000-2012 in the UK. Robustness test, with different measures for multinationality.
The dependent variable is a dummy variable that takes the value of 1 if a firm pays dividends and 0 if it does not pay dividends. Co indicates the coefficient of the variable. * indicates a significance level at 10%, ** at 5% and *** at 1%. An L before the variable indicates that the variable is lagged by one year. IFRS is a dummy variable that takes the value of 1 if a firm uses IFRS as its accounting standard and 0 if it takes UK GAAP as its accounting standard. MULT is the international sales to total sales ratio. IFRS*MULT is an interaction variable between IFRS and MULT. DMULT0 is a dummy variable that takes the value of 1 if the international sales to total sales ratio is bigger than 0% and 0 if the ratio is 0%. IFRS*DMULT0 is an interaction variable between IFRS and DMULT. PROF is calculated as operating income divided by total assets. SIZE is calculated as the natural log of total assets. MTBR is the market to book value of the firm. ∆TA is calculated as the relative change in assets. DREP is a dummy variable with the value of 1 if a firm repurchases shares and 0 if the firm does not. REPU is calculated as repurchases of shares divided by total assets. DCRIS is a dummy variable that has a value of 1 from 2008-2012 and 0 from 2000-2007. DPRE is the dividend premium. Cons is the intercept of the regression equation. Total Obs is the total amount of observations for which all information was available. Prob indicates the significance of the equation and McF R2 is the McFadden R-squared.
Probit Logit Probit Logit
Co Co Co Co IFRS -0.059 -0.074 IFRS -0.129** -0.188* MULT -0.006*** -0.009*** DMULT0 -0.212*** -0.353*** IFRS*MULT 0.001 0.001 DSY*DMULT0 0.116 0.171 LPROF 7.678*** 14.341*** LPROF 7.706*** 14.430*** LSIZE 0.261*** 0.458*** LSIZE 0.244*** 0.424*** LMTBV -0.136*** -0.256*** LMTBV -0.141*** -0.264*** L∆TA -0.074 -0.134 L∆TA -0.100** -0.175** DREP 0.403*** 0.707*** DREP 0.408*** 0.717*** LREPU -0.211 -0.151 LREPU -0.227 -0.199 DCRIS -0.445*** -0.752*** DCRIS -0.481*** -0.822*** DPRE -0.171*** -0.285** DPRE -0.170*** -0.279** Cons -2.429*** -4.322*** Cons -2.252*** -3.986***
Total Obs 11184 11184 Total Obs 11184 11184
Prob(LR statistic) 0.434 0.000*** Prob(LR statistic) 0.426 0.000***
