“The Effect of Brand Value Changes on
Firm Value and its Liquidity”
Joost Korse 6056326
Master thesis: Business Economics, Finance Faculty Economics & Business
University of Amsterdam
Thesis supervisor: Mr. Dr. J.E. Ligterink
Abstract
Brand-building efforts, e.g. advertising, can lead to brand value changes, which might have an impact on the firm value of the brand-owning firm. Also, advertising can have an impact on investor behavior. Namely, advertising can lead to a greater visibility and it may also attract more uninformed investors. These two consequences might lead to an increase of the stock’s liquidity (Grullon, Kanatas and Weston, 2004). According to existing literature, strong brands appear to have a low beta and can therefore be safe havens in periods of high market volatility (Bris, Smit and Sorell, 2010). Therefore, this study examines the beta of global brands and analyzes if an announcement of brand values results in an immediate, significant abnormal stock return for the brand-owning firms. Besides, this study examines the relationship between brand value changes and advertising expenses on the liquidity of the firm for the period 2007-2013. This study uses brand value announcements provided by Interbrand and Brand-Finance, world’s leading consultancy firms. This study is unique since it examines the beta of global brand-owning firms combined with the analysis of the impact of brand value announcements on the market value of the brand owner. The latter combined with analyzing the relationship between advertising, brand value changes on liquidity makes this study unique. The results of this study show that global brands do not have a lower beta compared to the market. Also, the results show that brand value announcements do not have a significant impact on firm value. This can be disadvantageous for marketing managers, who are under pressure to justify their brand-building efforts. Also, this study shows that strong brands do not have a low beta and therefore cannot be seen as safe havens in periods of high market volatility. Investors who are seeking for save investments during economic downturn should take this into account. Furthermore, the results show that liquidity is not related with advertising expenses, change in brand value and brand ranking. Therefore it is unlikely that a firm’s advertising may attract uninformed investors to its stock. The reason that this study does not find significant results for these relationship might be due to a limited dataset or endogeneity problems.
1
Table of Contents
1 Introduction 2
2 Literature review 5
2.1. Brand value 5
2.1.1. Importance of brand value 5
2.1.2. Brand value and its relation to firm value 6
2.1.3. Characteristics of a global brand firm 9
2.2. Advertising 11
2.2.1. Reasons of advertising 11
2.2.2. The effects of advertising 13
2.3. Liquidity 16
2.3.1. Properties and characteristics of liquidity 16
2.3.2. Liquidity and asset prices 17
2.3.3. Liquidity and its determinants 18
2.3.4. Effect of advertising on liquidity 20
3. Hypotheses & Methodology 22
3.1. Hypotheses 22
3.2. Methodology 26
4. Data 34
4.1. Sample selection and variable description 34
4.2. Sample characteristics 37
5. Results 41
5.1. The effect of brand value announcements on stock return 41
5.2. The effect of advertising expenses and brand value changes on liquidity 44
6. Conclusion 47
References 50
2
1 Introduction
Nowadays there is an increasing pressure on marketing managers to justify the impact of their brand-building efforts on total firm value (Rust, Ambler, Carpenter, Kumar and Srivastava, 2004). The main goal of these efforts is to increase the brand value of the firm. Interbrand, world’s leading brand consultancy firm, publishes the global top 100 most valuable brands since 2001. Such an announcement could affect the market value of the brand-owning firm. Besides the effect on the market value of the firm, brand-building efforts such as advertising might also lead to a greater visibility of the firm which can influence investor behavior. The aim of this study is to analyze if an announcement of brand value results in an immediate, significant abnormal stock return for the brand-owning firm. Also this study examines the effect of advertising and brand value changes on the liquidity of this firm.
Besides justifying managers’ brand-building efforts, this can also be interesting because managers adjust firm advertising expenditures to influence investor behavior and short-term stock prices (Lou, 2009). Having a strong global brand might be interesting due to the fact that they have a lower beta compared to the market. Therefore, firms owning strong brands might be a safe haven in periods of high market volatility (Bris et al., 2010). Having a stock with a low beta compared to the market can also be an advantage, because Hong and Sraer (2012) show that low beta stocks outperforms high beta stocks. Increasing the brand value might be advantageous because of above mentioned reasoning. Therefore, this study focuses on the impact of brand value changes of global brand-owning firms and examines the beta of these firms. Moreover, firms with greater advertising expenditures have a larger number of both individual and institutional investors. These advertising expenditures lead to a greater visibility of the firm. This greater visibility can have a positive effect on brand recognition. Firms with a greater visibility have better liquidity of their common stock (Grullon et al., 2004). The question if brand-building efforts make sense is also relevant because of the ongoing debate on the balance sheet recognition of intangible assets (Lev, 2008). There are several earlier studies that have documented a positive relation between brand value estimates provided by brand consultancy firms and firm value (Rust et al., 2004; Barth, Clement, Foster and Kasznik, 1998).
This study will therefore answer the following questions: “To what extent does an announcement on brand values result in an immediate, significant positive abnormal stock return for the brand owning firm? And to what extent do advertising expenses and brand value changes effect the liquidity of this firm?” Besides that brand-building efforts, e.g. advertising,
3
can lead to brand value changes, which might have an impact on the firm value of the brand-owning firm, it can also have an impact on investor behavior. Namely, advertising can lead to a greater visibility and it may also attract more uninformed investors. These two consequences might lead to an increase of the stock’s liquidity (Grullon et al., 2004). Therefore this study also examines the relationship between advertising expenses and brand value changes on the liquidity of the firm. This study contributes to existing literature due to its uniqueness by examining the beta of global brand-owning firms combined with the analysis of the impact of brand value announcements on the market value of the brand owner. The latter combined with analyzing the relationship between advertising, brand value changes on liquidity makes this study unique. Furthermore, a limitation of existing research is that they only concentrate on brand value announcements provided by Interbrand and only focus on U.S. brands. This study also includes brand value announcements provided by Brand-Finance and non-U.S. brands. The research questions are answered by conducting an empirical research.
This study retrieves estimated brand values from Interbrand’s Best Global Brands lists and Brand-Finance’s Global top 100 over the period 2007, the year of Brand-Finance’s first publication of the Global top 100, to 2013. Besides the data about the brand value announcements, also all stock and market return data is needed. These data are collected from the Datastream database. To examine the impact of the brand value announcements, the abnormal returns are calculated using the stock and market returns and subsequently these are tested for significance. To test if global brands have a lower beta compared to the market this study uses the simple comparison of means method regarding these betas. To examine the relationship between advertising expenses, brand value changes on liquidity, this study uses a panel regression. This research measures liquidity using relative bid-ask spreads. The same sample period is used as for the impact of brand value announcements, namely 2007 – 2013. The liquidity measure is regressed on a yearly basis using OLS on the variables of interests and other control variables. The variable advertising expenses is collected from the Compustat database, while estimated brand values are retrieved as mentioned above. Control variables are collected from Datastream.
This study finds empirical evidence rejecting the theory that global brands have a lower beta than the market. Subsequently this study does not find evidence that externally provided brand value announcements have a direct impact on the firm value of the brand owner. No evidence has been found for announcements of both brand consultancy firms, i.e. Brand-Finance and Interbrand. Also, the relative brand value change and change in ranking is
4
not related to the (cumulative) abnormal returns. This might be due to the fact that stockholder may be able to anticipate a large fraction of the brand value estimates of Interbrand and Brand-Finance. Bris et al. (2010) mention that strong brands are safe havens in periods of high market volatility. However, this study shows that strong brands do not have a lower beta than the market. Also it is shown that the beta of global brands rise in periods of high market volatility. This could potentially lead to a decrease in the assessments of the added value of the brands during this period, therefore a change in brand value might not have an effect on firm value. Regarding the impact of advertising expenses and brand value changes on the dependent variable, i.e. the relative bid-ask spread, this study does not find any significant relationship. This study shows there is no significant relationship between advertising expenses and the relative bid-ask spread. Besides, this study shows there is no relationship between brand value changes and liquidity. Finally, it is shown that firms listed in the top 10 of the global top 100 do not have a better liquidity after adding all control variables. Summarized, the results show that brand value announcements do not have a significant impact on firm value and global brands do not have a lower beta compared to the market. This can be disadvantageous for marketing managers, who are under pressure to justify their brand-building efforts. Also, according to existing literature, strong brands can be safe havens in periods of high market volatility. This study shows that this is not the case. Investors who are seeking for save investments during economic downturn should take this into account. Furthermore, the results show that liquidity is not related with advertising expenses and changes in brand value. Therefore it is unlikely that a firm’s advertising may attract uninformed investors to its stock.
The remainder of this paper is organized as follows. Section two provides a brief overview of existing literature relevant for this topic. Section three develops the hypotheses based on the literature review and discusses the methodology to test these hypotheses. Description of the data and the variables that are used are presented in section four. Section five presents the results of this study. This study concludes with summarizing and discussing the results.
5
2 Literature review
In the first part of this chapter a brief overview of brand value itself and its relation to the stock price is given as it provides an important background to the empirical research that follows. Increasing brand value can be achieved by advertising. This is addressed in the second part of this chapter. Besides, existing literature about other reasons and effects of advertising is assessed. The last part of this chapter covers liquidity and its characteristics, since these are affected by advertising. Also the relation between liquidity and asset prices is addressed followed by setting out the determinants of liquidity. Finally, the effect of advertising on liquidity itself is discussed.
2.1. Brand value
2.1.1. Importance of brand value
The value of a brand is known as brand equity. Brand equity can be defined in various ways. According to Aaker (1991) brand equity can be defined as a well-known brand name that ensures the owner of the brand earning more money with selling its products under the well-known brand name than under a less well well-known name.
It can be defined in other ways, however, the value of the brand must be derived in the marketplace for consumers. They decide with their purchases which brands are more valuable than other brands (Keller, 2003). According to Keller there exists a differential effect that brand knowledge has on consumer response to marketing activities. Consumers respond more favorably to marketing activities when the consumer is familiar with the brand compared to when it is not. Strong brands have a memory encoding and storage advantage over less familiar brands in building brand awareness and image. Consumers develop a greater number of stronger links for familiar brands. Strong brands have better developed consumer knowledge structures. Therefore the chance that these links that make up this knowledge will be uniquely associated with the specific brand increases. This also works the other way around, a less strong brand has a less developed consumer knowledge structure. If a customer purchases a product of this brand, the associations of the purchase may end up being stored under the product category instead of under the brand (Keller, 2003).
Besides the fact that strong brands have better developed consumer knowledge structures, consumers have more confidence in the product when they get more familiar in a domain. This has to do with loss aversion, in case the consumer switches to a less known
6
brand he can incur losses. The possibility of a potential loss can be an advantage for strong brands. Consumer confidence may lead to greater use of favorable associations to help with decision making (Keller 2003).
Building strong brands has become popular for many firms because it yields a number of marketing advantages. If a customer faces a difficult situation, these brand advantages will be evident. Strong brands have better developed consumer knowledge structures. Also consumers have more confidence in products of a strong brand. These advantages facilitate decision making (Keller 2003).
2.1.2. Brand value and its relation to firm value
In the previous paragraph it is shown that a strong brand can have advantages. The theory explains why a strong brand, i.e. a high brand value, can be important. Investors are assumed to focus on financial measures and ignore less tangible non financial measures such as product quality, customer satisfaction and order lead time, whereas these less tangible non financial measures are the drivers of corporate success (Peters, 1981). Aaker and Jacobson (1994) show that the explanatory power of the product quality measure on the stock price can be compared to the explanatory power of ROI on the stock price. Therefore there can be an association between stock return and information contained in the quality measure.
Non financial measures as customer satisfaction and product quality is reflected in the brand value, therefore brand value can be a quality measure (Aaker and Jacobson, 1994). Bris et al. (2010) provide evidence that strong brands do not earn significantly higher stock returns. The brand values are provided by Interbrand. Interbrand publishes a yearly list of the top 100 world’s most valuable brands. They analyze the operating, financial, and market performance of firms included in Interbrand’s Global top 100 and compare them to a sample of non global brand firms by using a dummy variable if the company is listed in Interbrand’s top 100. The study is conducted in the period 2000-2008. They show that global brands do not display significant higher, risk-adjusted excess returns. Both in time-series analyses and cross-sectional regressions the alpha for global brands is not significantly different from zero. They argue that only certain firms in an industry implement the global brand strategy by selling a large array of products under the same brand. Other firms choose to have a portfolio of non-global brands. Being a non-global brand is an equilibrium strategy which is applied by some firms and not by other firms until market benefits of one of the strategies is equalized.
Barth et al. (1998) investigate the relation between brand value and firm value. The brand values are estimated by FinancialWorld. They used a sample of 1,204 brand values
7
relating to 1991 to 1996 fiscal years. They find that brand value estimates are significantly positively related to prices and returns while they adjust for book value and net income by conducting cross-sectional regressions. The share price at year end is regressed on the book value of equity per share, earnings per share and the brand value estimates per share. They control for calendar time-specific effects by including time dummy variables. They find that the coefficient of the brand value estimates per share is significantly positive. Therefore, findings suggest that brand value estimates are relevant and sufficiently reliable to be reflected in share prices. However, brand value estimates could also be determined using share prices, therefore they also tested for simultaneity bias by estimating a system of simultaneous equations that treats the brand value estimates and market value of equity as jointly determined endogenous variables. Barth et al. (1998) show that the inferences are not attributable to simultaneity bias. They also find that estimates of these brand values are already incorporated in the share price. Due to the fact that brand values likely are relevant to investors, these findings should reduce the concerns about the reliability of the estimated brand values.
Madden, Fehle and Fournier (2006) also provide empirical evidence for the branding-shareholder value creation link. The brand values they use are again provided by Interbrand. While Barth et al. (1998) investigated the explanatory power of brand values, Madden et al. (1998) investigate if the world’s most valuable brands outperform the market for the years 1994 until 2001. They compare a portfolio containing firms that own brands that at least are reported once on Interbrand’s list. This portfolio is called the world most valuable brands portfolio. They compare the market performance of this portfolio with two benchmark portfolios, one containing all firms in the CRSP database, i.e. full market portfolio, and one portfolio containing all firms in the CRSP database excluding the firms listed in the most valuable brands portfolio, i.e. reduced market portfolio. They make use of the Fama and French method to estimate alpha, i.e. the degree of outperformance, of this top 100 (Madden et al. 1998). Besides, the beta is calculated using this method. Beta measures the degree of the asset’s sensitivity to market risk (Sharpe, 1964). The full market portfolio has an alpha of zero and a beta of one by design, because this portfolio is the proxy for market in itself. Subsequently, they find an alpha of 0.57% per month, which is statistically significantly different from the alphas of the two benchmark portfolios. Nearby, they find a beta of 0.85, which also is statistically significantly different from the two benchmarks. Madden et al. (2006) performed regressions with both a normal portfolio of world’s most valuable brands as a weighted one. In both cases the alpha is significantly positive. This indicates that strong
8
brands deliver greater returns to stockholders than a relevant benchmark does. Besides their outperformance they deliver these greater returns with a lower beta, i.e. lower risk. Madden et al. (2006) use a different sample period than Bris et al. (2010). Also the latter examines the operating, financial, and market performance of firms included in Interbrand’s Global top 100 in contrast to Madden et al. (2006) who only examine the market performance using the Fama and French method.
Dutordoir, Verbeeten and De Beijer (2010) examine if brand value announcements have an impact on firm value. The big difference with the previous mentioned studies is that the study of Dutordoir et al. (2010) concentrates on the brand value announcements itself while the other studies examine the relation between brand value and firm value controlling for simultaneity and omitted variables. Thus Dutordoir et al. (2010) are conducting an event study to examine the effect of an announcement on the share price. Interbrand and Brand-Finance claim that their estimates are based on publicly-available information as well on proprietary information. Also Brand-Finance uses news of his announcements in their campaigns as can be seen in Figure 1. This figure shows how media responds to the announcement of the rise in brand value of TCS, i.e. Tata Consultancy Services (Brand-Finance, 2013). In their campaign they claim the following: “TCS’s stock price soared on the release day adding over $350 million to TCS’s market cap. Analysts attributed the uplift to the announcement of the uplift in brand value.” Another interesting news item comes from The Edge Financial Daily that reported CIMB rose 3% after it was ranked number fourth in Asian’s annual ranking of the most valuable banking brands globally measured by Brand Finance (The Edge Financial Daily, 2014).
If brand value announcements indeed contain news, and thus can be seen as proprietary information, it should have effect on the stock price as soon as this information becomes available according to semi-strong market efficiency theory (Fama, 1970). The other studies mentioned earlier examine the relation between brand value and firm value using a panel regression. This difference is important because Dutordoir et al. (2010) show the direct effect of the brand value announcements on firm value, while the other studies show the relationship between brand value and firm value or they show the performance of global brands compared to the market.
Dutordoir et al. (2010) conduct an event study over the period 2001-2008 with in total 329 brand value announcements belonging to 54 different U.S. firms. They calculate the announcement-date abnormal return as the difference between the announcement-date stock return and the return on the CRSP value-weighted market index on that date. They show that
9
there is evidence of significant abnormal stock returns on brand value announcement dates. The magnitude of the abnormal stock return increases in the brand value change. Also in this study, Interbrand’s global top 100 is used.
Dutordoir et al. (2010) come to the same conclusion as Barth et al. (1998) and Madden et al. (2006). All three studies find a relationship between brand value and share price. Bris et al. (2010) do not confirm this relationship. This is possible because they use different performance measures and a different sample period compared to Barth et al. (1998) and Madden et al. (2006). Dutordoir et al. (2010) use the same sample period but they only examine the effect of a brand value announcement and not the brand value itself. It could be the case that Bris et al. (2010) find similar results if they only look at the market performance instead of also taking into account operating and financial performance.
2.1.3. Characteristics of a global brand firm
Bris et al. (2010) state that firms with a global brand have a low beta and may therefore be safe havens in periods of high market volatility when diversification is most needed. They argue that the intangible asset that is created by the brand can serve as a cushion in periods of high volatility. This is due to the fact that strong brands create customer value. This customer value is created by more name awareness, higher customer satisfaction, higher quality satisfaction, loyalty and customer lifetime value. This is in line with the reasoning of Rego, Billett and Morgan (2009), who argue systematic risk is expected to be lower for global brand firms because of three primary reasons. Firstly, global brands enable rapid product or service identification and it reduces the consumer search costs. Therefore repeated purchasing behavior is encouraged. Secondly, due to the characteristic that global brands create higher quality satisfaction, products of global brands are associated with lower price sensitivity. The third reason that Rego et al. (2009) put forward is that firms with a global brand are better-known. This corporate reputation effect signals lower risk to shareholders and debt holders. Moreover Bris et al. (2010) show that firms with a global brand take on less debt compared to firms without a global brand. This is due to the fact that global brands base their performance on a highly valuable intangible asset. The arguments of Bris et al. (2010) and Rego et al. (2009) will also hold during periods of high volatility. Therefore it can be expected that firms with a global brand will also have a low beta in periods of high volatility.
Despite the fact that systematic risk is expected to be lower for global brand firms Bris et al. (2010) show that these firms do not earn significantly higher risk-adjusted stock returns. However, as mentioned in the previous section, Madden et al. (2006) show that global brands
10
create value for their shareholders by yielding returns that are greater than the relevant benchmark with less risk. Both papers emphasize that global brands indeed have a lower beta. According to Madden et al. (2006) it is advantageous for firms having a global brand because it yields greater returns compared to the relevant benchmark with less risk. This could be a reason to justify managers’ brand-building efforts.
As mentioned earlier beta measures the degree of the asset’s sensitivity to market risk. The higher the degree of the asset’s sensitivity, the higher the expected return that is determined by the capital asset pricing model (Sharpe, 1964). There is sufficient evidence that this risk and return trade-off does often not hold, which can been seen in Figure 2. The figure shows that low risk stocks outperformed high risk stocks over the last thirty years (Hong and Sraer, 2012). This outperformance is significant. Hong and Sraer (2012) provide a theory that explains this high risk and low return phenomenon. In their research they allow for macro disagreement and they prohibit some investors from short selling. Macro disagreement can best be described as the analyst’s or households’ forecast divergence about for instance industrial production growth and inflation. In practice, macro disagreement actually occurs (Hong and Sraer, 2012). Also short-sales constraints emerges in practice. Many investors, such as retail mutual funds are simply prohibited to sell short. Only a small part of investors do sell short, for instance hedge funds. Due to short-sales constraints not all information is included in the stock’s price. Namely, the stock price will reflect the valuations that optimists bind to it, but it will not reflect the valuations that pessimists attach to it. This is due to the fact that pessimists do not participate in this market, i.e. they cannot sell short (Miller, 1977). Hong and Sraer (2012) put these two assumptions into the CAPM framework. They find that high beta assets are overpriced compared to low beta assets when macro disagreement is high. If macro disagreement is high, forecasts and expectations among analysts and households about macro variables have a large deviation. The impact of this high macro disagreement on forecasts of the assets is greater for assets with a high beta than for assets with a low beta. Forecasts of the earnings of the high beta will naturally diverge more than these of the low beta. Thus, beta strengthens the impact of the macro disagreement. Due to short-sales constraints and the high sensitivity of high beta stocks, the latter will only hold by optimists because pessimists do not participate in the market. Therefore high beta stocks are overpriced compared to low beta stocks. It is expected that arbitrageurs correct this mispricing, however their risk aversion limits their short interest. This in turn leads to overpricing in equilibrium (Hong and Sraer, 2012). They show that if aggregate disagreement is high enough, the relationship between risk and return takes on an inverted U-shape. This means that at first,
11
returns are increasing in beta until a certain cut-off point and then returns are decreasing by beta due to the higher impact of macro disagreement. The greater the disagreement, the lower the cut-off level for beta for which assets experience binding short-sales constraints.
Hong and Sraer (2012) test their predictions using monthly time-series of disagreement about market earnings and economic uncertainty provided by the I/B/E/S database for the period 1970-2010. U.S. stock return data are collected from the CRSP database. They measure the disagreement by taking the standard deviation of analysts’ long-term earnings growth forecasts. The aggregate disagreement measure is a beta-weighted average of these forecasts to match with their own theory. They show that the risk-return relationship indeed has an inverted-U shape in case of high disagreement. This leads to the conclusion that high beta assets are over-priced compared to low beta ones in case macro disagreement is high.
As mentioned before, global brands have a low beta on average. This is an advantage for firms with a global brand because Hong and Sraer (2012) show that low beta assets outperform high beta assets when macro-disagreement is high. This behavioral biases create an advantage for global brands This is in line with the findings of Madden et al. (2006) that global brands create value for their shareholders by yielding returns that are greater than the relevant benchmark with less risk. Managers can partly justify their brand-building efforts, because increasing the brand value of a firm can possibly lead to higher risk adjusted returns. Until now, the relation between brand value changes and firm value was discussed. The next section focuses on ways to increase brand value. In order to increase brand value, the firm must gain familiarity of the public. Advertising can be a way to do this. Therefore the next section discusses why firms advertise and what the effect is of advertising.
2.2. Advertising
2.2.1. Reasons of advertising
Nowadays there is an increasing pressure on marketing managers to justify the impact of their brand-building efforts on total firm value (Rust et al., 2004). This study already mentioned in section 2.1.1. that building strong brands has become popular for firms because it yields a number of marketing advantages. Building strong brands can be achieved by advertising (Rust et al., 2004).This paragraph discusses the reasons of advertising.
One of the reasons why managers advertise, is to give a signal to the customers. Managers’ information about the firm its products is likely to be superior to the customers,
12
there is asymmetric information between the managers and customers. To convince the customers their product is of high quality they should give a credible sign. A sign is credible only if it is supported by actions that would be too costly to take if the sign is untrue (Berk and Demarzo, 2007). Advertising can be a signal of product quality. Firms advertise to inform potential customers about the existence, characteristics, and prices of the products they offer. (Milgrom and Roberts, 1986). However, a great amount of advertising expenditures does not have this informational content. Despite the fact that this type of advertising does not have this informational content, firms spend a lot of money on this kind of advertising. The idea behind this is the fact that a particular brand of a good could be a signal of high quality. High quality brands can afford these types of advertising. Rational, informed consumers will respond positively to this advertising. Milgrom and Roberts (1986) constructed a model to show that uninformative advertising for an experience good could be a signal for product quality. Besides advertising they also allow for the possibility that the price of the product itself might be a signal. They conclude that a high-quality firm has a higher marginal benefit from attracting customers through uninformative advertising and thus are willing to advertise more. Because of the high costs of this advertising, the signal is credible.
Milgrom and Roberts (1986) show that uninformative advertising could be a signal for product quality. Due to the high costs of this advertising customers can conclude that the brand of a good is of high quality. Therefore advertising increases brand value. Another reason can be found from an investor’s perspective. Namely, such advertising can be carried out to make investors aware of the firm. The awareness of the brand itself could lead to more trading behavior. Grullon et al. (2004) show there is a positive relation between advertising expenses and liquidity. They take the relative bid-ask spread as a proxy for liquidity. This study and its results is further discussed in section 2.3.3. Lou (2009) also shows that an increase in advertising expenses leads to an increase in individual investor buying. Grullon et al. (2004) argue that the larger the visibility of the stock, the more it is expected that the ownership of the firm will have a larger breadth. Chen, Hong and Stein (2001) subsequently show that stocks experiencing declines in breadth of ownership underperform stocks for which breadth has increased. Next to the studies of Milgrom and Roberts (1986) and Rust et al. (2004) that find that advertising increase brand value, the three papers mentioned above support the idea that an increase in advertising expenses increases the breadth of ownership, liquidity and the stock return in short run. These papers and these relationships are further discussed in section 2.2.2. and 2.3.
13
2.2.2. The effects of advertising
Rust et al. (2004) mention that firms have to justify their advertising expenditures. As mentioned in the previous paragraph, one reason for these expenditures is to give a credible signal. The credible signal that is given by advertising can be seen as an investment. Managers can justify their brand-building efforts if the investment leads to a positive NPV (Milgrom and Roberts, 1986). Rust et al. (2004) investigated how this credible signal translates into real productivity. They propose different methods to measure the marketing productivity, i.e. return of the investment. Marketing productivity can be assessed by modeling the market impact of marketing expenditures. Another powerful method is to assess the impact of marketing expenditures on the value of the firm (Rust et al., 2004). Besides, Barth et al. (1998) show that brand value estimates are significantly positively associated with advertising expenses, brand operating margin and brand market share. Brand value estimates reflect information that is useful for investors beyond that reflected in the alternative measures associated with brand value. Besides this relationship between advertising and brand value, advertising have other effects.
Aaker and Jacobson (1994) investigate whether including advertising expenditures in the regression can explain away the relation between quality and stock return. They show that the association of stock return with quality does not stem from a joint association with advertising. They emphasize that this doesn’t mean the market does not incorporate the advertising expenditures. Their results indicate that the market does incorporate these expenditures to the extent that they influence current-term ROI.
Besides this direct effect of advertising expenditures on performance measures, Osinga, Leeflang, Srinivasan and Wieringa (2011) indicate that there is a positive relation to shareholder value and direct-to-consumer advertising (DTCA), even for companies with limited sales response. They show that DTCA not only increases stock return but also decreases systematic risk. DTCA does increase idiosyncratic risk because investors perceive DTCA as an risky investment. But the increase of this risk shouldn’t affect a well-diversified investor because this risk can be diversified. A managerial implication is that a firm should seek for a balance in their advertising expenditures to optimize shareholder value (Osinga et al, 2011).
So managers can adjust firm advertising expenditures to increase brand value, the performance of a firm or it can act as an effective signaling device. Also, a firm can influence investor behavior and short-term stock prices by adjusting these expenditures. Lou (2009) investigates this effect of marketing. He argues that advertising can affect stock returns in two
14
different ways. The first reason is that, in general, it is hard to short sell securities. Therefore investors seek more to stocks which they want to buy, than seeking for stocks they want to sell. Therefore investors’ buying decisions are more focused on popular stocks than on their selling decisions. An increase in advertising expenditures can therefore potentially lead to more buy orders than sell orders, which in the short run leads to an increase in firm value. The second reason is that investors might link a firm’s performance to its product market advertising. Advertising is often exaggerated and does not reflect, in an unbiased manner, all features of the product and firm. Investors might think that the firm has unexpected growth potentials, without thinking about whether this information is already incorporated in the share price (Lou, 2009). Above mentioned reasons both predict that an increase in advertising expenditures leads to a rise in the share price in the short run. Lou (2009) provides empirical evidence that managers adjust firm advertising expenditures to influence investor behavior and short-term stock prices. The paper shows that increased advertising spending is associated with individual investor buying and a rise in abnormal stock returns. This rise in abnormal stock return is reversed in subsequent years. Besides, Lou (2009) finds that managers increase their advertising expenditures when the potential benefits are the largest. These potential benefits are, for instance, before insider sales or new equity issuances. Increasing advertising expenditures before insider sales or new equity issuances causes the share price to rise. Then insider sales and new equity issuances will yield a higher return (Lou, 2009).
As mentioned before Grullon et al. (2004) argue that an increase of the visibility of the stock will result in an increase of the breadth ownership of the firm. Grullon et al. (2004) take advertising expenditures as proxy for firm visibility. This means that an increase of the firm’s advertising expenditures will lead to investors that are more aware of the existence of the firm. Hong et al. (2001) show that this breadth of ownership can be a valuation indicator. Stocks that experience a decline in breadth of ownership underperform those for which breadth has increased. Advertising expenditures therefore can have an indirect effect on the stock return of the firm. It increases the breadth of ownership which in turn affects the return of the stock. Hong et al. (2001) see breadth of ownership as a proxy for how tightly short-sales constraints bind.
Hong et al. (2001) investigate the relationship between a change in breadth of ownership and stock return in response to the theory of Miller (1977). Miller (1977) states that the stock price will reflect the valuations that optimists bind to it, but it will not reflect the valuations that pessimists attach to it. This is due to the fact that pessimists sit out of this market, i.e. they cannot sell short. Therefore short-sales constraints can have a significant
15
impact on market prices and expected returns. Several studies investigated the relationship between short-sales constraints and stock returns. However, the majority of stocks have no short interest outstanding, since many mutual funds are prohibited to take short positions. Therefore a portfolio containing stocks with high short interest will be very small, which in turn can potentially lead to a decrease in the power of any tests. To circumvent this shortcoming, Hong et al. (2001) use breadth of ownership as a proxy for short-sales constraints. The idea behind it is that when breadth is low, relatively more investors do not participate in this market. The negative information that these investors hold about the stock are kept off the market. Hence, they expect that reductions in the breadth of ownership should forecast lower subsequent returns and vice versa.
Hong et al. (2001) test their expectations using quarterly data on mutual fund holdings
coming from the Mutual Fund Common Stock Holding/Transactions database over the period 1979-1998. They compared the stocks whose change in breadth in the prior quarter places them in the lowest decile with those in the top decile. They found an underperformance of the lowest decile of 3.82% in the first six months after portfolio formation, and of 6.38% in the first twelve months. Also after controlling for other variables such as size, book-to-market and momentum they still find a significant underperformance of the lowest decile of 2.92% and 4.95% for respectively six and twelve months. Their findings correspond with their expectations.
Grullon et al. (2004) their results indicate that firms spending more on advertising have a larger number of both individual and institutional investors. They emphasize that also institutional investors are affected by advertising which are in most cases the main investors. This supports the idea that advertising can lead to a wider breadth of ownership which in turn can lead to higher stock returns, whereof the latter is confirmed by Hong et al. (2001).
This paragraph addressed the effect of advertising expenditures. One of the main reasons to advertise is to give a credible signal to consumers. How this credible signal eventually works out can be tested by examining the relation between the advertising expenditures and brand or firm value. Barth et al. (1998) show that advertising expenditures are associated with brand value. Aaker and Jacobson (1994) find that advertising affects the ROI. Osinga et al. (2011) indicate that there is a positive relation to shareholder value and direct-to-consumer advertising.
Besides the influence on consumer behavior, the behavior of investors can be affected. Lou (2009) finds that increasing advertising expenditures lead to a rise of the stock price in the short run. Grullon et al. (2004) find that an increase in advertising expenditures can lead to
16
a wider breadth of ownership. The latter may lead to higher stock returns according to Hong et al. (2001). Overall, the papers concerning the effect on consumer behavior show that increasing advertising expenditures leads to a higher brand value and an increase on ROI. The papers concerning the influence on investor behavior show also show corresponding results. Lou (2009) shows that an increase in advertising expenditures leads to a positive rise of the stock price in the short run, while Hong et al. (2001) show that a positive change in breadth, which can be caused by an increase in advertising expenditures, also leads to higher returns. Besides these effects on returns, breadth of ownership and brand or firm value, advertising has an effect on the liquidity of the stock, which is discussed in the next section.
2.3. Liquidity
2.3.1. Properties and characteristics of liquidity
This paragraph discusses what liquidity encompasses. Besides that, the sources of liquidity are addressed. Liquidity is a summary quality or attribute of a security or asset market and can be important valuing equity. Liquidity has a lot of definitions. According to Harris (2003) a market or security is liquid when you can trade large amounts without moving the price very much.
Following Amihud et al. (2005) liquidity can be defined or proxied as the cost of trading an asset. Illiquidity is caused by multiple sources. One of the sources is exogenous transaction costs. These costs are for example brokerage fees, order processing costs, or transaction taxes. These costs have to be paid every time an investor wants to engage in a transaction. The second source of illiquidity is demand pressure and inventory risk. Demand pressure arises when an investor wants to sell a security immediately, but at that moment there is no buyer. The investor can sell the security to a market maker. This market maker is willing to buy this security in anticipation of being able to sell the security later on to any other investor. In the meanwhile, the market maker is exposed to price risk, i.e. the price of the security can change. This is called inventory risk. The market maker wants to be compensated for this risk in the form of bid-ask prices, which is discussed later on (Amihud et al, 2005). The third source of why trading an asset can be costly is private information. Either the potential seller or buyer can have private information about the fundamentals of the firm. Trading with an informed counterparty will end up with a loss. Besides private information about the fundamentals of a firm, an investor can also have private information about order flow. Sometimes a party knows that an investor has to liquidate a huge position in an security.
17
This party knows that this will decrease the price. The party can therefore (short) sell the security and buy back the same security later on. The last source of illiquidity is the difficulty of searching for an counterparty that is willing to buy or sell the asset. Besides, once this counterparty is found, the price needs to be negotiated. This occurs in a less competitive environment because of a lack of alternative trading partners. This last source of illiquidity particularly occurs in the over-the-counter market. In such a market there is no central marketplace. The cost of this source comes from searching the counterparty, but also from price concessions that need to be made to succeed in a trade. All above mentioned costs of illiquidity affect asset pricing (Amihud et al., 2005).
The theory of Amihud et al. (2005) corresponds with the liquidity-based asset pricing model set out by Holmstrom and Tirole (2001) that assume the economy is capital constrained. Therefore collateralizable assets are in short supply and thus will command a premium, which is discussed in the next paragraph.
It is widely accepted that liquidity affects investors’ portfolio decisions. Investors care about expected returns net of trading costs. Less liquid assets need to provide a higher gross return compared to more liquid assets (Datar, Naik and Radcliffe, 1998). Therefore it is obvious that a liquid asset is advantageous for an investor.
This paragraph discussed liquidity, its attributes and its relation to returns. Liquidity cost can be expressed as the cost of trading an asset (Harris, 2003). Illiquidity is caused by multiple sources: exogenous transaction costs, demand pressure and inventory risk, private information and at last the difficulty of locating and pricing with a counterparty (Amihud et al., 2005). Because of these costs, investors ask higher gross returns for less liquid assets (Datar et al., 1998). The next paragraph addresses how this demand of investors translates into real returns and it discusses the impact of liquidity on asset prices.
2.3.2. Liquidity and asset prices
This paragraph provides empirical evidence to show the relation between liquidity and asset prices. In the previous paragraph it is explained that due to the costs of illiquidity, investors ask for higher gross returns for less liquid assets. This paragraph covers this relation using empirical evidence.
Amihud and Mendelson (1986) argue that liquidity is among the primary attributes of many investment plans and financial instruments. Portfolio managers adjust their portfolios to fit liquidity objectives and the horizon of the investment of their clients. They examine the effects of liquidity on asset pricing. If an investor wants to engage in a transaction he can
18
choose to either wait for a favorable price or to execute the order at the current bid or ask price. This leads to a either a buying premium or a selling concession. Garbade (1982) show that the relative bid-ask spread is negatively correlated with other liquidity characteristics. Examples of other liquidity characteristics are trading volume, number of shareholder, number of market makers that are trading the stock. The larger these characteristics are, the higher the liquidity. With the relative bid-ask spread it is the other way around. The higher the relative bid-ask spread, the higher the cost of trading a security and thus the more illiquid this security is.
Amihud and Mendelson (1986) expect that the return of an asset is increasing in the relative bid-ask spread. This is in line with what Datar et al. (1998) argue. Both papers indicate this is due to the fact that investors care about expected returns net of trading costs. Less liquid assets, i.e. a high relative bid-ask spread, need to provide a higher gross return compared to more liquid assets. Amihud and Mendelson (1986) also expect that investors with longer holding periods select assets with higher spreads. The cost of the illiquidity will relatively become less because of the longer holding period, therefore an investor that expects to hold the asset for a long period can gain by holding high-spread assets. Hence, they expect that expected returns net of trading costs increase with the holding period. To test their expectations they use data collected for NYSE stocks over the period 1961-1980. They find that average portfolio risk-adjusted returns increase with their relative bid-ask spread. The slope of this spread-return relationship decreases with the spread. This spread effect persists also including firm size as an explanatory variable. They emphasize that these findings are not an indication of market inefficiency. It is rather a response of the market to the existence of this spread.
This paragraph addressed the relation between liquidity and returns. Amihud and Mendelson (1986) prove there is a relation between liquidity and returns by showing that average portfolio risk-adjusted returns increase with their relative bid-ask spread. These findings do not indicate markets are inefficient.
2.3.3. Liquidity and its determinants
This paragraph addresses the determinants of liquidity. The degree of liquidity is related to different variables. This paragraph appoints these variables and explains why these variables have an impact on the liquidity. Also, empirical evidence is addressed to examine the relationship between these variables and the liquidity.
19
One of the major sources of illiquidity is risk. As explained by Amihud et al. (2005) and Holmstrom and Tirole (2001) demand pressure arises when an investor wants to sell a stock immediately, but at that time there is no purchaser. The investor can sell its shares to the market maker. The market maker buys the shares in anticipation of being able to sell the shares to another investor at a later moment. In the meantime, the market maker is exposed to price risk. The price of the stock can change, this is called inventory risk. The market maker wants to be compensated for this risk in the form of the bid-ask spread. The higher the volatility of these price changes, the higher the inventory risk. Therefore return volatility is a good proxy for risk. Benston and Hagerman (1974) explain that investors are fully compensated for bearing the systematic risk associated with it according to CAPM, therefore systematic risk should not be in the spread. Furthermore, unsystematic risk can be eliminated, however, market makers are not able to hold a perfectly diversified portfolio because the markets they make are limited. Grullon et al. (2004) confirm the relationship: return volatility is positively correlated with the relative bid-ask spread. Another control variable that is a proxy for risk is firm age. Older companies increase their corporate reputation. As Rego et al. (2009) argue the corporate reputation effect signals lower risk to shareholders and debt holders. Therefore, an increase in firm age results in a better liquidity, i.e. a lower relative bid-ask spread. Grullon et al. (2004) confirm this by showing a significant negative relation between firm age and liquidity.
Benston and Hagerman (1974) argue that the number of market makers and the number of shareholders are likely to be correlated with each other. Large companies have more shareholders and also more market makers that are interested in making a market. Due to the competition between market makers, the earnings, i.e. the spread, will be lower for such firms. These variables are correlated with firm size. Grullon et al. (2004) add to this that larger firms are more likely to have more media attention and analyst coverage. Also larger firms have more shares available to buy. Grullon et al. (2004) show that a negative relationship between the relative bid-ask spread and the market value of the firm exists.
The latter study also shows a negative relationship between share turnover and the relative bid-ask spread, i.e. a proxy for liquidity. They argue that stocks with a high share turnover, called liquid stocks, may be preferred by a larger group of investors. Benston and Hagerman (1974) further submit that as the volume increases so does the number of limit orders which facilitate immediate exchange. Seen from the side of the market makers this means that they relatively need less inventory per transaction. Theory suggests that the inventory of a dealer is less than proportional of the number of transactions. Therefore the
20
spread declines as the share turnover of the stock increases. Subsequently, share turnover is an important determinant of liquidity.
Benston and Hagerman (1974) explain that market participants relate transaction costs to the dollar amount traded. They argue that in the case only the price differs, i.e. ceteris paribus, traders would use limit orders to equalize the spread per dollar regardless of the price per share traded. Therefore spreads would be proportional to the share price. Benston and Hagerman (1974) add that this strict proportionality might not hold in the case of disproportionate broker costs. In this case arbitrage for low-priced stocks does not equalize the spread. Stocks with a higher share price might have better liquidity for above reasoning. Grullon et al. (2004) share this view and see share price as a possible determinant of liquidity.
Grullon et al. (2004) also show a negative relationship between the return on assets and the relative liquidity spread. However, this relationship is statistically not significant. The higher the return on the assets, the more likely it is that investors are attracted to these firms due to good performance (Grullon et al., 2004). Moreover, companies performing well will gain more press coverage. This increases the visibility of the firm. Therefore return on assets is likely to be a determinant of liquidity.
Another determinant of liquidity can be the exchange on which the stock is traded. Some exchanges are preferred by investors and market makers. As mentioned before, more competition between market makers will lower the spread. The study of Grullon et al. (2004) confirm this relationship. They added dummy variables indicating the firm’s primary exchange. The coefficients of these dummies are significant.
This study examines the relationship between advertising and liquidity. Therefore, other determinants of liquidity need to be known. This paragraph addressed these determinants of liquidity and the underlying theory. Empirical evidence was brought forward to support this theory. In the next section the impact of advertising expenses on liquidity is addressed.
2.3.4. Effect of advertising on liquidity
Section 2.2.2. elaborated that increased advertising spending is associated with individual investor trading and a rise in abnormal stock returns. In this paragraph additional arguments are given why this is the case. Moreover, empirical evidence is brought forward to show the impact of advertising on liquidity.
An investor can buy a firm’s stock, partly because of the familiarity of the stock. As mentioned before, Grullon et al. (2004) argue that an increase of the visibility of the stock will
21
result in an increase of the breadth of ownership of the firm. They take advertising expenditures as proxy for firm visibility. This means that an increase of the firm’s advertising expenditures will lead to investors that are more aware of the existence of the firm. Such investors base their investment decision on public information and thus can be seen as uninformed investors. If a greater visibility leads to a relative increase in uninformed investors, then an improvement in familiarity of a firm can have an positive effect on the stock’s liquidity (Grullon et al., 2004).
Grullon et al. (2004) investigate if firm visibility is related to ownership breadth and stock market liquidity. Grullon et al. (2004) test their expectations using yearly data on firms that have data available on advertising expenditures. Their study consists of 5,776 observations over the period 1993-1998. They use the relative bid-ask spread as proxy for stock liquidity. Eventually they regress the relative bid-ask spread on the variable of interest: advertising expenses. They include the following control variables: firm age, return on assets, market value, share price, share turnover, return volatility and two dummy variables that indicate the firm’s primary exchange. Their results show a positive relation between advertising expenditures and stock market liquidity. They find that firms that have greater advertising expenditures, have common stock with lower bid-ask spreads, smaller price impacts and greater depth. Their study does not address the issue if an improvement in the firm’s visibility also leads to an increase in firm value. Their results are in line with the research of Lou (2009) that showed that increasing advertising expenses leads to an increase in individual investor buying.
This paragraph discussed the relation between advertising expenditures and liquidity. Advertising leads to a greater visibility of the firm and a better liquidity. In the next chapter the hypotheses and methodology is discussed.
22
3. Hypotheses & Methodology
This chapter formulates the hypotheses based on the literature review. It also describes the methodology to test these hypotheses. First the hypotheses that are tested in the next chapter to give an answer to the main questions of this study are set out. In the second part the methods of testing the hypotheses are described.
3.1. Hypotheses
The impact of brand value announcements and its relation to beta
As mentioned in the literature review, Dutordoir et al. (2010) examined if brand value announcements of Interbrand for U.S. companies over the period 2001-2008 have an direct impact on firm value. This study examines the impact of brand value announcements on firm value and relates this to the beta of the firms listed in the top 100.
Besides that this study uses Interbrand’s brand value announcements, it also uses the brand value announcements of brand consultancy firm Brand-Finance. To conduct a broad research, this study uses the whole top 100 instead of using only U.S. firms for the years 2007-2013. The development of the hypothesis below is based on the theory explained in the literature review.
As explained in section 2.1.3. global brands have a low beta and can therefore be seen as safe havens in periods of high market volatility when diversification is most needed (Bris et al.,2010; Madden et al., 2006). Hong and Sraer (2012) show that lower beta stocks outperform high beta stocks due to short sales constraints. High beta stocks are overpriced compared to low beta stocks, especially when macro disagreement is high. Becoming a global brand or increasing the value of the global brand might lead to a lower beta. This can be a reason to justify managers’ brand-building efforts. Increasing the brand value might lead to a lower beta. According to Hong and Sraer (2012) this lower beta leads to less overpricing. Before testing the impact of brand value announcements on firm value, this study tests if global brands actually have a lower beta compared to the market. Therefore, the first hypothesis of this study is formulated as follows:
23
To test this hypothesis, this study calculates the beta of the firms whose brand occurs in the top 100 of the annual ranking of Brand-Finance and Interbrand for the years 2007-2013 and compare this with the beta of the market, which is equal to one by design. This hypothesis is tested using the simple comparison of means method. Subsequently, this can be an argument that changes in brand value matter for investors, because a lower beta leads to less overpricing (Hong and Sraer, 2012). Testing this hypothesis gives a better understanding of why brand value changes could have an impact on firm value. The latter is examined by testing the third hypothesis.
According to the literature review, it can be argued that a higher ranking is related to a lower beta. If global brands actually have a lower beta on average, it is expected that brands with the highest ranking also have a lower beta compared to the brands with the lowest ranking. This results in the second hypothesis:
H2: Stocks of firms listed in the top 25 of the global top 100 have a lower beta than those that are listed in the bottom 25.
This hypothesis is also tested using the simple comparison of means method. The mean of the beta of brands listed in the top 25 is compared with the beta of the brands listed in the bottom 25. Testing the second hypothesis can give insights about the impact of brand value changes on firm value. If firms with a higher ranking in the list have a lower beta, then this might be advantageous because low beta outperforms high beta when macro-disagreement is high. This behavioral biases create an advantage for global brands with a high ranking. This can eventually be a reason why changes in brand value are related to firm value. Barth et al. (1998) and Madden et al. (2006) test this relationship and indicate there is a positive relation between brand value and firm value. This results in the third hypothesis:
H3: Externally-provided brand value announcements have a direct significant impact on firm value at the announcement date.
To test this hypothesis this study analyses the abnormal returns of the firms for which a brand value will occur in the top 100 of the annual ranking of Brand-Finance and Interbrand at the announcement date. Besides examining the direct impact of the brand value announcements, this study also inspects the impact on firm value using different event windows. Subsequently, this study investigates if the magnitude of the abnormal returns increases in the brand value
24
change. This is elaborated in the methodology. If abnormal returns are significant it can be concluded that brand value announcements have a significant impact on firm value. In that case Dutordoir et al. (2010) their findings are supported by this study. Consequently, it can be said that brand value announcements indeed contain proprietary information of which the market deems the provided brand value information is reliable. If the third hypothesis is accepted, testing the first two hypotheses can possibly explain why brand value changes are related to firm value.
Another way to examine the effect of a brand value announcement can be performed by testing if the magnitude of the abnormal returns increases in the change in ranking of the brand. In this manner the relative performance of the brands is tested. Namely, in times of prosperity it can be the case that the value of the majority of the brands in the top 100 increase in value, while the ranking does not change. The distinction between excellent performance and good performance becomes less clear. This issue can partly solved by adding year fixed effects in the first regression. This is taken into account in the next section. However, to examine the effect of brand value announcements in another way, this study also investigates the effect of the change in ranking on the stock price of the brand. Besides examining the impact on firm value of brand value changes, this study also takes into account changes in ranking, as an alternative way of looking at the effect of brand value announcements. This is carried out, because stockholders can react to information other than information that is directly related to future cash flows like brand value. Guberman and Regev (2001) suggest that stockholders can also react to information such as rankings and indices. Therefore the following hypothesis is also tested:
H4: The magnitude of abnormal returns increases in the change in ranking of the brand.
This hypothesis is tested by analyzing the abnormal returns in the same way as the third hypothesis. Only in this part, the abnormal returns are related to the changes in ranking. Because it is likely that a rise in brand value goes along with a positive change in ranking, it is expected that a higher rating leads to a higher firm value.
25
The effect of advertising expenses and brand value changes on liquidity
The remaining part of this study focuses on the effect of advertising expenses and brand value changes on the liquidity of the stock. This study refers to this remaining part as the second part of the study. This section formulates the hypotheses regarding this part of the study.
Barth et al. (1998) show that higher advertising expenses are associated with a higher brand value. Also, Lou (2009) finds that increasing advertising expenditures lead to a rise of the stock price in the short run due to a rise of investor buying behavior. Next to Lou (2009), Grullon et al. (2004) argue that an investor can buy a firm’s stock, partly because of the familiarity of the stock. An increase of the visibility of the stock leads to a relative increase in uninformed investors which can have a positive effect on the stock’s liquidity. In that case, increasing advertising expenses should have an positive effect on the liquidity of the stock. This expectation is formulated by the following hypothesis:
H5: Advertising expenditures are positively associated with the liquidity of the stock.
To test this hypothesis this study regresses the relative bid-ask spread on the variable of interest: advertising expenditures. Following Amihud et al. (2005) liquidity can be defined or proxied as the cost of trading an asset. One of the most important sources of illiquidity is caused by inventory risk. As mentioned in the literature review, the market maker wants to be compensated for this risk in the form of bid-ask prices. The higher the bid-ask spread, the higher the cost of trading an asset and therefore the worse the liquidity is. Therefore, the bid-ask spread is a proper proxy for liquidity. In order to make spreads comparable, the spread is divided by the closing price, which leads to the relative bid-ask spread. This study uses the determinants mentioned in the literature review as control variables.
If a brand increases in value, it seems likely that the brand becomes more appealing for investors. In the same way advertising attracts a disproportionate number of investors who make their investment decisions based on familiarity rather than on more fundamental information (Grullon et al., 2004). In order to find additional support for the theory explained above, this study also investigates if a change in brand value affects the liquidity of the stock. Due to the positive association between advertising expenses and brand value, it is expected that the change in brand value is negatively related to the relative bid-ask spread. Note that the
26
lower the relative bid-ask spread is, the better the liquidity of the stock is. This expectation is formulated by the hypothesis below:
H6:The change in brand value is positively related to the liquidity of the stock.
.
Finally, to further examine the effect of the brand value on liquidity, this study investigates if firms that are listed in the top 10 of the global top 100 have a better liquidity. This leads to the following final hypothesis:
H7: Stocks of firms that are listed in the top 10 of the global top 100 have a better liquidity.
This hypothesis is supported by the same argument as the seventh hypothesis. Again, firms that have popular brands can attract investors that decide investing based on familiarity (Grullon et al., 2004). Firms that are in the top 10 of global top 100 might spend more money on advertising and are more popular for investors. Thus it is expected that these firms will have a better liquidity.
The first two hypotheses are tested using the comparison of means method. The third and fourth hypothesis are tested by conducting an event study whereby abnormal returns are calculated. These abnormal returns are related to brand value changes and changes in ranking. The last two hypotheses are tested using a regression-based analyses. The next section describes these methods in more depth.
3.2. Methodology
This study measures the impact of a brand value announcement on the firm value of the brand owner. Furthermore, the beta of firms listed in the top 100 is examined in detail. Moreover, this research examines the effect of advertising and brand value changes on liquidity. In this section the methodology, that examines the hypotheses formulated in the previous section, is discussed.
The impact of brand value announcements and its relation to beta
Before examining the impact of brand value announcements on firm value, the beta of firms listed in the top 100 is investigated. To test the first hypothesis, the average beta of firms
27
listed in the top 100 is compared to the beta of the market using the simple comparison of means method. The comparison can be formulated as follows:
H0: μ1 - μ2 > 0 and H1: μ1 - μ2 =< 0
Where μ1 is the mean beta of the market, which is equal to one by design, and μ2 is equal to
the average beta of all firms listed in the top 100 of Brand-Finance and Interbrand. As explained in the previous section, it is expected that the average beta of the global brands is lower than the beta of the market, i.e. one tailed hypothesis. Subsequently a t-statistic is calculated to test whether the result is significant:
𝑡𝑡 = (𝑥𝑥̅1− 𝑥𝑥̅2) − (𝜇𝜇1− 𝜇𝜇2) �𝑠𝑠12
𝑛𝑛1+ 𝑠𝑠
22
𝑛𝑛2
To test the second hypothesis, the mean of the beta of brands listed in the top 25 is compared with the beta of the brands listed in the bottom 25. This is done in the same way as for the
first hypothesis, where this time μ1 and μ2 is equal to the average beta of firms listed at
position 75-100 and position 1-25 of the global top 100, respectively.
To give an answer to the third and fourth hypotheses, an event study is conducted. This study uses the same methodology as Dutordoir et al. (2010). To examine the effect of a brand value announcement on firm value this study uses estimated brand values from Interbrand’s Best Global Brands lists and Brand-Finance’s Global top 100 over the period 2007, the year of Brand-Finance’s first publication of the Global top 100, to 2013. First, for each year, the change in brand value is calculated. The reports of both brand consultancy
firms provide an estimation of the brand value at 31st December of the year before the report
is released. Each year the reports of both brand consultancy firms are released on a different date. The aim of this study is to examine the effect on the stock prices of the firms that are listed in the global top 100 on the release date, which is called the event date. The identification of this event date is a critical step in an event study. Therefore the exact dates on which Interbrand and Brand-Finance release their findings to the public are needed. As the event date is determined, the abnormal returns are calculated to appraise the event’s impact. In case the market is closed on the release date, the first trading day is used as event date. For
