Valuation of Dutch and German listed equity
-- A multiples based approach --
Rijksuniversiteit Groningen
Faculty: Management and Organization
Acknowledgements
Nobody said that writing your masters thesis would be easy; in fact it is not easy at all. Finding a topic which could be empirically tested enhanced many ideas, but almost as many did not seemed practical. During the start of my thesis I have switched the subject of my thesis twice. Making a decision to change subjects is never easy; time and effort is lost. However, looking back it was a good choice since the topic at hand is practical and has lead to a proper thesis. Therefore, I am thankful for the people who supported me in those decisions: my professor Mr. Brunia and my girlfriend.
Abstract
Little is known about how the valuation of Dutch shares compares to other European stock exchanges. Although investors suspect that the Dutch shares are undervalued, few empirical evidence is at hand. This study compares the market valuation of the AEX index companies for the Dutch market to the DAX-30 index companies for the German market. A multiples analyses is performed which does not show any evidence of undervaluation of the Dutch shares compared to the German shares. On the contrary, the AEX companies seem to be valued higher by the capital market. Especially the Dutch non-financial companies (production and service companies) are valued higher compared to the German non-financial companies. The financial companies (insurance and banking companies) show a mixed picture. However, there does exist a valuation difference between the Dutch and German listed equity.
Table of contents
Acknowledgements Abstract
Tables and figures
Chapter I: Introduction………..…………..page 1 Chapter II: Theoretical background………..page 5
2.1 Valuation page 5
2.2 Multiples valuation page 6
2.3 Market valuation page 6
2.4 Valuation difference page 7
2.5 Related studies page 8
2.6 Relevant multiples page 14
Chapter III: Research methodology………...page 18
3.1 Data page 18
3.2 Methodology page 20
Chapter IV: Results of the T-tests………..page 23
4.1 The used multiples page 23
4.2 Significance test of the multiples page 25 4.3 Significance test of the performance measures page 27
4.4 Interpreting the results page 28
Chapter V: Rationale behind the valuation difference………..page 30
5.1 Company specific factors page 31
5.2 Market specific factors page 35
5.3 Country specific factors page 36
Tables and figures
Tables:
Table 4.1 Test result non-financials, 1998-2004………..………page 25 Table 4.2 Test result financials, 1998-2004……….page 26 Table 4.3 Performance measures non-financials, 1998-2004.…….……page 27 Table 4.4 Performance measures financials, 1998-2004...………..page 28 Table 5.1 Test results for size, 1998-2004.………..page 31 Table 5.2 Test result profitability, 1998-2004.………page 32 Table 5.3 Test results debt/equity, 1998-2004……….page 33
Figures:
Figure 1.1 Market capitalisation of the AEX and the DAX….………….page 2
Figure 1.2 Conceptual model……….…page 4
Chapter I: Introduction
“Dutch shares are cheap. One can not come to a different conclusion when comparing the price earnings ratios of Dutch listed companies to foreign listed companies. When looking at the market capitalisation over turnover, Dutch shares also appear to be undervalued” (FD 2).
The above quote states that Dutch listed equity is undervalued. The rational for the undervaluation is that the Dutch structure regime (and other factors, as will be discussed in this thesis) makes the Dutch listed equity less attractive to investors and therefore, the equity trades at a lower price (Visser, 2003). This phenomenon is referred to in the market as the “Dutch Discount” (Fogarty, 2003; 16).
The question is whether a difference in market valuation over turnover implies undervaluation. It is possible that there are other factors involved, for example profitability differences. Therefore, the topic of this thesis is not undervaluation, but ‘valuation difference’.
The markets under observation are the Dutch and German listed equity market, i.e. the companies which constitute the AEX index for the Netherlands and the companies which constitute the DAX-30 index (hereafter DAX) for the German equity market. The companies of these two indexes are listed in appendix I.
The reason these two markets are compared with each other is multiple:
1. The Dutch and German listed equity markets are, together with the UK and the French market, the four largest and most important European equity markets (VEB, 2004).
2. The Dutch and German economies are interconnected; Germany is the largest trading partner of the Netherlands. The trade between the two countries had a value of over 100 billion Euro’s in 2004 (www.cbs.nl).
3. Figure 1.1 displays the market capitalisation of the companies which constitute the AEX (Dutch companies) and the DAX (German companies).
-100 200 300 400 500 600 700 800 900 1998 1999 2000 2001 2002 2003 2004 2005 Year E U R B illio n AEX DAX
Figure 1.1: market capitalisation of the AEX and the DAX (www.euronext.com)
period (2002 until 2005) compared to the Dutch companies. In the overall period the AEX loses 6%, where the DAX gained 36%. Chapter 4 discusses whether this leads to different valuation multiples for the German companies.
4. The relevancy of this study lies, besides the above three points, in a broader context, namely the European Union (EU). The EU strives towards integrated European financial market (Ballas and Hevas, 2004). This implies that the European listed equity should be valued equally, unless there still exist boundaries and country differences. This study handles the question whether the European market is integrated or segregated, based on a comparison of Dutch and German listed equity.
For the above reasons, it is interesting to see how well the two markets (Dutch and German) are integrated; therefore my research question will be the following:
Are there valuation differences between Dutch and German listed equity?
Dutch Discount . Country specific factors Market specific factors Company specific factors Valuation Difference Theory . Methodology . Ch ap te r V C h a p te rs I I - IV Related research Hypothesis . Multiples . Method .
Figure 1.2: conceptual model
The chapters 2 until 4 discuss the topic ‘valuation difference’, and analyse whether the Dutch shares suffer such a difference. As mentioned above, a valuation difference does not imply undervaluation (as in the Dutch Discount). The fifth chapter discusses and analyses factors which could explain a valuation difference, but do not imply a Dutch Discount, and thereby undervaluation. The final chapter is the conclusion and recommendations.
Chapter II: Theoretical background
This chapter provides background information regarding valuation and relevant theory. Attention is paid to the two relevant valuation methods for this thesis, namely the multiples valuation and the market valuation. Thereafter, valuation differences and related studies are discussed which lead to the hypothesis. In the final section of this chapter a list of relevant multiples and performance measures is composed, based on the analysed literature.
2.1 Valuation
opportunity cost of debt” (Copeland, 2000;136). The equity value of a company equals the value of its operations and operating assets, less the value of the debt and non-operating liabilities (Copeland, 2000;137). The following two sections discuss the two relevant valuation methods for this thesis, i.e. the multiples valuation and the market valuation.
2.2 Multiples valuation
The commonly used multiples approach assumes a direct proportionality between the price and the value driver. The price represents the market value, i.e. the market capitalisation or the EV of a listed company. The value driver is equal to an accounting measure as: turnover, EBITDA, EBIT or others. The advantage of a multiple approach compared to other valuation approaches are that it is effective (i.e. easy to communicate), relative easy to calculate and it is based on market values. Although multiples are effective, they are often used complementary to a comprehensive valuation method (Lui a.o., 2002; 1) like the discounted cash-flow analyses.
2.3 Market valuation
the equity value is equal to what the investors expect to be the future cash flow1 from operations, including growth options and the use of debt financing. As mentioned above, the equity value (market capitalisation) plus the net debt (interest bearing debt minus excess cash) equals the enterprise value (EV) of a listed firm. The higher the enterprise value, the higher the expected future cash flow of that company is. Investors influence the enterprise value by adjusting the equity value trough buying or selling shares of the company.
2.4 Valuation difference
As mentioned in the introduction, this thesis is about a valuation difference instead of undervaluation. It is possible that one company’s market capitalisation is 7 times its EBIT, where another company’s market capitalisation equals 8 times its EBIT. This does not necessarily mean that the first company is undervalued relative to the second company; it does mean that there is a valuation difference between the two companies. A valuation difference should have a logical explanation, in this case the second company could use a different amount of debt financing relative to the first company. Judging only on multiples is dangerous, especially when one is claiming that Dutch shares are undervalued because their market capitalisation over turnover is low compared to other countries (FD 2). This implies that one cannot judge solely on multiples regarding the undervaluation of a share before looking at other explaining factors. The most one can say is that there exists a valuation difference.
1
2.5 Related studies
Until now, little is known about the integration or segmentation of the European equity markets. It is known that the European Union (EU) is striving for integrated capital markets, and co-ordinate the policy by a set of rules, harmonising financial statements. Differences remain, which can be attributed to cultural factors such as corporate governance and institutional differences that refer to the tax authorities. Despite the differences, the EU objective remains integrated capital markets, and therefore many common fundamentals drive the EU capital markets (Ballas and Hevas, 2004; 4).
This part discusses related research, which describe about the same topic as this thesis. First, American research is discussed, then the focus will be on European research and finally, the Dutch situation will be analysed.
American studies
Another study of cross-country differences for the US and Canadian equity market is performed by King and Segal (2003). The authors compared the two equity markets with the intention to find out whether the Canadian equity trades at a discount compared to the US equity. With a database of over 5000 companies, including 160 Canadian interlisted2 firms, controlled for the following factors:
Company specific factors: include size, profitability, industry, cost of equity and secondary market liquidity;
Market specific factors: industry composition and return of the stock market;
Country specific factor: accounting differences.
King and Segal used a multiples method to come to their conclusions. The authors found that after controlling for the above factors, the Canadian equity market remained undervalued compared to the US equity market. The conclusion was that the US and Canadian market were not fully integrated. The multiples used by King and Segal (2003) are: market to book value, price earnings, EV / EBITDA, ROE and ROA.
Harrison states in his 2004 paper that investors need to have a global view instead of a local view. Valuations of investors are often driven by a local view and due to this ‘home bias’ capital is sub-optimal allocated. In this manner, free lunches are available for investors, but not eaten. The author uses the example of Exxon. The company sees Royal Dutch and Total as its competitors, where investors see Wall-Mart as the competitor and make their investment decision accordingly. With Total trading at half the valuation of
2
Exxon, yet growing twice as fast, the investment decision should be clear. Due to the home bias, investors do not exploit open opportunities. The conclusion of this article is that cross-border valuation anomalies occur due to a home-bias (Harrison, 2004).
European studies
Joos and Lang (1994) investigated whether accounting differences lead to differences in market valuation for the French, German, Dutch and UK listed companies. They found that differences in the accounting systems lead to differences in the market valuation of accounting data and that results in a valuation difference (when looking at multiples). In an attempt to explain these differences, they showed that neither macro-economic factors nor the composition of the sample does explain the cross-country differences. The effort of the EU, regarding integrated reporting standards, did not seem to affect the differences as well (at that moment). The consequence of this is that cross-country comparison of accounting data is that differences in accounting systems can lead to valuation differences. The multiples used in their study are: ROE (book value), earnings / price and book to market value.
valuation of accounting income and the book value of equity are significant for the four largest European indexes3. The multiples used by the authors are MC / earnings and the market to book ratio.
The Dutch situation: investor beliefs
This section pays attention to the Dutch situation. Although there is only little research available, (Dutch) investors have a clear opinion on the valuation of Dutch shares; they presume that the Dutch listed equity is undervalued (van Mullingen, 2004).
Looking at the Dutch equity market and, no hard evidence regarding the undervaluation of Dutch listed equity is available. What one can find are rumours amongst investors that the Dutch listed equity is undervalued and investors refer to this phenomenon as the Dutch Discount (van Mulligen, 2004). The belief amongst Dutch investors is that due to the following points, Dutch shares are undervalued compared to other European stocks:
The structure regime: the structure regime entails that shareholders have less influence on the company’s decisions (compared to abroad companies) since they have no voting rights (van der Poel, 2003) and therefore, the equity is valued lower (FD 1);
Dutch scandals: Dutch scandals as the accounting fraud at Ahold, affect the trust of the investors and makes them less willing to invest in Dutch shares (Annink and Claassen, 2004);
According to research done by the Dutch investment bank Kempen & Co the Dutch equity is undervalued due to high volatility of the market (FD 1). The Dutch market has a high volatility for two reasons:
3
1. The dollar sensitivity of the market: high dollar sensitivity makes the market more volatile (VEB, 2004);
2. The composure of the AEX: financials have a high weight in the AEX. Financials are more volatile since they invest in equity themselves (Conijn, 2003).
Indicators for the Dutch Discount are according to Mr. van der Stee from Van Lanschot Bankiers (FD 2), the price / earnings (PE) ratio and the market-capitalisation over turnover.
Besides the thoughts amongst the national investors, foreign investors seem to share the view of the Dutch Discount. Dutch companies are said to be cheap due to the Dutch Discount, which goes as far as a discount of 20% (Rolvink, 2005A). The Dutch listed companies perform well. However, this performance is not always incorporated in the stock rating (Rolvink, 2005B).
Due to the ‘code Tabaksblat’4, companies can not hold (hostile) investors outside forever since the entry barriers are reduced. These lower barriers and the presumed lower price (Dutch discount) make the Dutch companies more attractive to (foreign) investors. It is expected that more companies will be delisted from the Dutch stock exchange in the upcoming years, mainly by foreign investors (Rolvink, 2005A). Examples of these delistings are Vendex KBB by the US private equity party KKR (Elshout, 2004) and a more recent example is the bid of KKR and other venture capitalists on the shares of the Dutch listed company VNU, which is expected to be delisted in the upcoming months (Wuijster, 2006).
4
Conclusions regarding the related research
The described research shows that the US equity and the Canadian equity market are not perfect substitutes. Even after controlling for company, market and country specific factors, the Canadian equity remains undervalued. Besides, investors do not fully exploit the opportunities abroad due to the home bias. Moreover, valuation differences (in the EU) can partially be explained by the industry composition of a market, although the efforts of the EU for integrated financial statements did not show any effect in the mid-nineties. As for the Netherlands, investors assume that the Dutch shares are undervalued, but these assumptions are based on feelings, rather than on empirical research.
Hypothesis
This study analyses whether the Dutch equity is undervalued, or as I call it whether there is a valuation difference between the Dutch and German listed equity market. The hypothesis below will be tested in this thesis.
(Ho) Dutch and German listed equity is valued equally by the capital market. (Ha) Dutch and German listed equity is not valued equally by the capital market.
Based on the above research, a list of multiples and performance measures is formed in the next section. These multiples and performance measures are applied in this thesis.
2.6 Relevant multiples
In the research as described in the previous section, different authors used different multiples when looking at valuation differences. Despite that the multiples are based on accounting numbers and therefore sensitive to fraud, they do provide an element to compare market valuations of different companies (or countries).
Most research paid attention to the market valuation of earnings. The following multiples will be used to analyse the market valuation of earnings:
1. Enterprise Value / EBITDA; 2. Enterprise Value / EBIT.
The enterprise value is the total company value, i.e. debt and equity. The two accounting numbers (EBITDA and EBIT) do not take financing differences (I = interest) and tax differences (T = taxes) into account. The result are not influenced by differences in the tax regime and differences in the financing mix. A distinction is made between EBIT and EBITDA. The EBIT multiple states how the operating result (= EBIT) is valued by the capital market. Since the EBIT multiple is affected by differences amortisation and/or depreciation, the EBITDA multiple is used as well.
Another often used multiple in the above research is market to book value (M-t-B). This thesis uses the market to book value as well:
3. Market Value / Book Value.
lower (M-t-B < 1) compared to the book value. A high M-t-B value (>1) means that the company has created (substantial) wealth for its shareholders, a low M-t-B value (<1) means that shareholder value has been destroyed. The advantage of the M-t-B multiple is that it can be used when a company does not make profit in a given period (Vanbuggenhout and Bruggeman, 2002).
The price-earnings (PE) ratio is also an often used multiple in the above researches. This multiple displays how many times an investor is willing to pay for the earnings:
4. Price / Earnings.
The price per share (or market capitalisation) is divided by the earnings per share (or earnings). The multiple is irrelevant when the company under study does not make profit (Vanbuggenhout and Bruggeman, 2002).
According to Mr. van der Stee (FD2), turnover multiples are relevant. Therefore this research includes a turnover multiple as well:
5. Enterprise Value / Turnover.
The above 5 multiples are extracted from the described literature. These 5 multiples are used to prove whether there exists a valuation difference between the Dutch and German listed equity, or not. Besides the above multiples, the fourth chapter also pays attention to performance measures. These performance measures could indicate why a possible valuation difference exists. Performance measures are lagging indicators and do not contain ‘forward looking’ information. However, they might point out reasons behind a possible valuation difference. The applied performance measures are the following:
1. Return on Assets (ROA)
The ROA tells how much profit is made with 1 Euro of assets. It indicates how efficient a company uses its assets. However, production companies (high capital intensity) are expected to have a lower ROA than service companies (low capital intensity). Therefore, comparing these two types of companies based on the ROA is incorrect. The ROA is calculates with the following formula: EBIT / Net Assets (total assets net of current liabilities).
2. Return on equity (ROE)
3. Operating profit margin (OPM)
The operating profit margin calculates how much a company earns relative to its sales. Since EBIT is used, differences in the financing mix and the tax regime do not affect the result. The operating profit margin is calculated by the following formula: EBIT / Turnover.
4. Return on invested capital (ROIC)
ROIC provides a good picture of the historical performance of a company. It measures the return for the investors of a company and is calculated with the following formula: Net income (NOPAT) / Invested Capital (Total assets – excess cash and non interest bearing liabilities).
Chapter III: Research methodology
This chapter discusses the methodology for the empirical research. The following elements will be discussed:
Data;
Methodology.
3.1 Data
This thesis is focused on the Dutch and German markets for listed equity. The ‘large caps’ are analysed in this thesis; the companies that constitute the AEX-index for the Dutch equity and the companies that constitute the DAX-30-index for the German equity. The AEX is formed of 23 companies; the DAX is formed 30 companies. The composure of the indexes is shown in appendix I (the appendix also displays the index composure, i.e. the relative weights of the companies in the indexes).
The data of the companies that constitute the AEX index and the DAX index are analysed to uncover whether there exist valuation differences or not. The period under study is from 1998 until 2004, since accounting data over 2005 are not available at the moment of this study. The time frame for the accounting data is determined by the availability of data in the database AMADEUS5, which is the source of the accounting data for the non-financials (production and service companies). BANKSCOPE6 provided the data of the financials (banks and insurance companies).
5
AMADEUS is a database from Bureau van Dijk which provides the annual accounts and other financial relevant information of almost every European company. Access to this database was available trough the network of the RijksUniversiteit Groningen.
6
The AEX and the DAX are composed of companies in different industries. A distinction is made between financial and non-financial companies. Financials have a different business model and thereby different annual accounts. Therefore, financials are not comparable with non-financials. The pie charts in figure 3.1 display the distribution in the indexes between financials and non-financials, based on their weights in the indexes per December 31, 2005. AEX Financial Non-financials DAX Financial Non-financials
Figure 3.1: distribution financials vs. non-financials
The above figure indicates that although the AEX entails four financials and the DAX five, the financials have a higher weight (40%) in the index compared to the DAX (24%). As mentioned in section 2.5 (related studies) the investors assume that for this reason the AEX companies have a lower market valuation compared to other European stock exchanges.
The entire population (AEX and DAX) is split up in two separate populations, i.e. a population non-financials and a population financials. The population of non-financials is formed of 19 companies from the AEX and 25 from the DAX. The population of financial companies is formed of 4 companies from the AEX and 5 companies from the DAX. For all companies, end year data is used.
multiples imply that investors are willing to pay for a negative return. Since investors also account for earnings which they expect in the future, the share price has a multiple year horizon. The relative high and negative ratios provide a biased picture. Figure 3.2 shows the acceptation ranges for the different multiples. The ranges are based on a paper by Lane and Reingold (2003), who looked at the trends of valuation multiples. For the period from 1995 until 2002 they analysed different US valuation multiples.
0 5 10 15 20 25 EV / Turnover EV / EBITDA EV / EBIT M-t-B* P / E M u lt ip le Range
*M-t-B can also be negative
Figure 3.2: acceptation ranges for multiples
At least 50% of the observations should fall within the acceptation range (per multiple, combined of the Dutch and German market) in order use the multiple for this analysis.
3.2 Methodology
The T-test assumes a normal distribution, although financial data or often not normal distributed (Brooks, 2002; 3). Therefore an adjustment is made which is called the ‘Welch-Aspin statistic’ (Reed III, 2005). With this formula, the requirements for the normal T-test are no longer relevant. The adjusted formula for the t-statistic is displayed in figure 3.3.
Figure 3.3: t-statistic formula (Reed III, 2005)
In the above formula, N1 and N2 are the sample sizes, and are the sample means,
and and are the sample variances7. The requirement of normality is no longer relevant since there are more than 30 degrees of freedom. With more than 30 degrees of freedom, the data follows a more normal distribution (Reed III, 2005).
The reciprocal formula (enterprise value / EBIT will be EBIT / enterprise value) is used for some multiples to reduce the effect of outliers (Lui a.o. 2002). This will only be applied for the EBIT, EBITDA and PE multiples, since he other multiples close to one (using the reciprocal formula does not lead to a smaller effect of outliers).
The t-statistic (calculated with the formula in figure 3.2) is compared to the critical value (displayed in appendix II). The significance level for the empirical tests is 5%. The degrees of freedom are calculated as: N1+N2-2. Since it is a two-tailed test is used (Ha: Dutch means are unequal to German means), a 5% significance (alpha) level implies an
7
acceptation range of 0.95 as expressed in figure 3.4 and a change of 5% for a type I error8.
Figure 3.4: acceptation and rejection region for a two-sampled test
R eje ction are a: Alp ha/2 = 0.025
R eje ction are a: Alp ha/2 = 0.025 Ac ce ptation are a:
1 - alp ha = 0.95
The result (whether the t-statistic exceeds the critical value) provides the conclusion regarding the significance of the differences in means, and whether to except or reject the null hypothesis.
8
Chapter IV: Results of the T-tests
The aim of this study, as already mentioned, is to test whether the AEX and the DAX companies are valued differently. This chapter discusses the results of the two-sample t-tests, and shows whether the means are significantly different. Thereafter, attention is paid to the interpretation of the results. First, it is discussed whether the multiples from the second chapter can be used in the analyses; are their enough ‘accepted’ scores given acceptation ranges?
4.1 The used multiples
In the second chapter the multiples that are used for the empirical research have been introduced. Based on the available data, some multiples seem less useful compared to others.
As mentioned in the third chapter, multiples which are out of the acceptation ranges as expressed in figure 3.2. are excluded. The PE multiple is excluded for the population non-financials since less than 50% of the PE multiples fall within the ‘acceptation range’. Figure 4.1 displays the percentage of accepted scores per multiple for the non-financials.
0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% EV / Turnover EV / EBITDA EV / EBIT M-t-B P / E Mu lt ip le Percentage DAX AEX
The other multiples provide enough ‘accepted’ scores (>50%), and therefore they are used in this thesis.
For the population financials not all multiples from chapter 2 can be used. Due to the difference business model of financials (they enhance turnover from the liabilities side of their balance sheet), it is impossible to calculate a representative EBITDA or EBIT number. Instead profit before tax (PBT) and net profit (NP) are used. For this reason, the acceptation ranges change as well, as expressed in figure 4.2. Again, these ranges are based on the paper of Lane and Reingold (2003).
0 5 10 15 20 25 MC / Turnover MC / PBT MC / NP M-t-B* P / E M u lt ip le Range
*M-t-B can also be negative
The use of debt is difficult for financials as well; financials borrow and lent money at the same time, calculating a representative amount of debt usage (for the EV) is therefore impossible. Therefore, the EV is replaced by the market capitalisation (MC).
Figure 4.3 displays the percentage of accepted scores for the population financials.
0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% MC / Turnover MC / PBT MC / NP P / E M-t-B M u lt iple Percentage DAX AEX
For the DAX companies, there are insufficient scores (<50%) for the P/E, MC/PBT and the MC/NP multiple. However, combined with the AEX companies there are over 50% accepted scores. Therefore, all above multiples will be used.
4.2 Significance test of the multiples
This section discusses the results of the t-tests. First the results for the non-financial companies are discussed, thereafter the result for the financial companies are discussed.
Non-financial companies
Table 4.1 displays the results of the t-tests for the non-financial companies. The extended version of this table is displayed in appendix 3. On all 4 used multiples the average for the AEX is higher compared to the DAX and the differences are significant at a 5% significance level. Difference (AEX - DAX) EV / Turnover 1,42 1,04 0,38* EV / EBITDA*** 9,33 6,89 2,44* EV / EBIT*** 12,73 9,74 2,99* M-t-B 2,00 1,55 0,45*
* Differences are significant with a 5% alpha level *** Calculated with the reciprocal formula
DAX AEX
Table 4.1: test results non-financials, 1998-2004
Financial companies
In table 4.2 the results of the significance tests for the population financials are displayed. The extended version of this table is displayed in appendix 4.
Difference (AEX - DAX) MC / Turnover 1,97 1,325 0,65* MC / PBT*** 8,65 12,36 (3,71)** MC / NP*** 11,14 12,74 (1,60)** P / E*** 11,14 12,74 (1,60)** M-t-B 1,781 1,53 0,25**
* Differences are significant with a 5% alpha level ** Not significant with a 5% alpha level
*** Calculated with the reciprocal formula
AEX DAX
Table 4.2: test results financials, 1998-2004
Table 4.2 provides a mixed picture. On average, the German financial companies are valued higher on 3 of the 5 multiples, however non are significant. For the other 2 multiples the Dutch financial companies are valued higher (only 1 significant). I can reject the Ho hypothesis for the population financials, since the results indicate that the Dutch and German financial companies are not valued equally. It cannot be stated whether the Dutch or German financial companies are valued higher due to the mixed results.
4.3 Significance test of the performance measures
In the second chapter, 4 different performance measures were described. This section analyses the result of these performance measures. All results are calculated with the formula as explained in the third chapter.
Non-financial companies
Table 4.3 contains the results of the performance measures for the non-financial companies. The extended version of this table is displayed in appendix 5. As opposed to the results of the multiples, the German financials perform better on the performance measures compared to the Dutch companies. Despite that the German firms are more profitable, have a higher ROA, ROE and ROIC, the capital market values the German companies lower than the Dutch companies. However, the differences are only significant for 2 of the 4 performance measures. German firms ‘perform better’ on the performance measures but the market values them lower compared to the Dutch firms. The ‘good’ performance of German firms is not (fully) incorporated in the market valuation relative to the Dutch firms.
Difference (AEX - DAX) ROA 0,074 0,121 (0,047)* ROE 0,040 0,180 (0,14)** OPM 0,055 0,089 (0,035)* ROIC 0,054 0,062 (0,008)**
* Differences are significant with a 5% alpha level ** Not significant with a 5% alpha level
AEX DAX
Financial companies
Table 4.4 contains the results of the performance measures for the financial companies. The extended version of this table is displayed in appendix 6.
Difference (AEX - DAX)
ROA 0,009 0,002 0,007*
ROE 0,135 0,064 0,071*
OPM 0,172 0,043 0,129*
* Differences are significant with a 5% alpha level
AEX DAX
Table 4.4: performance measures financials, 1998-2004
As mentioned, financials have a different business model. Therefore it is not possible to calculate a representative amount of debt. The ROIC is for that reason not used for the financial companies. The result of the ROA, ROE and OPM indicate that the Dutch financials perform significantly better compared to the German financials as opposed to the multiples, which showed a mixed picture. As with the non-financial companies, the ‘good’ performance of the Dutch financials is not (fully) incorporated in the market valuations.
4.4 Interpreting the results
When interpreting the results from the above analysis, we should keep in mind that the results have some limitations, which are:
The study only involved Dutch and German companies and can therefore not be extrapolated to the entire European Union. Therefore, I cannot judge whether the thoughts amongst investors as expressed in paragraph 2.5 about the Dutch Discount are true or not true;
The study only involved large caps. When medium and small caps were included different results could be obtained;
The stock markets fluctuate over time. The timeframe under study is from 1998 until 2004, a different timeframe may enhance different results.
Chapter V: Rationale behind the valuation difference
The previous chapter indicated that there exists a valuation difference between the Dutch and German companies; however there might be several logical explanations. This chapter looks at some possible explanations of the valuation difference.
Figure 5.1 is based on a working paper of King and Segal (2003) who compared the Canadian listed equity with the U.S. listed equity. The structure regime (a country specific factor) is added, since it is a specific Dutch country factor and it was not included in the research of King and Segal (2003). They found that despite controlling for the factors below (except for the structure regime) that the Canadian equity market is valued significantly lower compared to the U.S. market.
Valuation Difference
Company specific factors Market specific factors Country specific factors - Size
- Industry - Profitability - Cost of equity - Secondary market liquidity
- Return of the stock market - Industry composition
- Structure regime - Accounting differences
valuation difference to see whether the factors from figure 5.1 can explain the lower market valuation of the German shares.
5.1 Company specific factors
Figure 5.1 displays the company specific factors which could be accountable for (part) of the valuation difference as founded in the previous chapter. Below, the company specific factors are discussed.
Size
The bigger, the better. The larger a company is, the more it is able to profit from operations of scale and scope and is thereby more profitable (Besanko, 2000; 75). Based on the above statement, larger companies should be valued higher. As a proxy for size, turnover and total assets are used. Since the previous chapter showed that Dutch companies are valued higher by the market than German firms, it is expected that Dutch firms are larger as well. Table 5.1 expresses the results of a two-sample t-test, calculated in the same manner as explained in the third chapter. The extended version of this table is displayed in appendix 7.
Difference (AEX - DAX)
Assets*** 5.414 11.453 (6,039)*
Turnover*** 5.415 8.362 (2,947)*
* Differences are significant with a 5% alpha level *** Calculated with the reciprocal formula
EUR Billion AEX DAX
The results indicate that on average, the German companies are significantly larger compared to Dutch companies. Therefore, a difference in size does not seem to explain the valuation difference.
Industry
A production company can have a different valuation multiple than a services company. King and Segal (2003) adjusted their population by comparing companies within the same industry. Since the population under study is to small for such an adjustment, this distinction can not be made. However, a distinction is made between the financials and non financials; this is discussed in section 5.2: market specific factors.
Profitability
As mentioned, investors value cash. A more profitable firm should therefore be valued higher on a turnover multiple. The previous chapter already showed that the German non-financial companies are more profitable than the Dutch non-non-financial companies; the Dutch financials are more profitable than their German competitors. Table 5.2 displays the results for the AEX and DAX companies (financials and non-financials), calculated as EBIT / Turnover9 with the formula form the third chapter. The extended version of this table is displayed in appendix 8.
Difference (AEX-DAX)
Profitability 0,050 0,063 (0.013)**
** Not significant with a 5% alpha level
AEX DAX
Table 5.2: test results profitability, 1998-2004
9
Table 5.2 shows that the means of the combined result of the financials and the non-financials are not significantly different for the Dutch and German companies when looking at the profitability. Despite that the results are not significant, the German companies have a higher mean profitability (0,063). Differences in profitability do not seem to explain (part of) the valuation difference.
Cost of equity
The cost of equity for a firm is the expected return required by investors to induce them to hold the equity. The firm’s equity beta (measure of risk) increases linearly as the amount of debt financing increases, and therefore the cost of equity financing increases linearly as a function of the leverage ratio (Grinblatt & Titman, 2002; 383). In table 5.3 the debt ratio is calculated with the following formula: Debt / (Debt + Equity). In this formula the debt is the book value and the equity is the market value. The results are calculated with the formula as explained in the third chapter. The extended version of this table is displayed in appendix 9.
Difference (AEX - DAX)
Debt / Equity**** 0,364 0,405 (0,042)**
** Not significant with a 5% alpha level **** Only non-financial companies
AEX DAX
Table 5.3: test results debt/equity, 1998-2004
Since 3 of the 5 used multiples for the non-financial companies are entity value multiples, WACC is a more important indicator than the cost of equity. A company with relative more debt has a lower WACC since part of the cost of financing is borne by the government due to the fact that interest is tax deductible (Grinblatt & Titman, 2002; 481)10. Moreover, the higher debt level reduces the agency cost and the free cash flow11 problem as described by Jensen (1986). The theory implies that with a higher debt level the management is forced (less slack) to make better investment decisions (positive net present value projects) since a relative large part of the operating cash flow is paid to the debt holders.
The higher cost of equity for the German firms can thus be an explaining factor for the equity value multiples (P/E and M-t-B). However, the German companies are likely to make better investment decisions and have a lower WACC compares to the Dutch firms, therefore differences in the WACC do not explain the differences in the entity value multiples (since Dutch firms are on average valued higher by the capital market).
Secondary market liquidity
According to a study by the ECB (2004), the secondary market liquidity has started to improve in the European Union. The trend is the largest for the United Kingdom and the Netherlands, Germany is an exception regarding this trend (ECB, 2004). This means that Germany is falling behind on the Netherlands regarding the secondary market liquidity.
10
This does not imply that using as many debt as possible is the best, since a higher leverage ratio leads to a higher default rate and thereby cost of distress.
11
Therefore, the lower secondary market liquidity can be a reason for the lower market valuation of German shares compared to Dutch shares.
5.2 Market specific factors
Factors specific to a (stock) market can influence firm valuation. According to figure 5.1, the market specific factors are: return of the stock market and industry composition.
Return of the market
Figure 1.1 already looked at the development of the market capitalisation of the Dutch and German companies. Although the German stock exchange showed a fiercer negative trend, the DAX outperformed the AEX index in the overall period. Since the performance of a market is positively related to the firm valuation (since the equity value increases) the historical market performance can not be an explaining factor because the AEX companies are valued higher by the capital markets compared to the DAX companies. However, investors are forward looking. They accounting for earnings which they expect in the future. A possible explanation therefore could be that investors expect that the Dutch market will perform better compared to the German markets. Therefore, differences in the expected market return could be an explaining factor, although this cannot be quantified.
Industry composition
Financials are more volatile since they invest in equity themselves (FD1). Since financials have a relative high weight in the AEX (40% for the AEX vs. 24% for the DAX, see figure 3.1), the valuation effects of financial companies can be stronger for the Netherlands compared to other countries. As the results of chapter 4 indicated, the Dutch financials are valued lower when looking at the PE ratio, PBT and NP multiples, however the Dutch financials are have a higher market to book ratio and are valued higher on the turnover multiple. Therefore, the Dutch financials are not undervalued relative to their German competitors and the argument that the relative large weight of financials in the AEX is an explaining factor for the Dutch discount and the founded valuation difference in chapter 4 is not true.
5.3 Country Specific factors
According to King and Segal (2003) accounting differences are the country specific factor for valuation since “valuation depends critically on the inputs derived from a company’s financial statement” (King and Segal, 2003;6). Another country specific factor for the Netherlands is the structure regime. Since it is a specific Dutch factor, King and Segal did not take this factor into account.
Structure regime
2005). Since the German shares are founded to be valued lower by the capital market, the structure regime (although its influence is fading) is not an explaining factor for the founded valuation difference.
Accounting differences
In section 2.5 (related studies) a paper of Joos and Lang (1994) was discussed. The conclusion was that there existed accounting difference within the European Union. But the European Union is striving towards integrated capital markets (Ballas and Hevas, 2004). Arce and Mora (2002) performed a comparable study as Joos and Lang did in 1994; the conclusion is that there are valuation differences, but they can be attributed to both differences of accounting practices and differences in institutional factors. Accounting differences can thus influence the market valuation of companies, although the efforts of the EU for equal reporting standards does seem to have an effect towards more equal accounting systems.
Besides the above possible explanation, Copeland and others (2000) argue that there are two key value drivers, which are: growth and return on invested capital (relative to the cost of capital).
since they use relative more debt), differences in ROIC do not explain the higher market valuation of the Dutch companies. Differences in the expected growth rate could be an explaining factor. However, expectations are hard to quantify, as with the expected market return (section 5.2).
Chapter VI: Conclusions and recommendations
In the first chapter I have stated my research question which was: Are there valuation differences between Dutch and German listed equity?
Yes there are. Based on the empirical research, the companies that constitute the AEX and the DAX index show signs of a valuation difference. As opposed to the believes amongst Dutch investors, the Dutch companies are valued higher compared to the German companies. This result is most visible for the non-financial companies; financial companies show a mixed picture.
The following conclusions are derived from the empirical research:
The Dutch non-financial companies are not undervalued compared to the German non-financial companies. In fact, the Dutch companies are on average valued (significantly) higher by the capital market (chapter 4).
The market valuation of financial companies shows a mixed picture; the Dutch financials are valued higher compared to the German financials on the turnover multiple and the M-t-B value. The German firms are valued higher on PBT and NP multiple, and have a higher PE multiple (chapter 4).
As figure 1.1 indicated, the DAX index gained more compared to the AEX index in the period under study. As the performance measures of the non-financials indicated (table 4.3), the German firms perform better compared to the Dutch firms. However, the good performance is not fully incorporated in the market valuation of the German companies (as table 4.1 indicated), this implies a valuation difference.
Explaining factors for the valuation difference could lay in the cost of equity, secondary market liquidity, expected market return, accounting differences and the expected growth (chapter 5).
Given the research boundaries, the above conclusion is reached. This however, does not imply (as mentioned in section 4.4) that with another timeframe, multiples and indexes the same conclusion will be reached. The results cannot be extrapolated to the entire European Union, or the beliefs about the Dutch Discount whatsoever. It only implies the Dutch and German listed equity (based on the DAX and the AEX) do not show signs of the Dutch Discount, and the Dutch and German markets are not integrated, as the EU is striving to.
Information sources
Books and articles:
Arce, M., & A. Mora., Emperical evidence of the effect of European accounting differences on the stock market valuation o earnings and book value, The European Accounting Review, 11 (3), page 573 – 599.
Annink, H.M. & L. Claassen, Corporate Governance in de financiële sector, Jaarboek Compliance 2004, 2004.
Ballas, A.A., and D.L. Hevas, European differences in the valuation of earnings and book value: regulation effects or industry effects, 2004.
Besanko, B., D. Dranove, and M. Shanly, Economics of Strategy, John Wiley & Sons, Inc., USA, 2000.
Brooks, C., Introductory econometrics for finance, Cambridge University Press, 2002.
Copeland, T., T. Koller, and J. Murrin, Valuation, measuring and managing the value of companies, McKinsey & Company, Inc., third edition, 2000.
ECB (Europese Centrale Bank), The Euro Bond Market Study, Frankfurt am Main, December 2004.
Fogarty, K., 4th Meeting of the Eurasian Corporate Governance Roundtable, The Global Corporate Governance Forum, 2003.
Harrison, P., Value, country by country, Canadian Investment Review, summer2004, Vol, 17, issue 2. Database: Business Source Premier.
Jensen, M.C., Agency cost of the free cash flow, corporate finances and takeovers, American Economic Review, May 1986, Vol 76, No 2, page 323-329
Joos, P., and M. Lang, The Effects of accounting diversity: evidence from the European Union, Journal of Accounting Research, Volume 32, 1994, USA.
King, M.R., and D. Segal, Valuation of Canadian- vs. U.S.-Listed Equity: Is There a Discount? Bank of Canada, working paper 2003-6.
Lane, R., and J. Reingold, What’s your company worth now? INC, July 2003, page 71 - 104
Lui, J., D. Nissim, and J. Thomas, Equity valuation using multiples, Journal of Accounting research, 2002.
Reed III, J.F., Contributions to two-sample statistics, Journal of Applied Statistics, Vol. 32, No. 1, page 37-44, January 2005.
Sachse, H., and S. Leijten, The Netherlands: corporate governance, Business Source Premier, 2003.
Vanbuggenhout, G., and P. Bruggeman, Een diversiteit aan waarderingsmethoden: het bos door de bomen, Financiering en accounting - ideeën en inzichten, 2002 page 257-261
VEB12, VEB beursjaaroverzicht 2004, VEB, 31/12/2004.
Newspapers:
Booth, L., and X. Zhao, Competitiveness Part V: Corporate profits go North, Financial post, Monday, June 10 2002, page 11.
Conijn, F., Conjunctuur bepaalt koers van de Nederlandse aandelen, Het Financieele Dagblad, 14/10/2003.
Elshout, P., Vechtlust, Het Financieele Dagblad, 16/07/04.
Klok, P., Een tien met een griffel voor goed gedrag, De Volkskrant 08/03/2005.
Mulligen, F., van,. AEX maatsaf voor beleid ABN Amro Netherlands Fund, Het Financieele Dagblad, 18/9/2004.
Poel, J., van der, Structuurregime is geen ‘best practice’, Het Financieele Dagblad, 19/07/03.
Rolvink, F., De jager en zijn Hollandse prooi, Het Financieele Dagblad, 16/12/2005A, page 13.
Rolvink, F., Nederlandse beursfonds is een gewilde overnameprooi, Het Financieele Dagblad, 16/12/2005B, page 1.
Visser, C., Geef Tabaksblat permanent vervolg in een forum, Het Financieele Dagblad, 14/07/2003.
Wuijster, R., Bubble-risico met private equity? Het Financieele Dagblad, 25/02/2006.
Zevenbergen, B., Beleggers eisen meer greep op beursfonds, Het Financieele Dagblad, 16/11/2005.
FD 1, Nederlandse bedrijfscultuur overschat, Het Financieele Dagblad 14/10/2003.
FD 2, Dutch discount, Het Financieele Dagblad 14/06/2005.
12
Internet and databases:
AMADEUS, Bureau van Dijk.
BANKSCOPE, Bureau van Dijk.
www.marktmeter.com
www.deutsche-boerse.com
www.euronext.com
www.fd.nl
Appendices
Appendix 1: Overview of the index weights of the AEX and the
DAX-30 page 46
Appendix 2: Critical values of t-distributions for different probability
levels and degrees of freedom page 47
Appendix 3: Extended version of table 4.1 page 48
Appendix 4: Extended version of table 4.2 page 49 Appendix 5: Extended version of table 4.3 page 50 Appendix 6: Extended version of table 4.4 page 51 Appendix 7: Extended version of table 5.1 page 52
Appendix 8: Extended version of table 5.2 page 52
Appendix I
Overview of the index weights of the AEX and the DAX-30.
DAX 1/1/2006 E.ON 10.59% SIEMENS 10.17% ALLIANZ* 8.74% DEUTSHCE BANK* 7.49% DAIMLERCHRYSLER 7.19% DT. TELEKOM 6.09% BASF 5.69% SAP 5.56% RWE 5.12% BAYER 4.38% MUENCH.RUEKVERS* 3.93% COMMERZBANK* 2.67% DEUTSCHE POST 2.60%
BAY. MOTOREN WERKE 2.07%
CONTINENTAL 1.91% SCHERING 1.68% DEUTSCHE BOERSE 1.63% ADIDAS-SALOMON 1.41% VOLKSWAGEN 1.29% THYSSENKRUPP 1.25% MAN 1.11% HVB GROUP* 1.08% METRO 0.98% LUFTHANSA 0.98% LINDE 0.92% HENKEL 0.88% INFINEON TECH. 0.84% TUI 0.68% ALTANA 0.54% FRESEN.MED.CARE 0.51% www.deutsche-boerse.com AEX 1/1/2006 ING* 15.91%
ROYAL DUTCH SHELL 14.81%
ABN AMRO* 9.39% PHILIPS 9.02% FORTIS* 8.76% UNILEVER 8.67% AEGON* 5.63% KPN 5.10% TNT 2.96% AKZO NOBEL 2.91% AHOLD 2.60% REED ELSEVIER 2.17% ASML 2.12% VNU 1.76% DSM 1.71% HEINEKEN 1.59% NUMICO 1.47% WOLTERS KLUWER 1.32% SBM OFFSHORE 0.60% VEDIOR 0.54% BUHRMANN 0.43% HAGEMEYER 0.31% GETRONICS 0.21% www.marktmeter.com
* Financials companies. These companies have different annual accounts and cannot be compared to
Appendix 2
Critical values of t-distributions for different probability levels and degrees of freedom.
Appendix 3:
Extended version of table 4.1.
Difference (AEX - DAX) EV / Turnover 1,42 1,04 0,38* N 118 155 VarP 0,72 0,61 T-statistic Critical value EV / EBITDA*** 9,33 6,89 2,44* N 91 130 VarP 0,001 0,006 T-statistic Critical value EV / EBIT*** 12,73 9,74 2,99* N 84 123 VarP 0,000 0,004 T-statistic Critical value M-t-B 2,00 1,55 0,45* N 75 133 VarP 0,86 0,40 T-statistic Critical value
* Differences are significant with a 5% alpha level *** Calculated with the reciprocal formula
Appendix 4:
Extended version of table 4.2.
Difference (AEX - DAX) MC / Turnover 1,97 1,325 0,65* N 24 32 VarP 0,622 0,401 T-statistic Critical value MC / PBT*** 8,65 12,36 (3,71)** N 26 17 VarP 0,001 0,001 T-statistic Critical value MC / NP*** 11,14 12,74 (1,60)** N 26 13 VarP 0,001 0,002 T-statistic Critical value P / E*** 11,14 12,74 (1,60)** N 26 13 VarP 0,001 0,002 T-statistic Critical value M-t-B 1,781 1,53 0,25** N 25 33 VarP 0,50 0,625 T-statistic Critical value
* Differences are significant with a 5% alpha level ** Not significant with a 5% alpha level
*** Calculated with the reciprocal formula
Appendix 5:
Extended version of table 4.3
Difference (AEX - DAX) ROA 0,074 0,121 (0,047)* N 133 166 VarP 0,016 0,036 T-statistic Critical value ROE 0,040 0,180 (0,14)** N 133 168 VarP 0,739 0,098 T-statistic Critical value OPM 0,055 0,089 (0,035)* N 133 171 VarP 0,020 0,010 T-statistic Critical value ROIC 0,054 0,062 (0,008)** N 133 168 VarP 0,739 0,098 T-statistic Critical value
* Differences are significant with a 5% alpha level ** Not significant with a 5% alpha level
Appendix 6
Extended version of table 4.4
Difference (AEX - DAX) ROA 0,009 0,002 0,007* N 28 35 VarP 0,000 0,000 T-statistic Critical value ROE 0,135 0,064 0,071* N 28 35 VarP 0,002 0,007 T-statistic Critical value OPM 0,172 0,043 0,129* N 28 35 VarP 0,007 0,02 T-statistic Critical value
* Differences are significant with a 5% alpha level
Appendix 7
Extended version of table 5.1
Difference (AEX - DAX) Assets*** 5.414 11.453 (6,039)* N 161 206 VarP 0,0000 0,0000 T-statistic 4,49 Critical value 1,96 Turnover*** 5.415 8.362 (2,947)* N 161 206 VarP 0,0000 0,0000 T-statistic 2,80 Critical value 1,96
* Differences are significant with a 5% alpha level *** Calculated with the reciprocal formula
EUR Billion AEX DAX
Appendix 8
Extended version of table 5.2
Difference (AEX-DAX) Profitability 0,050 0,063 (0.013)** N 161 200 VarP 0,017 0,017 T-statistic - 0,959 Critical value 1,960
** Not significant with a 5% alpha level
Appendix 9
Extended version of table 5.3
Difference (AEX - DAX) Debt / Equity**** 0,364 0,405 (0,042)** N 133 35 VarP 0,028 0,061 T-statistic Critical value
** Not significant with a 5% alpha level **** Only non-financial companies
-1,74 1,96
