The Effect of Customer Satisfaction on Stock Returns

(1)

Stock Returns

Can Abnormal Returns be Achieved by Investing in Firms with High

Customer Satisfaction?

Yory Wollerich

July 2010

UNIVERSITY OF GRONINGEN

Faculty of Economics and Business

MSc Finance

(2)

University of Groningen ii Master Thesis Finance

Customer Satisfaction and Stock Returns

Can Abnormal returns be achieved by investing in firms with high customer satisfaction?

ABSTRACT

In this paper, the relation between customer satisfaction and stock returns is examined. Several stock portfolios are composed based on customer satisfaction (ACSI) scores. With the use of the four factor model proposed by Carhart (1997) the abnormal returns of the different portfolios are examined. Conclusions that can be drawn based on the analysis is that abnormal returns can be achieved by investing in firms that excel in customer satisfaction. Furthermore, the systematic market risk from a high customer satisfaction portfolio is considerably lower compared to the other portfolios. Moreover, with the use of a GARCH-in-mean model it is examined whether customer satisfaction can be classified as a risk factor. The relation between the conditional variance and the conditional mean is not confirmed and consequently customer satisfaction cannot be interpreted as a risk factor.

JOURNAL OF ECONOMIC LITERATURE CODES C22, C31, G12, M31

KEYNOTES

Customer Satisfaction, Asset Pricing, Accountability, Abnormal Returns, Risk Factors, Portfolio Analysis,Factor models, GARCH-in-Mean models

CHARACTERISTICS AUTHOR: Date: 13st of July

Author : Yory Wollerich

Adress: Onstahof 10, 9403 PN, Assen Mobile number : 0031- (0)618037634

(3)

University of Groningen iii Master Thesis Finance

PREFACE

During my master thesis in Marketing and Marketing Research, about the effect of advertising on sales, I became interested in the marketing-finance interface. For this reason I decided to add a master in Finance to my Curriculum. After an exciting semester with courses like portfolio management, corporate valuation and behavioral finance, my actual goal was not yet reached. Reasonably but regrettably, the influence of marketing and the marketing-finance relations were completely ignored during the courses. However, with special regards to Jaap Wieringa, my master thesis topic really enabled me to further specialize in this interesting field.

During the process of writing my thesis, several persons were very helpful. First of all, I would like to thank my thesis supervisors Jaap Wieringa and Auke Plantinga. Both for providing me the needed advices and for triggering my interests during their courses. Mainly due to the courses Marketing Model Building and Advanced Marketing Research, statistics became a good friend of mine. Moreover, the accessibility and applicability of the Portfolio Management Course positively modified my prejudices about stock and bond trading. In addition, I would like to thank my girlfriend, parents and brother for their moral support.

(4)

iv

TABLE OF CONTENT

1. Introduction ...1

2. Literature ...3

3. Research methodology ...8

4. Data and descriptive statistics ... 12

5. Results ... 15

6. Conclusion and discussion... 19

7. References ... 22

Appendix 1: National ACSI scores per quarter ... 27

Appendix 2: Results three factor scores percentiles ... 28

(5)

University of Groningen 1 Master Thesis Finance 1. INTRODUCTION

Webster, Malter and Ganesan (2005) conclude, based on several in-depth interviews with marketing executives and chief executive officers, a marked fall-off in the influence, stature and significance of the corporate marketing department. Even more alarming evidence comes from Nath and Mahajan’s (2008) study, which analyzes a multi-industry sample of 167 firms and concludes that the CMO´s presence in top management teams has almost no impact on firm performance. In addition, a study of the FTSE 100 index firms in the United Kingdom conducted by Simms (2001) reveals that only thirteen chief-executive officers (CEO) have a marketing background compared with twenty-six executives with a financial background. As a consequence, marketers are more often confronted with CEOs with a financial background. However, Dougherty (1992) states that the interaction between marketing and finance can best be characterized as “two different worlds”. This is confirmed by De Ruyter and Wetzels (2000) who describe the main issue that constrains the marketing-finance relationship: both departments are focusing on different stakeholders. The finance department is mainly concerned with shareholders and money providing institutions while marketing’s focus is on customers, suppliers and competitors.

Marketing’s inability to measure the effects of its actions is the most important cause for the decline of marketing’s influence (Verhoef and Leeflang, 2008). Kumar and Shah (2009) mention, (based on the CMO Council Report 2007) that CMOs do not demonstrate adequate returns on investments and consequently fail to show the true potential of marketing. The importance of

accountability has been confirmed by Moorman and Rust (1999) who show a positive relation

(mediated by market orientation) between accountability and marketing’s influence within the firm. In order to recapture top management’s respect and enhance marketing’s influence in the firm, the financial accountability of marketing strategies becomes extremely important (Doyle 2000).

(6)

University of Groningen 2 Master Thesis Finance

delayed (Srinivasan and Hanssens, 2009). Moreover, Fornell (2007) states, based on the considerable difference between a firm’s book and market value, that most economic value creation today does not get recorded on the balance sheet.

One of the most important “intermediate” performance metrics is customer satisfaction. Customer satisfaction is viewed as a measure of the size, loyalty and the quality of the customer base of a firm (Fornell et al 2006, Morgan and Rego 2006). Fornell (2007) states in his book “The Satisfied Customer” that the health of customer relationships can be considered the major indicator for future firm performance. The importance of customer satisfaction is underlined by Peppers and Rogers (2005) who state that customers these days are the ultimate scarce resource. On the other hand, increased nervousness about future earnings makes organizations reluctant to make long-term investments in marketing (customer relations and brand equity), switching instead to tactics from which they know that they will drive short-term sales (Grande 2006). As a consequence, it is important to show the financial relevance of customer satisfaction to shareholders.

Therefore, Claes Fornell, founder of the American Customer Satisfaction Index (ACSI) and the Claes Fornell Institution (CFI) group, started to relate his satisfaction index to the stock market (Fornell et al., 2006). Based on a portfolio of above industry-average customer satisfaction firms, Fornell et al. (2006) claimed to continuously outperform the market for 8 years. Moreover, while stock markets around the world were plunging during recent years, the year-to-date return on a portfolio of high customer satisfaction companies hovers around zero. Companies with high-customer satisfaction scores have beaten the S&P 500 market index continuously, they produced higher stock returns and their stock values and cash flows have been less volatile (Hart 2007). Moreover, Hart (2007) and Fornell et al (2006) question how this might be possible, given efficient market theory, which states that it is not possible to consistently outperform the market without being exposed to additional risk (Grinblatt and Titman, 2002, p76).

(7)

The rest of this thesis is structured as follows. The second section provides an overview of the literature, the third section elucidate the research methodology. In section four, the data and the descriptive statistics will be provided. Section 5 presents the results and section 6 the discussion and conclusions.

2. LITERATURE REVIEW

According to Fornell (et al. 2006, 2007) both neoclassical economics and marketing theory provide a general case for a positive association between customer satisfaction and stock prices. Fornell (2007) summarizes that real economic growth depends on the productivity of economic resources and the quality of the output that those resources generate. During the first half of the twentieth century, the economy was industrial and manufacturing oriented. As a consequence of the limited customer choice, mainly due to geographical boundaries, mass production was the key to growth. Currently however, consumer utility rather than productivity can be considered as the real standard for economic growth (Fornell 2007). Moreover, in his search for a quantitative measure of utility Hicks (1939) considered utility and satisfaction as equivalent.

(8)

Fornell (2007) states that a dissatisfied buyer will not remain a customer unless there is nowhere else to go or it is too expensive (switching threshold) to get there.

From a marketing perspective, the link between customer satisfaction and subsequent customer behavior is widely accepted and credibly established (Anderson and Mansi, 2009). Customer satisfaction increases behavior with positive economic consequences while reducing behavior with negative economic consequences. Bolton, Lemon and Verhoef (2004) discuss the effect of customer satisfaction on three key customer related behaviors that reflect the length, depth, and breadth of the customer-service organization relationship. First, the length or duration of a relationship corresponds to customer retention (or defection) and is defined as the probability that a customer continues the relationship with the organization (Kumar and Reinartz, 2006). Bolton (1998) indicates that customer satisfaction ratings are positively related to the duration of the relationship. Moreover, Szymanski and Henard (2001) conclude based on a meta analysis of 50 empirical studies that satisfaction affects the likelihood that consumers will buy the product again. Second, the depth of the relation is reflected in the frequency of service usage over time, or the decisions to upgrade and purchase premium products (Bolton, Lemon and Verhoef, 2004). Bolton and Lemon (1999) find causal evidence that high customer satisfaction leads to high usage levels in future periods. Third, the breadth of a relationship is reflected in cross-buying from a company over time. If a customer is satisfied with a company’s product or service, customers are more willing to purchase additional services (Anderson, Fornell and Rust, 1997).

Another very important positive consequence of customer satisfaction is word-of-mouth. It is widely accepted that satisfied customers will engage in word-of-mouth favorable to the firm (Reicheld and Sasser, 1990). Moreover, Schlesinger and Heskett (1991), Lam et al. (2004) and Szymanski and Henard (2001) found a significant positive relation between customer satisfaction and word-of-mouth. However, Anderson (1998) concluded that dissatisfied customers do engage in greater (negative) word-of-mouth than satisfied customers engage in (positive) word-of-mouth. Word-of-mouth, subsequently leads to the acquisition of new customers (Trusov, Bucklin and Pauwels, 2009), and according to Reichheld (2006) to long-term business performance. The final revenue enhancing effect of customer satisfaction is related to pricing. Highly satisfied customers are willing to pay premium prices (Homburg, Koschate, and Hoyer, 2005) and are less price elastic (Fornell et al, 1996)

(9)

efficiency of future advertising and promotion investments. Luo and Homburg (2007) argue that this phenomenon is caused by the possibility that customer satisfaction generates free word-of-mouth advertising and saves subsequent marketing costs. Moreover, Steenkamp, Nijs, Hanssens and Dekimpe (2005) mention that word-of-mouth may also serve as a countervailing strategy rather than the traditional policy of firms to confront advertising attacks from competitors. The next cost-effective outcome of customer satisfaction according to Luo and Homburg (2007) is employee related. Companies with high customer satisfaction scores face less difficulties with attracting and retaining high-qualified employees and managers and employees who are confronted with highly satisfied customers will develop a higher level of future job satisfaction (Ping, 1993). A rational explanation for the customer/employee satisfaction association, and another costs related outcome, is the number of complaints. Fornell et al. (1996) conclude that customer satisfaction leads to fewer complaints and consequently fewer costs regarding handling and managing complaints. Pugh (2001) mentions that a person who expresses positive and negative emotions can produce a corresponding change in the observer’s emotional state (employee satisfaction). Moreover, Anderson, Fornell and Lehman (1994) state that a firm that consistently provides high customer satisfaction should have fewer resources devoted to handling returns and reworking defective items.

Finally, some marketing scientists relate customer satisfaction to a overall business performance measure. For example, Anderson, Fornell and Mazvancheryl (2004) found a strong relationship between customer satisfaction and Tobin’s Q. Reichheld and Teal (1996) state that customer satisfaction is positively related to long-term growth and Naumann and Hoisington (2001) report positive associations among customer satisfaction, market share and productivity measures. Regarding the effect of customer satisfaction on shareholder value, Srivastava, Shervani and Fahey (1998) provided the first conceptual framework. According to Srivastava et al (1998) shareholders value consists of the present value of cash flows during the value growth period and the long-term residual value. This is in line with common financial theory, indicating that stock prices should reflect all expected future discounted cash flows to the investor.

(10)

faster market penetration (Anderson, Fornell, and Mazvancheryl (2004), speed of buyer response to marketing efforts (Fornell et al. 2006), and positive word of mouth and recommendations (Reicheld, 2006). Third, the reduction in risk is accounted for by the increased customer retention (Anderson, Fornell and Mazvancheryl, 2004) and the diminished influence of competitors’ efforts and external environmental shocks on business performance (Gruca and Rego, 2005). All together, customer satisfaction reduces volatility by stabilizing corporate revenues (Harrison-Walker and Perdue, 2007). Mainly because the cash flows from satisfied and loyal customers are less susceptible to competitive activities (Srivastava et al. 1998). Moreover, a stable customer base reduces the risk associated with future cash flows (Anderson et al, 2004) and creates shareholder value. Besides, new customers are attracted due to the free word-of-mouth and consequently the long-term growth rate will increase. Increases in the long-term growth rate in earnings impacts the firm’s valuation and consequently stock prices.

The first empirical evidence regarding the customer satisfaction and stock price association is provided by Fornell et al. (2006). In order to confirm his expectations Fornell et al. (2006) performed three different analyses. The first analysis regressed the market value of equity as a function of the book value of the assets, the book value of the liabilities and the customer satisfaction index score. Fornell et al. (2006) found the customer satisfaction coefficient to be strongly significant and reveals a satisfaction-elasticity of 4.6%. As a consequence, Fornell et al. (2006) states that customer satisfaction truly is an economic asset left of the balance sheet and not fully reflected in the recorded assets.

(11)

The third study concerned a portfolio study, in which Fornell et al (2006) formed a stock portfolio based on ACSI information. Fornell et al. (2006) included those firms who were in the top 20% of ACSI (relative to competition). The portfolio was rebalanced every quarter when new information about a firm’s satisfaction scores was made available. The customer satisfaction portfolio (top 20%) generated a cumulative return (between 1997 and 2003) of 40% (dividends and transaction costs excluded) compared with 21% for the Dow Jones Industrial Average, 13% for the S&P500 and 9% for the NASDAQ. eMoreover, the abnormal returns are not due to a compensation for risk. The beta associated with the portfolio is namely 0.78 and thus substantially less than the market. Also the book-to-market ratio cannot explain the abnormal returns because the calculated B/V was 0.41 for the portfolio and 0.42 for the stocks not in the portfolio. Nonetheless, these results are based on a paper portfolio that is back tested which may exhibit several biases. Therefore, Fornell et al (2006) provided additional evidence to the study by changing from a hypothetical paper portfolio to an actual stock portfolio, based on straightforward trading rules. Again, this portfolio outperformed the S&P every year and generated a +75% return compared with -19% for the S&P 500. Yet again, these returns are not associated with higher risk (the beta is 0.76).

In addition to the study of Fornell et al (2006), Jacobson and Mizik (2009b) reexamine the financial market’s misprice of customer satisfaction (i.e. firms with high customer satisfaction scores are posited to earn positive future-period abnormal stock returns). Jacobson and Mizik (2009a) analyze the abnormal returns of the portfolios based on a four-factor model (Carhart, 1997). The authors did not find significant abnormal returns for all firms, except for computer/internet firms. Therefore, Jacobson and Mizik (2009a) conclude that the statistically evidence of financial market mispricing is limited to firms in the computer and internet sector. However, Fornell et al. (2009a) updated the actual portfolio results for the period 2004-2009, using the same four-factor CAPM model and did find significant abnormal returns. O’Sullivan, Hutchinson and O’Connell (2009) continued researching on the conflicting evidence on whether the stock market (mis) prices the value of customer satisfaction. Their results indicate higher cumulative returns for the top 20% compared with the lowest 80%. Moreover, although the alpha for the top 20% firms is positive, it is not significant. In line with Jacobson and Mizik (2009a), O’Sullivan, Hutchinson and O’Connell (2009) conclude that the trading strategies do not provide compelling evidence that the market mis-prices the value of customer satisfaction.

(12)

other three portfolio combination along with the S&P 500. These results differ from Jacobson and Mizik (2009a) who use the exact same methodology and research design but fail to find significant abnormal returns. Jacobson and Mizik (2009a) exemplify the conflicting results by the selection of firms from the ACSI database to match with the CRSP data. In addition, another explanation can be the use of the relative (to competitors) ACSI-scores or the absolute scores. Moreover, Tuli and Bharadwaj (2009) analyzed the effect of customer satisfaction on (downside) idiosyncratic and (downside) systematic risk. They use the Fama-French three factor model to obtain the measures of risk for each firm. These risk measures are used as dependent variable and are explained by the customer satisfaction scores and other confounding variables. Their results indicate that positive changes in customer satisfaction significantly leads to negative changes in the systematic, downside systematic, idiosyncratic, and downside idiosyncratic risk. Satisfied customers insulate a firm’s stock return form market movements and lowers the volatility of the stock’s return.

All together, scientists are still uncertain about the stock market’s (mis)pricing of customer satisfaction. Although the use of comparable or slightly different estimation models, several authors reveal completely different conclusions.

3. RESEARCH METHODOLOGY

In order to test the relationship between customer satisfaction and stock returns, portfolios are formed solely based on American Customer Satisfaction Index (ACSI) scores. Before elaborating on the portfolio composition rules, it is essential to clarify briefly the methodology behind the ACSI. The ACSI reports scores on a 0-100 scale for more than 200 companies. These companies can be classified in ten economic sectors and 44 industries. The ACSI-scores for each individual firm are updated once a year, however the publication date depends on the economic sector in which the company is subdivided. For example, the sector utilities is updated each year in May (1st quarter scores), E-business in August (2nd quarter scores), Manufacturing in November (3rd quarter scores) and retail in February (4th quarter scores).

(13)

In this paper however, several different trading rules and their impacts on the portfolio returns are considered. First, both absolute ACSI scores (from 0-100) and relative ACSI scores (compared to competitors in similar industry) are considered. The relative ACSI scores are computed as follows (O’Sullivan et al 2009, Fornell et al 2006):

The hypothesis is that the relative ACSI score based portfolio outperforms the absolute portfolios. Mainly because it automatically corrects for industry influences and from every industry the best performing firms are chosen. In addition, in order to have a diversified portfolio of reasonable size, Fornell et al. (2006) selected firms in the top 20% of the relative ACSI scores. However, this 20% split sounds arbitrary and therefore this paper also considers a 33% and 50% split. Moreover, the top 20% is not only compared to the lowest 80% (and consequently top 33% with lowest 66%), but also to percentiles.

When a firm’s ACSI score satisfies the selection criteria, the corresponding stock is included in the portfolio on the first of the next month after the publication date. For example, the utility 1st quarter scores are published on the second Tuesday of May, and consequently included in the portfolio on the first of June. After the inclusion, the stock was held for at least a year or more depending on whether it met the criteria again in the following year. All together it should be mentioned that the portfolio compositions are based on a hypothetical paper portfolio rather than a real-world portfolio.

(14)

model with two additional systematic risk factors. The first risk factor (SMB) takes the size of a firm into account while the other factor deals with the book-to-market ratio (HML). The SMB (small minus big) factor controls for the additional returns that can be achieved by investing in small stocks and the corresponding higher risk. Moreover, the HML (high minus low) controls for the historic excess returns of value stocks over growth stocks. All together, the specified three factor model is as follows:

Where, the betas indicate to what extent the portfolio returns are due to the corresponding risk factor (market factor, size factor and book-to-market factor). The estimated intercept (αi), referred to as Jensen’s Alpha, can be interpreted as a risk-adjusted return. This alpha is also called the abnormal return, that part of the return unrelated to the risk-factors. The sign and the significance of the alpha should be considered in order to determine whether the customer satisfaction based portfolios outperform the market without taking additional risks.

In the four factor model, a momentum factor is added to the three factor model. The momentum factor refers to momentum investing, buying stocks that are “winners” and selling stocks stat are “losers”. Besides, portfolios that perform well at time t-1 tend to continue to generate abnormal returns in year t. This is among others documented by Carhart (1997) and classified as a risk factor. However, not all academic publications confirm that the momentum factor can be considered as a risk factor. For example, Charoenrook and Conrad (2005) find that the evidence for momentum is not consistent with a risk-based explanation in the mean-variance setting. Nevertheless, the four factor model is specified as follows:

The additional beta added to the model refers to the additional risk considered by investing a larger proportion of the portfolio in winning stocks.

(15)

additional risk factor. Although this claim seems ungrounded, we refer to Fama and French’s inclusion of the size and book-to-market factors: “Although size and book-to-market equity seems like ad hoc variables for explaining average stock returns, we have reasons to expect that they proxy for common risk factors in returns”. Their argumentation for the factors is that they are related to economic fundamentals like earnings and profitability. The literature section of this paper clearly indicates that customer satisfaction is also strongly related to economic fundamentals like revenue, costs and profitability. As a consequence, it is reasonable to consider customer satisfaction as a risk factor. The question however remains whether the risk can be considered systematic or idiosyncratic risk. In order to be classified as systematic risk, some part of the stock’s return variation cannot be diversified away (Sharpe, 1964). When customer satisfaction fluctuations are due to economic fundamentals or business cycle, it can be considered systematic. The only major difference actually is that risk factors consider additional risk as an explanation for abnormal returns. However, Fornell et al (2006) state: “high returns, low risk”.

In order to examine customer satisfaction as a risk factor the methodology developed by Charoenrook and Conrad (2005) is used. Charoenrook and Conrad (2005) form portfolios based on the considered risk factor (here customer satisfaction). Moreover, they subtract the portfolio with the lowest score on the risk factor from the portfolio with the highest score. For example, the high minus low (HML) or small minus big (SMB) portfolio, and here the good minus bad satisfaction portfolio (GMB). For this GMB portfolio, the relation between the conditional mean and the conditional variance is examined. This relation is considered by using a (E-)(G)ARCH-in mean model. In the standard (E-)(G)ARCH models the assumption of constant variance of the error terms is already relaxed and allows the conditional variance to depend on previous own lags (GARCH). The “in-mean” variant specification however enters the conditional variance of asset returns into the conditional mean equation. By specifying an “in-mean” model, the model adapts for the reward investors should receive for taking additional risks, the risk factor.

The specified GARCH(1,1) in-mean model is:

(16)

and GARCH coefficients test for time-varying conditional variance of the GMB portfolio. After controlling for the lower part of the equation, the upper portion shows whether there is a relation between the estimated conditional variance and conditional mean ( ). If customer satisfaction should be considered as a risk factor, the significance and positive sign of this delta is important.

Moreover, Charoenrook and Conrad (2005) do not conclude anything about the best “in-mean” model. Therefore, besides the GARCH-in mean also an ARCH(2)-in-mean and EGARCH(1,1)in-mean are established. The ARCH(2) model includes two previous lags ( in the conditional variance equation and the EGARCH (1,1) extends the GARCH (1,1) model with an asymmetric effect ( in the conditional variance.

Specification ARCH(2): Specification EGARCH(1,1): Where X = 1 if <0 = 0 if otherwise

4. DATA AND DESCRIPTIVE STATISTICS

(17)

All together, the 307 company brands are reduced to 214 for which the ACSI data were available for at least once in the fourteen year. However, not all brands are measured every year. On average 120 companies were available for inclusion in the portfolio. Although only limited remarks are made by other authors about the total number of companies, we have substantial evidence to conclude that they are in line with this paper. First, O’Sullivan et al (2009) mention that between February 1997 and May 2003 the average number of firms was 93 (75 minimum and 125 maximum). In our study the average number in this time period was 92 (74 minimum and 122 maximum). Moreover, Fornell’s top 20% portfolio consists of 20-26 firms, in the similar period in our study this was 20-27. In appendix 1, the national ACSI scores are presented per quarter.

The remaining variables (the market index, SMB-, HML-, and momentum- factor, and the risk-free rate) are all provided by the Kenneth French website1. The descriptive statistics of these variables are provided in table 4.2. In addition, the descriptive statistics of the portfolio containing all stocks available for inclusion in a ACSI based portfolio are also provided. The descriptive statistics already indicate that the total portfolio’s monthly return (0,942%) is considerably higher than the market portfolio (0,615%). This is however in line with O’Sullivan where the total portfolio’s return was 0,92% and for the market 0,73% between 1996 and 2006. Graph 4.1 and 4.2 show the absolute and cumulative returns of both the market portfolio and the total portfolio (portfolio including all available stocks for selection).

1

http://mba.tuck.dartmouth.edu/pages/faculty/ken.french/

Table 4.1

Connecting the brand’s ACSI scores provided to the company’s stock returns

The table describes how the ACSI data is connected to the stock data. Sometimes, ACSI scores are measured for different divisions of the same parent, or for a division and the parent etc. In order to properly reproduce the paper it is essential how these brands are connected to their parent.

Action Example

0- No record in datastream Drop the record Mars is a privately owned company

1- Clear Match Take return Coca Cola brand and Coca Cola company

2- Multiple divisions (parent not covered)

Take average and assign to parent’s stock

Red Lobster and Olive Garden are owned by Darden Restaurant

3- Single division and parent is covered

Drop the single division ACSI score and assign parent’s ACSI score to stock

Kraft Food is subsidiary of Phillip Morris, both are measured on ACSI.

4- Single division but parent is not covered

Take single division as representative for parent.

Geico (measured) is subsidiary of Berkshire Hathaway (not measured)

5- Merger with overlapping ACSI observations

The overlapping records for the surviving entity are retained

HP acquired Compaq but both brands are separately measured by ACSI.

(18)

Figure 4.1

Absolute and cumulative returns of the Total portfolio and the market portfolio

Between the sixth month of 1996 and the twelfth month of 2009 the returns of the market portfolio and the returns of the total portfolio are analyzed. The total portfolio consists of all the stocks available for inclusion in the specified portfolios. In other words, it consists of all stocks for which both ACSI and return data was available in the corresponding time period.

Table 4.2 Descriptive statistics

The table shows the average monthly return and standard deviation of the market portfolio, small-minus-big factor, high-minus-low factor, momentum factor and the risk free rate presented by K. French’s website. Moreover, the descriptive statistics of the total portfolio (including all available stocks based on ACSI data.

Market Portfolio

SMB HML MOM Risk Free Total portfolio Average monthly return (%) 0.615 -0.057 0.552 0.451 0.274 0.942

Monthly Standard Deviation (%) 4.945 3.966 3.736 6.234 0.158 5.381

(19)

University of Groningen 15 Master Thesis Finance 5. RESULTS

Based on the customer satisfaction selection criteria, the total portfolio is divided in different portfolios which are all analyzed based on their returns and standard deviations (see table 5.1). In the upper part of the table, the portfolios are composed based on the absolute ACSI scores and in the lower part it are based on relative ACSI scores. The major difference between the two methods is the correction for industry effects. The total portfolio is divided in 2 (best 50% versus worst 50%, 3 (1st 33%, 2nd 33% and third 33%) and 5 (1st till 5th 20%) parts. It can be concluded that the portfolios based on best relative ACSI scores generate higher returns compared to the portfolios based on best absolute ACSI scores. Most notable is the 1,233% monthly average return of the portfolio with the 20% best performing Firms based on relative ACSI scores. Even more outstanding is the corresponding standard deviation which is considerably lower compared to the other 20%-percentiles. All together, the portfolios with the best ACSI performing firms seem to outperform the portfolios with the low ACSI scores.

However, average returns and standard deviation alone are not sufficient to conclude whether the portfolios really achieve abnormal returns during 1996 and 2009. Therefore, the three- and four- factor models are estimated. Based on the significance of the momentum factor and the higher adjusted R-squared for the four factor model, only the results of the four factor model are presented in table 5.2. The results of the three factor model are shows in appendix 2.

Table 5.1

Descriptive statistics of the absolute and relative ACSI based portfolio

The table shows the monthly average return and standard deviation of the different portfolios based on the ACSI data. Part A. shows the portfolios based on absolute ACSI scores while part B. shows the portfolios based on relative ACSI scores. The relative scores are corrected for the industry average.

A. Portfolios based on absolute ACSI scores All firms 100% Best 50% Worst 50% Best 33% Second 33% Worst 33% 1st 20% 2nd 20% 3rd 20% 4th 20% 5th 20% Average return (%) 0.942 0.974 0.907 0.972 0.882 0.989 1.043 1.057 0.863 0.841 0.909 Standard deviation (%) 5.381 5.024 6.145 4.717 5.656 6.791 4.906 5.137 6.060 5.961 7.575 N 163 163 163 163 163 163 163 163 163 163 163

(20)

Results Four-factor model

The table shows the results of the four factor model: for the different ACSI based portfolios. The constant indicates whether the portfolio gains an abnormal

return above the risk free rate. Moreover, the market beta, Small minus big (SMB), High minus low (HML) and momentum betas (MOM) indicate the extent till which the portfolio is exposed to the corresponding systematic risk. The betas and corresponding t-statistics are presented. ***,**, ***,**, and * denote the significance level at 1%, 5% and 10%.

A. Portfolios based on absolute ACSI scores Alpha Market Beta SMB Beta HML Beta MOM Beta Adj. R-squared F-value All firms (100%) 0.003 (1.53) 0.784 (17.95)*** 0.131 (2.49)** 0.421 (7.22)*** -0.288 (-8.59)*** 0.789 152.17*** Best 50% 0.004 (1.63) 0.709 (14.45)*** 0.112 (1.86)* 0.369 (5.6)*** -0.230 (-6.07)*** 0.701 92.59*** Worst 50% (0.003) (1.23) 0.874 (18.84)*** 0.148 (2.59)** 0.455 (7.35)*** -0.373 (-10.44)*** 0.817 181.46*** 1st 33% 0.004 (1.856)* 0.646 (13.49)*** 0.101 (1.721)* 0.323 (5.04)*** -0.223 (-6.05)*** 0.669 82.83*** 2nd 33% 0.002 (0.894) 0.795 (14.96)*** 0.124 (1.89)* 0.456 (6.43)*** -0.277 (-6.78)*** 0.716 103.32*** 3rd 33% 0.003 (1.282) 0.942 (16.626)*** 0.165 (2.376)** 0.460 (6.09)*** -0.402 (-9.23)*** 0.776 141.30*** 1st 20% 0.005 (2.09)** 0.671 (12.99)*** 0.100 (1.58) 0.274 (3.971)*** -0.218 (-5.49)*** 0.644 74.21*** 2nd 20% 0.004 (1.68)* 0.694 (13.12)*** 0.123 (1.90)* 0.459 (6.49)*** -0.232 (-5.69)*** 0.660 79.50*** 3rd 20% 0.002 (0.624) 0.817 (12.43)*** 0.105 (1.30) 0.447 (5.09)*** -0.260 (-5.13)*** 0.622 67.70*** 4th 20% 0.003 (1.130) 0.823 (16.82)*** 0.189 (3.15)*** 0.376 (5.78)*** -0.359 (-9.53)*** 0.784 148.28*** 5th 20% 0.002 (0.552) 0.995 (13.44)*** 0.163 (1.79)* 0.557 (5.64)*** -0.419 (-7.359)*** 0.692 92.05***

B. Portfolios based on relative ACSI scores Alpha Market Beta SMB

(21)

The results in table 5.2 clearly indicate that abnormal returns can be achieved by investing in firms that perform well in customer satisfaction relatively to their competitors. The most obvious evidence underlining this conclusion is the Jensen’s alpha (constant in four factor model) of 0,007 (t-statistic 3,30 p<0,01) concerning the portfolio with the best 20% firms based on relative ACSI scores. Moreover, investing in the best 50% or best 33% based on the relative scores can results in an abnormal return. Nevertheless, these are not statistically significant (p<0.05) but can be considered as an indication. Based on the absolute score portfolios, the evidence is less powerful. Investing in the top 20% will result in 0.005 abnormal return (t-value is 2.09, p<0,05).

Moreover, the market betas of the portfolios with high customer satisfaction results (both based on relative and absolute ACSI scores) are considerably lower compared to the other portfolios. Therefore, investing in these firms results in less systematic market risk exposure. In addition, the risk and additional return related to investing in value stocks rather than growth stocks is also to some extent lower in the higher satisfaction portfolios. The betas concerning the excess returns from investing in small firms compared with large firms are not consistent. For some portfolios it is significant while for others it is not, regardless of the customer satisfaction scores. The final factor, that distinguished the three factor from the four factor model is the momentum factor. For all models the momentum factor is significant and negative, indicating a negative relation with the momentum based portfolio. The adjusted R-squared of the models are all between 0.622 and 0.817 indicating that the included variables explain a substantial part of the variance in the dependent variable. Based on the Jensen’s Alpha’s derived from the highest relative ACSI score portfolios compared with the corresponding returns from the highest absolute ACSI score portfolios, it can be concluded that a portfolio composition based on the relative instead of abnormal scores will result in an abnormal return of 0.007. Moreover, narrowing the portfolio selection range (best 50% to best 33% to best 20%) results in higher abnormal returns.

(22)

Table 5.5

Descriptive statistics and analysis GARCH, ARCH and EGARCH models

This table presents the time series tests of the customer satisfaction-factor mimicking portfolio. Each month we construct 3 (33%) or 5 (20%) satisfaction based portfolios. The satisfaction-factor-mimicking portfolio is the return of the most satisfied portfolio minus the return of the most unsatisfied portfolio. First, the average monthly returns are presented in the upper part. In the lower part, the estimates of an ARCH (2,2), GARCH (1,1) and EGARCH (1,1) in mean models are presented. The coefficient and corresponding z-statistics are shown. ***,**, and * denote the significance level at 1%, 5% and 10% respectively.

A. Models based on highest 33% - lowest 33% Mean 0.0019

T-statistic 0.919

N 163

GARCH (1.1) ARCH (2.2) EGARCH (1.1) I.MEAN EQUATION Intercept 0.296 (0.156) 0.042 (1.733)* 0.370 (0.397) Delta -11.576 (-0.154) -1.599 (1.608) -14.486 (-0.398)

II. CONDITIONAL VARIANCE QUATION Intercept 0.000 (2.510)*** 0.000 (7.650)*** -7.406 (-3.202)*** ARCH 1 0.016 (0.148) 0.088 (1.124) 0.034 (0.375) ARCH 2 0.142 (1.561) GARCH 1 -0.031 -0.004 (-0.015) (-0.079) Theta -0.003 (-0.231) AIC -4.446 -4.454 -4.436

B. Portfolios based on highest 20% - lowest 20% Mean 0.0028

T-statistic 1.011

N 163

GARCH (1.1) ARCH (2.2) EGARCH (1.1) I.MEAN EQUATION Intercept 0.082 (1.066) 0.068 (1.238) 0.063 (1.327) Delta -2.272 (-1.011) -1.874 (-1.159) -1.736 (-1.249)

(23)

The mean of the satisfaction-factor-mimicking portfolio is 0.0019 for the 33%- and 0.0028 for the 20% portfolio. The corresponding t-statistics indicate that the average of the portfolios do not significantly differ from zero.

The results of the (E)(G)ARCH-in-mean models are not outstanding. First, none of the presented models provide evidence that the conditional variance should vary through time. Neither the ARCH nor GARCH effects statistically differ from zero and consequently the current error term does not depend on the previous error term. Moreover, neither are the EGARCH effects statistically significant and as a result asymmetries in the factor-mimicking-portfolios’ volatility are not proven. The only significant term in the conditional variance equation (II-part) is the intercept. However, rounded to tree decimals it is still approximately zero. All together, there is insufficient evidence to conclude that the variance varies over time.

The second part of the conclusions concerns the mean equation, the I-part of the results. A factor that can be considered as a risk factor has a positive and significant delta in the mean equation. The significance of the delta indicates that a relation exists between the conditional variance and the conditional mean while a positive sign points out that higher risk are rewarded with higher returns. Although none of the deltas in the mean equation significantly differ from zero, it is notable that all the estimated deltas are negative. The corresponding significance levels are 0.312 (GARCH), 0.246 (ARCH), and 0.211 (EGARCH). Far from being statistically significant but still obvious because usually these signs are supposed to be positive. All together, it can be concluded that investing in stocks from firms with high customer satisfaction levels does not lead to additional risks.

6. CONCLUSION AND DISCUSSION

In this paper, the relation between customer satisfaction and stock returns is examined. Several stock portfolios are composed based on customer satisfaction (ACSI) scores. With the use of the four factor model proposed by Carhart (1997) the abnormal returns of the different portfolios are examined. Moreover, with the use of a GARCH-in-mean model it is examined whether customer satisfaction can be classified as a risk factor.

(24)

the market return, and the absolute returns are positive, the Jensen’s Alpha is not significant. The different outcomes of the studies can, among others, be caused by the different time-intervals. For example, in both Fornell (2009a) and this paper the returns between 1996 and 2009/2010 are analyzed, while O’Sullivan’s portfolio was ceased in 2006. Moreover, another clarification might be the brand-firms classification and assumptions regarding this relation. In this paper, the six basic rules of Ittner, Larcker and Taylor (2009) are followed whereas other studies do not mention anything about their classification. Significant deviations in the results might be derived by different classification rules. In addition, one of the rules concerns the availability of multiple division’s ACSI scores but the absence of the parent’s score. For example, msnbc.com and msn.com are available but Microsoft is not. In this paper, based on the average of msnbc.com and msn.com the Microsoft stock is bought or sold. It is however questionable whether the satisfaction levels of these brands are representative of Microsoft’s customer satisfaction scores.

Furthermore, investing in firms that perform very well on customer satisfaction in comparison with their competitors will results in higher returns compared with firms that perform well on an absolute level. Based on relative scores, the portfolios are corrected for industry influences and consequently the best performing firm per industry is considered for the portfolio. Fornell et al. (2006) already mentioned that the extent to which buyers can reward and punish firms depends on availability of substitutes. For the energy industry, the satisfaction levels are considerably lower compared with the personal care and cleaning industry. However, customers cannot punish all energy firms by consuming shampoo rather than electricity. The customer will choose that energy firm that maximizes his utility compared with the competitors. Moreover, dissatisfied buyers will not remain a customer unless there is nowhere else to go or it is too expensive to get there.

Regarding the risk of the portfolios, it can be concluded that the portfolios with the highest satisfaction scores face less systematic market risk compared with the other portfolios. This conclusion is supported by Gruca and Rego (2005) who mention the diminishing influence of external shocks on business performance for firms with high satisfaction levels. Therefore, Fornell (2009a) analyzed his betas during bear markets and bull markets and found that the corresponding betas were significantly lower during down markets. In other words, firms with high satisfaction levels face less market risk during down markets and more during up markets.

(25)

conditional variance and the conditional mean of this portfolio is examined with the use of a (E)(G)ARCH-in-mean model. None of the hypothesis regarding the conditional mean and variance are rejected and consequently customer satisfaction cannot be considered a risk factor. All though far from significant (t-value between -1 and -1.25), it is notable that all deltas in the mean equation are negative. Indicating that higher risk in this case leads to lower returns or lower risk leads to higher returns. Therefore, Fornell’s (2006) statement about customer satisfaction: “higher returns, lower risks” is neither proven nor completely rejected.

One of the marginal notes regarding the study concerns the availability or absence of some ACSI data. For example, the gasoline industry is measured and presented until 2000 and thereafter it is only available on industry aggregated level. Therefore, none of the gasoline firms is included in any portfolio after 2000 and this will certainly cause some deviations in the results. Moreover, the scores are updated once a year, ignoring fluctuations during the year.

For further research it would be very interesting to consider other customer satisfaction scores from different methodologies in different countries. For example, the customer performance index in the Netherlands from the Customer Insight Center. Moreover, a model with the systematic market beta as a time-varying parameter would provide insights in the beta surfing behavior of the customer satisfaction portfolio. Besides, it would be interesting to examine whether a maximum point exists regarding the effects of customer satisfaction. For example, is it worthwhile improving absolute/relative satisfaction levels for the best performing firms or is this discarded money?

(26)

University of Groningen 22 Master Thesis Finance 7. REFERENCES

Aksoy, L., Cooil, B., Groening, C., Keiningham, T.L., Yalcin, A. (2008)."&D2&". Journal of Marketing, Vol. 72, 105-122

Anderson, E.W. (1998)."Customer Satisfaction and Word of Mouth". Journal of Service Research, 1 (1), 1-14

Anderson, E.W., Fornell, C., Lehman, D.R. (1994)."Customer Satisfaction, Market Share, and Profitability: Findings from Sweden". Journal of Marketing, Vol. 58 (July), 53-66

Anderson, E.W., Fornell, C., Mazvancheryl S.K. (2004)."Customer Satisfaction and Shareholder Value". Journal of Marketing, Vol. 68 (Oktober), 172-185

Anderson, E.W., Fornell, C., Rust, R.T. (1997)."Customer Satisfaction, productivity, and profitability: differences between goods and services". Marketing Science, Vol. 16 (2), 53-66

Anderson, E.W., Mansi, S.A. (2009)."Does Customer Satisfaction Matter to Investors? Findings from the Bond Market". Journal of Marketing Research, Vol. 46 (October), 703-714

Blackshaw, P. Nazzaro, M. (2006)."Consumer Generated Media; 101 Word-of-Mouth in the Age of Web-Fortified Consumer". Nielzen BuzzMetrics, 2d edition (spring),

Bolton R.N., Lemon, K.N. (1999)."A Dynamic Model of Customer's Relationship with a Continuous Service Provider: The Role of Satisfaction". Marketing Science, Vol. 17 (1)

Bolton, R.N. (1998)."A Dynamic Model of the Duration of the Customer's Relationship With a Continuous Service Provider: “The Role of Satisfaction". Marketing Science, Vol. 17 (issue 1), 45-65

Bolton, R.N., Lemon, K.N., Verhoef, P.C. (2004)."The Theoretical Underpinnings of Customer Asset Management: A Framework and Propositions for Future Research". Journal of Academy of Marketing Science, Volume 32 (no. 3), 271-292

Brealey, R.A., Myers, S.C., Marcus, A.J. (2004)."Fundamentals of Corporate Finance". United States, McGraw-Hill Irwin,

Carhart, M.M., (1997)."On Persistence in Mutual Fund Performance". Journal of Finance, Vol. 52 (no. 1), 57-82

(27)

Dougherty, D. (1992)."Interpretive Barriers to Succesfull Product Innovations in Large Firms". Organization Science, (3), 179-202

Doyle P. (2000)."Value-Based Marketing". Chisester, Uk, John Wiley & Sons

Fama, E.F., French, K.R. (1993)."Common Risk Factors in the Return on Stocks and Bonds". Journal of Financial Economics, Volume 33, 3-56

Fama, E.F., French, K.R. (1992)."The Cross-Section of Expected Stock Returns.". Journal of finance, Vol. 47 (no. 1), 427-465

Fornell, F., Mithas, S., Morgeson III, F.V. (2009a)."The Economic and Statistical Significance of Stock Returns on Customer Satisfaction. Marketing Science, Vol. 28 (no. 5), 820-825

Fornell, F., Mithas, S., Morgeson III, F.V. (2009b)."The Statistical Significance of Portfolio Returns". International Journal of Market Research, Vol. 26, 162-163

Fornell, C. (2007)."The Satisfied Customer: Winners and Losers in the Battle for Buyer Preference". Palgrave MacMillan, United States, St. Martin's Press LLC

Fornell, C. Mithas, S., Morgeson III, F.V., Krishnan, M.S. (2006)."Customer Satisfaction and Stock Prices: High Return, Low Risk. Journal of Marketing, Vol. 70 (January), 3-14

Grande, C. (2006)."Business Keep Tight Grip on Advertising amid Economy Fears". Financial Times, July 17,

Grinblatt, M., Titman, S. (2002)."Financial Markets and Corporate Strategy". United States, McGraw-Hill Irwin,

Gruce, T.S., Rego, L.L. (2005)."Customer Satisfaction, Cash Flow, and Shareholder Value". Journal of Marketing, Vol. 69 (July), 115-130

Harrison-Walken, L.J., Perdue, G. (2007)."The Role of Marketing in the Valuation of a Firm: Exploring the Underlying Mechanism". Journal of Strategic Marketing, Volume 15 issue 5, 377-386

Hart, C.W. (2007)."Beating the Market with Customer Satisfaction. Harvard Business Review, March, 30-32

(28)

Homburg, C., Koschate, N., Hoyer, W.D. (2005)."Do Satisfied Customers Really Pay More? A Study of the Relationship Between Customer Satisfaction and Willingness to Pay". Journal of Marketing, Vol. 69 (April), 84-97

Ittner, C.D., Larcker, D.F. (1998)."Are Nonfinancial Measures Leading Indicators of Financial Performance? An Analysis of Customer Satisfaction". Journal of Accounting Research, Vol. 36, 1-35

Ittner, C.D., Larcker, D.F., Taylor, D. (2009)."The Stock Pricing of Customer Satisfaction. Marketing Science, Vol. 28 (no. 5), 826-835

Jacobson, R., Mizik, N. (2009a)."Customer Satisfaction-Based Mispricing: Issues and Misconceptions". Marketing Science, Vol. 28 (no. 5), 836-845

Jacobson, R., Mizik, N. (2009b)."The Financial Markets and Customer Satisfaction: Reexamining Possible Financial Market Mispricing of Customer Satisfaction. Marketing Science, Vol. 28 (no. 5), 810-819

Kotler, K., Armstrong, G., Saunders, J., Wong, V. (2002)."Principles of Marketing". London, Prenctice Hall

Kumar, V., Reinartz, W.J. (2006)."Customer Relationship Management; A Database Approach". United States, John Wiley & Sons Inc.,

Kumar, V., Shah, D. (2009)."Expanding the Role of Marketing: From Customer Equity to Market Capitalization. Journal of Marketing, Vol. 73 (November 2009), 119-136

Lam, S.Y., Shankar, V., Erramili, M.K., Murthy, B. (2004)."Customer Value, Satisfaction, Loyalty, Switching Costs: an Illustration from Business-to-Business service context. Journal of Academy of Marketing Science, Vol. 32, 293-311

Luo, X., Homburg, C. (2007)."Neglected Outcomes of Customer Satisfaction. Journal of Marketing, Vol. 71 , 133-149

Luo, X., Homburg, C. (2008)."Satisfaction, Complaint, and the Stock Value Gap. Journal of Marketing, Vol. 72 (July), 29-43

Moorman, C., Rust, R.T. (1999)."The Role of Marketing". Journal of Marketing, (63), 180-197

(29)

Narayanan, S., Desiraju, R., Chintagunta, P.K. (2004)."Return on Investment Implications for Pharmaceuticals promotional expenditures: the Role of Marketing Mix Interactions". Journal of Marketing, Volume 68 Oktober, 90-105

Nath, P., Mahajan, V. (2008)."Chief Marketing Officers: A Study of Their Presence in Firms' Top Management Teams. Journal of Marketing, Vol. 72 issue 1, 65-81

Nauman, E., Hoisington, S.H. (2001)."Customer Centered Six Sigma: Linking Customers, Process Improvement, and Financial Results. Milwaukee, ASQ Quality Press

O'Sullivan, D., Hutchinson, M.C., O'Connel, V. (2009)."Empirical Evidence of the Stock Market's (mis)pricing of Customer Satisfaction". International Journal of Market Research, Vol. 26, 154-161

Pauwels, K., Silva-Risso, J., Srinivasan, S., Hanssens, D.M. (2004)."New Products, Sales Promotions, and Firm Value: the Case of the Automobile Industry". Journal of Marketing, Volume 68 Oktober, 142-156

Peppers, D. Rogers, M. (2005)."Return on Customer: Creating Maximum Value from Your Scarcest Resource". New York, Currency Doubleday,

Ping jr., R.A. (1993)."The Effects of Satisfaction and Structural Constraints on Retailing Exiting, Voice, Loyalty, Opportunism, and Neglected". Journal of Retailing, Vol. 69 (no. 3), 320-352

Pugh, D.S. (2001)."Services with a Smile, Emotional Contagion in the Service Encounter". Academy of Management Journal, Vol. 44 (no. 5), 1018-1027

Rechheld, F.F. (2006)."The Ultimate Question: Should you recommend our company". United States, Business Bibliotheek,

Reichheld, F. Teal, T. (1996)."The Loyalty Effect: The Hidden Force Behind Growth, Profits, and Lasting Value". Boston, Harvard Business School Press,

Reichheld, F., Sasser, W. (1990)."Zero Defections: Quality Comes to Service. Harvard Business Review, Vol. 68 (5), 105-111

Rust, R.T., Lemon, K.N. and Zeithaml, V.A. (2004)."Return on Marketing: Using Customer Equity to Focus Marketing Strategy". Journal of Marketing, Volume 68 January, 109-127

(30)

Schlesinger L., Hesket, J. (1991)."The Service-Driven Service Company. Harvard Business Review, Vol. 69 (5), 71-81

Sharpe, W.F. (1964)."Capital Asset Prices - A theory of Market Equilibrium Under Conditions of Risk. Journal of Finance, Vol. 19, 425-442

Simms, J. (2001)."Do We Need More Marketing CEOs?". Marketing, April, 24-25

Srinivasan, S., Hanssens, D.M. (2009)."Marketing and Firm Value: Metrics, Methods, Findings, and Future Directions. Journal of Marketing Research, Vol. 46 (June), 293-312

Srivastava, R.K., Shervani, T.A., Fahey L. (1998)."Market-Based Assets and Shareholders Value: A Framework for Analysis". Journal of Marketing, Volume 62 (January),

Steenkamp, J.B.E.M., Nijs, V.R., Hanssens, D.M, DeKimpe, M.G. (2005)."Competitive Reactions to Advertising and Promotion Attacks. Marketing Science, Vol. 24 (no. 1), 35-54

Szymanski D.M., Henard, D.H. (2001)."Customer Satisfaction: A Meta-Analysis of the Empirical Evidence". Journal of the Academy of Marketing Science, Vol. 29 (no. 1), 16-35

Trusov, M., Bucklin, R.E., Pauwels, K. (2009)."Effects of Word-of0Mouth Versus Traditional Marketing: Findings from an Internet Social Networking Site". Journal of Marketing, Vol. 73 (nr. 5), 90-102

Tuli, K.R., Bharadwaj, S.G. (2009)."Customer Satisfaction and Stock Return Risk. Journal of Marketing, Vol. 73 (November), 184-197

Verhoef, P.C., Leeflang, P.S.H. (2008)."Getting Marketing Back in the Boardroom: Understanding Drivers of Marketing Influence. MSI, Working Paper,

(31)

University of Groningen 27 Master Thesis Finance APPENDIX 1: NATIONAL ACSI SCORES PER QUARTER

(32)

University of Groningen 28 Master Thesis Finance APPENDIX 2: RESULTS THREE FACTOR SCORES PERCENTILES

The table shows the results of the four factor model:

for the different ACSI based portfolios. The constant indicates whether the portfolio gains an abnormal return above the risk free rate. Moreover, the market beta, Small minus big (SMB)and High minus low (HML) indicate the extent till which the portfolio is exposed to the corresponding systematic risk. The betas and corresponding t-statistics are presented. ***,**, ***,**, and * denote the significance level at 1%, 5% and 10%.

A. Portfolios based on absolute ACSI scores Alpha Market Beta SMB Beta HML Beta Adj. R-squared F-value All firms (100%) 0.001 (0.363) 0.912*** (18.443) 0.183*** (2.859) 0.509*** (7.360) 0.693 122.717*** Best 50% 0.002 (0.769) 0.812 (15.902)*** 0.154 (2.323)*** 0.439 (6.152)*** 0.624 90.668*** Worst 50% 0.000 (-0.089) 1.040 (18.422)*** 0.215 (2.941)*** 0.570 (7.209)*** 0.692 122.489*** 1st 33% 0.002 (0.982) 0.746 (14.980)*** 0.142 (2.194)*** 0.391 (5.614)*** 0.595 80.291*** 2nd 33% 0.000 (0.019) 0.919 (16.251)*** 0.174 (2.375)*** 0.541 (6.840)*** 0.636 95.447*** 3rd 33% 0.000 (0.077) 1.121 (17.040) 0.238 (2.792) 0.584 (6.341) 0.657 104.648*** 1st 20% 0.003 (1.287) 0.768 (14.555)*** 0.140 (2.045)*** 0.341 (4.615)*** 0.579 75.149*** 2nd 20% 0.002 (0.873) 0.797 (14.664)*** 0.166 (2.348)*** 0.530 (6.963)*** 0.592 79.479*** 3rd 20% 0.000 (-0.034) 0.933 (14.030)*** 0.152 (1.767)* 0.527 (5.662)*** 0.562 70.295*** 4th 20% 0.000 (-0.078) 0.983 (17.093)*** 0.254 (3.408)*** 0.486 (6.044)*** 0.662 106.982*** 5th 20% -0.001 (-0.345) 1.182 (14.717)*** 0.239 (2.293)*** 0.686 (6.104)*** 0.589 78.449***

B. Portfolios based on relative ACSI scores Alpha Market Beta SMB

(33)