Branding in turbulent times: the value of ‘strong’ brands for shareholders

(1)

Branding in turbulent times: the value of

‘strong’ brands for shareholders

Author: Inge Fakkert

(2)

1

Branding in Turbulent Times: the value of ‘strong’

brands for shareholders

Author: Inge Fakkert

University of Groningen Faculty of Economics and Business

Master Thesis

Marketing Management and Research

Date: August 2012

Address: Groevelaan 9, 9351DW-Leek Phone Number: 06-46081142 Email: ingefakkert@hotmail.com

Student Number: 1747797

(3)

2

| Abstract

(4)

5| Conclusion and recommendations 28

References 31

(5)

4

1| Introduction

Many recent analyses identify the link between branding and the financial performance of a firm. The analysis of this link is closely related to the increased corporate attention on branding, and the importance of branding within firms is also seen by the corporate leaders of an increasing number of firms. However, for the marketing executives it is still difficult to show the financial accountability of the marketing related activities, which includes branding (Madden, Fehle and Fournier, 2006). This increased interest in intangible assets such as branding is supported by the fact that firms are looking for competitive advantages that are unique and not easy to imitate.

When focusing on the intangible asset branding, there is a problem to be solved; does branding have a positive effect on shareholder value at any given circumstance/situation?

Regarding this problem some research has been done. There is empirical evidence that there is a link between branding and shareholder value. It was investigated that stronger brands deliver greater returns to the shareholders (Shrivastava, Shervani, Fahey, 1998) and also with less risk (Madden et al, 2006). The research of Madden et al. (2006), indicates the importance of building brand equity for firms. It is even mentioned that the research is a justification to provide long term investments for building brand equity.

There are however some aspects that needs to be investigated further with regard to the problem. The research of Madden et al (2006) has limited support, it is indicated within the research that it should be investigated whether the results hold across different types of firms, brands, industries and market conditions, which indicates that the results are not generalizable.

(6)

5

drops more than 50% each 10 years is 90% (Zhou and Zhu, 2010). The fact that a financial crisis is a problem in the market that returns on a regular basis, underpins the importance of investigating that element when a research is performed on the effect of brand equity on shareholder value. That these crisis periods happen on a regular base, has an impact on the long-term investments of a company. When there is a drop in the Dow Jones index at least every ten years it is important to investigate whether it is even worthwhile to invest in building a ‘strong’ brand on the long term. When taking notice of another crisis; the Dot-Com bubble, which took place during the late 1990’s, the crisis mainly affected the High-Tech companies such as Microsoft, Intel and IBM (DeLong and Magin, 2006). Therefore, when taking into account the crisis periods when measuring the effect of brand equity on shareholder value it should be noted that specific industries could have an effect on these outcomes as well.

It becomes clear that branding can have a positive effect on the performance of a firm and on shareholder value. There are however some variables that could change this effect, how do strong brands cope with these variables? Can strong brands cope better with crisis periods? Is the effect of these crisis periods on the performance for stronger brands less than for weaker brands? Do stronger brands recover faster than weaker brands? Based on the previous information the following aim of the research can be formulated:

To what extent are stock prices/returns of stronger brands, less affected by crisis periods compared to weaker brands? Is there a difference in this effect during the decline and recovery of a crisis period, and during different types of crisis periods?

(7)

6

(8)

7

2|Theoretical framework

The main focus of the research will be the relationship between brand equity and stock prices during crises periods. In short the three variables of interest will be explained; the first variable is brand equity, according to Aaker (1996) “the set of assets inherent in a brand that add value to a firm and its customers”. The set of assets consists of brand name awareness, brand loyalty, perceived quality and brand associations (Aaker, 1996; Bick, 2009). In the literature many definitions of brand equity can be found, the two characteristics that can be found in most definitions is the financial value for the firm (shareholders) and the customer value. The next variable is stock prices, stock can be seen as the ownership in a corporation by means of shares. Stock prices are a good tool to see whether corporations perform good or bad even when no new information is released. When for example agents get some positive signals about a corporation they buy those shares, which cause an increase in stock prices. This increase in stock price gives a signal to the investors that the company has good prospects (Peress, 2010). The third variable is crisis periods, which can be seen as one of the possible market conditions that could affect a company, and which is almost impossible to control (macro environmental factor) (Leeflang, 2003). During these periods the stock prices of the companies decline enormously. It is however difficult to cope with these crisis periods, as it is never exactly clear when the next crisis period will take place and how big the effect on the economy will be (Zhou and Zhu, 2010).

(9)

8 The relation between brand equity and stock prices

It is already stated in the introduction that there is a link between branding and the financial performance of a firm. A research by Barth, Clement, Foster and Kaszkik (1998), indicated that Interbrand values (linked to brand equity) are positively related to stock prices and stock returns. A research of Conchar, Crask and Zinkhan (2005), also indicated a significant positive relationship between branding activities such as advertising/promotion spending and the market value of the firm (financial performance of the firm).

Another important point that can be found in the relation between brand equity and stock prices is risk. It can be found in the theoretical literature that investments in branding (market based assets), should not only lead to increased returns but also to lower risks associated with these returns (Rego, Billet and Morgan, 2009). This lower risk also points out the increased value of these returns.

Of course branding itself does not have a direct (positive) effect on stock prices. However, branding has an effect on the behavior of consumers and investors regarding a brand. These ‘strong’ brands mostly have a good position in the market. Also in situations of uncertainty strong brands have many direct and indirect advantages (Hoeffler and Keller, 2003). The aspect of risk is of course very important for investors, when less risk is involved with ‘strong’ brands, investors are more likely to invest in them.

To conclude this section it can be said that branding has a positive effect on stock prices as ‘strong’ brands increase the willingness to buy for consumers. Besides, branding decreases the risk for investors to invest in ‘strong’ brands. These two elements both cause a positive increase in the stock prices of ‘strong’ brands.

The following two hypotheses are formulated to test whether statements made about the relation between brand equity and stock prices are supported within this research as well: H1| Strong brands have a positive effect on the stock prices of firms.

(10)

9 The influence of crisis periods on the relation between brand equity and stock prices

Whether the positive relation between brand equity and stock prices holds or not, there are also some possible moderators that could effect this relation and which forms, as said before, the main element of this research. Within the Madden et al., (2006) research the positive relation between brand equity and stock price was proven. It was however indicated in the limitations that it should be tested whether this relation holds during different market conditions.

One of the possible market conditions are crisis periods. Crisis periods of course have a negative effect on stock prices. For investors the risk to invest in firms is much higher during crisis periods than during economic stable times. This makes it more important for firms to prevent that their company is seen as a high risk investment. Besides, during these crisis periods firms will probable get less customers, which results in a lower performance of the firm.

Based on these reactions towards crisis periods it is interesting to see whether high brand equity still has a positive effect on stock prices. It sounds quite logical that ‘strong’ brands are more stable then ‘weaker’ brands and that investing in these firms/brands is without high risks for investors. The same can be said for customers, it sounds logical that they keep their trust in these ‘strong’ brands more easily than in ‘weaker’ brands. It is however the question whether these ‘logical’ reactions really happen when there is a period of economic instability.

The effect of crisis periods on the relation between brand equity and stock prices results in the following hypothesis:

(11)

10

3| Research design

Crisis Periods

To investigate whether strong brands have less problems during decline, and recover better when a crisis period ends, the specific crisis periods that took place between 2001 and 2011 will be determined. This is the first step in the research design, as this information is needed to create the different portfolios.

The first crisis period is the burst of the Dot-com bubble. The burst of the Dot-com bubble caused a worldwide recession. Before the burst there was an enormous increase in the stock prices of the internet sector and companies related to this sector. At March 10, 2000, the NASDAQ value reached his peak around a value of 5.000 (see figure 1), which meant more than double its value compared to the previous year. This peak indicates the end of a period of growth and is the start of a recession (crisis period)(Hall, Feldstein, Bernanke, Frankel, Gordon and Zarnowitz, 2001) only around April 4 there was a small recovery, however the decline continued shortly after that and continued until 2002. The period can be seen as a recession (crisis period) as the decline lasted more than a few months (Hall et al., 2001) Around October 2002 the recovery of the crisis period started. This recovery continued till the first quarter of 2004, afterwards the NASDAQ index became stable again (see figure 1).

Another tool to determine the burst of the Dot-com Bubble is by looking at the economic indicators. The economic indicator used in this research is the Gross Domestic Product (GDP) growth rate of the USA. The GDP is used for a long time to measure recessions, it however has it downturn as the values are calculated quarterly which is of course not as specific as a monthly measure. According to this tool we can speak of a recession when the GDP has a

(12)

11

negative value during two quarters on a row. Within figure 2 the GDP growth rate for the years 2000 till 2004 are presented. What can be seen in figure 2 is that there are not two quarters in a row that the GDP has a negative value. Therefore according to this tool there was no recession.

When combining the NASDAQ index and the GDP, and take into account the aim of the research which is related to stock returns, the Dot-com Bubble should be used in this research as the NASDAQ index showed the enormous decrease in stock returns. However, the GDP gives an indication that the first crisis period is subject to some limitations.

The second crisis period is the Credit crunch. What happens during a credit crunch is a reduction in the availability of credits. The crisis escalated in 2008, however it already ‘slowly’ started in August 2007, when the central bank had to intervene to provide liquidity (Soros, 2008). In September 2008 the Credit crunch escalated when numerous large banks got into trouble. Around September 2009 it was indicated by the OESO that the Credit Crunch came to an end. During that period most economies slightly recovered or stabilized. The NASDAQ index (see figure 1) shows that the crisis started at the end of 2007 and lasted till the beginning of 2009. The recovery after the Credit Crunch is partly visible in the NASDAQ index (see figure 1). It is however difficult to see when it precisely stabilized. For the Credit Crunch the GDP is calculated as well (see figure 3) which is known till the beginning of 2011. Within this time period we can obviously speak of a recession, there are more than two quarters on a row that the GDP has a negative value. During 2008 there is a small recovery,

(13)

12

however afterwards the GDP value is again negative during two quarters on a row. In the beginning of 2009 the recovery of the Credit Crunch starts till the beginning of 2010 when the GDP value becomes stable.

When combining the different tools the crisis period ends in the beginning of 2009 according to the NASDAQ index and the GDP, only the OESO indicated a slightly different ending time for the credit crunch.

Based on the NASDAQ index and the GDP the following time periods will be used for investigating how ‘strong’ brands cope with crisis periods:

March 2000 till October 2002 – Burst Dot-com Bubble November 2002 till March 2004 – Recovery Dot-com Bubble August 2007 till March 2009 – Credit Crunch

April 2009 till April 2010 – Recovery Credit Crunch

Portfolio creation

The second step in the research design is the creation of portfolios. Portfolio analysis is used more and more in marketing, and became an accepted method in marketing research (Madden et al, 2006; Sorescu, Shankar and Kushwaha, 2007; O’Sullivan, Hutchinson and O’Connell, 2009). Within this research the Interbrand list of most valuable brands will be used as a valuation method for the different brands. Interbrand publishes a yearly list of the top 100 most valuable brands. This list is probable the most well-known and widely used brand valuation method (Haigh and Perrier, 1997; Madden et al, 2006). Around September of each

(14)

13

year Interbrand publishes a new list of most valuable brands. In September of 2001 the first list was published, and the data is known till 2011. This means that the burst of the dot-com bubble is not completely included in the data. Besides, some of the companies that occur in one of the two portfolios cannot be used in the research because; 1) private companies (Levi’s, Armani) cannot be used as no stock return information is publicly known, 2) brands that are part of a firm which has more than one brand (Gillette and Pamper are both brands of Proctor and Gamble).

To be able to create the different portfolios it is important to identify the criteria that will be used to place the different companies from the Interbrand list into groups. It already becomes clear in the aim of the research that within this research we want to compare strong/valuable brands versus weaker/less valuable brands. Therefore within this research the brands are classified in one of the two portfolios based on the following criteria:

 High or low brand value.

The portfolio allocation is an intersection of the previous mentioned criteria, and results in two distinct portfolio groups:

Portfolio 1: Brands that are placed in the top (brand value higher than 15 million dollar) of the Interbrand list of most valuable brands

Portfolio 2: Brands that are placed in the bottom (brand value lower than 10 million dollar) of the Interbrand list of most valuable brands

The reason to choose for the value of 15 million dollar for the top brands and 10 million dollar for the bottom brands is to be able to have substantial sized portfolios, however at the same time a significant difference between the top and bottom brands.

Besides this criteria and the allocation in two distinct portfolio groups, there are also two different scenarios:

(15)

14

that the portfolios only change in the month September during economic stable times. Therefore the years that the portfolios will change is 2004, 2005, 2006, 2007, 2010 and 2011 The second crisis period starts slowly around august 2007, however as there are no enormous effects on the economy at that point the portfolio is still adjusted in 2007.

The second scenario is a bit more complex. Here the portfolios change each time the Interbrand list of most valuable brands is released. This means that the portfolios change each year in the month September, also during the two crises periods.

Table 1 gives an overview of the different scenarios and their portfolios:

Scenario 1 Scenario 2

Portfolio 1 Top brands; the same brands

during crisis periods,

portfolio changes during economic stable times

Top brands; changes each year (September)

Portfolio 2 Bottom brands; the same

brands during crisis periods, portfolio changes during economic stable times

Bottom brands; changes each year (September)

Within this research SAS 9.2 is used for doing the analyses. In appendix 1 (scenario 1) and appendix 2 (scenario 2), the codes can be found which were used for creating the portfolios. To be able to investigate whether strong brands cope better with crisis periods then weaker brands, an extra portfolio is created. This portfolio should give an indication of the overall market performance and is the benchmark portfolio for this research, in this research the S&P 500. The S&P 500 is an index that gives the most reliable indication on the developments in the USA stock market.

Descriptive statistics

(16)

15

differ per year between 2001 and 2011. In appendix 3 the brands included in scenario 1 (portfolio 1 and 2) can be found and in appendix 4 the brands included in scenario 2 (portfolio 1 and 2).

In the figures 4 till 7 the sectors of portfolio 1 and 2 of both scenarios are presented. The percentages for the different sectors indicate the presence of the sectors in each portfolio over the years (2001 till 2011).

Figure 4| Portfolio 1/Scenario 1 Figure 5| Portfolio 2/Scenario 1

Figure 6| Portfolio1/Scenario 2 Figure 7| Portfolio 2/Scenario 2

Within all the four portfolios electronics is an important sector. For the portfolios 1 in both scenarios business services and financial services are important sectors. It is interesting to see that Media is quite a large sector in portfolio 1 of scenario 2, however is not even present in portfolio 1 of scenario 1. For the portfolios 2 in both scenarios FMCG and Financial services are important sectors. The differences between both portfolios 2 are less than for both portfolios 1. The main difference between the portfolios 1 and 2, is the absence of FMCG in both portfolios 1.

Stock Returns

(17)

16

reflect all available information about a firm (Madden et al., 2006). Related to this research, when a crisis period occurs (event) the behavior of the stock price can be investigated, and could help to give an indication in the market forecast of the firms’ future income (Madden et al., 2006). The stock returns of the portfolios as a whole will be calculated instead of the individual firms. When the stock returns are calculated for each portfolio it is possible to diversify away most of the idiosyncratic risk (Aksoy, Cooil, Groening, Keiningham and Yalcin, 2008). The only portfolio that has no idiosyncratic risk at all is the market portfolio, besides this market portfolio has no abnormal returns. To be able to test the different hypotheses it is needed that at some points the portfolios are non-diversified, this makes it possible to see whether some of the portfolios earn abnormal returns during the decline and recovery of crisis periods (Mashruwala, Rajgopal and Shevlin, 2006).

To calculate the stock returns of the different brands per month the following formula is used:

R = LN(Pᵗ/Pᵗˉ¹)*100% Where: R = Stock return per month

P = Adjusted close price of the month

P ¹ = Adjusted close price of the month one month before P

The reason to present the stock returns per month in percentages is to make each company in the portfolio equally important. Otherwise a company with very high or low stock returns will have an enormous effect on the stock returns of the overall portfolio. The adjusted close price is calculated in the following way: alternate monthly price derived from the daily prices divided by the CFACPR (cumulative factor to adjust price) value. The codes to link the portfolios to the stock data in SAS can be found in appendix 5.

Risk

Risk is an important element to take into account when stock returns are used to investigate the possible importance of branding in creating shareholder value. The stock returns should therefore be adjusted for risk to be able to make meaningful conclusions.

(18)

17

The following formula is used to calculate it:

R = the stock return in month t at the risk free rate = the excess return on the S&P 500

= the error term

= the intercept of the regression = the slope of the regression

The alphas and betas can be used to test the differences between the stock performances of the portfolios. A normal intercept is zero, in this situation the case for the benchmark portfolio (S&P500). When a portfolio has a positive alpha it means that it outperforms the benchmark portfolio, when the alpha is negative it performs worse than the benchmark portfolio. For the Beta, one means that the risk is as expected, this is again the case for the benchmark portfolio. When the beta is lower than one the risk is lower than the overall market (a risk free company has a beta of zero). When the beta is higher than one the risk is higher than the overall market. It can be said that companies with low betas have stock returns that are less sensitive to things happening in the market (Chambers and Lacey, 2008).

As it should be tested whether there is a significant difference between the stock returns of the different portfolios and the benchmark portfolio a regression analysis is performed.

Buy and Hold

Another aspect within this research is the use of Eventus/Buy and Hold. To see whether shareholder value is created, the event study method is a valuable tool. With Eventus it is possible to see whether an event - in this case a crisis period – causes abnormal movements in the prices of the stock (Madden et al., 2006).

Within the Buy and Hold research only the portfolios that remain stable during the crisis periods are used, as it is buy and hold. Therefore the portfolios of the first scenario are used. There are some adjustments needed in the dataset. A text file needs to be created with permno codes as an identification for the different brands. The event time should be added, in this case the date that the crisis started. And a sign to identify in which group (portfolio) the brand is located. An example of the codes and dates are included in appendix 6.

(19)

18

(20)

19

Figure 8| Monthly stock returns portfolios and benchmark portfolio 4| Results

The first step in this research part is to compare the overall performance of the 4 portfolios and the benchmark portfolio. This will be a comparison from September 2001 till December 2011. Figure 8 is an overview of the average monthly return of the different portfolios and the benchmark portfolio. The line in the graph is an indication how much the stock return of the different portfolios increase/decrease over the indicated time frame in percentages. All the lines start at 100%.

(21)

20

for portfolio 2 of both scenarios (scenario 1; -0,049% and scenario 2; -0,024%) the average monthly stock returns are quite similar to the portfolios 1 of scenario 1 and 2. Looking at figure 8 it becomes clear that the differences between the portfolios 1 of both scenarios and the portfolios 2 of both scenarios are minimal. When the statistical tests (regression codes SAS, appendix 7) were conducted, not all the results were statistically significant at an alpha of .05. The model as a whole is significant however, when focusing on the individual variables (portfolios) only the benchmark portfolio and portfolio 2 of scenario 2 are significant. The problem in this model is that the different portfolios are highly correlated. This is quite logical as portfolio 1 of scenario 1 and 2 are the same during economic stable times, the same counts for portfolio 2 of scenario 1 and 2. Only during crisis periods all the portfolios are different. Therefore, - although at some points the different portfolios are highly correlated – all the portfolios will be used for the research as the crisis periods are the main point of interest.

Within the rest of the research the benchmark portfolio will be the independent variable and the portfolios the dependent variable. When comparing the benchmark portfolio with all the portfolios the market portfolio was significant with an alpha of .05 in all cases. This means that the independent variable (benchmark) is a good predictor of the dependent variables (portfolios).

By comparing the stock returns of the different portfolios the first step was taken in investigating whether brands with high brand equity have a positive effect on stock prices/returns of firms. To be able to say something about the difference in risk between high and low brand equity and to make meaningful conclusions the CAPM regression results are presented for the complete time frame as well.

Portfolio Alpha Market Beta

Portfolio 1 Scenario 1 -0,29 1,24

(22)

21

The performances of all the portfolios are quite similar. The benchmark portfolio has an alpha of 0, all portfolios have a negative alpha which means that they perform worse than the benchmark portfolio (see table 2). And all portfolios have a beta higher then 1 (see table 2). The market has a beta of one which means that all portfolios have more risk, this is caused by the fact that the stock returns of the portfolios have more extreme changes in value. Overall brands with high brand equity have no extra positive effect on stock prices, and there is also no reduced risk associated with the high brand equity. The adjusted R-squares of all the portfolios are between ,80 and ,90 this means that 80 till 90% of the variance in the portfolios is explained by the benchmark portfolio. This confirms the previous made statement; ‘the benchmark portfolio is a good predictor of the portfolios’.

The results presented here do not support the research of Madden et al., (2006). Both aspects; that stronger brands deliver higher returns to shareholders and the second aspect with less risk are not supported in this research. A point to take into account is that the brands in the different portfolios are part of the benchmark portfolio as well. The high risks could indicate that the portfolios are highly sensitive for things happening in the market.

The next step in the research is to investigate whether strong brands perform better during crisis periods, and have a better recovery. There are two crisis periods which will be investigated during the crisis period and during the recovery. This means that 4 periods will be investigated. First with the CAPM regression, and afterwards with Buy and Hold.

CAPM

Burst Dot-com Bubble (Sept. 2001 – Oct. 2002)

(23)

22

Portfolio 1 Scenario 1 0,21 1,39

When looking at the alpha, which are all positive it can be said that when the benchmark portfolio is 0, the value of portfolio 1 is ,21 for scenario 1 and ,07 for scenario 2 for portfolio 2 it is ,64 for scenario 1 and ,70 for scenario 2 (see table 3). Therefore it can be concluded that during the Dot-com bubble the benchmark portfolio is outperformed (not significant) by all portfolios however with significant more risk, the stock returns of the portfolios change enormously, with in this case a positive result. With an interesting note that the low brand equity portfolio (portfolio 2) outperformed the high brand equity portfolio (portfolio 1) and also with less risk.

Recovery Dot-com Bubble (Nov. 2002 – March 2004)

During the Dot-com Bubble high brand equity was a helpful tool to be less effected by the crisis with regard to stock returns. How is this situation for the recovery? When comparing the average monthly return of the portfolios they are all positive. Portfolio 1 has an average monthly return of 1,5% for scenario 1 and 1,6% for scenario 2, portfolio 2 has an average monthly return of 2,0% for scenario 1 and 1,9% for scenario 2 and the benchmark portfolio has an average monthly return of 1,5%. This indicates that portfolio 2 the ‘weaker’ brands has the best recovery after the Dot-com Bubble, besides portfolio 1 of both scenarios perform better than the benchmark portfolio as well.

(24)

23

When looking at the CAPM analysis, it is interesting to see that portfolio 2 outperforms (not significant) the benchmark portfolio and portfolio 1. Portfolio 2 has a positive alpha of ,49 for scenario 1 and ,53 for scenario 2 and the alpha of portfolio 1 is negative -,55 for scenario 1 and 2 (see table 4). Besides there is also significant less risk associated with portfolio 2 compared to the benchmark portfolio. To conclude it can be said that portfolio 2 of both scenarios (weaker brands) recover better after the Dot-com Bubble and with significant less risk than the benchmark portfolio and portfolio 1.

During the burst of the Dot-com Bubble and the recovery afterwards the portfolios with ‘weaker’ brands performed very good. Of course they were affected during the burst, however less than the benchmark portfolio and the ‘stronger’ brands portfolios. The reason that the ‘weaker’ brands are less affected during the dot-com bubble could be that less high tech brands are present in those portfolios. It is known that mainly brands such as Microsoft and IBM were affected during this crises, and these brands are of course present in the S&P 500 as well. The extreme risk associated with both portfolios 1 can be related to the presence of the high tech companies in the portfolios as well. In general it was good to change the portfolios during the crisis period (including recovery) for the ‘weaker’ brands portfolio, this portfolio performed slightly better and with less risk. The reason for this better performance could be that because the differences between the brands in the ‘weaker’ portfolios are minimal, that the brands that perform really bad are removed from the ‘weaker’ portfolios sooner (during the crisis period). The ‘stronger’ brands portfolios have really high interbrand values and although performing bad won’t have such a decrease in value that they are removed from the portfolios during the crisis period.

(25)

24 Credit Crunch (Aug. 2007 – March 2009)

The second crisis period investigated is the Credit Crunch. During this period the ‘weaker’ brands are more affected by the crisis then during the first crisis period. The average monthly return of portfolio 2 is -4,4% for scenario 1 and -4,3% for scenario 2 compared to -3,6% for portfolio 1 in both scenarios and the benchmark portfolio has an average monthly return of -2,8%. For the second time (dot-com bubble) the ‘strong’ and ‘weaker’ brands are more affected by the crisis period then the benchmark portfolio according to the average monthly stock returns.

When looking at the CAPM analysis, it becomes clear that all portfolios perform worse than the market portfolio (not significant). During this crisis period mainly portfolio 2 (weaker brands) is effected, an alpha of -,31 for scenario 1 and -,40 for scenario 2 (see table 5). The alpha of portfolio 1 is -,04 for both scenarios (see table 5). To conclude it can be said that the benchmark portfolio outperforms portfolio 1 and 2 with significant less risk.

Recovery Credit Crunch (April 2009 – April 2010)

The last period investigated in this research is the recovery of the credit crunch. It is interesting to see that again the ‘weaker’ brands portfolios outperform the benchmark portfolio during the recovery looking only at the average monthly returns and the ‘weaker’ brands outperform the ‘stronger’ brands. The average monthly returns of portfolio 2 is 4,3% for scenario 1 and 4,2% for scenario 2, for portfolio 1 it is 3,6% for scenario 1 and 3,8% for scenario 2. And the recovery is slightly lower for the Benchmark portfolio, namely 3,2%.

(26)

25

During the recovery of the credit crunch, the portfolios 1 and 2 are outperformed by the benchmark portfolio. All portfolios perform worse than the benchmark portfolio (not significant). To see the difference, when the alpha of the benchmark portfolio is 0, the value of portfolio 1 is -,82 for scenario 1 and -,51 for scenario 2 and for portfolio 2, -,39 for scenario 1 and -,70 for scenario 2 (see table 6). The adjusted R-square of all portfolios during the 4 time periods was around the .80. and .90. This proves again that the benchmark portfolio is a good predictor of all portfolios.

The risk during this crisis period is much higher for the portfolios, mainly for the ‘weaker’ brands portfolios. It could be that in time of uncertainty about credits it is positive to have a ‘stronger’ brand, it might give some extra trust for investors. For the ‘stronger’ brands portfolio it was positive to change the brands in the portfolio during the recovery of the credit crunch, however for the ‘weaker’ brands portfolios it was the other way around. It could be that the ‘stronger’ brand portfolios was positively affected by some upcoming brands such as Apple and when a brand is still performing very bad such as Citi it did not stay in the portfolio during the whole recovery.

Overall it really depends on the crisis period how severe the effect is on the different portfolios in relation to the benchmark portfolio. And it seems that the sector is an important aspect to take into account!

(27)

26 Eventus

Within Eventus the Buy and Hold Abnormal returns are used. In general these percentages can be used to measure how accurate a model is. However, it is also possible to use it so see what the effect of specific events is on stock prices. To be more specific, to see whether an event causes abnormal movements in the prices of the stock.

The Dot-com Bubble

The results of the BHAR during the Dot-com Bubble and during the recovery are presented in table 7. In table 7 it becomes clear that the burst of the Dot-com Bubble has negative effects on stock prices of portfolio 1. This is quite similar for portfolio 2, it is only less severe. The BHAR results specifically indicate that when investing in the brands of portfolio 1 during the Dot-com Bubble, that at the end of that period the returns of the investment are 18,96% less compared to similar brands not present in this portfolio, for portfolio 2 that percentage is only 0,26%. During the recovery the percentages of portfolio 1 and 2 are positive, during the recovery portfolio 1 has an increase of 7,61% and portfolio 2 an increase of 13,45%.

Portfolio Buy and Hold Abnormal return

Portfolio 1 (crisis time) -18,96%

Portfolio 1 (recovery) 7,61%

That the first portfolio is so much affected during the Dot-com bubble is most likely caused by the presence of Microsoft, Intel and IBM within that portfolio. It was already stated during the CAPM research that those brands are highly affected during this crisis period.

The Credit Crunch

The results of the BHAR during the Credit Crunch and during the recovery are presented in table 8. Again the results are negative for portfolio 1, also during the recovery. The crisis

(28)

27

period has slightly negative effect on the stock prices of the brands in portfolio 2 during the crisis and during the recovery it has a positive effect. During the Credit Crunch the BHAR specifically indicates that when investing in the brands of portfolio 1, that at the end of that period the returns of the investment are 12,11% less compared to similar brands not present in this portfolio. For portfolio 2 this is only -2,14%. For portfolio 2 the recovery of the credit crunch has a positive effect on the investments of 14,75% compared to similar brands not present in the portfolio.

Portfolio Buy and Hold Abnormal return

Portfolio 1 (recovery) -5,20%

Compare CAPM and Buy and Hold

When combining both the CAPM regression analysis and the results of the analysis with Buy and Hold it becomes clear that ‘strong’ brands are not less affected by crisis periods than ‘weaker’ brands in both CAPM and Buy and Hold. The only main difference between CAPM and Buy and Hold is the outcomes for portfolio 2 during the credit crunch. For CAPM these results are negative and for Eventus they are positive. It should however be noted that within CAPM the stock returns are compared with the benchmark portfolio (S&P500) and within Buy and Hold with similar brands that are not present in portfolio 2. It could be that there are brands similar to portfolio 2 that performed even worse during the Credit Crunch.

(29)

28

5| Conclusion and recommendations

The aim of this research was to investigate whether brands with high brand equity perform better during crisis periods then brands with lower brand equity. The reason to investigate this was to see whether it is useful to invest in branding on the long term to create shareholder value, although it is known that the probability of a crisis period each ten years is 90% (Zhou and Zhu, 2010).

The first step in the research was to investigate whether the ‘strong’ brands performed better than the ‘weaker’ brands during the whole time period (2002 till 2011). The research of Madden et al., (2006) indicated that stronger brands provided higher returns to the shareholders with less risk. The investigation of Madden et al., (2006) is not supported within this research. From 2002 till 2011 the stronger brands did not outperform the weaker brands. For both portfolios the average monthly returns over the years were slightly negative compared to the benchmark portfolio (S&P500). Therefore, H1| Strong brands have a

positive effect on the stock prices of firms and H2| Strong brands have a positive effect on the stock prices of firms with less associated risk compared to weaker brands are rejected.

The main focus of this research was to see whether stronger brands cope better with crisis periods (including recovery) then weaker brands. Two crises periods were investigated within this research, the Dot-com Bubble and the Credit Crunch. Two completely different crisis periods resulting in completely different results.

The first crisis period investigated was the burst of the Dot-com Bubble. During this crisis period both the stronger and weaker brands had a negative average monthly return compared to the benchmark portfolio. An interesting outcome was that during the recovery of the Dot-com Bubble the weaker brands outperformed the stronger brands and the benchmark portfolio. It should however be noted that the weaker brands are also on the Interbrand list of most valuable brands. The brand equity of these brands is lower, however still stronger than many other brands in the world.

(30)

29

Eventus it is indicated that the weaker brands portfolio is a positive/worthwhile investment in relation to comparable brands.

When summarizing the overall results of the crisis periods it can be concluded that H3| The

stock p ic s of fi ms wit ‘st ong’ b ands ( ig b and quity) a l ss affected by crisis p iods compa d to fi ms wit ‘w ak ’ b ands (low b and quity) should be rejected.

When giving an answer to the aim of the research: To what extent are stock prices/returns of

stronger brands, less affected by crisis periods compared to weaker brands? Is there a difference in this effect during the decline and recovery of a crisis period, and during different types of crisis periods? It becomes clear that the stronger brands are more affected by the

crisis periods than the weaker brands. The stronger brands are effected severely by both crisis periods, which could be an indication for the overall (2002 till 2011) negative monthly stock returns of these brands. The brands listed in the bottom of the interbrand list of most valuable brands had a positive recovery after the crisis periods and mainly after the Dot-com Bubble. However, these brands are severely affected during the crisis periods as well.

It is recommended that companies should not intensively invest in branding. To a certain extent it could have a positive effect during the recovery of a crisis period. However, when the brand becomes really strong (high brand equity) the investments in branding could have a negative effect on the stock prices/returns, especially during crises periods.

Limitations and further research

One of the main limitations within this research is the fact that the industries within the portfolios are extremely different. The main industry in one portfolio was completely missing in the other portfolio. The effects of crises periods on the different industries are most likely to be different. Therefore when performing a research related to this research it is interesting to see what the effect of the different industries is on the results presented here. It is however difficult to include all industries in this follow up research as there are many brands with ‘medium’ brand equity however less brands with really high brand equity. Some industries can simply not be found in the top of the interbrand list of most valuable brands.

(31)

30

the Interbrand list of most valuable brands. And maybe there are even more methods to divide brands in different portfolios.

(32)

31

| References

Aaker, D.A. (1996) Building Strong Brands, New York: The Free Press

Aksoy, L. Cooil, B. Groening, C. Keiningham, T.L. Yalcin, A. (2008), The long-term stock market valuation of customer satisfaction. Journal of Marketing, Vol. 74(4), pp 105-122 Barth, M.E. Clement, M.B. Foster, G. Kaszkik, R. (1998), Brand values and capital market valuation. Review of Accounting studies, Vol. 3, pp. 41-68

Bick, G.N.C. (2009), Increasing shareholder value through building customer and brand equity, Journal of Marketing Management, Vol. 25, pp. 117-141

Chambers, D.R. Lacey, N.J. (2008), Modern corporate finance: Theory & Practice. Fifth edition, Hayden McNeill Publishing inc.

Conchar, M.P. Crask, M.R. Zinkhan, G.M. (2005), Market valuation models of the effect of advertising and promotional spending: a review and Meta-analysis. Journal of the Academy of Marketing Science, Vol. 33(4), pp. 445-460

DeLong, J.B. Magin, K. (2006), A short note on the size of the Dot-Com Bubble. NBER working paper series, National Bureau of Economic Research

Haigh, D. Perrier, R. (1997), Valuation of Trademarks and Brand Names. Brand Valuation, 3th edition Raymond Perrier, London, Premier books, pp. 19-24

Hall, R. Feldstein, M. Bernanke, B. Frankel, J. Gordon, R. Zarnowitz, V. (2001), The Business-Cycle Peak of March 2001. National Bureau of Economic Research.

Hoeffler, S. Keller, L.K. (2003), The marketing advantages of strong brands. Brand Management, Vol. 10(6), pp. 421-445

Leeflang, P.S.H. (2003), Marketing. Wolters-Noordhoff bv

Madden, T.J. Fehle, F. Fournier, S. (2006), Brands Matter: An empirical demonstration of the creation of shareholder value through branding, Journal of the Academy of Marketing Science, Vol. 34(2), pp. 224-235

Mashruwala, C. Rajgopal, S. Shevlin, T. (2006), Why is the accrual anomaly not arbitraged away? The role of idiosyncratic risk and transaction costs. Journal of Accounting and Economics, Vol. 42(1-2), pp. 3-33

Mitchell, M. and Stafford, E. (2000), Managerial decisions and long-term stock price performance, Journal of Business, Vol. 73, pp. 287-329

(33)

32

Peress, J. (2010), Learning From Stock Prices and Economic Growth. Insead and CEPR Rego, L.L. Billett, M.T. Morgan, N.A. (2009), Consumer-based brand equity and firm risk. Journal of Marketing, Vol. 73, pp. 47-60

Sorescu, A.B. Shankar, V. Kushawa, T. (2007), New product preannouncements and shareholder value: Don’t make promises you can’t keep. Journal of Marketing Research, Vol. 44(3), pp. 468-489

Soros, G. (2008), The Crash of 2008 and what it means. PublicAffairs, Perseus Books Group

Srivastava, R.K. Shervani, T.A. Fahey, L. (1998), Market-Based assets and Shareholder value: A framework for analysis, Journal of Marketing, Vol. 62 (Jan), pp. 2-18

Zhou, G. & Zhu, Y. (2010), Is the recent financial crisis really a ‘One-in-a-century’ event? Financial Analysist Journal, pp. 24-27

(34)

33

|Appendices

Appendix 1| Creation portfolios scenario 1

data have;

input Brand &$10. Year Interbrand_Value :commax. GVKEY Ticker_symbol$ Sector$;

cards;

;

run;

data p1 p2;

set have(where=(year in (2001200420052006200720102011)));

if Interbrand_Value gt 15thenoutput p1; elseif Interbrand_Value lt 10thenoutput p2;

run;

procsql;

createtable year asselectdistinct year from have;

createtable portfolios1 as

select a.year,Brand ,Interbrand_Value ,GVKEY,Ticker_symbol, sector from (select year from year where year in (200120022003)) as a, (select * from p1 where year=2001)

union all corresponding

select * from p1 where year in (200420052006200720102011) union all corresponding

select a.year,Brand ,Interbrand_Value ,GVKEY,Ticker_symbol, sector

from (select year from year where year notin (20012002200320042005200620072010 2011)) as a,

(select * from p1 where year=2007);

createtable portfolios2 as

select a.year,Brand ,Interbrand_Value ,GVKEY,Ticker_symbol, sector from (select year from year where year in (200120022003)) as a, (select * from p2 where year=2001)

union all corresponding

select * from p2 where year in (200420052006200720102011) union all corresponding

select a.year,Brand ,Interbrand_Value ,GVKEY,Ticker_symbol, sector

from (select year from year where year notin (20012002200320042005200620072010 2011)) as a,

(select * from p2 where year=2007);

(35)

34

Appendix 2| Creation portfolios scenario 2

data have;

cards;

;

run;

procsortdata=have;

by brand year;

data p1 p2; set have; by brand;

if Interbrand_Value>15thenoutput p1;

if Interbrand_Value<10thenoutput p2;

(36)

35

Appendix 3| Brands per portfolio scenario 1

Within these portfolios the brands should be equal during the crisis periods. For the brands in red this is not the case.

Portfolio1

2001 2002 2003 2004 2005 2006

Microsoft Microsoft Microsoft Microsoft Microsoft Microsoft AT&T AT&T AT&T Coca Cola Coca Cola Coca Cola Coca Cola Coca Cola Coca Cola GE GE GE

GE GE GE IBM IBM IBM

IBM IBM IBM Marlboro Marlboro Marlboro Marlboro Marlboro Marlboro Disney Disney Disney

Ford Ford Ford HP HP HP Disney Disney Disney McDonalds McDonalds McDonalds McDonalds McDonalds HP American Express American Express American Express

Sony Sony McDonalds Intel Intel Intel Merrill Lynch Merrill Lynch Sony Citi Honda Honda American Express American Express Merrill Lynch Cisco Citi Citi

Intel Intel American Express Nokia Cisco Cisco Citi Citi Intel Nokia Nokia Cisco Cisco Citi

Nokia Nokia Cisco Nokia

2007 2008 2009 2010 2011

Microsoft Microsoft Microsoft Microsoft Microsoft Coca Cola Coca Cola Coca Cola Coca Cola Coca Cola

GE GE GE GE GE

IBM IBM IBM IBM IBM Marlboro Marlboro Marlboro Marlboro Oracle

Disney Disney Disney Disney Disney

HP HP HP HP HP

McDonalds McDonalds McDonalds McDonalds McDonalds American Express American Express American Express Apple Apple

Intel Intel Intel Intel Intel Honda Honda Honda Honda Honda

Citi Citi Citi Cisco Cisco Cisco Cisco Cisco Nokia Nokia Nokia Mercedes-Benz Mercedes-Benz Google Google Google Nokia Nokia

(37)

36

Portfolio 2

2001 2002 2003 2004 2005 2006

Amazon.com Amazon.com Amazon.com Accenture Accenture Accenture AOL AOL AOL Amazon.com Amazon.com Amazon.com Apple Apple Apple Apple Apple Apple AVON AVON AVON AVON AVON AVON

Barbie Barbie Barbie Boeing BP BP Benetton Benetton Benetton BP Canon Canon

Boeing Boeing Boeing Canon Caterpilla Caterpillar BP BP BP Caterpillar Colgate Colgate

Burger King Canon Canon Colgate EBAY EBAY

Canon Colgate Colgate EBAY Gap Gap Colgate Dell Dell Estee Lauder Goldman Sachs Goldman Sachs

Dell FedEx FedEx Gap Google Heinz FedEx Gap Gap Goldman Sachs Heinz ING

Gap Goldman Sachs Goldman Sachs Heinz ING Johnson & Johnson Goldman Sachs Gucci Gucci HSBC J.P. Morgan Kellogg's

Gucci Heinz Heinz ING Johnson & Johnson Kleenex Heinz Hilton Hilton J.P. Morgan Kellogg's Kodak Hilton Jack Daniels Jack Daniels Johnson & Johnson Kleenex Kraft Jack Daniels Kellogg's Kellogg's Kellogg's Kodak LG

Kellogg's Kleenex Kleenex Kleenex Kraft Morgan Stanley Kleenex Kraft Kraft Kodak LG MTV

Merck Merck Merck Kraft MTV Nissan Mobil Mobil Mobil Merck Nissan Novartis

MTV MTV MTV Mobil Novartis Pfizer Nike Nike Nike MTV Pfizer Philips NIVEA NIVEA NIVEA Nike Philips Reuters

Panasonic Panasonic Pfizer Nissan Reuters Shell Pfizer Pfizer Philips Philips SAP Siemens Philips Philips Ralph Lauren Ralph Lauren Siemens Starbucks Ralph Laur Ralph Lauren Reuters Reuters Starbucks Tiffany & CO

Reuters Reuters SAP SAP Tiffany & CO UBS Starbucks SAP Siemens Siemens UBS Wrigley Texas Instruments Siemens Starbucks Starbucks UPS Xerox

Tiffany & CO Starbucks Texas Instruments Tiffany & CO Wrigley Yahoo Wall Street Journal Texas Instruments Tiffany & CO TIME Xerox

Wrigley Tiffany & CO Wall Street Journal UBS Yahoo Xerox Wall Street Journal Wrigley Wrigley

(38)

37

2007 2008 2009 2010 2011

Accenture Accenture Accenture 3M 3M AIG AIG AIG Accenture Accenture Allianz Allianz Allianz Adobe Adobe Amazon.com Amazon.com Amazon.com Amazon.com AVON AVON AVON AVON AVON Barclays

AXA AXA AXA Barclays Blackberry BP BP BP Blackberry Caterpillar Caterpillar Caterpillar Caterpillar Caterpillar Colgate

Colgate Colgate Colgate Colgate Citi EBAY EBAY EBAY Campbells Credit Suisse

Ford Ford Ford Citi Dell Gap Gap Gap Credit Suisse EBAY Harley Davidson Harley Davidson Harley Davidson Dell Ford

Heinz Heinz Heinz EBAY Gap Hertz Hermes Hermes Ford Goldman Sachs

ING Hertz Hertz Gap Harley Davidson Johnson & Johnson ING ING Goldman Sachs Heinz

Kellogg's Johnson & Johnson Johnson & Johnson Harley Davidson Jack Daniels Kleenex Kellogg's Kellogg's Heinz John Deere

Kodak Kleenex Kleenex Jack Daniels Johnson & Johnson Kraft Kodak Kodak Johnson & Johnson Kleenex

LG Kraft Kraft Kleenex Morgan Stanley MTV LG LG Morgan Stanley MTV Nissan MTV MTV MTV Panasonic

Philips Nissan Panasonic Panasonic Philips Ralph Lauren Philips Philips Philips Santander

Reuters Ralph Lauren Ralph Lauren Santander Shell Shell Shell Shell Shell Sony Siemens Siemens Siemens Siemens Siemens Starbucks Starbucks Starbucks Starbucks Starbucks Tiffany & CO Tiffany & Co Tiffany & Co Thomson Reuters Thomson Reuters

UBS UBS UBS Tiffany & Co Tiffany & Co Wrigley Wrigley Xerox UBS UBS

Xerox Xerox Yahoo Visa Visa Yahoo Yahoo Xerox Xerox

(39)

38

Appendix 4| Brands per portfolio scenario 2 Portfolio 1

2001 2002 2003 2004 2005 2006

American Express American Express American Express American Express American Express American Express AT&T AT&T Cisco Cisco Cisco Cisco

Cisco Cisco Citi Citi Citi Citi Citi Citi Coca Cola Coca Cola Coca Cola Coca Cola Coca Cola Coca Cola Disney Disney Disney Disney

Disney Disney Ford GE GE GE Ford Ford GE HP Honda Honda

GE GE Honda IBM HP HP IBM Honda HP Intel IBM IBM Intel IBM IBM Marlboro Intel Intel Marlboro Intel Intel McDonalds Marlboro Marlboro McDonalds Marlboro Marlboro Microsoft McDonalds McDonalds Merrill Lynch McDonalds McDonalds Nokia Microsoft Microsoft

Microsoft Microsoft Microsoft Nokia Nokia Nokia Nokia Nokia

Sony

2007 2008 2009 2010 2011

American Express American Express Apple Apple Apple Cisco Cisco Cisco Cisco Cisco Citi Citi Coca Cola Coca Cola Coca Cola Coca Cola Coca Cola Disney Disney Disney

Disney Disney GE GE GE GE GE Google Google Google Google Google Honda Honda Honda Honda Honda HP HP HP

HP HP IBM IBM IBM IBM IBM Intel Intel Intel Intel Intel Marlboro Marlboro McDonalds Marlboro Marlboro McDonalds McDonalds Microsoft McDonalds McDonalds Mercedes-Benz Microsoft Nokia

Microsoft Mercedes-Benz Microsoft Nokia Oracle Nokia Microsoft Nokia

(40)

39

Portfolio 2

2001 2002 2003 2004 2005 2006

Amazon.com Accenture Accenture Accenture Accenture Accenture AOL Amazon.com Amazon.com Amazon.com Amazon.com Amazon.com Apple AOL AOL Apple Apple Apple AVON Apple Apple AVON AVON AVON

Barbie AVON AVON Boeing BP BP Benetton Barbie Barbie BP Canon Canon

Boeing Boeing Boeing Canon Caterpillar Caterpillar BP BP BP Caterpillar Colgate Colgate Burger King Canon Canon Colgate EBAY EBAY Canon Caterpillar Caterpillar EBAY Gap Gap Colgate Colgate Colgate Estee Laud Goldman Sachs Goldman Sachs

Dell Dell FedEx Gap Google Heinz FedEx FedEx Gap Goldman Sachs Heinz ING

Gap Gap Goldman Sachs Heinz ING Johnson & Johnson Goldman Sachs Goldman Sachs Gucci HSBC J.P. Morgan Kellogg's

Gucci Gucci Heinz ING Johnson & Johnson Kleenex Heinz Heinz HSBC J.P. Morgan Kellogg's Kodak Hilton J.P. Morgan J.P. Morgan Johnson & Johnson Kleenex Kraft Jack Daniels Jack Daniels Jack Daniels Kellogg's Kodak LG

Kellogg's Johnson & Johnson Johnson & Johnson Kleenex Kraft Morgan Stanley Kleenex Kellogg's Kellogg's Kodak LG MTV

Merck Kleenex Kleenex Kraft MTV Nissan Mobil Kodak Kodak Merck Nissan Novartis

MTV Kraft Kraft Mobil Novartis Pfizer Nike Merck Merck MTV Pfizer Philips NIVEA Mobil Mobil Nike Philips Reuters Panasonic MTV MTV Nissan Reuters Shell

Pepsi Nike Nike Philips SAP Siemens Pfizer NIVEA Nissan Ralph Lauren Siemens Starbucks Philips Panasonic NIVEA Reuters Starbucks Tiffany & Co Ralph Lauren Pepsi Philips SAP Tiffany & Co UBS

Reuters Pfizer Ralph Lauren Siemens UBS Wrigley Starbucks Philips Reuters Starbucks UPS Xerox Texas Instruments Ralph Lauren SAP Tiffany & Co Wrigley Yahoo

Tiffany & Co Reuters Starbucks TIME Xerox Wall Street Journal SAP Tiffany & Co UBS Yahoo

Wrigley Starbucks Wall Street Journal Wrigley Xerox Tiffany & Co Wrigley Xerox Yahoo Wall Street Journal Xerox Yahoo

Wrigley Yahoo Xerox

(41)

40

2007 2008 2009 2010 2011

Accenture Accenture Accenture 3M 3M AIG AIG Adobe Accenture Accenture Allianz Allianz Allianz Adobe Adobe Amazon.com Amazon.com Amazon.com Amazon.com AVON AVON AVON AVON AVON Barclays

AXA AXA AXA Barclays BlackBerry BP BlackBerry BlackBerry BlackBerry Caterpillar Caterpillar BP BP Campbells Citi

Colgate Caterpillar Burger Kin Caterpillar Colgate EBAY Colgate Campbells Citi Credit Suisse

Ford EBAY Caterpillar Colgate Dell Gap FedEx Colgate Credit Suisse EBAY Harley Davidson Ford EBAY Dell Ford

Heinz Gap Ford EBAY Gap Hertz Harley Davidson Gap Ford Goldman Sachs

ING Heinz Goldman Sachs Gap Harley Davidson Johnson & Johnson Hermes Harley Davidson Goldman Sachs Heinz

Kellogg's ING Heinz Harley Davidson Jack Daniels Kleenex Johnson & Johnson Hermes Heinz John Deere

Kodak Kellogg's J.P. Morgan Jack Daniels Johnson & Johnson Kraft Kleenex Johnson & Johnson Johnson & Johnson Kleenex

LG Marriott Kleenex Kleenex Morgan Stanley MTV Morgan Stanley Morgan Stanley Morgan Stanley MTV Nissan MTV MTV MTV Panasonic Philips Philips Panasonic Panasonic Philips Ralph Lauren Shell Philips Philips Santander

Reuters Siemens Ralph Laur Santander Shell Shell Starbucks Shell Shell Siemens Siemens Tiffany & Co Siemens Siemens Sony Starbucks UBS Starbucks Starbucks Starbucks Tiffany & Co Wrigley Thomson Reuters Thomson Reuters Thomson Reuters

UBS Xerox Tiffany & Co Tiffany & Co Tiffany & Co Wrigley Yahoo UBS UBS UBS

Xerox Visa Visa Visa Yahoo Xerox Xerox Xerox

(42)

41

Appendix 5| Connect portfolios and stock prices

data monthly;

informat date ddmmyy10. ;

format date ddmmyy10.;

input Date Ticker_Symbol $ Price $ CFACPR $;

cards;

;

data yearly;

cards;

;

procsql;

createtable linked asselecta.*,Date, Price, CFACPR from yearly as a,monthly as b where year(date)=year and a.Ticker_Symbol=b.Ticker_Symbol;

quit;

procprintdata=linked(obs=10);

(43)

42

Appendix 6| Example text file eventus Crisis period 2

(44)

(45)

44

Appendix 7| Regression codes SAS

In this situation Markt is the dependent variable, main part of the research it is however the other way around.

procregdata = Mylib.Marktportfolios;

model Markt = P1 P2 P3 P4;